ftrace.c 144.5 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 1189 1190 1191 1192 1193 1194 1195 1196
static __always_inline unsigned long
ftrace_hash_key(struct ftrace_hash *hash, unsigned long ip)
{
	if (hash->size_bits > 0)
		return hash_long(ip, hash->size_bits);

	return 0;
}

1197 1198 1199
/* Only use this function if ftrace_hash_empty() has already been tested */
static __always_inline struct ftrace_func_entry *
__ftrace_lookup_ip(struct ftrace_hash *hash, unsigned long ip)
1200 1201 1202 1203 1204
{
	unsigned long key;
	struct ftrace_func_entry *entry;
	struct hlist_head *hhd;

1205
	key = ftrace_hash_key(hash, ip);
1206 1207
	hhd = &hash->buckets[key];

1208
	hlist_for_each_entry_rcu_notrace(entry, hhd, hlist) {
1209 1210 1211 1212 1213 1214
		if (entry->ip == ip)
			return entry;
	}
	return NULL;
}

1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233
/**
 * ftrace_lookup_ip - Test to see if an ip exists in an ftrace_hash
 * @hash: The hash to look at
 * @ip: The instruction pointer to test
 *
 * Search a given @hash to see if a given instruction pointer (@ip)
 * exists in it.
 *
 * Returns the entry that holds the @ip if found. NULL otherwise.
 */
struct ftrace_func_entry *
ftrace_lookup_ip(struct ftrace_hash *hash, unsigned long ip)
{
	if (ftrace_hash_empty(hash))
		return NULL;

	return __ftrace_lookup_ip(hash, ip);
}

1234 1235
static void __add_hash_entry(struct ftrace_hash *hash,
			     struct ftrace_func_entry *entry)
1236 1237 1238 1239
{
	struct hlist_head *hhd;
	unsigned long key;

1240
	key = ftrace_hash_key(hash, entry->ip);
1241 1242 1243
	hhd = &hash->buckets[key];
	hlist_add_head(&entry->hlist, hhd);
	hash->count++;
1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255
}

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);
1256 1257 1258 1259 1260

	return 0;
}

static void
1261
free_hash_entry(struct ftrace_hash *hash,
1262 1263 1264 1265 1266 1267 1268
		  struct ftrace_func_entry *entry)
{
	hlist_del(&entry->hlist);
	kfree(entry);
	hash->count--;
}

1269 1270 1271 1272 1273 1274 1275 1276
static void
remove_hash_entry(struct ftrace_hash *hash,
		  struct ftrace_func_entry *entry)
{
	hlist_del(&entry->hlist);
	hash->count--;
}

1277 1278 1279
static void ftrace_hash_clear(struct ftrace_hash *hash)
{
	struct hlist_head *hhd;
1280
	struct hlist_node *tn;
1281 1282 1283 1284
	struct ftrace_func_entry *entry;
	int size = 1 << hash->size_bits;
	int i;

1285 1286 1287
	if (!hash->count)
		return;

1288 1289
	for (i = 0; i < size; i++) {
		hhd = &hash->buckets[i];
1290
		hlist_for_each_entry_safe(entry, tn, hhd, hlist)
1291
			free_hash_entry(hash, entry);
1292 1293 1294 1295
	}
	FTRACE_WARN_ON(hash->count);
}

1296 1297 1298 1299 1300 1301 1302 1303 1304
static void free_ftrace_hash(struct ftrace_hash *hash)
{
	if (!hash || hash == EMPTY_HASH)
		return;
	ftrace_hash_clear(hash);
	kfree(hash->buckets);
	kfree(hash);
}

1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319
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);
}

1320 1321
void ftrace_free_filter(struct ftrace_ops *ops)
{
1322
	ftrace_ops_init(ops);
1323 1324
	free_ftrace_hash(ops->func_hash->filter_hash);
	free_ftrace_hash(ops->func_hash->notrace_hash);
1325 1326
}

1327 1328 1329 1330 1331 1332 1333 1334 1335 1336
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;
1337
	hash->buckets = kcalloc(size, sizeof(*hash->buckets), GFP_KERNEL);
1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362

	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? */
1363
	if (ftrace_hash_empty(hash))
1364 1365 1366 1367
		return new_hash;

	size = 1 << hash->size_bits;
	for (i = 0; i < size; i++) {
1368
		hlist_for_each_entry(entry, &hash->buckets[i], hlist) {
1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383
			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;
}

1384
static void
1385
ftrace_hash_rec_disable_modify(struct ftrace_ops *ops, int filter_hash);
1386
static void
1387
ftrace_hash_rec_enable_modify(struct ftrace_ops *ops, int filter_hash);
1388

1389 1390 1391
static int ftrace_hash_ipmodify_update(struct ftrace_ops *ops,
				       struct ftrace_hash *new_hash);

1392 1393
static struct ftrace_hash *
__ftrace_hash_move(struct ftrace_hash *src)
1394 1395
{
	struct ftrace_func_entry *entry;
1396
	struct hlist_node *tn;
1397
	struct hlist_head *hhd;
1398
	struct ftrace_hash *new_hash;
1399 1400 1401 1402 1403
	int size = src->count;
	int bits = 0;
	int i;

	/*
1404
	 * If the new source is empty, just return the empty_hash.
1405
	 */
1406 1407
	if (!src->count)
		return EMPTY_HASH;
1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418

	/*
	 * 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;

1419 1420
	new_hash = alloc_ftrace_hash(bits);
	if (!new_hash)
1421
		return NULL;
1422 1423 1424 1425

	size = 1 << src->size_bits;
	for (i = 0; i < size; i++) {
		hhd = &src->buckets[i];
1426
		hlist_for_each_entry_safe(entry, tn, hhd, hlist) {
1427
			remove_hash_entry(src, entry);
1428
			__add_hash_entry(new_hash, entry);
1429 1430 1431
		}
	}

1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449
	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;

1450 1451 1452 1453 1454 1455 1456 1457 1458 1459
	/* 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;
		}
	}

1460 1461 1462 1463
	/*
	 * Remove the current set, update the hash and add
	 * them back.
	 */
1464
	ftrace_hash_rec_disable_modify(ops, enable);
1465

1466 1467
	rcu_assign_pointer(*dst, new_hash);

1468
	ftrace_hash_rec_enable_modify(ops, enable);
1469

1470
	return 0;
1471 1472
}

1473 1474 1475 1476 1477 1478 1479 1480 1481 1482
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) ||
1483
		__ftrace_lookup_ip(hash->filter_hash, ip)) &&
1484
		(ftrace_hash_empty(hash->notrace_hash) ||
1485
		 !__ftrace_lookup_ip(hash->notrace_hash, ip));
1486 1487
}

1488 1489 1490 1491 1492 1493 1494 1495
/*
 * 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.
1496 1497 1498
 *
 * This needs to be called with preemption disabled as
 * the hashes are freed with call_rcu_sched().
1499 1500
 */
static int
1501
ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip, void *regs)
1502
{
1503
	struct ftrace_ops_hash hash;
1504 1505
	int ret;

1506 1507 1508 1509 1510 1511 1512 1513 1514 1515
#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

1516 1517
	hash.filter_hash = rcu_dereference_raw_notrace(ops->func_hash->filter_hash);
	hash.notrace_hash = rcu_dereference_raw_notrace(ops->func_hash->notrace_hash);
1518

1519
	if (hash_contains_ip(ip, &hash))
1520 1521 1522 1523 1524 1525 1526
		ret = 1;
	else
		ret = 0;

	return ret;
}

1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540
/*
 * 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()		\
		}				\
	}

1541 1542 1543

static int ftrace_cmp_recs(const void *a, const void *b)
{
1544 1545
	const struct dyn_ftrace *key = a;
	const struct dyn_ftrace *rec = b;
1546

1547
	if (key->flags < rec->ip)
1548
		return -1;
1549 1550
	if (key->ip >= rec->ip + MCOUNT_INSN_SIZE)
		return 1;
1551 1552 1553
	return 0;
}

1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566
/**
 * 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)
1567 1568 1569
{
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
1570
	struct dyn_ftrace key;
1571

1572 1573
	key.ip = start;
	key.flags = end;	/* overload flags, as it is unsigned long */
1574 1575

	for (pg = ftrace_pages_start; pg; pg = pg->next) {
1576 1577
		if (end < pg->records[0].ip ||
		    start >= (pg->records[pg->index - 1].ip + MCOUNT_INSN_SIZE))
1578
			continue;
1579 1580 1581 1582
		rec = bsearch(&key, pg->records, pg->index,
			      sizeof(struct dyn_ftrace),
			      ftrace_cmp_recs);
		if (rec)
1583
			return rec->ip;
1584
	}
1585 1586 1587 1588

	return 0;
}

1589 1590 1591 1592
/**
 * ftrace_location - return true if the ip giving is a traced location
 * @ip: the instruction pointer to check
 *
1593
 * Returns rec->ip if @ip given is a pointer to a ftrace location.
1594 1595 1596 1597
 * 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.
 */
1598
unsigned long ftrace_location(unsigned long ip)
1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612
{
	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.
 */
1613
int ftrace_text_reserved(const void *start, const void *end)
1614
{
1615 1616 1617 1618 1619 1620
	unsigned long ret;

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

	return (int)!!ret;
1621 1622
}

1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642
/* 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;
}

1643
static bool __ftrace_hash_rec_update(struct ftrace_ops *ops,
1644 1645 1646 1647 1648 1649 1650
				     int filter_hash,
				     bool inc)
{
	struct ftrace_hash *hash;
	struct ftrace_hash *other_hash;
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
1651
	bool update = false;
1652 1653 1654 1655 1656
	int count = 0;
	int all = 0;

	/* Only update if the ops has been registered */
	if (!(ops->flags & FTRACE_OPS_FL_ENABLED))
1657
		return false;
1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670

	/*
	 * 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) {
1671 1672
		hash = ops->func_hash->filter_hash;
		other_hash = ops->func_hash->notrace_hash;
1673
		if (ftrace_hash_empty(hash))
1674 1675 1676
			all = 1;
	} else {
		inc = !inc;
1677 1678
		hash = ops->func_hash->notrace_hash;
		other_hash = ops->func_hash->filter_hash;
1679 1680 1681 1682
		/*
		 * If the notrace hash has no items,
		 * then there's nothing to do.
		 */
1683
		if (ftrace_hash_empty(hash))
1684
			return false;
1685 1686 1687 1688 1689 1690 1691
	}

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

1692 1693 1694
		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

1695 1696 1697 1698 1699
		if (all) {
			/*
			 * Only the filter_hash affects all records.
			 * Update if the record is not in the notrace hash.
			 */
1700
			if (!other_hash || !ftrace_lookup_ip(other_hash, rec->ip))
1701 1702
				match = 1;
		} else {
1703 1704
			in_hash = !!ftrace_lookup_ip(hash, rec->ip);
			in_other_hash = !!ftrace_lookup_ip(other_hash, rec->ip);
1705 1706

			/*
1707 1708
			 * If filter_hash is set, we want to match all functions
			 * that are in the hash but not in the other hash.
1709
			 *
1710 1711 1712 1713 1714
			 * 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.
1715 1716 1717 1718
			 */
			if (filter_hash && in_hash && !in_other_hash)
				match = 1;
			else if (!filter_hash && in_hash &&
1719
				 (in_other_hash || ftrace_hash_empty(other_hash)))
1720 1721 1722 1723 1724 1725 1726
				match = 1;
		}
		if (!match)
			continue;

		if (inc) {
			rec->flags++;
1727
			if (FTRACE_WARN_ON(ftrace_rec_count(rec) == FTRACE_REF_MAX))
1728
				return false;
1729 1730 1731 1732 1733 1734

			/*
			 * 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.
			 */
1735
			if (ftrace_rec_count(rec) == 1 && ops->trampoline)
1736
				rec->flags |= FTRACE_FL_TRAMP;
1737
			else
1738 1739 1740
				/*
				 * If we are adding another function callback
				 * to this function, and the previous had a
1741 1742
				 * custom trampoline in use, then we need to go
				 * back to the default trampoline.
1743
				 */
1744
				rec->flags &= ~FTRACE_FL_TRAMP;
1745

1746 1747 1748 1749 1750 1751
			/*
			 * 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;
1752
		} else {
1753
			if (FTRACE_WARN_ON(ftrace_rec_count(rec) == 0))
1754
				return false;
1755
			rec->flags--;
1756

1757 1758 1759 1760 1761 1762
			/*
			 * 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.
			 */
1763
			if (ftrace_rec_count(rec) > 0 &&
1764 1765 1766 1767 1768
			    rec->flags & FTRACE_FL_REGS &&
			    ops->flags & FTRACE_OPS_FL_SAVE_REGS) {
				if (!test_rec_ops_needs_regs(rec))
					rec->flags &= ~FTRACE_FL_REGS;
			}
1769

1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780
			/*
			 * 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;

1781 1782 1783 1784
			/*
			 * flags will be cleared in ftrace_check_record()
			 * if rec count is zero.
			 */
1785 1786
		}
		count++;
1787 1788 1789 1790

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

1791 1792
		/* Shortcut, if we handled all records, we are done. */
		if (!all && count == hash->count)
1793
			return update;
1794
	} while_for_each_ftrace_rec();
1795 1796

	return update;
1797 1798
}

1799
static bool ftrace_hash_rec_disable(struct ftrace_ops *ops,
1800 1801
				    int filter_hash)
{
1802
	return __ftrace_hash_rec_update(ops, filter_hash, 0);
1803 1804
}

1805
static bool ftrace_hash_rec_enable(struct ftrace_ops *ops,
1806 1807
				   int filter_hash)
{
1808
	return __ftrace_hash_rec_update(ops, filter_hash, 1);
1809 1810
}

1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845
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);
}

1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878
/*
 * 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) {
1879 1880 1881 1882

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904
		/* 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) {
1905 1906 1907 1908

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

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 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961
		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);
}

1962
static void print_ip_ins(const char *fmt, const unsigned char *p)
1963 1964 1965 1966 1967 1968 1969 1970 1971
{
	int i;

	printk(KERN_CONT "%s", fmt);

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

1972 1973
static struct ftrace_ops *
ftrace_find_tramp_ops_any(struct dyn_ftrace *rec);
1974 1975
static struct ftrace_ops *
ftrace_find_tramp_ops_next(struct dyn_ftrace *rec, struct ftrace_ops *ops);
1976

1977
enum ftrace_bug_type ftrace_bug_type;
1978
const void *ftrace_expected;
1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999

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;
	}
}

2000 2001 2002
/**
 * ftrace_bug - report and shutdown function tracer
 * @failed: The failed type (EFAULT, EINVAL, EPERM)
2003
 * @rec: The record that failed
2004 2005 2006 2007 2008 2009 2010 2011
 *
 * 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
 */
2012
void ftrace_bug(int failed, struct dyn_ftrace *rec)
2013
{
2014 2015
	unsigned long ip = rec ? rec->ip : 0;

2016 2017 2018 2019 2020 2021 2022 2023 2024 2025
	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);
2026
		print_ip_ins(" actual:   ", (unsigned char *)ip);
2027
		pr_cont("\n");
2028 2029 2030 2031
		if (ftrace_expected) {
			print_ip_ins(" expected: ", ftrace_expected);
			pr_cont("\n");
		}
2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042
		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);
	}
2043
	print_bug_type();
2044 2045 2046 2047 2048 2049 2050 2051
	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);
2052 2053 2054 2055 2056 2057 2058 2059
			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
2060 2061 2062 2063
				pr_cont("\ttramp: ERROR!");

		}
		ip = ftrace_get_addr_curr(rec);
2064
		pr_cont("\n expected tramp: %lx\n", ip);
2065
	}
2066 2067
}

2068
static int ftrace_check_record(struct dyn_ftrace *rec, int enable, int update)
2069
{
2070
	unsigned long flag = 0UL;
2071

2072 2073
	ftrace_bug_type = FTRACE_BUG_UNKNOWN;

2074 2075 2076
	if (rec->flags & FTRACE_FL_DISABLED)
		return FTRACE_UPDATE_IGNORE;

S
Steven Rostedt 已提交
2077
	/*
2078
	 * If we are updating calls:
S
Steven Rostedt 已提交
2079
	 *
2080 2081
	 *   If the record has a ref count, then we need to enable it
	 *   because someone is using it.
S
Steven Rostedt 已提交
2082
	 *
2083 2084
	 *   Otherwise we make sure its disabled.
	 *
2085
	 * If we are disabling calls, then disable all records that
2086
	 * are enabled.
S
Steven Rostedt 已提交
2087
	 */
2088
	if (enable && ftrace_rec_count(rec))
2089
		flag = FTRACE_FL_ENABLED;
S
Steven Rostedt 已提交
2090

2091
	/*
2092 2093 2094
	 * 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.
2095
	 */
2096 2097 2098 2099 2100 2101 2102 2103 2104
	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;
	}
2105

2106 2107
	/* If the state of this record hasn't changed, then do nothing */
	if ((rec->flags & FTRACE_FL_ENABLED) == flag)
2108
		return FTRACE_UPDATE_IGNORE;
S
Steven Rostedt 已提交
2109

2110
	if (flag) {
2111 2112 2113 2114
		/* Save off if rec is being enabled (for return value) */
		flag ^= rec->flags & FTRACE_FL_ENABLED;

		if (update) {
2115
			rec->flags |= FTRACE_FL_ENABLED;
2116 2117 2118 2119 2120 2121
			if (flag & FTRACE_FL_REGS) {
				if (rec->flags & FTRACE_FL_REGS)
					rec->flags |= FTRACE_FL_REGS_EN;
				else
					rec->flags &= ~FTRACE_FL_REGS_EN;
			}
2122 2123 2124 2125 2126 2127
			if (flag & FTRACE_FL_TRAMP) {
				if (rec->flags & FTRACE_FL_TRAMP)
					rec->flags |= FTRACE_FL_TRAMP_EN;
				else
					rec->flags &= ~FTRACE_FL_TRAMP_EN;
			}
2128 2129 2130 2131 2132 2133 2134
		}

		/*
		 * 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
2135
		 *   from the save regs, to a non-save regs function or
2136
		 *   vice versa, or from a trampoline call.
2137
		 */
2138 2139
		if (flag & FTRACE_FL_ENABLED) {
			ftrace_bug_type = FTRACE_BUG_CALL;
2140
			return FTRACE_UPDATE_MAKE_CALL;
2141
		}
2142

2143
		ftrace_bug_type = FTRACE_BUG_UPDATE;
2144
		return FTRACE_UPDATE_MODIFY_CALL;
2145 2146
	}

2147 2148
	if (update) {
		/* If there's no more users, clear all flags */
2149
		if (!ftrace_rec_count(rec))
2150 2151
			rec->flags = 0;
		else
2152 2153 2154 2155 2156 2157
			/*
			 * 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);
2158
	}
2159

2160
	ftrace_bug_type = FTRACE_BUG_NOP;
2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190
	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);
}

2191 2192 2193 2194
static struct ftrace_ops *
ftrace_find_tramp_ops_any(struct dyn_ftrace *rec)
{
	struct ftrace_ops *op;
2195
	unsigned long ip = rec->ip;
2196 2197 2198 2199 2200 2201

	do_for_each_ftrace_op(op, ftrace_ops_list) {

		if (!op->trampoline)
			continue;

2202
		if (hash_contains_ip(ip, op->func_hash))
2203 2204 2205 2206 2207 2208
			return op;
	} while_for_each_ftrace_op(op);

	return NULL;
}

2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226
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;
}

2227 2228 2229 2230
static struct ftrace_ops *
ftrace_find_tramp_ops_curr(struct dyn_ftrace *rec)
{
	struct ftrace_ops *op;
2231
	unsigned long ip = rec->ip;
2232

2233 2234 2235 2236 2237 2238 2239 2240
	/*
	 * 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))
2241 2242 2243
			return removed_ops;
	}

2244 2245 2246 2247 2248 2249 2250
	/*
	 * 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
2251 2252 2253 2254 2255 2256 2257 2258 2259 2260
	 * 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.
2261
	 */
2262
	do_for_each_ftrace_op(op, ftrace_ops_list) {
2263 2264 2265 2266 2267 2268 2269 2270 2271

		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)
2272 2273
			continue;

2274

2275
		/*
2276 2277 2278
		 * If the ops is being modified and is in the old
		 * hash, then it is probably being removed from this
		 * function.
2279 2280 2281
		 */
		if ((op->flags & FTRACE_OPS_FL_MODIFYING) &&
		    hash_contains_ip(ip, &op->old_hash))
2282
			return op;
2283 2284 2285 2286 2287 2288 2289 2290
		/*
		 * 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;
2291 2292 2293 2294 2295 2296 2297 2298 2299 2300

	} 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;
2301
	unsigned long ip = rec->ip;
2302 2303 2304

	do_for_each_ftrace_op(op, ftrace_ops_list) {
		/* pass rec in as regs to have non-NULL val */
2305
		if (hash_contains_ip(ip, op->func_hash))
2306 2307 2308 2309 2310 2311
			return op;
	} while_for_each_ftrace_op(op);

	return NULL;
}

2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323
/**
 * 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)
{
2324 2325 2326 2327 2328 2329
	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)) {
2330 2331
			pr_warn("Bad trampoline accounting at: %p (%pS) (%lx)\n",
				(void *)rec->ip, (void *)rec->ip, rec->flags);
2332 2333 2334 2335 2336 2337
			/* Ftrace is shutting down, return anything */
			return (unsigned long)FTRACE_ADDR;
		}
		return ops->trampoline;
	}

2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355
	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)
{
2356 2357 2358 2359 2360 2361
	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)) {
2362 2363
			pr_warn("Bad trampoline accounting at: %p (%pS)\n",
				(void *)rec->ip, (void *)rec->ip);
2364 2365 2366 2367 2368 2369
			/* Ftrace is shutting down, return anything */
			return (unsigned long)FTRACE_ADDR;
		}
		return ops->trampoline;
	}

2370 2371 2372 2373 2374 2375
	if (rec->flags & FTRACE_FL_REGS_EN)
		return (unsigned long)FTRACE_REGS_ADDR;
	else
		return (unsigned long)FTRACE_ADDR;
}

2376 2377 2378
static int
__ftrace_replace_code(struct dyn_ftrace *rec, int enable)
{
2379
	unsigned long ftrace_old_addr;
2380 2381 2382
	unsigned long ftrace_addr;
	int ret;

2383
	ftrace_addr = ftrace_get_addr_new(rec);
2384

2385 2386 2387 2388
	/* 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);
2389

2390 2391
	ftrace_bug_type = FTRACE_BUG_UNKNOWN;

2392 2393 2394 2395 2396
	switch (ret) {
	case FTRACE_UPDATE_IGNORE:
		return 0;

	case FTRACE_UPDATE_MAKE_CALL:
2397
		ftrace_bug_type = FTRACE_BUG_CALL;
2398
		return ftrace_make_call(rec, ftrace_addr);
2399 2400

	case FTRACE_UPDATE_MAKE_NOP:
2401
		ftrace_bug_type = FTRACE_BUG_NOP;
2402
		return ftrace_make_nop(NULL, rec, ftrace_old_addr);
2403 2404

	case FTRACE_UPDATE_MODIFY_CALL:
2405
		ftrace_bug_type = FTRACE_BUG_UPDATE;
2406
		return ftrace_modify_call(rec, ftrace_old_addr, ftrace_addr);
2407 2408
	}

2409
	return -1; /* unknow ftrace bug */
2410 2411
}

2412
void __weak ftrace_replace_code(int enable)
2413 2414 2415
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
S
Steven Rostedt 已提交
2416
	int failed;
2417

2418 2419 2420
	if (unlikely(ftrace_disabled))
		return;

2421
	do_for_each_ftrace_rec(pg, rec) {
2422 2423 2424 2425

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

2426
		failed = __ftrace_replace_code(rec, enable);
2427
		if (failed) {
2428
			ftrace_bug(failed, rec);
2429 2430
			/* Stop processing */
			return;
2431
		}
2432
	} while_for_each_ftrace_rec();
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 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506
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];
}

2507
static int
2508
ftrace_code_disable(struct module *mod, struct dyn_ftrace *rec)
2509
{
2510
	int ret;
2511

2512 2513 2514
	if (unlikely(ftrace_disabled))
		return 0;

2515
	ret = ftrace_make_nop(mod, rec, MCOUNT_ADDR);
2516
	if (ret) {
2517
		ftrace_bug_type = FTRACE_BUG_INIT;
2518
		ftrace_bug(ret, rec);
2519
		return 0;
2520
	}
2521
	return 1;
2522 2523
}

2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541
/*
 * 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;
}

2542
void ftrace_modify_all_code(int command)
2543
{
2544
	int update = command & FTRACE_UPDATE_TRACE_FUNC;
2545
	int err = 0;
2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556

	/*
	 * 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.
	 */
2557 2558 2559 2560 2561
	if (update) {
		err = ftrace_update_ftrace_func(ftrace_ops_list_func);
		if (FTRACE_WARN_ON(err))
			return;
	}
2562

2563
	if (command & FTRACE_UPDATE_CALLS)
2564
		ftrace_replace_code(1);
2565
	else if (command & FTRACE_DISABLE_CALLS)
2566 2567
		ftrace_replace_code(0);

2568 2569 2570 2571 2572 2573
	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);
2574 2575 2576
		err = ftrace_update_ftrace_func(ftrace_trace_function);
		if (FTRACE_WARN_ON(err))
			return;
2577
	}
2578

2579
	if (command & FTRACE_START_FUNC_RET)
2580
		err = ftrace_enable_ftrace_graph_caller();
2581
	else if (command & FTRACE_STOP_FUNC_RET)
2582 2583
		err = ftrace_disable_ftrace_graph_caller();
	FTRACE_WARN_ON(err);
2584 2585 2586 2587 2588 2589 2590
}

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

	ftrace_modify_all_code(*command);
2591

2592
	return 0;
2593 2594
}

2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618
/**
 * 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
2619
static void ftrace_run_update_code(int command)
2620
{
2621 2622 2623 2624 2625 2626 2627
	int ret;

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

2628 2629 2630 2631 2632 2633 2634 2635
	/*
	 * 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);

2636 2637
	ret = ftrace_arch_code_modify_post_process();
	FTRACE_WARN_ON(ret);
2638 2639
}

2640
static void ftrace_run_modify_code(struct ftrace_ops *ops, int command,
2641
				   struct ftrace_ops_hash *old_hash)
2642 2643
{
	ops->flags |= FTRACE_OPS_FL_MODIFYING;
2644 2645
	ops->old_hash.filter_hash = old_hash->filter_hash;
	ops->old_hash.notrace_hash = old_hash->notrace_hash;
2646
	ftrace_run_update_code(command);
2647
	ops->old_hash.filter_hash = NULL;
2648
	ops->old_hash.notrace_hash = NULL;
2649 2650 2651
	ops->flags &= ~FTRACE_OPS_FL_MODIFYING;
}

2652
static ftrace_func_t saved_ftrace_func;
2653
static int ftrace_start_up;
2654

2655 2656 2657 2658
void __weak arch_ftrace_trampoline_free(struct ftrace_ops *ops)
{
}

2659
static void per_cpu_ops_free(struct ftrace_ops *ops)
2660 2661 2662 2663
{
	free_percpu(ops->disabled);
}

2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675
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);
}
2676

2677 2678 2679 2680 2681 2682 2683
static void ftrace_startup_all(int command)
{
	update_all_ops = true;
	ftrace_startup_enable(command);
	update_all_ops = false;
}

2684
static int ftrace_startup(struct ftrace_ops *ops, int command)
2685
{
2686
	int ret;
2687

2688
	if (unlikely(ftrace_disabled))
2689
		return -ENODEV;
2690

2691 2692 2693 2694
	ret = __register_ftrace_function(ops);
	if (ret)
		return ret;

2695
	ftrace_start_up++;
2696

2697 2698 2699 2700 2701 2702 2703 2704 2705
	/*
	 * 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;
2706

2707 2708 2709 2710 2711 2712 2713 2714 2715
	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;
	}

2716 2717
	if (ftrace_hash_rec_enable(ops, 1))
		command |= FTRACE_UPDATE_CALLS;
2718

2719
	ftrace_startup_enable(command);
2720

2721 2722
	ops->flags &= ~FTRACE_OPS_FL_ADDING;

2723
	return 0;
2724 2725
}

2726
static int ftrace_shutdown(struct ftrace_ops *ops, int command)
2727
{
2728
	int ret;
2729

2730
	if (unlikely(ftrace_disabled))
2731 2732 2733 2734 2735
		return -ENODEV;

	ret = __unregister_ftrace_function(ops);
	if (ret)
		return ret;
2736

2737
	ftrace_start_up--;
2738 2739 2740 2741 2742 2743 2744
	/*
	 * 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);

2745 2746
	/* Disabling ipmodify never fails */
	ftrace_hash_ipmodify_disable(ops);
2747

2748 2749
	if (ftrace_hash_rec_disable(ops, 1))
		command |= FTRACE_UPDATE_CALLS;
2750

2751
	ops->flags &= ~FTRACE_OPS_FL_ENABLED;
2752

2753 2754 2755 2756
	if (saved_ftrace_func != ftrace_trace_function) {
		saved_ftrace_func = ftrace_trace_function;
		command |= FTRACE_UPDATE_TRACE_FUNC;
	}
2757

2758 2759
	if (!command || !ftrace_enabled) {
		/*
2760
		 * If these are per_cpu ops, they still need their
2761 2762 2763 2764
		 * per_cpu field freed. Since, function tracing is
		 * not currently active, we can just free them
		 * without synchronizing all CPUs.
		 */
2765 2766
		if (ops->flags & FTRACE_OPS_FL_PER_CPU)
			per_cpu_ops_free(ops);
2767
		return 0;
2768
	}
2769

2770 2771 2772 2773
	/*
	 * If the ops uses a trampoline, then it needs to be
	 * tested first on update.
	 */
2774
	ops->flags |= FTRACE_OPS_FL_REMOVING;
2775 2776
	removed_ops = ops;

2777 2778 2779 2780
	/* 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;

2781
	ftrace_run_update_code(command);
2782

2783 2784 2785 2786 2787 2788 2789 2790 2791
	/*
	 * 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) {
2792
			if (FTRACE_WARN_ON_ONCE(rec->flags & ~FTRACE_FL_DISABLED))
2793 2794 2795 2796 2797
				pr_warn("  %pS flags:%lx\n",
					(void *)rec->ip, rec->flags);
		} while_for_each_ftrace_rec();
	}

2798 2799 2800 2801
	ops->old_hash.filter_hash = NULL;
	ops->old_hash.notrace_hash = NULL;

	removed_ops = NULL;
2802
	ops->flags &= ~FTRACE_OPS_FL_REMOVING;
2803

2804 2805 2806
	/*
	 * Dynamic ops may be freed, we must make sure that all
	 * callers are done before leaving this function.
2807
	 * The same goes for freeing the per_cpu data of the per_cpu
2808 2809 2810 2811 2812 2813 2814 2815 2816 2817
	 * 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.
	 */
2818
	if (ops->flags & (FTRACE_OPS_FL_DYNAMIC | FTRACE_OPS_FL_PER_CPU)) {
2819 2820
		schedule_on_each_cpu(ftrace_sync);

2821 2822
		arch_ftrace_trampoline_free(ops);

2823 2824
		if (ops->flags & FTRACE_OPS_FL_PER_CPU)
			per_cpu_ops_free(ops);
2825 2826
	}

2827
	return 0;
2828 2829
}

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

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

2837 2838
	/* Force update next time */
	saved_ftrace_func = NULL;
2839
	/* ftrace_start_up is true if we want ftrace running */
2840 2841 2842 2843
	if (ftrace_start_up) {
		command = FTRACE_UPDATE_CALLS;
		if (ftrace_graph_active)
			command |= FTRACE_START_FUNC_RET;
2844
		ftrace_startup_enable(command);
2845
	}
2846 2847
}

I
Ingo Molnar 已提交
2848
static void ftrace_shutdown_sysctl(void)
2849
{
2850 2851
	int command;

2852 2853 2854
	if (unlikely(ftrace_disabled))
		return;

2855
	/* ftrace_start_up is true if ftrace is running */
2856 2857 2858 2859 2860 2861
	if (ftrace_start_up) {
		command = FTRACE_DISABLE_CALLS;
		if (ftrace_graph_active)
			command |= FTRACE_STOP_FUNC_RET;
		ftrace_run_update_code(command);
	}
2862 2863
}

2864
static u64		ftrace_update_time;
2865 2866
unsigned long		ftrace_update_tot_cnt;

2867
static inline int ops_traces_mod(struct ftrace_ops *ops)
2868
{
2869 2870 2871 2872
	/*
	 * Filter_hash being empty will default to trace module.
	 * But notrace hash requires a test of individual module functions.
	 */
2873 2874
	return ftrace_hash_empty(ops->func_hash->filter_hash) &&
		ftrace_hash_empty(ops->func_hash->notrace_hash);
2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890
}

/*
 * 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;

2891
	/* If ops traces all then it includes this function */
2892
	if (ops_traces_mod(ops))
2893
		return 1;
2894 2895

	/* The function must be in the filter */
2896
	if (!ftrace_hash_empty(ops->func_hash->filter_hash) &&
2897
	    !__ftrace_lookup_ip(ops->func_hash->filter_hash, rec->ip))
2898
		return 0;
2899

2900
	/* If in notrace hash, we ignore it too */
2901
	if (ftrace_lookup_ip(ops->func_hash->notrace_hash, rec->ip))
2902 2903 2904 2905 2906
		return 0;

	return 1;
}

2907
static int ftrace_update_code(struct module *mod, struct ftrace_page *new_pgs)
2908
{
2909
	struct ftrace_page *pg;
2910
	struct dyn_ftrace *p;
2911
	u64 start, stop;
2912
	unsigned long update_cnt = 0;
2913
	unsigned long rec_flags = 0;
2914
	int i;
2915

2916 2917
	start = ftrace_now(raw_smp_processor_id());

2918
	/*
2919 2920 2921 2922 2923 2924 2925 2926 2927
	 * 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.
2928
	 */
2929 2930
	if (mod)
		rec_flags |= FTRACE_FL_DISABLED;
2931

2932
	for (pg = new_pgs; pg; pg = pg->next) {
2933

2934
		for (i = 0; i < pg->index; i++) {
2935

2936 2937 2938
			/* If something went wrong, bail without enabling anything */
			if (unlikely(ftrace_disabled))
				return -1;
2939

2940
			p = &pg->records[i];
2941
			p->flags = rec_flags;
2942

2943 2944 2945 2946 2947 2948
			/*
			 * Do the initial record conversion from mcount jump
			 * to the NOP instructions.
			 */
			if (!ftrace_code_disable(mod, p))
				break;
2949

2950
			update_cnt++;
2951
		}
2952 2953
	}

I
Ingo Molnar 已提交
2954
	stop = ftrace_now(raw_smp_processor_id());
2955
	ftrace_update_time = stop - start;
2956
	ftrace_update_tot_cnt += update_cnt;
2957

2958 2959 2960
	return 0;
}

2961
static int ftrace_allocate_records(struct ftrace_page *pg, int count)
2962
{
2963
	int order;
2964 2965
	int cnt;

2966 2967 2968 2969
	if (WARN_ON(!count))
		return -EINVAL;

	order = get_count_order(DIV_ROUND_UP(count, ENTRIES_PER_PAGE));
2970 2971

	/*
2972 2973
	 * We want to fill as much as possible. No more than a page
	 * may be empty.
2974
	 */
2975 2976
	while ((PAGE_SIZE << order) / ENTRY_SIZE >= count + ENTRIES_PER_PAGE)
		order--;
2977

2978 2979
 again:
	pg->records = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, order);
2980

2981 2982 2983 2984 2985 2986 2987
	if (!pg->records) {
		/* if we can't allocate this size, try something smaller */
		if (!order)
			return -ENOMEM;
		order >>= 1;
		goto again;
	}
2988

2989 2990
	cnt = (PAGE_SIZE << order) / ENTRY_SIZE;
	pg->size = cnt;
2991

2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024
	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)
3025 3026
			break;

3027 3028 3029 3030
		pg->next = kzalloc(sizeof(*pg), GFP_KERNEL);
		if (!pg->next)
			goto free_pages;

3031 3032 3033
		pg = pg->next;
	}

3034 3035 3036
	return start_pg;

 free_pages:
3037 3038
	pg = start_pg;
	while (pg) {
3039 3040 3041 3042 3043 3044 3045 3046 3047 3048
		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;
}

3049 3050 3051
#define FTRACE_BUFF_MAX (KSYM_SYMBOL_LEN+4) /* room for wildcards */

struct ftrace_iterator {
3052
	loff_t				pos;
3053 3054 3055 3056 3057
	loff_t				func_pos;
	struct ftrace_page		*pg;
	struct dyn_ftrace		*func;
	struct ftrace_func_probe	*probe;
	struct trace_parser		parser;
3058
	struct ftrace_hash		*hash;
3059
	struct ftrace_ops		*ops;
3060 3061 3062
	int				hidx;
	int				idx;
	unsigned			flags;
3063 3064
};

3065
static void *
3066
t_hash_next(struct seq_file *m, loff_t *pos)
3067 3068
{
	struct ftrace_iterator *iter = m->private;
3069
	struct hlist_node *hnd = NULL;
3070 3071 3072
	struct hlist_head *hhd;

	(*pos)++;
3073
	iter->pos = *pos;
3074

3075 3076
	if (iter->probe)
		hnd = &iter->probe->node;
3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098
 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;
		}
	}

3099 3100 3101 3102 3103 3104
	if (WARN_ON_ONCE(!hnd))
		return NULL;

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

	return iter;
3105 3106 3107 3108 3109 3110
}

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
3111 3112
	loff_t l;

3113 3114 3115
	if (!(iter->flags & FTRACE_ITER_DO_HASH))
		return NULL;

3116 3117
	if (iter->func_pos > *pos)
		return NULL;
3118

L
Li Zefan 已提交
3119
	iter->hidx = 0;
3120
	for (l = 0; l <= (*pos - iter->func_pos); ) {
3121
		p = t_hash_next(m, &l);
L
Li Zefan 已提交
3122 3123 3124
		if (!p)
			break;
	}
3125 3126 3127
	if (!p)
		return NULL;

3128 3129 3130
	/* Only set this if we have an item */
	iter->flags |= FTRACE_ITER_HASH;

3131
	return iter;
3132 3133
}

3134 3135
static int
t_hash_show(struct seq_file *m, struct ftrace_iterator *iter)
3136
{
S
Steven Rostedt 已提交
3137
	struct ftrace_func_probe *rec;
3138

3139 3140 3141
	rec = iter->probe;
	if (WARN_ON_ONCE(!rec))
		return -EIO;
3142

3143 3144 3145
	if (rec->ops->print)
		return rec->ops->print(m, rec->ip, rec->ops, rec->data);

3146
	seq_printf(m, "%ps:%ps", (void *)rec->ip, (void *)rec->ops->func);
3147 3148 3149 3150 3151 3152 3153 3154

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

	return 0;
}

I
Ingo Molnar 已提交
3155
static void *
3156 3157 3158
t_next(struct seq_file *m, void *v, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
3159
	struct ftrace_ops *ops = iter->ops;
3160 3161
	struct dyn_ftrace *rec = NULL;

3162 3163 3164
	if (unlikely(ftrace_disabled))
		return NULL;

3165
	if (iter->flags & FTRACE_ITER_HASH)
3166
		return t_hash_next(m, pos);
3167

3168
	(*pos)++;
3169
	iter->pos = iter->func_pos = *pos;
3170

3171
	if (iter->flags & FTRACE_ITER_PRINTALL)
3172
		return t_hash_start(m, pos);
3173

3174 3175 3176 3177 3178 3179 3180 3181 3182
 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++];
3183
		if (((iter->flags & FTRACE_ITER_FILTER) &&
3184
		     !(ftrace_lookup_ip(ops->func_hash->filter_hash, rec->ip))) ||
S
Steven Rostedt 已提交
3185

3186
		    ((iter->flags & FTRACE_ITER_NOTRACE) &&
3187
		     !ftrace_lookup_ip(ops->func_hash->notrace_hash, rec->ip)) ||
3188 3189

		    ((iter->flags & FTRACE_ITER_ENABLED) &&
3190
		     !(rec->flags & FTRACE_FL_ENABLED))) {
3191

3192 3193 3194 3195 3196
			rec = NULL;
			goto retry;
		}
	}

3197
	if (!rec)
3198
		return t_hash_start(m, pos);
3199 3200 3201 3202

	iter->func = rec;

	return iter;
3203 3204
}

3205 3206 3207 3208
static void reset_iter_read(struct ftrace_iterator *iter)
{
	iter->pos = 0;
	iter->func_pos = 0;
3209
	iter->flags &= ~(FTRACE_ITER_PRINTALL | FTRACE_ITER_HASH);
3210 3211 3212 3213 3214
}

static void *t_start(struct seq_file *m, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
3215
	struct ftrace_ops *ops = iter->ops;
3216
	void *p = NULL;
3217
	loff_t l;
3218

3219
	mutex_lock(&ftrace_lock);
3220 3221 3222 3223

	if (unlikely(ftrace_disabled))
		return NULL;

3224 3225 3226 3227 3228 3229
	/*
	 * If an lseek was done, then reset and start from beginning.
	 */
	if (*pos < iter->pos)
		reset_iter_read(iter);

3230 3231 3232 3233 3234
	/*
	 * For set_ftrace_filter reading, if we have the filter
	 * off, we can short cut and just print out that all
	 * functions are enabled.
	 */
3235
	if ((iter->flags & FTRACE_ITER_FILTER &&
3236
	     ftrace_hash_empty(ops->func_hash->filter_hash)) ||
3237
	    (iter->flags & FTRACE_ITER_NOTRACE &&
3238
	     ftrace_hash_empty(ops->func_hash->notrace_hash))) {
3239
		if (*pos > 0)
3240
			return t_hash_start(m, pos);
3241
		iter->flags |= FTRACE_ITER_PRINTALL;
3242 3243
		/* reset in case of seek/pread */
		iter->flags &= ~FTRACE_ITER_HASH;
3244 3245 3246
		return iter;
	}

3247 3248 3249
	if (iter->flags & FTRACE_ITER_HASH)
		return t_hash_start(m, pos);

3250 3251 3252 3253 3254
	/*
	 * 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.
	 */
3255 3256 3257 3258 3259 3260
	iter->pg = ftrace_pages_start;
	iter->idx = 0;
	for (l = 0; l <= *pos; ) {
		p = t_next(m, p, &l);
		if (!p)
			break;
3261
	}
3262

3263 3264
	if (!p)
		return t_hash_start(m, pos);
3265 3266

	return iter;
3267 3268 3269 3270
}

static void t_stop(struct seq_file *m, void *p)
{
3271
	mutex_unlock(&ftrace_lock);
3272 3273
}

3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289
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);
}

3290 3291
static int t_show(struct seq_file *m, void *v)
{
3292
	struct ftrace_iterator *iter = m->private;
3293
	struct dyn_ftrace *rec;
3294

3295
	if (iter->flags & FTRACE_ITER_HASH)
3296
		return t_hash_show(m, iter);
3297

3298
	if (iter->flags & FTRACE_ITER_PRINTALL) {
3299
		if (iter->flags & FTRACE_ITER_NOTRACE)
3300
			seq_puts(m, "#### no functions disabled ####\n");
3301
		else
3302
			seq_puts(m, "#### all functions enabled ####\n");
3303 3304 3305
		return 0;
	}

3306 3307
	rec = iter->func;

3308 3309 3310
	if (!rec)
		return 0;

3311
	seq_printf(m, "%ps", (void *)rec->ip);
3312
	if (iter->flags & FTRACE_ITER_ENABLED) {
3313
		struct ftrace_ops *ops;
3314

3315
		seq_printf(m, " (%ld)%s%s",
3316
			   ftrace_rec_count(rec),
3317 3318
			   rec->flags & FTRACE_FL_REGS ? " R" : "  ",
			   rec->flags & FTRACE_FL_IPMODIFY ? " I" : "  ");
3319
		if (rec->flags & FTRACE_FL_TRAMP_EN) {
3320
			ops = ftrace_find_tramp_ops_any(rec);
3321 3322 3323 3324 3325
			if (ops) {
				do {
					seq_printf(m, "\ttramp: %pS (%pS)",
						   (void *)ops->trampoline,
						   (void *)ops->func);
3326
					add_trampoline_func(m, ops, rec);
3327 3328 3329
					ops = ftrace_find_tramp_ops_next(rec, ops);
				} while (ops);
			} else
3330
				seq_puts(m, "\ttramp: ERROR!");
3331 3332
		} else {
			add_trampoline_func(m, NULL, rec);
3333 3334 3335
		}
	}	

3336
	seq_putc(m, '\n');
3337 3338 3339 3340

	return 0;
}

J
James Morris 已提交
3341
static const struct seq_operations show_ftrace_seq_ops = {
3342 3343 3344 3345 3346 3347
	.start = t_start,
	.next = t_next,
	.stop = t_stop,
	.show = t_show,
};

I
Ingo Molnar 已提交
3348
static int
3349 3350 3351 3352
ftrace_avail_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;

3353 3354 3355
	if (unlikely(ftrace_disabled))
		return -ENODEV;

3356 3357 3358 3359
	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
3360
	}
3361

3362
	return iter ? 0 : -ENOMEM;
3363 3364
}

3365 3366 3367 3368 3369
static int
ftrace_enabled_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;

3370 3371 3372 3373 3374
	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;
3375 3376
	}

3377
	return iter ? 0 : -ENOMEM;
3378 3379
}

3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392
/**
 * 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.
3393
 * tracing_lseek() should be used as the lseek routine, and
3394 3395 3396
 * release must call ftrace_regex_release().
 */
int
3397
ftrace_regex_open(struct ftrace_ops *ops, int flag,
3398
		  struct inode *inode, struct file *file)
3399 3400
{
	struct ftrace_iterator *iter;
3401
	struct ftrace_hash *hash;
3402 3403
	int ret = 0;

3404 3405
	ftrace_ops_init(ops);

3406 3407 3408
	if (unlikely(ftrace_disabled))
		return -ENODEV;

3409 3410 3411 3412
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
	if (!iter)
		return -ENOMEM;

3413 3414 3415 3416 3417
	if (trace_parser_get_init(&iter->parser, FTRACE_BUFF_MAX)) {
		kfree(iter);
		return -ENOMEM;
	}

3418 3419 3420
	iter->ops = ops;
	iter->flags = flag;

3421
	mutex_lock(&ops->func_hash->regex_lock);
3422

3423
	if (flag & FTRACE_ITER_NOTRACE)
3424
		hash = ops->func_hash->notrace_hash;
3425
	else
3426
		hash = ops->func_hash->filter_hash;
3427

3428
	if (file->f_mode & FMODE_WRITE) {
3429 3430 3431 3432 3433 3434 3435
		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);

3436 3437 3438
		if (!iter->hash) {
			trace_parser_put(&iter->parser);
			kfree(iter);
3439 3440
			ret = -ENOMEM;
			goto out_unlock;
3441 3442
		}
	}
3443

3444 3445 3446 3447 3448 3449 3450
	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;
3451
		} else {
3452 3453
			/* Failed */
			free_ftrace_hash(iter->hash);
3454
			trace_parser_put(&iter->parser);
3455
			kfree(iter);
3456
		}
3457 3458
	} else
		file->private_data = iter;
3459 3460

 out_unlock:
3461
	mutex_unlock(&ops->func_hash->regex_lock);
3462 3463 3464 3465

	return ret;
}

3466 3467 3468
static int
ftrace_filter_open(struct inode *inode, struct file *file)
{
3469 3470 3471
	struct ftrace_ops *ops = inode->i_private;

	return ftrace_regex_open(ops,
3472 3473
			FTRACE_ITER_FILTER | FTRACE_ITER_DO_HASH,
			inode, file);
3474 3475 3476 3477 3478
}

static int
ftrace_notrace_open(struct inode *inode, struct file *file)
{
3479 3480 3481
	struct ftrace_ops *ops = inode->i_private;

	return ftrace_regex_open(ops, FTRACE_ITER_NOTRACE,
3482
				 inode, file);
3483 3484
}

3485 3486 3487 3488 3489 3490 3491
/* Type for quick search ftrace basic regexes (globs) from filter_parse_regex */
struct ftrace_glob {
	char *search;
	unsigned len;
	int type;
};

3492 3493 3494 3495 3496 3497 3498 3499 3500 3501
/*
 * 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;
}

3502
static int ftrace_match(char *str, struct ftrace_glob *g)
3503 3504
{
	int matched = 0;
3505
	int slen;
3506

3507 3508
	str = arch_ftrace_match_adjust(str, g->search);

3509
	switch (g->type) {
3510
	case MATCH_FULL:
3511
		if (strcmp(str, g->search) == 0)
3512 3513 3514
			matched = 1;
		break;
	case MATCH_FRONT_ONLY:
3515
		if (strncmp(str, g->search, g->len) == 0)
3516 3517 3518
			matched = 1;
		break;
	case MATCH_MIDDLE_ONLY:
3519
		if (strstr(str, g->search))
3520 3521 3522
			matched = 1;
		break;
	case MATCH_END_ONLY:
3523
		slen = strlen(str);
3524 3525
		if (slen >= g->len &&
		    memcmp(str + slen - g->len, g->search, g->len) == 0)
3526 3527
			matched = 1;
		break;
3528 3529 3530 3531
	case MATCH_GLOB:
		if (glob_match(g->search, str))
			matched = 1;
		break;
3532 3533 3534 3535 3536
	}

	return matched;
}

3537
static int
3538
enter_record(struct ftrace_hash *hash, struct dyn_ftrace *rec, int clear_filter)
3539
{
3540 3541 3542
	struct ftrace_func_entry *entry;
	int ret = 0;

3543
	entry = ftrace_lookup_ip(hash, rec->ip);
3544
	if (clear_filter) {
3545 3546 3547
		/* Do nothing if it doesn't exist */
		if (!entry)
			return 0;
3548

3549
		free_hash_entry(hash, entry);
3550 3551 3552 3553
	} else {
		/* Do nothing if it exists */
		if (entry)
			return 0;
3554

3555
		ret = add_hash_entry(hash, rec->ip);
3556 3557
	}
	return ret;
3558 3559
}

3560
static int
D
Dmitry Safonov 已提交
3561 3562
ftrace_match_record(struct dyn_ftrace *rec, struct ftrace_glob *func_g,
		struct ftrace_glob *mod_g, int exclude_mod)
3563 3564
{
	char str[KSYM_SYMBOL_LEN];
3565 3566 3567 3568
	char *modname;

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

D
Dmitry Safonov 已提交
3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588
	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)
3589 3590
			return 0;

D
Dmitry Safonov 已提交
3591
func_match:
3592
		/* blank search means to match all funcs in the mod */
3593
		if (!func_g->len)
3594 3595
			return 1;
	}
3596

3597
	return ftrace_match(str, func_g);
3598 3599
}

3600
static int
3601
match_records(struct ftrace_hash *hash, char *func, int len, char *mod)
3602 3603 3604
{
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
3605
	struct ftrace_glob func_g = { .type = MATCH_FULL };
D
Dmitry Safonov 已提交
3606 3607 3608
	struct ftrace_glob mod_g = { .type = MATCH_FULL };
	struct ftrace_glob *mod_match = (mod) ? &mod_g : NULL;
	int exclude_mod = 0;
3609
	int found = 0;
3610
	int ret;
3611
	int clear_filter;
3612

D
Dmitry Safonov 已提交
3613
	if (func) {
3614 3615 3616
		func_g.type = filter_parse_regex(func, len, &func_g.search,
						 &clear_filter);
		func_g.len = strlen(func_g.search);
3617
	}
3618

D
Dmitry Safonov 已提交
3619 3620 3621 3622
	if (mod) {
		mod_g.type = filter_parse_regex(mod, strlen(mod),
				&mod_g.search, &exclude_mod);
		mod_g.len = strlen(mod_g.search);
3623
	}
3624

3625
	mutex_lock(&ftrace_lock);
3626

3627 3628
	if (unlikely(ftrace_disabled))
		goto out_unlock;
3629

3630
	do_for_each_ftrace_rec(pg, rec) {
3631 3632 3633 3634

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

D
Dmitry Safonov 已提交
3635
		if (ftrace_match_record(rec, &func_g, mod_match, exclude_mod)) {
3636
			ret = enter_record(hash, rec, clear_filter);
3637 3638 3639 3640
			if (ret < 0) {
				found = ret;
				goto out_unlock;
			}
3641
			found = 1;
3642 3643
		}
	} while_for_each_ftrace_rec();
3644
 out_unlock:
3645
	mutex_unlock(&ftrace_lock);
3646 3647

	return found;
3648 3649
}

3650
static int
3651
ftrace_match_records(struct ftrace_hash *hash, char *buff, int len)
3652
{
3653
	return match_records(hash, buff, len, NULL);
3654 3655 3656
}


3657 3658 3659 3660 3661 3662
/*
 * We register the module command as a template to show others how
 * to register the a command as well.
 */

static int
3663
ftrace_mod_callback(struct ftrace_hash *hash,
3664
		    char *func, char *cmd, char *module, int enable)
3665
{
3666
	int ret;
3667 3668 3669 3670 3671 3672 3673 3674

	/*
	 * 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.
	 */
3675
	ret = match_records(hash, func, strlen(func), module);
3676
	if (!ret)
3677
		return -EINVAL;
3678 3679 3680
	if (ret < 0)
		return ret;
	return 0;
3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691
}

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);
}
3692
core_initcall(ftrace_mod_cmd_init);
3693

3694
static void function_trace_probe_call(unsigned long ip, unsigned long parent_ip,
3695
				      struct ftrace_ops *op, struct pt_regs *pt_regs)
3696
{
S
Steven Rostedt 已提交
3697
	struct ftrace_func_probe *entry;
3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 3711 3712
	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.
	 */
3713
	preempt_disable_notrace();
3714
	hlist_for_each_entry_rcu_notrace(entry, hhd, node) {
3715 3716 3717
		if (entry->ip == ip)
			entry->ops->func(ip, parent_ip, &entry->data);
	}
3718
	preempt_enable_notrace();
3719 3720
}

S
Steven Rostedt 已提交
3721
static struct ftrace_ops trace_probe_ops __read_mostly =
3722
{
3723
	.func		= function_trace_probe_call,
3724
	.flags		= FTRACE_OPS_FL_INITIALIZED,
3725
	INIT_OPS_HASH(trace_probe_ops)
3726 3727
};

S
Steven Rostedt 已提交
3728
static int ftrace_probe_registered;
3729

3730
static void __enable_ftrace_function_probe(struct ftrace_ops_hash *old_hash)
3731
{
3732
	int ret;
3733 3734
	int i;

3735 3736 3737
	if (ftrace_probe_registered) {
		/* still need to update the function call sites */
		if (ftrace_enabled)
3738 3739
			ftrace_run_modify_code(&trace_probe_ops, FTRACE_UPDATE_CALLS,
					       old_hash);
3740
		return;
3741
	}
3742 3743 3744 3745 3746 3747 3748 3749 3750 3751

	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;

3752
	ret = ftrace_startup(&trace_probe_ops, 0);
3753

S
Steven Rostedt 已提交
3754
	ftrace_probe_registered = 1;
3755 3756
}

S
Steven Rostedt 已提交
3757
static void __disable_ftrace_function_probe(void)
3758 3759 3760
{
	int i;

S
Steven Rostedt 已提交
3761
	if (!ftrace_probe_registered)
3762 3763 3764 3765 3766 3767 3768 3769 3770
		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 */
3771
	ftrace_shutdown(&trace_probe_ops, 0);
3772

S
Steven Rostedt 已提交
3773
	ftrace_probe_registered = 0;
3774 3775 3776
}


3777
static void ftrace_free_entry(struct ftrace_func_probe *entry)
3778 3779
{
	if (entry->ops->free)
3780
		entry->ops->free(entry->ops, entry->ip, &entry->data);
3781 3782 3783 3784
	kfree(entry);
}

int
S
Steven Rostedt 已提交
3785
register_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
3786 3787
			      void *data)
{
3788
	struct ftrace_ops_hash old_hash_ops;
S
Steven Rostedt 已提交
3789
	struct ftrace_func_probe *entry;
3790
	struct ftrace_glob func_g;
3791
	struct ftrace_hash **orig_hash = &trace_probe_ops.func_hash->filter_hash;
3792
	struct ftrace_hash *old_hash = *orig_hash;
3793
	struct ftrace_hash *hash;
3794 3795
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
3796
	int not;
S
Steven Rostedt 已提交
3797
	unsigned long key;
3798
	int count = 0;
3799
	int ret;
3800

3801 3802 3803
	func_g.type = filter_parse_regex(glob, strlen(glob),
			&func_g.search, &not);
	func_g.len = strlen(func_g.search);
3804

S
Steven Rostedt 已提交
3805
	/* we do not support '!' for function probes */
3806 3807 3808
	if (WARN_ON(not))
		return -EINVAL;

3809
	mutex_lock(&trace_probe_ops.func_hash->regex_lock);
3810

3811 3812 3813 3814
	old_hash_ops.filter_hash = old_hash;
	/* Probes only have filters */
	old_hash_ops.notrace_hash = NULL;

3815
	hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, old_hash);
3816 3817
	if (!hash) {
		count = -ENOMEM;
3818
		goto out;
3819 3820 3821 3822
	}

	if (unlikely(ftrace_disabled)) {
		count = -ENODEV;
3823
		goto out;
3824
	}
3825

3826 3827
	mutex_lock(&ftrace_lock);

3828
	do_for_each_ftrace_rec(pg, rec) {
3829

3830 3831 3832
		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

D
Dmitry Safonov 已提交
3833
		if (!ftrace_match_record(rec, &func_g, NULL, 0))
3834 3835 3836 3837
			continue;

		entry = kmalloc(sizeof(*entry), GFP_KERNEL);
		if (!entry) {
S
Steven Rostedt 已提交
3838
			/* If we did not process any, then return error */
3839 3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852
			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.
		 */
3853 3854
		if (ops->init) {
			if (ops->init(ops, rec->ip, &entry->data) < 0) {
3855 3856 3857 3858 3859 3860
				/* caller does not like this func */
				kfree(entry);
				continue;
			}
		}

3861 3862 3863 3864 3865 3866 3867
		ret = enter_record(hash, rec, 0);
		if (ret < 0) {
			kfree(entry);
			count = ret;
			goto out_unlock;
		}

3868 3869 3870 3871 3872 3873 3874
		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();
3875 3876

	ret = ftrace_hash_move(&trace_probe_ops, 1, orig_hash, hash);
3877

3878
	__enable_ftrace_function_probe(&old_hash_ops);
3879

3880 3881 3882
	if (!ret)
		free_ftrace_hash_rcu(old_hash);
	else
3883 3884
		count = ret;

3885
 out_unlock:
3886 3887
	mutex_unlock(&ftrace_lock);
 out:
3888
	mutex_unlock(&trace_probe_ops.func_hash->regex_lock);
3889
	free_ftrace_hash(hash);
3890 3891 3892 3893 3894

	return count;
}

enum {
S
Steven Rostedt 已提交
3895 3896
	PROBE_TEST_FUNC		= 1,
	PROBE_TEST_DATA		= 2
3897 3898 3899
};

static void
S
Steven Rostedt 已提交
3900
__unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
3901 3902
				  void *data, int flags)
{
3903
	struct ftrace_func_entry *rec_entry;
S
Steven Rostedt 已提交
3904
	struct ftrace_func_probe *entry;
3905
	struct ftrace_func_probe *p;
3906
	struct ftrace_glob func_g;
3907
	struct ftrace_hash **orig_hash = &trace_probe_ops.func_hash->filter_hash;
3908
	struct ftrace_hash *old_hash = *orig_hash;
3909
	struct list_head free_list;
3910
	struct ftrace_hash *hash;
3911
	struct hlist_node *tmp;
3912
	char str[KSYM_SYMBOL_LEN];
3913
	int i, ret;
3914

3915
	if (glob && (strcmp(glob, "*") == 0 || !strlen(glob)))
3916
		func_g.search = NULL;
3917
	else if (glob) {
3918 3919
		int not;

3920 3921 3922 3923
		func_g.type = filter_parse_regex(glob, strlen(glob),
						 &func_g.search, &not);
		func_g.len = strlen(func_g.search);
		func_g.search = glob;
3924

S
Steven Rostedt 已提交
3925
		/* we do not support '!' for function probes */
3926 3927 3928 3929
		if (WARN_ON(not))
			return;
	}

3930
	mutex_lock(&trace_probe_ops.func_hash->regex_lock);
3931 3932 3933 3934 3935 3936

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

3937 3938
	INIT_LIST_HEAD(&free_list);

3939 3940 3941
	for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) {
		struct hlist_head *hhd = &ftrace_func_hash[i];

3942
		hlist_for_each_entry_safe(entry, tmp, hhd, node) {
3943 3944

			/* break up if statements for readability */
S
Steven Rostedt 已提交
3945
			if ((flags & PROBE_TEST_FUNC) && entry->ops != ops)
3946 3947
				continue;

S
Steven Rostedt 已提交
3948
			if ((flags & PROBE_TEST_DATA) && entry->data != data)
3949 3950 3951
				continue;

			/* do this last, since it is the most expensive */
3952
			if (func_g.search) {
3953 3954
				kallsyms_lookup(entry->ip, NULL, NULL,
						NULL, str);
3955
				if (!ftrace_match(str, &func_g))
3956 3957 3958
					continue;
			}

3959 3960 3961 3962 3963
			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);

3964
			hlist_del_rcu(&entry->node);
3965
			list_add(&entry->free_list, &free_list);
3966 3967
		}
	}
3968
	mutex_lock(&ftrace_lock);
S
Steven Rostedt 已提交
3969
	__disable_ftrace_function_probe();
3970 3971 3972 3973
	/*
	 * Remove after the disable is called. Otherwise, if the last
	 * probe is removed, a null hash means *all enabled*.
	 */
3974
	ret = ftrace_hash_move(&trace_probe_ops, 1, orig_hash, hash);
3975
	synchronize_sched();
3976 3977 3978
	if (!ret)
		free_ftrace_hash_rcu(old_hash);

3979 3980 3981 3982
	list_for_each_entry_safe(entry, p, &free_list, free_list) {
		list_del(&entry->free_list);
		ftrace_free_entry(entry);
	}
3983
	mutex_unlock(&ftrace_lock);
3984

3985
 out_unlock:
3986
	mutex_unlock(&trace_probe_ops.func_hash->regex_lock);
3987
	free_ftrace_hash(hash);
3988 3989 3990
}

void
S
Steven Rostedt 已提交
3991
unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
3992 3993
				void *data)
{
S
Steven Rostedt 已提交
3994 3995
	__unregister_ftrace_function_probe(glob, ops, data,
					  PROBE_TEST_FUNC | PROBE_TEST_DATA);
3996 3997 3998
}

void
S
Steven Rostedt 已提交
3999
unregister_ftrace_function_probe_func(char *glob, struct ftrace_probe_ops *ops)
4000
{
S
Steven Rostedt 已提交
4001
	__unregister_ftrace_function_probe(glob, ops, NULL, PROBE_TEST_FUNC);
4002 4003
}

S
Steven Rostedt 已提交
4004
void unregister_ftrace_function_probe_all(char *glob)
4005
{
S
Steven Rostedt 已提交
4006
	__unregister_ftrace_function_probe(glob, NULL, NULL, 0);
4007 4008
}

4009 4010 4011
static LIST_HEAD(ftrace_commands);
static DEFINE_MUTEX(ftrace_cmd_mutex);

4012 4013 4014 4015 4016
/*
 * Currently we only register ftrace commands from __init, so mark this
 * __init too.
 */
__init int register_ftrace_command(struct ftrace_func_command *cmd)
4017 4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034
{
	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;
}

4035 4036 4037 4038 4039
/*
 * Currently we only unregister ftrace commands from __init, so mark
 * this __init too.
 */
__init int unregister_ftrace_command(struct ftrace_func_command *cmd)
4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050 4051 4052 4053 4054 4055 4056 4057
{
	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;
}

4058 4059
static int ftrace_process_regex(struct ftrace_hash *hash,
				char *buff, int len, int enable)
4060
{
4061
	char *func, *command, *next = buff;
S
Steven Rostedt 已提交
4062
	struct ftrace_func_command *p;
4063
	int ret = -EINVAL;
4064 4065 4066 4067

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

	if (!next) {
4068
		ret = ftrace_match_records(hash, func, len);
4069 4070 4071 4072 4073
		if (!ret)
			ret = -EINVAL;
		if (ret < 0)
			return ret;
		return 0;
4074 4075
	}

4076
	/* command found */
4077 4078 4079

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

4080 4081 4082
	mutex_lock(&ftrace_cmd_mutex);
	list_for_each_entry(p, &ftrace_commands, list) {
		if (strcmp(p->name, command) == 0) {
4083
			ret = p->func(hash, func, command, next, enable);
4084 4085
			goto out_unlock;
		}
4086
	}
4087 4088
 out_unlock:
	mutex_unlock(&ftrace_cmd_mutex);
4089

4090
	return ret;
4091 4092
}

I
Ingo Molnar 已提交
4093
static ssize_t
4094 4095
ftrace_regex_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos, int enable)
4096 4097
{
	struct ftrace_iterator *iter;
4098 4099
	struct trace_parser *parser;
	ssize_t ret, read;
4100

4101
	if (!cnt)
4102 4103 4104 4105 4106 4107 4108 4109
		return 0;

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

4110
	if (unlikely(ftrace_disabled))
4111 4112 4113
		return -ENODEV;

	/* iter->hash is a local copy, so we don't need regex_lock */
4114

4115 4116
	parser = &iter->parser;
	read = trace_get_user(parser, ubuf, cnt, ppos);
4117

4118
	if (read >= 0 && trace_parser_loaded(parser) &&
4119
	    !trace_parser_cont(parser)) {
4120
		ret = ftrace_process_regex(iter->hash, parser->buffer,
4121
					   parser->idx, enable);
4122
		trace_parser_clear(parser);
4123
		if (ret < 0)
4124
			goto out;
4125
	}
4126 4127

	ret = read;
4128
 out:
4129 4130 4131
	return ret;
}

4132
ssize_t
4133 4134 4135 4136 4137 4138
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);
}

4139
ssize_t
4140 4141 4142 4143 4144 4145
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);
}

4146
static int
4147 4148 4149 4150 4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162 4163 4164
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);
}

4165
static void ftrace_ops_update_code(struct ftrace_ops *ops,
4166
				   struct ftrace_ops_hash *old_hash)
4167
{
4168 4169 4170 4171 4172 4173
	struct ftrace_ops *op;

	if (!ftrace_enabled)
		return;

	if (ops->flags & FTRACE_OPS_FL_ENABLED) {
4174
		ftrace_run_modify_code(ops, FTRACE_UPDATE_CALLS, old_hash);
4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193
		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);
4194 4195
}

4196 4197 4198
static int
ftrace_set_hash(struct ftrace_ops *ops, unsigned char *buf, int len,
		unsigned long ip, int remove, int reset, int enable)
4199
{
4200
	struct ftrace_hash **orig_hash;
4201
	struct ftrace_ops_hash old_hash_ops;
4202
	struct ftrace_hash *old_hash;
4203
	struct ftrace_hash *hash;
4204
	int ret;
4205

4206
	if (unlikely(ftrace_disabled))
4207
		return -ENODEV;
4208

4209
	mutex_lock(&ops->func_hash->regex_lock);
4210

4211
	if (enable)
4212
		orig_hash = &ops->func_hash->filter_hash;
4213
	else
4214
		orig_hash = &ops->func_hash->notrace_hash;
4215

4216 4217 4218 4219 4220
	if (reset)
		hash = alloc_ftrace_hash(FTRACE_HASH_DEFAULT_BITS);
	else
		hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, *orig_hash);

4221 4222 4223 4224
	if (!hash) {
		ret = -ENOMEM;
		goto out_regex_unlock;
	}
4225

4226 4227 4228 4229
	if (buf && !ftrace_match_records(hash, buf, len)) {
		ret = -EINVAL;
		goto out_regex_unlock;
	}
4230 4231 4232 4233 4234
	if (ip) {
		ret = ftrace_match_addr(hash, ip, remove);
		if (ret < 0)
			goto out_regex_unlock;
	}
4235 4236

	mutex_lock(&ftrace_lock);
4237
	old_hash = *orig_hash;
4238 4239
	old_hash_ops.filter_hash = ops->func_hash->filter_hash;
	old_hash_ops.notrace_hash = ops->func_hash->notrace_hash;
4240
	ret = ftrace_hash_move(ops, enable, orig_hash, hash);
4241
	if (!ret) {
4242
		ftrace_ops_update_code(ops, &old_hash_ops);
4243 4244
		free_ftrace_hash_rcu(old_hash);
	}
4245 4246
	mutex_unlock(&ftrace_lock);

4247
 out_regex_unlock:
4248
	mutex_unlock(&ops->func_hash->regex_lock);
4249 4250 4251

	free_ftrace_hash(hash);
	return ret;
4252 4253
}

4254 4255 4256 4257 4258 4259 4260 4261 4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273
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)
{
4274
	ftrace_ops_init(ops);
4275 4276 4277 4278
	return ftrace_set_addr(ops, ip, remove, reset, 1);
}
EXPORT_SYMBOL_GPL(ftrace_set_filter_ip);

4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295
/**
 * 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);

4296 4297 4298 4299 4300 4301 4302
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);
}

4303 4304
/**
 * ftrace_set_filter - set a function to filter on in ftrace
4305 4306 4307 4308 4309 4310 4311 4312
 * @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.
 */
4313
int ftrace_set_filter(struct ftrace_ops *ops, unsigned char *buf,
4314 4315
		       int len, int reset)
{
4316
	ftrace_ops_init(ops);
4317
	return ftrace_set_regex(ops, buf, len, reset, 1);
4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331
}
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.
 */
4332
int ftrace_set_notrace(struct ftrace_ops *ops, unsigned char *buf,
4333 4334
			int len, int reset)
{
4335
	ftrace_ops_init(ops);
4336
	return ftrace_set_regex(ops, buf, len, reset, 0);
4337 4338 4339
}
EXPORT_SYMBOL_GPL(ftrace_set_notrace);
/**
4340
 * ftrace_set_global_filter - set a function to filter on with global tracers
4341 4342 4343 4344 4345 4346 4347
 * @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.
 */
4348
void ftrace_set_global_filter(unsigned char *buf, int len, int reset)
4349
{
4350
	ftrace_set_regex(&global_ops, buf, len, reset, 1);
4351
}
4352
EXPORT_SYMBOL_GPL(ftrace_set_global_filter);
4353

4354
/**
4355
 * ftrace_set_global_notrace - set a function to not trace with global tracers
4356 4357 4358 4359 4360 4361 4362 4363
 * @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.
 */
4364
void ftrace_set_global_notrace(unsigned char *buf, int len, int reset)
4365
{
4366
	ftrace_set_regex(&global_ops, buf, len, reset, 0);
4367
}
4368
EXPORT_SYMBOL_GPL(ftrace_set_global_notrace);
4369

4370 4371 4372 4373 4374 4375 4376
/*
 * 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;

4377 4378 4379
/* Used by function selftest to not test if filter is set */
bool ftrace_filter_param __initdata;

4380 4381
static int __init set_ftrace_notrace(char *str)
{
4382
	ftrace_filter_param = true;
4383
	strlcpy(ftrace_notrace_buf, str, FTRACE_FILTER_SIZE);
4384 4385 4386 4387 4388 4389
	return 1;
}
__setup("ftrace_notrace=", set_ftrace_notrace);

static int __init set_ftrace_filter(char *str)
{
4390
	ftrace_filter_param = true;
4391
	strlcpy(ftrace_filter_buf, str, FTRACE_FILTER_SIZE);
4392 4393 4394 4395
	return 1;
}
__setup("ftrace_filter=", set_ftrace_filter);

4396
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
4397
static char ftrace_graph_buf[FTRACE_FILTER_SIZE] __initdata;
4398
static char ftrace_graph_notrace_buf[FTRACE_FILTER_SIZE] __initdata;
4399
static int ftrace_graph_set_hash(struct ftrace_hash *hash, char *buffer);
4400

4401 4402 4403
static unsigned long save_global_trampoline;
static unsigned long save_global_flags;

4404 4405
static int __init set_graph_function(char *str)
{
4406
	strlcpy(ftrace_graph_buf, str, FTRACE_FILTER_SIZE);
4407 4408 4409 4410
	return 1;
}
__setup("ftrace_graph_filter=", set_graph_function);

4411 4412 4413 4414 4415 4416
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);
4417 4418 4419 4420 4421 4422 4423 4424 4425

static int __init set_graph_max_depth_function(char *str)
{
	if (!str)
		return 0;
	fgraph_max_depth = simple_strtoul(str, NULL, 0);
	return 1;
}
__setup("ftrace_graph_max_depth=", set_graph_max_depth_function);
4426 4427

static void __init set_ftrace_early_graph(char *buf, int enable)
4428 4429 4430
{
	int ret;
	char *func;
4431
	struct ftrace_hash *hash;
4432

4433 4434 4435
	hash = alloc_ftrace_hash(FTRACE_HASH_DEFAULT_BITS);
	if (WARN_ON(!hash))
		return;
4436 4437 4438 4439

	while (buf) {
		func = strsep(&buf, ",");
		/* we allow only one expression at a time */
4440
		ret = ftrace_graph_set_hash(hash, func);
4441 4442 4443 4444
		if (ret)
			printk(KERN_DEBUG "ftrace: function %s not "
					  "traceable\n", func);
	}
4445 4446 4447 4448 4449

	if (enable)
		ftrace_graph_hash = hash;
	else
		ftrace_graph_notrace_hash = hash;
4450 4451 4452
}
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

4453 4454
void __init
ftrace_set_early_filter(struct ftrace_ops *ops, char *buf, int enable)
4455 4456 4457
{
	char *func;

4458 4459
	ftrace_ops_init(ops);

4460 4461
	while (buf) {
		func = strsep(&buf, ",");
4462
		ftrace_set_regex(ops, func, strlen(func), 0, enable);
4463 4464 4465 4466 4467 4468
	}
}

static void __init set_ftrace_early_filters(void)
{
	if (ftrace_filter_buf[0])
4469
		ftrace_set_early_filter(&global_ops, ftrace_filter_buf, 1);
4470
	if (ftrace_notrace_buf[0])
4471
		ftrace_set_early_filter(&global_ops, ftrace_notrace_buf, 0);
4472 4473
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	if (ftrace_graph_buf[0])
4474 4475 4476
		set_ftrace_early_graph(ftrace_graph_buf, 1);
	if (ftrace_graph_notrace_buf[0])
		set_ftrace_early_graph(ftrace_graph_notrace_buf, 0);
4477
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */
4478 4479
}

4480
int ftrace_regex_release(struct inode *inode, struct file *file)
4481 4482
{
	struct seq_file *m = (struct seq_file *)file->private_data;
4483
	struct ftrace_ops_hash old_hash_ops;
4484
	struct ftrace_iterator *iter;
4485
	struct ftrace_hash **orig_hash;
4486
	struct ftrace_hash *old_hash;
4487
	struct trace_parser *parser;
4488
	int filter_hash;
4489
	int ret;
4490 4491 4492 4493 4494 4495 4496

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

4497 4498 4499
	parser = &iter->parser;
	if (trace_parser_loaded(parser)) {
		parser->buffer[parser->idx] = 0;
4500
		ftrace_match_records(iter->hash, parser->buffer, parser->idx);
4501 4502
	}

4503 4504
	trace_parser_put(parser);

4505
	mutex_lock(&iter->ops->func_hash->regex_lock);
4506

4507
	if (file->f_mode & FMODE_WRITE) {
4508 4509 4510
		filter_hash = !!(iter->flags & FTRACE_ITER_FILTER);

		if (filter_hash)
4511
			orig_hash = &iter->ops->func_hash->filter_hash;
4512
		else
4513
			orig_hash = &iter->ops->func_hash->notrace_hash;
4514

4515
		mutex_lock(&ftrace_lock);
4516
		old_hash = *orig_hash;
4517 4518
		old_hash_ops.filter_hash = iter->ops->func_hash->filter_hash;
		old_hash_ops.notrace_hash = iter->ops->func_hash->notrace_hash;
4519 4520
		ret = ftrace_hash_move(iter->ops, filter_hash,
				       orig_hash, iter->hash);
4521
		if (!ret) {
4522
			ftrace_ops_update_code(iter->ops, &old_hash_ops);
4523 4524
			free_ftrace_hash_rcu(old_hash);
		}
4525 4526
		mutex_unlock(&ftrace_lock);
	}
4527

4528
	mutex_unlock(&iter->ops->func_hash->regex_lock);
4529 4530
	free_ftrace_hash(iter->hash);
	kfree(iter);
4531

4532 4533 4534
	return 0;
}

4535
static const struct file_operations ftrace_avail_fops = {
4536 4537 4538
	.open = ftrace_avail_open,
	.read = seq_read,
	.llseek = seq_lseek,
L
Li Zefan 已提交
4539
	.release = seq_release_private,
4540 4541
};

4542 4543 4544 4545 4546 4547 4548
static const struct file_operations ftrace_enabled_fops = {
	.open = ftrace_enabled_open,
	.read = seq_read,
	.llseek = seq_lseek,
	.release = seq_release_private,
};

4549
static const struct file_operations ftrace_filter_fops = {
4550
	.open = ftrace_filter_open,
L
Lai Jiangshan 已提交
4551
	.read = seq_read,
4552
	.write = ftrace_filter_write,
4553
	.llseek = tracing_lseek,
4554
	.release = ftrace_regex_release,
4555 4556
};

4557
static const struct file_operations ftrace_notrace_fops = {
4558
	.open = ftrace_notrace_open,
L
Lai Jiangshan 已提交
4559
	.read = seq_read,
4560
	.write = ftrace_notrace_write,
4561
	.llseek = tracing_lseek,
4562
	.release = ftrace_regex_release,
4563 4564
};

4565 4566 4567 4568
#ifdef CONFIG_FUNCTION_GRAPH_TRACER

static DEFINE_MUTEX(graph_lock);

4569 4570 4571 4572 4573 4574 4575
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,
};
4576

4577 4578
#define FTRACE_GRAPH_EMPTY	((void *)1)

4579
struct ftrace_graph_data {
4580 4581 4582 4583 4584 4585 4586
	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;
	const struct seq_operations	*seq_ops;
	struct trace_parser		parser;
4587 4588
};

4589
static void *
4590
__g_next(struct seq_file *m, loff_t *pos)
4591
{
4592
	struct ftrace_graph_data *fgd = m->private;
4593 4594 4595
	struct ftrace_func_entry *entry = fgd->entry;
	struct hlist_head *head;
	int i, idx = fgd->idx;
4596

4597
	if (*pos >= fgd->hash->count)
4598
		return NULL;
4599 4600 4601 4602 4603 4604 4605 4606 4607 4608 4609 4610 4611 4612 4613 4614 4615 4616 4617

	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;
4618
}
4619

4620 4621 4622 4623 4624
static void *
g_next(struct seq_file *m, void *v, loff_t *pos)
{
	(*pos)++;
	return __g_next(m, pos);
4625 4626 4627 4628
}

static void *g_start(struct seq_file *m, loff_t *pos)
{
4629 4630
	struct ftrace_graph_data *fgd = m->private;

4631 4632
	mutex_lock(&graph_lock);

4633 4634 4635 4636 4637 4638 4639
	if (fgd->type == GRAPH_FILTER_FUNCTION)
		fgd->hash = rcu_dereference_protected(ftrace_graph_hash,
					lockdep_is_held(&graph_lock));
	else
		fgd->hash = rcu_dereference_protected(ftrace_graph_notrace_hash,
					lockdep_is_held(&graph_lock));

4640
	/* Nothing, tell g_show to print all functions are enabled */
4641
	if (ftrace_hash_empty(fgd->hash) && !*pos)
4642
		return FTRACE_GRAPH_EMPTY;
4643

4644 4645
	fgd->idx = 0;
	fgd->entry = NULL;
4646
	return __g_next(m, pos);
4647 4648 4649 4650 4651 4652 4653 4654 4655
}

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

static int g_show(struct seq_file *m, void *v)
{
4656
	struct ftrace_func_entry *entry = v;
4657

4658
	if (!entry)
4659 4660
		return 0;

4661
	if (entry == FTRACE_GRAPH_EMPTY) {
4662 4663
		struct ftrace_graph_data *fgd = m->private;

4664
		if (fgd->type == GRAPH_FILTER_FUNCTION)
4665
			seq_puts(m, "#### all functions enabled ####\n");
4666
		else
4667
			seq_puts(m, "#### no functions disabled ####\n");
4668 4669 4670
		return 0;
	}

4671
	seq_printf(m, "%ps\n", (void *)entry->ip);
4672 4673 4674 4675

	return 0;
}

J
James Morris 已提交
4676
static const struct seq_operations ftrace_graph_seq_ops = {
4677 4678 4679 4680 4681 4682 4683
	.start = g_start,
	.next = g_next,
	.stop = g_stop,
	.show = g_show,
};

static int
4684 4685
__ftrace_graph_open(struct inode *inode, struct file *file,
		    struct ftrace_graph_data *fgd)
4686 4687
{
	int ret = 0;
4688
	struct ftrace_hash *new_hash = NULL;
4689

4690 4691 4692
	if (file->f_mode & FMODE_WRITE) {
		const int size_bits = FTRACE_HASH_DEFAULT_BITS;

4693 4694 4695
		if (trace_parser_get_init(&fgd->parser, FTRACE_BUFF_MAX))
			return -ENOMEM;

4696 4697 4698 4699 4700 4701 4702 4703 4704
		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;
		}
4705 4706
	}

4707
	if (file->f_mode & FMODE_READ) {
4708
		ret = seq_open(file, &ftrace_graph_seq_ops);
4709 4710 4711
		if (!ret) {
			struct seq_file *m = file->private_data;
			m->private = fgd;
4712 4713 4714 4715
		} else {
			/* Failed */
			free_ftrace_hash(new_hash);
			new_hash = NULL;
4716 4717 4718
		}
	} else
		file->private_data = fgd;
4719

4720
out:
4721 4722 4723
	if (ret < 0 && file->f_mode & FMODE_WRITE)
		trace_parser_put(&fgd->parser);

4724
	fgd->new_hash = new_hash;
4725 4726 4727 4728 4729 4730 4731 4732

	/*
	 * All uses of fgd->hash must be taken with the graph_lock
	 * held. The graph_lock is going to be released, so force
	 * fgd->hash to be reinitialized when it is taken again.
	 */
	fgd->hash = NULL;

4733 4734 4735
	return ret;
}

4736 4737 4738 4739
static int
ftrace_graph_open(struct inode *inode, struct file *file)
{
	struct ftrace_graph_data *fgd;
4740
	int ret;
4741 4742 4743 4744 4745 4746 4747 4748

	if (unlikely(ftrace_disabled))
		return -ENODEV;

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

4749 4750
	mutex_lock(&graph_lock);

4751 4752
	fgd->hash = rcu_dereference_protected(ftrace_graph_hash,
					lockdep_is_held(&graph_lock));
4753
	fgd->type = GRAPH_FILTER_FUNCTION;
4754 4755
	fgd->seq_ops = &ftrace_graph_seq_ops;

4756 4757 4758 4759 4760 4761
	ret = __ftrace_graph_open(inode, file, fgd);
	if (ret < 0)
		kfree(fgd);

	mutex_unlock(&graph_lock);
	return ret;
4762 4763
}

4764 4765 4766 4767
static int
ftrace_graph_notrace_open(struct inode *inode, struct file *file)
{
	struct ftrace_graph_data *fgd;
4768
	int ret;
4769 4770 4771 4772 4773 4774 4775 4776

	if (unlikely(ftrace_disabled))
		return -ENODEV;

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

4777 4778
	mutex_lock(&graph_lock);

4779 4780
	fgd->hash = rcu_dereference_protected(ftrace_graph_notrace_hash,
					lockdep_is_held(&graph_lock));
4781
	fgd->type = GRAPH_FILTER_NOTRACE;
4782 4783
	fgd->seq_ops = &ftrace_graph_seq_ops;

4784 4785 4786 4787 4788 4789
	ret = __ftrace_graph_open(inode, file, fgd);
	if (ret < 0)
		kfree(fgd);

	mutex_unlock(&graph_lock);
	return ret;
4790 4791
}

4792 4793 4794
static int
ftrace_graph_release(struct inode *inode, struct file *file)
{
4795
	struct ftrace_graph_data *fgd;
4796 4797 4798
	struct ftrace_hash *old_hash, *new_hash;
	struct trace_parser *parser;
	int ret = 0;
4799

4800 4801 4802
	if (file->f_mode & FMODE_READ) {
		struct seq_file *m = file->private_data;

4803
		fgd = m->private;
4804
		seq_release(inode, file);
4805
	} else {
4806
		fgd = file->private_data;
4807 4808
	}

4809 4810 4811 4812 4813 4814 4815 4816 4817 4818 4819 4820 4821 4822 4823 4824 4825 4826 4827 4828 4829 4830 4831 4832 4833 4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846 4847 4848

	if (file->f_mode & FMODE_WRITE) {

		parser = &fgd->parser;

		if (trace_parser_loaded((parser))) {
			parser->buffer[parser->idx] = 0;
			ret = ftrace_graph_set_hash(fgd->new_hash,
						    parser->buffer);
		}

		trace_parser_put(parser);

		new_hash = __ftrace_hash_move(fgd->new_hash);
		if (!new_hash) {
			ret = -ENOMEM;
			goto out;
		}

		mutex_lock(&graph_lock);

		if (fgd->type == GRAPH_FILTER_FUNCTION) {
			old_hash = rcu_dereference_protected(ftrace_graph_hash,
					lockdep_is_held(&graph_lock));
			rcu_assign_pointer(ftrace_graph_hash, new_hash);
		} else {
			old_hash = rcu_dereference_protected(ftrace_graph_notrace_hash,
					lockdep_is_held(&graph_lock));
			rcu_assign_pointer(ftrace_graph_notrace_hash, new_hash);
		}

		mutex_unlock(&graph_lock);

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

		free_ftrace_hash(old_hash);
	}

 out:
4849 4850 4851
	kfree(fgd->new_hash);
	kfree(fgd);

4852
	return ret;
4853 4854
}

4855
static int
4856
ftrace_graph_set_hash(struct ftrace_hash *hash, char *buffer)
4857
{
4858
	struct ftrace_glob func_g;
4859 4860
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
4861
	struct ftrace_func_entry *entry;
4862
	int fail = 1;
4863
	int not;
4864

4865
	/* decode regex */
4866 4867
	func_g.type = filter_parse_regex(buffer, strlen(buffer),
					 &func_g.search, &not);
4868

4869
	func_g.len = strlen(func_g.search);
4870

4871
	mutex_lock(&ftrace_lock);
4872 4873 4874 4875 4876 4877

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

4878 4879
	do_for_each_ftrace_rec(pg, rec) {

4880 4881 4882
		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

D
Dmitry Safonov 已提交
4883
		if (ftrace_match_record(rec, &func_g, NULL, 0)) {
4884
			entry = ftrace_lookup_ip(hash, rec->ip);
4885 4886 4887

			if (!not) {
				fail = 0;
4888 4889 4890 4891 4892

				if (entry)
					continue;
				if (add_hash_entry(hash, rec->ip) < 0)
					goto out;
4893
			} else {
4894 4895
				if (entry) {
					free_hash_entry(hash, entry);
4896 4897 4898
					fail = 0;
				}
			}
4899
		}
4900
	} while_for_each_ftrace_rec();
4901
out:
4902
	mutex_unlock(&ftrace_lock);
4903

4904 4905 4906 4907
	if (fail)
		return -EINVAL;

	return 0;
4908 4909 4910 4911 4912 4913
}

static ssize_t
ftrace_graph_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
4914
	ssize_t read, ret = 0;
4915
	struct ftrace_graph_data *fgd = file->private_data;
4916
	struct trace_parser *parser;
4917

4918
	if (!cnt)
4919 4920
		return 0;

4921 4922 4923 4924 4925 4926
	/* Read mode uses seq functions */
	if (file->f_mode & FMODE_READ) {
		struct seq_file *m = file->private_data;
		fgd = m->private;
	}

4927
	parser = &fgd->parser;
4928

4929
	read = trace_get_user(parser, ubuf, cnt, ppos);
4930

4931 4932
	if (read >= 0 && trace_parser_loaded(parser) &&
	    !trace_parser_cont(parser)) {
4933

4934
		ret = ftrace_graph_set_hash(fgd->new_hash,
4935 4936
					    parser->buffer);
		trace_parser_clear(parser);
4937 4938
	}

4939 4940
	if (!ret)
		ret = read;
4941

4942 4943 4944 4945
	return ret;
}

static const struct file_operations ftrace_graph_fops = {
4946 4947 4948
	.open		= ftrace_graph_open,
	.read		= seq_read,
	.write		= ftrace_graph_write,
4949
	.llseek		= tracing_lseek,
4950
	.release	= ftrace_graph_release,
4951
};
4952 4953 4954 4955 4956

static const struct file_operations ftrace_graph_notrace_fops = {
	.open		= ftrace_graph_notrace_open,
	.read		= seq_read,
	.write		= ftrace_graph_write,
4957
	.llseek		= tracing_lseek,
4958 4959
	.release	= ftrace_graph_release,
};
4960 4961
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

4962 4963 4964 4965 4966 4967 4968 4969 4970 4971 4972 4973 4974 4975 4976 4977 4978 4979 4980 4981 4982 4983 4984 4985 4986 4987 4988 4989 4990 4991
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);
}

4992
static __init int ftrace_init_dyn_tracefs(struct dentry *d_tracer)
4993 4994
{

4995 4996
	trace_create_file("available_filter_functions", 0444,
			d_tracer, NULL, &ftrace_avail_fops);
4997

4998 4999 5000
	trace_create_file("enabled_functions", 0444,
			d_tracer, NULL, &ftrace_enabled_fops);

5001
	ftrace_create_filter_files(&global_ops, d_tracer);
5002

5003
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
5004
	trace_create_file("set_graph_function", 0444, d_tracer,
5005 5006
				    NULL,
				    &ftrace_graph_fops);
5007 5008 5009
	trace_create_file("set_graph_notrace", 0444, d_tracer,
				    NULL,
				    &ftrace_graph_notrace_fops);
5010 5011
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

5012 5013 5014
	return 0;
}

5015
static int ftrace_cmp_ips(const void *a, const void *b)
5016
{
5017 5018
	const unsigned long *ipa = a;
	const unsigned long *ipb = b;
5019

5020 5021 5022 5023 5024 5025 5026
	if (*ipa > *ipb)
		return 1;
	if (*ipa < *ipb)
		return -1;
	return 0;
}

5027
static int ftrace_process_locs(struct module *mod,
5028
			       unsigned long *start,
5029 5030
			       unsigned long *end)
{
5031
	struct ftrace_page *start_pg;
5032
	struct ftrace_page *pg;
5033
	struct dyn_ftrace *rec;
5034
	unsigned long count;
5035 5036
	unsigned long *p;
	unsigned long addr;
5037
	unsigned long flags = 0; /* Shut up gcc */
5038 5039 5040 5041 5042 5043 5044
	int ret = -ENOMEM;

	count = end - start;

	if (!count)
		return 0;

5045
	sort(start, count, sizeof(*start),
5046
	     ftrace_cmp_ips, NULL);
5047

5048 5049
	start_pg = ftrace_allocate_pages(count);
	if (!start_pg)
5050
		return -ENOMEM;
5051

S
Steven Rostedt 已提交
5052
	mutex_lock(&ftrace_lock);
5053

5054 5055 5056 5057 5058
	/*
	 * 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.
	 */
5059 5060 5061
	if (!mod) {
		WARN_ON(ftrace_pages || ftrace_pages_start);
		/* First initialization */
5062
		ftrace_pages = ftrace_pages_start = start_pg;
5063
	} else {
5064
		if (!ftrace_pages)
5065
			goto out;
5066

5067 5068 5069 5070
		if (WARN_ON(ftrace_pages->next)) {
			/* Hmm, we have free pages? */
			while (ftrace_pages->next)
				ftrace_pages = ftrace_pages->next;
5071
		}
5072

5073
		ftrace_pages->next = start_pg;
5074 5075
	}

5076
	p = start;
5077
	pg = start_pg;
5078 5079
	while (p < end) {
		addr = ftrace_call_adjust(*p++);
5080 5081 5082 5083 5084 5085 5086 5087
		/*
		 * 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;
5088 5089 5090 5091 5092 5093 5094 5095 5096 5097

		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;
5098 5099
	}

5100 5101 5102 5103 5104 5105
	/* We should have used all pages */
	WARN_ON(pg->next);

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

5106
	/*
5107 5108 5109 5110 5111 5112
	 * 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.
5113
	 */
5114 5115
	if (!mod)
		local_irq_save(flags);
5116
	ftrace_update_code(mod, start_pg);
5117 5118
	if (!mod)
		local_irq_restore(flags);
5119 5120
	ret = 0;
 out:
S
Steven Rostedt 已提交
5121
	mutex_unlock(&ftrace_lock);
5122

5123
	return ret;
5124 5125
}

5126
#ifdef CONFIG_MODULES
5127 5128 5129

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

5130 5131 5132 5133 5134 5135 5136 5137 5138 5139 5140 5141 5142
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;
}

5143
void ftrace_release_mod(struct module *mod)
5144 5145
{
	struct dyn_ftrace *rec;
5146
	struct ftrace_page **last_pg;
5147
	struct ftrace_page *pg;
5148
	int order;
5149

5150 5151
	mutex_lock(&ftrace_lock);

5152
	if (ftrace_disabled)
5153
		goto out_unlock;
5154

5155 5156 5157 5158 5159 5160 5161
	/*
	 * 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];
5162
		if (within_module_core(rec->ip, mod)) {
5163
			/*
5164 5165
			 * As core pages are first, the first
			 * page should never be a module page.
5166
			 */
5167 5168 5169 5170 5171 5172 5173 5174
			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;
5175 5176 5177
			order = get_count_order(pg->size / ENTRIES_PER_PAGE);
			free_pages((unsigned long)pg->records, order);
			kfree(pg);
5178 5179 5180
		} else
			last_pg = &pg->next;
	}
5181
 out_unlock:
5182 5183 5184
	mutex_unlock(&ftrace_lock);
}

5185
void ftrace_module_enable(struct module *mod)
5186 5187 5188 5189 5190 5191 5192 5193 5194 5195 5196 5197 5198 5199 5200 5201 5202 5203 5204 5205 5206 5207 5208 5209 5210 5211 5212 5213 5214 5215 5216 5217 5218 5219 5220 5221 5222 5223 5224 5225 5226 5227 5228 5229 5230 5231 5232 5233 5234 5235 5236 5237 5238 5239 5240 5241 5242 5243 5244 5245 5246 5247 5248 5249 5250 5251 5252 5253
{
	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);
}

5254
void ftrace_module_init(struct module *mod)
5255
{
5256
	if (ftrace_disabled || !mod->num_ftrace_callsites)
5257
		return;
5258

5259 5260
	ftrace_process_locs(mod, mod->ftrace_callsites,
			    mod->ftrace_callsites + mod->num_ftrace_callsites);
5261
}
5262 5263
#endif /* CONFIG_MODULES */

5264 5265
void __init ftrace_init(void)
{
5266 5267
	extern unsigned long __start_mcount_loc[];
	extern unsigned long __stop_mcount_loc[];
5268
	unsigned long count, flags;
5269 5270 5271
	int ret;

	local_irq_save(flags);
5272
	ret = ftrace_dyn_arch_init();
5273
	local_irq_restore(flags);
5274
	if (ret)
5275 5276 5277
		goto failed;

	count = __stop_mcount_loc - __start_mcount_loc;
5278 5279
	if (!count) {
		pr_info("ftrace: No functions to be traced?\n");
5280
		goto failed;
5281 5282 5283 5284
	}

	pr_info("ftrace: allocating %ld entries in %ld pages\n",
		count, count / ENTRIES_PER_PAGE + 1);
5285 5286 5287

	last_ftrace_enabled = ftrace_enabled = 1;

5288
	ret = ftrace_process_locs(NULL,
5289
				  __start_mcount_loc,
5290 5291
				  __stop_mcount_loc);

5292 5293
	set_ftrace_early_filters();

5294 5295 5296 5297 5298
	return;
 failed:
	ftrace_disabled = 1;
}

5299 5300 5301 5302 5303 5304 5305
/* 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)
{
5306 5307 5308 5309 5310 5311 5312 5313 5314 5315 5316

/*
 * 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
5317 5318 5319
	/* Currently, only non dynamic ops can have a trampoline */
	if (ops->flags & FTRACE_OPS_FL_DYNAMIC)
		return;
5320
#endif
5321 5322 5323 5324

	arch_ftrace_update_trampoline(ops);
}

5325
#else
5326

5327
static struct ftrace_ops global_ops = {
5328
	.func			= ftrace_stub,
5329 5330 5331
	.flags			= FTRACE_OPS_FL_RECURSION_SAFE |
				  FTRACE_OPS_FL_INITIALIZED |
				  FTRACE_OPS_FL_PID,
5332 5333
};

5334 5335 5336 5337 5338
static int __init ftrace_nodyn_init(void)
{
	ftrace_enabled = 1;
	return 0;
}
5339
core_initcall(ftrace_nodyn_init);
5340

5341
static inline int ftrace_init_dyn_tracefs(struct dentry *d_tracer) { return 0; }
5342
static inline void ftrace_startup_enable(int command) { }
5343
static inline void ftrace_startup_all(int command) { }
5344
/* Keep as macros so we do not need to define the commands */
5345 5346 5347 5348 5349 5350
# define ftrace_startup(ops, command)					\
	({								\
		int ___ret = __register_ftrace_function(ops);		\
		if (!___ret)						\
			(ops)->flags |= FTRACE_OPS_FL_ENABLED;		\
		___ret;							\
5351
	})
5352 5353 5354 5355 5356 5357 5358
# define ftrace_shutdown(ops, command)					\
	({								\
		int ___ret = __unregister_ftrace_function(ops);		\
		if (!___ret)						\
			(ops)->flags &= ~FTRACE_OPS_FL_ENABLED;		\
		___ret;							\
	})
5359

I
Ingo Molnar 已提交
5360 5361
# define ftrace_startup_sysctl()	do { } while (0)
# define ftrace_shutdown_sysctl()	do { } while (0)
5362 5363

static inline int
5364
ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip, void *regs)
5365 5366 5367 5368
{
	return 1;
}

5369 5370 5371 5372
static void ftrace_update_trampoline(struct ftrace_ops *ops)
{
}

5373 5374
#endif /* CONFIG_DYNAMIC_FTRACE */

5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 5390 5391 5392 5393 5394 5395 5396 5397
__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;
}

5398 5399
static inline void
__ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
5400
		       struct ftrace_ops *ignored, struct pt_regs *regs)
5401
{
5402
	struct ftrace_ops *op;
5403
	int bit;
5404

5405 5406 5407
	bit = trace_test_and_set_recursion(TRACE_LIST_START, TRACE_LIST_MAX);
	if (bit < 0)
		return;
5408

5409 5410 5411 5412 5413
	/*
	 * Some of the ops may be dynamically allocated,
	 * they must be freed after a synchronize_sched().
	 */
	preempt_disable_notrace();
5414

5415
	do_for_each_ftrace_op(op, ftrace_ops_list) {
5416 5417 5418 5419 5420 5421 5422 5423 5424 5425 5426 5427 5428 5429
		/*
		 * 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)) {
		    
5430 5431
			if (FTRACE_WARN_ON(!op->func)) {
				pr_warn("op=%p %pS\n", op, op);
5432 5433
				goto out;
			}
5434
			op->func(ip, parent_ip, op, regs);
5435
		}
5436
	} while_for_each_ftrace_op(op);
5437
out:
5438
	preempt_enable_notrace();
5439
	trace_clear_recursion(bit);
5440 5441
}

5442 5443 5444 5445 5446
/*
 * 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.
5447 5448 5449
 * 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
5450 5451
 * being NULL, or CONFIG_DYNAMIC_FTRACE_WITH_REGS.
 * Note, CONFIG_DYNAMIC_FTRACE_WITH_REGS expects a full regs to be saved.
5452
 * An architecture can pass partial regs with ftrace_ops and still
L
Li Bin 已提交
5453
 * set the ARCH_SUPPORTS_FTRACE_OPS.
5454 5455 5456
 */
#if ARCH_SUPPORTS_FTRACE_OPS
static void ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
5457
				 struct ftrace_ops *op, struct pt_regs *regs)
5458
{
5459
	__ftrace_ops_list_func(ip, parent_ip, NULL, regs);
5460 5461 5462 5463
}
#else
static void ftrace_ops_no_ops(unsigned long ip, unsigned long parent_ip)
{
5464
	__ftrace_ops_list_func(ip, parent_ip, NULL, NULL);
5465 5466 5467
}
#endif

5468 5469
/*
 * If there's only one function registered but it does not support
5470 5471
 * recursion, needs RCU protection and/or requires per cpu handling, then
 * this function will be called by the mcount trampoline.
5472
 */
5473
static void ftrace_ops_assist_func(unsigned long ip, unsigned long parent_ip,
5474 5475 5476 5477
				   struct ftrace_ops *op, struct pt_regs *regs)
{
	int bit;

5478 5479 5480
	if ((op->flags & FTRACE_OPS_FL_RCU) && !rcu_is_watching())
		return;

5481 5482 5483 5484
	bit = trace_test_and_set_recursion(TRACE_LIST_START, TRACE_LIST_MAX);
	if (bit < 0)
		return;

5485
	preempt_disable_notrace();
5486

5487 5488 5489 5490 5491 5492
	if (!(op->flags & FTRACE_OPS_FL_PER_CPU) ||
	    !ftrace_function_local_disabled(op)) {
		op->func(ip, parent_ip, op, regs);
	}

	preempt_enable_notrace();
5493 5494 5495
	trace_clear_recursion(bit);
}

5496 5497 5498 5499 5500 5501 5502
/**
 * 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
5503
 * ftrace_ops_assist_func() instead.
5504 5505 5506 5507 5508 5509
 *
 * Returns the function that the trampoline should call for @ops.
 */
ftrace_func_t ftrace_ops_get_func(struct ftrace_ops *ops)
{
	/*
5510 5511
	 * 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.
5512
	 */
5513 5514 5515
	if (!(ops->flags & FTRACE_OPS_FL_RECURSION_SAFE) ||
	    ops->flags & (FTRACE_OPS_FL_RCU | FTRACE_OPS_FL_PER_CPU))
		return ftrace_ops_assist_func;
5516 5517 5518 5519

	return ops->func;
}

5520 5521 5522
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
5523
{
5524 5525
	struct trace_array *tr = data;
	struct trace_pid_list *pid_list;
S
Steven Rostedt 已提交
5526

5527
	pid_list = rcu_dereference_sched(tr->function_pids);
5528

5529 5530
	this_cpu_write(tr->trace_buffer.data->ftrace_ignore_pid,
		       trace_ignore_this_task(pid_list, next));
S
Steven Rostedt 已提交
5531 5532
}

5533
static void clear_ftrace_pids(struct trace_array *tr)
5534
{
5535 5536
	struct trace_pid_list *pid_list;
	int cpu;
5537

5538 5539 5540 5541
	pid_list = rcu_dereference_protected(tr->function_pids,
					     lockdep_is_held(&ftrace_lock));
	if (!pid_list)
		return;
5542

5543
	unregister_trace_sched_switch(ftrace_filter_pid_sched_switch_probe, tr);
5544

5545 5546
	for_each_possible_cpu(cpu)
		per_cpu_ptr(tr->trace_buffer.data, cpu)->ftrace_ignore_pid = false;
S
Steven Rostedt 已提交
5547

5548
	rcu_assign_pointer(tr->function_pids, NULL);
S
Steven Rostedt 已提交
5549

5550 5551
	/* Wait till all users are no longer using pid filtering */
	synchronize_sched();
5552

5553
	trace_free_pid_list(pid_list);
5554 5555
}

5556
static void ftrace_pid_reset(struct trace_array *tr)
5557
{
5558
	mutex_lock(&ftrace_lock);
5559
	clear_ftrace_pids(tr);
S
Steven Rostedt 已提交
5560

5561
	ftrace_update_pid_func();
5562
	ftrace_startup_all(0);
5563 5564 5565 5566

	mutex_unlock(&ftrace_lock);
}

5567 5568
/* Greater than any max PID */
#define FTRACE_NO_PIDS		(void *)(PID_MAX_LIMIT + 1)
5569

5570
static void *fpid_start(struct seq_file *m, loff_t *pos)
5571
	__acquires(RCU)
5572
{
5573 5574 5575
	struct trace_pid_list *pid_list;
	struct trace_array *tr = m->private;

5576
	mutex_lock(&ftrace_lock);
5577 5578 5579
	rcu_read_lock_sched();

	pid_list = rcu_dereference_sched(tr->function_pids);
5580

5581 5582
	if (!pid_list)
		return !(*pos) ? FTRACE_NO_PIDS : NULL;
5583

5584
	return trace_pid_start(pid_list, pos);
5585 5586 5587 5588
}

static void *fpid_next(struct seq_file *m, void *v, loff_t *pos)
{
5589 5590 5591 5592
	struct trace_array *tr = m->private;
	struct trace_pid_list *pid_list = rcu_dereference_sched(tr->function_pids);

	if (v == FTRACE_NO_PIDS)
5593 5594
		return NULL;

5595
	return trace_pid_next(pid_list, v, pos);
5596 5597 5598
}

static void fpid_stop(struct seq_file *m, void *p)
5599
	__releases(RCU)
5600
{
5601
	rcu_read_unlock_sched();
5602 5603 5604 5605 5606
	mutex_unlock(&ftrace_lock);
}

static int fpid_show(struct seq_file *m, void *v)
{
5607
	if (v == FTRACE_NO_PIDS) {
5608
		seq_puts(m, "no pid\n");
5609 5610 5611
		return 0;
	}

5612
	return trace_pid_show(m, v);
5613 5614 5615 5616 5617 5618 5619 5620 5621 5622 5623 5624
}

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)
{
5625 5626
	struct trace_array *tr = inode->i_private;
	struct seq_file *m;
5627 5628
	int ret = 0;

5629 5630 5631
	if (trace_array_get(tr) < 0)
		return -ENODEV;

5632 5633
	if ((file->f_mode & FMODE_WRITE) &&
	    (file->f_flags & O_TRUNC))
5634
		ftrace_pid_reset(tr);
5635

5636 5637 5638 5639 5640 5641 5642 5643
	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;
	}
5644 5645 5646 5647

	return ret;
}

5648 5649 5650 5651 5652 5653 5654 5655 5656 5657 5658 5659 5660 5661 5662 5663
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));
}

5664 5665 5666 5667
static ssize_t
ftrace_pid_write(struct file *filp, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
5668 5669 5670 5671 5672
	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;
5673

5674 5675 5676 5677 5678 5679 5680 5681 5682 5683 5684
	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;
5685

5686
	rcu_assign_pointer(tr->function_pids, pid_list);
5687

5688 5689 5690 5691 5692 5693 5694
	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);
	}
5695

5696
	/*
5697 5698 5699
	 * 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.
5700
	 */
5701
	on_each_cpu(ignore_task_cpu, tr, 1);
5702

5703 5704 5705 5706
	ftrace_update_pid_func();
	ftrace_startup_all(0);
 out:
	mutex_unlock(&ftrace_lock);
5707

5708 5709
	if (ret > 0)
		*ppos += ret;
5710

5711
	return ret;
5712
}
5713

5714 5715 5716
static int
ftrace_pid_release(struct inode *inode, struct file *file)
{
5717
	struct trace_array *tr = inode->i_private;
5718

5719 5720 5721
	trace_array_put(tr);

	return seq_release(inode, file);
5722 5723
}

5724
static const struct file_operations ftrace_pid_fops = {
5725 5726 5727
	.open		= ftrace_pid_open,
	.write		= ftrace_pid_write,
	.read		= seq_read,
5728
	.llseek		= tracing_lseek,
5729
	.release	= ftrace_pid_release,
5730 5731
};

5732
void ftrace_init_tracefs(struct trace_array *tr, struct dentry *d_tracer)
5733
{
5734
	trace_create_file("set_ftrace_pid", 0644, d_tracer,
5735
			    tr, &ftrace_pid_fops);
5736 5737
}

5738 5739 5740 5741 5742 5743 5744 5745 5746 5747
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
5748
/**
5749
 * ftrace_kill - kill ftrace
S
Steven Rostedt 已提交
5750 5751 5752 5753 5754
 *
 * 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.
 */
5755
void ftrace_kill(void)
S
Steven Rostedt 已提交
5756 5757 5758 5759 5760 5761
{
	ftrace_disabled = 1;
	ftrace_enabled = 0;
	clear_ftrace_function();
}

5762 5763 5764 5765 5766 5767 5768 5769
/**
 * Test if ftrace is dead or not.
 */
int ftrace_is_dead(void)
{
	return ftrace_disabled;
}

5770
/**
5771 5772
 * register_ftrace_function - register a function for profiling
 * @ops - ops structure that holds the function for profiling.
5773
 *
5774 5775 5776 5777 5778 5779
 * 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.
5780
 */
5781
int register_ftrace_function(struct ftrace_ops *ops)
5782
{
5783
	int ret = -1;
5784

5785 5786
	ftrace_ops_init(ops);

S
Steven Rostedt 已提交
5787
	mutex_lock(&ftrace_lock);
5788

5789
	ret = ftrace_startup(ops, 0);
5790

S
Steven Rostedt 已提交
5791
	mutex_unlock(&ftrace_lock);
5792

5793
	return ret;
5794
}
5795
EXPORT_SYMBOL_GPL(register_ftrace_function);
5796 5797

/**
5798
 * unregister_ftrace_function - unregister a function for profiling.
5799 5800 5801 5802 5803 5804 5805 5806
 * @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
5807
	mutex_lock(&ftrace_lock);
5808
	ret = ftrace_shutdown(ops, 0);
S
Steven Rostedt 已提交
5809
	mutex_unlock(&ftrace_lock);
5810 5811 5812

	return ret;
}
5813
EXPORT_SYMBOL_GPL(unregister_ftrace_function);
5814

I
Ingo Molnar 已提交
5815
int
5816
ftrace_enable_sysctl(struct ctl_table *table, int write,
5817
		     void __user *buffer, size_t *lenp,
5818 5819
		     loff_t *ppos)
{
5820
	int ret = -ENODEV;
5821

S
Steven Rostedt 已提交
5822
	mutex_lock(&ftrace_lock);
5823

5824 5825 5826 5827
	if (unlikely(ftrace_disabled))
		goto out;

	ret = proc_dointvec(table, write, buffer, lenp, ppos);
5828

5829
	if (ret || !write || (last_ftrace_enabled == !!ftrace_enabled))
5830 5831
		goto out;

5832
	last_ftrace_enabled = !!ftrace_enabled;
5833 5834 5835 5836

	if (ftrace_enabled) {

		/* we are starting ftrace again */
5837 5838
		if (ftrace_ops_list != &ftrace_list_end)
			update_ftrace_function();
5839

5840 5841
		ftrace_startup_sysctl();

5842 5843 5844 5845 5846 5847 5848 5849
	} else {
		/* stopping ftrace calls (just send to ftrace_stub) */
		ftrace_trace_function = ftrace_stub;

		ftrace_shutdown_sysctl();
	}

 out:
S
Steven Rostedt 已提交
5850
	mutex_unlock(&ftrace_lock);
5851
	return ret;
5852
}
5853

5854
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
5855

5856 5857 5858 5859
static struct ftrace_ops graph_ops = {
	.func			= ftrace_stub,
	.flags			= FTRACE_OPS_FL_RECURSION_SAFE |
				   FTRACE_OPS_FL_INITIALIZED |
5860
				   FTRACE_OPS_FL_PID |
5861 5862 5863
				   FTRACE_OPS_FL_STUB,
#ifdef FTRACE_GRAPH_TRAMP_ADDR
	.trampoline		= FTRACE_GRAPH_TRAMP_ADDR,
5864
	/* trampoline_size is only needed for dynamically allocated tramps */
5865 5866 5867 5868
#endif
	ASSIGN_OPS_HASH(graph_ops, &global_ops.local_hash)
};

5869 5870 5871 5872 5873 5874 5875 5876 5877 5878
void ftrace_graph_sleep_time_control(bool enable)
{
	fgraph_sleep_time = enable;
}

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

5879 5880 5881 5882 5883
int ftrace_graph_entry_stub(struct ftrace_graph_ent *trace)
{
	return 0;
}

5884 5885 5886
/* The callbacks that hook a function */
trace_func_graph_ret_t ftrace_graph_return =
			(trace_func_graph_ret_t)ftrace_stub;
5887
trace_func_graph_ent_t ftrace_graph_entry = ftrace_graph_entry_stub;
5888
static trace_func_graph_ent_t __ftrace_graph_entry = ftrace_graph_entry_stub;
5889 5890 5891 5892 5893 5894 5895 5896 5897 5898 5899 5900 5901 5902 5903 5904 5905 5906 5907 5908 5909

/* 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;
		}
	}

5910
	read_lock(&tasklist_lock);
5911 5912 5913 5914 5915 5916 5917
	do_each_thread(g, t) {
		if (start == end) {
			ret = -EAGAIN;
			goto unlock;
		}

		if (t->ret_stack == NULL) {
5918
			atomic_set(&t->tracing_graph_pause, 0);
5919
			atomic_set(&t->trace_overrun, 0);
5920 5921 5922 5923
			t->curr_ret_stack = -1;
			/* Make sure the tasks see the -1 first: */
			smp_wmb();
			t->ret_stack = ret_stack_list[start++];
5924 5925 5926 5927
		}
	} while_each_thread(g, t);

unlock:
5928
	read_unlock(&tasklist_lock);
5929 5930 5931 5932 5933 5934
free:
	for (i = start; i < end; i++)
		kfree(ret_stack_list[i]);
	return ret;
}

5935
static void
5936
ftrace_graph_probe_sched_switch(void *ignore, bool preempt,
5937
			struct task_struct *prev, struct task_struct *next)
5938 5939 5940 5941
{
	unsigned long long timestamp;
	int index;

5942 5943 5944 5945
	/*
	 * Does the user want to count the time a function was asleep.
	 * If so, do not update the time stamps.
	 */
5946
	if (fgraph_sleep_time)
5947 5948
		return;

5949 5950 5951 5952 5953 5954 5955 5956 5957 5958 5959 5960 5961 5962 5963 5964 5965 5966
	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;
}

5967
/* Allocate a return stack for each task */
5968
static int start_graph_tracing(void)
5969 5970
{
	struct ftrace_ret_stack **ret_stack_list;
5971
	int ret, cpu;
5972 5973 5974 5975 5976 5977 5978 5979

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

	if (!ret_stack_list)
		return -ENOMEM;

5980
	/* The cpu_boot init_task->ret_stack will never be freed */
5981 5982
	for_each_online_cpu(cpu) {
		if (!idle_task(cpu)->ret_stack)
5983
			ftrace_graph_init_idle_task(idle_task(cpu), cpu);
5984
	}
5985

5986 5987 5988 5989
	do {
		ret = alloc_retstack_tasklist(ret_stack_list);
	} while (ret == -EAGAIN);

5990
	if (!ret) {
5991
		ret = register_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
5992 5993 5994 5995 5996
		if (ret)
			pr_info("ftrace_graph: Couldn't activate tracepoint"
				" probe to kernel_sched_switch\n");
	}

5997 5998 5999 6000
	kfree(ret_stack_list);
	return ret;
}

6001 6002 6003 6004 6005 6006 6007 6008 6009 6010 6011 6012 6013 6014 6015 6016 6017 6018 6019 6020 6021
/*
 * 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;
}

6022 6023 6024 6025 6026 6027 6028 6029 6030 6031 6032 6033 6034 6035 6036 6037
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)
{
6038 6039 6040 6041 6042 6043 6044 6045 6046 6047 6048 6049 6050 6051 6052 6053 6054 6055 6056
	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)
6057
		ftrace_graph_entry = ftrace_graph_entry_test;
6058 6059
	else
		ftrace_graph_entry = __ftrace_graph_entry;
6060 6061
}

6062 6063 6064 6065
static struct notifier_block ftrace_suspend_notifier = {
	.notifier_call = ftrace_suspend_notifier_call,
};

6066 6067
int register_ftrace_graph(trace_func_graph_ret_t retfunc,
			trace_func_graph_ent_t entryfunc)
6068
{
6069 6070
	int ret = 0;

S
Steven Rostedt 已提交
6071
	mutex_lock(&ftrace_lock);
6072

6073
	/* we currently allow only one tracer registered at a time */
6074
	if (ftrace_graph_active) {
6075 6076 6077 6078
		ret = -EBUSY;
		goto out;
	}

6079 6080
	register_pm_notifier(&ftrace_suspend_notifier);

6081
	ftrace_graph_active++;
6082
	ret = start_graph_tracing();
6083
	if (ret) {
6084
		ftrace_graph_active--;
6085 6086
		goto out;
	}
6087

6088
	ftrace_graph_return = retfunc;
6089 6090 6091 6092 6093 6094 6095 6096 6097 6098

	/*
	 * 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();
6099

6100
	ret = ftrace_startup(&graph_ops, FTRACE_START_FUNC_RET);
6101
out:
S
Steven Rostedt 已提交
6102
	mutex_unlock(&ftrace_lock);
6103
	return ret;
6104 6105
}

6106
void unregister_ftrace_graph(void)
6107
{
S
Steven Rostedt 已提交
6108
	mutex_lock(&ftrace_lock);
6109

6110
	if (unlikely(!ftrace_graph_active))
6111 6112
		goto out;

6113
	ftrace_graph_active--;
6114
	ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub;
6115
	ftrace_graph_entry = ftrace_graph_entry_stub;
6116
	__ftrace_graph_entry = ftrace_graph_entry_stub;
6117
	ftrace_shutdown(&graph_ops, FTRACE_STOP_FUNC_RET);
6118
	unregister_pm_notifier(&ftrace_suspend_notifier);
6119
	unregister_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
6120

6121 6122 6123 6124 6125 6126 6127 6128 6129 6130 6131
#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

6132
 out:
S
Steven Rostedt 已提交
6133
	mutex_unlock(&ftrace_lock);
6134
}
6135

6136 6137 6138 6139 6140 6141 6142 6143
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
6144
	/* make curr_ret_stack visible before we add the ret_stack */
6145 6146 6147 6148 6149 6150 6151 6152 6153 6154 6155 6156 6157 6158 6159 6160 6161 6162 6163 6164 6165 6166 6167 6168 6169 6170 6171 6172 6173 6174 6175 6176 6177 6178
	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);
	}
}

6179
/* Allocate a return stack for newly created task */
6180
void ftrace_graph_init_task(struct task_struct *t)
6181
{
6182 6183
	/* Make sure we do not use the parent ret_stack */
	t->ret_stack = NULL;
6184
	t->curr_ret_stack = -1;
6185

6186
	if (ftrace_graph_active) {
6187 6188 6189
		struct ftrace_ret_stack *ret_stack;

		ret_stack = kmalloc(FTRACE_RETFUNC_DEPTH
6190 6191
				* sizeof(struct ftrace_ret_stack),
				GFP_KERNEL);
6192
		if (!ret_stack)
6193
			return;
6194
		graph_init_task(t, ret_stack);
6195
	}
6196 6197
}

6198
void ftrace_graph_exit_task(struct task_struct *t)
6199
{
6200 6201
	struct ftrace_ret_stack	*ret_stack = t->ret_stack;

6202
	t->ret_stack = NULL;
6203 6204 6205 6206
	/* NULL must become visible to IRQs before we free it: */
	barrier();

	kfree(ret_stack);
6207
}
6208
#endif