ftrace.c 160.9 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
#include <linux/stop_machine.h>
#include <linux/clocksource.h>
18
#include <linux/sched/task.h>
19
#include <linux/kallsyms.h>
20
#include <linux/seq_file.h>
21
#include <linux/suspend.h>
22
#include <linux/tracefs.h>
23
#include <linux/hardirq.h>
I
Ingo Molnar 已提交
24
#include <linux/kthread.h>
25
#include <linux/uaccess.h>
26
#include <linux/bsearch.h>
27
#include <linux/module.h>
I
Ingo Molnar 已提交
28
#include <linux/ftrace.h>
29
#include <linux/sysctl.h>
30
#include <linux/slab.h>
31
#include <linux/ctype.h>
32
#include <linux/sort.h>
33
#include <linux/list.h>
34
#include <linux/hash.h>
35
#include <linux/rcupdate.h>
36

37
#include <trace/events/sched.h>
38

39
#include <asm/sections.h>
40
#include <asm/setup.h>
41

42
#include "trace_output.h"
S
Steven Rostedt 已提交
43
#include "trace_stat.h"
44

45
#define FTRACE_WARN_ON(cond)			\
46 47 48
	({					\
		int ___r = cond;		\
		if (WARN_ON(___r))		\
49
			ftrace_kill();		\
50 51
		___r;				\
	})
52 53

#define FTRACE_WARN_ON_ONCE(cond)		\
54 55 56
	({					\
		int ___r = cond;		\
		if (WARN_ON_ONCE(___r))		\
57
			ftrace_kill();		\
58 59
		___r;				\
	})
60

61 62 63
/* hash bits for specific function selection */
#define FTRACE_HASH_BITS 7
#define FTRACE_FUNC_HASHSIZE (1 << FTRACE_HASH_BITS)
64 65
#define FTRACE_HASH_DEFAULT_BITS 10
#define FTRACE_HASH_MAX_BITS 12
66

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

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

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

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

94
static bool ftrace_pids_enabled(struct ftrace_ops *ops)
95
{
96 97 98 99 100 101 102 103
	struct trace_array *tr;

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

	tr = ops->private;

	return tr->function_pids != NULL;
104 105 106 107
}

static void ftrace_update_trampoline(struct ftrace_ops *ops);

108 109 110 111 112 113
/*
 * ftrace_disabled is set when an anomaly is discovered.
 * ftrace_disabled is much stronger than ftrace_enabled.
 */
static int ftrace_disabled __read_mostly;

114
static DEFINE_MUTEX(ftrace_lock);
115

116
static struct ftrace_ops __rcu *ftrace_ops_list __read_mostly = &ftrace_list_end;
117
ftrace_func_t ftrace_trace_function __read_mostly = ftrace_stub;
118
static struct ftrace_ops global_ops;
119

120 121
#if ARCH_SUPPORTS_FTRACE_OPS
static void ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
122
				 struct ftrace_ops *op, struct pt_regs *regs);
123 124 125 126 127
#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
128

129 130
/*
 * Traverse the ftrace_global_list, invoking all entries.  The reason that we
131
 * can use rcu_dereference_raw_notrace() is that elements removed from this list
132
 * are simply leaked, so there is no need to interact with a grace-period
133
 * mechanism.  The rcu_dereference_raw_notrace() calls are needed to handle
134 135 136 137 138
 * concurrent insertions into the ftrace_global_list.
 *
 * Silly Alpha and silly pointer-speculation compiler optimizations!
 */
#define do_for_each_ftrace_op(op, list)			\
139
	op = rcu_dereference_raw_notrace(list);			\
140 141 142 143 144 145
	do

/*
 * Optimized for just a single item in the list (as that is the normal case).
 */
#define while_for_each_ftrace_op(op)				\
146
	while (likely(op = rcu_dereference_raw_notrace((op)->next)) &&	\
147 148
	       unlikely((op) != &ftrace_list_end))

149 150 151 152
static inline void ftrace_ops_init(struct ftrace_ops *ops)
{
#ifdef CONFIG_DYNAMIC_FTRACE
	if (!(ops->flags & FTRACE_OPS_FL_INITIALIZED)) {
153 154
		mutex_init(&ops->local_hash.regex_lock);
		ops->func_hash = &ops->local_hash;
155 156 157 158 159
		ops->flags |= FTRACE_OPS_FL_INITIALIZED;
	}
#endif
}

160 161 162 163 164 165 166 167 168 169 170 171
/**
 * 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);

172 173 174 175 176
	for (ops = rcu_dereference_protected(ftrace_ops_list,
					     lockdep_is_held(&ftrace_lock));
	     ops != &ftrace_list_end;
	     ops = rcu_dereference_protected(ops->next,
					     lockdep_is_held(&ftrace_lock)))
177 178 179 180 181 182 183
		cnt++;

	mutex_unlock(&ftrace_lock);

	return cnt;
}

184
static void ftrace_pid_func(unsigned long ip, unsigned long parent_ip,
185
			    struct ftrace_ops *op, struct pt_regs *regs)
186
{
187 188 189
	struct trace_array *tr = op->private;

	if (tr && this_cpu_read(tr->trace_buffer.data->ftrace_ignore_pid))
190 191
		return;

192
	op->saved_func(ip, parent_ip, op, regs);
193 194
}

195
/**
196
 * clear_ftrace_function - reset the ftrace function
197
 *
198 199
 * This NULLs the ftrace function and in essence stops
 * tracing.  There may be lag
200
 */
201
void clear_ftrace_function(void)
202
{
203 204 205
	ftrace_trace_function = ftrace_stub;
}

206
static void per_cpu_ops_disable_all(struct ftrace_ops *ops)
207 208 209 210 211 212 213
{
	int cpu;

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

214
static int per_cpu_ops_alloc(struct ftrace_ops *ops)
215 216 217
{
	int __percpu *disabled;

218 219 220
	if (WARN_ON_ONCE(!(ops->flags & FTRACE_OPS_FL_PER_CPU)))
		return -EINVAL;

221 222 223 224 225
	disabled = alloc_percpu(int);
	if (!disabled)
		return -ENOMEM;

	ops->disabled = disabled;
226
	per_cpu_ops_disable_all(ops);
227 228 229
	return 0;
}

230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246
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();
}

247 248
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static void update_function_graph_func(void);
249 250 251 252 253

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

254 255 256 257
#else
static inline void update_function_graph_func(void) { }
#endif

258 259 260 261

static ftrace_func_t ftrace_ops_get_list_func(struct ftrace_ops *ops)
{
	/*
262
	 * If this is a dynamic, RCU, or per CPU ops, or we force list func,
263 264
	 * then it needs to call the list anyway.
	 */
265 266
	if (ops->flags & (FTRACE_OPS_FL_DYNAMIC | FTRACE_OPS_FL_PER_CPU |
			  FTRACE_OPS_FL_RCU) || FTRACE_FORCE_LIST_FUNC)
267 268 269 270 271
		return ftrace_ops_list_func;

	return ftrace_ops_get_func(ops);
}

272 273 274 275
static void update_ftrace_function(void)
{
	ftrace_func_t func;

276 277 278 279 280
	/*
	 * 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.
	 */
281 282
	set_function_trace_op = rcu_dereference_protected(ftrace_ops_list,
						lockdep_is_held(&ftrace_lock));
283 284

	/* If there's no ftrace_ops registered, just call the stub function */
285
	if (set_function_trace_op == &ftrace_list_end) {
286 287
		func = ftrace_stub;

288 289
	/*
	 * If we are at the end of the list and this ops is
290 291
	 * recursion safe and not dynamic and the arch supports passing ops,
	 * then have the mcount trampoline call the function directly.
292
	 */
293 294
	} else if (rcu_dereference_protected(ftrace_ops_list->next,
			lockdep_is_held(&ftrace_lock)) == &ftrace_list_end) {
295
		func = ftrace_ops_get_list_func(ftrace_ops_list);
296

297 298
	} else {
		/* Just use the default ftrace_ops */
299
		set_function_trace_op = &ftrace_list_end;
300
		func = ftrace_ops_list_func;
301
	}
302

303 304
	update_function_graph_func();

305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347
	/* 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 */

348 349 350
	ftrace_trace_function = func;
}

351 352 353 354 355
int using_ftrace_ops_list_func(void)
{
	return ftrace_trace_function == ftrace_ops_list_func;
}

356 357
static void add_ftrace_ops(struct ftrace_ops __rcu **list,
			   struct ftrace_ops *ops)
358
{
359 360
	rcu_assign_pointer(ops->next, *list);

361
	/*
362
	 * We are entering ops into the list but another
363 364
	 * CPU might be walking that list. We need to make sure
	 * the ops->next pointer is valid before another CPU sees
365
	 * the ops pointer included into the list.
366
	 */
367
	rcu_assign_pointer(*list, ops);
368 369
}

370 371
static int remove_ftrace_ops(struct ftrace_ops __rcu **list,
			     struct ftrace_ops *ops)
372 373 374 375
{
	struct ftrace_ops **p;

	/*
376 377
	 * If we are removing the last function, then simply point
	 * to the ftrace_stub.
378
	 */
379 380 381 382
	if (rcu_dereference_protected(*list,
			lockdep_is_held(&ftrace_lock)) == ops &&
	    rcu_dereference_protected(ops->next,
			lockdep_is_held(&ftrace_lock)) == &ftrace_list_end) {
383
		*list = &ftrace_list_end;
S
Steven Rostedt 已提交
384
		return 0;
385 386
	}

387
	for (p = list; *p != &ftrace_list_end; p = &(*p)->next)
388 389 390
		if (*p == ops)
			break;

S
Steven Rostedt 已提交
391 392
	if (*p != ops)
		return -1;
393 394

	*p = (*p)->next;
395 396
	return 0;
}
397

398 399
static void ftrace_update_trampoline(struct ftrace_ops *ops);

400 401
static int __register_ftrace_function(struct ftrace_ops *ops)
{
402 403 404
	if (ops->flags & FTRACE_OPS_FL_DELETED)
		return -EINVAL;

405 406 407
	if (WARN_ON(ops->flags & FTRACE_OPS_FL_ENABLED))
		return -EBUSY;

408
#ifndef CONFIG_DYNAMIC_FTRACE_WITH_REGS
409 410 411 412 413 414 415 416 417 418 419 420 421
	/*
	 * 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

422 423 424
	if (!core_kernel_data((unsigned long)ops))
		ops->flags |= FTRACE_OPS_FL_DYNAMIC;

425 426
	if (ops->flags & FTRACE_OPS_FL_PER_CPU) {
		if (per_cpu_ops_alloc(ops))
427
			return -ENOMEM;
428 429 430
	}

	add_ftrace_ops(&ftrace_ops_list, ops);
431

432 433 434
	/* Always save the function, and reset at unregistering */
	ops->saved_func = ops->func;

435
	if (ftrace_pids_enabled(ops))
436 437
		ops->func = ftrace_pid_func;

438 439
	ftrace_update_trampoline(ops);

440 441 442 443 444 445 446 447 448 449
	if (ftrace_enabled)
		update_ftrace_function();

	return 0;
}

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

450 451 452
	if (WARN_ON(!(ops->flags & FTRACE_OPS_FL_ENABLED)))
		return -EBUSY;

453
	ret = remove_ftrace_ops(&ftrace_ops_list, ops);
454

455 456
	if (ret < 0)
		return ret;
457

458 459
	if (ftrace_enabled)
		update_ftrace_function();
460

461 462
	ops->func = ops->saved_func;

S
Steven Rostedt 已提交
463
	return 0;
464 465
}

466 467
static void ftrace_update_pid_func(void)
{
468 469
	struct ftrace_ops *op;

470
	/* Only do something if we are tracing something */
471
	if (ftrace_trace_function == ftrace_stub)
472
		return;
473

474 475
	do_for_each_ftrace_op(op, ftrace_ops_list) {
		if (op->flags & FTRACE_OPS_FL_PID) {
476 477
			op->func = ftrace_pids_enabled(op) ?
				ftrace_pid_func : op->saved_func;
478 479 480 481
			ftrace_update_trampoline(op);
		}
	} while_for_each_ftrace_op(op);

482
	update_ftrace_function();
483 484
}

485 486 487 488 489
#ifdef CONFIG_FUNCTION_PROFILER
struct ftrace_profile {
	struct hlist_node		node;
	unsigned long			ip;
	unsigned long			counter;
490 491
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	unsigned long long		time;
492
	unsigned long long		time_squared;
493
#endif
494 495
};

496 497 498 499
struct ftrace_profile_page {
	struct ftrace_profile_page	*next;
	unsigned long			index;
	struct ftrace_profile		records[];
500 501
};

502 503 504 505 506 507 508 509
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;
};

510 511
#define PROFILE_RECORDS_SIZE						\
	(PAGE_SIZE - offsetof(struct ftrace_profile_page, records))
512

513 514
#define PROFILES_PER_PAGE					\
	(PROFILE_RECORDS_SIZE / sizeof(struct ftrace_profile))
515

516 517 518
static int ftrace_profile_enabled __read_mostly;

/* ftrace_profile_lock - synchronize the enable and disable of the profiler */
S
Steven Rostedt 已提交
519 520
static DEFINE_MUTEX(ftrace_profile_lock);

521
static DEFINE_PER_CPU(struct ftrace_profile_stat, ftrace_profile_stats);
522

523 524
#define FTRACE_PROFILE_HASH_BITS 10
#define FTRACE_PROFILE_HASH_SIZE (1 << FTRACE_PROFILE_HASH_BITS)
525

S
Steven Rostedt 已提交
526 527 528
static void *
function_stat_next(void *v, int idx)
{
529 530
	struct ftrace_profile *rec = v;
	struct ftrace_profile_page *pg;
S
Steven Rostedt 已提交
531

532
	pg = (struct ftrace_profile_page *)((unsigned long)rec & PAGE_MASK);
S
Steven Rostedt 已提交
533 534

 again:
L
Li Zefan 已提交
535 536 537
	if (idx != 0)
		rec++;

S
Steven Rostedt 已提交
538 539 540 541 542
	if ((void *)rec >= (void *)&pg->records[pg->index]) {
		pg = pg->next;
		if (!pg)
			return NULL;
		rec = &pg->records[0];
543 544
		if (!rec->counter)
			goto again;
S
Steven Rostedt 已提交
545 546 547 548 549 550 551
	}

	return rec;
}

static void *function_stat_start(struct tracer_stat *trace)
{
552 553 554 555 556 557 558
	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
559 560
}

561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576
#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
577 578
static int function_stat_cmp(void *p1, void *p2)
{
579 580
	struct ftrace_profile *a = p1;
	struct ftrace_profile *b = p2;
S
Steven Rostedt 已提交
581 582 583 584 585 586 587 588

	if (a->counter < b->counter)
		return -1;
	if (a->counter > b->counter)
		return 1;
	else
		return 0;
}
589
#endif
S
Steven Rostedt 已提交
590 591 592

static int function_stat_headers(struct seq_file *m)
{
593
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
594 595 596 597
	seq_puts(m, "  Function                               "
		 "Hit    Time            Avg             s^2\n"
		    "  --------                               "
		 "---    ----            ---             ---\n");
598
#else
599 600
	seq_puts(m, "  Function                               Hit\n"
		    "  --------                               ---\n");
601
#endif
S
Steven Rostedt 已提交
602 603 604 605 606
	return 0;
}

static int function_stat_show(struct seq_file *m, void *v)
{
607
	struct ftrace_profile *rec = v;
S
Steven Rostedt 已提交
608
	char str[KSYM_SYMBOL_LEN];
609
	int ret = 0;
610
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
611 612
	static struct trace_seq s;
	unsigned long long avg;
613
	unsigned long long stddev;
614
#endif
615 616 617 618 619 620 621
	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
622

623 624 625 626 627 628 629
#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
630
	kallsyms_lookup(rec->ip, NULL, NULL, NULL, str);
631 632 633
	seq_printf(m, "  %-30.30s  %10lu", str, rec->counter);

#ifdef CONFIG_FUNCTION_GRAPH_TRACER
634
	seq_puts(m, "    ");
635

636 637 638 639
	/* Sample standard deviation (s^2) */
	if (rec->counter <= 1)
		stddev = 0;
	else {
640 641 642 643 644 645 646
		/*
		 * 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;

647 648 649 650
		/*
		 * Divide only 1000 for ns^2 -> us^2 conversion.
		 * trace_print_graph_duration will divide 1000 again.
		 */
651
		do_div(stddev, rec->counter * (rec->counter - 1) * 1000);
652 653
	}

654 655 656 657
	trace_seq_init(&s);
	trace_print_graph_duration(rec->time, &s);
	trace_seq_puts(&s, "    ");
	trace_print_graph_duration(avg, &s);
658 659
	trace_seq_puts(&s, "    ");
	trace_print_graph_duration(stddev, &s);
660 661 662
	trace_print_seq(m, &s);
#endif
	seq_putc(m, '\n');
663 664
out:
	mutex_unlock(&ftrace_profile_lock);
S
Steven Rostedt 已提交
665

666
	return ret;
S
Steven Rostedt 已提交
667 668
}

669
static void ftrace_profile_reset(struct ftrace_profile_stat *stat)
S
Steven Rostedt 已提交
670
{
671
	struct ftrace_profile_page *pg;
S
Steven Rostedt 已提交
672

673
	pg = stat->pages = stat->start;
S
Steven Rostedt 已提交
674

675 676 677 678
	while (pg) {
		memset(pg->records, 0, PROFILE_RECORDS_SIZE);
		pg->index = 0;
		pg = pg->next;
S
Steven Rostedt 已提交
679 680
	}

681
	memset(stat->hash, 0,
682 683
	       FTRACE_PROFILE_HASH_SIZE * sizeof(struct hlist_head));
}
S
Steven Rostedt 已提交
684

685
int ftrace_profile_pages_init(struct ftrace_profile_stat *stat)
686 687
{
	struct ftrace_profile_page *pg;
688 689
	int functions;
	int pages;
690
	int i;
S
Steven Rostedt 已提交
691

692
	/* If we already allocated, do nothing */
693
	if (stat->pages)
694
		return 0;
S
Steven Rostedt 已提交
695

696 697
	stat->pages = (void *)get_zeroed_page(GFP_KERNEL);
	if (!stat->pages)
698
		return -ENOMEM;
S
Steven Rostedt 已提交
699

700 701 702 703 704 705 706 707 708 709 710 711 712
#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

713
	pg = stat->start = stat->pages;
S
Steven Rostedt 已提交
714

715 716
	pages = DIV_ROUND_UP(functions, PROFILES_PER_PAGE);

717
	for (i = 1; i < pages; i++) {
718 719
		pg->next = (void *)get_zeroed_page(GFP_KERNEL);
		if (!pg->next)
720
			goto out_free;
721 722 723 724
		pg = pg->next;
	}

	return 0;
725 726 727 728 729 730 731 732 733 734 735 736 737 738

 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
739 740
}

741
static int ftrace_profile_init_cpu(int cpu)
S
Steven Rostedt 已提交
742
{
743
	struct ftrace_profile_stat *stat;
744
	int size;
S
Steven Rostedt 已提交
745

746 747 748
	stat = &per_cpu(ftrace_profile_stats, cpu);

	if (stat->hash) {
749
		/* If the profile is already created, simply reset it */
750
		ftrace_profile_reset(stat);
751 752
		return 0;
	}
S
Steven Rostedt 已提交
753

754 755 756 757 758
	/*
	 * 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
759

760
	stat->hash = kzalloc(sizeof(struct hlist_head) * size, GFP_KERNEL);
761

762
	if (!stat->hash)
763 764
		return -ENOMEM;

765
	/* Preallocate the function profiling pages */
766 767 768
	if (ftrace_profile_pages_init(stat) < 0) {
		kfree(stat->hash);
		stat->hash = NULL;
769 770 771 772
		return -ENOMEM;
	}

	return 0;
S
Steven Rostedt 已提交
773 774
}

775 776 777 778 779
static int ftrace_profile_init(void)
{
	int cpu;
	int ret = 0;

780
	for_each_possible_cpu(cpu) {
781 782 783 784 785 786 787 788
		ret = ftrace_profile_init_cpu(cpu);
		if (ret)
			break;
	}

	return ret;
}

789
/* interrupts must be disabled */
790 791
static struct ftrace_profile *
ftrace_find_profiled_func(struct ftrace_profile_stat *stat, unsigned long ip)
S
Steven Rostedt 已提交
792
{
793
	struct ftrace_profile *rec;
S
Steven Rostedt 已提交
794 795 796
	struct hlist_head *hhd;
	unsigned long key;

797
	key = hash_long(ip, FTRACE_PROFILE_HASH_BITS);
798
	hhd = &stat->hash[key];
S
Steven Rostedt 已提交
799 800 801 802

	if (hlist_empty(hhd))
		return NULL;

803
	hlist_for_each_entry_rcu_notrace(rec, hhd, node) {
S
Steven Rostedt 已提交
804
		if (rec->ip == ip)
805 806 807 808 809 810
			return rec;
	}

	return NULL;
}

811 812
static void ftrace_add_profile(struct ftrace_profile_stat *stat,
			       struct ftrace_profile *rec)
813 814 815
{
	unsigned long key;

816
	key = hash_long(rec->ip, FTRACE_PROFILE_HASH_BITS);
817
	hlist_add_head_rcu(&rec->node, &stat->hash[key]);
818 819
}

820 821 822
/*
 * The memory is already allocated, this simply finds a new record to use.
 */
823
static struct ftrace_profile *
824
ftrace_profile_alloc(struct ftrace_profile_stat *stat, unsigned long ip)
825 826 827
{
	struct ftrace_profile *rec = NULL;

828
	/* prevent recursion (from NMIs) */
829
	if (atomic_inc_return(&stat->disabled) != 1)
830 831 832
		goto out;

	/*
833 834
	 * Try to find the function again since an NMI
	 * could have added it
835
	 */
836
	rec = ftrace_find_profiled_func(stat, ip);
837
	if (rec)
838
		goto out;
839

840 841 842 843
	if (stat->pages->index == PROFILES_PER_PAGE) {
		if (!stat->pages->next)
			goto out;
		stat->pages = stat->pages->next;
S
Steven Rostedt 已提交
844
	}
845

846
	rec = &stat->pages->records[stat->pages->index++];
847
	rec->ip = ip;
848
	ftrace_add_profile(stat, rec);
849

S
Steven Rostedt 已提交
850
 out:
851
	atomic_dec(&stat->disabled);
S
Steven Rostedt 已提交
852 853 854 855 856

	return rec;
}

static void
857
function_profile_call(unsigned long ip, unsigned long parent_ip,
858
		      struct ftrace_ops *ops, struct pt_regs *regs)
S
Steven Rostedt 已提交
859
{
860
	struct ftrace_profile_stat *stat;
861
	struct ftrace_profile *rec;
S
Steven Rostedt 已提交
862 863 864 865 866 867
	unsigned long flags;

	if (!ftrace_profile_enabled)
		return;

	local_irq_save(flags);
868

869
	stat = this_cpu_ptr(&ftrace_profile_stats);
870
	if (!stat->hash || !ftrace_profile_enabled)
871 872 873
		goto out;

	rec = ftrace_find_profiled_func(stat, ip);
874
	if (!rec) {
875
		rec = ftrace_profile_alloc(stat, ip);
876 877 878
		if (!rec)
			goto out;
	}
S
Steven Rostedt 已提交
879 880 881 882 883 884

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

885 886 887
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static int profile_graph_entry(struct ftrace_graph_ent *trace)
{
888 889
	int index = trace->depth;

890
	function_profile_call(trace->func, 0, NULL, NULL);
891

892 893 894 895
	/* If function graph is shutting down, ret_stack can be NULL */
	if (!current->ret_stack)
		return 0;

896 897 898
	if (index >= 0 && index < FTRACE_RETFUNC_DEPTH)
		current->ret_stack[index].subtime = 0;

899 900 901 902 903
	return 1;
}

static void profile_graph_return(struct ftrace_graph_ret *trace)
{
904
	struct ftrace_profile_stat *stat;
905
	unsigned long long calltime;
906
	struct ftrace_profile *rec;
907
	unsigned long flags;
908 909

	local_irq_save(flags);
910
	stat = this_cpu_ptr(&ftrace_profile_stats);
911
	if (!stat->hash || !ftrace_profile_enabled)
912 913
		goto out;

914 915 916 917
	/* If the calltime was zero'd ignore it */
	if (!trace->calltime)
		goto out;

918 919
	calltime = trace->rettime - trace->calltime;

920
	if (!fgraph_graph_time) {
921 922 923 924 925 926 927 928 929 930 931 932 933 934
		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;
	}

935
	rec = ftrace_find_profiled_func(stat, trace->func);
936
	if (rec) {
937
		rec->time += calltime;
938 939
		rec->time_squared += calltime * calltime;
	}
940

941
 out:
942 943 944 945 946 947 948 949 950 951 952 953 954 955
	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
956
static struct ftrace_ops ftrace_profile_ops __read_mostly = {
957
	.func		= function_profile_call,
958
	.flags		= FTRACE_OPS_FL_RECURSION_SAFE | FTRACE_OPS_FL_INITIALIZED,
959
	INIT_OPS_HASH(ftrace_profile_ops)
S
Steven Rostedt 已提交
960 961
};

962 963 964 965 966 967 968 969 970 971 972
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
973 974 975 976 977 978 979
static ssize_t
ftrace_profile_write(struct file *filp, const char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
	unsigned long val;
	int ret;

980 981
	ret = kstrtoul_from_user(ubuf, cnt, 10, &val);
	if (ret)
S
Steven Rostedt 已提交
982 983 984 985 986 987 988
		return ret;

	val = !!val;

	mutex_lock(&ftrace_profile_lock);
	if (ftrace_profile_enabled ^ val) {
		if (val) {
989 990 991 992 993 994
			ret = ftrace_profile_init();
			if (ret < 0) {
				cnt = ret;
				goto out;
			}

995 996 997 998 999
			ret = register_ftrace_profiler();
			if (ret < 0) {
				cnt = ret;
				goto out;
			}
S
Steven Rostedt 已提交
1000 1001 1002
			ftrace_profile_enabled = 1;
		} else {
			ftrace_profile_enabled = 0;
1003 1004 1005 1006
			/*
			 * unregister_ftrace_profiler calls stop_machine
			 * so this acts like an synchronize_sched.
			 */
1007
			unregister_ftrace_profiler();
S
Steven Rostedt 已提交
1008 1009
		}
	}
1010
 out:
S
Steven Rostedt 已提交
1011 1012
	mutex_unlock(&ftrace_profile_lock);

1013
	*ppos += cnt;
S
Steven Rostedt 已提交
1014 1015 1016 1017

	return cnt;
}

1018 1019 1020 1021
static ssize_t
ftrace_profile_read(struct file *filp, char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
1022
	char buf[64];		/* big enough to hold a number */
1023 1024 1025 1026 1027 1028
	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
1029 1030 1031 1032
static const struct file_operations ftrace_profile_fops = {
	.open		= tracing_open_generic,
	.read		= ftrace_profile_read,
	.write		= ftrace_profile_write,
1033
	.llseek		= default_llseek,
S
Steven Rostedt 已提交
1034 1035
};

1036 1037
/* used to initialize the real stat files */
static struct tracer_stat function_stats __initdata = {
1038 1039 1040 1041 1042 1043
	.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
1044 1045
};

1046
static __init void ftrace_profile_tracefs(struct dentry *d_tracer)
S
Steven Rostedt 已提交
1047
{
1048
	struct ftrace_profile_stat *stat;
S
Steven Rostedt 已提交
1049
	struct dentry *entry;
1050
	char *name;
S
Steven Rostedt 已提交
1051
	int ret;
1052 1053 1054 1055 1056
	int cpu;

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

1057
		name = kasprintf(GFP_KERNEL, "function%d", cpu);
1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077
		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
1078 1079
	}

1080
	entry = tracefs_create_file("function_profile_enabled", 0644,
S
Steven Rostedt 已提交
1081 1082
				    d_tracer, NULL, &ftrace_profile_fops);
	if (!entry)
1083
		pr_warn("Could not create tracefs 'function_profile_enabled' entry\n");
S
Steven Rostedt 已提交
1084 1085 1086
}

#else /* CONFIG_FUNCTION_PROFILER */
1087
static __init void ftrace_profile_tracefs(struct dentry *d_tracer)
S
Steven Rostedt 已提交
1088 1089 1090 1091
{
}
#endif /* CONFIG_FUNCTION_PROFILER */

1092 1093
static struct pid * const ftrace_swapper_pid = &init_struct_pid;

1094 1095 1096 1097 1098 1099
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static int ftrace_graph_active;
#else
# define ftrace_graph_active 0
#endif

1100 1101
#ifdef CONFIG_DYNAMIC_FTRACE

1102 1103
static struct ftrace_ops *removed_ops;

1104 1105 1106 1107 1108 1109
/*
 * 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;

1110 1111 1112 1113
#ifndef CONFIG_FTRACE_MCOUNT_RECORD
# error Dynamic ftrace depends on MCOUNT_RECORD
#endif

1114 1115 1116 1117 1118
struct ftrace_func_entry {
	struct hlist_node hlist;
	unsigned long ip;
};

1119 1120 1121 1122 1123
struct ftrace_func_probe {
	struct ftrace_probe_ops	*probe_ops;
	struct ftrace_ops	ops;
	struct trace_array	*tr;
	struct list_head	list;
1124
	void			*data;
1125 1126 1127
	int			ref;
};

1128 1129 1130 1131 1132 1133 1134 1135 1136
/*
 * 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,
1137
};
1138
#define EMPTY_HASH	((struct ftrace_hash *)&empty_hash)
1139

1140
static struct ftrace_ops global_ops = {
1141 1142 1143 1144 1145
	.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 |
1146 1147
					  FTRACE_OPS_FL_INITIALIZED |
					  FTRACE_OPS_FL_PID,
1148 1149
};

1150 1151
/*
 * This is used by __kernel_text_address() to return true if the
1152
 * address is on a dynamically allocated trampoline that would
1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186
 * 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;
}

1187 1188
struct ftrace_page {
	struct ftrace_page	*next;
1189
	struct dyn_ftrace	*records;
1190
	int			index;
1191
	int			size;
1192 1193
};

1194 1195
#define ENTRY_SIZE sizeof(struct dyn_ftrace)
#define ENTRIES_PER_PAGE (PAGE_SIZE / ENTRY_SIZE)
1196 1197 1198 1199 1200 1201 1202

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

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

1203 1204 1205 1206 1207 1208 1209 1210 1211
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;
}

1212 1213 1214
/* 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)
1215 1216 1217 1218 1219
{
	unsigned long key;
	struct ftrace_func_entry *entry;
	struct hlist_head *hhd;

1220
	key = ftrace_hash_key(hash, ip);
1221 1222
	hhd = &hash->buckets[key];

1223
	hlist_for_each_entry_rcu_notrace(entry, hhd, hlist) {
1224 1225 1226 1227 1228 1229
		if (entry->ip == ip)
			return entry;
	}
	return NULL;
}

1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248
/**
 * 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);
}

1249 1250
static void __add_hash_entry(struct ftrace_hash *hash,
			     struct ftrace_func_entry *entry)
1251 1252 1253 1254
{
	struct hlist_head *hhd;
	unsigned long key;

1255
	key = ftrace_hash_key(hash, entry->ip);
1256 1257 1258
	hhd = &hash->buckets[key];
	hlist_add_head(&entry->hlist, hhd);
	hash->count++;
1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270
}

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);
1271 1272 1273 1274 1275

	return 0;
}

static void
1276
free_hash_entry(struct ftrace_hash *hash,
1277 1278 1279 1280 1281 1282 1283
		  struct ftrace_func_entry *entry)
{
	hlist_del(&entry->hlist);
	kfree(entry);
	hash->count--;
}

1284 1285 1286 1287
static void
remove_hash_entry(struct ftrace_hash *hash,
		  struct ftrace_func_entry *entry)
{
1288
	hlist_del_rcu(&entry->hlist);
1289 1290 1291
	hash->count--;
}

1292 1293 1294
static void ftrace_hash_clear(struct ftrace_hash *hash)
{
	struct hlist_head *hhd;
1295
	struct hlist_node *tn;
1296 1297 1298 1299
	struct ftrace_func_entry *entry;
	int size = 1 << hash->size_bits;
	int i;

1300 1301 1302
	if (!hash->count)
		return;

1303 1304
	for (i = 0; i < size; i++) {
		hhd = &hash->buckets[i];
1305
		hlist_for_each_entry_safe(entry, tn, hhd, hlist)
1306
			free_hash_entry(hash, entry);
1307 1308 1309 1310
	}
	FTRACE_WARN_ON(hash->count);
}

1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332
static void free_ftrace_mod(struct ftrace_mod_load *ftrace_mod)
{
	list_del(&ftrace_mod->list);
	kfree(ftrace_mod->module);
	kfree(ftrace_mod->func);
	kfree(ftrace_mod);
}

static void clear_ftrace_mod_list(struct list_head *head)
{
	struct ftrace_mod_load *p, *n;

	/* stack tracer isn't supported yet */
	if (!head)
		return;

	mutex_lock(&ftrace_lock);
	list_for_each_entry_safe(p, n, head, list)
		free_ftrace_mod(p);
	mutex_unlock(&ftrace_lock);
}

1333 1334 1335 1336 1337 1338 1339 1340 1341
static void free_ftrace_hash(struct ftrace_hash *hash)
{
	if (!hash || hash == EMPTY_HASH)
		return;
	ftrace_hash_clear(hash);
	kfree(hash->buckets);
	kfree(hash);
}

1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356
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);
}

1357 1358
void ftrace_free_filter(struct ftrace_ops *ops)
{
1359
	ftrace_ops_init(ops);
1360 1361
	free_ftrace_hash(ops->func_hash->filter_hash);
	free_ftrace_hash(ops->func_hash->notrace_hash);
1362 1363
}

1364 1365 1366 1367 1368 1369 1370 1371 1372 1373
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;
1374
	hash->buckets = kcalloc(size, sizeof(*hash->buckets), GFP_KERNEL);
1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385

	if (!hash->buckets) {
		kfree(hash);
		return NULL;
	}

	hash->size_bits = size_bits;

	return hash;
}

1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414

static int ftrace_add_mod(struct trace_array *tr,
			  const char *func, const char *module,
			  int enable)
{
	struct ftrace_mod_load *ftrace_mod;
	struct list_head *mod_head = enable ? &tr->mod_trace : &tr->mod_notrace;

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

	ftrace_mod->func = kstrdup(func, GFP_KERNEL);
	ftrace_mod->module = kstrdup(module, GFP_KERNEL);
	ftrace_mod->enable = enable;

	if (!ftrace_mod->func || !ftrace_mod->module)
		goto out_free;

	list_add(&ftrace_mod->list, mod_head);

	return 0;

 out_free:
	free_ftrace_mod(ftrace_mod);

	return -ENOMEM;
}

1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427
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;

1428 1429 1430
	if (hash)
		new_hash->flags = hash->flags;

1431
	/* Empty hash? */
1432
	if (ftrace_hash_empty(hash))
1433 1434 1435 1436
		return new_hash;

	size = 1 << hash->size_bits;
	for (i = 0; i < size; i++) {
1437
		hlist_for_each_entry(entry, &hash->buckets[i], hlist) {
1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452
			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;
}

1453
static void
1454
ftrace_hash_rec_disable_modify(struct ftrace_ops *ops, int filter_hash);
1455
static void
1456
ftrace_hash_rec_enable_modify(struct ftrace_ops *ops, int filter_hash);
1457

1458 1459 1460
static int ftrace_hash_ipmodify_update(struct ftrace_ops *ops,
				       struct ftrace_hash *new_hash);

1461 1462
static struct ftrace_hash *
__ftrace_hash_move(struct ftrace_hash *src)
1463 1464
{
	struct ftrace_func_entry *entry;
1465
	struct hlist_node *tn;
1466
	struct hlist_head *hhd;
1467
	struct ftrace_hash *new_hash;
1468 1469 1470 1471 1472
	int size = src->count;
	int bits = 0;
	int i;

	/*
1473
	 * If the new source is empty, just return the empty_hash.
1474
	 */
1475
	if (ftrace_hash_empty(src))
1476
		return EMPTY_HASH;
1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487

	/*
	 * 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;

1488 1489
	new_hash = alloc_ftrace_hash(bits);
	if (!new_hash)
1490
		return NULL;
1491

1492 1493
	new_hash->flags = src->flags;

1494 1495 1496
	size = 1 << src->size_bits;
	for (i = 0; i < size; i++) {
		hhd = &src->buckets[i];
1497
		hlist_for_each_entry_safe(entry, tn, hhd, hlist) {
1498
			remove_hash_entry(src, entry);
1499
			__add_hash_entry(new_hash, entry);
1500 1501 1502
		}
	}

1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520
	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;

1521 1522 1523 1524 1525 1526 1527 1528 1529 1530
	/* 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;
		}
	}

1531 1532 1533 1534
	/*
	 * Remove the current set, update the hash and add
	 * them back.
	 */
1535
	ftrace_hash_rec_disable_modify(ops, enable);
1536

1537 1538
	rcu_assign_pointer(*dst, new_hash);

1539
	ftrace_hash_rec_enable_modify(ops, enable);
1540

1541
	return 0;
1542 1543
}

1544 1545 1546 1547 1548 1549 1550 1551 1552 1553
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) ||
1554
		__ftrace_lookup_ip(hash->filter_hash, ip)) &&
1555
		(ftrace_hash_empty(hash->notrace_hash) ||
1556
		 !__ftrace_lookup_ip(hash->notrace_hash, ip));
1557 1558
}

1559 1560 1561 1562 1563 1564 1565 1566
/*
 * 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.
1567 1568 1569
 *
 * This needs to be called with preemption disabled as
 * the hashes are freed with call_rcu_sched().
1570 1571
 */
static int
1572
ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip, void *regs)
1573
{
1574
	struct ftrace_ops_hash hash;
1575 1576
	int ret;

1577 1578 1579 1580 1581 1582 1583 1584 1585 1586
#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

1587 1588
	rcu_assign_pointer(hash.filter_hash, ops->func_hash->filter_hash);
	rcu_assign_pointer(hash.notrace_hash, ops->func_hash->notrace_hash);
1589

1590
	if (hash_contains_ip(ip, &hash))
1591 1592 1593 1594 1595 1596 1597
		ret = 1;
	else
		ret = 0;

	return ret;
}

1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611
/*
 * 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()		\
		}				\
	}

1612 1613 1614

static int ftrace_cmp_recs(const void *a, const void *b)
{
1615 1616
	const struct dyn_ftrace *key = a;
	const struct dyn_ftrace *rec = b;
1617

1618
	if (key->flags < rec->ip)
1619
		return -1;
1620 1621
	if (key->ip >= rec->ip + MCOUNT_INSN_SIZE)
		return 1;
1622 1623 1624
	return 0;
}

1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637
/**
 * 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)
1638 1639 1640
{
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
1641
	struct dyn_ftrace key;
1642

1643 1644
	key.ip = start;
	key.flags = end;	/* overload flags, as it is unsigned long */
1645 1646

	for (pg = ftrace_pages_start; pg; pg = pg->next) {
1647 1648
		if (end < pg->records[0].ip ||
		    start >= (pg->records[pg->index - 1].ip + MCOUNT_INSN_SIZE))
1649
			continue;
1650 1651 1652 1653
		rec = bsearch(&key, pg->records, pg->index,
			      sizeof(struct dyn_ftrace),
			      ftrace_cmp_recs);
		if (rec)
1654
			return rec->ip;
1655
	}
1656 1657 1658 1659

	return 0;
}

1660 1661 1662 1663
/**
 * ftrace_location - return true if the ip giving is a traced location
 * @ip: the instruction pointer to check
 *
1664
 * Returns rec->ip if @ip given is a pointer to a ftrace location.
1665 1666 1667 1668
 * 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.
 */
1669
unsigned long ftrace_location(unsigned long ip)
1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683
{
	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.
 */
1684
int ftrace_text_reserved(const void *start, const void *end)
1685
{
1686 1687 1688 1689 1690 1691
	unsigned long ret;

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

	return (int)!!ret;
1692 1693
}

1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713
/* 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;
}

1714
static bool __ftrace_hash_rec_update(struct ftrace_ops *ops,
1715 1716 1717 1718 1719 1720 1721
				     int filter_hash,
				     bool inc)
{
	struct ftrace_hash *hash;
	struct ftrace_hash *other_hash;
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
1722
	bool update = false;
1723
	int count = 0;
1724
	int all = false;
1725 1726 1727

	/* Only update if the ops has been registered */
	if (!(ops->flags & FTRACE_OPS_FL_ENABLED))
1728
		return false;
1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741

	/*
	 * 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) {
1742 1743
		hash = ops->func_hash->filter_hash;
		other_hash = ops->func_hash->notrace_hash;
1744
		if (ftrace_hash_empty(hash))
1745
			all = true;
1746 1747
	} else {
		inc = !inc;
1748 1749
		hash = ops->func_hash->notrace_hash;
		other_hash = ops->func_hash->filter_hash;
1750 1751 1752 1753
		/*
		 * If the notrace hash has no items,
		 * then there's nothing to do.
		 */
1754
		if (ftrace_hash_empty(hash))
1755
			return false;
1756 1757 1758 1759 1760 1761 1762
	}

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

1763 1764 1765
		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

1766 1767 1768 1769 1770
		if (all) {
			/*
			 * Only the filter_hash affects all records.
			 * Update if the record is not in the notrace hash.
			 */
1771
			if (!other_hash || !ftrace_lookup_ip(other_hash, rec->ip))
1772 1773
				match = 1;
		} else {
1774 1775
			in_hash = !!ftrace_lookup_ip(hash, rec->ip);
			in_other_hash = !!ftrace_lookup_ip(other_hash, rec->ip);
1776 1777

			/*
1778 1779
			 * If filter_hash is set, we want to match all functions
			 * that are in the hash but not in the other hash.
1780
			 *
1781 1782 1783 1784 1785
			 * 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.
1786 1787 1788 1789
			 */
			if (filter_hash && in_hash && !in_other_hash)
				match = 1;
			else if (!filter_hash && in_hash &&
1790
				 (in_other_hash || ftrace_hash_empty(other_hash)))
1791 1792 1793 1794 1795 1796 1797
				match = 1;
		}
		if (!match)
			continue;

		if (inc) {
			rec->flags++;
1798
			if (FTRACE_WARN_ON(ftrace_rec_count(rec) == FTRACE_REF_MAX))
1799
				return false;
1800 1801 1802 1803 1804 1805

			/*
			 * 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.
			 */
1806
			if (ftrace_rec_count(rec) == 1 && ops->trampoline)
1807
				rec->flags |= FTRACE_FL_TRAMP;
1808
			else
1809 1810 1811
				/*
				 * If we are adding another function callback
				 * to this function, and the previous had a
1812 1813
				 * custom trampoline in use, then we need to go
				 * back to the default trampoline.
1814
				 */
1815
				rec->flags &= ~FTRACE_FL_TRAMP;
1816

1817 1818 1819 1820 1821 1822
			/*
			 * 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;
1823
		} else {
1824
			if (FTRACE_WARN_ON(ftrace_rec_count(rec) == 0))
1825
				return false;
1826
			rec->flags--;
1827

1828 1829 1830 1831 1832 1833
			/*
			 * 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.
			 */
1834
			if (ftrace_rec_count(rec) > 0 &&
1835 1836 1837 1838 1839
			    rec->flags & FTRACE_FL_REGS &&
			    ops->flags & FTRACE_OPS_FL_SAVE_REGS) {
				if (!test_rec_ops_needs_regs(rec))
					rec->flags &= ~FTRACE_FL_REGS;
			}
1840

1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851
			/*
			 * 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;

1852 1853 1854 1855
			/*
			 * flags will be cleared in ftrace_check_record()
			 * if rec count is zero.
			 */
1856 1857
		}
		count++;
1858 1859 1860 1861

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

1862 1863
		/* Shortcut, if we handled all records, we are done. */
		if (!all && count == hash->count)
1864
			return update;
1865
	} while_for_each_ftrace_rec();
1866 1867

	return update;
1868 1869
}

1870
static bool ftrace_hash_rec_disable(struct ftrace_ops *ops,
1871 1872
				    int filter_hash)
{
1873
	return __ftrace_hash_rec_update(ops, filter_hash, 0);
1874 1875
}

1876
static bool ftrace_hash_rec_enable(struct ftrace_ops *ops,
1877 1878
				   int filter_hash)
{
1879
	return __ftrace_hash_rec_update(ops, filter_hash, 1);
1880 1881
}

1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916
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);
}

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
/*
 * 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) {
1950 1951 1952 1953

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
		/* 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) {
1976 1977 1978 1979

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032
		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);
}

2033
static void print_ip_ins(const char *fmt, const unsigned char *p)
2034 2035 2036 2037 2038 2039 2040 2041 2042
{
	int i;

	printk(KERN_CONT "%s", fmt);

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

2043 2044
static struct ftrace_ops *
ftrace_find_tramp_ops_any(struct dyn_ftrace *rec);
2045 2046
static struct ftrace_ops *
ftrace_find_tramp_ops_next(struct dyn_ftrace *rec, struct ftrace_ops *ops);
2047

2048
enum ftrace_bug_type ftrace_bug_type;
2049
const void *ftrace_expected;
2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070

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;
	}
}

2071 2072 2073
/**
 * ftrace_bug - report and shutdown function tracer
 * @failed: The failed type (EFAULT, EINVAL, EPERM)
2074
 * @rec: The record that failed
2075 2076 2077 2078 2079 2080 2081 2082
 *
 * 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
 */
2083
void ftrace_bug(int failed, struct dyn_ftrace *rec)
2084
{
2085 2086
	unsigned long ip = rec ? rec->ip : 0;

2087 2088 2089 2090 2091 2092 2093 2094 2095 2096
	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);
2097
		print_ip_ins(" actual:   ", (unsigned char *)ip);
2098
		pr_cont("\n");
2099 2100 2101 2102
		if (ftrace_expected) {
			print_ip_ins(" expected: ", ftrace_expected);
			pr_cont("\n");
		}
2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113
		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);
	}
2114
	print_bug_type();
2115 2116 2117 2118 2119 2120 2121 2122
	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);
2123 2124 2125 2126 2127 2128 2129 2130
			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
2131 2132 2133 2134
				pr_cont("\ttramp: ERROR!");

		}
		ip = ftrace_get_addr_curr(rec);
2135
		pr_cont("\n expected tramp: %lx\n", ip);
2136
	}
2137 2138
}

2139
static int ftrace_check_record(struct dyn_ftrace *rec, int enable, int update)
2140
{
2141
	unsigned long flag = 0UL;
2142

2143 2144
	ftrace_bug_type = FTRACE_BUG_UNKNOWN;

2145 2146 2147
	if (rec->flags & FTRACE_FL_DISABLED)
		return FTRACE_UPDATE_IGNORE;

S
Steven Rostedt 已提交
2148
	/*
2149
	 * If we are updating calls:
S
Steven Rostedt 已提交
2150
	 *
2151 2152
	 *   If the record has a ref count, then we need to enable it
	 *   because someone is using it.
S
Steven Rostedt 已提交
2153
	 *
2154 2155
	 *   Otherwise we make sure its disabled.
	 *
2156
	 * If we are disabling calls, then disable all records that
2157
	 * are enabled.
S
Steven Rostedt 已提交
2158
	 */
2159
	if (enable && ftrace_rec_count(rec))
2160
		flag = FTRACE_FL_ENABLED;
S
Steven Rostedt 已提交
2161

2162
	/*
2163 2164 2165
	 * 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.
2166
	 */
2167 2168 2169 2170 2171 2172 2173 2174 2175
	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;
	}
2176

2177 2178
	/* If the state of this record hasn't changed, then do nothing */
	if ((rec->flags & FTRACE_FL_ENABLED) == flag)
2179
		return FTRACE_UPDATE_IGNORE;
S
Steven Rostedt 已提交
2180

2181
	if (flag) {
2182 2183 2184 2185
		/* Save off if rec is being enabled (for return value) */
		flag ^= rec->flags & FTRACE_FL_ENABLED;

		if (update) {
2186
			rec->flags |= FTRACE_FL_ENABLED;
2187 2188 2189 2190 2191 2192
			if (flag & FTRACE_FL_REGS) {
				if (rec->flags & FTRACE_FL_REGS)
					rec->flags |= FTRACE_FL_REGS_EN;
				else
					rec->flags &= ~FTRACE_FL_REGS_EN;
			}
2193 2194 2195 2196 2197 2198
			if (flag & FTRACE_FL_TRAMP) {
				if (rec->flags & FTRACE_FL_TRAMP)
					rec->flags |= FTRACE_FL_TRAMP_EN;
				else
					rec->flags &= ~FTRACE_FL_TRAMP_EN;
			}
2199 2200 2201 2202 2203 2204 2205
		}

		/*
		 * 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
2206
		 *   from the save regs, to a non-save regs function or
2207
		 *   vice versa, or from a trampoline call.
2208
		 */
2209 2210
		if (flag & FTRACE_FL_ENABLED) {
			ftrace_bug_type = FTRACE_BUG_CALL;
2211
			return FTRACE_UPDATE_MAKE_CALL;
2212
		}
2213

2214
		ftrace_bug_type = FTRACE_BUG_UPDATE;
2215
		return FTRACE_UPDATE_MODIFY_CALL;
2216 2217
	}

2218 2219
	if (update) {
		/* If there's no more users, clear all flags */
2220
		if (!ftrace_rec_count(rec))
2221 2222
			rec->flags = 0;
		else
2223 2224 2225 2226 2227 2228
			/*
			 * 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);
2229
	}
2230

2231
	ftrace_bug_type = FTRACE_BUG_NOP;
2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261
	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);
}

2262 2263 2264 2265
static struct ftrace_ops *
ftrace_find_tramp_ops_any(struct dyn_ftrace *rec)
{
	struct ftrace_ops *op;
2266
	unsigned long ip = rec->ip;
2267 2268 2269 2270 2271 2272

	do_for_each_ftrace_op(op, ftrace_ops_list) {

		if (!op->trampoline)
			continue;

2273
		if (hash_contains_ip(ip, op->func_hash))
2274 2275 2276 2277 2278 2279
			return op;
	} while_for_each_ftrace_op(op);

	return NULL;
}

2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297
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;
}

2298 2299 2300 2301
static struct ftrace_ops *
ftrace_find_tramp_ops_curr(struct dyn_ftrace *rec)
{
	struct ftrace_ops *op;
2302
	unsigned long ip = rec->ip;
2303

2304 2305 2306 2307 2308 2309 2310 2311
	/*
	 * 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))
2312 2313 2314
			return removed_ops;
	}

2315 2316 2317 2318 2319 2320 2321
	/*
	 * 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
2322 2323 2324 2325 2326 2327 2328 2329 2330 2331
	 * 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.
2332
	 */
2333
	do_for_each_ftrace_op(op, ftrace_ops_list) {
2334 2335 2336 2337 2338 2339 2340 2341 2342

		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)
2343 2344
			continue;

2345

2346
		/*
2347 2348 2349
		 * If the ops is being modified and is in the old
		 * hash, then it is probably being removed from this
		 * function.
2350 2351 2352
		 */
		if ((op->flags & FTRACE_OPS_FL_MODIFYING) &&
		    hash_contains_ip(ip, &op->old_hash))
2353
			return op;
2354 2355 2356 2357 2358 2359 2360 2361
		/*
		 * 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;
2362 2363 2364 2365 2366 2367 2368 2369 2370 2371

	} 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;
2372
	unsigned long ip = rec->ip;
2373 2374 2375

	do_for_each_ftrace_op(op, ftrace_ops_list) {
		/* pass rec in as regs to have non-NULL val */
2376
		if (hash_contains_ip(ip, op->func_hash))
2377 2378 2379 2380 2381 2382
			return op;
	} while_for_each_ftrace_op(op);

	return NULL;
}

2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394
/**
 * 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)
{
2395 2396 2397 2398 2399 2400
	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)) {
2401 2402
			pr_warn("Bad trampoline accounting at: %p (%pS) (%lx)\n",
				(void *)rec->ip, (void *)rec->ip, rec->flags);
2403 2404 2405 2406 2407 2408
			/* Ftrace is shutting down, return anything */
			return (unsigned long)FTRACE_ADDR;
		}
		return ops->trampoline;
	}

2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426
	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)
{
2427 2428 2429 2430 2431 2432
	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)) {
2433 2434
			pr_warn("Bad trampoline accounting at: %p (%pS)\n",
				(void *)rec->ip, (void *)rec->ip);
2435 2436 2437 2438 2439 2440
			/* Ftrace is shutting down, return anything */
			return (unsigned long)FTRACE_ADDR;
		}
		return ops->trampoline;
	}

2441 2442 2443 2444 2445 2446
	if (rec->flags & FTRACE_FL_REGS_EN)
		return (unsigned long)FTRACE_REGS_ADDR;
	else
		return (unsigned long)FTRACE_ADDR;
}

2447 2448 2449
static int
__ftrace_replace_code(struct dyn_ftrace *rec, int enable)
{
2450
	unsigned long ftrace_old_addr;
2451 2452 2453
	unsigned long ftrace_addr;
	int ret;

2454
	ftrace_addr = ftrace_get_addr_new(rec);
2455

2456 2457 2458 2459
	/* 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);
2460

2461 2462
	ftrace_bug_type = FTRACE_BUG_UNKNOWN;

2463 2464 2465 2466 2467
	switch (ret) {
	case FTRACE_UPDATE_IGNORE:
		return 0;

	case FTRACE_UPDATE_MAKE_CALL:
2468
		ftrace_bug_type = FTRACE_BUG_CALL;
2469
		return ftrace_make_call(rec, ftrace_addr);
2470 2471

	case FTRACE_UPDATE_MAKE_NOP:
2472
		ftrace_bug_type = FTRACE_BUG_NOP;
2473
		return ftrace_make_nop(NULL, rec, ftrace_old_addr);
2474 2475

	case FTRACE_UPDATE_MODIFY_CALL:
2476
		ftrace_bug_type = FTRACE_BUG_UPDATE;
2477
		return ftrace_modify_call(rec, ftrace_old_addr, ftrace_addr);
2478 2479
	}

2480
	return -1; /* unknow ftrace bug */
2481 2482
}

2483
void __weak ftrace_replace_code(int enable)
2484 2485 2486
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
S
Steven Rostedt 已提交
2487
	int failed;
2488

2489 2490 2491
	if (unlikely(ftrace_disabled))
		return;

2492
	do_for_each_ftrace_rec(pg, rec) {
2493 2494 2495 2496

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

2497
		failed = __ftrace_replace_code(rec, enable);
2498
		if (failed) {
2499
			ftrace_bug(failed, rec);
2500 2501
			/* Stop processing */
			return;
2502
		}
2503
	} while_for_each_ftrace_rec();
2504 2505
}

2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577
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];
}

2578
static int
2579
ftrace_code_disable(struct module *mod, struct dyn_ftrace *rec)
2580
{
2581
	int ret;
2582

2583 2584 2585
	if (unlikely(ftrace_disabled))
		return 0;

2586
	ret = ftrace_make_nop(mod, rec, MCOUNT_ADDR);
2587
	if (ret) {
2588
		ftrace_bug_type = FTRACE_BUG_INIT;
2589
		ftrace_bug(ret, rec);
2590
		return 0;
2591
	}
2592
	return 1;
2593 2594
}

2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612
/*
 * 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;
}

2613
void ftrace_modify_all_code(int command)
2614
{
2615
	int update = command & FTRACE_UPDATE_TRACE_FUNC;
2616
	int err = 0;
2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627

	/*
	 * 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.
	 */
2628 2629 2630 2631 2632
	if (update) {
		err = ftrace_update_ftrace_func(ftrace_ops_list_func);
		if (FTRACE_WARN_ON(err))
			return;
	}
2633

2634
	if (command & FTRACE_UPDATE_CALLS)
2635
		ftrace_replace_code(1);
2636
	else if (command & FTRACE_DISABLE_CALLS)
2637 2638
		ftrace_replace_code(0);

2639 2640 2641 2642 2643 2644
	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);
2645 2646 2647
		err = ftrace_update_ftrace_func(ftrace_trace_function);
		if (FTRACE_WARN_ON(err))
			return;
2648
	}
2649

2650
	if (command & FTRACE_START_FUNC_RET)
2651
		err = ftrace_enable_ftrace_graph_caller();
2652
	else if (command & FTRACE_STOP_FUNC_RET)
2653 2654
		err = ftrace_disable_ftrace_graph_caller();
	FTRACE_WARN_ON(err);
2655 2656 2657 2658 2659 2660 2661
}

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

	ftrace_modify_all_code(*command);
2662

2663
	return 0;
2664 2665
}

2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689
/**
 * 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
2690
static void ftrace_run_update_code(int command)
2691
{
2692 2693 2694 2695 2696 2697 2698
	int ret;

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

2699 2700 2701 2702 2703 2704 2705 2706
	/*
	 * 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);

2707 2708
	ret = ftrace_arch_code_modify_post_process();
	FTRACE_WARN_ON(ret);
2709 2710
}

2711
static void ftrace_run_modify_code(struct ftrace_ops *ops, int command,
2712
				   struct ftrace_ops_hash *old_hash)
2713 2714
{
	ops->flags |= FTRACE_OPS_FL_MODIFYING;
2715 2716
	ops->old_hash.filter_hash = old_hash->filter_hash;
	ops->old_hash.notrace_hash = old_hash->notrace_hash;
2717
	ftrace_run_update_code(command);
2718
	ops->old_hash.filter_hash = NULL;
2719
	ops->old_hash.notrace_hash = NULL;
2720 2721 2722
	ops->flags &= ~FTRACE_OPS_FL_MODIFYING;
}

2723
static ftrace_func_t saved_ftrace_func;
2724
static int ftrace_start_up;
2725

2726 2727 2728 2729
void __weak arch_ftrace_trampoline_free(struct ftrace_ops *ops)
{
}

2730
static void per_cpu_ops_free(struct ftrace_ops *ops)
2731 2732 2733 2734
{
	free_percpu(ops->disabled);
}

2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746
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);
}
2747

2748 2749 2750 2751 2752 2753 2754
static void ftrace_startup_all(int command)
{
	update_all_ops = true;
	ftrace_startup_enable(command);
	update_all_ops = false;
}

2755
static int ftrace_startup(struct ftrace_ops *ops, int command)
2756
{
2757
	int ret;
2758

2759
	if (unlikely(ftrace_disabled))
2760
		return -ENODEV;
2761

2762 2763 2764 2765
	ret = __register_ftrace_function(ops);
	if (ret)
		return ret;

2766
	ftrace_start_up++;
2767

2768 2769 2770 2771 2772 2773 2774 2775 2776
	/*
	 * 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;
2777

2778 2779 2780 2781 2782 2783 2784 2785 2786
	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;
	}

2787 2788
	if (ftrace_hash_rec_enable(ops, 1))
		command |= FTRACE_UPDATE_CALLS;
2789

2790
	ftrace_startup_enable(command);
2791

2792 2793
	ops->flags &= ~FTRACE_OPS_FL_ADDING;

2794
	return 0;
2795 2796
}

2797
static int ftrace_shutdown(struct ftrace_ops *ops, int command)
2798
{
2799
	int ret;
2800

2801
	if (unlikely(ftrace_disabled))
2802 2803 2804 2805 2806
		return -ENODEV;

	ret = __unregister_ftrace_function(ops);
	if (ret)
		return ret;
2807

2808
	ftrace_start_up--;
2809 2810 2811 2812 2813 2814 2815
	/*
	 * 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);

2816 2817
	/* Disabling ipmodify never fails */
	ftrace_hash_ipmodify_disable(ops);
2818

2819 2820
	if (ftrace_hash_rec_disable(ops, 1))
		command |= FTRACE_UPDATE_CALLS;
2821

2822
	ops->flags &= ~FTRACE_OPS_FL_ENABLED;
2823

2824 2825 2826 2827
	if (saved_ftrace_func != ftrace_trace_function) {
		saved_ftrace_func = ftrace_trace_function;
		command |= FTRACE_UPDATE_TRACE_FUNC;
	}
2828

2829 2830
	if (!command || !ftrace_enabled) {
		/*
2831 2832
		 * If these are dynamic or per_cpu ops, they still
		 * need their data freed. Since, function tracing is
2833 2834 2835
		 * not currently active, we can just free them
		 * without synchronizing all CPUs.
		 */
2836 2837 2838
		if (ops->flags & (FTRACE_OPS_FL_DYNAMIC | FTRACE_OPS_FL_PER_CPU))
			goto free_ops;

2839
		return 0;
2840
	}
2841

2842 2843 2844 2845
	/*
	 * If the ops uses a trampoline, then it needs to be
	 * tested first on update.
	 */
2846
	ops->flags |= FTRACE_OPS_FL_REMOVING;
2847 2848
	removed_ops = ops;

2849 2850 2851 2852
	/* 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;

2853
	ftrace_run_update_code(command);
2854

2855 2856 2857 2858
	/*
	 * If there's no more ops registered with ftrace, run a
	 * sanity check to make sure all rec flags are cleared.
	 */
2859 2860
	if (rcu_dereference_protected(ftrace_ops_list,
			lockdep_is_held(&ftrace_lock)) == &ftrace_list_end) {
2861 2862 2863 2864
		struct ftrace_page *pg;
		struct dyn_ftrace *rec;

		do_for_each_ftrace_rec(pg, rec) {
2865
			if (FTRACE_WARN_ON_ONCE(rec->flags & ~FTRACE_FL_DISABLED))
2866 2867 2868 2869 2870
				pr_warn("  %pS flags:%lx\n",
					(void *)rec->ip, rec->flags);
		} while_for_each_ftrace_rec();
	}

2871 2872 2873 2874
	ops->old_hash.filter_hash = NULL;
	ops->old_hash.notrace_hash = NULL;

	removed_ops = NULL;
2875
	ops->flags &= ~FTRACE_OPS_FL_REMOVING;
2876

2877 2878 2879
	/*
	 * Dynamic ops may be freed, we must make sure that all
	 * callers are done before leaving this function.
2880
	 * The same goes for freeing the per_cpu data of the per_cpu
2881 2882
	 * ops.
	 */
2883
	if (ops->flags & (FTRACE_OPS_FL_DYNAMIC | FTRACE_OPS_FL_PER_CPU)) {
2884 2885 2886 2887 2888 2889 2890 2891
		/*
		 * 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.
		 */
2892 2893
		schedule_on_each_cpu(ftrace_sync);

2894 2895 2896 2897 2898 2899 2900 2901 2902 2903
		/*
		 * When the kernel is preeptive, tasks can be preempted
		 * while on a ftrace trampoline. Just scheduling a task on
		 * a CPU is not good enough to flush them. Calling
		 * synchornize_rcu_tasks() will wait for those tasks to
		 * execute and either schedule voluntarily or enter user space.
		 */
		if (IS_ENABLED(CONFIG_PREEMPT))
			synchronize_rcu_tasks();

2904
 free_ops:
2905 2906
		arch_ftrace_trampoline_free(ops);

2907 2908
		if (ops->flags & FTRACE_OPS_FL_PER_CPU)
			per_cpu_ops_free(ops);
2909 2910
	}

2911
	return 0;
2912 2913
}

I
Ingo Molnar 已提交
2914
static void ftrace_startup_sysctl(void)
2915
{
2916 2917
	int command;

2918 2919 2920
	if (unlikely(ftrace_disabled))
		return;

2921 2922
	/* Force update next time */
	saved_ftrace_func = NULL;
2923
	/* ftrace_start_up is true if we want ftrace running */
2924 2925 2926 2927
	if (ftrace_start_up) {
		command = FTRACE_UPDATE_CALLS;
		if (ftrace_graph_active)
			command |= FTRACE_START_FUNC_RET;
2928
		ftrace_startup_enable(command);
2929
	}
2930 2931
}

I
Ingo Molnar 已提交
2932
static void ftrace_shutdown_sysctl(void)
2933
{
2934 2935
	int command;

2936 2937 2938
	if (unlikely(ftrace_disabled))
		return;

2939
	/* ftrace_start_up is true if ftrace is running */
2940 2941 2942 2943 2944 2945
	if (ftrace_start_up) {
		command = FTRACE_DISABLE_CALLS;
		if (ftrace_graph_active)
			command |= FTRACE_STOP_FUNC_RET;
		ftrace_run_update_code(command);
	}
2946 2947
}

2948
static u64		ftrace_update_time;
2949 2950
unsigned long		ftrace_update_tot_cnt;

2951
static inline int ops_traces_mod(struct ftrace_ops *ops)
2952
{
2953 2954 2955 2956
	/*
	 * Filter_hash being empty will default to trace module.
	 * But notrace hash requires a test of individual module functions.
	 */
2957 2958
	return ftrace_hash_empty(ops->func_hash->filter_hash) &&
		ftrace_hash_empty(ops->func_hash->notrace_hash);
2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974
}

/*
 * 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;

2975
	/* If ops traces all then it includes this function */
2976
	if (ops_traces_mod(ops))
2977
		return 1;
2978 2979

	/* The function must be in the filter */
2980
	if (!ftrace_hash_empty(ops->func_hash->filter_hash) &&
2981
	    !__ftrace_lookup_ip(ops->func_hash->filter_hash, rec->ip))
2982
		return 0;
2983

2984
	/* If in notrace hash, we ignore it too */
2985
	if (ftrace_lookup_ip(ops->func_hash->notrace_hash, rec->ip))
2986 2987 2988 2989 2990
		return 0;

	return 1;
}

2991
static int ftrace_update_code(struct module *mod, struct ftrace_page *new_pgs)
2992
{
2993
	struct ftrace_page *pg;
2994
	struct dyn_ftrace *p;
2995
	u64 start, stop;
2996
	unsigned long update_cnt = 0;
2997
	unsigned long rec_flags = 0;
2998
	int i;
2999

3000 3001
	start = ftrace_now(raw_smp_processor_id());

3002
	/*
3003 3004 3005 3006 3007 3008 3009 3010 3011
	 * 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.
3012
	 */
3013 3014
	if (mod)
		rec_flags |= FTRACE_FL_DISABLED;
3015

3016
	for (pg = new_pgs; pg; pg = pg->next) {
3017

3018
		for (i = 0; i < pg->index; i++) {
3019

3020 3021 3022
			/* If something went wrong, bail without enabling anything */
			if (unlikely(ftrace_disabled))
				return -1;
3023

3024
			p = &pg->records[i];
3025
			p->flags = rec_flags;
3026

3027 3028 3029 3030 3031 3032
			/*
			 * Do the initial record conversion from mcount jump
			 * to the NOP instructions.
			 */
			if (!ftrace_code_disable(mod, p))
				break;
3033

3034
			update_cnt++;
3035
		}
3036 3037
	}

I
Ingo Molnar 已提交
3038
	stop = ftrace_now(raw_smp_processor_id());
3039
	ftrace_update_time = stop - start;
3040
	ftrace_update_tot_cnt += update_cnt;
3041

3042 3043 3044
	return 0;
}

3045
static int ftrace_allocate_records(struct ftrace_page *pg, int count)
3046
{
3047
	int order;
3048 3049
	int cnt;

3050 3051 3052 3053
	if (WARN_ON(!count))
		return -EINVAL;

	order = get_count_order(DIV_ROUND_UP(count, ENTRIES_PER_PAGE));
3054 3055

	/*
3056 3057
	 * We want to fill as much as possible. No more than a page
	 * may be empty.
3058
	 */
3059 3060
	while ((PAGE_SIZE << order) / ENTRY_SIZE >= count + ENTRIES_PER_PAGE)
		order--;
3061

3062 3063
 again:
	pg->records = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, order);
3064

3065 3066 3067 3068 3069 3070 3071
	if (!pg->records) {
		/* if we can't allocate this size, try something smaller */
		if (!order)
			return -ENOMEM;
		order >>= 1;
		goto again;
	}
3072

3073 3074
	cnt = (PAGE_SIZE << order) / ENTRY_SIZE;
	pg->size = cnt;
3075

3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108
	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)
3109 3110
			break;

3111 3112 3113 3114
		pg->next = kzalloc(sizeof(*pg), GFP_KERNEL);
		if (!pg->next)
			goto free_pages;

3115 3116 3117
		pg = pg->next;
	}

3118 3119 3120
	return start_pg;

 free_pages:
3121 3122
	pg = start_pg;
	while (pg) {
3123 3124 3125 3126 3127 3128 3129 3130 3131 3132
		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;
}

3133 3134 3135
#define FTRACE_BUFF_MAX (KSYM_SYMBOL_LEN+4) /* room for wildcards */

struct ftrace_iterator {
3136
	loff_t				pos;
3137
	loff_t				func_pos;
3138
	loff_t				mod_pos;
3139 3140 3141
	struct ftrace_page		*pg;
	struct dyn_ftrace		*func;
	struct ftrace_func_probe	*probe;
3142
	struct ftrace_func_entry	*probe_entry;
3143
	struct trace_parser		parser;
3144
	struct ftrace_hash		*hash;
3145
	struct ftrace_ops		*ops;
3146 3147
	struct trace_array		*tr;
	struct list_head		*mod_list;
3148
	int				pidx;
3149 3150
	int				idx;
	unsigned			flags;
3151 3152
};

3153
static void *
3154
t_probe_next(struct seq_file *m, loff_t *pos)
3155 3156
{
	struct ftrace_iterator *iter = m->private;
3157
	struct trace_array *tr = iter->ops->private;
3158
	struct list_head *func_probes;
3159 3160
	struct ftrace_hash *hash;
	struct list_head *next;
3161
	struct hlist_node *hnd = NULL;
3162
	struct hlist_head *hhd;
3163
	int size;
3164 3165

	(*pos)++;
3166
	iter->pos = *pos;
3167

3168
	if (!tr)
3169 3170
		return NULL;

3171 3172
	func_probes = &tr->func_probes;
	if (list_empty(func_probes))
3173 3174
		return NULL;

3175
	if (!iter->probe) {
3176
		next = func_probes->next;
3177
		iter->probe = list_entry(next, struct ftrace_func_probe, list);
3178 3179 3180 3181 3182 3183 3184 3185 3186 3187
	}

	if (iter->probe_entry)
		hnd = &iter->probe_entry->hlist;

	hash = iter->probe->ops.func_hash->filter_hash;
	size = 1 << hash->size_bits;

 retry:
	if (iter->pidx >= size) {
3188
		if (iter->probe->list.next == func_probes)
3189 3190
			return NULL;
		next = iter->probe->list.next;
3191
		iter->probe = list_entry(next, struct ftrace_func_probe, list);
3192 3193 3194 3195 3196 3197
		hash = iter->probe->ops.func_hash->filter_hash;
		size = 1 << hash->size_bits;
		iter->pidx = 0;
	}

	hhd = &hash->buckets[iter->pidx];
3198 3199

	if (hlist_empty(hhd)) {
3200
		iter->pidx++;
3201 3202 3203 3204 3205 3206 3207 3208 3209
		hnd = NULL;
		goto retry;
	}

	if (!hnd)
		hnd = hhd->first;
	else {
		hnd = hnd->next;
		if (!hnd) {
3210
			iter->pidx++;
3211 3212 3213 3214
			goto retry;
		}
	}

3215 3216 3217
	if (WARN_ON_ONCE(!hnd))
		return NULL;

3218
	iter->probe_entry = hlist_entry(hnd, struct ftrace_func_entry, hlist);
3219 3220

	return iter;
3221 3222
}

3223
static void *t_probe_start(struct seq_file *m, loff_t *pos)
3224 3225 3226
{
	struct ftrace_iterator *iter = m->private;
	void *p = NULL;
L
Li Zefan 已提交
3227 3228
	loff_t l;

3229
	if (!(iter->flags & FTRACE_ITER_DO_PROBES))
3230 3231
		return NULL;

3232
	if (iter->mod_pos > *pos)
3233
		return NULL;
3234

3235 3236 3237
	iter->probe = NULL;
	iter->probe_entry = NULL;
	iter->pidx = 0;
3238
	for (l = 0; l <= (*pos - iter->mod_pos); ) {
3239
		p = t_probe_next(m, &l);
L
Li Zefan 已提交
3240 3241 3242
		if (!p)
			break;
	}
3243 3244 3245
	if (!p)
		return NULL;

3246
	/* Only set this if we have an item */
3247
	iter->flags |= FTRACE_ITER_PROBE;
3248

3249
	return iter;
3250 3251
}

3252
static int
3253
t_probe_show(struct seq_file *m, struct ftrace_iterator *iter)
3254
{
3255
	struct ftrace_func_entry *probe_entry;
3256 3257
	struct ftrace_probe_ops *probe_ops;
	struct ftrace_func_probe *probe;
3258

3259 3260
	probe = iter->probe;
	probe_entry = iter->probe_entry;
3261

3262
	if (WARN_ON_ONCE(!probe || !probe_entry))
3263
		return -EIO;
3264

3265
	probe_ops = probe->probe_ops;
3266

3267
	if (probe_ops->print)
3268
		return probe_ops->print(m, probe_entry->ip, probe_ops, probe->data);
3269

3270 3271
	seq_printf(m, "%ps:%ps\n", (void *)probe_entry->ip,
		   (void *)probe_ops->func);
3272 3273 3274 3275

	return 0;
}

3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351
static void *
t_mod_next(struct seq_file *m, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
	struct trace_array *tr = iter->tr;

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

	iter->mod_list = iter->mod_list->next;

	if (iter->mod_list == &tr->mod_trace ||
	    iter->mod_list == &tr->mod_notrace) {
		iter->flags &= ~FTRACE_ITER_MOD;
		return NULL;
	}

	iter->mod_pos = *pos;

	return iter;
}

static void *t_mod_start(struct seq_file *m, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
	void *p = NULL;
	loff_t l;

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

	iter->mod_pos = iter->func_pos;

	/* probes are only available if tr is set */
	if (!iter->tr)
		return NULL;

	for (l = 0; l <= (*pos - iter->func_pos); ) {
		p = t_mod_next(m, &l);
		if (!p)
			break;
	}
	if (!p) {
		iter->flags &= ~FTRACE_ITER_MOD;
		return t_probe_start(m, pos);
	}

	/* Only set this if we have an item */
	iter->flags |= FTRACE_ITER_MOD;

	return iter;
}

static int
t_mod_show(struct seq_file *m, struct ftrace_iterator *iter)
{
	struct ftrace_mod_load *ftrace_mod;
	struct trace_array *tr = iter->tr;

	if (WARN_ON_ONCE(!iter->mod_list) ||
			 iter->mod_list == &tr->mod_trace ||
			 iter->mod_list == &tr->mod_notrace)
		return -EIO;

	ftrace_mod = list_entry(iter->mod_list, struct ftrace_mod_load, list);

	if (ftrace_mod->func)
		seq_printf(m, "%s", ftrace_mod->func);
	else
		seq_putc(m, '*');

	seq_printf(m, ":mod:%s\n", ftrace_mod->module);

	return 0;
}

I
Ingo Molnar 已提交
3352
static void *
3353
t_func_next(struct seq_file *m, loff_t *pos)
3354 3355 3356 3357 3358
{
	struct ftrace_iterator *iter = m->private;
	struct dyn_ftrace *rec = NULL;

	(*pos)++;
3359

3360 3361 3362 3363 3364 3365 3366 3367 3368
 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++];
3369 3370
		if (((iter->flags & (FTRACE_ITER_FILTER | FTRACE_ITER_NOTRACE)) &&
		     !ftrace_lookup_ip(iter->hash, rec->ip)) ||
3371 3372

		    ((iter->flags & FTRACE_ITER_ENABLED) &&
3373
		     !(rec->flags & FTRACE_FL_ENABLED))) {
3374

3375 3376 3377 3378 3379
			rec = NULL;
			goto retry;
		}
	}

3380
	if (!rec)
3381
		return NULL;
3382

3383
	iter->pos = iter->func_pos = *pos;
3384 3385 3386
	iter->func = rec;

	return iter;
3387 3388
}

3389 3390 3391 3392
static void *
t_next(struct seq_file *m, void *v, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
3393
	loff_t l = *pos; /* t_probe_start() must use original pos */
3394 3395 3396 3397 3398
	void *ret;

	if (unlikely(ftrace_disabled))
		return NULL;

3399 3400
	if (iter->flags & FTRACE_ITER_PROBE)
		return t_probe_next(m, pos);
3401

3402 3403 3404
	if (iter->flags & FTRACE_ITER_MOD)
		return t_mod_next(m, pos);

3405
	if (iter->flags & FTRACE_ITER_PRINTALL) {
3406
		/* next must increment pos, and t_probe_start does not */
3407
		(*pos)++;
3408
		return t_mod_start(m, &l);
3409 3410 3411 3412 3413
	}

	ret = t_func_next(m, pos);

	if (!ret)
3414
		return t_mod_start(m, &l);
3415 3416 3417 3418

	return ret;
}

3419 3420 3421 3422
static void reset_iter_read(struct ftrace_iterator *iter)
{
	iter->pos = 0;
	iter->func_pos = 0;
3423
	iter->flags &= ~(FTRACE_ITER_PRINTALL | FTRACE_ITER_PROBE | FTRACE_ITER_MOD);
3424 3425 3426 3427 3428 3429
}

static void *t_start(struct seq_file *m, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
	void *p = NULL;
3430
	loff_t l;
3431

3432
	mutex_lock(&ftrace_lock);
3433 3434 3435 3436

	if (unlikely(ftrace_disabled))
		return NULL;

3437 3438 3439 3440 3441 3442
	/*
	 * If an lseek was done, then reset and start from beginning.
	 */
	if (*pos < iter->pos)
		reset_iter_read(iter);

3443 3444 3445 3446 3447
	/*
	 * For set_ftrace_filter reading, if we have the filter
	 * off, we can short cut and just print out that all
	 * functions are enabled.
	 */
3448 3449
	if ((iter->flags & (FTRACE_ITER_FILTER | FTRACE_ITER_NOTRACE)) &&
	    ftrace_hash_empty(iter->hash)) {
3450
		iter->func_pos = 1; /* Account for the message */
3451
		if (*pos > 0)
3452
			return t_mod_start(m, pos);
3453
		iter->flags |= FTRACE_ITER_PRINTALL;
3454
		/* reset in case of seek/pread */
3455
		iter->flags &= ~FTRACE_ITER_PROBE;
3456 3457 3458
		return iter;
	}

3459 3460
	if (iter->flags & FTRACE_ITER_MOD)
		return t_mod_start(m, pos);
3461

3462 3463 3464 3465 3466
	/*
	 * 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.
	 */
3467 3468 3469
	iter->pg = ftrace_pages_start;
	iter->idx = 0;
	for (l = 0; l <= *pos; ) {
3470
		p = t_func_next(m, &l);
3471 3472
		if (!p)
			break;
3473
	}
3474

3475
	if (!p)
3476
		return t_mod_start(m, pos);
3477 3478

	return iter;
3479 3480 3481 3482
}

static void t_stop(struct seq_file *m, void *p)
{
3483
	mutex_unlock(&ftrace_lock);
3484 3485
}

3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501
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);
}

3502 3503
static int t_show(struct seq_file *m, void *v)
{
3504
	struct ftrace_iterator *iter = m->private;
3505
	struct dyn_ftrace *rec;
3506

3507 3508
	if (iter->flags & FTRACE_ITER_PROBE)
		return t_probe_show(m, iter);
3509

3510 3511 3512
	if (iter->flags & FTRACE_ITER_MOD)
		return t_mod_show(m, iter);

3513
	if (iter->flags & FTRACE_ITER_PRINTALL) {
3514
		if (iter->flags & FTRACE_ITER_NOTRACE)
3515
			seq_puts(m, "#### no functions disabled ####\n");
3516
		else
3517
			seq_puts(m, "#### all functions enabled ####\n");
3518 3519 3520
		return 0;
	}

3521 3522
	rec = iter->func;

3523 3524 3525
	if (!rec)
		return 0;

3526
	seq_printf(m, "%ps", (void *)rec->ip);
3527
	if (iter->flags & FTRACE_ITER_ENABLED) {
3528
		struct ftrace_ops *ops;
3529

3530
		seq_printf(m, " (%ld)%s%s",
3531
			   ftrace_rec_count(rec),
3532 3533
			   rec->flags & FTRACE_FL_REGS ? " R" : "  ",
			   rec->flags & FTRACE_FL_IPMODIFY ? " I" : "  ");
3534
		if (rec->flags & FTRACE_FL_TRAMP_EN) {
3535
			ops = ftrace_find_tramp_ops_any(rec);
3536 3537 3538 3539 3540
			if (ops) {
				do {
					seq_printf(m, "\ttramp: %pS (%pS)",
						   (void *)ops->trampoline,
						   (void *)ops->func);
3541
					add_trampoline_func(m, ops, rec);
3542 3543 3544
					ops = ftrace_find_tramp_ops_next(rec, ops);
				} while (ops);
			} else
3545
				seq_puts(m, "\ttramp: ERROR!");
3546 3547
		} else {
			add_trampoline_func(m, NULL, rec);
3548 3549 3550
		}
	}	

3551
	seq_putc(m, '\n');
3552 3553 3554 3555

	return 0;
}

J
James Morris 已提交
3556
static const struct seq_operations show_ftrace_seq_ops = {
3557 3558 3559 3560 3561 3562
	.start = t_start,
	.next = t_next,
	.stop = t_stop,
	.show = t_show,
};

I
Ingo Molnar 已提交
3563
static int
3564 3565 3566 3567
ftrace_avail_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;

3568 3569 3570
	if (unlikely(ftrace_disabled))
		return -ENODEV;

3571
	iter = __seq_open_private(file, &show_ftrace_seq_ops, sizeof(*iter));
3572 3573
	if (!iter)
		return -ENOMEM;
3574

3575 3576 3577 3578
	iter->pg = ftrace_pages_start;
	iter->ops = &global_ops;

	return 0;
3579 3580
}

3581 3582 3583 3584 3585
static int
ftrace_enabled_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;

3586
	iter = __seq_open_private(file, &show_ftrace_seq_ops, sizeof(*iter));
3587 3588
	if (!iter)
		return -ENOMEM;
3589

3590 3591 3592 3593 3594
	iter->pg = ftrace_pages_start;
	iter->flags = FTRACE_ITER_ENABLED;
	iter->ops = &global_ops;

	return 0;
3595 3596
}

3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609
/**
 * 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.
3610
 * tracing_lseek() should be used as the lseek routine, and
3611 3612 3613
 * release must call ftrace_regex_release().
 */
int
3614
ftrace_regex_open(struct ftrace_ops *ops, int flag,
3615
		  struct inode *inode, struct file *file)
3616 3617
{
	struct ftrace_iterator *iter;
3618
	struct ftrace_hash *hash;
3619 3620
	struct list_head *mod_head;
	struct trace_array *tr = ops->private;
3621 3622
	int ret = 0;

3623 3624
	ftrace_ops_init(ops);

3625 3626 3627
	if (unlikely(ftrace_disabled))
		return -ENODEV;

3628 3629 3630 3631
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
	if (!iter)
		return -ENOMEM;

3632 3633 3634 3635 3636
	if (trace_parser_get_init(&iter->parser, FTRACE_BUFF_MAX)) {
		kfree(iter);
		return -ENOMEM;
	}

3637 3638
	iter->ops = ops;
	iter->flags = flag;
3639
	iter->tr = tr;
3640

3641
	mutex_lock(&ops->func_hash->regex_lock);
3642

3643
	if (flag & FTRACE_ITER_NOTRACE) {
3644
		hash = ops->func_hash->notrace_hash;
3645
		mod_head = tr ? &tr->mod_notrace : NULL;
3646
	} else {
3647
		hash = ops->func_hash->filter_hash;
3648
		mod_head = tr ? &tr->mod_trace : NULL;
3649
	}
3650

3651 3652
	iter->mod_list = mod_head;

3653
	if (file->f_mode & FMODE_WRITE) {
3654 3655
		const int size_bits = FTRACE_HASH_DEFAULT_BITS;

3656
		if (file->f_flags & O_TRUNC) {
3657
			iter->hash = alloc_ftrace_hash(size_bits);
3658 3659
			clear_ftrace_mod_list(mod_head);
	        } else {
3660
			iter->hash = alloc_and_copy_ftrace_hash(size_bits, hash);
3661
		}
3662

3663 3664 3665
		if (!iter->hash) {
			trace_parser_put(&iter->parser);
			kfree(iter);
3666 3667
			ret = -ENOMEM;
			goto out_unlock;
3668
		}
3669 3670
	} else
		iter->hash = hash;
3671

3672 3673 3674 3675 3676 3677 3678
	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;
3679
		} else {
3680 3681
			/* Failed */
			free_ftrace_hash(iter->hash);
3682
			trace_parser_put(&iter->parser);
3683
			kfree(iter);
3684
		}
3685 3686
	} else
		file->private_data = iter;
3687 3688

 out_unlock:
3689
	mutex_unlock(&ops->func_hash->regex_lock);
3690 3691 3692 3693

	return ret;
}

3694 3695 3696
static int
ftrace_filter_open(struct inode *inode, struct file *file)
{
3697 3698 3699
	struct ftrace_ops *ops = inode->i_private;

	return ftrace_regex_open(ops,
3700
			FTRACE_ITER_FILTER | FTRACE_ITER_DO_PROBES,
3701
			inode, file);
3702 3703 3704 3705 3706
}

static int
ftrace_notrace_open(struct inode *inode, struct file *file)
{
3707 3708 3709
	struct ftrace_ops *ops = inode->i_private;

	return ftrace_regex_open(ops, FTRACE_ITER_NOTRACE,
3710
				 inode, file);
3711 3712
}

3713 3714 3715 3716 3717 3718 3719
/* Type for quick search ftrace basic regexes (globs) from filter_parse_regex */
struct ftrace_glob {
	char *search;
	unsigned len;
	int type;
};

3720 3721 3722 3723 3724 3725 3726 3727 3728 3729
/*
 * 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;
}

3730
static int ftrace_match(char *str, struct ftrace_glob *g)
3731 3732
{
	int matched = 0;
3733
	int slen;
3734

3735 3736
	str = arch_ftrace_match_adjust(str, g->search);

3737
	switch (g->type) {
3738
	case MATCH_FULL:
3739
		if (strcmp(str, g->search) == 0)
3740 3741 3742
			matched = 1;
		break;
	case MATCH_FRONT_ONLY:
3743
		if (strncmp(str, g->search, g->len) == 0)
3744 3745 3746
			matched = 1;
		break;
	case MATCH_MIDDLE_ONLY:
3747
		if (strstr(str, g->search))
3748 3749 3750
			matched = 1;
		break;
	case MATCH_END_ONLY:
3751
		slen = strlen(str);
3752 3753
		if (slen >= g->len &&
		    memcmp(str + slen - g->len, g->search, g->len) == 0)
3754 3755
			matched = 1;
		break;
3756 3757 3758 3759
	case MATCH_GLOB:
		if (glob_match(g->search, str))
			matched = 1;
		break;
3760 3761 3762 3763 3764
	}

	return matched;
}

3765
static int
3766
enter_record(struct ftrace_hash *hash, struct dyn_ftrace *rec, int clear_filter)
3767
{
3768 3769 3770
	struct ftrace_func_entry *entry;
	int ret = 0;

3771
	entry = ftrace_lookup_ip(hash, rec->ip);
3772
	if (clear_filter) {
3773 3774 3775
		/* Do nothing if it doesn't exist */
		if (!entry)
			return 0;
3776

3777
		free_hash_entry(hash, entry);
3778 3779 3780 3781
	} else {
		/* Do nothing if it exists */
		if (entry)
			return 0;
3782

3783
		ret = add_hash_entry(hash, rec->ip);
3784 3785
	}
	return ret;
3786 3787
}

3788
static int
D
Dmitry Safonov 已提交
3789 3790
ftrace_match_record(struct dyn_ftrace *rec, struct ftrace_glob *func_g,
		struct ftrace_glob *mod_g, int exclude_mod)
3791 3792
{
	char str[KSYM_SYMBOL_LEN];
3793 3794 3795 3796
	char *modname;

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

D
Dmitry Safonov 已提交
3797 3798 3799 3800 3801 3802
	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 */
3803
			if (!exclude_mod != !modname)
D
Dmitry Safonov 已提交
3804 3805 3806 3807
				goto func_match;
			return 0;
		}

3808 3809 3810 3811 3812 3813 3814 3815
		/*
		 * exclude_mod is set to trace everything but the given
		 * module. If it is set and the module matches, then
		 * return 0. If it is not set, and the module doesn't match
		 * also return 0. Otherwise, check the function to see if
		 * that matches.
		 */
		if (!mod_matches == !exclude_mod)
3816
			return 0;
D
Dmitry Safonov 已提交
3817
func_match:
3818
		/* blank search means to match all funcs in the mod */
3819
		if (!func_g->len)
3820 3821
			return 1;
	}
3822

3823
	return ftrace_match(str, func_g);
3824 3825
}

3826
static int
3827
match_records(struct ftrace_hash *hash, char *func, int len, char *mod)
3828 3829 3830
{
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
3831
	struct ftrace_glob func_g = { .type = MATCH_FULL };
D
Dmitry Safonov 已提交
3832 3833 3834
	struct ftrace_glob mod_g = { .type = MATCH_FULL };
	struct ftrace_glob *mod_match = (mod) ? &mod_g : NULL;
	int exclude_mod = 0;
3835
	int found = 0;
3836
	int ret;
3837
	int clear_filter = 0;
3838

D
Dmitry Safonov 已提交
3839
	if (func) {
3840 3841 3842
		func_g.type = filter_parse_regex(func, len, &func_g.search,
						 &clear_filter);
		func_g.len = strlen(func_g.search);
3843
	}
3844

D
Dmitry Safonov 已提交
3845 3846 3847 3848
	if (mod) {
		mod_g.type = filter_parse_regex(mod, strlen(mod),
				&mod_g.search, &exclude_mod);
		mod_g.len = strlen(mod_g.search);
3849
	}
3850

3851
	mutex_lock(&ftrace_lock);
3852

3853 3854
	if (unlikely(ftrace_disabled))
		goto out_unlock;
3855

3856
	do_for_each_ftrace_rec(pg, rec) {
3857 3858 3859 3860

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

D
Dmitry Safonov 已提交
3861
		if (ftrace_match_record(rec, &func_g, mod_match, exclude_mod)) {
3862
			ret = enter_record(hash, rec, clear_filter);
3863 3864 3865 3866
			if (ret < 0) {
				found = ret;
				goto out_unlock;
			}
3867
			found = 1;
3868 3869
		}
	} while_for_each_ftrace_rec();
3870
 out_unlock:
3871
	mutex_unlock(&ftrace_lock);
3872 3873

	return found;
3874 3875
}

3876
static int
3877
ftrace_match_records(struct ftrace_hash *hash, char *buff, int len)
3878
{
3879
	return match_records(hash, buff, len, NULL);
3880 3881
}

3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931
static void ftrace_ops_update_code(struct ftrace_ops *ops,
				   struct ftrace_ops_hash *old_hash)
{
	struct ftrace_ops *op;

	if (!ftrace_enabled)
		return;

	if (ops->flags & FTRACE_OPS_FL_ENABLED) {
		ftrace_run_modify_code(ops, FTRACE_UPDATE_CALLS, old_hash);
		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);
}

static int ftrace_hash_move_and_update_ops(struct ftrace_ops *ops,
					   struct ftrace_hash **orig_hash,
					   struct ftrace_hash *hash,
					   int enable)
{
	struct ftrace_ops_hash old_hash_ops;
	struct ftrace_hash *old_hash;
	int ret;

	old_hash = *orig_hash;
	old_hash_ops.filter_hash = ops->func_hash->filter_hash;
	old_hash_ops.notrace_hash = ops->func_hash->notrace_hash;
	ret = ftrace_hash_move(ops, enable, orig_hash, hash);
	if (!ret) {
		ftrace_ops_update_code(ops, &old_hash_ops);
		free_ftrace_hash_rcu(old_hash);
	}
	return ret;
}
3932

3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970
static bool module_exists(const char *module)
{
	/* All modules have the symbol __this_module */
	const char this_mod[] = "__this_module";
	const int modname_size = MAX_PARAM_PREFIX_LEN + sizeof(this_mod) + 1;
	char modname[modname_size + 1];
	unsigned long val;
	int n;

	n = snprintf(modname, modname_size + 1, "%s:%s", module, this_mod);

	if (n > modname_size)
		return false;

	val = module_kallsyms_lookup_name(modname);
	return val != 0;
}

static int cache_mod(struct trace_array *tr,
		     const char *func, char *module, int enable)
{
	struct ftrace_mod_load *ftrace_mod, *n;
	struct list_head *head = enable ? &tr->mod_trace : &tr->mod_notrace;
	int ret;

	mutex_lock(&ftrace_lock);

	/* We do not cache inverse filters */
	if (func[0] == '!') {
		func++;
		ret = -EINVAL;

		/* Look to remove this hash */
		list_for_each_entry_safe(ftrace_mod, n, head, list) {
			if (strcmp(ftrace_mod->module, module) != 0)
				continue;

			/* no func matches all */
3971
			if (strcmp(func, "*") == 0 ||
3972 3973 3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994
			    (ftrace_mod->func &&
			     strcmp(ftrace_mod->func, func) == 0)) {
				ret = 0;
				free_ftrace_mod(ftrace_mod);
				continue;
			}
		}
		goto out;
	}

	ret = -EINVAL;
	/* We only care about modules that have not been loaded yet */
	if (module_exists(module))
		goto out;

	/* Save this string off, and execute it when the module is loaded */
	ret = ftrace_add_mod(tr, func, module, enable);
 out:
	mutex_unlock(&ftrace_lock);

	return ret;
}

3995 3996 3997 3998
static int
ftrace_set_regex(struct ftrace_ops *ops, unsigned char *buf, int len,
		 int reset, int enable);

3999
#ifdef CONFIG_MODULES
4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018
static void process_mod_list(struct list_head *head, struct ftrace_ops *ops,
			     char *mod, bool enable)
{
	struct ftrace_mod_load *ftrace_mod, *n;
	struct ftrace_hash **orig_hash, *new_hash;
	LIST_HEAD(process_mods);
	char *func;
	int ret;

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

	if (enable)
		orig_hash = &ops->func_hash->filter_hash;
	else
		orig_hash = &ops->func_hash->notrace_hash;

	new_hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS,
					      *orig_hash);
	if (!new_hash)
4019
		goto out; /* warn? */
4020 4021 4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050 4051 4052 4053 4054

	mutex_lock(&ftrace_lock);

	list_for_each_entry_safe(ftrace_mod, n, head, list) {

		if (strcmp(ftrace_mod->module, mod) != 0)
			continue;

		if (ftrace_mod->func)
			func = kstrdup(ftrace_mod->func, GFP_KERNEL);
		else
			func = kstrdup("*", GFP_KERNEL);

		if (!func) /* warn? */
			continue;

		list_del(&ftrace_mod->list);
		list_add(&ftrace_mod->list, &process_mods);

		/* Use the newly allocated func, as it may be "*" */
		kfree(ftrace_mod->func);
		ftrace_mod->func = func;
	}

	mutex_unlock(&ftrace_lock);

	list_for_each_entry_safe(ftrace_mod, n, &process_mods, list) {

		func = ftrace_mod->func;

		/* Grabs ftrace_lock, which is why we have this extra step */
		match_records(new_hash, func, strlen(func), mod);
		free_ftrace_mod(ftrace_mod);
	}

4055 4056 4057
	if (enable && list_empty(head))
		new_hash->flags &= ~FTRACE_HASH_FL_MOD;

4058 4059 4060 4061 4062 4063
	mutex_lock(&ftrace_lock);

	ret = ftrace_hash_move_and_update_ops(ops, orig_hash,
					      new_hash, enable);
	mutex_unlock(&ftrace_lock);

4064
 out:
4065 4066 4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084 4085 4086 4087 4088 4089
	mutex_unlock(&ops->func_hash->regex_lock);

	free_ftrace_hash(new_hash);
}

static void process_cached_mods(const char *mod_name)
{
	struct trace_array *tr;
	char *mod;

	mod = kstrdup(mod_name, GFP_KERNEL);
	if (!mod)
		return;

	mutex_lock(&trace_types_lock);
	list_for_each_entry(tr, &ftrace_trace_arrays, list) {
		if (!list_empty(&tr->mod_trace))
			process_mod_list(&tr->mod_trace, tr->ops, mod, true);
		if (!list_empty(&tr->mod_notrace))
			process_mod_list(&tr->mod_notrace, tr->ops, mod, false);
	}
	mutex_unlock(&trace_types_lock);

	kfree(mod);
}
4090
#endif
4091

4092 4093 4094 4095 4096 4097
/*
 * We register the module command as a template to show others how
 * to register the a command as well.
 */

static int
4098
ftrace_mod_callback(struct trace_array *tr, struct ftrace_hash *hash,
4099
		    char *func_orig, char *cmd, char *module, int enable)
4100
{
4101
	char *func;
4102
	int ret;
4103

4104 4105 4106 4107 4108
	/* match_records() modifies func, and we need the original */
	func = kstrdup(func_orig, GFP_KERNEL);
	if (!func)
		return -ENOMEM;

4109 4110 4111 4112 4113 4114 4115
	/*
	 * 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.
	 */
4116
	ret = match_records(hash, func, strlen(func), module);
4117 4118
	kfree(func);

4119
	if (!ret)
4120
		return cache_mod(tr, func_orig, module, enable);
4121 4122 4123
	if (ret < 0)
		return ret;
	return 0;
4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134
}

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);
}
4135
core_initcall(ftrace_mod_cmd_init);
4136

4137
static void function_trace_probe_call(unsigned long ip, unsigned long parent_ip,
4138
				      struct ftrace_ops *op, struct pt_regs *pt_regs)
4139
{
4140
	struct ftrace_probe_ops *probe_ops;
4141
	struct ftrace_func_probe *probe;
4142

4143 4144
	probe = container_of(op, struct ftrace_func_probe, ops);
	probe_ops = probe->probe_ops;
4145 4146 4147 4148 4149 4150

	/*
	 * 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.
	 */
4151
	preempt_disable_notrace();
4152
	probe_ops->func(ip, parent_ip, probe->tr, probe_ops, probe->data);
4153
	preempt_enable_notrace();
4154 4155
}

4156 4157 4158
struct ftrace_func_map {
	struct ftrace_func_entry	entry;
	void				*data;
4159 4160
};

4161 4162 4163
struct ftrace_func_mapper {
	struct ftrace_hash		hash;
};
4164

4165 4166 4167 4168 4169 4170
/**
 * allocate_ftrace_func_mapper - allocate a new ftrace_func_mapper
 *
 * Returns a ftrace_func_mapper descriptor that can be used to map ips to data.
 */
struct ftrace_func_mapper *allocate_ftrace_func_mapper(void)
4171
{
4172
	struct ftrace_hash *hash;
4173

4174 4175 4176 4177 4178 4179 4180 4181
	/*
	 * The mapper is simply a ftrace_hash, but since the entries
	 * in the hash are not ftrace_func_entry type, we define it
	 * as a separate structure.
	 */
	hash = alloc_ftrace_hash(FTRACE_HASH_DEFAULT_BITS);
	return (struct ftrace_func_mapper *)hash;
}
4182

4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193 4194 4195 4196 4197 4198
/**
 * ftrace_func_mapper_find_ip - Find some data mapped to an ip
 * @mapper: The mapper that has the ip maps
 * @ip: the instruction pointer to find the data for
 *
 * Returns the data mapped to @ip if found otherwise NULL. The return
 * is actually the address of the mapper data pointer. The address is
 * returned for use cases where the data is no bigger than a long, and
 * the user can use the data pointer as its data instead of having to
 * allocate more memory for the reference.
 */
void **ftrace_func_mapper_find_ip(struct ftrace_func_mapper *mapper,
				  unsigned long ip)
{
	struct ftrace_func_entry *entry;
	struct ftrace_func_map *map;
4199

4200 4201 4202
	entry = ftrace_lookup_ip(&mapper->hash, ip);
	if (!entry)
		return NULL;
4203

4204 4205
	map = (struct ftrace_func_map *)entry;
	return &map->data;
4206 4207
}

4208 4209 4210 4211 4212 4213 4214 4215 4216 4217
/**
 * ftrace_func_mapper_add_ip - Map some data to an ip
 * @mapper: The mapper that has the ip maps
 * @ip: The instruction pointer address to map @data to
 * @data: The data to map to @ip
 *
 * Returns 0 on succes otherwise an error.
 */
int ftrace_func_mapper_add_ip(struct ftrace_func_mapper *mapper,
			      unsigned long ip, void *data)
4218
{
4219 4220
	struct ftrace_func_entry *entry;
	struct ftrace_func_map *map;
4221

4222 4223 4224
	entry = ftrace_lookup_ip(&mapper->hash, ip);
	if (entry)
		return -EBUSY;
4225

4226 4227 4228
	map = kmalloc(sizeof(*map), GFP_KERNEL);
	if (!map)
		return -ENOMEM;
4229

4230 4231
	map->entry.ip = ip;
	map->data = data;
4232

4233
	__add_hash_entry(&mapper->hash, &map->entry);
4234

4235 4236
	return 0;
}
4237

4238 4239 4240 4241 4242 4243 4244 4245 4246 4247 4248 4249
/**
 * ftrace_func_mapper_remove_ip - Remove an ip from the mapping
 * @mapper: The mapper that has the ip maps
 * @ip: The instruction pointer address to remove the data from
 *
 * Returns the data if it is found, otherwise NULL.
 * Note, if the data pointer is used as the data itself, (see 
 * ftrace_func_mapper_find_ip(), then the return value may be meaningless,
 * if the data pointer was set to zero.
 */
void *ftrace_func_mapper_remove_ip(struct ftrace_func_mapper *mapper,
				   unsigned long ip)
4250
{
4251 4252 4253 4254 4255 4256 4257 4258 4259 4260 4261 4262
	struct ftrace_func_entry *entry;
	struct ftrace_func_map *map;
	void *data;

	entry = ftrace_lookup_ip(&mapper->hash, ip);
	if (!entry)
		return NULL;

	map = (struct ftrace_func_map *)entry;
	data = map->data;

	remove_hash_entry(&mapper->hash, entry);
4263
	kfree(entry);
4264 4265 4266 4267 4268 4269 4270 4271 4272 4273 4274 4275 4276 4277 4278 4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295 4296

	return data;
}

/**
 * free_ftrace_func_mapper - free a mapping of ips and data
 * @mapper: The mapper that has the ip maps
 * @free_func: A function to be called on each data item.
 *
 * This is used to free the function mapper. The @free_func is optional
 * and can be used if the data needs to be freed as well.
 */
void free_ftrace_func_mapper(struct ftrace_func_mapper *mapper,
			     ftrace_mapper_func free_func)
{
	struct ftrace_func_entry *entry;
	struct ftrace_func_map *map;
	struct hlist_head *hhd;
	int size = 1 << mapper->hash.size_bits;
	int i;

	if (free_func && mapper->hash.count) {
		for (i = 0; i < size; i++) {
			hhd = &mapper->hash.buckets[i];
			hlist_for_each_entry(entry, hhd, hlist) {
				map = (struct ftrace_func_map *)entry;
				free_func(map);
			}
		}
	}
	free_ftrace_hash(&mapper->hash);
}

4297 4298 4299 4300 4301 4302 4303 4304 4305 4306 4307 4308 4309
static void release_probe(struct ftrace_func_probe *probe)
{
	struct ftrace_probe_ops *probe_ops;

	mutex_lock(&ftrace_lock);

	WARN_ON(probe->ref <= 0);

	/* Subtract the ref that was used to protect this instance */
	probe->ref--;

	if (!probe->ref) {
		probe_ops = probe->probe_ops;
4310 4311 4312 4313 4314 4315
		/*
		 * Sending zero as ip tells probe_ops to free
		 * the probe->data itself
		 */
		if (probe_ops->free)
			probe_ops->free(probe_ops, probe->tr, 0, probe->data);
4316 4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328
		list_del(&probe->list);
		kfree(probe);
	}
	mutex_unlock(&ftrace_lock);
}

static void acquire_probe_locked(struct ftrace_func_probe *probe)
{
	/*
	 * Add one ref to keep it from being freed when releasing the
	 * ftrace_lock mutex.
	 */
	probe->ref++;
4329 4330 4331
}

int
4332
register_ftrace_function_probe(char *glob, struct trace_array *tr,
4333 4334
			       struct ftrace_probe_ops *probe_ops,
			       void *data)
4335
{
4336
	struct ftrace_func_entry *entry;
4337
	struct ftrace_func_probe *probe;
4338 4339
	struct ftrace_hash **orig_hash;
	struct ftrace_hash *old_hash;
4340
	struct ftrace_hash *hash;
4341
	int count = 0;
4342
	int size;
4343
	int ret;
4344
	int i;
4345

4346
	if (WARN_ON(!tr))
4347 4348
		return -EINVAL;

4349 4350
	/* We do not support '!' for function probes */
	if (WARN_ON(glob[0] == '!'))
4351 4352
		return -EINVAL;

4353

4354 4355 4356 4357 4358
	mutex_lock(&ftrace_lock);
	/* Check if the probe_ops is already registered */
	list_for_each_entry(probe, &tr->func_probes, list) {
		if (probe->probe_ops == probe_ops)
			break;
4359
	}
4360 4361 4362 4363 4364 4365 4366 4367 4368 4369 4370
	if (&probe->list == &tr->func_probes) {
		probe = kzalloc(sizeof(*probe), GFP_KERNEL);
		if (!probe) {
			mutex_unlock(&ftrace_lock);
			return -ENOMEM;
		}
		probe->probe_ops = probe_ops;
		probe->ops.func = function_trace_probe_call;
		probe->tr = tr;
		ftrace_ops_init(&probe->ops);
		list_add(&probe->list, &tr->func_probes);
4371
	}
4372

4373
	acquire_probe_locked(probe);
4374

4375
	mutex_unlock(&ftrace_lock);
4376

4377
	mutex_lock(&probe->ops.func_hash->regex_lock);
4378

4379
	orig_hash = &probe->ops.func_hash->filter_hash;
4380 4381
	old_hash = *orig_hash;
	hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, old_hash);
4382

4383
	ret = ftrace_match_records(hash, glob, strlen(glob));
4384

4385 4386 4387
	/* Nothing found? */
	if (!ret)
		ret = -EINVAL;
4388

4389 4390
	if (ret < 0)
		goto out;
4391

4392 4393 4394 4395
	size = 1 << hash->size_bits;
	for (i = 0; i < size; i++) {
		hlist_for_each_entry(entry, &hash->buckets[i], hlist) {
			if (ftrace_lookup_ip(old_hash, entry->ip))
4396
				continue;
4397 4398 4399 4400 4401
			/*
			 * 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.
			 */
4402 4403
			if (probe_ops->init) {
				ret = probe_ops->init(probe_ops, tr,
4404 4405 4406 4407 4408 4409 4410
						      entry->ip, data,
						      &probe->data);
				if (ret < 0) {
					if (probe_ops->free && count)
						probe_ops->free(probe_ops, tr,
								0, probe->data);
					probe->data = NULL;
4411
					goto out;
4412
				}
4413
			}
4414
			count++;
4415
		}
4416
	}
4417

4418
	mutex_lock(&ftrace_lock);
4419

4420 4421 4422 4423 4424
	if (!count) {
		/* Nothing was added? */
		ret = -EINVAL;
		goto out_unlock;
	}
4425

4426 4427
	ret = ftrace_hash_move_and_update_ops(&probe->ops, orig_hash,
					      hash, 1);
4428
	if (ret < 0)
4429
		goto err_unlock;
4430

4431 4432
	/* One ref for each new function traced */
	probe->ref += count;
4433

4434 4435
	if (!(probe->ops.flags & FTRACE_OPS_FL_ENABLED))
		ret = ftrace_startup(&probe->ops, 0);
4436

4437
 out_unlock:
4438
	mutex_unlock(&ftrace_lock);
4439

4440
	if (!ret)
4441
		ret = count;
4442
 out:
4443
	mutex_unlock(&probe->ops.func_hash->regex_lock);
4444
	free_ftrace_hash(hash);
4445

4446
	release_probe(probe);
4447

4448
	return ret;
4449

4450
 err_unlock:
4451
	if (!probe_ops->free || !count)
4452 4453 4454 4455 4456 4457 4458
		goto out_unlock;

	/* Failed to do the move, need to call the free functions */
	for (i = 0; i < size; i++) {
		hlist_for_each_entry(entry, &hash->buckets[i], hlist) {
			if (ftrace_lookup_ip(old_hash, entry->ip))
				continue;
4459
			probe_ops->free(probe_ops, tr, entry->ip, probe->data);
4460 4461 4462
		}
	}
	goto out_unlock;
4463 4464
}

4465
int
4466 4467
unregister_ftrace_function_probe_func(char *glob, struct trace_array *tr,
				      struct ftrace_probe_ops *probe_ops)
4468
{
4469
	struct ftrace_ops_hash old_hash_ops;
4470
	struct ftrace_func_entry *entry;
4471
	struct ftrace_func_probe *probe;
4472
	struct ftrace_glob func_g;
4473 4474 4475
	struct ftrace_hash **orig_hash;
	struct ftrace_hash *old_hash;
	struct ftrace_hash *hash = NULL;
4476
	struct hlist_node *tmp;
4477
	struct hlist_head hhd;
4478
	char str[KSYM_SYMBOL_LEN];
4479 4480
	int count = 0;
	int i, ret = -ENODEV;
4481
	int size;
4482

4483
	if (!glob || !strlen(glob) || !strcmp(glob, "*"))
4484
		func_g.search = NULL;
4485
	else {
4486 4487
		int not;

4488 4489 4490 4491
		func_g.type = filter_parse_regex(glob, strlen(glob),
						 &func_g.search, &not);
		func_g.len = strlen(func_g.search);
		func_g.search = glob;
4492

S
Steven Rostedt 已提交
4493
		/* we do not support '!' for function probes */
4494
		if (WARN_ON(not))
4495
			return -EINVAL;
4496 4497
	}

4498 4499 4500 4501 4502
	mutex_lock(&ftrace_lock);
	/* Check if the probe_ops is already registered */
	list_for_each_entry(probe, &tr->func_probes, list) {
		if (probe->probe_ops == probe_ops)
			break;
4503
	}
4504 4505 4506 4507 4508 4509 4510 4511 4512 4513
	if (&probe->list == &tr->func_probes)
		goto err_unlock_ftrace;

	ret = -EINVAL;
	if (!(probe->ops.flags & FTRACE_OPS_FL_INITIALIZED))
		goto err_unlock_ftrace;

	acquire_probe_locked(probe);

	mutex_unlock(&ftrace_lock);
4514

4515
	mutex_lock(&probe->ops.func_hash->regex_lock);
4516

4517
	orig_hash = &probe->ops.func_hash->filter_hash;
4518 4519 4520 4521
	old_hash = *orig_hash;

	if (ftrace_hash_empty(old_hash))
		goto out_unlock;
4522

4523 4524 4525 4526
	old_hash_ops.filter_hash = old_hash;
	/* Probes only have filters */
	old_hash_ops.notrace_hash = NULL;

4527
	ret = -ENOMEM;
4528
	hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, old_hash);
4529 4530 4531
	if (!hash)
		goto out_unlock;

4532
	INIT_HLIST_HEAD(&hhd);
4533

4534 4535 4536
	size = 1 << hash->size_bits;
	for (i = 0; i < size; i++) {
		hlist_for_each_entry_safe(entry, tmp, &hash->buckets[i], hlist) {
4537

4538
			if (func_g.search) {
4539 4540
				kallsyms_lookup(entry->ip, NULL, NULL,
						NULL, str);
4541
				if (!ftrace_match(str, &func_g))
4542 4543
					continue;
			}
4544
			count++;
4545 4546
			remove_hash_entry(hash, entry);
			hlist_add_head(&entry->hlist, &hhd);
4547 4548
		}
	}
4549 4550

	/* Nothing found? */
4551
	if (!count) {
4552 4553 4554 4555
		ret = -EINVAL;
		goto out_unlock;
	}

4556
	mutex_lock(&ftrace_lock);
4557

4558
	WARN_ON(probe->ref < count);
4559

4560
	probe->ref -= count;
4561

4562 4563 4564 4565
	if (ftrace_hash_empty(hash))
		ftrace_shutdown(&probe->ops, 0);

	ret = ftrace_hash_move_and_update_ops(&probe->ops, orig_hash,
4566
					      hash, 1);
4567 4568

	/* still need to update the function call sites */
4569
	if (ftrace_enabled && !ftrace_hash_empty(hash))
4570
		ftrace_run_modify_code(&probe->ops, FTRACE_UPDATE_CALLS,
4571
				       &old_hash_ops);
4572
	synchronize_sched();
4573

4574 4575
	hlist_for_each_entry_safe(entry, tmp, &hhd, hlist) {
		hlist_del(&entry->hlist);
4576
		if (probe_ops->free)
4577
			probe_ops->free(probe_ops, tr, entry->ip, probe->data);
4578
		kfree(entry);
4579
	}
4580
	mutex_unlock(&ftrace_lock);
4581

4582
 out_unlock:
4583
	mutex_unlock(&probe->ops.func_hash->regex_lock);
4584
	free_ftrace_hash(hash);
4585

4586
	release_probe(probe);
4587

4588
	return ret;
4589

4590 4591
 err_unlock_ftrace:
	mutex_unlock(&ftrace_lock);
4592
	return ret;
4593 4594
}

4595 4596 4597 4598 4599 4600 4601 4602
void clear_ftrace_function_probes(struct trace_array *tr)
{
	struct ftrace_func_probe *probe, *n;

	list_for_each_entry_safe(probe, n, &tr->func_probes, list)
		unregister_ftrace_function_probe_func(NULL, tr, probe->probe_ops);
}

4603 4604 4605
static LIST_HEAD(ftrace_commands);
static DEFINE_MUTEX(ftrace_cmd_mutex);

4606 4607 4608 4609 4610
/*
 * Currently we only register ftrace commands from __init, so mark this
 * __init too.
 */
__init int register_ftrace_command(struct ftrace_func_command *cmd)
4611 4612 4613 4614 4615 4616 4617 4618 4619 4620 4621 4622 4623 4624 4625 4626 4627 4628
{
	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;
}

4629 4630 4631 4632 4633
/*
 * Currently we only unregister ftrace commands from __init, so mark
 * this __init too.
 */
__init int unregister_ftrace_command(struct ftrace_func_command *cmd)
4634 4635 4636 4637 4638 4639 4640 4641 4642 4643 4644 4645 4646 4647 4648 4649 4650 4651
{
	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;
}

4652
static int ftrace_process_regex(struct ftrace_iterator *iter,
4653
				char *buff, int len, int enable)
4654
{
4655
	struct ftrace_hash *hash = iter->hash;
4656
	struct trace_array *tr = iter->ops->private;
4657
	char *func, *command, *next = buff;
S
Steven Rostedt 已提交
4658
	struct ftrace_func_command *p;
4659
	int ret = -EINVAL;
4660 4661 4662 4663

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

	if (!next) {
4664
		ret = ftrace_match_records(hash, func, len);
4665 4666 4667 4668 4669
		if (!ret)
			ret = -EINVAL;
		if (ret < 0)
			return ret;
		return 0;
4670 4671
	}

4672
	/* command found */
4673 4674 4675

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

4676 4677 4678
	mutex_lock(&ftrace_cmd_mutex);
	list_for_each_entry(p, &ftrace_commands, list) {
		if (strcmp(p->name, command) == 0) {
4679
			ret = p->func(tr, hash, func, command, next, enable);
4680 4681
			goto out_unlock;
		}
4682
	}
4683 4684
 out_unlock:
	mutex_unlock(&ftrace_cmd_mutex);
4685

4686
	return ret;
4687 4688
}

I
Ingo Molnar 已提交
4689
static ssize_t
4690 4691
ftrace_regex_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos, int enable)
4692 4693
{
	struct ftrace_iterator *iter;
4694 4695
	struct trace_parser *parser;
	ssize_t ret, read;
4696

4697
	if (!cnt)
4698 4699 4700 4701 4702 4703 4704 4705
		return 0;

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

4706
	if (unlikely(ftrace_disabled))
4707 4708 4709
		return -ENODEV;

	/* iter->hash is a local copy, so we don't need regex_lock */
4710

4711 4712
	parser = &iter->parser;
	read = trace_get_user(parser, ubuf, cnt, ppos);
4713

4714
	if (read >= 0 && trace_parser_loaded(parser) &&
4715
	    !trace_parser_cont(parser)) {
4716
		ret = ftrace_process_regex(iter, parser->buffer,
4717
					   parser->idx, enable);
4718
		trace_parser_clear(parser);
4719
		if (ret < 0)
4720
			goto out;
4721
	}
4722 4723

	ret = read;
4724
 out:
4725 4726 4727
	return ret;
}

4728
ssize_t
4729 4730 4731 4732 4733 4734
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);
}

4735
ssize_t
4736 4737 4738 4739 4740 4741
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);
}

4742
static int
4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758 4759 4760 4761 4762 4763
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);
}

static int
ftrace_set_hash(struct ftrace_ops *ops, unsigned char *buf, int len,
		unsigned long ip, int remove, int reset, int enable)
4764
{
4765
	struct ftrace_hash **orig_hash;
4766
	struct ftrace_hash *hash;
4767
	int ret;
4768

4769
	if (unlikely(ftrace_disabled))
4770
		return -ENODEV;
4771

4772
	mutex_lock(&ops->func_hash->regex_lock);
4773

4774
	if (enable)
4775
		orig_hash = &ops->func_hash->filter_hash;
4776
	else
4777
		orig_hash = &ops->func_hash->notrace_hash;
4778

4779 4780 4781 4782 4783
	if (reset)
		hash = alloc_ftrace_hash(FTRACE_HASH_DEFAULT_BITS);
	else
		hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, *orig_hash);

4784 4785 4786 4787
	if (!hash) {
		ret = -ENOMEM;
		goto out_regex_unlock;
	}
4788

4789 4790 4791 4792
	if (buf && !ftrace_match_records(hash, buf, len)) {
		ret = -EINVAL;
		goto out_regex_unlock;
	}
4793 4794 4795 4796 4797
	if (ip) {
		ret = ftrace_match_addr(hash, ip, remove);
		if (ret < 0)
			goto out_regex_unlock;
	}
4798 4799

	mutex_lock(&ftrace_lock);
4800
	ret = ftrace_hash_move_and_update_ops(ops, orig_hash, hash, enable);
4801 4802
	mutex_unlock(&ftrace_lock);

4803
 out_regex_unlock:
4804
	mutex_unlock(&ops->func_hash->regex_lock);
4805 4806 4807

	free_ftrace_hash(hash);
	return ret;
4808 4809
}

4810 4811 4812 4813 4814 4815 4816 4817 4818 4819 4820 4821 4822 4823 4824 4825 4826 4827 4828 4829
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)
{
4830
	ftrace_ops_init(ops);
4831 4832 4833 4834
	return ftrace_set_addr(ops, ip, remove, reset, 1);
}
EXPORT_SYMBOL_GPL(ftrace_set_filter_ip);

4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846 4847 4848 4849 4850 4851
/**
 * 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);

4852 4853 4854 4855 4856 4857 4858
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);
}

4859 4860
/**
 * ftrace_set_filter - set a function to filter on in ftrace
4861 4862 4863 4864 4865 4866 4867 4868
 * @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.
 */
4869
int ftrace_set_filter(struct ftrace_ops *ops, unsigned char *buf,
4870 4871
		       int len, int reset)
{
4872
	ftrace_ops_init(ops);
4873
	return ftrace_set_regex(ops, buf, len, reset, 1);
4874 4875 4876 4877 4878 4879 4880 4881 4882 4883 4884 4885 4886 4887
}
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.
 */
4888
int ftrace_set_notrace(struct ftrace_ops *ops, unsigned char *buf,
4889 4890
			int len, int reset)
{
4891
	ftrace_ops_init(ops);
4892
	return ftrace_set_regex(ops, buf, len, reset, 0);
4893 4894 4895
}
EXPORT_SYMBOL_GPL(ftrace_set_notrace);
/**
4896
 * ftrace_set_global_filter - set a function to filter on with global tracers
4897 4898 4899 4900 4901 4902 4903
 * @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.
 */
4904
void ftrace_set_global_filter(unsigned char *buf, int len, int reset)
4905
{
4906
	ftrace_set_regex(&global_ops, buf, len, reset, 1);
4907
}
4908
EXPORT_SYMBOL_GPL(ftrace_set_global_filter);
4909

4910
/**
4911
 * ftrace_set_global_notrace - set a function to not trace with global tracers
4912 4913 4914 4915 4916 4917 4918 4919
 * @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.
 */
4920
void ftrace_set_global_notrace(unsigned char *buf, int len, int reset)
4921
{
4922
	ftrace_set_regex(&global_ops, buf, len, reset, 0);
4923
}
4924
EXPORT_SYMBOL_GPL(ftrace_set_global_notrace);
4925

4926 4927 4928 4929 4930 4931 4932
/*
 * 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;

4933 4934 4935
/* Used by function selftest to not test if filter is set */
bool ftrace_filter_param __initdata;

4936 4937
static int __init set_ftrace_notrace(char *str)
{
4938
	ftrace_filter_param = true;
4939
	strlcpy(ftrace_notrace_buf, str, FTRACE_FILTER_SIZE);
4940 4941 4942 4943 4944 4945
	return 1;
}
__setup("ftrace_notrace=", set_ftrace_notrace);

static int __init set_ftrace_filter(char *str)
{
4946
	ftrace_filter_param = true;
4947
	strlcpy(ftrace_filter_buf, str, FTRACE_FILTER_SIZE);
4948 4949 4950 4951
	return 1;
}
__setup("ftrace_filter=", set_ftrace_filter);

4952
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
4953
static char ftrace_graph_buf[FTRACE_FILTER_SIZE] __initdata;
4954
static char ftrace_graph_notrace_buf[FTRACE_FILTER_SIZE] __initdata;
4955
static int ftrace_graph_set_hash(struct ftrace_hash *hash, char *buffer);
4956

4957 4958 4959
static unsigned long save_global_trampoline;
static unsigned long save_global_flags;

4960 4961
static int __init set_graph_function(char *str)
{
4962
	strlcpy(ftrace_graph_buf, str, FTRACE_FILTER_SIZE);
4963 4964 4965 4966
	return 1;
}
__setup("ftrace_graph_filter=", set_graph_function);

4967 4968 4969 4970 4971 4972 4973
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);

4974 4975 4976 4977 4978 4979 4980 4981
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);
4982 4983

static void __init set_ftrace_early_graph(char *buf, int enable)
4984 4985 4986
{
	int ret;
	char *func;
4987
	struct ftrace_hash *hash;
4988

4989 4990 4991
	hash = alloc_ftrace_hash(FTRACE_HASH_DEFAULT_BITS);
	if (WARN_ON(!hash))
		return;
4992 4993 4994 4995

	while (buf) {
		func = strsep(&buf, ",");
		/* we allow only one expression at a time */
4996
		ret = ftrace_graph_set_hash(hash, func);
4997 4998 4999 5000
		if (ret)
			printk(KERN_DEBUG "ftrace: function %s not "
					  "traceable\n", func);
	}
5001 5002 5003 5004 5005

	if (enable)
		ftrace_graph_hash = hash;
	else
		ftrace_graph_notrace_hash = hash;
5006 5007 5008
}
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

5009 5010
void __init
ftrace_set_early_filter(struct ftrace_ops *ops, char *buf, int enable)
5011 5012 5013
{
	char *func;

5014 5015
	ftrace_ops_init(ops);

5016 5017
	while (buf) {
		func = strsep(&buf, ",");
5018
		ftrace_set_regex(ops, func, strlen(func), 0, enable);
5019 5020 5021 5022 5023 5024
	}
}

static void __init set_ftrace_early_filters(void)
{
	if (ftrace_filter_buf[0])
5025
		ftrace_set_early_filter(&global_ops, ftrace_filter_buf, 1);
5026
	if (ftrace_notrace_buf[0])
5027
		ftrace_set_early_filter(&global_ops, ftrace_notrace_buf, 0);
5028 5029
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	if (ftrace_graph_buf[0])
5030 5031 5032
		set_ftrace_early_graph(ftrace_graph_buf, 1);
	if (ftrace_graph_notrace_buf[0])
		set_ftrace_early_graph(ftrace_graph_notrace_buf, 0);
5033
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */
5034 5035
}

5036
int ftrace_regex_release(struct inode *inode, struct file *file)
5037 5038 5039
{
	struct seq_file *m = (struct seq_file *)file->private_data;
	struct ftrace_iterator *iter;
5040
	struct ftrace_hash **orig_hash;
5041
	struct trace_parser *parser;
5042
	int filter_hash;
5043
	int ret;
5044 5045 5046 5047 5048 5049 5050

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

5051 5052 5053
	parser = &iter->parser;
	if (trace_parser_loaded(parser)) {
		parser->buffer[parser->idx] = 0;
5054
		ftrace_match_records(iter->hash, parser->buffer, parser->idx);
5055 5056
	}

5057 5058
	trace_parser_put(parser);

5059
	mutex_lock(&iter->ops->func_hash->regex_lock);
5060

5061
	if (file->f_mode & FMODE_WRITE) {
5062 5063
		filter_hash = !!(iter->flags & FTRACE_ITER_FILTER);

5064
		if (filter_hash) {
5065
			orig_hash = &iter->ops->func_hash->filter_hash;
5066
			if (iter->tr && !list_empty(&iter->tr->mod_trace))
5067 5068
				iter->hash->flags |= FTRACE_HASH_FL_MOD;
		} else
5069
			orig_hash = &iter->ops->func_hash->notrace_hash;
5070

5071
		mutex_lock(&ftrace_lock);
5072 5073
		ret = ftrace_hash_move_and_update_ops(iter->ops, orig_hash,
						      iter->hash, filter_hash);
5074
		mutex_unlock(&ftrace_lock);
5075 5076 5077
	} else {
		/* For read only, the hash is the ops hash */
		iter->hash = NULL;
5078
	}
5079

5080
	mutex_unlock(&iter->ops->func_hash->regex_lock);
5081 5082
	free_ftrace_hash(iter->hash);
	kfree(iter);
5083

5084 5085 5086
	return 0;
}

5087
static const struct file_operations ftrace_avail_fops = {
5088 5089 5090
	.open = ftrace_avail_open,
	.read = seq_read,
	.llseek = seq_lseek,
L
Li Zefan 已提交
5091
	.release = seq_release_private,
5092 5093
};

5094 5095 5096 5097 5098 5099 5100
static const struct file_operations ftrace_enabled_fops = {
	.open = ftrace_enabled_open,
	.read = seq_read,
	.llseek = seq_lseek,
	.release = seq_release_private,
};

5101
static const struct file_operations ftrace_filter_fops = {
5102
	.open = ftrace_filter_open,
L
Lai Jiangshan 已提交
5103
	.read = seq_read,
5104
	.write = ftrace_filter_write,
5105
	.llseek = tracing_lseek,
5106
	.release = ftrace_regex_release,
5107 5108
};

5109
static const struct file_operations ftrace_notrace_fops = {
5110
	.open = ftrace_notrace_open,
L
Lai Jiangshan 已提交
5111
	.read = seq_read,
5112
	.write = ftrace_notrace_write,
5113
	.llseek = tracing_lseek,
5114
	.release = ftrace_regex_release,
5115 5116
};

5117 5118 5119 5120
#ifdef CONFIG_FUNCTION_GRAPH_TRACER

static DEFINE_MUTEX(graph_lock);

5121 5122 5123 5124 5125 5126 5127
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,
};
5128

5129 5130
#define FTRACE_GRAPH_EMPTY	((void *)1)

5131
struct ftrace_graph_data {
5132 5133 5134 5135 5136 5137 5138
	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;
5139 5140
};

5141
static void *
5142
__g_next(struct seq_file *m, loff_t *pos)
5143
{
5144
	struct ftrace_graph_data *fgd = m->private;
5145 5146 5147
	struct ftrace_func_entry *entry = fgd->entry;
	struct hlist_head *head;
	int i, idx = fgd->idx;
5148

5149
	if (*pos >= fgd->hash->count)
5150
		return NULL;
5151 5152 5153 5154 5155 5156 5157 5158 5159 5160 5161 5162 5163 5164 5165 5166 5167 5168 5169

	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;
5170
}
5171

5172 5173 5174 5175 5176
static void *
g_next(struct seq_file *m, void *v, loff_t *pos)
{
	(*pos)++;
	return __g_next(m, pos);
5177 5178 5179 5180
}

static void *g_start(struct seq_file *m, loff_t *pos)
{
5181 5182
	struct ftrace_graph_data *fgd = m->private;

5183 5184
	mutex_lock(&graph_lock);

5185 5186 5187 5188 5189 5190 5191
	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));

5192
	/* Nothing, tell g_show to print all functions are enabled */
5193
	if (ftrace_hash_empty(fgd->hash) && !*pos)
5194
		return FTRACE_GRAPH_EMPTY;
5195

5196 5197
	fgd->idx = 0;
	fgd->entry = NULL;
5198
	return __g_next(m, pos);
5199 5200 5201 5202 5203 5204 5205 5206 5207
}

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

static int g_show(struct seq_file *m, void *v)
{
5208
	struct ftrace_func_entry *entry = v;
5209

5210
	if (!entry)
5211 5212
		return 0;

5213
	if (entry == FTRACE_GRAPH_EMPTY) {
5214 5215
		struct ftrace_graph_data *fgd = m->private;

5216
		if (fgd->type == GRAPH_FILTER_FUNCTION)
5217
			seq_puts(m, "#### all functions enabled ####\n");
5218
		else
5219
			seq_puts(m, "#### no functions disabled ####\n");
5220 5221 5222
		return 0;
	}

5223
	seq_printf(m, "%ps\n", (void *)entry->ip);
5224 5225 5226 5227

	return 0;
}

J
James Morris 已提交
5228
static const struct seq_operations ftrace_graph_seq_ops = {
5229 5230 5231 5232 5233 5234 5235
	.start = g_start,
	.next = g_next,
	.stop = g_stop,
	.show = g_show,
};

static int
5236 5237
__ftrace_graph_open(struct inode *inode, struct file *file,
		    struct ftrace_graph_data *fgd)
5238 5239
{
	int ret = 0;
5240
	struct ftrace_hash *new_hash = NULL;
5241

5242 5243 5244
	if (file->f_mode & FMODE_WRITE) {
		const int size_bits = FTRACE_HASH_DEFAULT_BITS;

5245 5246 5247
		if (trace_parser_get_init(&fgd->parser, FTRACE_BUFF_MAX))
			return -ENOMEM;

5248 5249 5250 5251 5252 5253 5254 5255 5256
		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;
		}
5257 5258
	}

5259
	if (file->f_mode & FMODE_READ) {
5260
		ret = seq_open(file, &ftrace_graph_seq_ops);
5261 5262 5263
		if (!ret) {
			struct seq_file *m = file->private_data;
			m->private = fgd;
5264 5265 5266 5267
		} else {
			/* Failed */
			free_ftrace_hash(new_hash);
			new_hash = NULL;
5268 5269 5270
		}
	} else
		file->private_data = fgd;
5271

5272
out:
5273 5274 5275
	if (ret < 0 && file->f_mode & FMODE_WRITE)
		trace_parser_put(&fgd->parser);

5276
	fgd->new_hash = new_hash;
5277 5278 5279 5280 5281 5282 5283 5284

	/*
	 * 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;

5285 5286 5287
	return ret;
}

5288 5289 5290 5291
static int
ftrace_graph_open(struct inode *inode, struct file *file)
{
	struct ftrace_graph_data *fgd;
5292
	int ret;
5293 5294 5295 5296 5297 5298 5299 5300

	if (unlikely(ftrace_disabled))
		return -ENODEV;

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

5301 5302
	mutex_lock(&graph_lock);

5303 5304
	fgd->hash = rcu_dereference_protected(ftrace_graph_hash,
					lockdep_is_held(&graph_lock));
5305
	fgd->type = GRAPH_FILTER_FUNCTION;
5306 5307
	fgd->seq_ops = &ftrace_graph_seq_ops;

5308 5309 5310 5311 5312 5313
	ret = __ftrace_graph_open(inode, file, fgd);
	if (ret < 0)
		kfree(fgd);

	mutex_unlock(&graph_lock);
	return ret;
5314 5315
}

5316 5317 5318 5319
static int
ftrace_graph_notrace_open(struct inode *inode, struct file *file)
{
	struct ftrace_graph_data *fgd;
5320
	int ret;
5321 5322 5323 5324 5325 5326 5327 5328

	if (unlikely(ftrace_disabled))
		return -ENODEV;

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

5329 5330
	mutex_lock(&graph_lock);

5331 5332
	fgd->hash = rcu_dereference_protected(ftrace_graph_notrace_hash,
					lockdep_is_held(&graph_lock));
5333
	fgd->type = GRAPH_FILTER_NOTRACE;
5334 5335
	fgd->seq_ops = &ftrace_graph_seq_ops;

5336 5337 5338 5339 5340 5341
	ret = __ftrace_graph_open(inode, file, fgd);
	if (ret < 0)
		kfree(fgd);

	mutex_unlock(&graph_lock);
	return ret;
5342 5343
}

5344 5345 5346
static int
ftrace_graph_release(struct inode *inode, struct file *file)
{
5347
	struct ftrace_graph_data *fgd;
5348 5349 5350
	struct ftrace_hash *old_hash, *new_hash;
	struct trace_parser *parser;
	int ret = 0;
5351

5352 5353 5354
	if (file->f_mode & FMODE_READ) {
		struct seq_file *m = file->private_data;

5355
		fgd = m->private;
5356
		seq_release(inode, file);
5357
	} else {
5358
		fgd = file->private_data;
5359 5360
	}

5361 5362 5363 5364 5365 5366 5367 5368 5369 5370 5371 5372 5373 5374 5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 5390 5391 5392 5393 5394 5395 5396 5397 5398 5399 5400

	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:
5401
	free_ftrace_hash(fgd->new_hash);
5402 5403
	kfree(fgd);

5404
	return ret;
5405 5406
}

5407
static int
5408
ftrace_graph_set_hash(struct ftrace_hash *hash, char *buffer)
5409
{
5410
	struct ftrace_glob func_g;
5411 5412
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
5413
	struct ftrace_func_entry *entry;
5414
	int fail = 1;
5415
	int not;
5416

5417
	/* decode regex */
5418 5419
	func_g.type = filter_parse_regex(buffer, strlen(buffer),
					 &func_g.search, &not);
5420

5421
	func_g.len = strlen(func_g.search);
5422

5423
	mutex_lock(&ftrace_lock);
5424 5425 5426 5427 5428 5429

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

5430 5431
	do_for_each_ftrace_rec(pg, rec) {

5432 5433 5434
		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

D
Dmitry Safonov 已提交
5435
		if (ftrace_match_record(rec, &func_g, NULL, 0)) {
5436
			entry = ftrace_lookup_ip(hash, rec->ip);
5437 5438 5439

			if (!not) {
				fail = 0;
5440 5441 5442 5443 5444

				if (entry)
					continue;
				if (add_hash_entry(hash, rec->ip) < 0)
					goto out;
5445
			} else {
5446 5447
				if (entry) {
					free_hash_entry(hash, entry);
5448 5449 5450
					fail = 0;
				}
			}
5451
		}
5452
	} while_for_each_ftrace_rec();
5453
out:
5454
	mutex_unlock(&ftrace_lock);
5455

5456 5457 5458 5459
	if (fail)
		return -EINVAL;

	return 0;
5460 5461 5462 5463 5464 5465
}

static ssize_t
ftrace_graph_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
5466
	ssize_t read, ret = 0;
5467
	struct ftrace_graph_data *fgd = file->private_data;
5468
	struct trace_parser *parser;
5469

5470
	if (!cnt)
5471 5472
		return 0;

5473 5474 5475 5476 5477 5478
	/* Read mode uses seq functions */
	if (file->f_mode & FMODE_READ) {
		struct seq_file *m = file->private_data;
		fgd = m->private;
	}

5479
	parser = &fgd->parser;
5480

5481
	read = trace_get_user(parser, ubuf, cnt, ppos);
5482

5483 5484
	if (read >= 0 && trace_parser_loaded(parser) &&
	    !trace_parser_cont(parser)) {
5485

5486
		ret = ftrace_graph_set_hash(fgd->new_hash,
5487 5488
					    parser->buffer);
		trace_parser_clear(parser);
5489 5490
	}

5491 5492
	if (!ret)
		ret = read;
5493

5494 5495 5496 5497
	return ret;
}

static const struct file_operations ftrace_graph_fops = {
5498 5499 5500
	.open		= ftrace_graph_open,
	.read		= seq_read,
	.write		= ftrace_graph_write,
5501
	.llseek		= tracing_lseek,
5502
	.release	= ftrace_graph_release,
5503
};
5504 5505 5506 5507 5508

static const struct file_operations ftrace_graph_notrace_fops = {
	.open		= ftrace_graph_notrace_open,
	.read		= seq_read,
	.write		= ftrace_graph_write,
5509
	.llseek		= tracing_lseek,
5510 5511
	.release	= ftrace_graph_release,
};
5512 5513
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

5514 5515 5516 5517 5518 5519 5520 5521 5522 5523 5524 5525 5526 5527 5528 5529 5530 5531 5532 5533 5534 5535 5536 5537 5538 5539 5540 5541 5542 5543
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);
}

5544
static __init int ftrace_init_dyn_tracefs(struct dentry *d_tracer)
5545 5546
{

5547 5548
	trace_create_file("available_filter_functions", 0444,
			d_tracer, NULL, &ftrace_avail_fops);
5549

5550 5551 5552
	trace_create_file("enabled_functions", 0444,
			d_tracer, NULL, &ftrace_enabled_fops);

5553
	ftrace_create_filter_files(&global_ops, d_tracer);
5554

5555
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
5556
	trace_create_file("set_graph_function", 0444, d_tracer,
5557 5558
				    NULL,
				    &ftrace_graph_fops);
5559 5560 5561
	trace_create_file("set_graph_notrace", 0444, d_tracer,
				    NULL,
				    &ftrace_graph_notrace_fops);
5562 5563
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

5564 5565 5566
	return 0;
}

5567
static int ftrace_cmp_ips(const void *a, const void *b)
5568
{
5569 5570
	const unsigned long *ipa = a;
	const unsigned long *ipb = b;
5571

5572 5573 5574 5575 5576 5577 5578
	if (*ipa > *ipb)
		return 1;
	if (*ipa < *ipb)
		return -1;
	return 0;
}

5579
static int ftrace_process_locs(struct module *mod,
5580
			       unsigned long *start,
5581 5582
			       unsigned long *end)
{
5583
	struct ftrace_page *start_pg;
5584
	struct ftrace_page *pg;
5585
	struct dyn_ftrace *rec;
5586
	unsigned long count;
5587 5588
	unsigned long *p;
	unsigned long addr;
5589
	unsigned long flags = 0; /* Shut up gcc */
5590 5591 5592 5593 5594 5595 5596
	int ret = -ENOMEM;

	count = end - start;

	if (!count)
		return 0;

5597
	sort(start, count, sizeof(*start),
5598
	     ftrace_cmp_ips, NULL);
5599

5600 5601
	start_pg = ftrace_allocate_pages(count);
	if (!start_pg)
5602
		return -ENOMEM;
5603

S
Steven Rostedt 已提交
5604
	mutex_lock(&ftrace_lock);
5605

5606 5607 5608 5609 5610
	/*
	 * 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.
	 */
5611 5612 5613
	if (!mod) {
		WARN_ON(ftrace_pages || ftrace_pages_start);
		/* First initialization */
5614
		ftrace_pages = ftrace_pages_start = start_pg;
5615
	} else {
5616
		if (!ftrace_pages)
5617
			goto out;
5618

5619 5620 5621 5622
		if (WARN_ON(ftrace_pages->next)) {
			/* Hmm, we have free pages? */
			while (ftrace_pages->next)
				ftrace_pages = ftrace_pages->next;
5623
		}
5624

5625
		ftrace_pages->next = start_pg;
5626 5627
	}

5628
	p = start;
5629
	pg = start_pg;
5630 5631
	while (p < end) {
		addr = ftrace_call_adjust(*p++);
5632 5633 5634 5635 5636 5637 5638 5639
		/*
		 * 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;
5640 5641 5642 5643 5644 5645 5646 5647 5648 5649

		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;
5650 5651
	}

5652 5653 5654 5655 5656 5657
	/* We should have used all pages */
	WARN_ON(pg->next);

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

5658
	/*
5659 5660 5661 5662 5663 5664
	 * 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.
5665
	 */
5666 5667
	if (!mod)
		local_irq_save(flags);
5668
	ftrace_update_code(mod, start_pg);
5669 5670
	if (!mod)
		local_irq_restore(flags);
5671 5672
	ret = 0;
 out:
S
Steven Rostedt 已提交
5673
	mutex_unlock(&ftrace_lock);
5674

5675
	return ret;
5676 5677
}

5678
#ifdef CONFIG_MODULES
5679 5680 5681

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

5682 5683 5684 5685 5686 5687 5688 5689 5690 5691 5692 5693 5694
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;
}

5695 5696 5697 5698 5699 5700 5701 5702 5703 5704 5705 5706 5707 5708 5709 5710 5711 5712 5713 5714 5715 5716 5717 5718 5719 5720 5721 5722 5723 5724 5725 5726 5727 5728 5729 5730 5731 5732 5733 5734
static void
clear_mod_from_hash(struct ftrace_page *pg, struct ftrace_hash *hash)
{
	struct ftrace_func_entry *entry;
	struct dyn_ftrace *rec;
	int i;

	if (ftrace_hash_empty(hash))
		return;

	for (i = 0; i < pg->index; i++) {
		rec = &pg->records[i];
		entry = __ftrace_lookup_ip(hash, rec->ip);
		/*
		 * Do not allow this rec to match again.
		 * Yeah, it may waste some memory, but will be removed
		 * if/when the hash is modified again.
		 */
		if (entry)
			entry->ip = 0;
	}
}

/* Clear any records from hashs */
static void clear_mod_from_hashes(struct ftrace_page *pg)
{
	struct trace_array *tr;

	mutex_lock(&trace_types_lock);
	list_for_each_entry(tr, &ftrace_trace_arrays, list) {
		if (!tr->ops || !tr->ops->func_hash)
			continue;
		mutex_lock(&tr->ops->func_hash->regex_lock);
		clear_mod_from_hash(pg, tr->ops->func_hash->filter_hash);
		clear_mod_from_hash(pg, tr->ops->func_hash->notrace_hash);
		mutex_unlock(&tr->ops->func_hash->regex_lock);
	}
	mutex_unlock(&trace_types_lock);
}

5735
void ftrace_release_mod(struct module *mod)
5736 5737
{
	struct dyn_ftrace *rec;
5738
	struct ftrace_page **last_pg;
5739
	struct ftrace_page *tmp_page = NULL;
5740
	struct ftrace_page *pg;
5741
	int order;
5742

5743 5744
	mutex_lock(&ftrace_lock);

5745
	if (ftrace_disabled)
5746
		goto out_unlock;
5747

5748 5749 5750 5751 5752 5753 5754
	/*
	 * 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];
5755 5756
		if (within_module_core(rec->ip, mod) ||
		    within_module_init(rec->ip, mod)) {
5757
			/*
5758 5759
			 * As core pages are first, the first
			 * page should never be a module page.
5760
			 */
5761 5762 5763 5764 5765 5766 5767
			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);

5768
			ftrace_update_tot_cnt -= pg->index;
5769
			*last_pg = pg->next;
5770 5771 5772

			pg->next = tmp_page;
			tmp_page = pg;
5773 5774 5775
		} else
			last_pg = &pg->next;
	}
5776
 out_unlock:
5777
	mutex_unlock(&ftrace_lock);
5778 5779 5780 5781 5782 5783 5784 5785 5786 5787 5788

	for (pg = tmp_page; pg; pg = tmp_page) {

		/* Needs to be called outside of ftrace_lock */
		clear_mod_from_hashes(pg);

		order = get_count_order(pg->size / ENTRIES_PER_PAGE);
		free_pages((unsigned long)pg->records, order);
		tmp_page = pg->next;
		kfree(pg);
	}
5789 5790
}

5791
void ftrace_module_enable(struct module *mod)
5792 5793 5794 5795 5796 5797 5798 5799 5800 5801 5802 5803 5804 5805 5806 5807 5808 5809 5810 5811 5812 5813 5814 5815 5816 5817 5818 5819 5820 5821 5822 5823 5824
{
	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.
		 */
5825 5826
		if (!within_module_core(rec->ip, mod) &&
		    !within_module_init(rec->ip, mod))
5827 5828 5829 5830 5831 5832 5833 5834 5835 5836 5837 5838 5839 5840 5841 5842 5843 5844 5845 5846 5847 5848 5849 5850 5851 5852 5853 5854 5855 5856 5857 5858
			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);
5859 5860

	process_cached_mods(mod->name);
5861 5862
}

5863
void ftrace_module_init(struct module *mod)
5864
{
5865
	if (ftrace_disabled || !mod->num_ftrace_callsites)
5866
		return;
5867

5868 5869
	ftrace_process_locs(mod, mod->ftrace_callsites,
			    mod->ftrace_callsites + mod->num_ftrace_callsites);
5870
}
5871 5872
#endif /* CONFIG_MODULES */

5873
void ftrace_free_mem(void *start_ptr, void *end_ptr)
5874
{
5875 5876
	unsigned long start = (unsigned long)(start_ptr);
	unsigned long end = (unsigned long)(end_ptr);
5877 5878 5879 5880 5881 5882 5883 5884 5885 5886 5887 5888 5889 5890 5891 5892 5893 5894 5895 5896 5897 5898
	struct ftrace_page **last_pg = &ftrace_pages_start;
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
	struct dyn_ftrace key;
	int order;

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

	mutex_lock(&ftrace_lock);

	for (pg = ftrace_pages_start; pg; last_pg = &pg->next, pg = *last_pg) {
		if (end < pg->records[0].ip ||
		    start >= (pg->records[pg->index - 1].ip + MCOUNT_INSN_SIZE))
			continue;
 again:
		rec = bsearch(&key, pg->records, pg->index,
			      sizeof(struct dyn_ftrace),
			      ftrace_cmp_recs);
		if (!rec)
			continue;
		pg->index--;
5899
		ftrace_update_tot_cnt--;
5900 5901 5902 5903 5904 5905 5906 5907 5908 5909 5910 5911 5912 5913 5914 5915 5916 5917
		if (!pg->index) {
			*last_pg = pg->next;
			order = get_count_order(pg->size / ENTRIES_PER_PAGE);
			free_pages((unsigned long)pg->records, order);
			kfree(pg);
			pg = container_of(last_pg, struct ftrace_page, next);
			if (!(*last_pg))
				ftrace_pages = pg;
			continue;
		}
		memmove(rec, rec + 1,
			(pg->index - (rec - pg->records)) * sizeof(*rec));
		/* More than one function may be in this block */
		goto again;
	}
	mutex_unlock(&ftrace_lock);
}

5918 5919 5920 5921 5922 5923 5924 5925
void __init ftrace_free_init_mem(void)
{
	void *start = (void *)(&__init_begin);
	void *end = (void *)(&__init_end);

	ftrace_free_mem(start, end);
}

5926 5927
void __init ftrace_init(void)
{
5928 5929
	extern unsigned long __start_mcount_loc[];
	extern unsigned long __stop_mcount_loc[];
5930
	unsigned long count, flags;
5931 5932 5933
	int ret;

	local_irq_save(flags);
5934
	ret = ftrace_dyn_arch_init();
5935
	local_irq_restore(flags);
5936
	if (ret)
5937 5938 5939
		goto failed;

	count = __stop_mcount_loc - __start_mcount_loc;
5940 5941
	if (!count) {
		pr_info("ftrace: No functions to be traced?\n");
5942
		goto failed;
5943 5944 5945 5946
	}

	pr_info("ftrace: allocating %ld entries in %ld pages\n",
		count, count / ENTRIES_PER_PAGE + 1);
5947 5948 5949

	last_ftrace_enabled = ftrace_enabled = 1;

5950
	ret = ftrace_process_locs(NULL,
5951
				  __start_mcount_loc,
5952 5953
				  __stop_mcount_loc);

5954 5955
	set_ftrace_early_filters();

5956 5957 5958 5959 5960
	return;
 failed:
	ftrace_disabled = 1;
}

5961 5962 5963 5964 5965 5966 5967 5968 5969 5970
/* 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)
{
	arch_ftrace_update_trampoline(ops);
}

5971 5972 5973
void ftrace_init_trace_array(struct trace_array *tr)
{
	INIT_LIST_HEAD(&tr->func_probes);
5974 5975
	INIT_LIST_HEAD(&tr->mod_trace);
	INIT_LIST_HEAD(&tr->mod_notrace);
5976
}
5977
#else
5978

5979
static struct ftrace_ops global_ops = {
5980
	.func			= ftrace_stub,
5981 5982 5983
	.flags			= FTRACE_OPS_FL_RECURSION_SAFE |
				  FTRACE_OPS_FL_INITIALIZED |
				  FTRACE_OPS_FL_PID,
5984 5985
};

5986 5987 5988 5989 5990
static int __init ftrace_nodyn_init(void)
{
	ftrace_enabled = 1;
	return 0;
}
5991
core_initcall(ftrace_nodyn_init);
5992

5993
static inline int ftrace_init_dyn_tracefs(struct dentry *d_tracer) { return 0; }
5994
static inline void ftrace_startup_enable(int command) { }
5995
static inline void ftrace_startup_all(int command) { }
5996
/* Keep as macros so we do not need to define the commands */
5997 5998 5999 6000 6001 6002
# define ftrace_startup(ops, command)					\
	({								\
		int ___ret = __register_ftrace_function(ops);		\
		if (!___ret)						\
			(ops)->flags |= FTRACE_OPS_FL_ENABLED;		\
		___ret;							\
6003
	})
6004 6005 6006 6007 6008 6009 6010
# define ftrace_shutdown(ops, command)					\
	({								\
		int ___ret = __unregister_ftrace_function(ops);		\
		if (!___ret)						\
			(ops)->flags &= ~FTRACE_OPS_FL_ENABLED;		\
		___ret;							\
	})
6011

I
Ingo Molnar 已提交
6012 6013
# define ftrace_startup_sysctl()	do { } while (0)
# define ftrace_shutdown_sysctl()	do { } while (0)
6014 6015

static inline int
6016
ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip, void *regs)
6017 6018 6019 6020
{
	return 1;
}

6021 6022 6023 6024
static void ftrace_update_trampoline(struct ftrace_ops *ops)
{
}

6025 6026
#endif /* CONFIG_DYNAMIC_FTRACE */

6027 6028 6029 6030
__init void ftrace_init_global_array_ops(struct trace_array *tr)
{
	tr->ops = &global_ops;
	tr->ops->private = tr;
6031
	ftrace_init_trace_array(tr);
6032 6033 6034 6035 6036 6037 6038 6039 6040 6041 6042 6043 6044 6045 6046 6047 6048 6049 6050
}

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;
}

6051 6052
static inline void
__ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
6053
		       struct ftrace_ops *ignored, struct pt_regs *regs)
6054
{
6055
	struct ftrace_ops *op;
6056
	int bit;
6057

6058 6059 6060
	bit = trace_test_and_set_recursion(TRACE_LIST_START, TRACE_LIST_MAX);
	if (bit < 0)
		return;
6061

6062 6063 6064 6065 6066
	/*
	 * Some of the ops may be dynamically allocated,
	 * they must be freed after a synchronize_sched().
	 */
	preempt_disable_notrace();
6067

6068
	do_for_each_ftrace_op(op, ftrace_ops_list) {
6069 6070 6071 6072 6073 6074 6075 6076 6077 6078 6079 6080 6081 6082
		/*
		 * 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)) {
		    
6083 6084
			if (FTRACE_WARN_ON(!op->func)) {
				pr_warn("op=%p %pS\n", op, op);
6085 6086
				goto out;
			}
6087
			op->func(ip, parent_ip, op, regs);
6088
		}
6089
	} while_for_each_ftrace_op(op);
6090
out:
6091
	preempt_enable_notrace();
6092
	trace_clear_recursion(bit);
6093 6094
}

6095 6096 6097 6098 6099
/*
 * 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.
6100 6101 6102
 * 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
6103 6104
 * being NULL, or CONFIG_DYNAMIC_FTRACE_WITH_REGS.
 * Note, CONFIG_DYNAMIC_FTRACE_WITH_REGS expects a full regs to be saved.
6105
 * An architecture can pass partial regs with ftrace_ops and still
L
Li Bin 已提交
6106
 * set the ARCH_SUPPORTS_FTRACE_OPS.
6107 6108 6109
 */
#if ARCH_SUPPORTS_FTRACE_OPS
static void ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
6110
				 struct ftrace_ops *op, struct pt_regs *regs)
6111
{
6112
	__ftrace_ops_list_func(ip, parent_ip, NULL, regs);
6113 6114 6115 6116
}
#else
static void ftrace_ops_no_ops(unsigned long ip, unsigned long parent_ip)
{
6117
	__ftrace_ops_list_func(ip, parent_ip, NULL, NULL);
6118 6119 6120
}
#endif

6121 6122
/*
 * If there's only one function registered but it does not support
6123 6124
 * recursion, needs RCU protection and/or requires per cpu handling, then
 * this function will be called by the mcount trampoline.
6125
 */
6126
static void ftrace_ops_assist_func(unsigned long ip, unsigned long parent_ip,
6127 6128 6129 6130
				   struct ftrace_ops *op, struct pt_regs *regs)
{
	int bit;

6131 6132 6133
	if ((op->flags & FTRACE_OPS_FL_RCU) && !rcu_is_watching())
		return;

6134 6135 6136 6137
	bit = trace_test_and_set_recursion(TRACE_LIST_START, TRACE_LIST_MAX);
	if (bit < 0)
		return;

6138
	preempt_disable_notrace();
6139

6140 6141 6142 6143 6144 6145
	if (!(op->flags & FTRACE_OPS_FL_PER_CPU) ||
	    !ftrace_function_local_disabled(op)) {
		op->func(ip, parent_ip, op, regs);
	}

	preempt_enable_notrace();
6146 6147 6148
	trace_clear_recursion(bit);
}

6149 6150 6151 6152 6153 6154 6155
/**
 * 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
6156
 * ftrace_ops_assist_func() instead.
6157 6158 6159 6160 6161 6162
 *
 * Returns the function that the trampoline should call for @ops.
 */
ftrace_func_t ftrace_ops_get_func(struct ftrace_ops *ops)
{
	/*
6163 6164
	 * 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.
6165
	 */
6166 6167 6168
	if (!(ops->flags & FTRACE_OPS_FL_RECURSION_SAFE) ||
	    ops->flags & (FTRACE_OPS_FL_RCU | FTRACE_OPS_FL_PER_CPU))
		return ftrace_ops_assist_func;
6169 6170 6171 6172

	return ops->func;
}

6173 6174 6175
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
6176
{
6177 6178
	struct trace_array *tr = data;
	struct trace_pid_list *pid_list;
S
Steven Rostedt 已提交
6179

6180
	pid_list = rcu_dereference_sched(tr->function_pids);
6181

6182 6183
	this_cpu_write(tr->trace_buffer.data->ftrace_ignore_pid,
		       trace_ignore_this_task(pid_list, next));
S
Steven Rostedt 已提交
6184 6185
}

6186 6187 6188 6189 6190 6191 6192 6193 6194 6195 6196 6197 6198 6199 6200 6201 6202 6203 6204 6205 6206 6207 6208 6209 6210 6211 6212 6213 6214 6215 6216 6217 6218 6219 6220 6221 6222
static void
ftrace_pid_follow_sched_process_fork(void *data,
				     struct task_struct *self,
				     struct task_struct *task)
{
	struct trace_pid_list *pid_list;
	struct trace_array *tr = data;

	pid_list = rcu_dereference_sched(tr->function_pids);
	trace_filter_add_remove_task(pid_list, self, task);
}

static void
ftrace_pid_follow_sched_process_exit(void *data, struct task_struct *task)
{
	struct trace_pid_list *pid_list;
	struct trace_array *tr = data;

	pid_list = rcu_dereference_sched(tr->function_pids);
	trace_filter_add_remove_task(pid_list, NULL, task);
}

void ftrace_pid_follow_fork(struct trace_array *tr, bool enable)
{
	if (enable) {
		register_trace_sched_process_fork(ftrace_pid_follow_sched_process_fork,
						  tr);
		register_trace_sched_process_exit(ftrace_pid_follow_sched_process_exit,
						  tr);
	} else {
		unregister_trace_sched_process_fork(ftrace_pid_follow_sched_process_fork,
						    tr);
		unregister_trace_sched_process_exit(ftrace_pid_follow_sched_process_exit,
						    tr);
	}
}

6223
static void clear_ftrace_pids(struct trace_array *tr)
6224
{
6225 6226
	struct trace_pid_list *pid_list;
	int cpu;
6227

6228 6229 6230 6231
	pid_list = rcu_dereference_protected(tr->function_pids,
					     lockdep_is_held(&ftrace_lock));
	if (!pid_list)
		return;
6232

6233
	unregister_trace_sched_switch(ftrace_filter_pid_sched_switch_probe, tr);
6234

6235 6236
	for_each_possible_cpu(cpu)
		per_cpu_ptr(tr->trace_buffer.data, cpu)->ftrace_ignore_pid = false;
S
Steven Rostedt 已提交
6237

6238
	rcu_assign_pointer(tr->function_pids, NULL);
S
Steven Rostedt 已提交
6239

6240 6241
	/* Wait till all users are no longer using pid filtering */
	synchronize_sched();
6242

6243
	trace_free_pid_list(pid_list);
6244 6245
}

6246 6247 6248 6249 6250 6251 6252 6253 6254
void ftrace_clear_pids(struct trace_array *tr)
{
	mutex_lock(&ftrace_lock);

	clear_ftrace_pids(tr);

	mutex_unlock(&ftrace_lock);
}

6255
static void ftrace_pid_reset(struct trace_array *tr)
6256
{
6257
	mutex_lock(&ftrace_lock);
6258
	clear_ftrace_pids(tr);
S
Steven Rostedt 已提交
6259

6260
	ftrace_update_pid_func();
6261
	ftrace_startup_all(0);
6262 6263 6264 6265

	mutex_unlock(&ftrace_lock);
}

6266 6267
/* Greater than any max PID */
#define FTRACE_NO_PIDS		(void *)(PID_MAX_LIMIT + 1)
6268

6269
static void *fpid_start(struct seq_file *m, loff_t *pos)
6270
	__acquires(RCU)
6271
{
6272 6273 6274
	struct trace_pid_list *pid_list;
	struct trace_array *tr = m->private;

6275
	mutex_lock(&ftrace_lock);
6276 6277 6278
	rcu_read_lock_sched();

	pid_list = rcu_dereference_sched(tr->function_pids);
6279

6280 6281
	if (!pid_list)
		return !(*pos) ? FTRACE_NO_PIDS : NULL;
6282

6283
	return trace_pid_start(pid_list, pos);
6284 6285 6286 6287
}

static void *fpid_next(struct seq_file *m, void *v, loff_t *pos)
{
6288 6289 6290 6291
	struct trace_array *tr = m->private;
	struct trace_pid_list *pid_list = rcu_dereference_sched(tr->function_pids);

	if (v == FTRACE_NO_PIDS)
6292 6293
		return NULL;

6294
	return trace_pid_next(pid_list, v, pos);
6295 6296 6297
}

static void fpid_stop(struct seq_file *m, void *p)
6298
	__releases(RCU)
6299
{
6300
	rcu_read_unlock_sched();
6301 6302 6303 6304 6305
	mutex_unlock(&ftrace_lock);
}

static int fpid_show(struct seq_file *m, void *v)
{
6306
	if (v == FTRACE_NO_PIDS) {
6307
		seq_puts(m, "no pid\n");
6308 6309 6310
		return 0;
	}

6311
	return trace_pid_show(m, v);
6312 6313 6314 6315 6316 6317 6318 6319 6320 6321 6322 6323
}

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)
{
6324 6325
	struct trace_array *tr = inode->i_private;
	struct seq_file *m;
6326 6327
	int ret = 0;

6328 6329 6330
	if (trace_array_get(tr) < 0)
		return -ENODEV;

6331 6332
	if ((file->f_mode & FMODE_WRITE) &&
	    (file->f_flags & O_TRUNC))
6333
		ftrace_pid_reset(tr);
6334

6335 6336 6337 6338 6339 6340 6341 6342
	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;
	}
6343 6344 6345 6346

	return ret;
}

6347 6348 6349 6350 6351 6352 6353 6354 6355 6356 6357 6358 6359 6360 6361 6362
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));
}

6363 6364 6365 6366
static ssize_t
ftrace_pid_write(struct file *filp, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
6367 6368 6369 6370 6371
	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;
6372

6373 6374 6375 6376 6377 6378 6379 6380 6381 6382 6383
	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;
6384

6385
	rcu_assign_pointer(tr->function_pids, pid_list);
6386

6387 6388 6389 6390 6391 6392 6393
	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);
	}
6394

6395
	/*
6396 6397 6398
	 * 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.
6399
	 */
6400
	on_each_cpu(ignore_task_cpu, tr, 1);
6401

6402 6403 6404 6405
	ftrace_update_pid_func();
	ftrace_startup_all(0);
 out:
	mutex_unlock(&ftrace_lock);
6406

6407 6408
	if (ret > 0)
		*ppos += ret;
6409

6410
	return ret;
6411
}
6412

6413 6414 6415
static int
ftrace_pid_release(struct inode *inode, struct file *file)
{
6416
	struct trace_array *tr = inode->i_private;
6417

6418 6419 6420
	trace_array_put(tr);

	return seq_release(inode, file);
6421 6422
}

6423
static const struct file_operations ftrace_pid_fops = {
6424 6425 6426
	.open		= ftrace_pid_open,
	.write		= ftrace_pid_write,
	.read		= seq_read,
6427
	.llseek		= tracing_lseek,
6428
	.release	= ftrace_pid_release,
6429 6430
};

6431
void ftrace_init_tracefs(struct trace_array *tr, struct dentry *d_tracer)
6432
{
6433
	trace_create_file("set_ftrace_pid", 0644, d_tracer,
6434
			    tr, &ftrace_pid_fops);
6435 6436
}

6437 6438 6439 6440 6441 6442 6443 6444 6445 6446
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
6447
/**
6448
 * ftrace_kill - kill ftrace
S
Steven Rostedt 已提交
6449 6450 6451 6452 6453
 *
 * 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.
 */
6454
void ftrace_kill(void)
S
Steven Rostedt 已提交
6455 6456 6457 6458 6459 6460
{
	ftrace_disabled = 1;
	ftrace_enabled = 0;
	clear_ftrace_function();
}

6461 6462 6463 6464 6465 6466 6467 6468
/**
 * Test if ftrace is dead or not.
 */
int ftrace_is_dead(void)
{
	return ftrace_disabled;
}

6469
/**
6470 6471
 * register_ftrace_function - register a function for profiling
 * @ops - ops structure that holds the function for profiling.
6472
 *
6473 6474 6475 6476 6477 6478
 * 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.
6479
 */
6480
int register_ftrace_function(struct ftrace_ops *ops)
6481
{
6482
	int ret = -1;
6483

6484 6485
	ftrace_ops_init(ops);

S
Steven Rostedt 已提交
6486
	mutex_lock(&ftrace_lock);
6487

6488
	ret = ftrace_startup(ops, 0);
6489

S
Steven Rostedt 已提交
6490
	mutex_unlock(&ftrace_lock);
6491

6492
	return ret;
6493
}
6494
EXPORT_SYMBOL_GPL(register_ftrace_function);
6495 6496

/**
6497
 * unregister_ftrace_function - unregister a function for profiling.
6498 6499 6500 6501 6502 6503 6504 6505
 * @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
6506
	mutex_lock(&ftrace_lock);
6507
	ret = ftrace_shutdown(ops, 0);
S
Steven Rostedt 已提交
6508
	mutex_unlock(&ftrace_lock);
6509 6510 6511

	return ret;
}
6512
EXPORT_SYMBOL_GPL(unregister_ftrace_function);
6513

I
Ingo Molnar 已提交
6514
int
6515
ftrace_enable_sysctl(struct ctl_table *table, int write,
6516
		     void __user *buffer, size_t *lenp,
6517 6518
		     loff_t *ppos)
{
6519
	int ret = -ENODEV;
6520

S
Steven Rostedt 已提交
6521
	mutex_lock(&ftrace_lock);
6522

6523 6524 6525 6526
	if (unlikely(ftrace_disabled))
		goto out;

	ret = proc_dointvec(table, write, buffer, lenp, ppos);
6527

6528
	if (ret || !write || (last_ftrace_enabled == !!ftrace_enabled))
6529 6530
		goto out;

6531
	last_ftrace_enabled = !!ftrace_enabled;
6532 6533 6534 6535

	if (ftrace_enabled) {

		/* we are starting ftrace again */
6536 6537
		if (rcu_dereference_protected(ftrace_ops_list,
			lockdep_is_held(&ftrace_lock)) != &ftrace_list_end)
6538
			update_ftrace_function();
6539

6540 6541
		ftrace_startup_sysctl();

6542 6543 6544 6545 6546 6547 6548 6549
	} else {
		/* stopping ftrace calls (just send to ftrace_stub) */
		ftrace_trace_function = ftrace_stub;

		ftrace_shutdown_sysctl();
	}

 out:
S
Steven Rostedt 已提交
6550
	mutex_unlock(&ftrace_lock);
6551
	return ret;
6552
}
6553

6554
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
6555

6556 6557 6558 6559
static struct ftrace_ops graph_ops = {
	.func			= ftrace_stub,
	.flags			= FTRACE_OPS_FL_RECURSION_SAFE |
				   FTRACE_OPS_FL_INITIALIZED |
6560
				   FTRACE_OPS_FL_PID |
6561 6562 6563
				   FTRACE_OPS_FL_STUB,
#ifdef FTRACE_GRAPH_TRAMP_ADDR
	.trampoline		= FTRACE_GRAPH_TRAMP_ADDR,
6564
	/* trampoline_size is only needed for dynamically allocated tramps */
6565 6566 6567 6568
#endif
	ASSIGN_OPS_HASH(graph_ops, &global_ops.local_hash)
};

6569 6570 6571 6572 6573 6574 6575 6576 6577 6578
void ftrace_graph_sleep_time_control(bool enable)
{
	fgraph_sleep_time = enable;
}

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

6579 6580 6581 6582 6583
int ftrace_graph_entry_stub(struct ftrace_graph_ent *trace)
{
	return 0;
}

6584 6585 6586
/* The callbacks that hook a function */
trace_func_graph_ret_t ftrace_graph_return =
			(trace_func_graph_ret_t)ftrace_stub;
6587
trace_func_graph_ent_t ftrace_graph_entry = ftrace_graph_entry_stub;
6588
static trace_func_graph_ent_t __ftrace_graph_entry = ftrace_graph_entry_stub;
6589 6590 6591 6592 6593 6594 6595 6596 6597 6598 6599 6600 6601 6602 6603 6604 6605 6606 6607 6608 6609

/* 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;
		}
	}

6610
	read_lock(&tasklist_lock);
6611 6612 6613 6614 6615 6616 6617
	do_each_thread(g, t) {
		if (start == end) {
			ret = -EAGAIN;
			goto unlock;
		}

		if (t->ret_stack == NULL) {
6618
			atomic_set(&t->tracing_graph_pause, 0);
6619
			atomic_set(&t->trace_overrun, 0);
6620 6621 6622 6623
			t->curr_ret_stack = -1;
			/* Make sure the tasks see the -1 first: */
			smp_wmb();
			t->ret_stack = ret_stack_list[start++];
6624 6625 6626 6627
		}
	} while_each_thread(g, t);

unlock:
6628
	read_unlock(&tasklist_lock);
6629 6630 6631 6632 6633 6634
free:
	for (i = start; i < end; i++)
		kfree(ret_stack_list[i]);
	return ret;
}

6635
static void
6636
ftrace_graph_probe_sched_switch(void *ignore, bool preempt,
6637
			struct task_struct *prev, struct task_struct *next)
6638 6639 6640 6641
{
	unsigned long long timestamp;
	int index;

6642 6643 6644 6645
	/*
	 * Does the user want to count the time a function was asleep.
	 * If so, do not update the time stamps.
	 */
6646
	if (fgraph_sleep_time)
6647 6648
		return;

6649 6650 6651 6652 6653 6654 6655 6656 6657 6658 6659 6660 6661 6662 6663 6664 6665 6666
	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;
}

6667
/* Allocate a return stack for each task */
6668
static int start_graph_tracing(void)
6669 6670
{
	struct ftrace_ret_stack **ret_stack_list;
6671
	int ret, cpu;
6672 6673 6674 6675 6676 6677 6678 6679

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

	if (!ret_stack_list)
		return -ENOMEM;

6680
	/* The cpu_boot init_task->ret_stack will never be freed */
6681 6682
	for_each_online_cpu(cpu) {
		if (!idle_task(cpu)->ret_stack)
6683
			ftrace_graph_init_idle_task(idle_task(cpu), cpu);
6684
	}
6685

6686 6687 6688 6689
	do {
		ret = alloc_retstack_tasklist(ret_stack_list);
	} while (ret == -EAGAIN);

6690
	if (!ret) {
6691
		ret = register_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
6692 6693 6694 6695 6696
		if (ret)
			pr_info("ftrace_graph: Couldn't activate tracepoint"
				" probe to kernel_sched_switch\n");
	}

6697 6698 6699 6700
	kfree(ret_stack_list);
	return ret;
}

6701 6702 6703 6704 6705 6706 6707 6708 6709 6710 6711 6712 6713 6714 6715 6716 6717 6718 6719 6720 6721
/*
 * 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;
}

6722 6723 6724 6725 6726 6727 6728 6729 6730 6731 6732 6733 6734 6735 6736 6737
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)
{
6738 6739 6740 6741 6742 6743 6744 6745 6746 6747 6748 6749 6750 6751 6752 6753 6754 6755 6756
	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)
6757
		ftrace_graph_entry = ftrace_graph_entry_test;
6758 6759
	else
		ftrace_graph_entry = __ftrace_graph_entry;
6760 6761
}

6762 6763 6764 6765
static struct notifier_block ftrace_suspend_notifier = {
	.notifier_call = ftrace_suspend_notifier_call,
};

6766 6767
int register_ftrace_graph(trace_func_graph_ret_t retfunc,
			trace_func_graph_ent_t entryfunc)
6768
{
6769 6770
	int ret = 0;

S
Steven Rostedt 已提交
6771
	mutex_lock(&ftrace_lock);
6772

6773
	/* we currently allow only one tracer registered at a time */
6774
	if (ftrace_graph_active) {
6775 6776 6777 6778
		ret = -EBUSY;
		goto out;
	}

6779 6780
	register_pm_notifier(&ftrace_suspend_notifier);

6781
	ftrace_graph_active++;
6782
	ret = start_graph_tracing();
6783
	if (ret) {
6784
		ftrace_graph_active--;
6785 6786
		goto out;
	}
6787

6788
	ftrace_graph_return = retfunc;
6789 6790 6791 6792 6793 6794 6795 6796 6797 6798

	/*
	 * 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();
6799

6800
	ret = ftrace_startup(&graph_ops, FTRACE_START_FUNC_RET);
6801
out:
S
Steven Rostedt 已提交
6802
	mutex_unlock(&ftrace_lock);
6803
	return ret;
6804 6805
}

6806
void unregister_ftrace_graph(void)
6807
{
S
Steven Rostedt 已提交
6808
	mutex_lock(&ftrace_lock);
6809

6810
	if (unlikely(!ftrace_graph_active))
6811 6812
		goto out;

6813
	ftrace_graph_active--;
6814
	ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub;
6815
	ftrace_graph_entry = ftrace_graph_entry_stub;
6816
	__ftrace_graph_entry = ftrace_graph_entry_stub;
6817
	ftrace_shutdown(&graph_ops, FTRACE_STOP_FUNC_RET);
6818
	unregister_pm_notifier(&ftrace_suspend_notifier);
6819
	unregister_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
6820

6821 6822 6823 6824 6825 6826 6827 6828 6829 6830 6831
#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

6832
 out:
S
Steven Rostedt 已提交
6833
	mutex_unlock(&ftrace_lock);
6834
}
6835

6836 6837 6838 6839 6840 6841 6842 6843
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
6844
	/* make curr_ret_stack visible before we add the ret_stack */
6845 6846 6847 6848 6849 6850 6851 6852 6853 6854 6855 6856 6857 6858 6859 6860 6861 6862 6863 6864 6865 6866 6867 6868 6869 6870 6871 6872 6873 6874 6875 6876 6877 6878
	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);
	}
}

6879
/* Allocate a return stack for newly created task */
6880
void ftrace_graph_init_task(struct task_struct *t)
6881
{
6882 6883
	/* Make sure we do not use the parent ret_stack */
	t->ret_stack = NULL;
6884
	t->curr_ret_stack = -1;
6885

6886
	if (ftrace_graph_active) {
6887 6888 6889
		struct ftrace_ret_stack *ret_stack;

		ret_stack = kmalloc(FTRACE_RETFUNC_DEPTH
6890 6891
				* sizeof(struct ftrace_ret_stack),
				GFP_KERNEL);
6892
		if (!ret_stack)
6893
			return;
6894
		graph_init_task(t, ret_stack);
6895
	}
6896 6897
}

6898
void ftrace_graph_exit_task(struct task_struct *t)
6899
{
6900 6901
	struct ftrace_ret_stack	*ret_stack = t->ret_stack;

6902
	t->ret_stack = NULL;
6903 6904 6905 6906
	/* NULL must become visible to IRQs before we free it: */
	barrier();

	kfree(ret_stack);
6907
}
6908
#endif