ftrace.c 150.9 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 *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
}

160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180
/**
 * ftrace_nr_registered_ops - return number of ops registered
 *
 * Returns the number of ftrace_ops registered and tracing functions
 */
int ftrace_nr_registered_ops(void)
{
	struct ftrace_ops *ops;
	int cnt = 0;

	mutex_lock(&ftrace_lock);

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

	mutex_unlock(&ftrace_lock);

	return cnt;
}

181
static void ftrace_pid_func(unsigned long ip, unsigned long parent_ip,
182
			    struct ftrace_ops *op, struct pt_regs *regs)
183
{
184 185 186
	struct trace_array *tr = op->private;

	if (tr && this_cpu_read(tr->trace_buffer.data->ftrace_ignore_pid))
187 188
		return;

189
	op->saved_func(ip, parent_ip, op, regs);
190 191
}

192
/**
193
 * clear_ftrace_function - reset the ftrace function
194
 *
195 196
 * This NULLs the ftrace function and in essence stops
 * tracing.  There may be lag
197
 */
198
void clear_ftrace_function(void)
199
{
200 201 202
	ftrace_trace_function = ftrace_stub;
}

203
static void per_cpu_ops_disable_all(struct ftrace_ops *ops)
204 205 206 207 208 209 210
{
	int cpu;

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

211
static int per_cpu_ops_alloc(struct ftrace_ops *ops)
212 213 214
{
	int __percpu *disabled;

215 216 217
	if (WARN_ON_ONCE(!(ops->flags & FTRACE_OPS_FL_PER_CPU)))
		return -EINVAL;

218 219 220 221 222
	disabled = alloc_percpu(int);
	if (!disabled)
		return -ENOMEM;

	ops->disabled = disabled;
223
	per_cpu_ops_disable_all(ops);
224 225 226
	return 0;
}

227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243
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();
}

244 245
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static void update_function_graph_func(void);
246 247 248 249 250

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

251 252 253 254
#else
static inline void update_function_graph_func(void) { }
#endif

255 256 257 258

static ftrace_func_t ftrace_ops_get_list_func(struct ftrace_ops *ops)
{
	/*
259
	 * If this is a dynamic, RCU, or per CPU ops, or we force list func,
260 261
	 * then it needs to call the list anyway.
	 */
262 263
	if (ops->flags & (FTRACE_OPS_FL_DYNAMIC | FTRACE_OPS_FL_PER_CPU |
			  FTRACE_OPS_FL_RCU) || FTRACE_FORCE_LIST_FUNC)
264 265 266 267 268
		return ftrace_ops_list_func;

	return ftrace_ops_get_func(ops);
}

269 270 271 272
static void update_ftrace_function(void)
{
	ftrace_func_t func;

273 274 275 276 277 278 279 280 281 282 283
	/*
	 * Prepare the ftrace_ops that the arch callback will use.
	 * If there's only one ftrace_ops registered, the ftrace_ops_list
	 * will point to the ops we want.
	 */
	set_function_trace_op = ftrace_ops_list;

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

284 285
	/*
	 * If we are at the end of the list and this ops is
286 287
	 * recursion safe and not dynamic and the arch supports passing ops,
	 * then have the mcount trampoline call the function directly.
288
	 */
289
	} else if (ftrace_ops_list->next == &ftrace_list_end) {
290
		func = ftrace_ops_get_list_func(ftrace_ops_list);
291

292 293
	} else {
		/* Just use the default ftrace_ops */
294
		set_function_trace_op = &ftrace_list_end;
295
		func = ftrace_ops_list_func;
296
	}
297

298 299
	update_function_graph_func();

300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342
	/* 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 */

343 344 345
	ftrace_trace_function = func;
}

346 347 348 349 350
int using_ftrace_ops_list_func(void)
{
	return ftrace_trace_function == ftrace_ops_list_func;
}

351
static void add_ftrace_ops(struct ftrace_ops **list, struct ftrace_ops *ops)
352
{
353
	ops->next = *list;
354
	/*
355
	 * We are entering ops into the list but another
356 357
	 * CPU might be walking that list. We need to make sure
	 * the ops->next pointer is valid before another CPU sees
358
	 * the ops pointer included into the list.
359
	 */
360
	rcu_assign_pointer(*list, ops);
361 362
}

363
static int remove_ftrace_ops(struct ftrace_ops **list, struct ftrace_ops *ops)
364 365 366 367
{
	struct ftrace_ops **p;

	/*
368 369
	 * If we are removing the last function, then simply point
	 * to the ftrace_stub.
370
	 */
371 372
	if (*list == ops && ops->next == &ftrace_list_end) {
		*list = &ftrace_list_end;
S
Steven Rostedt 已提交
373
		return 0;
374 375
	}

376
	for (p = list; *p != &ftrace_list_end; p = &(*p)->next)
377 378 379
		if (*p == ops)
			break;

S
Steven Rostedt 已提交
380 381
	if (*p != ops)
		return -1;
382 383

	*p = (*p)->next;
384 385
	return 0;
}
386

387 388
static void ftrace_update_trampoline(struct ftrace_ops *ops);

389 390
static int __register_ftrace_function(struct ftrace_ops *ops)
{
391 392 393
	if (ops->flags & FTRACE_OPS_FL_DELETED)
		return -EINVAL;

394 395 396
	if (WARN_ON(ops->flags & FTRACE_OPS_FL_ENABLED))
		return -EBUSY;

397
#ifndef CONFIG_DYNAMIC_FTRACE_WITH_REGS
398 399 400 401 402 403 404 405 406 407 408 409 410
	/*
	 * 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

411 412 413
	if (!core_kernel_data((unsigned long)ops))
		ops->flags |= FTRACE_OPS_FL_DYNAMIC;

414 415
	if (ops->flags & FTRACE_OPS_FL_PER_CPU) {
		if (per_cpu_ops_alloc(ops))
416
			return -ENOMEM;
417 418 419
	}

	add_ftrace_ops(&ftrace_ops_list, ops);
420

421 422 423
	/* Always save the function, and reset at unregistering */
	ops->saved_func = ops->func;

424
	if (ftrace_pids_enabled(ops))
425 426
		ops->func = ftrace_pid_func;

427 428
	ftrace_update_trampoline(ops);

429 430 431 432 433 434 435 436 437 438
	if (ftrace_enabled)
		update_ftrace_function();

	return 0;
}

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

439 440 441
	if (WARN_ON(!(ops->flags & FTRACE_OPS_FL_ENABLED)))
		return -EBUSY;

442
	ret = remove_ftrace_ops(&ftrace_ops_list, ops);
443

444 445
	if (ret < 0)
		return ret;
446

447 448
	if (ftrace_enabled)
		update_ftrace_function();
449

450 451
	ops->func = ops->saved_func;

S
Steven Rostedt 已提交
452
	return 0;
453 454
}

455 456
static void ftrace_update_pid_func(void)
{
457 458
	struct ftrace_ops *op;

459
	/* Only do something if we are tracing something */
460
	if (ftrace_trace_function == ftrace_stub)
461
		return;
462

463 464
	do_for_each_ftrace_op(op, ftrace_ops_list) {
		if (op->flags & FTRACE_OPS_FL_PID) {
465 466
			op->func = ftrace_pids_enabled(op) ?
				ftrace_pid_func : op->saved_func;
467 468 469 470
			ftrace_update_trampoline(op);
		}
	} while_for_each_ftrace_op(op);

471
	update_ftrace_function();
472 473
}

474 475 476 477 478
#ifdef CONFIG_FUNCTION_PROFILER
struct ftrace_profile {
	struct hlist_node		node;
	unsigned long			ip;
	unsigned long			counter;
479 480
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	unsigned long long		time;
481
	unsigned long long		time_squared;
482
#endif
483 484
};

485 486 487 488
struct ftrace_profile_page {
	struct ftrace_profile_page	*next;
	unsigned long			index;
	struct ftrace_profile		records[];
489 490
};

491 492 493 494 495 496 497 498
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;
};

499 500
#define PROFILE_RECORDS_SIZE						\
	(PAGE_SIZE - offsetof(struct ftrace_profile_page, records))
501

502 503
#define PROFILES_PER_PAGE					\
	(PROFILE_RECORDS_SIZE / sizeof(struct ftrace_profile))
504

505 506 507
static int ftrace_profile_enabled __read_mostly;

/* ftrace_profile_lock - synchronize the enable and disable of the profiler */
S
Steven Rostedt 已提交
508 509
static DEFINE_MUTEX(ftrace_profile_lock);

510
static DEFINE_PER_CPU(struct ftrace_profile_stat, ftrace_profile_stats);
511

512 513
#define FTRACE_PROFILE_HASH_BITS 10
#define FTRACE_PROFILE_HASH_SIZE (1 << FTRACE_PROFILE_HASH_BITS)
514

S
Steven Rostedt 已提交
515 516 517
static void *
function_stat_next(void *v, int idx)
{
518 519
	struct ftrace_profile *rec = v;
	struct ftrace_profile_page *pg;
S
Steven Rostedt 已提交
520

521
	pg = (struct ftrace_profile_page *)((unsigned long)rec & PAGE_MASK);
S
Steven Rostedt 已提交
522 523

 again:
L
Li Zefan 已提交
524 525 526
	if (idx != 0)
		rec++;

S
Steven Rostedt 已提交
527 528 529 530 531
	if ((void *)rec >= (void *)&pg->records[pg->index]) {
		pg = pg->next;
		if (!pg)
			return NULL;
		rec = &pg->records[0];
532 533
		if (!rec->counter)
			goto again;
S
Steven Rostedt 已提交
534 535 536 537 538 539 540
	}

	return rec;
}

static void *function_stat_start(struct tracer_stat *trace)
{
541 542 543 544 545 546 547
	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
548 549
}

550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565
#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
566 567
static int function_stat_cmp(void *p1, void *p2)
{
568 569
	struct ftrace_profile *a = p1;
	struct ftrace_profile *b = p2;
S
Steven Rostedt 已提交
570 571 572 573 574 575 576 577

	if (a->counter < b->counter)
		return -1;
	if (a->counter > b->counter)
		return 1;
	else
		return 0;
}
578
#endif
S
Steven Rostedt 已提交
579 580 581

static int function_stat_headers(struct seq_file *m)
{
582
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
583 584 585 586
	seq_puts(m, "  Function                               "
		 "Hit    Time            Avg             s^2\n"
		    "  --------                               "
		 "---    ----            ---             ---\n");
587
#else
588 589
	seq_puts(m, "  Function                               Hit\n"
		    "  --------                               ---\n");
590
#endif
S
Steven Rostedt 已提交
591 592 593 594 595
	return 0;
}

static int function_stat_show(struct seq_file *m, void *v)
{
596
	struct ftrace_profile *rec = v;
S
Steven Rostedt 已提交
597
	char str[KSYM_SYMBOL_LEN];
598
	int ret = 0;
599
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
600 601
	static struct trace_seq s;
	unsigned long long avg;
602
	unsigned long long stddev;
603
#endif
604 605 606 607 608 609 610
	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
611

612 613 614 615 616 617 618
#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
619
	kallsyms_lookup(rec->ip, NULL, NULL, NULL, str);
620 621 622
	seq_printf(m, "  %-30.30s  %10lu", str, rec->counter);

#ifdef CONFIG_FUNCTION_GRAPH_TRACER
623
	seq_puts(m, "    ");
624

625 626 627 628
	/* Sample standard deviation (s^2) */
	if (rec->counter <= 1)
		stddev = 0;
	else {
629 630 631 632 633 634 635
		/*
		 * 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;

636 637 638 639
		/*
		 * Divide only 1000 for ns^2 -> us^2 conversion.
		 * trace_print_graph_duration will divide 1000 again.
		 */
640
		do_div(stddev, rec->counter * (rec->counter - 1) * 1000);
641 642
	}

643 644 645 646
	trace_seq_init(&s);
	trace_print_graph_duration(rec->time, &s);
	trace_seq_puts(&s, "    ");
	trace_print_graph_duration(avg, &s);
647 648
	trace_seq_puts(&s, "    ");
	trace_print_graph_duration(stddev, &s);
649 650 651
	trace_print_seq(m, &s);
#endif
	seq_putc(m, '\n');
652 653
out:
	mutex_unlock(&ftrace_profile_lock);
S
Steven Rostedt 已提交
654

655
	return ret;
S
Steven Rostedt 已提交
656 657
}

658
static void ftrace_profile_reset(struct ftrace_profile_stat *stat)
S
Steven Rostedt 已提交
659
{
660
	struct ftrace_profile_page *pg;
S
Steven Rostedt 已提交
661

662
	pg = stat->pages = stat->start;
S
Steven Rostedt 已提交
663

664 665 666 667
	while (pg) {
		memset(pg->records, 0, PROFILE_RECORDS_SIZE);
		pg->index = 0;
		pg = pg->next;
S
Steven Rostedt 已提交
668 669
	}

670
	memset(stat->hash, 0,
671 672
	       FTRACE_PROFILE_HASH_SIZE * sizeof(struct hlist_head));
}
S
Steven Rostedt 已提交
673

674
int ftrace_profile_pages_init(struct ftrace_profile_stat *stat)
675 676
{
	struct ftrace_profile_page *pg;
677 678
	int functions;
	int pages;
679
	int i;
S
Steven Rostedt 已提交
680

681
	/* If we already allocated, do nothing */
682
	if (stat->pages)
683
		return 0;
S
Steven Rostedt 已提交
684

685 686
	stat->pages = (void *)get_zeroed_page(GFP_KERNEL);
	if (!stat->pages)
687
		return -ENOMEM;
S
Steven Rostedt 已提交
688

689 690 691 692 693 694 695 696 697 698 699 700 701
#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

702
	pg = stat->start = stat->pages;
S
Steven Rostedt 已提交
703

704 705
	pages = DIV_ROUND_UP(functions, PROFILES_PER_PAGE);

706
	for (i = 1; i < pages; i++) {
707 708
		pg->next = (void *)get_zeroed_page(GFP_KERNEL);
		if (!pg->next)
709
			goto out_free;
710 711 712 713
		pg = pg->next;
	}

	return 0;
714 715 716 717 718 719 720 721 722 723 724 725 726 727

 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
728 729
}

730
static int ftrace_profile_init_cpu(int cpu)
S
Steven Rostedt 已提交
731
{
732
	struct ftrace_profile_stat *stat;
733
	int size;
S
Steven Rostedt 已提交
734

735 736 737
	stat = &per_cpu(ftrace_profile_stats, cpu);

	if (stat->hash) {
738
		/* If the profile is already created, simply reset it */
739
		ftrace_profile_reset(stat);
740 741
		return 0;
	}
S
Steven Rostedt 已提交
742

743 744 745 746 747
	/*
	 * 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
748

749
	stat->hash = kzalloc(sizeof(struct hlist_head) * size, GFP_KERNEL);
750

751
	if (!stat->hash)
752 753
		return -ENOMEM;

754
	/* Preallocate the function profiling pages */
755 756 757
	if (ftrace_profile_pages_init(stat) < 0) {
		kfree(stat->hash);
		stat->hash = NULL;
758 759 760 761
		return -ENOMEM;
	}

	return 0;
S
Steven Rostedt 已提交
762 763
}

764 765 766 767 768
static int ftrace_profile_init(void)
{
	int cpu;
	int ret = 0;

769
	for_each_possible_cpu(cpu) {
770 771 772 773 774 775 776 777
		ret = ftrace_profile_init_cpu(cpu);
		if (ret)
			break;
	}

	return ret;
}

778
/* interrupts must be disabled */
779 780
static struct ftrace_profile *
ftrace_find_profiled_func(struct ftrace_profile_stat *stat, unsigned long ip)
S
Steven Rostedt 已提交
781
{
782
	struct ftrace_profile *rec;
S
Steven Rostedt 已提交
783 784 785
	struct hlist_head *hhd;
	unsigned long key;

786
	key = hash_long(ip, FTRACE_PROFILE_HASH_BITS);
787
	hhd = &stat->hash[key];
S
Steven Rostedt 已提交
788 789 790 791

	if (hlist_empty(hhd))
		return NULL;

792
	hlist_for_each_entry_rcu_notrace(rec, hhd, node) {
S
Steven Rostedt 已提交
793
		if (rec->ip == ip)
794 795 796 797 798 799
			return rec;
	}

	return NULL;
}

800 801
static void ftrace_add_profile(struct ftrace_profile_stat *stat,
			       struct ftrace_profile *rec)
802 803 804
{
	unsigned long key;

805
	key = hash_long(rec->ip, FTRACE_PROFILE_HASH_BITS);
806
	hlist_add_head_rcu(&rec->node, &stat->hash[key]);
807 808
}

809 810 811
/*
 * The memory is already allocated, this simply finds a new record to use.
 */
812
static struct ftrace_profile *
813
ftrace_profile_alloc(struct ftrace_profile_stat *stat, unsigned long ip)
814 815 816
{
	struct ftrace_profile *rec = NULL;

817
	/* prevent recursion (from NMIs) */
818
	if (atomic_inc_return(&stat->disabled) != 1)
819 820 821
		goto out;

	/*
822 823
	 * Try to find the function again since an NMI
	 * could have added it
824
	 */
825
	rec = ftrace_find_profiled_func(stat, ip);
826
	if (rec)
827
		goto out;
828

829 830 831 832
	if (stat->pages->index == PROFILES_PER_PAGE) {
		if (!stat->pages->next)
			goto out;
		stat->pages = stat->pages->next;
S
Steven Rostedt 已提交
833
	}
834

835
	rec = &stat->pages->records[stat->pages->index++];
836
	rec->ip = ip;
837
	ftrace_add_profile(stat, rec);
838

S
Steven Rostedt 已提交
839
 out:
840
	atomic_dec(&stat->disabled);
S
Steven Rostedt 已提交
841 842 843 844 845

	return rec;
}

static void
846
function_profile_call(unsigned long ip, unsigned long parent_ip,
847
		      struct ftrace_ops *ops, struct pt_regs *regs)
S
Steven Rostedt 已提交
848
{
849
	struct ftrace_profile_stat *stat;
850
	struct ftrace_profile *rec;
S
Steven Rostedt 已提交
851 852 853 854 855 856
	unsigned long flags;

	if (!ftrace_profile_enabled)
		return;

	local_irq_save(flags);
857

858
	stat = this_cpu_ptr(&ftrace_profile_stats);
859
	if (!stat->hash || !ftrace_profile_enabled)
860 861 862
		goto out;

	rec = ftrace_find_profiled_func(stat, ip);
863
	if (!rec) {
864
		rec = ftrace_profile_alloc(stat, ip);
865 866 867
		if (!rec)
			goto out;
	}
S
Steven Rostedt 已提交
868 869 870 871 872 873

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

874 875 876
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static int profile_graph_entry(struct ftrace_graph_ent *trace)
{
877 878
	int index = trace->depth;

879
	function_profile_call(trace->func, 0, NULL, NULL);
880 881 882 883

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

884 885 886 887 888
	return 1;
}

static void profile_graph_return(struct ftrace_graph_ret *trace)
{
889
	struct ftrace_profile_stat *stat;
890
	unsigned long long calltime;
891
	struct ftrace_profile *rec;
892
	unsigned long flags;
893 894

	local_irq_save(flags);
895
	stat = this_cpu_ptr(&ftrace_profile_stats);
896
	if (!stat->hash || !ftrace_profile_enabled)
897 898
		goto out;

899 900 901 902
	/* If the calltime was zero'd ignore it */
	if (!trace->calltime)
		goto out;

903 904
	calltime = trace->rettime - trace->calltime;

905
	if (!fgraph_graph_time) {
906 907 908 909 910 911 912 913 914 915 916 917 918 919
		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;
	}

920
	rec = ftrace_find_profiled_func(stat, trace->func);
921
	if (rec) {
922
		rec->time += calltime;
923 924
		rec->time_squared += calltime * calltime;
	}
925

926
 out:
927 928 929 930 931 932 933 934 935 936 937 938 939 940
	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
941
static struct ftrace_ops ftrace_profile_ops __read_mostly = {
942
	.func		= function_profile_call,
943
	.flags		= FTRACE_OPS_FL_RECURSION_SAFE | FTRACE_OPS_FL_INITIALIZED,
944
	INIT_OPS_HASH(ftrace_profile_ops)
S
Steven Rostedt 已提交
945 946
};

947 948 949 950 951 952 953 954 955 956 957
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
958 959 960 961 962 963 964
static ssize_t
ftrace_profile_write(struct file *filp, const char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
	unsigned long val;
	int ret;

965 966
	ret = kstrtoul_from_user(ubuf, cnt, 10, &val);
	if (ret)
S
Steven Rostedt 已提交
967 968 969 970 971 972 973
		return ret;

	val = !!val;

	mutex_lock(&ftrace_profile_lock);
	if (ftrace_profile_enabled ^ val) {
		if (val) {
974 975 976 977 978 979
			ret = ftrace_profile_init();
			if (ret < 0) {
				cnt = ret;
				goto out;
			}

980 981 982 983 984
			ret = register_ftrace_profiler();
			if (ret < 0) {
				cnt = ret;
				goto out;
			}
S
Steven Rostedt 已提交
985 986 987
			ftrace_profile_enabled = 1;
		} else {
			ftrace_profile_enabled = 0;
988 989 990 991
			/*
			 * unregister_ftrace_profiler calls stop_machine
			 * so this acts like an synchronize_sched.
			 */
992
			unregister_ftrace_profiler();
S
Steven Rostedt 已提交
993 994
		}
	}
995
 out:
S
Steven Rostedt 已提交
996 997
	mutex_unlock(&ftrace_profile_lock);

998
	*ppos += cnt;
S
Steven Rostedt 已提交
999 1000 1001 1002

	return cnt;
}

1003 1004 1005 1006
static ssize_t
ftrace_profile_read(struct file *filp, char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
1007
	char buf[64];		/* big enough to hold a number */
1008 1009 1010 1011 1012 1013
	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
1014 1015 1016 1017
static const struct file_operations ftrace_profile_fops = {
	.open		= tracing_open_generic,
	.read		= ftrace_profile_read,
	.write		= ftrace_profile_write,
1018
	.llseek		= default_llseek,
S
Steven Rostedt 已提交
1019 1020
};

1021 1022
/* used to initialize the real stat files */
static struct tracer_stat function_stats __initdata = {
1023 1024 1025 1026 1027 1028
	.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
1029 1030
};

1031
static __init void ftrace_profile_tracefs(struct dentry *d_tracer)
S
Steven Rostedt 已提交
1032
{
1033
	struct ftrace_profile_stat *stat;
S
Steven Rostedt 已提交
1034
	struct dentry *entry;
1035
	char *name;
S
Steven Rostedt 已提交
1036
	int ret;
1037 1038 1039 1040 1041
	int cpu;

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

1042
		name = kasprintf(GFP_KERNEL, "function%d", cpu);
1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062
		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
1063 1064
	}

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

#else /* CONFIG_FUNCTION_PROFILER */
1072
static __init void ftrace_profile_tracefs(struct dentry *d_tracer)
S
Steven Rostedt 已提交
1073 1074 1075 1076
{
}
#endif /* CONFIG_FUNCTION_PROFILER */

1077 1078
static struct pid * const ftrace_swapper_pid = &init_struct_pid;

1079 1080 1081 1082 1083 1084
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static int ftrace_graph_active;
#else
# define ftrace_graph_active 0
#endif

1085 1086
#ifdef CONFIG_DYNAMIC_FTRACE

1087 1088
static struct ftrace_ops *removed_ops;

1089 1090 1091 1092 1093 1094
/*
 * 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;

1095 1096 1097 1098
#ifndef CONFIG_FTRACE_MCOUNT_RECORD
# error Dynamic ftrace depends on MCOUNT_RECORD
#endif

1099 1100 1101 1102 1103
struct ftrace_func_entry {
	struct hlist_node hlist;
	unsigned long ip;
};

1104 1105 1106 1107 1108 1109 1110 1111
struct ftrace_func_probe {
	struct ftrace_probe_ops	*probe_ops;
	struct ftrace_ops	ops;
	struct trace_array	*tr;
	struct list_head	list;
	int			ref;
};

1112 1113 1114 1115 1116 1117 1118 1119 1120
/*
 * 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,
1121
};
1122
#define EMPTY_HASH	((struct ftrace_hash *)&empty_hash)
1123

1124
static struct ftrace_ops global_ops = {
1125 1126 1127 1128 1129
	.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 |
1130 1131
					  FTRACE_OPS_FL_INITIALIZED |
					  FTRACE_OPS_FL_PID,
1132 1133
};

1134 1135
/*
 * This is used by __kernel_text_address() to return true if the
1136
 * address is on a dynamically allocated trampoline that would
1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170
 * 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;
}

1171 1172
struct ftrace_page {
	struct ftrace_page	*next;
1173
	struct dyn_ftrace	*records;
1174
	int			index;
1175
	int			size;
1176 1177
};

1178 1179
#define ENTRY_SIZE sizeof(struct dyn_ftrace)
#define ENTRIES_PER_PAGE (PAGE_SIZE / ENTRY_SIZE)
1180 1181 1182 1183 1184 1185 1186

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

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

1187 1188 1189 1190 1191 1192 1193 1194 1195
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;
}

1196 1197 1198
/* 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)
1199 1200 1201 1202 1203
{
	unsigned long key;
	struct ftrace_func_entry *entry;
	struct hlist_head *hhd;

1204
	key = ftrace_hash_key(hash, ip);
1205 1206
	hhd = &hash->buckets[key];

1207
	hlist_for_each_entry_rcu_notrace(entry, hhd, hlist) {
1208 1209 1210 1211 1212 1213
		if (entry->ip == ip)
			return entry;
	}
	return NULL;
}

1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232
/**
 * 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);
}

1233 1234
static void __add_hash_entry(struct ftrace_hash *hash,
			     struct ftrace_func_entry *entry)
1235 1236 1237 1238
{
	struct hlist_head *hhd;
	unsigned long key;

1239
	key = ftrace_hash_key(hash, entry->ip);
1240 1241 1242
	hhd = &hash->buckets[key];
	hlist_add_head(&entry->hlist, hhd);
	hash->count++;
1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254
}

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);
1255 1256 1257 1258 1259

	return 0;
}

static void
1260
free_hash_entry(struct ftrace_hash *hash,
1261 1262 1263 1264 1265 1266 1267
		  struct ftrace_func_entry *entry)
{
	hlist_del(&entry->hlist);
	kfree(entry);
	hash->count--;
}

1268 1269 1270 1271
static void
remove_hash_entry(struct ftrace_hash *hash,
		  struct ftrace_func_entry *entry)
{
1272
	hlist_del_rcu(&entry->hlist);
1273 1274 1275
	hash->count--;
}

1276 1277 1278
static void ftrace_hash_clear(struct ftrace_hash *hash)
{
	struct hlist_head *hhd;
1279
	struct hlist_node *tn;
1280 1281 1282 1283
	struct ftrace_func_entry *entry;
	int size = 1 << hash->size_bits;
	int i;

1284 1285 1286
	if (!hash->count)
		return;

1287 1288
	for (i = 0; i < size; i++) {
		hhd = &hash->buckets[i];
1289
		hlist_for_each_entry_safe(entry, tn, hhd, hlist)
1290
			free_hash_entry(hash, entry);
1291 1292 1293 1294
	}
	FTRACE_WARN_ON(hash->count);
}

1295 1296 1297 1298 1299 1300 1301 1302 1303
static void free_ftrace_hash(struct ftrace_hash *hash)
{
	if (!hash || hash == EMPTY_HASH)
		return;
	ftrace_hash_clear(hash);
	kfree(hash->buckets);
	kfree(hash);
}

1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318
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);
}

1319 1320
void ftrace_free_filter(struct ftrace_ops *ops)
{
1321
	ftrace_ops_init(ops);
1322 1323
	free_ftrace_hash(ops->func_hash->filter_hash);
	free_ftrace_hash(ops->func_hash->notrace_hash);
1324 1325
}

1326 1327 1328 1329 1330 1331 1332 1333 1334 1335
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;
1336
	hash->buckets = kcalloc(size, sizeof(*hash->buckets), GFP_KERNEL);
1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361

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

	hash->size_bits = size_bits;

	return hash;
}

static struct ftrace_hash *
alloc_and_copy_ftrace_hash(int size_bits, struct ftrace_hash *hash)
{
	struct ftrace_func_entry *entry;
	struct ftrace_hash *new_hash;
	int size;
	int ret;
	int i;

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

	/* Empty hash? */
1362
	if (ftrace_hash_empty(hash))
1363 1364 1365 1366
		return new_hash;

	size = 1 << hash->size_bits;
	for (i = 0; i < size; i++) {
1367
		hlist_for_each_entry(entry, &hash->buckets[i], hlist) {
1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382
			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;
}

1383
static void
1384
ftrace_hash_rec_disable_modify(struct ftrace_ops *ops, int filter_hash);
1385
static void
1386
ftrace_hash_rec_enable_modify(struct ftrace_ops *ops, int filter_hash);
1387

1388 1389 1390
static int ftrace_hash_ipmodify_update(struct ftrace_ops *ops,
				       struct ftrace_hash *new_hash);

1391 1392
static struct ftrace_hash *
__ftrace_hash_move(struct ftrace_hash *src)
1393 1394
{
	struct ftrace_func_entry *entry;
1395
	struct hlist_node *tn;
1396
	struct hlist_head *hhd;
1397
	struct ftrace_hash *new_hash;
1398 1399 1400 1401 1402
	int size = src->count;
	int bits = 0;
	int i;

	/*
1403
	 * If the new source is empty, just return the empty_hash.
1404
	 */
1405 1406
	if (!src->count)
		return EMPTY_HASH;
1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417

	/*
	 * 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;

1418 1419
	new_hash = alloc_ftrace_hash(bits);
	if (!new_hash)
1420
		return NULL;
1421 1422 1423 1424

	size = 1 << src->size_bits;
	for (i = 0; i < size; i++) {
		hhd = &src->buckets[i];
1425
		hlist_for_each_entry_safe(entry, tn, hhd, hlist) {
1426
			remove_hash_entry(src, entry);
1427
			__add_hash_entry(new_hash, entry);
1428 1429 1430
		}
	}

1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448
	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;

1449 1450 1451 1452 1453 1454 1455 1456 1457 1458
	/* 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;
		}
	}

1459 1460 1461 1462
	/*
	 * Remove the current set, update the hash and add
	 * them back.
	 */
1463
	ftrace_hash_rec_disable_modify(ops, enable);
1464

1465 1466
	rcu_assign_pointer(*dst, new_hash);

1467
	ftrace_hash_rec_enable_modify(ops, enable);
1468

1469
	return 0;
1470 1471
}

1472 1473 1474 1475 1476 1477 1478 1479 1480 1481
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) ||
1482
		__ftrace_lookup_ip(hash->filter_hash, ip)) &&
1483
		(ftrace_hash_empty(hash->notrace_hash) ||
1484
		 !__ftrace_lookup_ip(hash->notrace_hash, ip));
1485 1486
}

1487 1488 1489 1490 1491 1492 1493 1494
/*
 * 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.
1495 1496 1497
 *
 * This needs to be called with preemption disabled as
 * the hashes are freed with call_rcu_sched().
1498 1499
 */
static int
1500
ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip, void *regs)
1501
{
1502
	struct ftrace_ops_hash hash;
1503 1504
	int ret;

1505 1506 1507 1508 1509 1510 1511 1512 1513 1514
#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

1515 1516
	hash.filter_hash = rcu_dereference_raw_notrace(ops->func_hash->filter_hash);
	hash.notrace_hash = rcu_dereference_raw_notrace(ops->func_hash->notrace_hash);
1517

1518
	if (hash_contains_ip(ip, &hash))
1519 1520 1521 1522 1523 1524 1525
		ret = 1;
	else
		ret = 0;

	return ret;
}

1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539
/*
 * 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()		\
		}				\
	}

1540 1541 1542

static int ftrace_cmp_recs(const void *a, const void *b)
{
1543 1544
	const struct dyn_ftrace *key = a;
	const struct dyn_ftrace *rec = b;
1545

1546
	if (key->flags < rec->ip)
1547
		return -1;
1548 1549
	if (key->ip >= rec->ip + MCOUNT_INSN_SIZE)
		return 1;
1550 1551 1552
	return 0;
}

1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565
/**
 * 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)
1566 1567 1568
{
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
1569
	struct dyn_ftrace key;
1570

1571 1572
	key.ip = start;
	key.flags = end;	/* overload flags, as it is unsigned long */
1573 1574

	for (pg = ftrace_pages_start; pg; pg = pg->next) {
1575 1576
		if (end < pg->records[0].ip ||
		    start >= (pg->records[pg->index - 1].ip + MCOUNT_INSN_SIZE))
1577
			continue;
1578 1579 1580 1581
		rec = bsearch(&key, pg->records, pg->index,
			      sizeof(struct dyn_ftrace),
			      ftrace_cmp_recs);
		if (rec)
1582
			return rec->ip;
1583
	}
1584 1585 1586 1587

	return 0;
}

1588 1589 1590 1591
/**
 * ftrace_location - return true if the ip giving is a traced location
 * @ip: the instruction pointer to check
 *
1592
 * Returns rec->ip if @ip given is a pointer to a ftrace location.
1593 1594 1595 1596
 * 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.
 */
1597
unsigned long ftrace_location(unsigned long ip)
1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611
{
	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.
 */
1612
int ftrace_text_reserved(const void *start, const void *end)
1613
{
1614 1615 1616 1617 1618 1619
	unsigned long ret;

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

	return (int)!!ret;
1620 1621
}

1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641
/* 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;
}

1642
static bool __ftrace_hash_rec_update(struct ftrace_ops *ops,
1643 1644 1645 1646 1647 1648 1649
				     int filter_hash,
				     bool inc)
{
	struct ftrace_hash *hash;
	struct ftrace_hash *other_hash;
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
1650
	bool update = false;
1651 1652 1653 1654 1655
	int count = 0;
	int all = 0;

	/* Only update if the ops has been registered */
	if (!(ops->flags & FTRACE_OPS_FL_ENABLED))
1656
		return false;
1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669

	/*
	 * 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) {
1670 1671
		hash = ops->func_hash->filter_hash;
		other_hash = ops->func_hash->notrace_hash;
1672
		if (ftrace_hash_empty(hash))
1673 1674 1675
			all = 1;
	} else {
		inc = !inc;
1676 1677
		hash = ops->func_hash->notrace_hash;
		other_hash = ops->func_hash->filter_hash;
1678 1679 1680 1681
		/*
		 * If the notrace hash has no items,
		 * then there's nothing to do.
		 */
1682
		if (ftrace_hash_empty(hash))
1683
			return false;
1684 1685 1686 1687 1688 1689 1690
	}

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

1691 1692 1693
		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

1694 1695 1696 1697 1698
		if (all) {
			/*
			 * Only the filter_hash affects all records.
			 * Update if the record is not in the notrace hash.
			 */
1699
			if (!other_hash || !ftrace_lookup_ip(other_hash, rec->ip))
1700 1701
				match = 1;
		} else {
1702 1703
			in_hash = !!ftrace_lookup_ip(hash, rec->ip);
			in_other_hash = !!ftrace_lookup_ip(other_hash, rec->ip);
1704 1705

			/*
1706 1707
			 * If filter_hash is set, we want to match all functions
			 * that are in the hash but not in the other hash.
1708
			 *
1709 1710 1711 1712 1713
			 * 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.
1714 1715 1716 1717
			 */
			if (filter_hash && in_hash && !in_other_hash)
				match = 1;
			else if (!filter_hash && in_hash &&
1718
				 (in_other_hash || ftrace_hash_empty(other_hash)))
1719 1720 1721 1722 1723 1724 1725
				match = 1;
		}
		if (!match)
			continue;

		if (inc) {
			rec->flags++;
1726
			if (FTRACE_WARN_ON(ftrace_rec_count(rec) == FTRACE_REF_MAX))
1727
				return false;
1728 1729 1730 1731 1732 1733

			/*
			 * 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.
			 */
1734
			if (ftrace_rec_count(rec) == 1 && ops->trampoline)
1735
				rec->flags |= FTRACE_FL_TRAMP;
1736
			else
1737 1738 1739
				/*
				 * If we are adding another function callback
				 * to this function, and the previous had a
1740 1741
				 * custom trampoline in use, then we need to go
				 * back to the default trampoline.
1742
				 */
1743
				rec->flags &= ~FTRACE_FL_TRAMP;
1744

1745 1746 1747 1748 1749 1750
			/*
			 * 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;
1751
		} else {
1752
			if (FTRACE_WARN_ON(ftrace_rec_count(rec) == 0))
1753
				return false;
1754
			rec->flags--;
1755

1756 1757 1758 1759 1760 1761
			/*
			 * 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.
			 */
1762
			if (ftrace_rec_count(rec) > 0 &&
1763 1764 1765 1766 1767
			    rec->flags & FTRACE_FL_REGS &&
			    ops->flags & FTRACE_OPS_FL_SAVE_REGS) {
				if (!test_rec_ops_needs_regs(rec))
					rec->flags &= ~FTRACE_FL_REGS;
			}
1768

1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779
			/*
			 * 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;

1780 1781 1782 1783
			/*
			 * flags will be cleared in ftrace_check_record()
			 * if rec count is zero.
			 */
1784 1785
		}
		count++;
1786 1787 1788 1789

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

1790 1791
		/* Shortcut, if we handled all records, we are done. */
		if (!all && count == hash->count)
1792
			return update;
1793
	} while_for_each_ftrace_rec();
1794 1795

	return update;
1796 1797
}

1798
static bool ftrace_hash_rec_disable(struct ftrace_ops *ops,
1799 1800
				    int filter_hash)
{
1801
	return __ftrace_hash_rec_update(ops, filter_hash, 0);
1802 1803
}

1804
static bool ftrace_hash_rec_enable(struct ftrace_ops *ops,
1805 1806
				   int filter_hash)
{
1807
	return __ftrace_hash_rec_update(ops, filter_hash, 1);
1808 1809
}

1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844
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);
}

1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877
/*
 * 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) {
1878 1879 1880 1881

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903
		/* 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) {
1904 1905 1906 1907

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960
		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);
}

1961
static void print_ip_ins(const char *fmt, const unsigned char *p)
1962 1963 1964 1965 1966 1967 1968 1969 1970
{
	int i;

	printk(KERN_CONT "%s", fmt);

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

1971 1972
static struct ftrace_ops *
ftrace_find_tramp_ops_any(struct dyn_ftrace *rec);
1973 1974
static struct ftrace_ops *
ftrace_find_tramp_ops_next(struct dyn_ftrace *rec, struct ftrace_ops *ops);
1975

1976
enum ftrace_bug_type ftrace_bug_type;
1977
const void *ftrace_expected;
1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998

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;
	}
}

1999 2000 2001
/**
 * ftrace_bug - report and shutdown function tracer
 * @failed: The failed type (EFAULT, EINVAL, EPERM)
2002
 * @rec: The record that failed
2003 2004 2005 2006 2007 2008 2009 2010
 *
 * 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
 */
2011
void ftrace_bug(int failed, struct dyn_ftrace *rec)
2012
{
2013 2014
	unsigned long ip = rec ? rec->ip : 0;

2015 2016 2017 2018 2019 2020 2021 2022 2023 2024
	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);
2025
		print_ip_ins(" actual:   ", (unsigned char *)ip);
2026
		pr_cont("\n");
2027 2028 2029 2030
		if (ftrace_expected) {
			print_ip_ins(" expected: ", ftrace_expected);
			pr_cont("\n");
		}
2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041
		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);
	}
2042
	print_bug_type();
2043 2044 2045 2046 2047 2048 2049 2050
	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);
2051 2052 2053 2054 2055 2056 2057 2058
			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
2059 2060 2061 2062
				pr_cont("\ttramp: ERROR!");

		}
		ip = ftrace_get_addr_curr(rec);
2063
		pr_cont("\n expected tramp: %lx\n", ip);
2064
	}
2065 2066
}

2067
static int ftrace_check_record(struct dyn_ftrace *rec, int enable, int update)
2068
{
2069
	unsigned long flag = 0UL;
2070

2071 2072
	ftrace_bug_type = FTRACE_BUG_UNKNOWN;

2073 2074 2075
	if (rec->flags & FTRACE_FL_DISABLED)
		return FTRACE_UPDATE_IGNORE;

S
Steven Rostedt 已提交
2076
	/*
2077
	 * If we are updating calls:
S
Steven Rostedt 已提交
2078
	 *
2079 2080
	 *   If the record has a ref count, then we need to enable it
	 *   because someone is using it.
S
Steven Rostedt 已提交
2081
	 *
2082 2083
	 *   Otherwise we make sure its disabled.
	 *
2084
	 * If we are disabling calls, then disable all records that
2085
	 * are enabled.
S
Steven Rostedt 已提交
2086
	 */
2087
	if (enable && ftrace_rec_count(rec))
2088
		flag = FTRACE_FL_ENABLED;
S
Steven Rostedt 已提交
2089

2090
	/*
2091 2092 2093
	 * 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.
2094
	 */
2095 2096 2097 2098 2099 2100 2101 2102 2103
	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;
	}
2104

2105 2106
	/* If the state of this record hasn't changed, then do nothing */
	if ((rec->flags & FTRACE_FL_ENABLED) == flag)
2107
		return FTRACE_UPDATE_IGNORE;
S
Steven Rostedt 已提交
2108

2109
	if (flag) {
2110 2111 2112 2113
		/* Save off if rec is being enabled (for return value) */
		flag ^= rec->flags & FTRACE_FL_ENABLED;

		if (update) {
2114
			rec->flags |= FTRACE_FL_ENABLED;
2115 2116 2117 2118 2119 2120
			if (flag & FTRACE_FL_REGS) {
				if (rec->flags & FTRACE_FL_REGS)
					rec->flags |= FTRACE_FL_REGS_EN;
				else
					rec->flags &= ~FTRACE_FL_REGS_EN;
			}
2121 2122 2123 2124 2125 2126
			if (flag & FTRACE_FL_TRAMP) {
				if (rec->flags & FTRACE_FL_TRAMP)
					rec->flags |= FTRACE_FL_TRAMP_EN;
				else
					rec->flags &= ~FTRACE_FL_TRAMP_EN;
			}
2127 2128 2129 2130 2131 2132 2133
		}

		/*
		 * 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
2134
		 *   from the save regs, to a non-save regs function or
2135
		 *   vice versa, or from a trampoline call.
2136
		 */
2137 2138
		if (flag & FTRACE_FL_ENABLED) {
			ftrace_bug_type = FTRACE_BUG_CALL;
2139
			return FTRACE_UPDATE_MAKE_CALL;
2140
		}
2141

2142
		ftrace_bug_type = FTRACE_BUG_UPDATE;
2143
		return FTRACE_UPDATE_MODIFY_CALL;
2144 2145
	}

2146 2147
	if (update) {
		/* If there's no more users, clear all flags */
2148
		if (!ftrace_rec_count(rec))
2149 2150
			rec->flags = 0;
		else
2151 2152 2153 2154 2155 2156
			/*
			 * 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);
2157
	}
2158

2159
	ftrace_bug_type = FTRACE_BUG_NOP;
2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189
	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);
}

2190 2191 2192 2193
static struct ftrace_ops *
ftrace_find_tramp_ops_any(struct dyn_ftrace *rec)
{
	struct ftrace_ops *op;
2194
	unsigned long ip = rec->ip;
2195 2196 2197 2198 2199 2200

	do_for_each_ftrace_op(op, ftrace_ops_list) {

		if (!op->trampoline)
			continue;

2201
		if (hash_contains_ip(ip, op->func_hash))
2202 2203 2204 2205 2206 2207
			return op;
	} while_for_each_ftrace_op(op);

	return NULL;
}

2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225
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;
}

2226 2227 2228 2229
static struct ftrace_ops *
ftrace_find_tramp_ops_curr(struct dyn_ftrace *rec)
{
	struct ftrace_ops *op;
2230
	unsigned long ip = rec->ip;
2231

2232 2233 2234 2235 2236 2237 2238 2239
	/*
	 * 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))
2240 2241 2242
			return removed_ops;
	}

2243 2244 2245 2246 2247 2248 2249
	/*
	 * 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
2250 2251 2252 2253 2254 2255 2256 2257 2258 2259
	 * 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.
2260
	 */
2261
	do_for_each_ftrace_op(op, ftrace_ops_list) {
2262 2263 2264 2265 2266 2267 2268 2269 2270

		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)
2271 2272
			continue;

2273

2274
		/*
2275 2276 2277
		 * If the ops is being modified and is in the old
		 * hash, then it is probably being removed from this
		 * function.
2278 2279 2280
		 */
		if ((op->flags & FTRACE_OPS_FL_MODIFYING) &&
		    hash_contains_ip(ip, &op->old_hash))
2281
			return op;
2282 2283 2284 2285 2286 2287 2288 2289
		/*
		 * 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;
2290 2291 2292 2293 2294 2295 2296 2297 2298 2299

	} 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;
2300
	unsigned long ip = rec->ip;
2301 2302 2303

	do_for_each_ftrace_op(op, ftrace_ops_list) {
		/* pass rec in as regs to have non-NULL val */
2304
		if (hash_contains_ip(ip, op->func_hash))
2305 2306 2307 2308 2309 2310
			return op;
	} while_for_each_ftrace_op(op);

	return NULL;
}

2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322
/**
 * 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)
{
2323 2324 2325 2326 2327 2328
	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)) {
2329 2330
			pr_warn("Bad trampoline accounting at: %p (%pS) (%lx)\n",
				(void *)rec->ip, (void *)rec->ip, rec->flags);
2331 2332 2333 2334 2335 2336
			/* Ftrace is shutting down, return anything */
			return (unsigned long)FTRACE_ADDR;
		}
		return ops->trampoline;
	}

2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354
	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)
{
2355 2356 2357 2358 2359 2360
	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)) {
2361 2362
			pr_warn("Bad trampoline accounting at: %p (%pS)\n",
				(void *)rec->ip, (void *)rec->ip);
2363 2364 2365 2366 2367 2368
			/* Ftrace is shutting down, return anything */
			return (unsigned long)FTRACE_ADDR;
		}
		return ops->trampoline;
	}

2369 2370 2371 2372 2373 2374
	if (rec->flags & FTRACE_FL_REGS_EN)
		return (unsigned long)FTRACE_REGS_ADDR;
	else
		return (unsigned long)FTRACE_ADDR;
}

2375 2376 2377
static int
__ftrace_replace_code(struct dyn_ftrace *rec, int enable)
{
2378
	unsigned long ftrace_old_addr;
2379 2380 2381
	unsigned long ftrace_addr;
	int ret;

2382
	ftrace_addr = ftrace_get_addr_new(rec);
2383

2384 2385 2386 2387
	/* 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);
2388

2389 2390
	ftrace_bug_type = FTRACE_BUG_UNKNOWN;

2391 2392 2393 2394 2395
	switch (ret) {
	case FTRACE_UPDATE_IGNORE:
		return 0;

	case FTRACE_UPDATE_MAKE_CALL:
2396
		ftrace_bug_type = FTRACE_BUG_CALL;
2397
		return ftrace_make_call(rec, ftrace_addr);
2398 2399

	case FTRACE_UPDATE_MAKE_NOP:
2400
		ftrace_bug_type = FTRACE_BUG_NOP;
2401
		return ftrace_make_nop(NULL, rec, ftrace_old_addr);
2402 2403

	case FTRACE_UPDATE_MODIFY_CALL:
2404
		ftrace_bug_type = FTRACE_BUG_UPDATE;
2405
		return ftrace_modify_call(rec, ftrace_old_addr, ftrace_addr);
2406 2407
	}

2408
	return -1; /* unknow ftrace bug */
2409 2410
}

2411
void __weak ftrace_replace_code(int enable)
2412 2413 2414
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
S
Steven Rostedt 已提交
2415
	int failed;
2416

2417 2418 2419
	if (unlikely(ftrace_disabled))
		return;

2420
	do_for_each_ftrace_rec(pg, rec) {
2421 2422 2423 2424

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

2425
		failed = __ftrace_replace_code(rec, enable);
2426
		if (failed) {
2427
			ftrace_bug(failed, rec);
2428 2429
			/* Stop processing */
			return;
2430
		}
2431
	} while_for_each_ftrace_rec();
2432 2433
}

2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505
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];
}

2506
static int
2507
ftrace_code_disable(struct module *mod, struct dyn_ftrace *rec)
2508
{
2509
	int ret;
2510

2511 2512 2513
	if (unlikely(ftrace_disabled))
		return 0;

2514
	ret = ftrace_make_nop(mod, rec, MCOUNT_ADDR);
2515
	if (ret) {
2516
		ftrace_bug_type = FTRACE_BUG_INIT;
2517
		ftrace_bug(ret, rec);
2518
		return 0;
2519
	}
2520
	return 1;
2521 2522
}

2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540
/*
 * 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;
}

2541
void ftrace_modify_all_code(int command)
2542
{
2543
	int update = command & FTRACE_UPDATE_TRACE_FUNC;
2544
	int err = 0;
2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555

	/*
	 * 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.
	 */
2556 2557 2558 2559 2560
	if (update) {
		err = ftrace_update_ftrace_func(ftrace_ops_list_func);
		if (FTRACE_WARN_ON(err))
			return;
	}
2561

2562
	if (command & FTRACE_UPDATE_CALLS)
2563
		ftrace_replace_code(1);
2564
	else if (command & FTRACE_DISABLE_CALLS)
2565 2566
		ftrace_replace_code(0);

2567 2568 2569 2570 2571 2572
	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);
2573 2574 2575
		err = ftrace_update_ftrace_func(ftrace_trace_function);
		if (FTRACE_WARN_ON(err))
			return;
2576
	}
2577

2578
	if (command & FTRACE_START_FUNC_RET)
2579
		err = ftrace_enable_ftrace_graph_caller();
2580
	else if (command & FTRACE_STOP_FUNC_RET)
2581 2582
		err = ftrace_disable_ftrace_graph_caller();
	FTRACE_WARN_ON(err);
2583 2584 2585 2586 2587 2588 2589
}

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

	ftrace_modify_all_code(*command);
2590

2591
	return 0;
2592 2593
}

2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617
/**
 * 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
2618
static void ftrace_run_update_code(int command)
2619
{
2620 2621 2622 2623 2624 2625 2626
	int ret;

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

2627 2628 2629 2630 2631 2632 2633 2634
	/*
	 * 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);

2635 2636
	ret = ftrace_arch_code_modify_post_process();
	FTRACE_WARN_ON(ret);
2637 2638
}

2639
static void ftrace_run_modify_code(struct ftrace_ops *ops, int command,
2640
				   struct ftrace_ops_hash *old_hash)
2641 2642
{
	ops->flags |= FTRACE_OPS_FL_MODIFYING;
2643 2644
	ops->old_hash.filter_hash = old_hash->filter_hash;
	ops->old_hash.notrace_hash = old_hash->notrace_hash;
2645
	ftrace_run_update_code(command);
2646
	ops->old_hash.filter_hash = NULL;
2647
	ops->old_hash.notrace_hash = NULL;
2648 2649 2650
	ops->flags &= ~FTRACE_OPS_FL_MODIFYING;
}

2651
static ftrace_func_t saved_ftrace_func;
2652
static int ftrace_start_up;
2653

2654 2655 2656 2657
void __weak arch_ftrace_trampoline_free(struct ftrace_ops *ops)
{
}

2658
static void per_cpu_ops_free(struct ftrace_ops *ops)
2659 2660 2661 2662
{
	free_percpu(ops->disabled);
}

2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674
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);
}
2675

2676 2677 2678 2679 2680 2681 2682
static void ftrace_startup_all(int command)
{
	update_all_ops = true;
	ftrace_startup_enable(command);
	update_all_ops = false;
}

2683
static int ftrace_startup(struct ftrace_ops *ops, int command)
2684
{
2685
	int ret;
2686

2687
	if (unlikely(ftrace_disabled))
2688
		return -ENODEV;
2689

2690 2691 2692 2693
	ret = __register_ftrace_function(ops);
	if (ret)
		return ret;

2694
	ftrace_start_up++;
2695

2696 2697 2698 2699 2700 2701 2702 2703 2704
	/*
	 * 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;
2705

2706 2707 2708 2709 2710 2711 2712 2713 2714
	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;
	}

2715 2716
	if (ftrace_hash_rec_enable(ops, 1))
		command |= FTRACE_UPDATE_CALLS;
2717

2718
	ftrace_startup_enable(command);
2719

2720 2721
	ops->flags &= ~FTRACE_OPS_FL_ADDING;

2722
	return 0;
2723 2724
}

2725
static int ftrace_shutdown(struct ftrace_ops *ops, int command)
2726
{
2727
	int ret;
2728

2729
	if (unlikely(ftrace_disabled))
2730 2731 2732 2733 2734
		return -ENODEV;

	ret = __unregister_ftrace_function(ops);
	if (ret)
		return ret;
2735

2736
	ftrace_start_up--;
2737 2738 2739 2740 2741 2742 2743
	/*
	 * 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);

2744 2745
	/* Disabling ipmodify never fails */
	ftrace_hash_ipmodify_disable(ops);
2746

2747 2748
	if (ftrace_hash_rec_disable(ops, 1))
		command |= FTRACE_UPDATE_CALLS;
2749

2750
	ops->flags &= ~FTRACE_OPS_FL_ENABLED;
2751

2752 2753 2754 2755
	if (saved_ftrace_func != ftrace_trace_function) {
		saved_ftrace_func = ftrace_trace_function;
		command |= FTRACE_UPDATE_TRACE_FUNC;
	}
2756

2757 2758
	if (!command || !ftrace_enabled) {
		/*
2759
		 * If these are per_cpu ops, they still need their
2760 2761 2762 2763
		 * per_cpu field freed. Since, function tracing is
		 * not currently active, we can just free them
		 * without synchronizing all CPUs.
		 */
2764 2765
		if (ops->flags & FTRACE_OPS_FL_PER_CPU)
			per_cpu_ops_free(ops);
2766
		return 0;
2767
	}
2768

2769 2770 2771 2772
	/*
	 * If the ops uses a trampoline, then it needs to be
	 * tested first on update.
	 */
2773
	ops->flags |= FTRACE_OPS_FL_REMOVING;
2774 2775
	removed_ops = ops;

2776 2777 2778 2779
	/* 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;

2780
	ftrace_run_update_code(command);
2781

2782 2783 2784 2785 2786 2787 2788 2789 2790
	/*
	 * If there's no more ops registered with ftrace, run a
	 * sanity check to make sure all rec flags are cleared.
	 */
	if (ftrace_ops_list == &ftrace_list_end) {
		struct ftrace_page *pg;
		struct dyn_ftrace *rec;

		do_for_each_ftrace_rec(pg, rec) {
2791
			if (FTRACE_WARN_ON_ONCE(rec->flags & ~FTRACE_FL_DISABLED))
2792 2793 2794 2795 2796
				pr_warn("  %pS flags:%lx\n",
					(void *)rec->ip, rec->flags);
		} while_for_each_ftrace_rec();
	}

2797 2798 2799 2800
	ops->old_hash.filter_hash = NULL;
	ops->old_hash.notrace_hash = NULL;

	removed_ops = NULL;
2801
	ops->flags &= ~FTRACE_OPS_FL_REMOVING;
2802

2803 2804 2805
	/*
	 * Dynamic ops may be freed, we must make sure that all
	 * callers are done before leaving this function.
2806
	 * The same goes for freeing the per_cpu data of the per_cpu
2807 2808
	 * ops.
	 */
2809
	if (ops->flags & (FTRACE_OPS_FL_DYNAMIC | FTRACE_OPS_FL_PER_CPU)) {
2810 2811 2812 2813 2814 2815 2816 2817
		/*
		 * We need to do a hard force of sched synchronization.
		 * This is because we use preempt_disable() to do RCU, but
		 * the function tracers can be called where RCU is not watching
		 * (like before user_exit()). We can not rely on the RCU
		 * infrastructure to do the synchronization, thus we must do it
		 * ourselves.
		 */
2818 2819
		schedule_on_each_cpu(ftrace_sync);

2820 2821 2822 2823 2824 2825 2826 2827 2828 2829
		/*
		 * 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();

2830 2831
		arch_ftrace_trampoline_free(ops);

2832 2833
		if (ops->flags & FTRACE_OPS_FL_PER_CPU)
			per_cpu_ops_free(ops);
2834 2835
	}

2836
	return 0;
2837 2838
}

I
Ingo Molnar 已提交
2839
static void ftrace_startup_sysctl(void)
2840
{
2841 2842
	int command;

2843 2844 2845
	if (unlikely(ftrace_disabled))
		return;

2846 2847
	/* Force update next time */
	saved_ftrace_func = NULL;
2848
	/* ftrace_start_up is true if we want ftrace running */
2849 2850 2851 2852
	if (ftrace_start_up) {
		command = FTRACE_UPDATE_CALLS;
		if (ftrace_graph_active)
			command |= FTRACE_START_FUNC_RET;
2853
		ftrace_startup_enable(command);
2854
	}
2855 2856
}

I
Ingo Molnar 已提交
2857
static void ftrace_shutdown_sysctl(void)
2858
{
2859 2860
	int command;

2861 2862 2863
	if (unlikely(ftrace_disabled))
		return;

2864
	/* ftrace_start_up is true if ftrace is running */
2865 2866 2867 2868 2869 2870
	if (ftrace_start_up) {
		command = FTRACE_DISABLE_CALLS;
		if (ftrace_graph_active)
			command |= FTRACE_STOP_FUNC_RET;
		ftrace_run_update_code(command);
	}
2871 2872
}

2873
static u64		ftrace_update_time;
2874 2875
unsigned long		ftrace_update_tot_cnt;

2876
static inline int ops_traces_mod(struct ftrace_ops *ops)
2877
{
2878 2879 2880 2881
	/*
	 * Filter_hash being empty will default to trace module.
	 * But notrace hash requires a test of individual module functions.
	 */
2882 2883
	return ftrace_hash_empty(ops->func_hash->filter_hash) &&
		ftrace_hash_empty(ops->func_hash->notrace_hash);
2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899
}

/*
 * 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;

2900
	/* If ops traces all then it includes this function */
2901
	if (ops_traces_mod(ops))
2902
		return 1;
2903 2904

	/* The function must be in the filter */
2905
	if (!ftrace_hash_empty(ops->func_hash->filter_hash) &&
2906
	    !__ftrace_lookup_ip(ops->func_hash->filter_hash, rec->ip))
2907
		return 0;
2908

2909
	/* If in notrace hash, we ignore it too */
2910
	if (ftrace_lookup_ip(ops->func_hash->notrace_hash, rec->ip))
2911 2912 2913 2914 2915
		return 0;

	return 1;
}

2916
static int ftrace_update_code(struct module *mod, struct ftrace_page *new_pgs)
2917
{
2918
	struct ftrace_page *pg;
2919
	struct dyn_ftrace *p;
2920
	u64 start, stop;
2921
	unsigned long update_cnt = 0;
2922
	unsigned long rec_flags = 0;
2923
	int i;
2924

2925 2926
	start = ftrace_now(raw_smp_processor_id());

2927
	/*
2928 2929 2930 2931 2932 2933 2934 2935 2936
	 * 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.
2937
	 */
2938 2939
	if (mod)
		rec_flags |= FTRACE_FL_DISABLED;
2940

2941
	for (pg = new_pgs; pg; pg = pg->next) {
2942

2943
		for (i = 0; i < pg->index; i++) {
2944

2945 2946 2947
			/* If something went wrong, bail without enabling anything */
			if (unlikely(ftrace_disabled))
				return -1;
2948

2949
			p = &pg->records[i];
2950
			p->flags = rec_flags;
2951

2952 2953 2954 2955 2956 2957
			/*
			 * Do the initial record conversion from mcount jump
			 * to the NOP instructions.
			 */
			if (!ftrace_code_disable(mod, p))
				break;
2958

2959
			update_cnt++;
2960
		}
2961 2962
	}

I
Ingo Molnar 已提交
2963
	stop = ftrace_now(raw_smp_processor_id());
2964
	ftrace_update_time = stop - start;
2965
	ftrace_update_tot_cnt += update_cnt;
2966

2967 2968 2969
	return 0;
}

2970
static int ftrace_allocate_records(struct ftrace_page *pg, int count)
2971
{
2972
	int order;
2973 2974
	int cnt;

2975 2976 2977 2978
	if (WARN_ON(!count))
		return -EINVAL;

	order = get_count_order(DIV_ROUND_UP(count, ENTRIES_PER_PAGE));
2979 2980

	/*
2981 2982
	 * We want to fill as much as possible. No more than a page
	 * may be empty.
2983
	 */
2984 2985
	while ((PAGE_SIZE << order) / ENTRY_SIZE >= count + ENTRIES_PER_PAGE)
		order--;
2986

2987 2988
 again:
	pg->records = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, order);
2989

2990 2991 2992 2993 2994 2995 2996
	if (!pg->records) {
		/* if we can't allocate this size, try something smaller */
		if (!order)
			return -ENOMEM;
		order >>= 1;
		goto again;
	}
2997

2998 2999
	cnt = (PAGE_SIZE << order) / ENTRY_SIZE;
	pg->size = cnt;
3000

3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033
	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)
3034 3035
			break;

3036 3037 3038 3039
		pg->next = kzalloc(sizeof(*pg), GFP_KERNEL);
		if (!pg->next)
			goto free_pages;

3040 3041 3042
		pg = pg->next;
	}

3043 3044 3045
	return start_pg;

 free_pages:
3046 3047
	pg = start_pg;
	while (pg) {
3048 3049 3050 3051 3052 3053 3054 3055 3056 3057
		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;
}

3058 3059 3060
#define FTRACE_BUFF_MAX (KSYM_SYMBOL_LEN+4) /* room for wildcards */

struct ftrace_iterator {
3061
	loff_t				pos;
3062 3063 3064
	loff_t				func_pos;
	struct ftrace_page		*pg;
	struct dyn_ftrace		*func;
3065
	struct ftrace_func_probe	*probe;
3066
	struct ftrace_func_entry	*probe_entry;
3067
	struct trace_parser		parser;
3068
	struct ftrace_hash		*hash;
3069
	struct ftrace_ops		*ops;
3070
	int				pidx;
3071 3072
	int				idx;
	unsigned			flags;
3073 3074
};

3075
static void *
3076
t_probe_next(struct seq_file *m, loff_t *pos)
3077 3078
{
	struct ftrace_iterator *iter = m->private;
3079 3080
	struct trace_array *tr = global_ops.private;
	struct list_head *func_probes;
3081 3082
	struct ftrace_hash *hash;
	struct list_head *next;
3083
	struct hlist_node *hnd = NULL;
3084
	struct hlist_head *hhd;
3085
	int size;
3086 3087

	(*pos)++;
3088
	iter->pos = *pos;
3089

3090 3091 3092 3093 3094
	if (!tr)
		return NULL;

	func_probes = &tr->func_probes;
	if (list_empty(func_probes))
3095 3096
		return NULL;

3097
	if (!iter->probe) {
3098
		next = func_probes->next;
3099
		iter->probe = list_entry(next, struct ftrace_func_probe, list);
3100 3101 3102 3103 3104 3105 3106 3107 3108 3109
	}

	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) {
3110
		if (iter->probe->list.next == func_probes)
3111 3112
			return NULL;
		next = iter->probe->list.next;
3113
		iter->probe = list_entry(next, struct ftrace_func_probe, list);
3114 3115 3116 3117 3118 3119
		hash = iter->probe->ops.func_hash->filter_hash;
		size = 1 << hash->size_bits;
		iter->pidx = 0;
	}

	hhd = &hash->buckets[iter->pidx];
3120 3121

	if (hlist_empty(hhd)) {
3122
		iter->pidx++;
3123 3124 3125 3126 3127 3128 3129 3130 3131
		hnd = NULL;
		goto retry;
	}

	if (!hnd)
		hnd = hhd->first;
	else {
		hnd = hnd->next;
		if (!hnd) {
3132
			iter->pidx++;
3133 3134 3135 3136
			goto retry;
		}
	}

3137 3138 3139
	if (WARN_ON_ONCE(!hnd))
		return NULL;

3140
	iter->probe_entry = hlist_entry(hnd, struct ftrace_func_entry, hlist);
3141 3142

	return iter;
3143 3144
}

3145
static void *t_probe_start(struct seq_file *m, loff_t *pos)
3146 3147 3148
{
	struct ftrace_iterator *iter = m->private;
	void *p = NULL;
L
Li Zefan 已提交
3149 3150
	loff_t l;

3151
	if (!(iter->flags & FTRACE_ITER_DO_PROBES))
3152 3153
		return NULL;

3154 3155
	if (iter->func_pos > *pos)
		return NULL;
3156

3157 3158 3159
	iter->probe = NULL;
	iter->probe_entry = NULL;
	iter->pidx = 0;
3160
	for (l = 0; l <= (*pos - iter->func_pos); ) {
3161
		p = t_probe_next(m, &l);
L
Li Zefan 已提交
3162 3163 3164
		if (!p)
			break;
	}
3165 3166 3167
	if (!p)
		return NULL;

3168
	/* Only set this if we have an item */
3169
	iter->flags |= FTRACE_ITER_PROBE;
3170

3171
	return iter;
3172 3173
}

3174
static int
3175
t_probe_show(struct seq_file *m, struct ftrace_iterator *iter)
3176
{
3177
	struct ftrace_func_entry *probe_entry;
3178 3179
	struct ftrace_probe_ops *probe_ops;
	struct ftrace_func_probe *probe;
3180

3181 3182 3183 3184
	probe = iter->probe;
	probe_entry = iter->probe_entry;

	if (WARN_ON_ONCE(!probe || !probe_entry))
3185
		return -EIO;
3186

3187 3188 3189 3190
	probe_ops = probe->probe_ops;

	if (probe_ops->print)
		return probe_ops->print(m, probe_entry->ip, probe_ops, NULL);
3191

3192 3193
	seq_printf(m, "%ps:%ps\n", (void *)probe_entry->ip,
		   (void *)probe_ops->func);
3194 3195 3196 3197

	return 0;
}

I
Ingo Molnar 已提交
3198
static void *
3199
t_func_next(struct seq_file *m, loff_t *pos)
3200 3201 3202 3203 3204
{
	struct ftrace_iterator *iter = m->private;
	struct dyn_ftrace *rec = NULL;

	(*pos)++;
3205

3206 3207 3208 3209 3210 3211 3212 3213 3214
 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++];
3215 3216
		if (((iter->flags & (FTRACE_ITER_FILTER | FTRACE_ITER_NOTRACE)) &&
		     !ftrace_lookup_ip(iter->hash, rec->ip)) ||
3217 3218

		    ((iter->flags & FTRACE_ITER_ENABLED) &&
3219
		     !(rec->flags & FTRACE_FL_ENABLED))) {
3220

3221 3222 3223 3224 3225
			rec = NULL;
			goto retry;
		}
	}

3226
	if (!rec)
3227
		return NULL;
3228

3229
	iter->pos = iter->func_pos = *pos;
3230 3231 3232
	iter->func = rec;

	return iter;
3233 3234
}

3235 3236 3237 3238
static void *
t_next(struct seq_file *m, void *v, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
3239
	loff_t l = *pos; /* t_hash_start() must use original pos */
3240 3241 3242 3243 3244
	void *ret;

	if (unlikely(ftrace_disabled))
		return NULL;

3245 3246
	if (iter->flags & FTRACE_ITER_PROBE)
		return t_probe_next(m, pos);
3247 3248

	if (iter->flags & FTRACE_ITER_PRINTALL) {
3249
		/* next must increment pos, and t_probe_start does not */
3250
		(*pos)++;
3251
		return t_probe_start(m, &l);
3252 3253 3254 3255 3256
	}

	ret = t_func_next(m, pos);

	if (!ret)
3257
		return t_probe_start(m, &l);
3258 3259 3260 3261

	return ret;
}

3262 3263 3264 3265
static void reset_iter_read(struct ftrace_iterator *iter)
{
	iter->pos = 0;
	iter->func_pos = 0;
3266
	iter->flags &= ~(FTRACE_ITER_PRINTALL | FTRACE_ITER_PROBE);
3267 3268 3269 3270 3271 3272
}

static void *t_start(struct seq_file *m, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
	void *p = NULL;
3273
	loff_t l;
3274

3275
	mutex_lock(&ftrace_lock);
3276 3277 3278 3279

	if (unlikely(ftrace_disabled))
		return NULL;

3280 3281 3282 3283 3284 3285
	/*
	 * If an lseek was done, then reset and start from beginning.
	 */
	if (*pos < iter->pos)
		reset_iter_read(iter);

3286 3287 3288 3289 3290
	/*
	 * For set_ftrace_filter reading, if we have the filter
	 * off, we can short cut and just print out that all
	 * functions are enabled.
	 */
3291 3292
	if ((iter->flags & (FTRACE_ITER_FILTER | FTRACE_ITER_NOTRACE)) &&
	    ftrace_hash_empty(iter->hash)) {
3293
		iter->func_pos = 1; /* Account for the message */
3294
		if (*pos > 0)
3295
			return t_probe_start(m, pos);
3296
		iter->flags |= FTRACE_ITER_PRINTALL;
3297
		/* reset in case of seek/pread */
3298
		iter->flags &= ~FTRACE_ITER_PROBE;
3299 3300 3301
		return iter;
	}

3302 3303
	if (iter->flags & FTRACE_ITER_PROBE)
		return t_probe_start(m, pos);
3304

3305 3306 3307 3308 3309
	/*
	 * 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.
	 */
3310 3311 3312
	iter->pg = ftrace_pages_start;
	iter->idx = 0;
	for (l = 0; l <= *pos; ) {
3313
		p = t_func_next(m, &l);
3314 3315
		if (!p)
			break;
3316
	}
3317

3318
	if (!p)
3319
		return t_probe_start(m, pos);
3320 3321

	return iter;
3322 3323 3324 3325
}

static void t_stop(struct seq_file *m, void *p)
{
3326
	mutex_unlock(&ftrace_lock);
3327 3328
}

3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344
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);
}

3345 3346
static int t_show(struct seq_file *m, void *v)
{
3347
	struct ftrace_iterator *iter = m->private;
3348
	struct dyn_ftrace *rec;
3349

3350 3351
	if (iter->flags & FTRACE_ITER_PROBE)
		return t_probe_show(m, iter);
3352

3353
	if (iter->flags & FTRACE_ITER_PRINTALL) {
3354
		if (iter->flags & FTRACE_ITER_NOTRACE)
3355
			seq_puts(m, "#### no functions disabled ####\n");
3356
		else
3357
			seq_puts(m, "#### all functions enabled ####\n");
3358 3359 3360
		return 0;
	}

3361 3362
	rec = iter->func;

3363 3364 3365
	if (!rec)
		return 0;

3366
	seq_printf(m, "%ps", (void *)rec->ip);
3367
	if (iter->flags & FTRACE_ITER_ENABLED) {
3368
		struct ftrace_ops *ops;
3369

3370
		seq_printf(m, " (%ld)%s%s",
3371
			   ftrace_rec_count(rec),
3372 3373
			   rec->flags & FTRACE_FL_REGS ? " R" : "  ",
			   rec->flags & FTRACE_FL_IPMODIFY ? " I" : "  ");
3374
		if (rec->flags & FTRACE_FL_TRAMP_EN) {
3375
			ops = ftrace_find_tramp_ops_any(rec);
3376 3377 3378 3379 3380
			if (ops) {
				do {
					seq_printf(m, "\ttramp: %pS (%pS)",
						   (void *)ops->trampoline,
						   (void *)ops->func);
3381
					add_trampoline_func(m, ops, rec);
3382 3383 3384
					ops = ftrace_find_tramp_ops_next(rec, ops);
				} while (ops);
			} else
3385
				seq_puts(m, "\ttramp: ERROR!");
3386 3387
		} else {
			add_trampoline_func(m, NULL, rec);
3388 3389 3390
		}
	}	

3391
	seq_putc(m, '\n');
3392 3393 3394 3395

	return 0;
}

J
James Morris 已提交
3396
static const struct seq_operations show_ftrace_seq_ops = {
3397 3398 3399 3400 3401 3402
	.start = t_start,
	.next = t_next,
	.stop = t_stop,
	.show = t_show,
};

I
Ingo Molnar 已提交
3403
static int
3404 3405 3406 3407
ftrace_avail_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;

3408 3409 3410
	if (unlikely(ftrace_disabled))
		return -ENODEV;

3411
	iter = __seq_open_private(file, &show_ftrace_seq_ops, sizeof(*iter));
3412 3413
	if (!iter)
		return -ENOMEM;
3414

3415 3416 3417 3418
	iter->pg = ftrace_pages_start;
	iter->ops = &global_ops;

	return 0;
3419 3420
}

3421 3422 3423 3424 3425
static int
ftrace_enabled_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;

3426
	iter = __seq_open_private(file, &show_ftrace_seq_ops, sizeof(*iter));
3427 3428
	if (!iter)
		return -ENOMEM;
3429

3430 3431 3432 3433 3434
	iter->pg = ftrace_pages_start;
	iter->flags = FTRACE_ITER_ENABLED;
	iter->ops = &global_ops;

	return 0;
3435 3436
}

3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449
/**
 * 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.
3450
 * tracing_lseek() should be used as the lseek routine, and
3451 3452 3453
 * release must call ftrace_regex_release().
 */
int
3454
ftrace_regex_open(struct ftrace_ops *ops, int flag,
3455
		  struct inode *inode, struct file *file)
3456 3457
{
	struct ftrace_iterator *iter;
3458
	struct ftrace_hash *hash;
3459 3460
	int ret = 0;

3461 3462
	ftrace_ops_init(ops);

3463 3464 3465
	if (unlikely(ftrace_disabled))
		return -ENODEV;

3466 3467 3468 3469
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
	if (!iter)
		return -ENOMEM;

3470 3471 3472 3473 3474
	if (trace_parser_get_init(&iter->parser, FTRACE_BUFF_MAX)) {
		kfree(iter);
		return -ENOMEM;
	}

3475 3476 3477
	iter->ops = ops;
	iter->flags = flag;

3478
	mutex_lock(&ops->func_hash->regex_lock);
3479

3480
	if (flag & FTRACE_ITER_NOTRACE)
3481
		hash = ops->func_hash->notrace_hash;
3482
	else
3483
		hash = ops->func_hash->filter_hash;
3484

3485
	if (file->f_mode & FMODE_WRITE) {
3486 3487 3488 3489 3490 3491 3492
		const int size_bits = FTRACE_HASH_DEFAULT_BITS;

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

3493 3494 3495
		if (!iter->hash) {
			trace_parser_put(&iter->parser);
			kfree(iter);
3496 3497
			ret = -ENOMEM;
			goto out_unlock;
3498
		}
3499 3500
	} else
		iter->hash = hash;
3501

3502 3503 3504 3505 3506 3507 3508
	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;
3509
		} else {
3510 3511
			/* Failed */
			free_ftrace_hash(iter->hash);
3512
			trace_parser_put(&iter->parser);
3513
			kfree(iter);
3514
		}
3515 3516
	} else
		file->private_data = iter;
3517 3518

 out_unlock:
3519
	mutex_unlock(&ops->func_hash->regex_lock);
3520 3521 3522 3523

	return ret;
}

3524 3525 3526
static int
ftrace_filter_open(struct inode *inode, struct file *file)
{
3527 3528 3529
	struct ftrace_ops *ops = inode->i_private;

	return ftrace_regex_open(ops,
3530
			FTRACE_ITER_FILTER | FTRACE_ITER_DO_PROBES,
3531
			inode, file);
3532 3533 3534 3535 3536
}

static int
ftrace_notrace_open(struct inode *inode, struct file *file)
{
3537 3538 3539
	struct ftrace_ops *ops = inode->i_private;

	return ftrace_regex_open(ops, FTRACE_ITER_NOTRACE,
3540
				 inode, file);
3541 3542
}

3543 3544 3545 3546 3547 3548 3549
/* Type for quick search ftrace basic regexes (globs) from filter_parse_regex */
struct ftrace_glob {
	char *search;
	unsigned len;
	int type;
};

3550 3551 3552 3553 3554 3555 3556 3557 3558 3559
/*
 * 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;
}

3560
static int ftrace_match(char *str, struct ftrace_glob *g)
3561 3562
{
	int matched = 0;
3563
	int slen;
3564

3565 3566
	str = arch_ftrace_match_adjust(str, g->search);

3567
	switch (g->type) {
3568
	case MATCH_FULL:
3569
		if (strcmp(str, g->search) == 0)
3570 3571 3572
			matched = 1;
		break;
	case MATCH_FRONT_ONLY:
3573
		if (strncmp(str, g->search, g->len) == 0)
3574 3575 3576
			matched = 1;
		break;
	case MATCH_MIDDLE_ONLY:
3577
		if (strstr(str, g->search))
3578 3579 3580
			matched = 1;
		break;
	case MATCH_END_ONLY:
3581
		slen = strlen(str);
3582 3583
		if (slen >= g->len &&
		    memcmp(str + slen - g->len, g->search, g->len) == 0)
3584 3585
			matched = 1;
		break;
3586 3587 3588 3589
	case MATCH_GLOB:
		if (glob_match(g->search, str))
			matched = 1;
		break;
3590 3591 3592 3593 3594
	}

	return matched;
}

3595
static int
3596
enter_record(struct ftrace_hash *hash, struct dyn_ftrace *rec, int clear_filter)
3597
{
3598 3599 3600
	struct ftrace_func_entry *entry;
	int ret = 0;

3601
	entry = ftrace_lookup_ip(hash, rec->ip);
3602
	if (clear_filter) {
3603 3604 3605
		/* Do nothing if it doesn't exist */
		if (!entry)
			return 0;
3606

3607
		free_hash_entry(hash, entry);
3608 3609 3610 3611
	} else {
		/* Do nothing if it exists */
		if (entry)
			return 0;
3612

3613
		ret = add_hash_entry(hash, rec->ip);
3614 3615
	}
	return ret;
3616 3617
}

3618
static int
D
Dmitry Safonov 已提交
3619 3620
ftrace_match_record(struct dyn_ftrace *rec, struct ftrace_glob *func_g,
		struct ftrace_glob *mod_g, int exclude_mod)
3621 3622
{
	char str[KSYM_SYMBOL_LEN];
3623 3624 3625 3626
	char *modname;

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

D
Dmitry Safonov 已提交
3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646
	if (mod_g) {
		int mod_matches = (modname) ? ftrace_match(modname, mod_g) : 0;

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

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

		if (mod_matches && exclude_mod)
3647 3648
			return 0;

D
Dmitry Safonov 已提交
3649
func_match:
3650
		/* blank search means to match all funcs in the mod */
3651
		if (!func_g->len)
3652 3653
			return 1;
	}
3654

3655
	return ftrace_match(str, func_g);
3656 3657
}

3658
static int
3659
match_records(struct ftrace_hash *hash, char *func, int len, char *mod)
3660 3661 3662
{
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
3663
	struct ftrace_glob func_g = { .type = MATCH_FULL };
D
Dmitry Safonov 已提交
3664 3665 3666
	struct ftrace_glob mod_g = { .type = MATCH_FULL };
	struct ftrace_glob *mod_match = (mod) ? &mod_g : NULL;
	int exclude_mod = 0;
3667
	int found = 0;
3668
	int ret;
3669
	int clear_filter;
3670

D
Dmitry Safonov 已提交
3671
	if (func) {
3672 3673 3674
		func_g.type = filter_parse_regex(func, len, &func_g.search,
						 &clear_filter);
		func_g.len = strlen(func_g.search);
3675
	}
3676

D
Dmitry Safonov 已提交
3677 3678 3679 3680
	if (mod) {
		mod_g.type = filter_parse_regex(mod, strlen(mod),
				&mod_g.search, &exclude_mod);
		mod_g.len = strlen(mod_g.search);
3681
	}
3682

3683
	mutex_lock(&ftrace_lock);
3684

3685 3686
	if (unlikely(ftrace_disabled))
		goto out_unlock;
3687

3688
	do_for_each_ftrace_rec(pg, rec) {
3689 3690 3691 3692

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

D
Dmitry Safonov 已提交
3693
		if (ftrace_match_record(rec, &func_g, mod_match, exclude_mod)) {
3694
			ret = enter_record(hash, rec, clear_filter);
3695 3696 3697 3698
			if (ret < 0) {
				found = ret;
				goto out_unlock;
			}
3699
			found = 1;
3700 3701
		}
	} while_for_each_ftrace_rec();
3702
 out_unlock:
3703
	mutex_unlock(&ftrace_lock);
3704 3705

	return found;
3706 3707
}

3708
static int
3709
ftrace_match_records(struct ftrace_hash *hash, char *buff, int len)
3710
{
3711
	return match_records(hash, buff, len, NULL);
3712 3713
}

3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763
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;
}
3764

3765 3766 3767 3768 3769 3770
/*
 * We register the module command as a template to show others how
 * to register the a command as well.
 */

static int
3771
ftrace_mod_callback(struct trace_array *tr, struct ftrace_hash *hash,
3772
		    char *func, char *cmd, char *module, int enable)
3773
{
3774
	int ret;
3775 3776 3777 3778 3779 3780 3781 3782

	/*
	 * 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.
	 */
3783
	ret = match_records(hash, func, strlen(func), module);
3784
	if (!ret)
3785
		return -EINVAL;
3786 3787 3788
	if (ret < 0)
		return ret;
	return 0;
3789 3790 3791 3792 3793 3794 3795 3796 3797 3798 3799
}

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);
}
3800
core_initcall(ftrace_mod_cmd_init);
3801

3802
static void function_trace_probe_call(unsigned long ip, unsigned long parent_ip,
3803
				      struct ftrace_ops *op, struct pt_regs *pt_regs)
3804
{
3805
	struct ftrace_probe_ops *probe_ops;
3806
	struct ftrace_func_probe *probe;
3807

3808 3809
	probe = container_of(op, struct ftrace_func_probe, ops);
	probe_ops = probe->probe_ops;
3810 3811 3812 3813 3814 3815

	/*
	 * 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.
	 */
3816
	preempt_disable_notrace();
3817
	probe_ops->func(ip, parent_ip, probe->tr, probe_ops, NULL);
3818
	preempt_enable_notrace();
3819 3820
}

3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834 3835 3836 3837 3838 3839 3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874 3875 3876 3877 3878 3879 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 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
struct ftrace_func_map {
	struct ftrace_func_entry	entry;
	void				*data;
};

struct ftrace_func_mapper {
	struct ftrace_hash		hash;
};

/**
 * 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)
{
	struct ftrace_hash *hash;

	/*
	 * 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;
}

/**
 * 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;

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

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

/**
 * 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)
{
	struct ftrace_func_entry *entry;
	struct ftrace_func_map *map;

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

	map = kmalloc(sizeof(*map), GFP_KERNEL);
	if (!map)
		return -ENOMEM;

	map->entry.ip = ip;
	map->data = data;

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

	return 0;
}

/**
 * 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)
{
	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);
	kfree(entry);

	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);
}

3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972 3973 3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989
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;
		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++;
}

3990
int
3991
register_ftrace_function_probe(char *glob, struct trace_array *tr,
3992 3993
			       struct ftrace_probe_ops *probe_ops,
			       void *data)
3994
{
3995
	struct ftrace_func_entry *entry;
3996
	struct ftrace_func_probe *probe;
3997 3998
	struct ftrace_hash **orig_hash;
	struct ftrace_hash *old_hash;
3999
	struct ftrace_hash *hash;
4000
	int count = 0;
4001
	int size;
4002
	int ret;
4003
	int i;
4004

4005 4006 4007
	if (WARN_ON(!tr))
		return -EINVAL;

4008 4009
	/* We do not support '!' for function probes */
	if (WARN_ON(glob[0] == '!'))
4010 4011
		return -EINVAL;

4012 4013 4014 4015 4016 4017 4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029

	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;
	}
	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);
4030
	}
4031

4032
	acquire_probe_locked(probe);
4033

4034 4035 4036 4037 4038
	mutex_unlock(&ftrace_lock);

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

	orig_hash = &probe->ops.func_hash->filter_hash;
4039 4040
	old_hash = *orig_hash;
	hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, old_hash);
4041

4042
	ret = ftrace_match_records(hash, glob, strlen(glob));
4043

4044 4045 4046
	/* Nothing found? */
	if (!ret)
		ret = -EINVAL;
4047

4048 4049
	if (ret < 0)
		goto out;
4050

4051 4052 4053 4054
	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))
4055
				continue;
4056 4057 4058 4059 4060
			/*
			 * 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.
			 */
4061 4062 4063
			if (probe_ops->init) {
				ret = probe_ops->init(probe_ops, tr,
						      entry->ip, data);
4064 4065
				if (ret < 0)
					goto out;
4066 4067
			}
			count++;
4068
		}
4069
	}
4070

4071
	mutex_lock(&ftrace_lock);
4072

4073 4074 4075 4076 4077 4078 4079 4080
	if (!count) {
		/* Nothing was added? */
		ret = -EINVAL;
		goto out_unlock;
	}

	ret = ftrace_hash_move_and_update_ops(&probe->ops, orig_hash,
					      hash, 1);
4081
	if (ret < 0)
4082
		goto err_unlock;
4083

4084 4085
	/* One ref for each new function traced */
	probe->ref += count;
4086

4087 4088
	if (!(probe->ops.flags & FTRACE_OPS_FL_ENABLED))
		ret = ftrace_startup(&probe->ops, 0);
4089

4090
 out_unlock:
4091
	mutex_unlock(&ftrace_lock);
4092

4093
	if (!ret)
4094
		ret = count;
4095
 out:
4096
	mutex_unlock(&probe->ops.func_hash->regex_lock);
4097
	free_ftrace_hash(hash);
4098

4099 4100
	release_probe(probe);

4101
	return ret;
4102 4103

 err_unlock:
4104
	if (!probe_ops->free || !count)
4105 4106 4107 4108 4109 4110 4111
		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;
4112
			probe_ops->free(probe_ops, tr, entry->ip, NULL);
4113 4114 4115
		}
	}
	goto out_unlock;
4116 4117
}

4118
int
4119 4120
unregister_ftrace_function_probe_func(char *glob, struct trace_array *tr,
				      struct ftrace_probe_ops *probe_ops)
4121
{
4122
	struct ftrace_ops_hash old_hash_ops;
4123
	struct ftrace_func_entry *entry;
4124
	struct ftrace_func_probe *probe;
4125
	struct ftrace_glob func_g;
4126 4127 4128
	struct ftrace_hash **orig_hash;
	struct ftrace_hash *old_hash;
	struct ftrace_hash *hash = NULL;
4129
	struct hlist_node *tmp;
4130
	struct hlist_head hhd;
4131
	char str[KSYM_SYMBOL_LEN];
4132 4133
	int count = 0;
	int i, ret = -ENODEV;
4134
	int size;
4135

4136
	if (glob && (strcmp(glob, "*") == 0 || !strlen(glob)))
4137
		func_g.search = NULL;
4138
	else if (glob) {
4139 4140
		int not;

4141 4142 4143 4144
		func_g.type = filter_parse_regex(glob, strlen(glob),
						 &func_g.search, &not);
		func_g.len = strlen(func_g.search);
		func_g.search = glob;
4145

S
Steven Rostedt 已提交
4146
		/* we do not support '!' for function probes */
4147
		if (WARN_ON(not))
4148
			return -EINVAL;
4149 4150
	}

4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162 4163 4164 4165 4166 4167 4168
	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;
	}
	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);

	mutex_lock(&probe->ops.func_hash->regex_lock);
4169

4170
	orig_hash = &probe->ops.func_hash->filter_hash;
4171 4172 4173 4174
	old_hash = *orig_hash;

	if (ftrace_hash_empty(old_hash))
		goto out_unlock;
4175

4176 4177 4178 4179
	old_hash_ops.filter_hash = old_hash;
	/* Probes only have filters */
	old_hash_ops.notrace_hash = NULL;

4180
	ret = -ENOMEM;
4181
	hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, old_hash);
4182 4183 4184
	if (!hash)
		goto out_unlock;

4185
	INIT_HLIST_HEAD(&hhd);
4186

4187 4188 4189
	size = 1 << hash->size_bits;
	for (i = 0; i < size; i++) {
		hlist_for_each_entry_safe(entry, tmp, &hash->buckets[i], hlist) {
4190

4191
			if (func_g.search) {
4192 4193
				kallsyms_lookup(entry->ip, NULL, NULL,
						NULL, str);
4194
				if (!ftrace_match(str, &func_g))
4195 4196
					continue;
			}
4197
			count++;
4198 4199
			remove_hash_entry(hash, entry);
			hlist_add_head(&entry->hlist, &hhd);
4200 4201
		}
	}
4202 4203

	/* Nothing found? */
4204
	if (!count) {
4205 4206 4207 4208
		ret = -EINVAL;
		goto out_unlock;
	}

4209
	mutex_lock(&ftrace_lock);
4210

4211
	WARN_ON(probe->ref < count);
4212

4213
	probe->ref -= count;
4214

4215 4216 4217 4218
	if (ftrace_hash_empty(hash))
		ftrace_shutdown(&probe->ops, 0);

	ret = ftrace_hash_move_and_update_ops(&probe->ops, orig_hash,
4219
					      hash, 1);
4220 4221

	/* still need to update the function call sites */
4222
	if (ftrace_enabled && !ftrace_hash_empty(hash))
4223
		ftrace_run_modify_code(&probe->ops, FTRACE_UPDATE_CALLS,
4224
				       &old_hash_ops);
4225
	synchronize_sched();
4226

4227 4228
	hlist_for_each_entry_safe(entry, tmp, &hhd, hlist) {
		hlist_del(&entry->hlist);
4229 4230
		if (probe_ops->free)
			probe_ops->free(probe_ops, tr, entry->ip, NULL);
4231
		kfree(entry);
4232
	}
4233
	mutex_unlock(&ftrace_lock);
4234

4235
 out_unlock:
4236
	mutex_unlock(&probe->ops.func_hash->regex_lock);
4237
	free_ftrace_hash(hash);
4238 4239 4240 4241 4242 4243 4244

	release_probe(probe);

	return ret;

 err_unlock_ftrace:
	mutex_unlock(&ftrace_lock);
4245
	return ret;
4246 4247
}

4248 4249 4250
static LIST_HEAD(ftrace_commands);
static DEFINE_MUTEX(ftrace_cmd_mutex);

4251 4252 4253 4254 4255
/*
 * Currently we only register ftrace commands from __init, so mark this
 * __init too.
 */
__init int register_ftrace_command(struct ftrace_func_command *cmd)
4256 4257 4258 4259 4260 4261 4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273
{
	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;
}

4274 4275 4276 4277 4278
/*
 * Currently we only unregister ftrace commands from __init, so mark
 * this __init too.
 */
__init int unregister_ftrace_command(struct ftrace_func_command *cmd)
4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295 4296
{
	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;
}

4297
static int ftrace_process_regex(struct ftrace_iterator *iter,
4298
				char *buff, int len, int enable)
4299
{
4300 4301
	struct ftrace_hash *hash = iter->hash;
	struct trace_array *tr = global_ops.private;
4302
	char *func, *command, *next = buff;
S
Steven Rostedt 已提交
4303
	struct ftrace_func_command *p;
4304
	int ret = -EINVAL;
4305 4306 4307 4308

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

	if (!next) {
4309
		ret = ftrace_match_records(hash, func, len);
4310 4311 4312 4313 4314
		if (!ret)
			ret = -EINVAL;
		if (ret < 0)
			return ret;
		return 0;
4315 4316
	}

4317
	/* command found */
4318 4319 4320

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

4321 4322 4323
	if (WARN_ON_ONCE(!tr))
		return -EINVAL;

4324 4325 4326
	mutex_lock(&ftrace_cmd_mutex);
	list_for_each_entry(p, &ftrace_commands, list) {
		if (strcmp(p->name, command) == 0) {
4327
			ret = p->func(tr, hash, func, command, next, enable);
4328 4329
			goto out_unlock;
		}
4330
	}
4331 4332
 out_unlock:
	mutex_unlock(&ftrace_cmd_mutex);
4333

4334
	return ret;
4335 4336
}

I
Ingo Molnar 已提交
4337
static ssize_t
4338 4339
ftrace_regex_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos, int enable)
4340 4341
{
	struct ftrace_iterator *iter;
4342 4343
	struct trace_parser *parser;
	ssize_t ret, read;
4344

4345
	if (!cnt)
4346 4347 4348 4349 4350 4351 4352 4353
		return 0;

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

4354
	if (unlikely(ftrace_disabled))
4355 4356 4357
		return -ENODEV;

	/* iter->hash is a local copy, so we don't need regex_lock */
4358

4359 4360
	parser = &iter->parser;
	read = trace_get_user(parser, ubuf, cnt, ppos);
4361

4362
	if (read >= 0 && trace_parser_loaded(parser) &&
4363
	    !trace_parser_cont(parser)) {
4364
		ret = ftrace_process_regex(iter, parser->buffer,
4365
					   parser->idx, enable);
4366
		trace_parser_clear(parser);
4367
		if (ret < 0)
4368
			goto out;
4369
	}
4370 4371

	ret = read;
4372
 out:
4373 4374 4375
	return ret;
}

4376
ssize_t
4377 4378 4379 4380 4381 4382
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);
}

4383
ssize_t
4384 4385 4386 4387 4388 4389
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);
}

4390
static int
4391 4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408 4409 4410 4411
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)
4412
{
4413
	struct ftrace_hash **orig_hash;
4414
	struct ftrace_hash *hash;
4415
	int ret;
4416

4417
	if (unlikely(ftrace_disabled))
4418
		return -ENODEV;
4419

4420
	mutex_lock(&ops->func_hash->regex_lock);
4421

4422
	if (enable)
4423
		orig_hash = &ops->func_hash->filter_hash;
4424
	else
4425
		orig_hash = &ops->func_hash->notrace_hash;
4426

4427 4428 4429 4430 4431
	if (reset)
		hash = alloc_ftrace_hash(FTRACE_HASH_DEFAULT_BITS);
	else
		hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, *orig_hash);

4432 4433 4434 4435
	if (!hash) {
		ret = -ENOMEM;
		goto out_regex_unlock;
	}
4436

4437 4438 4439 4440
	if (buf && !ftrace_match_records(hash, buf, len)) {
		ret = -EINVAL;
		goto out_regex_unlock;
	}
4441 4442 4443 4444 4445
	if (ip) {
		ret = ftrace_match_addr(hash, ip, remove);
		if (ret < 0)
			goto out_regex_unlock;
	}
4446 4447

	mutex_lock(&ftrace_lock);
4448
	ret = ftrace_hash_move_and_update_ops(ops, orig_hash, hash, enable);
4449 4450
	mutex_unlock(&ftrace_lock);

4451
 out_regex_unlock:
4452
	mutex_unlock(&ops->func_hash->regex_lock);
4453 4454 4455

	free_ftrace_hash(hash);
	return ret;
4456 4457
}

4458 4459 4460 4461 4462 4463 4464 4465 4466 4467 4468 4469 4470 4471 4472 4473 4474 4475 4476 4477
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)
{
4478
	ftrace_ops_init(ops);
4479 4480 4481 4482
	return ftrace_set_addr(ops, ip, remove, reset, 1);
}
EXPORT_SYMBOL_GPL(ftrace_set_filter_ip);

4483 4484 4485 4486 4487 4488 4489 4490 4491 4492 4493 4494 4495 4496 4497 4498 4499
/**
 * 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);

4500 4501 4502 4503 4504 4505 4506
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);
}

4507 4508
/**
 * ftrace_set_filter - set a function to filter on in ftrace
4509 4510 4511 4512 4513 4514 4515 4516
 * @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.
 */
4517
int ftrace_set_filter(struct ftrace_ops *ops, unsigned char *buf,
4518 4519
		       int len, int reset)
{
4520
	ftrace_ops_init(ops);
4521
	return ftrace_set_regex(ops, buf, len, reset, 1);
4522 4523 4524 4525 4526 4527 4528 4529 4530 4531 4532 4533 4534 4535
}
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.
 */
4536
int ftrace_set_notrace(struct ftrace_ops *ops, unsigned char *buf,
4537 4538
			int len, int reset)
{
4539
	ftrace_ops_init(ops);
4540
	return ftrace_set_regex(ops, buf, len, reset, 0);
4541 4542 4543
}
EXPORT_SYMBOL_GPL(ftrace_set_notrace);
/**
4544
 * ftrace_set_global_filter - set a function to filter on with global tracers
4545 4546 4547 4548 4549 4550 4551
 * @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.
 */
4552
void ftrace_set_global_filter(unsigned char *buf, int len, int reset)
4553
{
4554
	ftrace_set_regex(&global_ops, buf, len, reset, 1);
4555
}
4556
EXPORT_SYMBOL_GPL(ftrace_set_global_filter);
4557

4558
/**
4559
 * ftrace_set_global_notrace - set a function to not trace with global tracers
4560 4561 4562 4563 4564 4565 4566 4567
 * @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.
 */
4568
void ftrace_set_global_notrace(unsigned char *buf, int len, int reset)
4569
{
4570
	ftrace_set_regex(&global_ops, buf, len, reset, 0);
4571
}
4572
EXPORT_SYMBOL_GPL(ftrace_set_global_notrace);
4573

4574 4575 4576 4577 4578 4579 4580
/*
 * 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;

4581 4582 4583
/* Used by function selftest to not test if filter is set */
bool ftrace_filter_param __initdata;

4584 4585
static int __init set_ftrace_notrace(char *str)
{
4586
	ftrace_filter_param = true;
4587
	strlcpy(ftrace_notrace_buf, str, FTRACE_FILTER_SIZE);
4588 4589 4590 4591 4592 4593
	return 1;
}
__setup("ftrace_notrace=", set_ftrace_notrace);

static int __init set_ftrace_filter(char *str)
{
4594
	ftrace_filter_param = true;
4595
	strlcpy(ftrace_filter_buf, str, FTRACE_FILTER_SIZE);
4596 4597 4598 4599
	return 1;
}
__setup("ftrace_filter=", set_ftrace_filter);

4600
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
4601
static char ftrace_graph_buf[FTRACE_FILTER_SIZE] __initdata;
4602
static char ftrace_graph_notrace_buf[FTRACE_FILTER_SIZE] __initdata;
4603
static int ftrace_graph_set_hash(struct ftrace_hash *hash, char *buffer);
4604

4605 4606 4607
static unsigned long save_global_trampoline;
static unsigned long save_global_flags;

4608 4609
static int __init set_graph_function(char *str)
{
4610
	strlcpy(ftrace_graph_buf, str, FTRACE_FILTER_SIZE);
4611 4612 4613 4614
	return 1;
}
__setup("ftrace_graph_filter=", set_graph_function);

4615 4616 4617 4618 4619 4620 4621
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);

4622 4623 4624 4625 4626 4627 4628 4629
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);
4630 4631

static void __init set_ftrace_early_graph(char *buf, int enable)
4632 4633 4634
{
	int ret;
	char *func;
4635
	struct ftrace_hash *hash;
4636

4637 4638 4639
	hash = alloc_ftrace_hash(FTRACE_HASH_DEFAULT_BITS);
	if (WARN_ON(!hash))
		return;
4640 4641 4642 4643

	while (buf) {
		func = strsep(&buf, ",");
		/* we allow only one expression at a time */
4644
		ret = ftrace_graph_set_hash(hash, func);
4645 4646 4647 4648
		if (ret)
			printk(KERN_DEBUG "ftrace: function %s not "
					  "traceable\n", func);
	}
4649 4650 4651 4652 4653

	if (enable)
		ftrace_graph_hash = hash;
	else
		ftrace_graph_notrace_hash = hash;
4654 4655 4656
}
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

4657 4658
void __init
ftrace_set_early_filter(struct ftrace_ops *ops, char *buf, int enable)
4659 4660 4661
{
	char *func;

4662 4663
	ftrace_ops_init(ops);

4664 4665
	while (buf) {
		func = strsep(&buf, ",");
4666
		ftrace_set_regex(ops, func, strlen(func), 0, enable);
4667 4668 4669 4670 4671 4672
	}
}

static void __init set_ftrace_early_filters(void)
{
	if (ftrace_filter_buf[0])
4673
		ftrace_set_early_filter(&global_ops, ftrace_filter_buf, 1);
4674
	if (ftrace_notrace_buf[0])
4675
		ftrace_set_early_filter(&global_ops, ftrace_notrace_buf, 0);
4676 4677
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	if (ftrace_graph_buf[0])
4678 4679 4680
		set_ftrace_early_graph(ftrace_graph_buf, 1);
	if (ftrace_graph_notrace_buf[0])
		set_ftrace_early_graph(ftrace_graph_notrace_buf, 0);
4681
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */
4682 4683
}

4684
int ftrace_regex_release(struct inode *inode, struct file *file)
4685 4686 4687
{
	struct seq_file *m = (struct seq_file *)file->private_data;
	struct ftrace_iterator *iter;
4688
	struct ftrace_hash **orig_hash;
4689
	struct trace_parser *parser;
4690
	int filter_hash;
4691
	int ret;
4692 4693 4694 4695 4696 4697 4698

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

4699 4700 4701
	parser = &iter->parser;
	if (trace_parser_loaded(parser)) {
		parser->buffer[parser->idx] = 0;
4702
		ftrace_match_records(iter->hash, parser->buffer, parser->idx);
4703 4704
	}

4705 4706
	trace_parser_put(parser);

4707
	mutex_lock(&iter->ops->func_hash->regex_lock);
4708

4709
	if (file->f_mode & FMODE_WRITE) {
4710 4711 4712
		filter_hash = !!(iter->flags & FTRACE_ITER_FILTER);

		if (filter_hash)
4713
			orig_hash = &iter->ops->func_hash->filter_hash;
4714
		else
4715
			orig_hash = &iter->ops->func_hash->notrace_hash;
4716

4717
		mutex_lock(&ftrace_lock);
4718 4719
		ret = ftrace_hash_move_and_update_ops(iter->ops, orig_hash,
						      iter->hash, filter_hash);
4720
		mutex_unlock(&ftrace_lock);
4721 4722 4723
	} else {
		/* For read only, the hash is the ops hash */
		iter->hash = NULL;
4724
	}
4725

4726
	mutex_unlock(&iter->ops->func_hash->regex_lock);
4727 4728
	free_ftrace_hash(iter->hash);
	kfree(iter);
4729

4730 4731 4732
	return 0;
}

4733
static const struct file_operations ftrace_avail_fops = {
4734 4735 4736
	.open = ftrace_avail_open,
	.read = seq_read,
	.llseek = seq_lseek,
L
Li Zefan 已提交
4737
	.release = seq_release_private,
4738 4739
};

4740 4741 4742 4743 4744 4745 4746
static const struct file_operations ftrace_enabled_fops = {
	.open = ftrace_enabled_open,
	.read = seq_read,
	.llseek = seq_lseek,
	.release = seq_release_private,
};

4747
static const struct file_operations ftrace_filter_fops = {
4748
	.open = ftrace_filter_open,
L
Lai Jiangshan 已提交
4749
	.read = seq_read,
4750
	.write = ftrace_filter_write,
4751
	.llseek = tracing_lseek,
4752
	.release = ftrace_regex_release,
4753 4754
};

4755
static const struct file_operations ftrace_notrace_fops = {
4756
	.open = ftrace_notrace_open,
L
Lai Jiangshan 已提交
4757
	.read = seq_read,
4758
	.write = ftrace_notrace_write,
4759
	.llseek = tracing_lseek,
4760
	.release = ftrace_regex_release,
4761 4762
};

4763 4764 4765 4766
#ifdef CONFIG_FUNCTION_GRAPH_TRACER

static DEFINE_MUTEX(graph_lock);

4767 4768 4769 4770 4771 4772 4773
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,
};
4774

4775 4776
#define FTRACE_GRAPH_EMPTY	((void *)1)

4777
struct ftrace_graph_data {
4778 4779 4780 4781 4782 4783 4784
	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;
4785 4786
};

4787
static void *
4788
__g_next(struct seq_file *m, loff_t *pos)
4789
{
4790
	struct ftrace_graph_data *fgd = m->private;
4791 4792 4793
	struct ftrace_func_entry *entry = fgd->entry;
	struct hlist_head *head;
	int i, idx = fgd->idx;
4794

4795
	if (*pos >= fgd->hash->count)
4796
		return NULL;
4797 4798 4799 4800 4801 4802 4803 4804 4805 4806 4807 4808 4809 4810 4811 4812 4813 4814 4815

	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;
4816
}
4817

4818 4819 4820 4821 4822
static void *
g_next(struct seq_file *m, void *v, loff_t *pos)
{
	(*pos)++;
	return __g_next(m, pos);
4823 4824 4825 4826
}

static void *g_start(struct seq_file *m, loff_t *pos)
{
4827 4828
	struct ftrace_graph_data *fgd = m->private;

4829 4830
	mutex_lock(&graph_lock);

4831 4832 4833 4834 4835 4836 4837
	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));

4838
	/* Nothing, tell g_show to print all functions are enabled */
4839
	if (ftrace_hash_empty(fgd->hash) && !*pos)
4840
		return FTRACE_GRAPH_EMPTY;
4841

4842 4843
	fgd->idx = 0;
	fgd->entry = NULL;
4844
	return __g_next(m, pos);
4845 4846 4847 4848 4849 4850 4851 4852 4853
}

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

static int g_show(struct seq_file *m, void *v)
{
4854
	struct ftrace_func_entry *entry = v;
4855

4856
	if (!entry)
4857 4858
		return 0;

4859
	if (entry == FTRACE_GRAPH_EMPTY) {
4860 4861
		struct ftrace_graph_data *fgd = m->private;

4862
		if (fgd->type == GRAPH_FILTER_FUNCTION)
4863
			seq_puts(m, "#### all functions enabled ####\n");
4864
		else
4865
			seq_puts(m, "#### no functions disabled ####\n");
4866 4867 4868
		return 0;
	}

4869
	seq_printf(m, "%ps\n", (void *)entry->ip);
4870 4871 4872 4873

	return 0;
}

J
James Morris 已提交
4874
static const struct seq_operations ftrace_graph_seq_ops = {
4875 4876 4877 4878 4879 4880 4881
	.start = g_start,
	.next = g_next,
	.stop = g_stop,
	.show = g_show,
};

static int
4882 4883
__ftrace_graph_open(struct inode *inode, struct file *file,
		    struct ftrace_graph_data *fgd)
4884 4885
{
	int ret = 0;
4886
	struct ftrace_hash *new_hash = NULL;
4887

4888 4889 4890
	if (file->f_mode & FMODE_WRITE) {
		const int size_bits = FTRACE_HASH_DEFAULT_BITS;

4891 4892 4893
		if (trace_parser_get_init(&fgd->parser, FTRACE_BUFF_MAX))
			return -ENOMEM;

4894 4895 4896 4897 4898 4899 4900 4901 4902
		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;
		}
4903 4904
	}

4905
	if (file->f_mode & FMODE_READ) {
4906
		ret = seq_open(file, &ftrace_graph_seq_ops);
4907 4908 4909
		if (!ret) {
			struct seq_file *m = file->private_data;
			m->private = fgd;
4910 4911 4912 4913
		} else {
			/* Failed */
			free_ftrace_hash(new_hash);
			new_hash = NULL;
4914 4915 4916
		}
	} else
		file->private_data = fgd;
4917

4918
out:
4919 4920 4921
	if (ret < 0 && file->f_mode & FMODE_WRITE)
		trace_parser_put(&fgd->parser);

4922
	fgd->new_hash = new_hash;
4923 4924 4925 4926 4927 4928 4929 4930

	/*
	 * 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;

4931 4932 4933
	return ret;
}

4934 4935 4936 4937
static int
ftrace_graph_open(struct inode *inode, struct file *file)
{
	struct ftrace_graph_data *fgd;
4938
	int ret;
4939 4940 4941 4942 4943 4944 4945 4946

	if (unlikely(ftrace_disabled))
		return -ENODEV;

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

4947 4948
	mutex_lock(&graph_lock);

4949 4950
	fgd->hash = rcu_dereference_protected(ftrace_graph_hash,
					lockdep_is_held(&graph_lock));
4951
	fgd->type = GRAPH_FILTER_FUNCTION;
4952 4953
	fgd->seq_ops = &ftrace_graph_seq_ops;

4954 4955 4956 4957 4958 4959
	ret = __ftrace_graph_open(inode, file, fgd);
	if (ret < 0)
		kfree(fgd);

	mutex_unlock(&graph_lock);
	return ret;
4960 4961
}

4962 4963 4964 4965
static int
ftrace_graph_notrace_open(struct inode *inode, struct file *file)
{
	struct ftrace_graph_data *fgd;
4966
	int ret;
4967 4968 4969 4970 4971 4972 4973 4974

	if (unlikely(ftrace_disabled))
		return -ENODEV;

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

4975 4976
	mutex_lock(&graph_lock);

4977 4978
	fgd->hash = rcu_dereference_protected(ftrace_graph_notrace_hash,
					lockdep_is_held(&graph_lock));
4979
	fgd->type = GRAPH_FILTER_NOTRACE;
4980 4981
	fgd->seq_ops = &ftrace_graph_seq_ops;

4982 4983 4984 4985 4986 4987
	ret = __ftrace_graph_open(inode, file, fgd);
	if (ret < 0)
		kfree(fgd);

	mutex_unlock(&graph_lock);
	return ret;
4988 4989
}

4990 4991 4992
static int
ftrace_graph_release(struct inode *inode, struct file *file)
{
4993
	struct ftrace_graph_data *fgd;
4994 4995 4996
	struct ftrace_hash *old_hash, *new_hash;
	struct trace_parser *parser;
	int ret = 0;
4997

4998 4999 5000
	if (file->f_mode & FMODE_READ) {
		struct seq_file *m = file->private_data;

5001
		fgd = m->private;
5002
		seq_release(inode, file);
5003
	} else {
5004
		fgd = file->private_data;
5005 5006
	}

5007 5008 5009 5010 5011 5012 5013 5014 5015 5016 5017 5018 5019 5020 5021 5022 5023 5024 5025 5026 5027 5028 5029 5030 5031 5032 5033 5034 5035 5036 5037 5038 5039 5040 5041 5042 5043 5044 5045 5046

	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:
5047 5048 5049
	kfree(fgd->new_hash);
	kfree(fgd);

5050
	return ret;
5051 5052
}

5053
static int
5054
ftrace_graph_set_hash(struct ftrace_hash *hash, char *buffer)
5055
{
5056
	struct ftrace_glob func_g;
5057 5058
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
5059
	struct ftrace_func_entry *entry;
5060
	int fail = 1;
5061
	int not;
5062

5063
	/* decode regex */
5064 5065
	func_g.type = filter_parse_regex(buffer, strlen(buffer),
					 &func_g.search, &not);
5066

5067
	func_g.len = strlen(func_g.search);
5068

5069
	mutex_lock(&ftrace_lock);
5070 5071 5072 5073 5074 5075

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

5076 5077
	do_for_each_ftrace_rec(pg, rec) {

5078 5079 5080
		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

D
Dmitry Safonov 已提交
5081
		if (ftrace_match_record(rec, &func_g, NULL, 0)) {
5082
			entry = ftrace_lookup_ip(hash, rec->ip);
5083 5084 5085

			if (!not) {
				fail = 0;
5086 5087 5088 5089 5090

				if (entry)
					continue;
				if (add_hash_entry(hash, rec->ip) < 0)
					goto out;
5091
			} else {
5092 5093
				if (entry) {
					free_hash_entry(hash, entry);
5094 5095 5096
					fail = 0;
				}
			}
5097
		}
5098
	} while_for_each_ftrace_rec();
5099
out:
5100
	mutex_unlock(&ftrace_lock);
5101

5102 5103 5104 5105
	if (fail)
		return -EINVAL;

	return 0;
5106 5107 5108 5109 5110 5111
}

static ssize_t
ftrace_graph_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
5112
	ssize_t read, ret = 0;
5113
	struct ftrace_graph_data *fgd = file->private_data;
5114
	struct trace_parser *parser;
5115

5116
	if (!cnt)
5117 5118
		return 0;

5119 5120 5121 5122 5123 5124
	/* Read mode uses seq functions */
	if (file->f_mode & FMODE_READ) {
		struct seq_file *m = file->private_data;
		fgd = m->private;
	}

5125
	parser = &fgd->parser;
5126

5127
	read = trace_get_user(parser, ubuf, cnt, ppos);
5128

5129 5130
	if (read >= 0 && trace_parser_loaded(parser) &&
	    !trace_parser_cont(parser)) {
5131

5132
		ret = ftrace_graph_set_hash(fgd->new_hash,
5133 5134
					    parser->buffer);
		trace_parser_clear(parser);
5135 5136
	}

5137 5138
	if (!ret)
		ret = read;
5139

5140 5141 5142 5143
	return ret;
}

static const struct file_operations ftrace_graph_fops = {
5144 5145 5146
	.open		= ftrace_graph_open,
	.read		= seq_read,
	.write		= ftrace_graph_write,
5147
	.llseek		= tracing_lseek,
5148
	.release	= ftrace_graph_release,
5149
};
5150 5151 5152 5153 5154

static const struct file_operations ftrace_graph_notrace_fops = {
	.open		= ftrace_graph_notrace_open,
	.read		= seq_read,
	.write		= ftrace_graph_write,
5155
	.llseek		= tracing_lseek,
5156 5157
	.release	= ftrace_graph_release,
};
5158 5159
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

5160 5161 5162 5163 5164 5165 5166 5167 5168 5169 5170 5171 5172 5173 5174 5175 5176 5177 5178 5179 5180 5181 5182 5183 5184 5185 5186 5187 5188 5189
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);
}

5190
static __init int ftrace_init_dyn_tracefs(struct dentry *d_tracer)
5191 5192
{

5193 5194
	trace_create_file("available_filter_functions", 0444,
			d_tracer, NULL, &ftrace_avail_fops);
5195

5196 5197 5198
	trace_create_file("enabled_functions", 0444,
			d_tracer, NULL, &ftrace_enabled_fops);

5199
	ftrace_create_filter_files(&global_ops, d_tracer);
5200

5201
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
5202
	trace_create_file("set_graph_function", 0444, d_tracer,
5203 5204
				    NULL,
				    &ftrace_graph_fops);
5205 5206 5207
	trace_create_file("set_graph_notrace", 0444, d_tracer,
				    NULL,
				    &ftrace_graph_notrace_fops);
5208 5209
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

5210 5211 5212
	return 0;
}

5213
static int ftrace_cmp_ips(const void *a, const void *b)
5214
{
5215 5216
	const unsigned long *ipa = a;
	const unsigned long *ipb = b;
5217

5218 5219 5220 5221 5222 5223 5224
	if (*ipa > *ipb)
		return 1;
	if (*ipa < *ipb)
		return -1;
	return 0;
}

5225
static int ftrace_process_locs(struct module *mod,
5226
			       unsigned long *start,
5227 5228
			       unsigned long *end)
{
5229
	struct ftrace_page *start_pg;
5230
	struct ftrace_page *pg;
5231
	struct dyn_ftrace *rec;
5232
	unsigned long count;
5233 5234
	unsigned long *p;
	unsigned long addr;
5235
	unsigned long flags = 0; /* Shut up gcc */
5236 5237 5238 5239 5240 5241 5242
	int ret = -ENOMEM;

	count = end - start;

	if (!count)
		return 0;

5243
	sort(start, count, sizeof(*start),
5244
	     ftrace_cmp_ips, NULL);
5245

5246 5247
	start_pg = ftrace_allocate_pages(count);
	if (!start_pg)
5248
		return -ENOMEM;
5249

S
Steven Rostedt 已提交
5250
	mutex_lock(&ftrace_lock);
5251

5252 5253 5254 5255 5256
	/*
	 * 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.
	 */
5257 5258 5259
	if (!mod) {
		WARN_ON(ftrace_pages || ftrace_pages_start);
		/* First initialization */
5260
		ftrace_pages = ftrace_pages_start = start_pg;
5261
	} else {
5262
		if (!ftrace_pages)
5263
			goto out;
5264

5265 5266 5267 5268
		if (WARN_ON(ftrace_pages->next)) {
			/* Hmm, we have free pages? */
			while (ftrace_pages->next)
				ftrace_pages = ftrace_pages->next;
5269
		}
5270

5271
		ftrace_pages->next = start_pg;
5272 5273
	}

5274
	p = start;
5275
	pg = start_pg;
5276 5277
	while (p < end) {
		addr = ftrace_call_adjust(*p++);
5278 5279 5280 5281 5282 5283 5284 5285
		/*
		 * 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;
5286 5287 5288 5289 5290 5291 5292 5293 5294 5295

		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;
5296 5297
	}

5298 5299 5300 5301 5302 5303
	/* We should have used all pages */
	WARN_ON(pg->next);

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

5304
	/*
5305 5306 5307 5308 5309 5310
	 * 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.
5311
	 */
5312 5313
	if (!mod)
		local_irq_save(flags);
5314
	ftrace_update_code(mod, start_pg);
5315 5316
	if (!mod)
		local_irq_restore(flags);
5317 5318
	ret = 0;
 out:
S
Steven Rostedt 已提交
5319
	mutex_unlock(&ftrace_lock);
5320

5321
	return ret;
5322 5323
}

5324
#ifdef CONFIG_MODULES
5325 5326 5327

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

5328 5329 5330 5331 5332 5333 5334 5335 5336 5337 5338 5339 5340
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;
}

5341
void ftrace_release_mod(struct module *mod)
5342 5343
{
	struct dyn_ftrace *rec;
5344
	struct ftrace_page **last_pg;
5345
	struct ftrace_page *pg;
5346
	int order;
5347

5348 5349
	mutex_lock(&ftrace_lock);

5350
	if (ftrace_disabled)
5351
		goto out_unlock;
5352

5353 5354 5355 5356 5357 5358 5359
	/*
	 * 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];
5360
		if (within_module_core(rec->ip, mod)) {
5361
			/*
5362 5363
			 * As core pages are first, the first
			 * page should never be a module page.
5364
			 */
5365 5366 5367 5368 5369 5370 5371 5372
			if (WARN_ON(pg == ftrace_pages_start))
				goto out_unlock;

			/* Check if we are deleting the last page */
			if (pg == ftrace_pages)
				ftrace_pages = next_to_ftrace_page(last_pg);

			*last_pg = pg->next;
5373 5374 5375
			order = get_count_order(pg->size / ENTRIES_PER_PAGE);
			free_pages((unsigned long)pg->records, order);
			kfree(pg);
5376 5377 5378
		} else
			last_pg = &pg->next;
	}
5379
 out_unlock:
5380 5381 5382
	mutex_unlock(&ftrace_lock);
}

5383
void ftrace_module_enable(struct module *mod)
5384 5385 5386 5387 5388 5389 5390 5391 5392 5393 5394 5395 5396 5397 5398 5399 5400 5401 5402 5403 5404 5405 5406 5407 5408 5409 5410 5411 5412 5413 5414 5415 5416 5417 5418 5419 5420 5421 5422 5423 5424 5425 5426 5427 5428 5429 5430 5431 5432 5433 5434 5435 5436 5437 5438 5439 5440 5441 5442 5443 5444 5445 5446 5447 5448 5449 5450 5451
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;

	mutex_lock(&ftrace_lock);

	if (ftrace_disabled)
		goto out_unlock;

	/*
	 * If the tracing is enabled, go ahead and enable the record.
	 *
	 * The reason not to enable the record immediatelly is the
	 * inherent check of ftrace_make_nop/ftrace_make_call for
	 * correct previous instructions.  Making first the NOP
	 * conversion puts the module to the correct state, thus
	 * passing the ftrace_make_call check.
	 *
	 * We also delay this to after the module code already set the
	 * text to read-only, as we now need to set it back to read-write
	 * so that we can modify the text.
	 */
	if (ftrace_start_up)
		ftrace_arch_code_modify_prepare();

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

		cnt = 0;

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

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

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

	} while_for_each_ftrace_rec();

 out_loop:
	if (ftrace_start_up)
		ftrace_arch_code_modify_post_process();

 out_unlock:
	mutex_unlock(&ftrace_lock);
}

5452
void ftrace_module_init(struct module *mod)
5453
{
5454
	if (ftrace_disabled || !mod->num_ftrace_callsites)
5455
		return;
5456

5457 5458
	ftrace_process_locs(mod, mod->ftrace_callsites,
			    mod->ftrace_callsites + mod->num_ftrace_callsites);
5459
}
5460 5461
#endif /* CONFIG_MODULES */

5462
void __init ftrace_free_init_mem(void)
5463
{
5464 5465
	unsigned long start = (unsigned long)(&__init_begin);
	unsigned long end = (unsigned long)(&__init_end);
5466 5467 5468 5469 5470 5471 5472 5473 5474 5475 5476 5477 5478 5479 5480 5481 5482 5483 5484 5485 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 5499 5500 5501 5502 5503 5504 5505
	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--;
		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);
}

5506 5507
void __init ftrace_init(void)
{
5508 5509
	extern unsigned long __start_mcount_loc[];
	extern unsigned long __stop_mcount_loc[];
5510
	unsigned long count, flags;
5511 5512 5513
	int ret;

	local_irq_save(flags);
5514
	ret = ftrace_dyn_arch_init();
5515
	local_irq_restore(flags);
5516
	if (ret)
5517 5518 5519
		goto failed;

	count = __stop_mcount_loc - __start_mcount_loc;
5520 5521
	if (!count) {
		pr_info("ftrace: No functions to be traced?\n");
5522
		goto failed;
5523 5524 5525 5526
	}

	pr_info("ftrace: allocating %ld entries in %ld pages\n",
		count, count / ENTRIES_PER_PAGE + 1);
5527 5528 5529

	last_ftrace_enabled = ftrace_enabled = 1;

5530
	ret = ftrace_process_locs(NULL,
5531
				  __start_mcount_loc,
5532 5533
				  __stop_mcount_loc);

5534 5535
	set_ftrace_early_filters();

5536 5537 5538 5539 5540
	return;
 failed:
	ftrace_disabled = 1;
}

5541 5542 5543 5544 5545 5546 5547 5548 5549 5550
/* 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);
}

5551 5552 5553 5554
void ftrace_init_trace_array(struct trace_array *tr)
{
	INIT_LIST_HEAD(&tr->func_probes);
}
5555
#else
5556

5557
static struct ftrace_ops global_ops = {
5558
	.func			= ftrace_stub,
5559 5560 5561
	.flags			= FTRACE_OPS_FL_RECURSION_SAFE |
				  FTRACE_OPS_FL_INITIALIZED |
				  FTRACE_OPS_FL_PID,
5562 5563
};

5564 5565 5566 5567 5568
static int __init ftrace_nodyn_init(void)
{
	ftrace_enabled = 1;
	return 0;
}
5569
core_initcall(ftrace_nodyn_init);
5570

5571
static inline int ftrace_init_dyn_tracefs(struct dentry *d_tracer) { return 0; }
5572
static inline void ftrace_startup_enable(int command) { }
5573
static inline void ftrace_startup_all(int command) { }
5574
/* Keep as macros so we do not need to define the commands */
5575 5576 5577 5578 5579 5580
# define ftrace_startup(ops, command)					\
	({								\
		int ___ret = __register_ftrace_function(ops);		\
		if (!___ret)						\
			(ops)->flags |= FTRACE_OPS_FL_ENABLED;		\
		___ret;							\
5581
	})
5582 5583 5584 5585 5586 5587 5588
# define ftrace_shutdown(ops, command)					\
	({								\
		int ___ret = __unregister_ftrace_function(ops);		\
		if (!___ret)						\
			(ops)->flags &= ~FTRACE_OPS_FL_ENABLED;		\
		___ret;							\
	})
5589

I
Ingo Molnar 已提交
5590 5591
# define ftrace_startup_sysctl()	do { } while (0)
# define ftrace_shutdown_sysctl()	do { } while (0)
5592 5593

static inline int
5594
ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip, void *regs)
5595 5596 5597 5598
{
	return 1;
}

5599 5600 5601 5602
static void ftrace_update_trampoline(struct ftrace_ops *ops)
{
}

5603 5604
#endif /* CONFIG_DYNAMIC_FTRACE */

5605 5606 5607 5608
__init void ftrace_init_global_array_ops(struct trace_array *tr)
{
	tr->ops = &global_ops;
	tr->ops->private = tr;
5609
	ftrace_init_trace_array(tr);
5610 5611 5612 5613 5614 5615 5616 5617 5618 5619 5620 5621 5622 5623 5624 5625 5626 5627 5628
}

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;
}

5629 5630
static inline void
__ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
5631
		       struct ftrace_ops *ignored, struct pt_regs *regs)
5632
{
5633
	struct ftrace_ops *op;
5634
	int bit;
5635

5636 5637 5638
	bit = trace_test_and_set_recursion(TRACE_LIST_START, TRACE_LIST_MAX);
	if (bit < 0)
		return;
5639

5640 5641 5642 5643 5644
	/*
	 * Some of the ops may be dynamically allocated,
	 * they must be freed after a synchronize_sched().
	 */
	preempt_disable_notrace();
5645

5646
	do_for_each_ftrace_op(op, ftrace_ops_list) {
5647 5648 5649 5650 5651 5652 5653 5654 5655 5656 5657 5658 5659 5660
		/*
		 * 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)) {
		    
5661 5662
			if (FTRACE_WARN_ON(!op->func)) {
				pr_warn("op=%p %pS\n", op, op);
5663 5664
				goto out;
			}
5665
			op->func(ip, parent_ip, op, regs);
5666
		}
5667
	} while_for_each_ftrace_op(op);
5668
out:
5669
	preempt_enable_notrace();
5670
	trace_clear_recursion(bit);
5671 5672
}

5673 5674 5675 5676 5677
/*
 * 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.
5678 5679 5680
 * 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
5681 5682
 * being NULL, or CONFIG_DYNAMIC_FTRACE_WITH_REGS.
 * Note, CONFIG_DYNAMIC_FTRACE_WITH_REGS expects a full regs to be saved.
5683
 * An architecture can pass partial regs with ftrace_ops and still
L
Li Bin 已提交
5684
 * set the ARCH_SUPPORTS_FTRACE_OPS.
5685 5686 5687
 */
#if ARCH_SUPPORTS_FTRACE_OPS
static void ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
5688
				 struct ftrace_ops *op, struct pt_regs *regs)
5689
{
5690
	__ftrace_ops_list_func(ip, parent_ip, NULL, regs);
5691 5692 5693 5694
}
#else
static void ftrace_ops_no_ops(unsigned long ip, unsigned long parent_ip)
{
5695
	__ftrace_ops_list_func(ip, parent_ip, NULL, NULL);
5696 5697 5698
}
#endif

5699 5700
/*
 * If there's only one function registered but it does not support
5701 5702
 * recursion, needs RCU protection and/or requires per cpu handling, then
 * this function will be called by the mcount trampoline.
5703
 */
5704
static void ftrace_ops_assist_func(unsigned long ip, unsigned long parent_ip,
5705 5706 5707 5708
				   struct ftrace_ops *op, struct pt_regs *regs)
{
	int bit;

5709 5710 5711
	if ((op->flags & FTRACE_OPS_FL_RCU) && !rcu_is_watching())
		return;

5712 5713 5714 5715
	bit = trace_test_and_set_recursion(TRACE_LIST_START, TRACE_LIST_MAX);
	if (bit < 0)
		return;

5716
	preempt_disable_notrace();
5717

5718 5719 5720 5721 5722 5723
	if (!(op->flags & FTRACE_OPS_FL_PER_CPU) ||
	    !ftrace_function_local_disabled(op)) {
		op->func(ip, parent_ip, op, regs);
	}

	preempt_enable_notrace();
5724 5725 5726
	trace_clear_recursion(bit);
}

5727 5728 5729 5730 5731 5732 5733
/**
 * 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
5734
 * ftrace_ops_assist_func() instead.
5735 5736 5737 5738 5739 5740
 *
 * Returns the function that the trampoline should call for @ops.
 */
ftrace_func_t ftrace_ops_get_func(struct ftrace_ops *ops)
{
	/*
5741 5742
	 * 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.
5743
	 */
5744 5745 5746
	if (!(ops->flags & FTRACE_OPS_FL_RECURSION_SAFE) ||
	    ops->flags & (FTRACE_OPS_FL_RCU | FTRACE_OPS_FL_PER_CPU))
		return ftrace_ops_assist_func;
5747 5748 5749 5750

	return ops->func;
}

5751 5752 5753
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
5754
{
5755 5756
	struct trace_array *tr = data;
	struct trace_pid_list *pid_list;
S
Steven Rostedt 已提交
5757

5758
	pid_list = rcu_dereference_sched(tr->function_pids);
5759

5760 5761
	this_cpu_write(tr->trace_buffer.data->ftrace_ignore_pid,
		       trace_ignore_this_task(pid_list, next));
S
Steven Rostedt 已提交
5762 5763
}

5764 5765 5766 5767 5768 5769 5770 5771 5772 5773 5774 5775 5776 5777 5778 5779 5780 5781 5782 5783 5784 5785 5786 5787 5788 5789 5790 5791 5792 5793 5794 5795 5796 5797 5798 5799 5800
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);
	}
}

5801
static void clear_ftrace_pids(struct trace_array *tr)
5802
{
5803 5804
	struct trace_pid_list *pid_list;
	int cpu;
5805

5806 5807 5808 5809
	pid_list = rcu_dereference_protected(tr->function_pids,
					     lockdep_is_held(&ftrace_lock));
	if (!pid_list)
		return;
5810

5811
	unregister_trace_sched_switch(ftrace_filter_pid_sched_switch_probe, tr);
5812

5813 5814
	for_each_possible_cpu(cpu)
		per_cpu_ptr(tr->trace_buffer.data, cpu)->ftrace_ignore_pid = false;
S
Steven Rostedt 已提交
5815

5816
	rcu_assign_pointer(tr->function_pids, NULL);
S
Steven Rostedt 已提交
5817

5818 5819
	/* Wait till all users are no longer using pid filtering */
	synchronize_sched();
5820

5821
	trace_free_pid_list(pid_list);
5822 5823
}

5824
static void ftrace_pid_reset(struct trace_array *tr)
5825
{
5826
	mutex_lock(&ftrace_lock);
5827
	clear_ftrace_pids(tr);
S
Steven Rostedt 已提交
5828

5829
	ftrace_update_pid_func();
5830
	ftrace_startup_all(0);
5831 5832 5833 5834

	mutex_unlock(&ftrace_lock);
}

5835 5836
/* Greater than any max PID */
#define FTRACE_NO_PIDS		(void *)(PID_MAX_LIMIT + 1)
5837

5838
static void *fpid_start(struct seq_file *m, loff_t *pos)
5839
	__acquires(RCU)
5840
{
5841 5842 5843
	struct trace_pid_list *pid_list;
	struct trace_array *tr = m->private;

5844
	mutex_lock(&ftrace_lock);
5845 5846 5847
	rcu_read_lock_sched();

	pid_list = rcu_dereference_sched(tr->function_pids);
5848

5849 5850
	if (!pid_list)
		return !(*pos) ? FTRACE_NO_PIDS : NULL;
5851

5852
	return trace_pid_start(pid_list, pos);
5853 5854 5855 5856
}

static void *fpid_next(struct seq_file *m, void *v, loff_t *pos)
{
5857 5858 5859 5860
	struct trace_array *tr = m->private;
	struct trace_pid_list *pid_list = rcu_dereference_sched(tr->function_pids);

	if (v == FTRACE_NO_PIDS)
5861 5862
		return NULL;

5863
	return trace_pid_next(pid_list, v, pos);
5864 5865 5866
}

static void fpid_stop(struct seq_file *m, void *p)
5867
	__releases(RCU)
5868
{
5869
	rcu_read_unlock_sched();
5870 5871 5872 5873 5874
	mutex_unlock(&ftrace_lock);
}

static int fpid_show(struct seq_file *m, void *v)
{
5875
	if (v == FTRACE_NO_PIDS) {
5876
		seq_puts(m, "no pid\n");
5877 5878 5879
		return 0;
	}

5880
	return trace_pid_show(m, v);
5881 5882 5883 5884 5885 5886 5887 5888 5889 5890 5891 5892
}

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)
{
5893 5894
	struct trace_array *tr = inode->i_private;
	struct seq_file *m;
5895 5896
	int ret = 0;

5897 5898 5899
	if (trace_array_get(tr) < 0)
		return -ENODEV;

5900 5901
	if ((file->f_mode & FMODE_WRITE) &&
	    (file->f_flags & O_TRUNC))
5902
		ftrace_pid_reset(tr);
5903

5904 5905 5906 5907 5908 5909 5910 5911
	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;
	}
5912 5913 5914 5915

	return ret;
}

5916 5917 5918 5919 5920 5921 5922 5923 5924 5925 5926 5927 5928 5929 5930 5931
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));
}

5932 5933 5934 5935
static ssize_t
ftrace_pid_write(struct file *filp, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
5936 5937 5938 5939 5940
	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;
5941

5942 5943 5944 5945 5946 5947 5948 5949 5950 5951 5952
	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;
5953

5954
	rcu_assign_pointer(tr->function_pids, pid_list);
5955

5956 5957 5958 5959 5960 5961 5962
	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);
	}
5963

5964
	/*
5965 5966 5967
	 * 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.
5968
	 */
5969
	on_each_cpu(ignore_task_cpu, tr, 1);
5970

5971 5972 5973 5974
	ftrace_update_pid_func();
	ftrace_startup_all(0);
 out:
	mutex_unlock(&ftrace_lock);
5975

5976 5977
	if (ret > 0)
		*ppos += ret;
5978

5979
	return ret;
5980
}
5981

5982 5983 5984
static int
ftrace_pid_release(struct inode *inode, struct file *file)
{
5985
	struct trace_array *tr = inode->i_private;
5986

5987 5988 5989
	trace_array_put(tr);

	return seq_release(inode, file);
5990 5991
}

5992
static const struct file_operations ftrace_pid_fops = {
5993 5994 5995
	.open		= ftrace_pid_open,
	.write		= ftrace_pid_write,
	.read		= seq_read,
5996
	.llseek		= tracing_lseek,
5997
	.release	= ftrace_pid_release,
5998 5999
};

6000
void ftrace_init_tracefs(struct trace_array *tr, struct dentry *d_tracer)
6001
{
6002
	trace_create_file("set_ftrace_pid", 0644, d_tracer,
6003
			    tr, &ftrace_pid_fops);
6004 6005
}

6006 6007 6008 6009 6010 6011 6012 6013 6014 6015
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
6016
/**
6017
 * ftrace_kill - kill ftrace
S
Steven Rostedt 已提交
6018 6019 6020 6021 6022
 *
 * 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.
 */
6023
void ftrace_kill(void)
S
Steven Rostedt 已提交
6024 6025 6026 6027 6028 6029
{
	ftrace_disabled = 1;
	ftrace_enabled = 0;
	clear_ftrace_function();
}

6030 6031 6032 6033 6034 6035 6036 6037
/**
 * Test if ftrace is dead or not.
 */
int ftrace_is_dead(void)
{
	return ftrace_disabled;
}

6038
/**
6039 6040
 * register_ftrace_function - register a function for profiling
 * @ops - ops structure that holds the function for profiling.
6041
 *
6042 6043 6044 6045 6046 6047
 * 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.
6048
 */
6049
int register_ftrace_function(struct ftrace_ops *ops)
6050
{
6051
	int ret = -1;
6052

6053 6054
	ftrace_ops_init(ops);

S
Steven Rostedt 已提交
6055
	mutex_lock(&ftrace_lock);
6056

6057
	ret = ftrace_startup(ops, 0);
6058

S
Steven Rostedt 已提交
6059
	mutex_unlock(&ftrace_lock);
6060

6061
	return ret;
6062
}
6063
EXPORT_SYMBOL_GPL(register_ftrace_function);
6064 6065

/**
6066
 * unregister_ftrace_function - unregister a function for profiling.
6067 6068 6069 6070 6071 6072 6073 6074
 * @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
6075
	mutex_lock(&ftrace_lock);
6076
	ret = ftrace_shutdown(ops, 0);
S
Steven Rostedt 已提交
6077
	mutex_unlock(&ftrace_lock);
6078 6079 6080

	return ret;
}
6081
EXPORT_SYMBOL_GPL(unregister_ftrace_function);
6082

I
Ingo Molnar 已提交
6083
int
6084
ftrace_enable_sysctl(struct ctl_table *table, int write,
6085
		     void __user *buffer, size_t *lenp,
6086 6087
		     loff_t *ppos)
{
6088
	int ret = -ENODEV;
6089

S
Steven Rostedt 已提交
6090
	mutex_lock(&ftrace_lock);
6091

6092 6093 6094 6095
	if (unlikely(ftrace_disabled))
		goto out;

	ret = proc_dointvec(table, write, buffer, lenp, ppos);
6096

6097
	if (ret || !write || (last_ftrace_enabled == !!ftrace_enabled))
6098 6099
		goto out;

6100
	last_ftrace_enabled = !!ftrace_enabled;
6101 6102 6103 6104

	if (ftrace_enabled) {

		/* we are starting ftrace again */
6105 6106
		if (ftrace_ops_list != &ftrace_list_end)
			update_ftrace_function();
6107

6108 6109
		ftrace_startup_sysctl();

6110 6111 6112 6113 6114 6115 6116 6117
	} else {
		/* stopping ftrace calls (just send to ftrace_stub) */
		ftrace_trace_function = ftrace_stub;

		ftrace_shutdown_sysctl();
	}

 out:
S
Steven Rostedt 已提交
6118
	mutex_unlock(&ftrace_lock);
6119
	return ret;
6120
}
6121

6122
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
6123

6124 6125 6126 6127
static struct ftrace_ops graph_ops = {
	.func			= ftrace_stub,
	.flags			= FTRACE_OPS_FL_RECURSION_SAFE |
				   FTRACE_OPS_FL_INITIALIZED |
6128
				   FTRACE_OPS_FL_PID |
6129 6130 6131
				   FTRACE_OPS_FL_STUB,
#ifdef FTRACE_GRAPH_TRAMP_ADDR
	.trampoline		= FTRACE_GRAPH_TRAMP_ADDR,
6132
	/* trampoline_size is only needed for dynamically allocated tramps */
6133 6134 6135 6136
#endif
	ASSIGN_OPS_HASH(graph_ops, &global_ops.local_hash)
};

6137 6138 6139 6140 6141 6142 6143 6144 6145 6146
void ftrace_graph_sleep_time_control(bool enable)
{
	fgraph_sleep_time = enable;
}

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

6147 6148 6149 6150 6151
int ftrace_graph_entry_stub(struct ftrace_graph_ent *trace)
{
	return 0;
}

6152 6153 6154
/* The callbacks that hook a function */
trace_func_graph_ret_t ftrace_graph_return =
			(trace_func_graph_ret_t)ftrace_stub;
6155
trace_func_graph_ent_t ftrace_graph_entry = ftrace_graph_entry_stub;
6156
static trace_func_graph_ent_t __ftrace_graph_entry = ftrace_graph_entry_stub;
6157 6158 6159 6160 6161 6162 6163 6164 6165 6166 6167 6168 6169 6170 6171 6172 6173 6174 6175 6176 6177

/* 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;
		}
	}

6178
	read_lock(&tasklist_lock);
6179 6180 6181 6182 6183 6184 6185
	do_each_thread(g, t) {
		if (start == end) {
			ret = -EAGAIN;
			goto unlock;
		}

		if (t->ret_stack == NULL) {
6186
			atomic_set(&t->tracing_graph_pause, 0);
6187
			atomic_set(&t->trace_overrun, 0);
6188 6189 6190 6191
			t->curr_ret_stack = -1;
			/* Make sure the tasks see the -1 first: */
			smp_wmb();
			t->ret_stack = ret_stack_list[start++];
6192 6193 6194 6195
		}
	} while_each_thread(g, t);

unlock:
6196
	read_unlock(&tasklist_lock);
6197 6198 6199 6200 6201 6202
free:
	for (i = start; i < end; i++)
		kfree(ret_stack_list[i]);
	return ret;
}

6203
static void
6204
ftrace_graph_probe_sched_switch(void *ignore, bool preempt,
6205
			struct task_struct *prev, struct task_struct *next)
6206 6207 6208 6209
{
	unsigned long long timestamp;
	int index;

6210 6211 6212 6213
	/*
	 * Does the user want to count the time a function was asleep.
	 * If so, do not update the time stamps.
	 */
6214
	if (fgraph_sleep_time)
6215 6216
		return;

6217 6218 6219 6220 6221 6222 6223 6224 6225 6226 6227 6228 6229 6230 6231 6232 6233 6234
	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;
}

6235
/* Allocate a return stack for each task */
6236
static int start_graph_tracing(void)
6237 6238
{
	struct ftrace_ret_stack **ret_stack_list;
6239
	int ret, cpu;
6240 6241 6242 6243 6244 6245 6246 6247

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

	if (!ret_stack_list)
		return -ENOMEM;

6248
	/* The cpu_boot init_task->ret_stack will never be freed */
6249 6250
	for_each_online_cpu(cpu) {
		if (!idle_task(cpu)->ret_stack)
6251
			ftrace_graph_init_idle_task(idle_task(cpu), cpu);
6252
	}
6253

6254 6255 6256 6257
	do {
		ret = alloc_retstack_tasklist(ret_stack_list);
	} while (ret == -EAGAIN);

6258
	if (!ret) {
6259
		ret = register_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
6260 6261 6262 6263 6264
		if (ret)
			pr_info("ftrace_graph: Couldn't activate tracepoint"
				" probe to kernel_sched_switch\n");
	}

6265 6266 6267 6268
	kfree(ret_stack_list);
	return ret;
}

6269 6270 6271 6272 6273 6274 6275 6276 6277 6278 6279 6280 6281 6282 6283 6284 6285 6286 6287 6288 6289
/*
 * 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;
}

6290 6291 6292 6293 6294 6295 6296 6297 6298 6299 6300 6301 6302 6303 6304 6305
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)
{
6306 6307 6308 6309 6310 6311 6312 6313 6314 6315 6316 6317 6318 6319 6320 6321 6322 6323 6324
	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)
6325
		ftrace_graph_entry = ftrace_graph_entry_test;
6326 6327
	else
		ftrace_graph_entry = __ftrace_graph_entry;
6328 6329
}

6330 6331 6332 6333
static struct notifier_block ftrace_suspend_notifier = {
	.notifier_call = ftrace_suspend_notifier_call,
};

6334 6335
int register_ftrace_graph(trace_func_graph_ret_t retfunc,
			trace_func_graph_ent_t entryfunc)
6336
{
6337 6338
	int ret = 0;

S
Steven Rostedt 已提交
6339
	mutex_lock(&ftrace_lock);
6340

6341
	/* we currently allow only one tracer registered at a time */
6342
	if (ftrace_graph_active) {
6343 6344 6345 6346
		ret = -EBUSY;
		goto out;
	}

6347 6348
	register_pm_notifier(&ftrace_suspend_notifier);

6349
	ftrace_graph_active++;
6350
	ret = start_graph_tracing();
6351
	if (ret) {
6352
		ftrace_graph_active--;
6353 6354
		goto out;
	}
6355

6356
	ftrace_graph_return = retfunc;
6357 6358 6359 6360 6361 6362 6363 6364 6365 6366

	/*
	 * 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();
6367

6368
	ret = ftrace_startup(&graph_ops, FTRACE_START_FUNC_RET);
6369
out:
S
Steven Rostedt 已提交
6370
	mutex_unlock(&ftrace_lock);
6371
	return ret;
6372 6373
}

6374
void unregister_ftrace_graph(void)
6375
{
S
Steven Rostedt 已提交
6376
	mutex_lock(&ftrace_lock);
6377

6378
	if (unlikely(!ftrace_graph_active))
6379 6380
		goto out;

6381
	ftrace_graph_active--;
6382
	ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub;
6383
	ftrace_graph_entry = ftrace_graph_entry_stub;
6384
	__ftrace_graph_entry = ftrace_graph_entry_stub;
6385
	ftrace_shutdown(&graph_ops, FTRACE_STOP_FUNC_RET);
6386
	unregister_pm_notifier(&ftrace_suspend_notifier);
6387
	unregister_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
6388

6389 6390 6391 6392 6393 6394 6395 6396 6397 6398 6399
#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

6400
 out:
S
Steven Rostedt 已提交
6401
	mutex_unlock(&ftrace_lock);
6402
}
6403

6404 6405 6406 6407 6408 6409 6410 6411
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
6412
	/* make curr_ret_stack visible before we add the ret_stack */
6413 6414 6415 6416 6417 6418 6419 6420 6421 6422 6423 6424 6425 6426 6427 6428 6429 6430 6431 6432 6433 6434 6435 6436 6437 6438 6439 6440 6441 6442 6443 6444 6445 6446
	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);
	}
}

6447
/* Allocate a return stack for newly created task */
6448
void ftrace_graph_init_task(struct task_struct *t)
6449
{
6450 6451
	/* Make sure we do not use the parent ret_stack */
	t->ret_stack = NULL;
6452
	t->curr_ret_stack = -1;
6453

6454
	if (ftrace_graph_active) {
6455 6456 6457
		struct ftrace_ret_stack *ret_stack;

		ret_stack = kmalloc(FTRACE_RETFUNC_DEPTH
6458 6459
				* sizeof(struct ftrace_ret_stack),
				GFP_KERNEL);
6460
		if (!ret_stack)
6461
			return;
6462
		graph_init_task(t, ret_stack);
6463
	}
6464 6465
}

6466
void ftrace_graph_exit_task(struct task_struct *t)
6467
{
6468 6469
	struct ftrace_ret_stack	*ret_stack = t->ret_stack;

6470
	t->ret_stack = NULL;
6471 6472 6473 6474
	/* NULL must become visible to IRQs before we free it: */
	barrier();

	kfree(ret_stack);
6475
}
6476
#endif