trace_syscalls.c 15.5 KB
Newer Older
1
#include <trace/syscall.h>
2
#include <trace/events/syscalls.h>
3
#include <linux/kernel.h>
4
#include <linux/ftrace.h>
5
#include <linux/perf_event.h>
6 7 8 9 10
#include <asm/syscall.h>

#include "trace_output.h"
#include "trace.h"

11
static DEFINE_MUTEX(syscall_trace_lock);
12 13
static int sys_refcount_enter;
static int sys_refcount_exit;
14 15
static DECLARE_BITMAP(enabled_enter_syscalls, NR_syscalls);
static DECLARE_BITMAP(enabled_exit_syscalls, NR_syscalls);
16

17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79
extern unsigned long __start_syscalls_metadata[];
extern unsigned long __stop_syscalls_metadata[];

static struct syscall_metadata **syscalls_metadata;

static struct syscall_metadata *find_syscall_meta(unsigned long syscall)
{
	struct syscall_metadata *start;
	struct syscall_metadata *stop;
	char str[KSYM_SYMBOL_LEN];


	start = (struct syscall_metadata *)__start_syscalls_metadata;
	stop = (struct syscall_metadata *)__stop_syscalls_metadata;
	kallsyms_lookup(syscall, NULL, NULL, NULL, str);

	for ( ; start < stop; start++) {
		/*
		 * Only compare after the "sys" prefix. Archs that use
		 * syscall wrappers may have syscalls symbols aliases prefixed
		 * with "SyS" instead of "sys", leading to an unwanted
		 * mismatch.
		 */
		if (start->name && !strcmp(start->name + 3, str + 3))
			return start;
	}
	return NULL;
}

static struct syscall_metadata *syscall_nr_to_meta(int nr)
{
	if (!syscalls_metadata || nr >= NR_syscalls || nr < 0)
		return NULL;

	return syscalls_metadata[nr];
}

int syscall_name_to_nr(char *name)
{
	int i;

	if (!syscalls_metadata)
		return -1;

	for (i = 0; i < NR_syscalls; i++) {
		if (syscalls_metadata[i]) {
			if (!strcmp(syscalls_metadata[i]->name, name))
				return i;
		}
	}
	return -1;
}

void set_syscall_enter_id(int num, int id)
{
	syscalls_metadata[num]->enter_id = id;
}

void set_syscall_exit_id(int num, int id)
{
	syscalls_metadata[num]->exit_id = id;
}

80 81 82 83 84 85 86 87 88
enum print_line_t
print_syscall_enter(struct trace_iterator *iter, int flags)
{
	struct trace_seq *s = &iter->seq;
	struct trace_entry *ent = iter->ent;
	struct syscall_trace_enter *trace;
	struct syscall_metadata *entry;
	int i, ret, syscall;

89
	trace = (typeof(trace))ent;
90 91
	syscall = trace->nr;
	entry = syscall_nr_to_meta(syscall);
92

93 94 95
	if (!entry)
		goto end;

96 97 98 99 100
	if (entry->enter_id != ent->type) {
		WARN_ON_ONCE(1);
		goto end;
	}

101 102 103 104 105 106
	ret = trace_seq_printf(s, "%s(", entry->name);
	if (!ret)
		return TRACE_TYPE_PARTIAL_LINE;

	for (i = 0; i < entry->nb_args; i++) {
		/* parameter types */
107
		if (trace_flags & TRACE_ITER_VERBOSE) {
108 109 110 111 112
			ret = trace_seq_printf(s, "%s ", entry->types[i]);
			if (!ret)
				return TRACE_TYPE_PARTIAL_LINE;
		}
		/* parameter values */
113
		ret = trace_seq_printf(s, "%s: %lx%s", entry->args[i],
114
				       trace->args[i],
115
				       i == entry->nb_args - 1 ? "" : ", ");
116 117 118 119
		if (!ret)
			return TRACE_TYPE_PARTIAL_LINE;
	}

120 121 122 123
	ret = trace_seq_putc(s, ')');
	if (!ret)
		return TRACE_TYPE_PARTIAL_LINE;

124
end:
125 126 127 128
	ret =  trace_seq_putc(s, '\n');
	if (!ret)
		return TRACE_TYPE_PARTIAL_LINE;

129 130 131 132 133 134 135 136 137 138 139 140 141
	return TRACE_TYPE_HANDLED;
}

enum print_line_t
print_syscall_exit(struct trace_iterator *iter, int flags)
{
	struct trace_seq *s = &iter->seq;
	struct trace_entry *ent = iter->ent;
	struct syscall_trace_exit *trace;
	int syscall;
	struct syscall_metadata *entry;
	int ret;

142
	trace = (typeof(trace))ent;
143 144
	syscall = trace->nr;
	entry = syscall_nr_to_meta(syscall);
145

146 147 148 149 150
	if (!entry) {
		trace_seq_printf(s, "\n");
		return TRACE_TYPE_HANDLED;
	}

151 152 153 154 155
	if (entry->exit_id != ent->type) {
		WARN_ON_ONCE(1);
		return TRACE_TYPE_UNHANDLED;
	}

156 157 158 159 160 161 162 163
	ret = trace_seq_printf(s, "%s -> 0x%lx\n", entry->name,
				trace->ret);
	if (!ret)
		return TRACE_TYPE_PARTIAL_LINE;

	return TRACE_TYPE_HANDLED;
}

164 165 166 167 168
extern char *__bad_type_size(void);

#define SYSCALL_FIELD(type, name)					\
	sizeof(type) != sizeof(trace.name) ?				\
		__bad_type_size() :					\
169 170
		#type, #name, offsetof(typeof(trace), name),		\
		sizeof(trace.name), is_signed_type(type)
171

172
int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s)
173 174 175
{
	int i;
	int nr;
176
	int ret;
177
	struct syscall_metadata *entry;
178 179
	struct syscall_trace_enter trace;
	int offset = offsetof(struct syscall_trace_enter, args);
180

181
	nr = syscall_name_to_nr(call->data);
182 183 184
	entry = syscall_nr_to_meta(nr);

	if (!entry)
185 186
		return 0;

187 188
	ret = trace_seq_printf(s, "\tfield:%s %s;\toffset:%zu;\tsize:%zu;"
			       "\tsigned:%u;\n",
189 190 191
			       SYSCALL_FIELD(int, nr));
	if (!ret)
		return 0;
192 193 194 195 196 197

	for (i = 0; i < entry->nb_args; i++) {
		ret = trace_seq_printf(s, "\tfield:%s %s;", entry->types[i],
				        entry->args[i]);
		if (!ret)
			return 0;
198 199 200 201
		ret = trace_seq_printf(s, "\toffset:%d;\tsize:%zu;"
				       "\tsigned:%u;\n", offset,
				       sizeof(unsigned long),
				       is_signed_type(unsigned long));
202 203 204 205 206
		if (!ret)
			return 0;
		offset += sizeof(unsigned long);
	}

207
	trace_seq_puts(s, "\nprint fmt: \"");
208
	for (i = 0; i < entry->nb_args; i++) {
209
		ret = trace_seq_printf(s, "%s: 0x%%0%zulx%s", entry->args[i],
210
				        sizeof(unsigned long),
211
					i == entry->nb_args - 1 ? "" : ", ");
212 213 214
		if (!ret)
			return 0;
	}
215
	trace_seq_putc(s, '"');
216 217

	for (i = 0; i < entry->nb_args; i++) {
218 219
		ret = trace_seq_printf(s, ", ((unsigned long)(REC->%s))",
				       entry->args[i]);
220 221 222 223
		if (!ret)
			return 0;
	}

224
	return trace_seq_putc(s, '\n');
225 226
}

227 228 229 230 231 232
int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s)
{
	int ret;
	struct syscall_trace_exit trace;

	ret = trace_seq_printf(s,
233 234 235 236
			       "\tfield:%s %s;\toffset:%zu;\tsize:%zu;"
			       "\tsigned:%u;\n"
			       "\tfield:%s %s;\toffset:%zu;\tsize:%zu;"
			       "\tsigned:%u;\n",
237
			       SYSCALL_FIELD(int, nr),
238
			       SYSCALL_FIELD(long, ret));
239 240 241 242 243 244
	if (!ret)
		return 0;

	return trace_seq_printf(s, "\nprint fmt: \"0x%%lx\", REC->ret\n");
}

245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264
int syscall_enter_define_fields(struct ftrace_event_call *call)
{
	struct syscall_trace_enter trace;
	struct syscall_metadata *meta;
	int ret;
	int nr;
	int i;
	int offset = offsetof(typeof(trace), args);

	nr = syscall_name_to_nr(call->data);
	meta = syscall_nr_to_meta(nr);

	if (!meta)
		return 0;

	ret = trace_define_common_fields(call);
	if (ret)
		return ret;

	for (i = 0; i < meta->nb_args; i++) {
265 266
		ret = trace_define_field(call, meta->types[i],
					 meta->args[i], offset,
267 268
					 sizeof(unsigned long), 0,
					 FILTER_OTHER);
269 270 271 272 273 274 275 276 277 278 279 280 281 282 283
		offset += sizeof(unsigned long);
	}

	return ret;
}

int syscall_exit_define_fields(struct ftrace_event_call *call)
{
	struct syscall_trace_exit trace;
	int ret;

	ret = trace_define_common_fields(call);
	if (ret)
		return ret;

284
	ret = trace_define_field(call, SYSCALL_FIELD(long, ret),
285
				 FILTER_OTHER);
286 287 288 289

	return ret;
}

290
void ftrace_syscall_enter(struct pt_regs *regs, long id)
291
{
292 293 294
	struct syscall_trace_enter *entry;
	struct syscall_metadata *sys_data;
	struct ring_buffer_event *event;
295
	struct ring_buffer *buffer;
296
	int size;
297 298 299
	int syscall_nr;

	syscall_nr = syscall_get_nr(current, regs);
300 301
	if (syscall_nr < 0)
		return;
302 303
	if (!test_bit(syscall_nr, enabled_enter_syscalls))
		return;
304

305 306 307 308 309 310
	sys_data = syscall_nr_to_meta(syscall_nr);
	if (!sys_data)
		return;

	size = sizeof(*entry) + sizeof(unsigned long) * sys_data->nb_args;

311 312
	event = trace_current_buffer_lock_reserve(&buffer, sys_data->enter_id,
						  size, 0, 0);
313 314 315 316 317 318 319
	if (!event)
		return;

	entry = ring_buffer_event_data(event);
	entry->nr = syscall_nr;
	syscall_get_arguments(current, regs, 0, sys_data->nb_args, entry->args);

320 321 322
	if (!filter_current_check_discard(buffer, sys_data->enter_event,
					  entry, event))
		trace_current_buffer_unlock_commit(buffer, event, 0, 0);
323 324
}

325
void ftrace_syscall_exit(struct pt_regs *regs, long ret)
326
{
327 328 329
	struct syscall_trace_exit *entry;
	struct syscall_metadata *sys_data;
	struct ring_buffer_event *event;
330
	struct ring_buffer *buffer;
331 332 333
	int syscall_nr;

	syscall_nr = syscall_get_nr(current, regs);
334 335
	if (syscall_nr < 0)
		return;
336 337
	if (!test_bit(syscall_nr, enabled_exit_syscalls))
		return;
338

339 340 341 342
	sys_data = syscall_nr_to_meta(syscall_nr);
	if (!sys_data)
		return;

343
	event = trace_current_buffer_lock_reserve(&buffer, sys_data->exit_id,
344 345 346 347 348 349 350 351
				sizeof(*entry), 0, 0);
	if (!event)
		return;

	entry = ring_buffer_event_data(event);
	entry->nr = syscall_nr;
	entry->ret = syscall_get_return_value(current, regs);

352 353 354
	if (!filter_current_check_discard(buffer, sys_data->exit_event,
					  entry, event))
		trace_current_buffer_unlock_commit(buffer, event, 0, 0);
355 356
}

357
int reg_event_syscall_enter(struct ftrace_event_call *call)
358
{
359 360 361 362
	int ret = 0;
	int num;
	char *name;

363
	name = (char *)call->data;
364
	num = syscall_name_to_nr(name);
365
	if (num < 0 || num >= NR_syscalls)
366 367 368
		return -ENOSYS;
	mutex_lock(&syscall_trace_lock);
	if (!sys_refcount_enter)
369
		ret = register_trace_sys_enter(ftrace_syscall_enter);
370 371 372 373 374 375 376 377 378
	if (ret) {
		pr_info("event trace: Could not activate"
				"syscall entry trace point");
	} else {
		set_bit(num, enabled_enter_syscalls);
		sys_refcount_enter++;
	}
	mutex_unlock(&syscall_trace_lock);
	return ret;
379 380
}

381
void unreg_event_syscall_enter(struct ftrace_event_call *call)
382
{
383 384
	int num;
	char *name;
385

386
	name = (char *)call->data;
387
	num = syscall_name_to_nr(name);
388
	if (num < 0 || num >= NR_syscalls)
389 390 391 392 393
		return;
	mutex_lock(&syscall_trace_lock);
	sys_refcount_enter--;
	clear_bit(num, enabled_enter_syscalls);
	if (!sys_refcount_enter)
394
		unregister_trace_sys_enter(ftrace_syscall_enter);
395 396
	mutex_unlock(&syscall_trace_lock);
}
397

398
int reg_event_syscall_exit(struct ftrace_event_call *call)
399
{
400 401 402 403
	int ret = 0;
	int num;
	char *name;

404
	name = call->data;
405
	num = syscall_name_to_nr(name);
406
	if (num < 0 || num >= NR_syscalls)
407 408 409
		return -ENOSYS;
	mutex_lock(&syscall_trace_lock);
	if (!sys_refcount_exit)
410
		ret = register_trace_sys_exit(ftrace_syscall_exit);
411 412 413 414 415 416
	if (ret) {
		pr_info("event trace: Could not activate"
				"syscall exit trace point");
	} else {
		set_bit(num, enabled_exit_syscalls);
		sys_refcount_exit++;
417
	}
418 419 420
	mutex_unlock(&syscall_trace_lock);
	return ret;
}
421

422
void unreg_event_syscall_exit(struct ftrace_event_call *call)
423 424 425
{
	int num;
	char *name;
426

427
	name = call->data;
428
	num = syscall_name_to_nr(name);
429
	if (num < 0 || num >= NR_syscalls)
430 431 432 433 434
		return;
	mutex_lock(&syscall_trace_lock);
	sys_refcount_exit--;
	clear_bit(num, enabled_exit_syscalls);
	if (!sys_refcount_exit)
435
		unregister_trace_sys_exit(ftrace_syscall_exit);
436
	mutex_unlock(&syscall_trace_lock);
437
}
438 439 440 441 442 443 444 445

struct trace_event event_syscall_enter = {
	.trace			= print_syscall_enter,
};

struct trace_event event_syscall_exit = {
	.trace			= print_syscall_exit,
};
446

447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469
int __init init_ftrace_syscalls(void)
{
	struct syscall_metadata *meta;
	unsigned long addr;
	int i;

	syscalls_metadata = kzalloc(sizeof(*syscalls_metadata) *
					NR_syscalls, GFP_KERNEL);
	if (!syscalls_metadata) {
		WARN_ON(1);
		return -ENOMEM;
	}

	for (i = 0; i < NR_syscalls; i++) {
		addr = arch_syscall_addr(i);
		meta = find_syscall_meta(addr);
		syscalls_metadata[i] = meta;
	}

	return 0;
}
core_initcall(init_ftrace_syscalls);

470
#ifdef CONFIG_EVENT_PROFILE
471

472 473
static DECLARE_BITMAP(enabled_prof_enter_syscalls, NR_syscalls);
static DECLARE_BITMAP(enabled_prof_exit_syscalls, NR_syscalls);
474 475 476 477 478 479
static int sys_prof_refcount_enter;
static int sys_prof_refcount_exit;

static void prof_syscall_enter(struct pt_regs *regs, long id)
{
	struct syscall_metadata *sys_data;
480 481
	struct syscall_trace_enter *rec;
	unsigned long flags;
482
	char *trace_buf;
483
	char *raw_data;
484
	int syscall_nr;
485
	int rctx;
486
	int size;
487
	int cpu;
488 489 490 491 492 493 494 495 496

	syscall_nr = syscall_get_nr(current, regs);
	if (!test_bit(syscall_nr, enabled_prof_enter_syscalls))
		return;

	sys_data = syscall_nr_to_meta(syscall_nr);
	if (!sys_data)
		return;

497 498 499 500 501
	/* get the size after alignment with the u32 buffer size field */
	size = sizeof(unsigned long) * sys_data->nb_args + sizeof(*rec);
	size = ALIGN(size + sizeof(u32), sizeof(u64));
	size -= sizeof(u32);

502 503 504 505 506 507
	if (WARN_ONCE(size > FTRACE_MAX_PROFILE_SIZE,
		      "profile buffer not large enough"))
		return;

	/* Protect the per cpu buffer, begin the rcu read side */
	local_irq_save(flags);
508

509 510
	rctx = perf_swevent_get_recursion_context();
	if (rctx < 0)
511 512
		goto end_recursion;

513 514
	cpu = smp_processor_id();

515
	trace_buf = rcu_dereference(perf_trace_buf);
516

517
	if (!trace_buf)
518
		goto end;
519

520
	raw_data = per_cpu_ptr(trace_buf, cpu);
521 522 523 524 525 526 527 528 529 530

	/* zero the dead bytes from align to not leak stack to user */
	*(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL;

	rec = (struct syscall_trace_enter *) raw_data;
	tracing_generic_entry_update(&rec->ent, 0, 0);
	rec->ent.type = sys_data->enter_id;
	rec->nr = syscall_nr;
	syscall_get_arguments(current, regs, 0, sys_data->nb_args,
			       (unsigned long *)&rec->args);
531
	perf_tp_event(sys_data->enter_id, 0, 1, rec, size);
532 533

end:
534
	perf_swevent_put_recursion_context(rctx);
535
end_recursion:
536
	local_irq_restore(flags);
537 538 539 540 541 542 543 544
}

int reg_prof_syscall_enter(char *name)
{
	int ret = 0;
	int num;

	num = syscall_name_to_nr(name);
545
	if (num < 0 || num >= NR_syscalls)
546 547 548 549
		return -ENOSYS;

	mutex_lock(&syscall_trace_lock);
	if (!sys_prof_refcount_enter)
550
		ret = register_trace_sys_enter(prof_syscall_enter);
551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566
	if (ret) {
		pr_info("event trace: Could not activate"
				"syscall entry trace point");
	} else {
		set_bit(num, enabled_prof_enter_syscalls);
		sys_prof_refcount_enter++;
	}
	mutex_unlock(&syscall_trace_lock);
	return ret;
}

void unreg_prof_syscall_enter(char *name)
{
	int num;

	num = syscall_name_to_nr(name);
567
	if (num < 0 || num >= NR_syscalls)
568 569 570 571 572 573
		return;

	mutex_lock(&syscall_trace_lock);
	sys_prof_refcount_enter--;
	clear_bit(num, enabled_prof_enter_syscalls);
	if (!sys_prof_refcount_enter)
574
		unregister_trace_sys_enter(prof_syscall_enter);
575 576 577 578 579 580
	mutex_unlock(&syscall_trace_lock);
}

static void prof_syscall_exit(struct pt_regs *regs, long ret)
{
	struct syscall_metadata *sys_data;
581 582
	struct syscall_trace_exit *rec;
	unsigned long flags;
583
	int syscall_nr;
584
	char *trace_buf;
585
	char *raw_data;
586
	int rctx;
587 588
	int size;
	int cpu;
589 590 591 592 593 594 595 596 597

	syscall_nr = syscall_get_nr(current, regs);
	if (!test_bit(syscall_nr, enabled_prof_exit_syscalls))
		return;

	sys_data = syscall_nr_to_meta(syscall_nr);
	if (!sys_data)
		return;

598 599 600
	/* We can probably do that at build time */
	size = ALIGN(sizeof(*rec) + sizeof(u32), sizeof(u64));
	size -= sizeof(u32);
601

602 603 604 605 606 607 608 609 610 611
	/*
	 * Impossible, but be paranoid with the future
	 * How to put this check outside runtime?
	 */
	if (WARN_ONCE(size > FTRACE_MAX_PROFILE_SIZE,
		"exit event has grown above profile buffer size"))
		return;

	/* Protect the per cpu buffer, begin the rcu read side */
	local_irq_save(flags);
612

613 614
	rctx = perf_swevent_get_recursion_context();
	if (rctx < 0)
615 616
		goto end_recursion;

617 618
	cpu = smp_processor_id();

619
	trace_buf = rcu_dereference(perf_trace_buf);
620

621
	if (!trace_buf)
622 623
		goto end;

624
	raw_data = per_cpu_ptr(trace_buf, cpu);
625 626 627 628 629 630 631 632 633 634 635

	/* zero the dead bytes from align to not leak stack to user */
	*(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL;

	rec = (struct syscall_trace_exit *)raw_data;

	tracing_generic_entry_update(&rec->ent, 0, 0);
	rec->ent.type = sys_data->exit_id;
	rec->nr = syscall_nr;
	rec->ret = syscall_get_return_value(current, regs);

636
	perf_tp_event(sys_data->exit_id, 0, 1, rec, size);
637 638

end:
639
	perf_swevent_put_recursion_context(rctx);
640
end_recursion:
641
	local_irq_restore(flags);
642 643 644 645 646 647 648 649
}

int reg_prof_syscall_exit(char *name)
{
	int ret = 0;
	int num;

	num = syscall_name_to_nr(name);
650
	if (num < 0 || num >= NR_syscalls)
651 652 653 654
		return -ENOSYS;

	mutex_lock(&syscall_trace_lock);
	if (!sys_prof_refcount_exit)
655
		ret = register_trace_sys_exit(prof_syscall_exit);
656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671
	if (ret) {
		pr_info("event trace: Could not activate"
				"syscall entry trace point");
	} else {
		set_bit(num, enabled_prof_exit_syscalls);
		sys_prof_refcount_exit++;
	}
	mutex_unlock(&syscall_trace_lock);
	return ret;
}

void unreg_prof_syscall_exit(char *name)
{
	int num;

	num = syscall_name_to_nr(name);
672
	if (num < 0 || num >= NR_syscalls)
673 674 675 676 677 678
		return;

	mutex_lock(&syscall_trace_lock);
	sys_prof_refcount_exit--;
	clear_bit(num, enabled_prof_exit_syscalls);
	if (!sys_prof_refcount_exit)
679
		unregister_trace_sys_exit(prof_syscall_exit);
680 681 682 683 684 685
	mutex_unlock(&syscall_trace_lock);
}

#endif