session.c 60.2 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0
2
#include <errno.h>
3
#include <inttypes.h>
4
#include <linux/kernel.h>
5
#include <traceevent/event-parse.h>
6
#include <api/fs/fs.h>
7

8
#include <byteswap.h>
9 10
#include <unistd.h>
#include <sys/types.h>
11
#include <sys/mman.h>
12

13 14
#include "evlist.h"
#include "evsel.h"
15
#include "memswap.h"
16
#include "session.h"
17
#include "tool.h"
18
#include "sort.h"
19
#include "util.h"
20
#include "cpumap.h"
21
#include "perf_regs.h"
22
#include "asm/bug.h"
23
#include "auxtrace.h"
24
#include "thread.h"
25
#include "thread-stack.h"
26
#include "stat.h"
27

28 29 30 31 32
static int perf_session__deliver_event(struct perf_session *session,
				       union perf_event *event,
				       struct perf_sample *sample,
				       struct perf_tool *tool,
				       u64 file_offset);
33

34
static int perf_session__open(struct perf_session *session)
35
{
36
	struct perf_data *data = session->data;
37

38
	if (perf_session__read_header(session) < 0) {
39
		pr_err("incompatible file format (rerun with -v to learn more)\n");
40
		return -1;
41 42
	}

43
	if (perf_data__is_pipe(data))
44 45
		return 0;

46 47 48
	if (perf_header__has_feat(&session->header, HEADER_STAT))
		return 0;

49
	if (!perf_evlist__valid_sample_type(session->evlist)) {
50
		pr_err("non matching sample_type\n");
51
		return -1;
52 53
	}

54
	if (!perf_evlist__valid_sample_id_all(session->evlist)) {
55
		pr_err("non matching sample_id_all\n");
56
		return -1;
57 58
	}

59
	if (!perf_evlist__valid_read_format(session->evlist)) {
60
		pr_err("non matching read_format\n");
61
		return -1;
62 63
	}

64 65 66
	return 0;
}

67
void perf_session__set_id_hdr_size(struct perf_session *session)
68
{
69 70 71
	u16 id_hdr_size = perf_evlist__id_hdr_size(session->evlist);

	machines__set_id_hdr_size(&session->machines, id_hdr_size);
72 73
}

74
int perf_session__create_kernel_maps(struct perf_session *session)
75
{
76
	int ret = machine__create_kernel_maps(&session->machines.host);
77 78

	if (ret >= 0)
79
		ret = machines__create_guest_kernel_maps(&session->machines);
80 81 82
	return ret;
}

83
static void perf_session__destroy_kernel_maps(struct perf_session *session)
84
{
85
	machines__destroy_kernel_maps(&session->machines);
86 87
}

88 89 90 91
static bool perf_session__has_comm_exec(struct perf_session *session)
{
	struct perf_evsel *evsel;

92
	evlist__for_each_entry(session->evlist, evsel) {
93 94 95 96 97 98 99 100 101 102 103 104 105 106
		if (evsel->attr.comm_exec)
			return true;
	}

	return false;
}

static void perf_session__set_comm_exec(struct perf_session *session)
{
	bool comm_exec = perf_session__has_comm_exec(session);

	machines__set_comm_exec(&session->machines, comm_exec);
}

107
static int ordered_events__deliver_event(struct ordered_events *oe,
108
					 struct ordered_event *event)
109
{
110 111 112 113 114 115 116 117 118 119
	struct perf_sample sample;
	struct perf_session *session = container_of(oe, struct perf_session,
						    ordered_events);
	int ret = perf_evlist__parse_sample(session->evlist, event->event, &sample);

	if (ret) {
		pr_err("Can't parse sample, err = %d\n", ret);
		return ret;
	}

120 121
	return perf_session__deliver_event(session, event->event, &sample,
					   session->tool, event->file_offset);
122 123
}

124
struct perf_session *perf_session__new(struct perf_data *data,
125
				       bool repipe, struct perf_tool *tool)
126
{
127
	struct perf_session *session = zalloc(sizeof(*session));
128

129
	if (!session)
130 131
		goto out;

132
	session->repipe = repipe;
133
	session->tool   = tool;
134
	INIT_LIST_HEAD(&session->auxtrace_index);
135
	machines__init(&session->machines);
136
	ordered_events__init(&session->ordered_events, ordered_events__deliver_event);
137

138 139
	if (data) {
		if (perf_data__open(data))
140
			goto out_delete;
141

142
		session->data = data;
143

144
		if (perf_data__is_read(data)) {
145
			if (perf_session__open(session) < 0)
146 147
				goto out_close;

148 149 150 151
			/*
			 * set session attributes that are present in perf.data
			 * but not in pipe-mode.
			 */
152
			if (!data->is_pipe) {
153 154 155
				perf_session__set_id_hdr_size(session);
				perf_session__set_comm_exec(session);
			}
156
		}
157 158
	} else  {
		session->machines.host.env = &perf_env;
159 160
	}

161
	if (!data || perf_data__is_write(data)) {
162 163
		/*
		 * In O_RDONLY mode this will be performed when reading the
164
		 * kernel MMAP event, in perf_event__process_mmap().
165
		 */
166
		if (perf_session__create_kernel_maps(session) < 0)
167
			pr_warning("Cannot read kernel map\n");
168
	}
169

170 171 172 173
	/*
	 * In pipe-mode, evlist is empty until PERF_RECORD_HEADER_ATTR is
	 * processed, so perf_evlist__sample_id_all is not meaningful here.
	 */
174
	if ((!data || !data->is_pipe) && tool && tool->ordering_requires_timestamps &&
175
	    tool->ordered_events && !perf_evlist__sample_id_all(session->evlist)) {
176
		dump_printf("WARNING: No sample_id_all support, falling back to unordered processing\n");
177
		tool->ordered_events = false;
178
	}
179

180
	return session;
181 182

 out_close:
183
	perf_data__close(data);
184
 out_delete:
185
	perf_session__delete(session);
186
 out:
187
	return NULL;
188 189
}

190 191
static void perf_session__delete_threads(struct perf_session *session)
{
192
	machine__delete_threads(&session->machines.host);
193 194
}

195
void perf_session__delete(struct perf_session *session)
196
{
197 198
	if (session == NULL)
		return;
199
	auxtrace__free(session);
200
	auxtrace_index__free(&session->auxtrace_index);
201 202
	perf_session__destroy_kernel_maps(session);
	perf_session__delete_threads(session);
203
	perf_env__exit(&session->header.env);
204
	machines__exit(&session->machines);
205 206
	if (session->data)
		perf_data__close(session->data);
207
	free(session);
208
}
209

210 211 212
static int process_event_synth_tracing_data_stub(struct perf_tool *tool
						 __maybe_unused,
						 union perf_event *event
213 214 215
						 __maybe_unused,
						 struct perf_session *session
						__maybe_unused)
216 217 218 219 220
{
	dump_printf(": unhandled!\n");
	return 0;
}

221 222
static int process_event_synth_attr_stub(struct perf_tool *tool __maybe_unused,
					 union perf_event *event __maybe_unused,
223 224
					 struct perf_evlist **pevlist
					 __maybe_unused)
225 226 227 228 229
{
	dump_printf(": unhandled!\n");
	return 0;
}

230 231 232 233 234
static int process_event_synth_event_update_stub(struct perf_tool *tool __maybe_unused,
						 union perf_event *event __maybe_unused,
						 struct perf_evlist **pevlist
						 __maybe_unused)
{
235 236 237
	if (dump_trace)
		perf_event__fprintf_event_update(event, stdout);

238 239 240 241
	dump_printf(": unhandled!\n");
	return 0;
}

242 243 244 245 246
static int process_event_sample_stub(struct perf_tool *tool __maybe_unused,
				     union perf_event *event __maybe_unused,
				     struct perf_sample *sample __maybe_unused,
				     struct perf_evsel *evsel __maybe_unused,
				     struct machine *machine __maybe_unused)
247 248 249 250 251
{
	dump_printf(": unhandled!\n");
	return 0;
}

252 253 254 255
static int process_event_stub(struct perf_tool *tool __maybe_unused,
			      union perf_event *event __maybe_unused,
			      struct perf_sample *sample __maybe_unused,
			      struct machine *machine __maybe_unused)
256 257 258 259 260
{
	dump_printf(": unhandled!\n");
	return 0;
}

261 262
static int process_finished_round_stub(struct perf_tool *tool __maybe_unused,
				       union perf_event *event __maybe_unused,
263
				       struct ordered_events *oe __maybe_unused)
264 265 266 267 268
{
	dump_printf(": unhandled!\n");
	return 0;
}

269
static int process_finished_round(struct perf_tool *tool,
270
				  union perf_event *event,
271
				  struct ordered_events *oe);
272

273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293
static int skipn(int fd, off_t n)
{
	char buf[4096];
	ssize_t ret;

	while (n > 0) {
		ret = read(fd, buf, min(n, (off_t)sizeof(buf)));
		if (ret <= 0)
			return ret;
		n -= ret;
	}

	return 0;
}

static s64 process_event_auxtrace_stub(struct perf_tool *tool __maybe_unused,
				       union perf_event *event,
				       struct perf_session *session
				       __maybe_unused)
{
	dump_printf(": unhandled!\n");
294 295
	if (perf_data__is_pipe(session->data))
		skipn(perf_data__fd(session->data), event->auxtrace.size);
296 297 298
	return event->auxtrace.size;
}

299 300 301
static int process_event_op2_stub(struct perf_tool *tool __maybe_unused,
				  union perf_event *event __maybe_unused,
				  struct perf_session *session __maybe_unused)
302 303 304 305 306
{
	dump_printf(": unhandled!\n");
	return 0;
}

307 308 309 310 311 312

static
int process_event_thread_map_stub(struct perf_tool *tool __maybe_unused,
				  union perf_event *event __maybe_unused,
				  struct perf_session *session __maybe_unused)
{
313 314 315
	if (dump_trace)
		perf_event__fprintf_thread_map(event, stdout);

316 317 318 319
	dump_printf(": unhandled!\n");
	return 0;
}

320 321 322 323 324
static
int process_event_cpu_map_stub(struct perf_tool *tool __maybe_unused,
			       union perf_event *event __maybe_unused,
			       struct perf_session *session __maybe_unused)
{
325 326 327
	if (dump_trace)
		perf_event__fprintf_cpu_map(event, stdout);

328 329 330 331
	dump_printf(": unhandled!\n");
	return 0;
}

332 333 334 335 336
static
int process_event_stat_config_stub(struct perf_tool *tool __maybe_unused,
				   union perf_event *event __maybe_unused,
				   struct perf_session *session __maybe_unused)
{
337 338 339
	if (dump_trace)
		perf_event__fprintf_stat_config(event, stdout);

340 341 342 343
	dump_printf(": unhandled!\n");
	return 0;
}

J
Jiri Olsa 已提交
344 345 346 347 348
static int process_stat_stub(struct perf_tool *tool __maybe_unused,
			     union perf_event *event __maybe_unused,
			     struct perf_session *perf_session
			     __maybe_unused)
{
349 350 351
	if (dump_trace)
		perf_event__fprintf_stat(event, stdout);

J
Jiri Olsa 已提交
352 353 354 355
	dump_printf(": unhandled!\n");
	return 0;
}

356 357 358 359 360
static int process_stat_round_stub(struct perf_tool *tool __maybe_unused,
				   union perf_event *event __maybe_unused,
				   struct perf_session *perf_session
				   __maybe_unused)
{
361 362 363
	if (dump_trace)
		perf_event__fprintf_stat_round(event, stdout);

364 365 366 367
	dump_printf(": unhandled!\n");
	return 0;
}

368
void perf_tool__fill_defaults(struct perf_tool *tool)
369
{
370 371 372 373
	if (tool->sample == NULL)
		tool->sample = process_event_sample_stub;
	if (tool->mmap == NULL)
		tool->mmap = process_event_stub;
374 375
	if (tool->mmap2 == NULL)
		tool->mmap2 = process_event_stub;
376 377
	if (tool->comm == NULL)
		tool->comm = process_event_stub;
378 379
	if (tool->namespaces == NULL)
		tool->namespaces = process_event_stub;
380 381 382 383 384 385
	if (tool->fork == NULL)
		tool->fork = process_event_stub;
	if (tool->exit == NULL)
		tool->exit = process_event_stub;
	if (tool->lost == NULL)
		tool->lost = perf_event__process_lost;
386 387
	if (tool->lost_samples == NULL)
		tool->lost_samples = perf_event__process_lost_samples;
388 389
	if (tool->aux == NULL)
		tool->aux = perf_event__process_aux;
390 391
	if (tool->itrace_start == NULL)
		tool->itrace_start = perf_event__process_itrace_start;
392 393
	if (tool->context_switch == NULL)
		tool->context_switch = perf_event__process_switch;
394 395 396 397 398 399 400 401
	if (tool->read == NULL)
		tool->read = process_event_sample_stub;
	if (tool->throttle == NULL)
		tool->throttle = process_event_stub;
	if (tool->unthrottle == NULL)
		tool->unthrottle = process_event_stub;
	if (tool->attr == NULL)
		tool->attr = process_event_synth_attr_stub;
402 403
	if (tool->event_update == NULL)
		tool->event_update = process_event_synth_event_update_stub;
404 405 406
	if (tool->tracing_data == NULL)
		tool->tracing_data = process_event_synth_tracing_data_stub;
	if (tool->build_id == NULL)
407
		tool->build_id = process_event_op2_stub;
408
	if (tool->finished_round == NULL) {
409
		if (tool->ordered_events)
410
			tool->finished_round = process_finished_round;
411
		else
412
			tool->finished_round = process_finished_round_stub;
413
	}
A
Adrian Hunter 已提交
414
	if (tool->id_index == NULL)
415
		tool->id_index = process_event_op2_stub;
416
	if (tool->auxtrace_info == NULL)
417
		tool->auxtrace_info = process_event_op2_stub;
418 419
	if (tool->auxtrace == NULL)
		tool->auxtrace = process_event_auxtrace_stub;
420
	if (tool->auxtrace_error == NULL)
421
		tool->auxtrace_error = process_event_op2_stub;
422 423
	if (tool->thread_map == NULL)
		tool->thread_map = process_event_thread_map_stub;
424 425
	if (tool->cpu_map == NULL)
		tool->cpu_map = process_event_cpu_map_stub;
426 427
	if (tool->stat_config == NULL)
		tool->stat_config = process_event_stat_config_stub;
J
Jiri Olsa 已提交
428 429
	if (tool->stat == NULL)
		tool->stat = process_stat_stub;
430 431
	if (tool->stat_round == NULL)
		tool->stat_round = process_stat_round_stub;
432 433
	if (tool->time_conv == NULL)
		tool->time_conv = process_event_op2_stub;
434 435
	if (tool->feature == NULL)
		tool->feature = process_event_op2_stub;
436
}
437

438 439 440 441 442 443 444 445 446 447
static void swap_sample_id_all(union perf_event *event, void *data)
{
	void *end = (void *) event + event->header.size;
	int size = end - data;

	BUG_ON(size % sizeof(u64));
	mem_bswap_64(data, size);
}

static void perf_event__all64_swap(union perf_event *event,
448
				   bool sample_id_all __maybe_unused)
449
{
450 451
	struct perf_event_header *hdr = &event->header;
	mem_bswap_64(hdr + 1, event->header.size - sizeof(*hdr));
452 453
}

454
static void perf_event__comm_swap(union perf_event *event, bool sample_id_all)
455
{
456 457
	event->comm.pid = bswap_32(event->comm.pid);
	event->comm.tid = bswap_32(event->comm.tid);
458 459 460 461

	if (sample_id_all) {
		void *data = &event->comm.comm;

462
		data += PERF_ALIGN(strlen(data) + 1, sizeof(u64));
463 464
		swap_sample_id_all(event, data);
	}
465 466
}

467 468
static void perf_event__mmap_swap(union perf_event *event,
				  bool sample_id_all)
469
{
470 471 472 473 474
	event->mmap.pid	  = bswap_32(event->mmap.pid);
	event->mmap.tid	  = bswap_32(event->mmap.tid);
	event->mmap.start = bswap_64(event->mmap.start);
	event->mmap.len	  = bswap_64(event->mmap.len);
	event->mmap.pgoff = bswap_64(event->mmap.pgoff);
475 476 477 478

	if (sample_id_all) {
		void *data = &event->mmap.filename;

479
		data += PERF_ALIGN(strlen(data) + 1, sizeof(u64));
480 481
		swap_sample_id_all(event, data);
	}
482 483
}

484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502
static void perf_event__mmap2_swap(union perf_event *event,
				  bool sample_id_all)
{
	event->mmap2.pid   = bswap_32(event->mmap2.pid);
	event->mmap2.tid   = bswap_32(event->mmap2.tid);
	event->mmap2.start = bswap_64(event->mmap2.start);
	event->mmap2.len   = bswap_64(event->mmap2.len);
	event->mmap2.pgoff = bswap_64(event->mmap2.pgoff);
	event->mmap2.maj   = bswap_32(event->mmap2.maj);
	event->mmap2.min   = bswap_32(event->mmap2.min);
	event->mmap2.ino   = bswap_64(event->mmap2.ino);

	if (sample_id_all) {
		void *data = &event->mmap2.filename;

		data += PERF_ALIGN(strlen(data) + 1, sizeof(u64));
		swap_sample_id_all(event, data);
	}
}
503
static void perf_event__task_swap(union perf_event *event, bool sample_id_all)
504
{
505 506 507 508 509
	event->fork.pid	 = bswap_32(event->fork.pid);
	event->fork.tid	 = bswap_32(event->fork.tid);
	event->fork.ppid = bswap_32(event->fork.ppid);
	event->fork.ptid = bswap_32(event->fork.ptid);
	event->fork.time = bswap_64(event->fork.time);
510 511 512

	if (sample_id_all)
		swap_sample_id_all(event, &event->fork + 1);
513 514
}

515
static void perf_event__read_swap(union perf_event *event, bool sample_id_all)
516
{
517 518 519 520 521 522
	event->read.pid		 = bswap_32(event->read.pid);
	event->read.tid		 = bswap_32(event->read.tid);
	event->read.value	 = bswap_64(event->read.value);
	event->read.time_enabled = bswap_64(event->read.time_enabled);
	event->read.time_running = bswap_64(event->read.time_running);
	event->read.id		 = bswap_64(event->read.id);
523 524 525

	if (sample_id_all)
		swap_sample_id_all(event, &event->read + 1);
526 527
}

528 529 530 531 532 533 534 535 536 537
static void perf_event__aux_swap(union perf_event *event, bool sample_id_all)
{
	event->aux.aux_offset = bswap_64(event->aux.aux_offset);
	event->aux.aux_size   = bswap_64(event->aux.aux_size);
	event->aux.flags      = bswap_64(event->aux.flags);

	if (sample_id_all)
		swap_sample_id_all(event, &event->aux + 1);
}

538 539 540 541 542 543 544 545 546 547
static void perf_event__itrace_start_swap(union perf_event *event,
					  bool sample_id_all)
{
	event->itrace_start.pid	 = bswap_32(event->itrace_start.pid);
	event->itrace_start.tid	 = bswap_32(event->itrace_start.tid);

	if (sample_id_all)
		swap_sample_id_all(event, &event->itrace_start + 1);
}

548 549 550 551 552 553 554 555 556 557 558 559 560
static void perf_event__switch_swap(union perf_event *event, bool sample_id_all)
{
	if (event->header.type == PERF_RECORD_SWITCH_CPU_WIDE) {
		event->context_switch.next_prev_pid =
				bswap_32(event->context_switch.next_prev_pid);
		event->context_switch.next_prev_tid =
				bswap_32(event->context_switch.next_prev_tid);
	}

	if (sample_id_all)
		swap_sample_id_all(event, &event->context_switch + 1);
}

561 562 563 564 565 566 567 568 569 570 571
static void perf_event__throttle_swap(union perf_event *event,
				      bool sample_id_all)
{
	event->throttle.time	  = bswap_64(event->throttle.time);
	event->throttle.id	  = bswap_64(event->throttle.id);
	event->throttle.stream_id = bswap_64(event->throttle.stream_id);

	if (sample_id_all)
		swap_sample_id_all(event, &event->throttle + 1);
}

572 573 574 575 576 577 578 579 580 581
static u8 revbyte(u8 b)
{
	int rev = (b >> 4) | ((b & 0xf) << 4);
	rev = ((rev & 0xcc) >> 2) | ((rev & 0x33) << 2);
	rev = ((rev & 0xaa) >> 1) | ((rev & 0x55) << 1);
	return (u8) rev;
}

/*
 * XXX this is hack in attempt to carry flags bitfield
582
 * through endian village. ABI says:
583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603
 *
 * Bit-fields are allocated from right to left (least to most significant)
 * on little-endian implementations and from left to right (most to least
 * significant) on big-endian implementations.
 *
 * The above seems to be byte specific, so we need to reverse each
 * byte of the bitfield. 'Internet' also says this might be implementation
 * specific and we probably need proper fix and carry perf_event_attr
 * bitfield flags in separate data file FEAT_ section. Thought this seems
 * to work for now.
 */
static void swap_bitfield(u8 *p, unsigned len)
{
	unsigned i;

	for (i = 0; i < len; i++) {
		*p = revbyte(*p);
		p++;
	}
}

604 605 606 607 608
/* exported for swapping attributes in file header */
void perf_event__attr_swap(struct perf_event_attr *attr)
{
	attr->type		= bswap_32(attr->type);
	attr->size		= bswap_32(attr->size);
609 610 611 612 613 614 615 616 617

#define bswap_safe(f, n) 					\
	(attr->size > (offsetof(struct perf_event_attr, f) + 	\
		       sizeof(attr->f) * (n)))
#define bswap_field(f, sz) 			\
do { 						\
	if (bswap_safe(f, 0))			\
		attr->f = bswap_##sz(attr->f);	\
} while(0)
618
#define bswap_field_16(f) bswap_field(f, 16)
619 620 621 622 623 624 625 626 627 628 629 630 631 632 633
#define bswap_field_32(f) bswap_field(f, 32)
#define bswap_field_64(f) bswap_field(f, 64)

	bswap_field_64(config);
	bswap_field_64(sample_period);
	bswap_field_64(sample_type);
	bswap_field_64(read_format);
	bswap_field_32(wakeup_events);
	bswap_field_32(bp_type);
	bswap_field_64(bp_addr);
	bswap_field_64(bp_len);
	bswap_field_64(branch_sample_type);
	bswap_field_64(sample_regs_user);
	bswap_field_32(sample_stack_user);
	bswap_field_32(aux_watermark);
634
	bswap_field_16(sample_max_stack);
635 636 637 638 639 640 641 642 643 644 645 646

	/*
	 * After read_format are bitfields. Check read_format because
	 * we are unable to use offsetof on bitfield.
	 */
	if (bswap_safe(read_format, 1))
		swap_bitfield((u8 *) (&attr->read_format + 1),
			      sizeof(u64));
#undef bswap_field_64
#undef bswap_field_32
#undef bswap_field
#undef bswap_safe
647 648
}

649
static void perf_event__hdr_attr_swap(union perf_event *event,
650
				      bool sample_id_all __maybe_unused)
651 652 653
{
	size_t size;

654
	perf_event__attr_swap(&event->attr.attr);
655

656 657 658
	size = event->header.size;
	size -= (void *)&event->attr.id - (void *)event;
	mem_bswap_64(event->attr.id, size);
659 660
}

661 662 663 664 665 666 667
static void perf_event__event_update_swap(union perf_event *event,
					  bool sample_id_all __maybe_unused)
{
	event->event_update.type = bswap_64(event->event_update.type);
	event->event_update.id   = bswap_64(event->event_update.id);
}

668
static void perf_event__event_type_swap(union perf_event *event,
669
					bool sample_id_all __maybe_unused)
670
{
671 672
	event->event_type.event_type.event_id =
		bswap_64(event->event_type.event_type.event_id);
673 674
}

675
static void perf_event__tracing_data_swap(union perf_event *event,
676
					  bool sample_id_all __maybe_unused)
677
{
678
	event->tracing_data.size = bswap_32(event->tracing_data.size);
679 680
}

681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703
static void perf_event__auxtrace_info_swap(union perf_event *event,
					   bool sample_id_all __maybe_unused)
{
	size_t size;

	event->auxtrace_info.type = bswap_32(event->auxtrace_info.type);

	size = event->header.size;
	size -= (void *)&event->auxtrace_info.priv - (void *)event;
	mem_bswap_64(event->auxtrace_info.priv, size);
}

static void perf_event__auxtrace_swap(union perf_event *event,
				      bool sample_id_all __maybe_unused)
{
	event->auxtrace.size      = bswap_64(event->auxtrace.size);
	event->auxtrace.offset    = bswap_64(event->auxtrace.offset);
	event->auxtrace.reference = bswap_64(event->auxtrace.reference);
	event->auxtrace.idx       = bswap_32(event->auxtrace.idx);
	event->auxtrace.tid       = bswap_32(event->auxtrace.tid);
	event->auxtrace.cpu       = bswap_32(event->auxtrace.cpu);
}

704 705 706 707 708 709 710 711 712 713 714
static void perf_event__auxtrace_error_swap(union perf_event *event,
					    bool sample_id_all __maybe_unused)
{
	event->auxtrace_error.type = bswap_32(event->auxtrace_error.type);
	event->auxtrace_error.code = bswap_32(event->auxtrace_error.code);
	event->auxtrace_error.cpu  = bswap_32(event->auxtrace_error.cpu);
	event->auxtrace_error.pid  = bswap_32(event->auxtrace_error.pid);
	event->auxtrace_error.tid  = bswap_32(event->auxtrace_error.tid);
	event->auxtrace_error.ip   = bswap_64(event->auxtrace_error.ip);
}

715 716 717 718 719 720 721 722 723 724 725
static void perf_event__thread_map_swap(union perf_event *event,
					bool sample_id_all __maybe_unused)
{
	unsigned i;

	event->thread_map.nr = bswap_64(event->thread_map.nr);

	for (i = 0; i < event->thread_map.nr; i++)
		event->thread_map.entries[i].pid = bswap_64(event->thread_map.entries[i].pid);
}

726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761
static void perf_event__cpu_map_swap(union perf_event *event,
				     bool sample_id_all __maybe_unused)
{
	struct cpu_map_data *data = &event->cpu_map.data;
	struct cpu_map_entries *cpus;
	struct cpu_map_mask *mask;
	unsigned i;

	data->type = bswap_64(data->type);

	switch (data->type) {
	case PERF_CPU_MAP__CPUS:
		cpus = (struct cpu_map_entries *)data->data;

		cpus->nr = bswap_16(cpus->nr);

		for (i = 0; i < cpus->nr; i++)
			cpus->cpu[i] = bswap_16(cpus->cpu[i]);
		break;
	case PERF_CPU_MAP__MASK:
		mask = (struct cpu_map_mask *) data->data;

		mask->nr = bswap_16(mask->nr);
		mask->long_size = bswap_16(mask->long_size);

		switch (mask->long_size) {
		case 4: mem_bswap_32(&mask->mask, mask->nr); break;
		case 8: mem_bswap_64(&mask->mask, mask->nr); break;
		default:
			pr_err("cpu_map swap: unsupported long size\n");
		}
	default:
		break;
	}
}

762 763 764 765 766 767 768 769 770 771
static void perf_event__stat_config_swap(union perf_event *event,
					 bool sample_id_all __maybe_unused)
{
	u64 size;

	size  = event->stat_config.nr * sizeof(event->stat_config.data[0]);
	size += 1; /* nr item itself */
	mem_bswap_64(&event->stat_config.nr, size);
}

J
Jiri Olsa 已提交
772 773 774 775 776 777 778 779 780 781 782
static void perf_event__stat_swap(union perf_event *event,
				  bool sample_id_all __maybe_unused)
{
	event->stat.id     = bswap_64(event->stat.id);
	event->stat.thread = bswap_32(event->stat.thread);
	event->stat.cpu    = bswap_32(event->stat.cpu);
	event->stat.val    = bswap_64(event->stat.val);
	event->stat.ena    = bswap_64(event->stat.ena);
	event->stat.run    = bswap_64(event->stat.run);
}

783 784 785 786 787 788 789
static void perf_event__stat_round_swap(union perf_event *event,
					bool sample_id_all __maybe_unused)
{
	event->stat_round.type = bswap_64(event->stat_round.type);
	event->stat_round.time = bswap_64(event->stat_round.time);
}

790 791
typedef void (*perf_event__swap_op)(union perf_event *event,
				    bool sample_id_all);
792

793 794
static perf_event__swap_op perf_event__swap_ops[] = {
	[PERF_RECORD_MMAP]		  = perf_event__mmap_swap,
795
	[PERF_RECORD_MMAP2]		  = perf_event__mmap2_swap,
796 797 798 799 800
	[PERF_RECORD_COMM]		  = perf_event__comm_swap,
	[PERF_RECORD_FORK]		  = perf_event__task_swap,
	[PERF_RECORD_EXIT]		  = perf_event__task_swap,
	[PERF_RECORD_LOST]		  = perf_event__all64_swap,
	[PERF_RECORD_READ]		  = perf_event__read_swap,
801 802
	[PERF_RECORD_THROTTLE]		  = perf_event__throttle_swap,
	[PERF_RECORD_UNTHROTTLE]	  = perf_event__throttle_swap,
803
	[PERF_RECORD_SAMPLE]		  = perf_event__all64_swap,
804
	[PERF_RECORD_AUX]		  = perf_event__aux_swap,
805
	[PERF_RECORD_ITRACE_START]	  = perf_event__itrace_start_swap,
806
	[PERF_RECORD_LOST_SAMPLES]	  = perf_event__all64_swap,
807 808
	[PERF_RECORD_SWITCH]		  = perf_event__switch_swap,
	[PERF_RECORD_SWITCH_CPU_WIDE]	  = perf_event__switch_swap,
809
	[PERF_RECORD_HEADER_ATTR]	  = perf_event__hdr_attr_swap,
810 811 812
	[PERF_RECORD_HEADER_EVENT_TYPE]	  = perf_event__event_type_swap,
	[PERF_RECORD_HEADER_TRACING_DATA] = perf_event__tracing_data_swap,
	[PERF_RECORD_HEADER_BUILD_ID]	  = NULL,
A
Adrian Hunter 已提交
813
	[PERF_RECORD_ID_INDEX]		  = perf_event__all64_swap,
814 815
	[PERF_RECORD_AUXTRACE_INFO]	  = perf_event__auxtrace_info_swap,
	[PERF_RECORD_AUXTRACE]		  = perf_event__auxtrace_swap,
816
	[PERF_RECORD_AUXTRACE_ERROR]	  = perf_event__auxtrace_error_swap,
817
	[PERF_RECORD_THREAD_MAP]	  = perf_event__thread_map_swap,
818
	[PERF_RECORD_CPU_MAP]		  = perf_event__cpu_map_swap,
819
	[PERF_RECORD_STAT_CONFIG]	  = perf_event__stat_config_swap,
J
Jiri Olsa 已提交
820
	[PERF_RECORD_STAT]		  = perf_event__stat_swap,
821
	[PERF_RECORD_STAT_ROUND]	  = perf_event__stat_round_swap,
822
	[PERF_RECORD_EVENT_UPDATE]	  = perf_event__event_update_swap,
823
	[PERF_RECORD_TIME_CONV]		  = perf_event__all64_swap,
824
	[PERF_RECORD_HEADER_MAX]	  = NULL,
825 826
};

827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865
/*
 * When perf record finishes a pass on every buffers, it records this pseudo
 * event.
 * We record the max timestamp t found in the pass n.
 * Assuming these timestamps are monotonic across cpus, we know that if
 * a buffer still has events with timestamps below t, they will be all
 * available and then read in the pass n + 1.
 * Hence when we start to read the pass n + 2, we can safely flush every
 * events with timestamps below t.
 *
 *    ============ PASS n =================
 *       CPU 0         |   CPU 1
 *                     |
 *    cnt1 timestamps  |   cnt2 timestamps
 *          1          |         2
 *          2          |         3
 *          -          |         4  <--- max recorded
 *
 *    ============ PASS n + 1 ==============
 *       CPU 0         |   CPU 1
 *                     |
 *    cnt1 timestamps  |   cnt2 timestamps
 *          3          |         5
 *          4          |         6
 *          5          |         7 <---- max recorded
 *
 *      Flush every events below timestamp 4
 *
 *    ============ PASS n + 2 ==============
 *       CPU 0         |   CPU 1
 *                     |
 *    cnt1 timestamps  |   cnt2 timestamps
 *          6          |         8
 *          7          |         9
 *          -          |         10
 *
 *      Flush every events below timestamp 7
 *      etc...
 */
866
static int process_finished_round(struct perf_tool *tool __maybe_unused,
867
				  union perf_event *event __maybe_unused,
868
				  struct ordered_events *oe)
869
{
870 871
	if (dump_trace)
		fprintf(stdout, "\n");
872
	return ordered_events__flush(oe, OE_FLUSH__ROUND);
873 874
}

875
int perf_session__queue_event(struct perf_session *s, union perf_event *event,
876
			      u64 timestamp, u64 file_offset)
877
{
878
	return ordered_events__queue(&s->ordered_events, event, timestamp, file_offset);
879
}
880

K
Kan Liang 已提交
881
static void callchain__lbr_callstack_printf(struct perf_sample *sample)
882
{
K
Kan Liang 已提交
883 884 885
	struct ip_callchain *callchain = sample->callchain;
	struct branch_stack *lbr_stack = sample->branch_stack;
	u64 kernel_callchain_nr = callchain->nr;
886
	unsigned int i;
887

K
Kan Liang 已提交
888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933
	for (i = 0; i < kernel_callchain_nr; i++) {
		if (callchain->ips[i] == PERF_CONTEXT_USER)
			break;
	}

	if ((i != kernel_callchain_nr) && lbr_stack->nr) {
		u64 total_nr;
		/*
		 * LBR callstack can only get user call chain,
		 * i is kernel call chain number,
		 * 1 is PERF_CONTEXT_USER.
		 *
		 * The user call chain is stored in LBR registers.
		 * LBR are pair registers. The caller is stored
		 * in "from" register, while the callee is stored
		 * in "to" register.
		 * For example, there is a call stack
		 * "A"->"B"->"C"->"D".
		 * The LBR registers will recorde like
		 * "C"->"D", "B"->"C", "A"->"B".
		 * So only the first "to" register and all "from"
		 * registers are needed to construct the whole stack.
		 */
		total_nr = i + 1 + lbr_stack->nr + 1;
		kernel_callchain_nr = i + 1;

		printf("... LBR call chain: nr:%" PRIu64 "\n", total_nr);

		for (i = 0; i < kernel_callchain_nr; i++)
			printf("..... %2d: %016" PRIx64 "\n",
			       i, callchain->ips[i]);

		printf("..... %2d: %016" PRIx64 "\n",
		       (int)(kernel_callchain_nr), lbr_stack->entries[0].to);
		for (i = 0; i < lbr_stack->nr; i++)
			printf("..... %2d: %016" PRIx64 "\n",
			       (int)(i + kernel_callchain_nr + 1), lbr_stack->entries[i].from);
	}
}

static void callchain__printf(struct perf_evsel *evsel,
			      struct perf_sample *sample)
{
	unsigned int i;
	struct ip_callchain *callchain = sample->callchain;

934
	if (perf_evsel__has_branch_callstack(evsel))
K
Kan Liang 已提交
935 936 937
		callchain__lbr_callstack_printf(sample);

	printf("... FP chain: nr:%" PRIu64 "\n", callchain->nr);
938

K
Kan Liang 已提交
939
	for (i = 0; i < callchain->nr; i++)
940
		printf("..... %2d: %016" PRIx64 "\n",
K
Kan Liang 已提交
941
		       i, callchain->ips[i]);
942 943
}

944 945 946 947 948 949
static void branch_stack__printf(struct perf_sample *sample)
{
	uint64_t i;

	printf("... branch stack: nr:%" PRIu64 "\n", sample->branch_stack->nr);

950 951 952 953 954
	for (i = 0; i < sample->branch_stack->nr; i++) {
		struct branch_entry *e = &sample->branch_stack->entries[i];

		printf("..... %2"PRIu64": %016" PRIx64 " -> %016" PRIx64 " %hu cycles %s%s%s%s %x\n",
			i, e->from, e->to,
955
			(unsigned short)e->flags.cycles,
956 957 958 959 960 961
			e->flags.mispred ? "M" : " ",
			e->flags.predicted ? "P" : " ",
			e->flags.abort ? "A" : " ",
			e->flags.in_tx ? "T" : " ",
			(unsigned)e->flags.reserved);
	}
962 963
}

964 965 966 967 968 969 970 971 972 973 974 975
static void regs_dump__printf(u64 mask, u64 *regs)
{
	unsigned rid, i = 0;

	for_each_set_bit(rid, (unsigned long *) &mask, sizeof(mask) * 8) {
		u64 val = regs[i++];

		printf(".... %-5s 0x%" PRIx64 "\n",
		       perf_reg_name(rid), val);
	}
}

976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001
static const char *regs_abi[] = {
	[PERF_SAMPLE_REGS_ABI_NONE] = "none",
	[PERF_SAMPLE_REGS_ABI_32] = "32-bit",
	[PERF_SAMPLE_REGS_ABI_64] = "64-bit",
};

static inline const char *regs_dump_abi(struct regs_dump *d)
{
	if (d->abi > PERF_SAMPLE_REGS_ABI_64)
		return "unknown";

	return regs_abi[d->abi];
}

static void regs__printf(const char *type, struct regs_dump *regs)
{
	u64 mask = regs->mask;

	printf("... %s regs: mask 0x%" PRIx64 " ABI %s\n",
	       type,
	       mask,
	       regs_dump_abi(regs));

	regs_dump__printf(mask, regs->regs);
}

1002
static void regs_user__printf(struct perf_sample *sample)
1003 1004 1005
{
	struct regs_dump *user_regs = &sample->user_regs;

1006 1007 1008 1009 1010 1011 1012 1013 1014 1015
	if (user_regs->regs)
		regs__printf("user", user_regs);
}

static void regs_intr__printf(struct perf_sample *sample)
{
	struct regs_dump *intr_regs = &sample->intr_regs;

	if (intr_regs->regs)
		regs__printf("intr", intr_regs);
1016 1017 1018 1019 1020 1021 1022 1023
}

static void stack_user__printf(struct stack_dump *dump)
{
	printf("... ustack: size %" PRIu64 ", offset 0x%x\n",
	       dump->size, dump->offset);
}

1024
static void perf_evlist__print_tstamp(struct perf_evlist *evlist,
1025
				       union perf_event *event,
1026
				       struct perf_sample *sample)
1027
{
1028
	u64 sample_type = __perf_evlist__combined_sample_type(evlist);
1029

1030
	if (event->header.type != PERF_RECORD_SAMPLE &&
1031
	    !perf_evlist__sample_id_all(evlist)) {
1032 1033 1034 1035
		fputs("-1 -1 ", stdout);
		return;
	}

1036
	if ((sample_type & PERF_SAMPLE_CPU))
1037 1038
		printf("%u ", sample->cpu);

1039
	if (sample_type & PERF_SAMPLE_TIME)
1040
		printf("%" PRIu64 " ", sample->time);
1041 1042
}

1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072
static void sample_read__printf(struct perf_sample *sample, u64 read_format)
{
	printf("... sample_read:\n");

	if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
		printf("...... time enabled %016" PRIx64 "\n",
		       sample->read.time_enabled);

	if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
		printf("...... time running %016" PRIx64 "\n",
		       sample->read.time_running);

	if (read_format & PERF_FORMAT_GROUP) {
		u64 i;

		printf(".... group nr %" PRIu64 "\n", sample->read.group.nr);

		for (i = 0; i < sample->read.group.nr; i++) {
			struct sample_read_value *value;

			value = &sample->read.group.values[i];
			printf("..... id %016" PRIx64
			       ", value %016" PRIx64 "\n",
			       value->id, value->value);
		}
	} else
		printf("..... id %016" PRIx64 ", value %016" PRIx64 "\n",
			sample->read.one.id, sample->read.one.value);
}

1073
static void dump_event(struct perf_evlist *evlist, union perf_event *event,
1074
		       u64 file_offset, struct perf_sample *sample)
1075 1076 1077 1078
{
	if (!dump_trace)
		return;

1079 1080
	printf("\n%#" PRIx64 " [%#x]: event: %d\n",
	       file_offset, event->header.size, event->header.type);
1081 1082 1083 1084

	trace_event(event);

	if (sample)
1085
		perf_evlist__print_tstamp(evlist, event, sample);
1086

1087
	printf("%#" PRIx64 " [%#x]: PERF_RECORD_%s", file_offset,
1088
	       event->header.size, perf_event__name(event->header.type));
1089 1090
}

1091
static void dump_sample(struct perf_evsel *evsel, union perf_event *event,
1092
			struct perf_sample *sample)
1093
{
1094 1095
	u64 sample_type;

1096 1097 1098
	if (!dump_trace)
		return;

1099
	printf("(IP, 0x%x): %d/%d: %#" PRIx64 " period: %" PRIu64 " addr: %#" PRIx64 "\n",
1100
	       event->header.misc, sample->pid, sample->tid, sample->ip,
1101
	       sample->period, sample->addr);
1102

1103
	sample_type = evsel->attr.sample_type;
1104 1105

	if (sample_type & PERF_SAMPLE_CALLCHAIN)
K
Kan Liang 已提交
1106
		callchain__printf(evsel, sample);
1107

1108
	if ((sample_type & PERF_SAMPLE_BRANCH_STACK) && !perf_evsel__has_branch_callstack(evsel))
1109
		branch_stack__printf(sample);
1110 1111

	if (sample_type & PERF_SAMPLE_REGS_USER)
1112
		regs_user__printf(sample);
1113

1114 1115 1116
	if (sample_type & PERF_SAMPLE_REGS_INTR)
		regs_intr__printf(sample);

1117 1118
	if (sample_type & PERF_SAMPLE_STACK_USER)
		stack_user__printf(&sample->user_stack);
1119 1120 1121

	if (sample_type & PERF_SAMPLE_WEIGHT)
		printf("... weight: %" PRIu64 "\n", sample->weight);
1122 1123 1124

	if (sample_type & PERF_SAMPLE_DATA_SRC)
		printf(" . data_src: 0x%"PRIx64"\n", sample->data_src);
1125

1126 1127 1128
	if (sample_type & PERF_SAMPLE_PHYS_ADDR)
		printf(" .. phys_addr: 0x%"PRIx64"\n", sample->phys_addr);

1129 1130 1131
	if (sample_type & PERF_SAMPLE_TRANSACTION)
		printf("... transaction: %" PRIx64 "\n", sample->transaction);

1132 1133
	if (sample_type & PERF_SAMPLE_READ)
		sample_read__printf(sample, evsel->attr.read_format);
1134 1135
}

J
Jiri Olsa 已提交
1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159
static void dump_read(struct perf_evsel *evsel, union perf_event *event)
{
	struct read_event *read_event = &event->read;
	u64 read_format;

	if (!dump_trace)
		return;

	printf(": %d %d %s %" PRIu64 "\n", event->read.pid, event->read.tid,
	       evsel ? perf_evsel__name(evsel) : "FAIL",
	       event->read.value);

	read_format = evsel->attr.read_format;

	if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
		printf("... time enabled : %" PRIu64 "\n", read_event->time_enabled);

	if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
		printf("... time running : %" PRIu64 "\n", read_event->time_running);

	if (read_format & PERF_FORMAT_ID)
		printf("... id           : %" PRIu64 "\n", read_event->id);
}

1160
static struct machine *machines__find_for_cpumode(struct machines *machines,
1161 1162
					       union perf_event *event,
					       struct perf_sample *sample)
1163
{
1164
	struct machine *machine;
1165

1166
	if (perf_guest &&
1167 1168
	    ((sample->cpumode == PERF_RECORD_MISC_GUEST_KERNEL) ||
	     (sample->cpumode == PERF_RECORD_MISC_GUEST_USER))) {
1169 1170
		u32 pid;

1171 1172
		if (event->header.type == PERF_RECORD_MMAP
		    || event->header.type == PERF_RECORD_MMAP2)
1173 1174
			pid = event->mmap.pid;
		else
1175
			pid = sample->pid;
1176

1177
		machine = machines__find(machines, pid);
1178
		if (!machine)
1179
			machine = machines__findnew(machines, DEFAULT_GUEST_KERNEL_ID);
1180
		return machine;
1181
	}
1182

1183
	return &machines->host;
1184 1185
}

1186
static int deliver_sample_value(struct perf_evlist *evlist,
1187 1188 1189 1190 1191 1192
				struct perf_tool *tool,
				union perf_event *event,
				struct perf_sample *sample,
				struct sample_read_value *v,
				struct machine *machine)
{
1193
	struct perf_sample_id *sid = perf_evlist__id2sid(evlist, v->id);
1194 1195 1196 1197 1198 1199 1200 1201

	if (sid) {
		sample->id     = v->id;
		sample->period = v->value - sid->period;
		sid->period    = v->value;
	}

	if (!sid || sid->evsel == NULL) {
1202
		++evlist->stats.nr_unknown_id;
1203 1204 1205 1206 1207 1208
		return 0;
	}

	return tool->sample(tool, event, sample, sid->evsel, machine);
}

1209
static int deliver_sample_group(struct perf_evlist *evlist,
1210 1211 1212 1213 1214 1215 1216 1217 1218
				struct perf_tool *tool,
				union  perf_event *event,
				struct perf_sample *sample,
				struct machine *machine)
{
	int ret = -EINVAL;
	u64 i;

	for (i = 0; i < sample->read.group.nr; i++) {
1219
		ret = deliver_sample_value(evlist, tool, event, sample,
1220 1221 1222 1223 1224 1225 1226 1227 1228 1229
					   &sample->read.group.values[i],
					   machine);
		if (ret)
			break;
	}

	return ret;
}

static int
1230
 perf_evlist__deliver_sample(struct perf_evlist *evlist,
1231 1232 1233 1234 1235 1236 1237 1238 1239 1240
			     struct perf_tool *tool,
			     union  perf_event *event,
			     struct perf_sample *sample,
			     struct perf_evsel *evsel,
			     struct machine *machine)
{
	/* We know evsel != NULL. */
	u64 sample_type = evsel->attr.sample_type;
	u64 read_format = evsel->attr.read_format;

1241
	/* Standard sample delivery. */
1242 1243 1244 1245 1246
	if (!(sample_type & PERF_SAMPLE_READ))
		return tool->sample(tool, event, sample, evsel, machine);

	/* For PERF_SAMPLE_READ we have either single or group mode. */
	if (read_format & PERF_FORMAT_GROUP)
1247
		return deliver_sample_group(evlist, tool, event, sample,
1248 1249
					    machine);
	else
1250
		return deliver_sample_value(evlist, tool, event, sample,
1251 1252 1253
					    &sample->read.one, machine);
}

1254 1255 1256 1257 1258
static int machines__deliver_event(struct machines *machines,
				   struct perf_evlist *evlist,
				   union perf_event *event,
				   struct perf_sample *sample,
				   struct perf_tool *tool, u64 file_offset)
1259
{
1260
	struct perf_evsel *evsel;
1261
	struct machine *machine;
1262

1263
	dump_event(evlist, event, file_offset, sample);
1264

1265
	evsel = perf_evlist__id2evsel(evlist, sample->id);
1266

1267
	machine = machines__find_for_cpumode(machines, event, sample);
1268

1269 1270
	switch (event->header.type) {
	case PERF_RECORD_SAMPLE:
1271
		if (evsel == NULL) {
1272
			++evlist->stats.nr_unknown_id;
1273
			return 0;
1274
		}
1275
		dump_sample(evsel, event, sample);
1276
		if (machine == NULL) {
1277
			++evlist->stats.nr_unprocessable_samples;
1278
			return 0;
1279
		}
1280
		return perf_evlist__deliver_sample(evlist, tool, event, sample, evsel, machine);
1281
	case PERF_RECORD_MMAP:
1282
		return tool->mmap(tool, event, sample, machine);
1283
	case PERF_RECORD_MMAP2:
1284 1285
		if (event->header.misc & PERF_RECORD_MISC_PROC_MAP_PARSE_TIMEOUT)
			++evlist->stats.nr_proc_map_timeout;
1286
		return tool->mmap2(tool, event, sample, machine);
1287
	case PERF_RECORD_COMM:
1288
		return tool->comm(tool, event, sample, machine);
1289 1290
	case PERF_RECORD_NAMESPACES:
		return tool->namespaces(tool, event, sample, machine);
1291
	case PERF_RECORD_FORK:
1292
		return tool->fork(tool, event, sample, machine);
1293
	case PERF_RECORD_EXIT:
1294
		return tool->exit(tool, event, sample, machine);
1295
	case PERF_RECORD_LOST:
1296
		if (tool->lost == perf_event__process_lost)
1297
			evlist->stats.total_lost += event->lost.lost;
1298
		return tool->lost(tool, event, sample, machine);
1299 1300 1301 1302
	case PERF_RECORD_LOST_SAMPLES:
		if (tool->lost_samples == perf_event__process_lost_samples)
			evlist->stats.total_lost_samples += event->lost_samples.lost;
		return tool->lost_samples(tool, event, sample, machine);
1303
	case PERF_RECORD_READ:
J
Jiri Olsa 已提交
1304
		dump_read(evsel, event);
1305
		return tool->read(tool, event, sample, evsel, machine);
1306
	case PERF_RECORD_THROTTLE:
1307
		return tool->throttle(tool, event, sample, machine);
1308
	case PERF_RECORD_UNTHROTTLE:
1309
		return tool->unthrottle(tool, event, sample, machine);
1310
	case PERF_RECORD_AUX:
1311 1312 1313 1314 1315 1316
		if (tool->aux == perf_event__process_aux) {
			if (event->aux.flags & PERF_AUX_FLAG_TRUNCATED)
				evlist->stats.total_aux_lost += 1;
			if (event->aux.flags & PERF_AUX_FLAG_PARTIAL)
				evlist->stats.total_aux_partial += 1;
		}
1317
		return tool->aux(tool, event, sample, machine);
1318 1319
	case PERF_RECORD_ITRACE_START:
		return tool->itrace_start(tool, event, sample, machine);
1320 1321 1322
	case PERF_RECORD_SWITCH:
	case PERF_RECORD_SWITCH_CPU_WIDE:
		return tool->context_switch(tool, event, sample, machine);
1323
	default:
1324
		++evlist->stats.nr_unknown_events;
1325 1326 1327 1328
		return -1;
	}
}

1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346
static int perf_session__deliver_event(struct perf_session *session,
				       union perf_event *event,
				       struct perf_sample *sample,
				       struct perf_tool *tool,
				       u64 file_offset)
{
	int ret;

	ret = auxtrace__process_event(session, event, sample, tool);
	if (ret < 0)
		return ret;
	if (ret > 0)
		return 0;

	return machines__deliver_event(&session->machines, session->evlist,
				       event, sample, tool, file_offset);
}

1347 1348 1349
static s64 perf_session__process_user_event(struct perf_session *session,
					    union perf_event *event,
					    u64 file_offset)
1350
{
1351
	struct ordered_events *oe = &session->ordered_events;
1352
	struct perf_tool *tool = session->tool;
1353
	int fd = perf_data__fd(session->data);
1354 1355
	int err;

1356
	dump_event(session->evlist, event, file_offset, NULL);
1357

1358
	/* These events are processed right away */
1359
	switch (event->header.type) {
1360
	case PERF_RECORD_HEADER_ATTR:
1361
		err = tool->attr(tool, event, &session->evlist);
1362
		if (err == 0) {
1363
			perf_session__set_id_hdr_size(session);
1364 1365
			perf_session__set_comm_exec(session);
		}
1366
		return err;
1367 1368
	case PERF_RECORD_EVENT_UPDATE:
		return tool->event_update(tool, event, &session->evlist);
1369 1370 1371 1372 1373 1374
	case PERF_RECORD_HEADER_EVENT_TYPE:
		/*
		 * Depreceated, but we need to handle it for sake
		 * of old data files create in pipe mode.
		 */
		return 0;
1375 1376
	case PERF_RECORD_HEADER_TRACING_DATA:
		/* setup for reading amidst mmap */
1377
		lseek(fd, file_offset, SEEK_SET);
1378
		return tool->tracing_data(tool, event, session);
1379
	case PERF_RECORD_HEADER_BUILD_ID:
1380
		return tool->build_id(tool, event, session);
1381
	case PERF_RECORD_FINISHED_ROUND:
1382
		return tool->finished_round(tool, event, oe);
A
Adrian Hunter 已提交
1383 1384
	case PERF_RECORD_ID_INDEX:
		return tool->id_index(tool, event, session);
1385 1386 1387 1388 1389 1390
	case PERF_RECORD_AUXTRACE_INFO:
		return tool->auxtrace_info(tool, event, session);
	case PERF_RECORD_AUXTRACE:
		/* setup for reading amidst mmap */
		lseek(fd, file_offset + event->header.size, SEEK_SET);
		return tool->auxtrace(tool, event, session);
1391
	case PERF_RECORD_AUXTRACE_ERROR:
1392
		perf_session__auxtrace_error_inc(session, event);
1393
		return tool->auxtrace_error(tool, event, session);
1394 1395
	case PERF_RECORD_THREAD_MAP:
		return tool->thread_map(tool, event, session);
1396 1397
	case PERF_RECORD_CPU_MAP:
		return tool->cpu_map(tool, event, session);
1398 1399
	case PERF_RECORD_STAT_CONFIG:
		return tool->stat_config(tool, event, session);
J
Jiri Olsa 已提交
1400 1401
	case PERF_RECORD_STAT:
		return tool->stat(tool, event, session);
1402 1403
	case PERF_RECORD_STAT_ROUND:
		return tool->stat_round(tool, event, session);
1404 1405 1406
	case PERF_RECORD_TIME_CONV:
		session->time_conv = event->time_conv;
		return tool->time_conv(tool, event, session);
1407 1408
	case PERF_RECORD_HEADER_FEATURE:
		return tool->feature(tool, event, session);
1409
	default:
1410
		return -EINVAL;
1411
	}
1412 1413
}

1414 1415
int perf_session__deliver_synth_event(struct perf_session *session,
				      union perf_event *event,
1416
				      struct perf_sample *sample)
1417
{
1418
	struct perf_evlist *evlist = session->evlist;
1419
	struct perf_tool *tool = session->tool;
1420 1421

	events_stats__inc(&evlist->stats, event->header.type);
1422 1423

	if (event->header.type >= PERF_RECORD_USER_TYPE_START)
1424
		return perf_session__process_user_event(session, event, 0);
1425

1426
	return machines__deliver_event(&session->machines, evlist, event, sample, tool, 0);
1427 1428
}

1429 1430 1431 1432 1433 1434 1435 1436 1437
static void event_swap(union perf_event *event, bool sample_id_all)
{
	perf_event__swap_op swap;

	swap = perf_event__swap_ops[event->header.type];
	if (swap)
		swap(event, sample_id_all);
}

1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452
int perf_session__peek_event(struct perf_session *session, off_t file_offset,
			     void *buf, size_t buf_sz,
			     union perf_event **event_ptr,
			     struct perf_sample *sample)
{
	union perf_event *event;
	size_t hdr_sz, rest;
	int fd;

	if (session->one_mmap && !session->header.needs_swap) {
		event = file_offset - session->one_mmap_offset +
			session->one_mmap_addr;
		goto out_parse_sample;
	}

1453
	if (perf_data__is_pipe(session->data))
1454 1455
		return -1;

1456
	fd = perf_data__fd(session->data);
1457 1458 1459 1460 1461 1462
	hdr_sz = sizeof(struct perf_event_header);

	if (buf_sz < hdr_sz)
		return -1;

	if (lseek(fd, file_offset, SEEK_SET) == (off_t)-1 ||
1463
	    readn(fd, buf, hdr_sz) != (ssize_t)hdr_sz)
1464 1465 1466 1467 1468 1469 1470
		return -1;

	event = (union perf_event *)buf;

	if (session->header.needs_swap)
		perf_event_header__bswap(&event->header);

1471
	if (event->header.size < hdr_sz || event->header.size > buf_sz)
1472 1473 1474 1475
		return -1;

	rest = event->header.size - hdr_sz;

1476
	if (readn(fd, buf, rest) != (ssize_t)rest)
1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492
		return -1;

	if (session->header.needs_swap)
		event_swap(event, perf_evlist__sample_id_all(session->evlist));

out_parse_sample:

	if (sample && event->header.type < PERF_RECORD_USER_TYPE_START &&
	    perf_evlist__parse_sample(session->evlist, event, sample))
		return -1;

	*event_ptr = event;

	return 0;
}

1493
static s64 perf_session__process_event(struct perf_session *session,
1494
				       union perf_event *event, u64 file_offset)
1495
{
1496
	struct perf_evlist *evlist = session->evlist;
1497
	struct perf_tool *tool = session->tool;
1498
	struct perf_sample sample;
1499 1500
	int ret;

1501
	if (session->header.needs_swap)
1502
		event_swap(event, perf_evlist__sample_id_all(evlist));
1503 1504 1505 1506

	if (event->header.type >= PERF_RECORD_HEADER_MAX)
		return -EINVAL;

1507
	events_stats__inc(&evlist->stats, event->header.type);
1508 1509

	if (event->header.type >= PERF_RECORD_USER_TYPE_START)
1510
		return perf_session__process_user_event(session, event, file_offset);
1511

1512 1513 1514
	/*
	 * For all kernel events we get the sample data
	 */
1515
	ret = perf_evlist__parse_sample(evlist, event, &sample);
1516 1517
	if (ret)
		return ret;
1518

1519
	if (tool->ordered_events) {
1520
		ret = perf_session__queue_event(session, event, sample.time, file_offset);
1521 1522 1523 1524
		if (ret != -ETIME)
			return ret;
	}

1525 1526
	return perf_session__deliver_event(session, event, &sample, tool,
					   file_offset);
1527 1528
}

1529
void perf_event_header__bswap(struct perf_event_header *hdr)
1530
{
1531 1532 1533
	hdr->type = bswap_32(hdr->type);
	hdr->misc = bswap_16(hdr->misc);
	hdr->size = bswap_16(hdr->size);
1534 1535
}

1536 1537
struct thread *perf_session__findnew(struct perf_session *session, pid_t pid)
{
1538
	return machine__findnew_thread(&session->machines.host, -1, pid);
1539 1540
}

1541
int perf_session__register_idle_thread(struct perf_session *session)
1542
{
1543
	struct thread *thread;
1544
	int err = 0;
1545

1546
	thread = machine__findnew_thread(&session->machines.host, 0, 0);
1547
	if (thread == NULL || thread__set_comm(thread, "swapper", 0)) {
1548
		pr_err("problem inserting idle task.\n");
1549
		err = -1;
1550 1551
	}

1552 1553 1554 1555 1556
	if (thread == NULL || thread__set_namespaces(thread, 0, NULL)) {
		pr_err("problem inserting idle task.\n");
		err = -1;
	}

1557 1558 1559
	/* machine__findnew_thread() got the thread, so put it */
	thread__put(thread);
	return err;
1560 1561
}

1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579
static void
perf_session__warn_order(const struct perf_session *session)
{
	const struct ordered_events *oe = &session->ordered_events;
	struct perf_evsel *evsel;
	bool should_warn = true;

	evlist__for_each_entry(session->evlist, evsel) {
		if (evsel->attr.write_backward)
			should_warn = false;
	}

	if (!should_warn)
		return;
	if (oe->nr_unordered_events != 0)
		ui__warning("%u out of order events recorded.\n", oe->nr_unordered_events);
}

1580
static void perf_session__warn_about_errors(const struct perf_session *session)
1581
{
1582 1583 1584
	const struct events_stats *stats = &session->evlist->stats;

	if (session->tool->lost == perf_event__process_lost &&
1585
	    stats->nr_events[PERF_RECORD_LOST] != 0) {
1586 1587
		ui__warning("Processed %d events and lost %d chunks!\n\n"
			    "Check IO/CPU overload!\n\n",
1588 1589
			    stats->nr_events[0],
			    stats->nr_events[PERF_RECORD_LOST]);
1590 1591
	}

1592 1593 1594 1595 1596 1597
	if (session->tool->lost_samples == perf_event__process_lost_samples) {
		double drop_rate;

		drop_rate = (double)stats->total_lost_samples /
			    (double) (stats->nr_events[PERF_RECORD_SAMPLE] + stats->total_lost_samples);
		if (drop_rate > 0.05) {
1598
			ui__warning("Processed %" PRIu64 " samples and lost %3.2f%% samples!\n\n",
1599 1600 1601 1602 1603
				    stats->nr_events[PERF_RECORD_SAMPLE] + stats->total_lost_samples,
				    drop_rate * 100.0);
		}
	}

1604 1605 1606 1607 1608 1609 1610
	if (session->tool->aux == perf_event__process_aux &&
	    stats->total_aux_lost != 0) {
		ui__warning("AUX data lost %" PRIu64 " times out of %u!\n\n",
			    stats->total_aux_lost,
			    stats->nr_events[PERF_RECORD_AUX]);
	}

1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627
	if (session->tool->aux == perf_event__process_aux &&
	    stats->total_aux_partial != 0) {
		bool vmm_exclusive = false;

		(void)sysfs__read_bool("module/kvm_intel/parameters/vmm_exclusive",
		                       &vmm_exclusive);

		ui__warning("AUX data had gaps in it %" PRIu64 " times out of %u!\n\n"
		            "Are you running a KVM guest in the background?%s\n\n",
			    stats->total_aux_partial,
			    stats->nr_events[PERF_RECORD_AUX],
			    vmm_exclusive ?
			    "\nReloading kvm_intel module with vmm_exclusive=0\n"
			    "will reduce the gaps to only guest's timeslices." :
			    "");
	}

1628
	if (stats->nr_unknown_events != 0) {
1629 1630 1631 1632 1633
		ui__warning("Found %u unknown events!\n\n"
			    "Is this an older tool processing a perf.data "
			    "file generated by a more recent tool?\n\n"
			    "If that is not the case, consider "
			    "reporting to linux-kernel@vger.kernel.org.\n\n",
1634
			    stats->nr_unknown_events);
1635 1636
	}

1637
	if (stats->nr_unknown_id != 0) {
1638
		ui__warning("%u samples with id not present in the header\n",
1639
			    stats->nr_unknown_id);
1640 1641
	}

1642
	if (stats->nr_invalid_chains != 0) {
1643 1644 1645
		ui__warning("Found invalid callchains!\n\n"
			    "%u out of %u events were discarded for this reason.\n\n"
			    "Consider reporting to linux-kernel@vger.kernel.org.\n\n",
1646 1647
			    stats->nr_invalid_chains,
			    stats->nr_events[PERF_RECORD_SAMPLE]);
1648
	}
1649

1650
	if (stats->nr_unprocessable_samples != 0) {
1651 1652
		ui__warning("%u unprocessable samples recorded.\n"
			    "Do you have a KVM guest running and not using 'perf kvm'?\n",
1653
			    stats->nr_unprocessable_samples);
1654
	}
1655

1656
	perf_session__warn_order(session);
1657 1658

	events_stats__auxtrace_error_warn(stats);
1659 1660 1661 1662 1663 1664

	if (stats->nr_proc_map_timeout != 0) {
		ui__warning("%d map information files for pre-existing threads were\n"
			    "not processed, if there are samples for addresses they\n"
			    "will not be resolved, you may find out which are these\n"
			    "threads by running with -v and redirecting the output\n"
1665 1666 1667
			    "to a file.\n"
			    "The time limit to process proc map is too short?\n"
			    "Increase it by --proc-map-timeout\n",
1668 1669
			    stats->nr_proc_map_timeout);
	}
1670 1671
}

1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684
static int perf_session__flush_thread_stack(struct thread *thread,
					    void *p __maybe_unused)
{
	return thread_stack__flush(thread);
}

static int perf_session__flush_thread_stacks(struct perf_session *session)
{
	return machines__for_each_thread(&session->machines,
					 perf_session__flush_thread_stack,
					 NULL);
}

1685 1686
volatile int session_done;

1687
static int __perf_session__process_pipe_events(struct perf_session *session)
1688
{
1689
	struct ordered_events *oe = &session->ordered_events;
1690
	struct perf_tool *tool = session->tool;
1691
	int fd = perf_data__fd(session->data);
1692 1693 1694
	union perf_event *event;
	uint32_t size, cur_size = 0;
	void *buf = NULL;
1695
	s64 skip = 0;
1696
	u64 head;
1697
	ssize_t err;
1698 1699
	void *p;

1700
	perf_tool__fill_defaults(tool);
1701 1702

	head = 0;
1703 1704 1705 1706 1707
	cur_size = sizeof(union perf_event);

	buf = malloc(cur_size);
	if (!buf)
		return -errno;
1708
	ordered_events__set_copy_on_queue(oe, true);
1709
more:
1710
	event = buf;
1711
	err = readn(fd, event, sizeof(struct perf_event_header));
1712 1713 1714 1715 1716 1717 1718 1719
	if (err <= 0) {
		if (err == 0)
			goto done;

		pr_err("failed to read event header\n");
		goto out_err;
	}

1720
	if (session->header.needs_swap)
1721
		perf_event_header__bswap(&event->header);
1722

1723
	size = event->header.size;
1724 1725 1726 1727
	if (size < sizeof(struct perf_event_header)) {
		pr_err("bad event header size\n");
		goto out_err;
	}
1728

1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739
	if (size > cur_size) {
		void *new = realloc(buf, size);
		if (!new) {
			pr_err("failed to allocate memory to read event\n");
			goto out_err;
		}
		buf = new;
		cur_size = size;
		event = buf;
	}
	p = event;
1740 1741
	p += sizeof(struct perf_event_header);

1742
	if (size - sizeof(struct perf_event_header)) {
1743
		err = readn(fd, p, size - sizeof(struct perf_event_header));
1744 1745 1746 1747 1748
		if (err <= 0) {
			if (err == 0) {
				pr_err("unexpected end of event stream\n");
				goto done;
			}
1749

1750 1751 1752
			pr_err("failed to read event data\n");
			goto out_err;
		}
1753 1754
	}

1755
	if ((skip = perf_session__process_event(session, event, head)) < 0) {
1756
		pr_err("%#" PRIx64 " [%#x]: failed to process type: %d\n",
1757
		       head, event->header.size, event->header.type);
1758 1759
		err = -EINVAL;
		goto out_err;
1760 1761 1762 1763 1764 1765 1766 1767 1768 1769
	}

	head += size;

	if (skip > 0)
		head += skip;

	if (!session_done())
		goto more;
done:
1770
	/* do the final flush for ordered samples */
1771
	err = ordered_events__flush(oe, OE_FLUSH__FINAL);
1772 1773 1774
	if (err)
		goto out_err;
	err = auxtrace__flush_events(session, tool);
1775 1776 1777
	if (err)
		goto out_err;
	err = perf_session__flush_thread_stacks(session);
1778
out_err:
1779
	free(buf);
1780
	perf_session__warn_about_errors(session);
1781
	ordered_events__free(&session->ordered_events);
1782
	auxtrace__free_events(session);
1783 1784 1785
	return err;
}

1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803
static union perf_event *
fetch_mmaped_event(struct perf_session *session,
		   u64 head, size_t mmap_size, char *buf)
{
	union perf_event *event;

	/*
	 * Ensure we have enough space remaining to read
	 * the size of the event in the headers.
	 */
	if (head + sizeof(event->header) > mmap_size)
		return NULL;

	event = (union perf_event *)(buf + head);

	if (session->header.needs_swap)
		perf_event_header__bswap(&event->header);

1804 1805 1806 1807
	if (head + event->header.size > mmap_size) {
		/* We're not fetching the event so swap back again */
		if (session->header.needs_swap)
			perf_event_header__bswap(&event->header);
1808
		return NULL;
1809
	}
1810 1811 1812 1813

	return event;
}

1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825
/*
 * On 64bit we can mmap the data file in one go. No need for tiny mmap
 * slices. On 32bit we use 32MB.
 */
#if BITS_PER_LONG == 64
#define MMAP_SIZE ULLONG_MAX
#define NUM_MMAPS 1
#else
#define MMAP_SIZE (32 * 1024 * 1024ULL)
#define NUM_MMAPS 128
#endif

1826 1827
static int __perf_session__process_events(struct perf_session *session,
					  u64 data_offset, u64 data_size,
1828
					  u64 file_size)
1829
{
1830
	struct ordered_events *oe = &session->ordered_events;
1831
	struct perf_tool *tool = session->tool;
1832
	int fd = perf_data__fd(session->data);
1833
	u64 head, page_offset, file_offset, file_pos, size;
1834
	int err, mmap_prot, mmap_flags, map_idx = 0;
1835
	size_t	mmap_size;
1836
	char *buf, *mmaps[NUM_MMAPS];
1837
	union perf_event *event;
1838
	struct ui_progress prog;
1839
	s64 skip;
1840

1841
	perf_tool__fill_defaults(tool);
1842

1843 1844 1845
	page_offset = page_size * (data_offset / page_size);
	file_offset = page_offset;
	head = data_offset - page_offset;
1846

1847 1848 1849 1850
	if (data_size == 0)
		goto out;

	if (data_offset + data_size < file_size)
1851 1852
		file_size = data_offset + data_size;

1853
	ui_progress__init_size(&prog, file_size, "Processing events...");
1854

1855
	mmap_size = MMAP_SIZE;
1856
	if (mmap_size > file_size) {
1857
		mmap_size = file_size;
1858 1859
		session->one_mmap = true;
	}
1860

1861 1862
	memset(mmaps, 0, sizeof(mmaps));

1863 1864 1865
	mmap_prot  = PROT_READ;
	mmap_flags = MAP_SHARED;

1866
	if (session->header.needs_swap) {
1867 1868 1869
		mmap_prot  |= PROT_WRITE;
		mmap_flags = MAP_PRIVATE;
	}
1870
remap:
1871
	buf = mmap(NULL, mmap_size, mmap_prot, mmap_flags, fd,
1872
		   file_offset);
1873 1874 1875 1876 1877
	if (buf == MAP_FAILED) {
		pr_err("failed to mmap file\n");
		err = -errno;
		goto out_err;
	}
1878 1879
	mmaps[map_idx] = buf;
	map_idx = (map_idx + 1) & (ARRAY_SIZE(mmaps) - 1);
1880
	file_pos = file_offset + head;
1881 1882 1883 1884
	if (session->one_mmap) {
		session->one_mmap_addr = buf;
		session->one_mmap_offset = file_offset;
	}
1885 1886

more:
1887 1888
	event = fetch_mmaped_event(session, head, mmap_size, buf);
	if (!event) {
1889 1890 1891 1892
		if (mmaps[map_idx]) {
			munmap(mmaps[map_idx], mmap_size);
			mmaps[map_idx] = NULL;
		}
1893

1894 1895 1896
		page_offset = page_size * (head / page_size);
		file_offset += page_offset;
		head -= page_offset;
1897 1898 1899 1900 1901
		goto remap;
	}

	size = event->header.size;

1902
	if (size < sizeof(struct perf_event_header) ||
1903
	    (skip = perf_session__process_event(session, event, file_pos)) < 0) {
1904 1905 1906 1907 1908
		pr_err("%#" PRIx64 " [%#x]: failed to process type: %d\n",
		       file_offset + head, event->header.size,
		       event->header.type);
		err = -EINVAL;
		goto out_err;
1909 1910
	}

1911 1912 1913
	if (skip)
		size += skip;

1914
	head += size;
1915
	file_pos += size;
1916

1917
	ui_progress__update(&prog, size);
1918

1919
	if (session_done())
1920
		goto out;
1921

1922
	if (file_pos < file_size)
1923
		goto more;
1924

1925
out:
1926
	/* do the final flush for ordered samples */
1927
	err = ordered_events__flush(oe, OE_FLUSH__FINAL);
1928 1929 1930
	if (err)
		goto out_err;
	err = auxtrace__flush_events(session, tool);
1931 1932 1933
	if (err)
		goto out_err;
	err = perf_session__flush_thread_stacks(session);
1934
out_err:
N
Namhyung Kim 已提交
1935
	ui_progress__finish();
1936
	perf_session__warn_about_errors(session);
1937 1938 1939 1940 1941
	/*
	 * We may switching perf.data output, make ordered_events
	 * reusable.
	 */
	ordered_events__reinit(&session->ordered_events);
1942
	auxtrace__free_events(session);
1943
	session->one_mmap = false;
1944 1945
	return err;
}
1946

1947
int perf_session__process_events(struct perf_session *session)
1948
{
1949
	u64 size = perf_data__size(session->data);
1950 1951
	int err;

1952
	if (perf_session__register_idle_thread(session) < 0)
1953 1954
		return -ENOMEM;

1955
	if (!perf_data__is_pipe(session->data))
1956 1957
		err = __perf_session__process_events(session,
						     session->header.data_offset,
1958
						     session->header.data_size, size);
1959
	else
1960
		err = __perf_session__process_pipe_events(session);
1961

1962 1963 1964
	return err;
}

1965
bool perf_session__has_traces(struct perf_session *session, const char *msg)
1966
{
1967 1968
	struct perf_evsel *evsel;

1969
	evlist__for_each_entry(session->evlist, evsel) {
1970 1971
		if (evsel->attr.type == PERF_TYPE_TRACEPOINT)
			return true;
1972 1973
	}

1974 1975
	pr_err("No trace sample to read. Did you call 'perf %s'?\n", msg);
	return false;
1976
}
1977

1978 1979
int maps__set_kallsyms_ref_reloc_sym(struct map **maps,
				     const char *symbol_name, u64 addr)
1980 1981
{
	char *bracket;
1982
	int i;
1983 1984 1985 1986 1987
	struct ref_reloc_sym *ref;

	ref = zalloc(sizeof(struct ref_reloc_sym));
	if (ref == NULL)
		return -ENOMEM;
1988

1989 1990 1991
	ref->name = strdup(symbol_name);
	if (ref->name == NULL) {
		free(ref);
1992
		return -ENOMEM;
1993
	}
1994

1995
	bracket = strchr(ref->name, ']');
1996 1997 1998
	if (bracket)
		*bracket = '\0';

1999
	ref->addr = addr;
2000 2001

	for (i = 0; i < MAP__NR_TYPES; ++i) {
2002
		struct kmap *kmap = map__kmap(maps[i]);
2003 2004 2005

		if (!kmap)
			continue;
2006
		kmap->ref_reloc_sym = ref;
2007 2008
	}

2009 2010
	return 0;
}
2011

2012
size_t perf_session__fprintf_dsos(struct perf_session *session, FILE *fp)
2013
{
2014
	return machines__fprintf_dsos(&session->machines, fp);
2015
}
2016

2017
size_t perf_session__fprintf_dsos_buildid(struct perf_session *session, FILE *fp,
2018
					  bool (skip)(struct dso *dso, int parm), int parm)
2019
{
2020
	return machines__fprintf_dsos_buildid(&session->machines, fp, skip, parm);
2021
}
2022 2023 2024

size_t perf_session__fprintf_nr_events(struct perf_session *session, FILE *fp)
{
2025 2026 2027 2028 2029 2030
	size_t ret;
	const char *msg = "";

	if (perf_header__has_feat(&session->header, HEADER_AUXTRACE))
		msg = " (excludes AUX area (e.g. instruction trace) decoded / synthesized events)";

2031
	ret = fprintf(fp, "\nAggregated stats:%s\n", msg);
2032

2033
	ret += events_stats__fprintf(&session->evlist->stats, fp);
2034 2035
	return ret;
}
2036

2037 2038 2039 2040 2041 2042
size_t perf_session__fprintf(struct perf_session *session, FILE *fp)
{
	/*
	 * FIXME: Here we have to actually print all the machines in this
	 * session, not just the host...
	 */
2043
	return machine__fprintf(&session->machines.host, fp);
2044 2045
}

2046 2047 2048 2049 2050
struct perf_evsel *perf_session__find_first_evtype(struct perf_session *session,
					      unsigned int type)
{
	struct perf_evsel *pos;

2051
	evlist__for_each_entry(session->evlist, pos) {
2052 2053 2054 2055 2056 2057
		if (pos->attr.type == type)
			return pos;
	}
	return NULL;
}

2058 2059 2060
int perf_session__cpu_bitmap(struct perf_session *session,
			     const char *cpu_list, unsigned long *cpu_bitmap)
{
2061
	int i, err = -1;
2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072
	struct cpu_map *map;

	for (i = 0; i < PERF_TYPE_MAX; ++i) {
		struct perf_evsel *evsel;

		evsel = perf_session__find_first_evtype(session, i);
		if (!evsel)
			continue;

		if (!(evsel->attr.sample_type & PERF_SAMPLE_CPU)) {
			pr_err("File does not contain CPU events. "
2073
			       "Remove -C option to proceed.\n");
2074 2075 2076 2077 2078
			return -1;
		}
	}

	map = cpu_map__new(cpu_list);
2079 2080 2081 2082
	if (map == NULL) {
		pr_err("Invalid cpu_list\n");
		return -1;
	}
2083 2084 2085 2086 2087 2088 2089

	for (i = 0; i < map->nr; i++) {
		int cpu = map->map[i];

		if (cpu >= MAX_NR_CPUS) {
			pr_err("Requested CPU %d too large. "
			       "Consider raising MAX_NR_CPUS\n", cpu);
2090
			goto out_delete_map;
2091 2092 2093 2094 2095
		}

		set_bit(cpu, cpu_bitmap);
	}

2096 2097 2098
	err = 0;

out_delete_map:
2099
	cpu_map__put(map);
2100
	return err;
2101
}
2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112

void perf_session__fprintf_info(struct perf_session *session, FILE *fp,
				bool full)
{
	if (session == NULL || fp == NULL)
		return;

	fprintf(fp, "# ========\n");
	perf_header__fprintf_info(session, fp, full);
	fprintf(fp, "# ========\n#\n");
}
2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123


int __perf_session__set_tracepoints_handlers(struct perf_session *session,
					     const struct perf_evsel_str_handler *assocs,
					     size_t nr_assocs)
{
	struct perf_evsel *evsel;
	size_t i;
	int err;

	for (i = 0; i < nr_assocs; i++) {
2124 2125 2126 2127 2128
		/*
		 * Adding a handler for an event not in the session,
		 * just ignore it.
		 */
		evsel = perf_evlist__find_tracepoint_by_name(session->evlist, assocs[i].name);
2129
		if (evsel == NULL)
2130
			continue;
2131 2132

		err = -EEXIST;
2133
		if (evsel->handler != NULL)
2134
			goto out;
2135
		evsel->handler = assocs[i].handler;
2136 2137 2138 2139 2140 2141
	}

	err = 0;
out:
	return err;
}
A
Adrian Hunter 已提交
2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195

int perf_event__process_id_index(struct perf_tool *tool __maybe_unused,
				 union perf_event *event,
				 struct perf_session *session)
{
	struct perf_evlist *evlist = session->evlist;
	struct id_index_event *ie = &event->id_index;
	size_t i, nr, max_nr;

	max_nr = (ie->header.size - sizeof(struct id_index_event)) /
		 sizeof(struct id_index_entry);
	nr = ie->nr;
	if (nr > max_nr)
		return -EINVAL;

	if (dump_trace)
		fprintf(stdout, " nr: %zu\n", nr);

	for (i = 0; i < nr; i++) {
		struct id_index_entry *e = &ie->entries[i];
		struct perf_sample_id *sid;

		if (dump_trace) {
			fprintf(stdout,	" ... id: %"PRIu64, e->id);
			fprintf(stdout,	"  idx: %"PRIu64, e->idx);
			fprintf(stdout,	"  cpu: %"PRId64, e->cpu);
			fprintf(stdout,	"  tid: %"PRId64"\n", e->tid);
		}

		sid = perf_evlist__id2sid(evlist, e->id);
		if (!sid)
			return -ENOENT;
		sid->idx = e->idx;
		sid->cpu = e->cpu;
		sid->tid = e->tid;
	}
	return 0;
}

int perf_event__synthesize_id_index(struct perf_tool *tool,
				    perf_event__handler_t process,
				    struct perf_evlist *evlist,
				    struct machine *machine)
{
	union perf_event *ev;
	struct perf_evsel *evsel;
	size_t nr = 0, i = 0, sz, max_nr, n;
	int err;

	pr_debug2("Synthesizing id index\n");

	max_nr = (UINT16_MAX - sizeof(struct id_index_event)) /
		 sizeof(struct id_index_entry);

2196
	evlist__for_each_entry(evlist, evsel)
A
Adrian Hunter 已提交
2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208
		nr += evsel->ids;

	n = nr > max_nr ? max_nr : nr;
	sz = sizeof(struct id_index_event) + n * sizeof(struct id_index_entry);
	ev = zalloc(sz);
	if (!ev)
		return -ENOMEM;

	ev->id_index.header.type = PERF_RECORD_ID_INDEX;
	ev->id_index.header.size = sz;
	ev->id_index.nr = n;

2209
	evlist__for_each_entry(evlist, evsel) {
A
Adrian Hunter 已提交
2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249
		u32 j;

		for (j = 0; j < evsel->ids; j++) {
			struct id_index_entry *e;
			struct perf_sample_id *sid;

			if (i >= n) {
				err = process(tool, ev, NULL, machine);
				if (err)
					goto out_err;
				nr -= n;
				i = 0;
			}

			e = &ev->id_index.entries[i++];

			e->id = evsel->id[j];

			sid = perf_evlist__id2sid(evlist, e->id);
			if (!sid) {
				free(ev);
				return -ENOENT;
			}

			e->idx = sid->idx;
			e->cpu = sid->cpu;
			e->tid = sid->tid;
		}
	}

	sz = sizeof(struct id_index_event) + nr * sizeof(struct id_index_entry);
	ev->id_index.header.size = sz;
	ev->id_index.nr = nr;

	err = process(tool, ev, NULL, machine);
out_err:
	free(ev);

	return err;
}