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

7
#include <byteswap.h>
8 9
#include <unistd.h>
#include <sys/types.h>
10
#include <sys/mman.h>
11

12 13
#include "evlist.h"
#include "evsel.h"
14
#include "memswap.h"
15
#include "session.h"
16
#include "tool.h"
17
#include "sort.h"
18
#include "util.h"
19
#include "cpumap.h"
20
#include "perf_regs.h"
21
#include "asm/bug.h"
22
#include "auxtrace.h"
23
#include "thread.h"
24
#include "thread-stack.h"
25
#include "stat.h"
26

27 28 29 30 31
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);
32

33
static int perf_session__open(struct perf_session *session)
34
{
35
	struct perf_data_file *file = session->file;
36

37
	if (perf_session__read_header(session) < 0) {
38
		pr_err("incompatible file format (rerun with -v to learn more)\n");
39
		return -1;
40 41
	}

42 43 44
	if (perf_data_file__is_pipe(file))
		return 0;

45 46 47
	if (perf_header__has_feat(&session->header, HEADER_STAT))
		return 0;

48
	if (!perf_evlist__valid_sample_type(session->evlist)) {
49
		pr_err("non matching sample_type\n");
50
		return -1;
51 52
	}

53
	if (!perf_evlist__valid_sample_id_all(session->evlist)) {
54
		pr_err("non matching sample_id_all\n");
55
		return -1;
56 57
	}

58
	if (!perf_evlist__valid_read_format(session->evlist)) {
59
		pr_err("non matching read_format\n");
60
		return -1;
61 62
	}

63 64 65
	return 0;
}

66
void perf_session__set_id_hdr_size(struct perf_session *session)
67
{
68 69 70
	u16 id_hdr_size = perf_evlist__id_hdr_size(session->evlist);

	machines__set_id_hdr_size(&session->machines, id_hdr_size);
71 72
}

73
int perf_session__create_kernel_maps(struct perf_session *session)
74
{
75
	int ret = machine__create_kernel_maps(&session->machines.host);
76 77

	if (ret >= 0)
78
		ret = machines__create_guest_kernel_maps(&session->machines);
79 80 81
	return ret;
}

82
static void perf_session__destroy_kernel_maps(struct perf_session *session)
83
{
84
	machines__destroy_kernel_maps(&session->machines);
85 86
}

87 88 89 90
static bool perf_session__has_comm_exec(struct perf_session *session)
{
	struct perf_evsel *evsel;

91
	evlist__for_each_entry(session->evlist, evsel) {
92 93 94 95 96 97 98 99 100 101 102 103 104 105
		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);
}

106
static int ordered_events__deliver_event(struct ordered_events *oe,
107
					 struct ordered_event *event)
108
{
109 110 111 112 113 114 115 116 117 118
	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;
	}

119 120
	return perf_session__deliver_event(session, event->event, &sample,
					   session->tool, event->file_offset);
121 122
}

123 124
struct perf_session *perf_session__new(struct perf_data_file *file,
				       bool repipe, struct perf_tool *tool)
125
{
126
	struct perf_session *session = zalloc(sizeof(*session));
127

128
	if (!session)
129 130
		goto out;

131
	session->repipe = repipe;
132
	session->tool   = tool;
133
	INIT_LIST_HEAD(&session->auxtrace_index);
134
	machines__init(&session->machines);
135
	ordered_events__init(&session->ordered_events, ordered_events__deliver_event);
136

137 138
	if (file) {
		if (perf_data_file__open(file))
139
			goto out_delete;
140

141
		session->file = file;
142 143

		if (perf_data_file__is_read(file)) {
144
			if (perf_session__open(session) < 0)
145 146
				goto out_close;

147 148 149 150 151 152 153 154
			/*
			 * set session attributes that are present in perf.data
			 * but not in pipe-mode.
			 */
			if (!file->is_pipe) {
				perf_session__set_id_hdr_size(session);
				perf_session__set_comm_exec(session);
			}
155
		}
156 157
	} else  {
		session->machines.host.env = &perf_env;
158 159 160
	}

	if (!file || perf_data_file__is_write(file)) {
161 162
		/*
		 * In O_RDONLY mode this will be performed when reading the
163
		 * kernel MMAP event, in perf_event__process_mmap().
164
		 */
165
		if (perf_session__create_kernel_maps(session) < 0)
166
			pr_warning("Cannot read kernel map\n");
167
	}
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.
	 */
	if ((!file || !file->is_pipe) && tool && tool->ordering_requires_timestamps &&
174
	    tool->ordered_events && !perf_evlist__sample_id_all(session->evlist)) {
175
		dump_printf("WARNING: No sample_id_all support, falling back to unordered processing\n");
176
		tool->ordered_events = false;
177
	}
178

179
	return session;
180 181 182 183

 out_close:
	perf_data_file__close(file);
 out_delete:
184
	perf_session__delete(session);
185
 out:
186
	return NULL;
187 188
}

189 190
static void perf_session__delete_threads(struct perf_session *session)
{
191
	machine__delete_threads(&session->machines.host);
192 193
}

194
void perf_session__delete(struct perf_session *session)
195
{
196 197
	if (session == NULL)
		return;
198
	auxtrace__free(session);
199
	auxtrace_index__free(&session->auxtrace_index);
200 201
	perf_session__destroy_kernel_maps(session);
	perf_session__delete_threads(session);
202
	perf_env__exit(&session->header.env);
203 204 205 206
	machines__exit(&session->machines);
	if (session->file)
		perf_data_file__close(session->file);
	free(session);
207
}
208

209 210 211
static int process_event_synth_tracing_data_stub(struct perf_tool *tool
						 __maybe_unused,
						 union perf_event *event
212 213 214
						 __maybe_unused,
						 struct perf_session *session
						__maybe_unused)
215 216 217 218 219
{
	dump_printf(": unhandled!\n");
	return 0;
}

220 221
static int process_event_synth_attr_stub(struct perf_tool *tool __maybe_unused,
					 union perf_event *event __maybe_unused,
222 223
					 struct perf_evlist **pevlist
					 __maybe_unused)
224 225 226 227 228
{
	dump_printf(": unhandled!\n");
	return 0;
}

229 230 231 232 233
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)
{
234 235 236
	if (dump_trace)
		perf_event__fprintf_event_update(event, stdout);

237 238 239 240
	dump_printf(": unhandled!\n");
	return 0;
}

241 242 243 244 245
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)
246 247 248 249 250
{
	dump_printf(": unhandled!\n");
	return 0;
}

251 252 253 254
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)
255 256 257 258 259
{
	dump_printf(": unhandled!\n");
	return 0;
}

260 261
static int process_finished_round_stub(struct perf_tool *tool __maybe_unused,
				       union perf_event *event __maybe_unused,
262
				       struct ordered_events *oe __maybe_unused)
263 264 265 266 267
{
	dump_printf(": unhandled!\n");
	return 0;
}

268
static int process_finished_round(struct perf_tool *tool,
269
				  union perf_event *event,
270
				  struct ordered_events *oe);
271

272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297
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");
	if (perf_data_file__is_pipe(session->file))
		skipn(perf_data_file__fd(session->file), event->auxtrace.size);
	return event->auxtrace.size;
}

298 299 300
static int process_event_op2_stub(struct perf_tool *tool __maybe_unused,
				  union perf_event *event __maybe_unused,
				  struct perf_session *session __maybe_unused)
301 302 303 304 305
{
	dump_printf(": unhandled!\n");
	return 0;
}

306 307 308 309 310 311

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)
{
312 313 314
	if (dump_trace)
		perf_event__fprintf_thread_map(event, stdout);

315 316 317 318
	dump_printf(": unhandled!\n");
	return 0;
}

319 320 321 322 323
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)
{
324 325 326
	if (dump_trace)
		perf_event__fprintf_cpu_map(event, stdout);

327 328 329 330
	dump_printf(": unhandled!\n");
	return 0;
}

331 332 333 334 335
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)
{
336 337 338
	if (dump_trace)
		perf_event__fprintf_stat_config(event, stdout);

339 340 341 342
	dump_printf(": unhandled!\n");
	return 0;
}

J
Jiri Olsa 已提交
343 344 345 346 347
static int process_stat_stub(struct perf_tool *tool __maybe_unused,
			     union perf_event *event __maybe_unused,
			     struct perf_session *perf_session
			     __maybe_unused)
{
348 349 350
	if (dump_trace)
		perf_event__fprintf_stat(event, stdout);

J
Jiri Olsa 已提交
351 352 353 354
	dump_printf(": unhandled!\n");
	return 0;
}

355 356 357 358 359
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)
{
360 361 362
	if (dump_trace)
		perf_event__fprintf_stat_round(event, stdout);

363 364 365 366
	dump_printf(": unhandled!\n");
	return 0;
}

367
void perf_tool__fill_defaults(struct perf_tool *tool)
368
{
369 370 371 372
	if (tool->sample == NULL)
		tool->sample = process_event_sample_stub;
	if (tool->mmap == NULL)
		tool->mmap = process_event_stub;
373 374
	if (tool->mmap2 == NULL)
		tool->mmap2 = process_event_stub;
375 376 377 378 379 380 381 382
	if (tool->comm == NULL)
		tool->comm = process_event_stub;
	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;
383 384
	if (tool->lost_samples == NULL)
		tool->lost_samples = perf_event__process_lost_samples;
385 386
	if (tool->aux == NULL)
		tool->aux = perf_event__process_aux;
387 388
	if (tool->itrace_start == NULL)
		tool->itrace_start = perf_event__process_itrace_start;
389 390
	if (tool->context_switch == NULL)
		tool->context_switch = perf_event__process_switch;
391 392 393 394 395 396 397 398
	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;
399 400
	if (tool->event_update == NULL)
		tool->event_update = process_event_synth_event_update_stub;
401 402 403
	if (tool->tracing_data == NULL)
		tool->tracing_data = process_event_synth_tracing_data_stub;
	if (tool->build_id == NULL)
404
		tool->build_id = process_event_op2_stub;
405
	if (tool->finished_round == NULL) {
406
		if (tool->ordered_events)
407
			tool->finished_round = process_finished_round;
408
		else
409
			tool->finished_round = process_finished_round_stub;
410
	}
A
Adrian Hunter 已提交
411
	if (tool->id_index == NULL)
412
		tool->id_index = process_event_op2_stub;
413
	if (tool->auxtrace_info == NULL)
414
		tool->auxtrace_info = process_event_op2_stub;
415 416
	if (tool->auxtrace == NULL)
		tool->auxtrace = process_event_auxtrace_stub;
417
	if (tool->auxtrace_error == NULL)
418
		tool->auxtrace_error = process_event_op2_stub;
419 420
	if (tool->thread_map == NULL)
		tool->thread_map = process_event_thread_map_stub;
421 422
	if (tool->cpu_map == NULL)
		tool->cpu_map = process_event_cpu_map_stub;
423 424
	if (tool->stat_config == NULL)
		tool->stat_config = process_event_stat_config_stub;
J
Jiri Olsa 已提交
425 426
	if (tool->stat == NULL)
		tool->stat = process_stat_stub;
427 428
	if (tool->stat_round == NULL)
		tool->stat_round = process_stat_round_stub;
429 430
	if (tool->time_conv == NULL)
		tool->time_conv = process_event_op2_stub;
431 432
	if (tool->feature == NULL)
		tool->feature = process_event_op2_stub;
433
}
434

435 436 437 438 439 440 441 442 443 444
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,
445
				   bool sample_id_all __maybe_unused)
446
{
447 448
	struct perf_event_header *hdr = &event->header;
	mem_bswap_64(hdr + 1, event->header.size - sizeof(*hdr));
449 450
}

451
static void perf_event__comm_swap(union perf_event *event, bool sample_id_all)
452
{
453 454
	event->comm.pid = bswap_32(event->comm.pid);
	event->comm.tid = bswap_32(event->comm.tid);
455 456 457 458

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

459
		data += PERF_ALIGN(strlen(data) + 1, sizeof(u64));
460 461
		swap_sample_id_all(event, data);
	}
462 463
}

464 465
static void perf_event__mmap_swap(union perf_event *event,
				  bool sample_id_all)
466
{
467 468 469 470 471
	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);
472 473 474 475

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

476
		data += PERF_ALIGN(strlen(data) + 1, sizeof(u64));
477 478
		swap_sample_id_all(event, data);
	}
479 480
}

481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499
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);
	}
}
500
static void perf_event__task_swap(union perf_event *event, bool sample_id_all)
501
{
502 503 504 505 506
	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);
507 508 509

	if (sample_id_all)
		swap_sample_id_all(event, &event->fork + 1);
510 511
}

512
static void perf_event__read_swap(union perf_event *event, bool sample_id_all)
513
{
514 515 516 517 518 519
	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);
520 521 522

	if (sample_id_all)
		swap_sample_id_all(event, &event->read + 1);
523 524
}

525 526 527 528 529 530 531 532 533 534
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);
}

535 536 537 538 539 540 541 542 543 544
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);
}

545 546 547 548 549 550 551 552 553 554 555 556 557
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);
}

558 559 560 561 562 563 564 565 566 567 568
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);
}

569 570 571 572 573 574 575 576 577 578
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
579
 * through endian village. ABI says:
580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600
 *
 * 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++;
	}
}

601 602 603 604 605
/* 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);
606 607 608 609 610 611 612 613 614

#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)
615
#define bswap_field_16(f) bswap_field(f, 16)
616 617 618 619 620 621 622 623 624 625 626 627 628 629 630
#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);
631
	bswap_field_16(sample_max_stack);
632 633 634 635 636 637 638 639 640 641 642 643

	/*
	 * 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
644 645
}

646
static void perf_event__hdr_attr_swap(union perf_event *event,
647
				      bool sample_id_all __maybe_unused)
648 649 650
{
	size_t size;

651
	perf_event__attr_swap(&event->attr.attr);
652

653 654 655
	size = event->header.size;
	size -= (void *)&event->attr.id - (void *)event;
	mem_bswap_64(event->attr.id, size);
656 657
}

658 659 660 661 662 663 664
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);
}

665
static void perf_event__event_type_swap(union perf_event *event,
666
					bool sample_id_all __maybe_unused)
667
{
668 669
	event->event_type.event_type.event_id =
		bswap_64(event->event_type.event_type.event_id);
670 671
}

672
static void perf_event__tracing_data_swap(union perf_event *event,
673
					  bool sample_id_all __maybe_unused)
674
{
675
	event->tracing_data.size = bswap_32(event->tracing_data.size);
676 677
}

678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700
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);
}

701 702 703 704 705 706 707 708 709 710 711
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);
}

712 713 714 715 716 717 718 719 720 721 722
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);
}

723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758
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;
	}
}

759 760 761 762 763 764 765 766 767 768
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
769 770 771 772 773 774 775 776 777 778 779
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);
}

780 781 782 783 784 785 786
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);
}

787 788
typedef void (*perf_event__swap_op)(union perf_event *event,
				    bool sample_id_all);
789

790 791
static perf_event__swap_op perf_event__swap_ops[] = {
	[PERF_RECORD_MMAP]		  = perf_event__mmap_swap,
792
	[PERF_RECORD_MMAP2]		  = perf_event__mmap2_swap,
793 794 795 796 797
	[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,
798 799
	[PERF_RECORD_THROTTLE]		  = perf_event__throttle_swap,
	[PERF_RECORD_UNTHROTTLE]	  = perf_event__throttle_swap,
800
	[PERF_RECORD_SAMPLE]		  = perf_event__all64_swap,
801
	[PERF_RECORD_AUX]		  = perf_event__aux_swap,
802
	[PERF_RECORD_ITRACE_START]	  = perf_event__itrace_start_swap,
803
	[PERF_RECORD_LOST_SAMPLES]	  = perf_event__all64_swap,
804 805
	[PERF_RECORD_SWITCH]		  = perf_event__switch_swap,
	[PERF_RECORD_SWITCH_CPU_WIDE]	  = perf_event__switch_swap,
806
	[PERF_RECORD_HEADER_ATTR]	  = perf_event__hdr_attr_swap,
807 808 809
	[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
810
	[PERF_RECORD_ID_INDEX]		  = perf_event__all64_swap,
811 812
	[PERF_RECORD_AUXTRACE_INFO]	  = perf_event__auxtrace_info_swap,
	[PERF_RECORD_AUXTRACE]		  = perf_event__auxtrace_swap,
813
	[PERF_RECORD_AUXTRACE_ERROR]	  = perf_event__auxtrace_error_swap,
814
	[PERF_RECORD_THREAD_MAP]	  = perf_event__thread_map_swap,
815
	[PERF_RECORD_CPU_MAP]		  = perf_event__cpu_map_swap,
816
	[PERF_RECORD_STAT_CONFIG]	  = perf_event__stat_config_swap,
J
Jiri Olsa 已提交
817
	[PERF_RECORD_STAT]		  = perf_event__stat_swap,
818
	[PERF_RECORD_STAT_ROUND]	  = perf_event__stat_round_swap,
819
	[PERF_RECORD_EVENT_UPDATE]	  = perf_event__event_update_swap,
820
	[PERF_RECORD_TIME_CONV]		  = perf_event__all64_swap,
821
	[PERF_RECORD_HEADER_MAX]	  = NULL,
822 823
};

824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862
/*
 * 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...
 */
863
static int process_finished_round(struct perf_tool *tool __maybe_unused,
864
				  union perf_event *event __maybe_unused,
865
				  struct ordered_events *oe)
866
{
867 868
	if (dump_trace)
		fprintf(stdout, "\n");
869
	return ordered_events__flush(oe, OE_FLUSH__ROUND);
870 871
}

872 873
int perf_session__queue_event(struct perf_session *s, union perf_event *event,
			      struct perf_sample *sample, u64 file_offset)
874
{
875
	return ordered_events__queue(&s->ordered_events, event, sample, file_offset);
876
}
877

K
Kan Liang 已提交
878
static void callchain__lbr_callstack_printf(struct perf_sample *sample)
879
{
K
Kan Liang 已提交
880 881 882
	struct ip_callchain *callchain = sample->callchain;
	struct branch_stack *lbr_stack = sample->branch_stack;
	u64 kernel_callchain_nr = callchain->nr;
883
	unsigned int i;
884

K
Kan Liang 已提交
885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930
	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;

931
	if (perf_evsel__has_branch_callstack(evsel))
K
Kan Liang 已提交
932 933 934
		callchain__lbr_callstack_printf(sample);

	printf("... FP chain: nr:%" PRIu64 "\n", callchain->nr);
935

K
Kan Liang 已提交
936
	for (i = 0; i < callchain->nr; i++)
937
		printf("..... %2d: %016" PRIx64 "\n",
K
Kan Liang 已提交
938
		       i, callchain->ips[i]);
939 940
}

941 942 943 944 945 946
static void branch_stack__printf(struct perf_sample *sample)
{
	uint64_t i;

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

947 948 949 950 951
	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,
952
			(unsigned short)e->flags.cycles,
953 954 955 956 957 958
			e->flags.mispred ? "M" : " ",
			e->flags.predicted ? "P" : " ",
			e->flags.abort ? "A" : " ",
			e->flags.in_tx ? "T" : " ",
			(unsigned)e->flags.reserved);
	}
959 960
}

961 962 963 964 965 966 967 968 969 970 971 972
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);
	}
}

973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998
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);
}

999
static void regs_user__printf(struct perf_sample *sample)
1000 1001 1002
{
	struct regs_dump *user_regs = &sample->user_regs;

1003 1004 1005 1006 1007 1008 1009 1010 1011 1012
	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);
1013 1014 1015 1016 1017 1018 1019 1020
}

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

1021
static void perf_evlist__print_tstamp(struct perf_evlist *evlist,
1022
				       union perf_event *event,
1023
				       struct perf_sample *sample)
1024
{
1025
	u64 sample_type = __perf_evlist__combined_sample_type(evlist);
1026

1027
	if (event->header.type != PERF_RECORD_SAMPLE &&
1028
	    !perf_evlist__sample_id_all(evlist)) {
1029 1030 1031 1032
		fputs("-1 -1 ", stdout);
		return;
	}

1033
	if ((sample_type & PERF_SAMPLE_CPU))
1034 1035
		printf("%u ", sample->cpu);

1036
	if (sample_type & PERF_SAMPLE_TIME)
1037
		printf("%" PRIu64 " ", sample->time);
1038 1039
}

1040 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
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);
}

1070
static void dump_event(struct perf_evlist *evlist, union perf_event *event,
1071
		       u64 file_offset, struct perf_sample *sample)
1072 1073 1074 1075
{
	if (!dump_trace)
		return;

1076 1077
	printf("\n%#" PRIx64 " [%#x]: event: %d\n",
	       file_offset, event->header.size, event->header.type);
1078 1079 1080 1081

	trace_event(event);

	if (sample)
1082
		perf_evlist__print_tstamp(evlist, event, sample);
1083

1084
	printf("%#" PRIx64 " [%#x]: PERF_RECORD_%s", file_offset,
1085
	       event->header.size, perf_event__name(event->header.type));
1086 1087
}

1088
static void dump_sample(struct perf_evsel *evsel, union perf_event *event,
1089
			struct perf_sample *sample)
1090
{
1091 1092
	u64 sample_type;

1093 1094 1095
	if (!dump_trace)
		return;

1096
	printf("(IP, 0x%x): %d/%d: %#" PRIx64 " period: %" PRIu64 " addr: %#" PRIx64 "\n",
1097
	       event->header.misc, sample->pid, sample->tid, sample->ip,
1098
	       sample->period, sample->addr);
1099

1100
	sample_type = evsel->attr.sample_type;
1101 1102

	if (sample_type & PERF_SAMPLE_CALLCHAIN)
K
Kan Liang 已提交
1103
		callchain__printf(evsel, sample);
1104

1105
	if ((sample_type & PERF_SAMPLE_BRANCH_STACK) && !perf_evsel__has_branch_callstack(evsel))
1106
		branch_stack__printf(sample);
1107 1108

	if (sample_type & PERF_SAMPLE_REGS_USER)
1109
		regs_user__printf(sample);
1110

1111 1112 1113
	if (sample_type & PERF_SAMPLE_REGS_INTR)
		regs_intr__printf(sample);

1114 1115
	if (sample_type & PERF_SAMPLE_STACK_USER)
		stack_user__printf(&sample->user_stack);
1116 1117 1118

	if (sample_type & PERF_SAMPLE_WEIGHT)
		printf("... weight: %" PRIu64 "\n", sample->weight);
1119 1120 1121

	if (sample_type & PERF_SAMPLE_DATA_SRC)
		printf(" . data_src: 0x%"PRIx64"\n", sample->data_src);
1122

1123 1124 1125
	if (sample_type & PERF_SAMPLE_PHYS_ADDR)
		printf(" .. phys_addr: 0x%"PRIx64"\n", sample->phys_addr);

1126 1127 1128
	if (sample_type & PERF_SAMPLE_TRANSACTION)
		printf("... transaction: %" PRIx64 "\n", sample->transaction);

1129 1130
	if (sample_type & PERF_SAMPLE_READ)
		sample_read__printf(sample, evsel->attr.read_format);
1131 1132
}

J
Jiri Olsa 已提交
1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156
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);
}

1157
static struct machine *machines__find_for_cpumode(struct machines *machines,
1158 1159
					       union perf_event *event,
					       struct perf_sample *sample)
1160
{
1161
	struct machine *machine;
1162

1163
	if (perf_guest &&
1164 1165
	    ((sample->cpumode == PERF_RECORD_MISC_GUEST_KERNEL) ||
	     (sample->cpumode == PERF_RECORD_MISC_GUEST_USER))) {
1166 1167
		u32 pid;

1168 1169
		if (event->header.type == PERF_RECORD_MMAP
		    || event->header.type == PERF_RECORD_MMAP2)
1170 1171
			pid = event->mmap.pid;
		else
1172
			pid = sample->pid;
1173

1174
		machine = machines__find(machines, pid);
1175
		if (!machine)
1176
			machine = machines__findnew(machines, DEFAULT_GUEST_KERNEL_ID);
1177
		return machine;
1178
	}
1179

1180
	return &machines->host;
1181 1182
}

1183
static int deliver_sample_value(struct perf_evlist *evlist,
1184 1185 1186 1187 1188 1189
				struct perf_tool *tool,
				union perf_event *event,
				struct perf_sample *sample,
				struct sample_read_value *v,
				struct machine *machine)
{
1190
	struct perf_sample_id *sid = perf_evlist__id2sid(evlist, v->id);
1191 1192 1193 1194 1195 1196 1197 1198

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

	if (!sid || sid->evsel == NULL) {
1199
		++evlist->stats.nr_unknown_id;
1200 1201 1202 1203 1204 1205
		return 0;
	}

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

1206
static int deliver_sample_group(struct perf_evlist *evlist,
1207 1208 1209 1210 1211 1212 1213 1214 1215
				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++) {
1216
		ret = deliver_sample_value(evlist, tool, event, sample,
1217 1218 1219 1220 1221 1222 1223 1224 1225 1226
					   &sample->read.group.values[i],
					   machine);
		if (ret)
			break;
	}

	return ret;
}

static int
1227
 perf_evlist__deliver_sample(struct perf_evlist *evlist,
1228 1229 1230 1231 1232 1233 1234 1235 1236 1237
			     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;

1238
	/* Standard sample delivery. */
1239 1240 1241 1242 1243
	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)
1244
		return deliver_sample_group(evlist, tool, event, sample,
1245 1246
					    machine);
	else
1247
		return deliver_sample_value(evlist, tool, event, sample,
1248 1249 1250
					    &sample->read.one, machine);
}

1251 1252 1253 1254 1255
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)
1256
{
1257
	struct perf_evsel *evsel;
1258
	struct machine *machine;
1259

1260
	dump_event(evlist, event, file_offset, sample);
1261

1262
	evsel = perf_evlist__id2evsel(evlist, sample->id);
1263

1264
	machine = machines__find_for_cpumode(machines, event, sample);
1265

1266 1267
	switch (event->header.type) {
	case PERF_RECORD_SAMPLE:
1268
		if (evsel == NULL) {
1269
			++evlist->stats.nr_unknown_id;
1270
			return 0;
1271
		}
1272
		dump_sample(evsel, event, sample);
1273
		if (machine == NULL) {
1274
			++evlist->stats.nr_unprocessable_samples;
1275
			return 0;
1276
		}
1277
		return perf_evlist__deliver_sample(evlist, tool, event, sample, evsel, machine);
1278
	case PERF_RECORD_MMAP:
1279
		return tool->mmap(tool, event, sample, machine);
1280
	case PERF_RECORD_MMAP2:
1281 1282
		if (event->header.misc & PERF_RECORD_MISC_PROC_MAP_PARSE_TIMEOUT)
			++evlist->stats.nr_proc_map_timeout;
1283
		return tool->mmap2(tool, event, sample, machine);
1284
	case PERF_RECORD_COMM:
1285
		return tool->comm(tool, event, sample, machine);
1286 1287
	case PERF_RECORD_NAMESPACES:
		return tool->namespaces(tool, event, sample, machine);
1288
	case PERF_RECORD_FORK:
1289
		return tool->fork(tool, event, sample, machine);
1290
	case PERF_RECORD_EXIT:
1291
		return tool->exit(tool, event, sample, machine);
1292
	case PERF_RECORD_LOST:
1293
		if (tool->lost == perf_event__process_lost)
1294
			evlist->stats.total_lost += event->lost.lost;
1295
		return tool->lost(tool, event, sample, machine);
1296 1297 1298 1299
	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);
1300
	case PERF_RECORD_READ:
J
Jiri Olsa 已提交
1301
		dump_read(evsel, event);
1302
		return tool->read(tool, event, sample, evsel, machine);
1303
	case PERF_RECORD_THROTTLE:
1304
		return tool->throttle(tool, event, sample, machine);
1305
	case PERF_RECORD_UNTHROTTLE:
1306
		return tool->unthrottle(tool, event, sample, machine);
1307
	case PERF_RECORD_AUX:
1308 1309 1310 1311 1312 1313
		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;
		}
1314
		return tool->aux(tool, event, sample, machine);
1315 1316
	case PERF_RECORD_ITRACE_START:
		return tool->itrace_start(tool, event, sample, machine);
1317 1318 1319
	case PERF_RECORD_SWITCH:
	case PERF_RECORD_SWITCH_CPU_WIDE:
		return tool->context_switch(tool, event, sample, machine);
1320
	default:
1321
		++evlist->stats.nr_unknown_events;
1322 1323 1324 1325
		return -1;
	}
}

1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343
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);
}

1344 1345 1346
static s64 perf_session__process_user_event(struct perf_session *session,
					    union perf_event *event,
					    u64 file_offset)
1347
{
1348
	struct ordered_events *oe = &session->ordered_events;
1349
	struct perf_tool *tool = session->tool;
1350
	int fd = perf_data_file__fd(session->file);
1351 1352
	int err;

1353
	dump_event(session->evlist, event, file_offset, NULL);
1354

1355
	/* These events are processed right away */
1356
	switch (event->header.type) {
1357
	case PERF_RECORD_HEADER_ATTR:
1358
		err = tool->attr(tool, event, &session->evlist);
1359
		if (err == 0) {
1360
			perf_session__set_id_hdr_size(session);
1361 1362
			perf_session__set_comm_exec(session);
		}
1363
		return err;
1364 1365
	case PERF_RECORD_EVENT_UPDATE:
		return tool->event_update(tool, event, &session->evlist);
1366 1367 1368 1369 1370 1371
	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;
1372 1373
	case PERF_RECORD_HEADER_TRACING_DATA:
		/* setup for reading amidst mmap */
1374
		lseek(fd, file_offset, SEEK_SET);
1375
		return tool->tracing_data(tool, event, session);
1376
	case PERF_RECORD_HEADER_BUILD_ID:
1377
		return tool->build_id(tool, event, session);
1378
	case PERF_RECORD_FINISHED_ROUND:
1379
		return tool->finished_round(tool, event, oe);
A
Adrian Hunter 已提交
1380 1381
	case PERF_RECORD_ID_INDEX:
		return tool->id_index(tool, event, session);
1382 1383 1384 1385 1386 1387
	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);
1388
	case PERF_RECORD_AUXTRACE_ERROR:
1389
		perf_session__auxtrace_error_inc(session, event);
1390
		return tool->auxtrace_error(tool, event, session);
1391 1392
	case PERF_RECORD_THREAD_MAP:
		return tool->thread_map(tool, event, session);
1393 1394
	case PERF_RECORD_CPU_MAP:
		return tool->cpu_map(tool, event, session);
1395 1396
	case PERF_RECORD_STAT_CONFIG:
		return tool->stat_config(tool, event, session);
J
Jiri Olsa 已提交
1397 1398
	case PERF_RECORD_STAT:
		return tool->stat(tool, event, session);
1399 1400
	case PERF_RECORD_STAT_ROUND:
		return tool->stat_round(tool, event, session);
1401 1402 1403
	case PERF_RECORD_TIME_CONV:
		session->time_conv = event->time_conv;
		return tool->time_conv(tool, event, session);
1404 1405
	case PERF_RECORD_HEADER_FEATURE:
		return tool->feature(tool, event, session);
1406
	default:
1407
		return -EINVAL;
1408
	}
1409 1410
}

1411 1412
int perf_session__deliver_synth_event(struct perf_session *session,
				      union perf_event *event,
1413
				      struct perf_sample *sample)
1414
{
1415
	struct perf_evlist *evlist = session->evlist;
1416
	struct perf_tool *tool = session->tool;
1417 1418

	events_stats__inc(&evlist->stats, event->header.type);
1419 1420

	if (event->header.type >= PERF_RECORD_USER_TYPE_START)
1421
		return perf_session__process_user_event(session, event, 0);
1422

1423
	return machines__deliver_event(&session->machines, evlist, event, sample, tool, 0);
1424 1425
}

1426 1427 1428 1429 1430 1431 1432 1433 1434
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);
}

1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459
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;
	}

	if (perf_data_file__is_pipe(session->file))
		return -1;

	fd = perf_data_file__fd(session->file);
	hdr_sz = sizeof(struct perf_event_header);

	if (buf_sz < hdr_sz)
		return -1;

	if (lseek(fd, file_offset, SEEK_SET) == (off_t)-1 ||
1460
	    readn(fd, buf, hdr_sz) != (ssize_t)hdr_sz)
1461 1462 1463 1464 1465 1466 1467
		return -1;

	event = (union perf_event *)buf;

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

1468
	if (event->header.size < hdr_sz || event->header.size > buf_sz)
1469 1470 1471 1472
		return -1;

	rest = event->header.size - hdr_sz;

1473
	if (readn(fd, buf, rest) != (ssize_t)rest)
1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489
		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;
}

1490
static s64 perf_session__process_event(struct perf_session *session,
1491
				       union perf_event *event, u64 file_offset)
1492
{
1493
	struct perf_evlist *evlist = session->evlist;
1494
	struct perf_tool *tool = session->tool;
1495
	struct perf_sample sample;
1496 1497
	int ret;

1498
	if (session->header.needs_swap)
1499
		event_swap(event, perf_evlist__sample_id_all(evlist));
1500 1501 1502 1503

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

1504
	events_stats__inc(&evlist->stats, event->header.type);
1505 1506

	if (event->header.type >= PERF_RECORD_USER_TYPE_START)
1507
		return perf_session__process_user_event(session, event, file_offset);
1508

1509 1510 1511
	/*
	 * For all kernel events we get the sample data
	 */
1512
	ret = perf_evlist__parse_sample(evlist, event, &sample);
1513 1514
	if (ret)
		return ret;
1515

1516
	if (tool->ordered_events) {
1517
		ret = perf_session__queue_event(session, event, &sample, file_offset);
1518 1519 1520 1521
		if (ret != -ETIME)
			return ret;
	}

1522 1523
	return perf_session__deliver_event(session, event, &sample, tool,
					   file_offset);
1524 1525
}

1526
void perf_event_header__bswap(struct perf_event_header *hdr)
1527
{
1528 1529 1530
	hdr->type = bswap_32(hdr->type);
	hdr->misc = bswap_16(hdr->misc);
	hdr->size = bswap_16(hdr->size);
1531 1532
}

1533 1534
struct thread *perf_session__findnew(struct perf_session *session, pid_t pid)
{
1535
	return machine__findnew_thread(&session->machines.host, -1, pid);
1536 1537
}

1538
int perf_session__register_idle_thread(struct perf_session *session)
1539
{
1540
	struct thread *thread;
1541
	int err = 0;
1542

1543
	thread = machine__findnew_thread(&session->machines.host, 0, 0);
1544
	if (thread == NULL || thread__set_comm(thread, "swapper", 0)) {
1545
		pr_err("problem inserting idle task.\n");
1546
		err = -1;
1547 1548
	}

1549 1550 1551 1552 1553
	if (thread == NULL || thread__set_namespaces(thread, 0, NULL)) {
		pr_err("problem inserting idle task.\n");
		err = -1;
	}

1554 1555 1556
	/* machine__findnew_thread() got the thread, so put it */
	thread__put(thread);
	return err;
1557 1558
}

1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576
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);
}

1577
static void perf_session__warn_about_errors(const struct perf_session *session)
1578
{
1579 1580 1581
	const struct events_stats *stats = &session->evlist->stats;

	if (session->tool->lost == perf_event__process_lost &&
1582
	    stats->nr_events[PERF_RECORD_LOST] != 0) {
1583 1584
		ui__warning("Processed %d events and lost %d chunks!\n\n"
			    "Check IO/CPU overload!\n\n",
1585 1586
			    stats->nr_events[0],
			    stats->nr_events[PERF_RECORD_LOST]);
1587 1588
	}

1589 1590 1591 1592 1593 1594
	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) {
1595
			ui__warning("Processed %" PRIu64 " samples and lost %3.2f%% samples!\n\n",
1596 1597 1598 1599 1600
				    stats->nr_events[PERF_RECORD_SAMPLE] + stats->total_lost_samples,
				    drop_rate * 100.0);
		}
	}

1601 1602 1603 1604 1605 1606 1607
	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]);
	}

1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624
	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." :
			    "");
	}

1625
	if (stats->nr_unknown_events != 0) {
1626 1627 1628 1629 1630
		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",
1631
			    stats->nr_unknown_events);
1632 1633
	}

1634
	if (stats->nr_unknown_id != 0) {
1635
		ui__warning("%u samples with id not present in the header\n",
1636
			    stats->nr_unknown_id);
1637 1638
	}

1639
	if (stats->nr_invalid_chains != 0) {
1640 1641 1642
		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",
1643 1644
			    stats->nr_invalid_chains,
			    stats->nr_events[PERF_RECORD_SAMPLE]);
1645
	}
1646

1647
	if (stats->nr_unprocessable_samples != 0) {
1648 1649
		ui__warning("%u unprocessable samples recorded.\n"
			    "Do you have a KVM guest running and not using 'perf kvm'?\n",
1650
			    stats->nr_unprocessable_samples);
1651
	}
1652

1653
	perf_session__warn_order(session);
1654 1655

	events_stats__auxtrace_error_warn(stats);
1656 1657 1658 1659 1660 1661

	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"
1662 1663 1664
			    "to a file.\n"
			    "The time limit to process proc map is too short?\n"
			    "Increase it by --proc-map-timeout\n",
1665 1666
			    stats->nr_proc_map_timeout);
	}
1667 1668
}

1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681
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);
}

1682 1683
volatile int session_done;

1684
static int __perf_session__process_pipe_events(struct perf_session *session)
1685
{
1686
	struct ordered_events *oe = &session->ordered_events;
1687
	struct perf_tool *tool = session->tool;
1688
	int fd = perf_data_file__fd(session->file);
1689 1690 1691
	union perf_event *event;
	uint32_t size, cur_size = 0;
	void *buf = NULL;
1692
	s64 skip = 0;
1693
	u64 head;
1694
	ssize_t err;
1695 1696
	void *p;

1697
	perf_tool__fill_defaults(tool);
1698 1699

	head = 0;
1700 1701 1702 1703 1704
	cur_size = sizeof(union perf_event);

	buf = malloc(cur_size);
	if (!buf)
		return -errno;
1705
	ordered_events__set_copy_on_queue(oe, true);
1706
more:
1707
	event = buf;
1708
	err = readn(fd, event, sizeof(struct perf_event_header));
1709 1710 1711 1712 1713 1714 1715 1716
	if (err <= 0) {
		if (err == 0)
			goto done;

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

1717
	if (session->header.needs_swap)
1718
		perf_event_header__bswap(&event->header);
1719

1720
	size = event->header.size;
1721 1722 1723 1724
	if (size < sizeof(struct perf_event_header)) {
		pr_err("bad event header size\n");
		goto out_err;
	}
1725

1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736
	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;
1737 1738
	p += sizeof(struct perf_event_header);

1739
	if (size - sizeof(struct perf_event_header)) {
1740
		err = readn(fd, p, size - sizeof(struct perf_event_header));
1741 1742 1743 1744 1745
		if (err <= 0) {
			if (err == 0) {
				pr_err("unexpected end of event stream\n");
				goto done;
			}
1746

1747 1748 1749
			pr_err("failed to read event data\n");
			goto out_err;
		}
1750 1751
	}

1752
	if ((skip = perf_session__process_event(session, event, head)) < 0) {
1753
		pr_err("%#" PRIx64 " [%#x]: failed to process type: %d\n",
1754
		       head, event->header.size, event->header.type);
1755 1756
		err = -EINVAL;
		goto out_err;
1757 1758 1759 1760 1761 1762 1763 1764 1765 1766
	}

	head += size;

	if (skip > 0)
		head += skip;

	if (!session_done())
		goto more;
done:
1767
	/* do the final flush for ordered samples */
1768
	err = ordered_events__flush(oe, OE_FLUSH__FINAL);
1769 1770 1771
	if (err)
		goto out_err;
	err = auxtrace__flush_events(session, tool);
1772 1773 1774
	if (err)
		goto out_err;
	err = perf_session__flush_thread_stacks(session);
1775
out_err:
1776
	free(buf);
1777
	perf_session__warn_about_errors(session);
1778
	ordered_events__free(&session->ordered_events);
1779
	auxtrace__free_events(session);
1780 1781 1782
	return err;
}

1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800
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);

1801 1802 1803 1804
	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);
1805
		return NULL;
1806
	}
1807 1808 1809 1810

	return event;
}

1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822
/*
 * 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

1823 1824
static int __perf_session__process_events(struct perf_session *session,
					  u64 data_offset, u64 data_size,
1825
					  u64 file_size)
1826
{
1827
	struct ordered_events *oe = &session->ordered_events;
1828
	struct perf_tool *tool = session->tool;
1829
	int fd = perf_data_file__fd(session->file);
1830
	u64 head, page_offset, file_offset, file_pos, size;
1831
	int err, mmap_prot, mmap_flags, map_idx = 0;
1832
	size_t	mmap_size;
1833
	char *buf, *mmaps[NUM_MMAPS];
1834
	union perf_event *event;
1835
	struct ui_progress prog;
1836
	s64 skip;
1837

1838
	perf_tool__fill_defaults(tool);
1839

1840 1841 1842
	page_offset = page_size * (data_offset / page_size);
	file_offset = page_offset;
	head = data_offset - page_offset;
1843

1844 1845 1846 1847
	if (data_size == 0)
		goto out;

	if (data_offset + data_size < file_size)
1848 1849
		file_size = data_offset + data_size;

1850
	ui_progress__init(&prog, file_size, "Processing events...");
1851

1852
	mmap_size = MMAP_SIZE;
1853
	if (mmap_size > file_size) {
1854
		mmap_size = file_size;
1855 1856
		session->one_mmap = true;
	}
1857

1858 1859
	memset(mmaps, 0, sizeof(mmaps));

1860 1861 1862
	mmap_prot  = PROT_READ;
	mmap_flags = MAP_SHARED;

1863
	if (session->header.needs_swap) {
1864 1865 1866
		mmap_prot  |= PROT_WRITE;
		mmap_flags = MAP_PRIVATE;
	}
1867
remap:
1868
	buf = mmap(NULL, mmap_size, mmap_prot, mmap_flags, fd,
1869
		   file_offset);
1870 1871 1872 1873 1874
	if (buf == MAP_FAILED) {
		pr_err("failed to mmap file\n");
		err = -errno;
		goto out_err;
	}
1875 1876
	mmaps[map_idx] = buf;
	map_idx = (map_idx + 1) & (ARRAY_SIZE(mmaps) - 1);
1877
	file_pos = file_offset + head;
1878 1879 1880 1881
	if (session->one_mmap) {
		session->one_mmap_addr = buf;
		session->one_mmap_offset = file_offset;
	}
1882 1883

more:
1884 1885
	event = fetch_mmaped_event(session, head, mmap_size, buf);
	if (!event) {
1886 1887 1888 1889
		if (mmaps[map_idx]) {
			munmap(mmaps[map_idx], mmap_size);
			mmaps[map_idx] = NULL;
		}
1890

1891 1892 1893
		page_offset = page_size * (head / page_size);
		file_offset += page_offset;
		head -= page_offset;
1894 1895 1896 1897 1898
		goto remap;
	}

	size = event->header.size;

1899
	if (size < sizeof(struct perf_event_header) ||
1900
	    (skip = perf_session__process_event(session, event, file_pos)) < 0) {
1901 1902 1903 1904 1905
		pr_err("%#" PRIx64 " [%#x]: failed to process type: %d\n",
		       file_offset + head, event->header.size,
		       event->header.type);
		err = -EINVAL;
		goto out_err;
1906 1907
	}

1908 1909 1910
	if (skip)
		size += skip;

1911
	head += size;
1912
	file_pos += size;
1913

1914
	ui_progress__update(&prog, size);
1915

1916
	if (session_done())
1917
		goto out;
1918

1919
	if (file_pos < file_size)
1920
		goto more;
1921

1922
out:
1923
	/* do the final flush for ordered samples */
1924
	err = ordered_events__flush(oe, OE_FLUSH__FINAL);
1925 1926 1927
	if (err)
		goto out_err;
	err = auxtrace__flush_events(session, tool);
1928 1929 1930
	if (err)
		goto out_err;
	err = perf_session__flush_thread_stacks(session);
1931
out_err:
N
Namhyung Kim 已提交
1932
	ui_progress__finish();
1933
	perf_session__warn_about_errors(session);
1934 1935 1936 1937 1938
	/*
	 * We may switching perf.data output, make ordered_events
	 * reusable.
	 */
	ordered_events__reinit(&session->ordered_events);
1939
	auxtrace__free_events(session);
1940
	session->one_mmap = false;
1941 1942
	return err;
}
1943

1944
int perf_session__process_events(struct perf_session *session)
1945
{
1946
	u64 size = perf_data_file__size(session->file);
1947 1948
	int err;

1949
	if (perf_session__register_idle_thread(session) < 0)
1950 1951
		return -ENOMEM;

1952 1953 1954
	if (!perf_data_file__is_pipe(session->file))
		err = __perf_session__process_events(session,
						     session->header.data_offset,
1955
						     session->header.data_size, size);
1956
	else
1957
		err = __perf_session__process_pipe_events(session);
1958

1959 1960 1961
	return err;
}

1962
bool perf_session__has_traces(struct perf_session *session, const char *msg)
1963
{
1964 1965
	struct perf_evsel *evsel;

1966
	evlist__for_each_entry(session->evlist, evsel) {
1967 1968
		if (evsel->attr.type == PERF_TYPE_TRACEPOINT)
			return true;
1969 1970
	}

1971 1972
	pr_err("No trace sample to read. Did you call 'perf %s'?\n", msg);
	return false;
1973
}
1974

1975 1976
int maps__set_kallsyms_ref_reloc_sym(struct map **maps,
				     const char *symbol_name, u64 addr)
1977 1978
{
	char *bracket;
1979
	int i;
1980 1981 1982 1983 1984
	struct ref_reloc_sym *ref;

	ref = zalloc(sizeof(struct ref_reloc_sym));
	if (ref == NULL)
		return -ENOMEM;
1985

1986 1987 1988
	ref->name = strdup(symbol_name);
	if (ref->name == NULL) {
		free(ref);
1989
		return -ENOMEM;
1990
	}
1991

1992
	bracket = strchr(ref->name, ']');
1993 1994 1995
	if (bracket)
		*bracket = '\0';

1996
	ref->addr = addr;
1997 1998

	for (i = 0; i < MAP__NR_TYPES; ++i) {
1999
		struct kmap *kmap = map__kmap(maps[i]);
2000 2001 2002

		if (!kmap)
			continue;
2003
		kmap->ref_reloc_sym = ref;
2004 2005
	}

2006 2007
	return 0;
}
2008

2009
size_t perf_session__fprintf_dsos(struct perf_session *session, FILE *fp)
2010
{
2011
	return machines__fprintf_dsos(&session->machines, fp);
2012
}
2013

2014
size_t perf_session__fprintf_dsos_buildid(struct perf_session *session, FILE *fp,
2015
					  bool (skip)(struct dso *dso, int parm), int parm)
2016
{
2017
	return machines__fprintf_dsos_buildid(&session->machines, fp, skip, parm);
2018
}
2019 2020 2021

size_t perf_session__fprintf_nr_events(struct perf_session *session, FILE *fp)
{
2022 2023 2024 2025 2026 2027
	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)";

2028
	ret = fprintf(fp, "\nAggregated stats:%s\n", msg);
2029

2030
	ret += events_stats__fprintf(&session->evlist->stats, fp);
2031 2032
	return ret;
}
2033

2034 2035 2036 2037 2038 2039
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...
	 */
2040
	return machine__fprintf(&session->machines.host, fp);
2041 2042
}

2043 2044 2045 2046 2047
struct perf_evsel *perf_session__find_first_evtype(struct perf_session *session,
					      unsigned int type)
{
	struct perf_evsel *pos;

2048
	evlist__for_each_entry(session->evlist, pos) {
2049 2050 2051 2052 2053 2054
		if (pos->attr.type == type)
			return pos;
	}
	return NULL;
}

2055 2056 2057
int perf_session__cpu_bitmap(struct perf_session *session,
			     const char *cpu_list, unsigned long *cpu_bitmap)
{
2058
	int i, err = -1;
2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069
	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. "
2070
			       "Remove -C option to proceed.\n");
2071 2072 2073 2074 2075
			return -1;
		}
	}

	map = cpu_map__new(cpu_list);
2076 2077 2078 2079
	if (map == NULL) {
		pr_err("Invalid cpu_list\n");
		return -1;
	}
2080 2081 2082 2083 2084 2085 2086

	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);
2087
			goto out_delete_map;
2088 2089 2090 2091 2092
		}

		set_bit(cpu, cpu_bitmap);
	}

2093 2094 2095
	err = 0;

out_delete_map:
2096
	cpu_map__put(map);
2097
	return err;
2098
}
2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109

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");
}
2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120


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++) {
2121 2122 2123 2124 2125
		/*
		 * 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);
2126
		if (evsel == NULL)
2127
			continue;
2128 2129

		err = -EEXIST;
2130
		if (evsel->handler != NULL)
2131
			goto out;
2132
		evsel->handler = assocs[i].handler;
2133 2134 2135 2136 2137 2138
	}

	err = 0;
out:
	return err;
}
A
Adrian Hunter 已提交
2139 2140 2141 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

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);

2193
	evlist__for_each_entry(evlist, evsel)
A
Adrian Hunter 已提交
2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205
		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;

2206
	evlist__for_each_entry(evlist, evsel) {
A
Adrian Hunter 已提交
2207 2208 2209 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
		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;
}