event.c 42.1 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0
2
#include <dirent.h>
3
#include <errno.h>
4
#include <fcntl.h>
5
#include <inttypes.h>
6
#include <linux/kernel.h>
7
#include <linux/types.h>
8 9 10
#include <sys/types.h>
#include <sys/stat.h>
#include <unistd.h>
11
#include <uapi/linux/mman.h> /* To get things like MAP_HUGETLB even on older libc headers */
12
#include <api/fs/fs.h>
13
#include <linux/perf_event.h>
14 15
#include "event.h"
#include "debug.h"
16
#include "hist.h"
17
#include "machine.h"
18
#include "sort.h"
19
#include "string2.h"
20
#include "strlist.h"
21
#include "thread.h"
22
#include "thread_map.h"
23
#include "sane_ctype.h"
24
#include "symbol/kallsyms.h"
25 26
#include "asm/bug.h"
#include "stat.h"
27

28
static const char *perf_event__names[] = {
29 30
	[0]					= "TOTAL",
	[PERF_RECORD_MMAP]			= "MMAP",
31
	[PERF_RECORD_MMAP2]			= "MMAP2",
32 33 34 35 36 37 38 39
	[PERF_RECORD_LOST]			= "LOST",
	[PERF_RECORD_COMM]			= "COMM",
	[PERF_RECORD_EXIT]			= "EXIT",
	[PERF_RECORD_THROTTLE]			= "THROTTLE",
	[PERF_RECORD_UNTHROTTLE]		= "UNTHROTTLE",
	[PERF_RECORD_FORK]			= "FORK",
	[PERF_RECORD_READ]			= "READ",
	[PERF_RECORD_SAMPLE]			= "SAMPLE",
40
	[PERF_RECORD_AUX]			= "AUX",
41
	[PERF_RECORD_ITRACE_START]		= "ITRACE_START",
42
	[PERF_RECORD_LOST_SAMPLES]		= "LOST_SAMPLES",
43 44
	[PERF_RECORD_SWITCH]			= "SWITCH",
	[PERF_RECORD_SWITCH_CPU_WIDE]		= "SWITCH_CPU_WIDE",
45
	[PERF_RECORD_NAMESPACES]		= "NAMESPACES",
46 47 48 49 50
	[PERF_RECORD_HEADER_ATTR]		= "ATTR",
	[PERF_RECORD_HEADER_EVENT_TYPE]		= "EVENT_TYPE",
	[PERF_RECORD_HEADER_TRACING_DATA]	= "TRACING_DATA",
	[PERF_RECORD_HEADER_BUILD_ID]		= "BUILD_ID",
	[PERF_RECORD_FINISHED_ROUND]		= "FINISHED_ROUND",
A
Adrian Hunter 已提交
51
	[PERF_RECORD_ID_INDEX]			= "ID_INDEX",
52 53
	[PERF_RECORD_AUXTRACE_INFO]		= "AUXTRACE_INFO",
	[PERF_RECORD_AUXTRACE]			= "AUXTRACE",
54
	[PERF_RECORD_AUXTRACE_ERROR]		= "AUXTRACE_ERROR",
55
	[PERF_RECORD_THREAD_MAP]		= "THREAD_MAP",
56
	[PERF_RECORD_CPU_MAP]			= "CPU_MAP",
57
	[PERF_RECORD_STAT_CONFIG]		= "STAT_CONFIG",
J
Jiri Olsa 已提交
58
	[PERF_RECORD_STAT]			= "STAT",
59
	[PERF_RECORD_STAT_ROUND]		= "STAT_ROUND",
60
	[PERF_RECORD_EVENT_UPDATE]		= "EVENT_UPDATE",
61
	[PERF_RECORD_TIME_CONV]			= "TIME_CONV",
62
	[PERF_RECORD_HEADER_FEATURE]		= "FEATURE",
63 64
};

65 66 67 68 69 70 71 72 73 74
static const char *perf_ns__names[] = {
	[NET_NS_INDEX]		= "net",
	[UTS_NS_INDEX]		= "uts",
	[IPC_NS_INDEX]		= "ipc",
	[PID_NS_INDEX]		= "pid",
	[USER_NS_INDEX]		= "user",
	[MNT_NS_INDEX]		= "mnt",
	[CGROUP_NS_INDEX]	= "cgroup",
};

75
const char *perf_event__name(unsigned int id)
76
{
77
	if (id >= ARRAY_SIZE(perf_event__names))
78
		return "INVALID";
79
	if (!perf_event__names[id])
80
		return "UNKNOWN";
81
	return perf_event__names[id];
82 83
}

84 85 86 87 88 89 90
static const char *perf_ns__name(unsigned int id)
{
	if (id >= ARRAY_SIZE(perf_ns__names))
		return "UNKNOWN";
	return perf_ns__names[id];
}

91 92 93 94 95 96
static int perf_tool__process_synth_event(struct perf_tool *tool,
					  union perf_event *event,
					  struct machine *machine,
					  perf_event__handler_t process)
{
	struct perf_sample synth_sample = {
97 98 99 100 101 102
	.pid	   = -1,
	.tid	   = -1,
	.time	   = -1,
	.stream_id = -1,
	.cpu	   = -1,
	.period	   = 1,
103 104 105 106
	.cpumode   = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK,
	};

	return process(tool, event, &synth_sample, machine);
107 108
};

109 110
/*
 * Assumes that the first 4095 bytes of /proc/pid/stat contains
111
 * the comm, tgid and ppid.
112
 */
113 114
static int perf_event__get_comm_ids(pid_t pid, char *comm, size_t len,
				    pid_t *tgid, pid_t *ppid)
115 116
{
	char filename[PATH_MAX];
117 118
	char bf[4096];
	int fd;
119 120
	size_t size = 0;
	ssize_t n;
121
	char *name, *tgids, *ppids;
122 123 124

	*tgid = -1;
	*ppid = -1;
125 126 127

	snprintf(filename, sizeof(filename), "/proc/%d/status", pid);

128 129
	fd = open(filename, O_RDONLY);
	if (fd < 0) {
130
		pr_debug("couldn't open %s\n", filename);
131
		return -1;
132 133
	}

134 135 136
	n = read(fd, bf, sizeof(bf) - 1);
	close(fd);
	if (n <= 0) {
137
		pr_warning("Couldn't get COMM, tigd and ppid for pid %d\n",
138 139
			   pid);
		return -1;
140
	}
141
	bf[n] = '\0';
142

143 144
	name = strstr(bf, "Name:");
	tgids = strstr(bf, "Tgid:");
145
	ppids = strstr(bf, "PPid:");
146 147

	if (name) {
148 149
		char *nl;

150
		name += 5;  /* strlen("Name:") */
151 152 153 154 155 156
		name = ltrim(name);

		nl = strchr(name, '\n');
		if (nl)
			*nl = '\0';

157 158 159 160 161 162 163 164 165 166 167
		size = strlen(name);
		if (size >= len)
			size = len - 1;
		memcpy(comm, name, size);
		comm[size] = '\0';
	} else {
		pr_debug("Name: string not found for pid %d\n", pid);
	}

	if (tgids) {
		tgids += 5;  /* strlen("Tgid:") */
168
		*tgid = atoi(tgids);
169 170 171
	} else {
		pr_debug("Tgid: string not found for pid %d\n", pid);
	}
172

173 174 175 176 177 178 179 180
	if (ppids) {
		ppids += 5;  /* strlen("PPid:") */
		*ppid = atoi(ppids);
	} else {
		pr_debug("PPid: string not found for pid %d\n", pid);
	}

	return 0;
181 182
}

183 184 185
static int perf_event__prepare_comm(union perf_event *event, pid_t pid,
				    struct machine *machine,
				    pid_t *tgid, pid_t *ppid)
186 187
{
	size_t size;
188 189

	*ppid = -1;
190 191 192

	memset(&event->comm, 0, sizeof(event->comm));

193 194 195 196 197 198 199 200 201
	if (machine__is_host(machine)) {
		if (perf_event__get_comm_ids(pid, event->comm.comm,
					     sizeof(event->comm.comm),
					     tgid, ppid) != 0) {
			return -1;
		}
	} else {
		*tgid = machine->pid;
	}
202

203 204
	if (*tgid < 0)
		return -1;
205

206
	event->comm.pid = *tgid;
207
	event->comm.header.type = PERF_RECORD_COMM;
208 209

	size = strlen(event->comm.comm) + 1;
210
	size = PERF_ALIGN(size, sizeof(u64));
211
	memset(event->comm.comm + size, 0, machine->id_hdr_size);
212 213
	event->comm.header.size = (sizeof(event->comm) -
				(sizeof(event->comm.comm) - size) +
214
				machine->id_hdr_size);
215
	event->comm.tid = pid;
216 217

	return 0;
218 219
}

220
pid_t perf_event__synthesize_comm(struct perf_tool *tool,
221 222 223 224
					 union perf_event *event, pid_t pid,
					 perf_event__handler_t process,
					 struct machine *machine)
{
225
	pid_t tgid, ppid;
226

227 228
	if (perf_event__prepare_comm(event, pid, machine, &tgid, &ppid) != 0)
		return -1;
229

230
	if (perf_tool__process_synth_event(tool, event, machine, process) != 0)
231
		return -1;
232

233
	return tgid;
234 235
}

236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287
static void perf_event__get_ns_link_info(pid_t pid, const char *ns,
					 struct perf_ns_link_info *ns_link_info)
{
	struct stat64 st;
	char proc_ns[128];

	sprintf(proc_ns, "/proc/%u/ns/%s", pid, ns);
	if (stat64(proc_ns, &st) == 0) {
		ns_link_info->dev = st.st_dev;
		ns_link_info->ino = st.st_ino;
	}
}

int perf_event__synthesize_namespaces(struct perf_tool *tool,
				      union perf_event *event,
				      pid_t pid, pid_t tgid,
				      perf_event__handler_t process,
				      struct machine *machine)
{
	u32 idx;
	struct perf_ns_link_info *ns_link_info;

	if (!tool || !tool->namespace_events)
		return 0;

	memset(&event->namespaces, 0, (sizeof(event->namespaces) +
	       (NR_NAMESPACES * sizeof(struct perf_ns_link_info)) +
	       machine->id_hdr_size));

	event->namespaces.pid = tgid;
	event->namespaces.tid = pid;

	event->namespaces.nr_namespaces = NR_NAMESPACES;

	ns_link_info = event->namespaces.link_info;

	for (idx = 0; idx < event->namespaces.nr_namespaces; idx++)
		perf_event__get_ns_link_info(pid, perf_ns__name(idx),
					     &ns_link_info[idx]);

	event->namespaces.header.type = PERF_RECORD_NAMESPACES;

	event->namespaces.header.size = (sizeof(event->namespaces) +
			(NR_NAMESPACES * sizeof(struct perf_ns_link_info)) +
			machine->id_hdr_size);

	if (perf_tool__process_synth_event(tool, event, machine, process) != 0)
		return -1;

	return 0;
}

288
static int perf_event__synthesize_fork(struct perf_tool *tool,
289 290 291
				       union perf_event *event,
				       pid_t pid, pid_t tgid, pid_t ppid,
				       perf_event__handler_t process,
292 293 294 295
				       struct machine *machine)
{
	memset(&event->fork, 0, sizeof(event->fork) + machine->id_hdr_size);

296 297 298 299 300 301 302 303 304 305 306 307
	/*
	 * for main thread set parent to ppid from status file. For other
	 * threads set parent pid to main thread. ie., assume main thread
	 * spawns all threads in a process
	*/
	if (tgid == pid) {
		event->fork.ppid = ppid;
		event->fork.ptid = ppid;
	} else {
		event->fork.ppid = tgid;
		event->fork.ptid = tgid;
	}
308 309 310 311 312 313
	event->fork.pid  = tgid;
	event->fork.tid  = pid;
	event->fork.header.type = PERF_RECORD_FORK;

	event->fork.header.size = (sizeof(event->fork) + machine->id_hdr_size);

314
	if (perf_tool__process_synth_event(tool, event, machine, process) != 0)
315 316 317 318 319
		return -1;

	return 0;
}

320 321 322 323 324
int perf_event__synthesize_mmap_events(struct perf_tool *tool,
				       union perf_event *event,
				       pid_t pid, pid_t tgid,
				       perf_event__handler_t process,
				       struct machine *machine,
325 326
				       bool mmap_data,
				       unsigned int proc_map_timeout)
327 328 329
{
	char filename[PATH_MAX];
	FILE *fp;
330 331
	unsigned long long t;
	bool truncation = false;
332
	unsigned long long timeout = proc_map_timeout * 1000000ULL;
333
	int rc = 0;
334 335
	const char *hugetlbfs_mnt = hugetlbfs__mountpoint();
	int hugetlbfs_mnt_len = hugetlbfs_mnt ? strlen(hugetlbfs_mnt) : 0;
336

337 338 339
	if (machine__is_default_guest(machine))
		return 0;

340 341
	snprintf(filename, sizeof(filename), "%s/proc/%d/task/%d/maps",
		 machine->root_dir, pid, pid);
342 343 344 345 346 347 348 349 350 351

	fp = fopen(filename, "r");
	if (fp == NULL) {
		/*
		 * We raced with a task exiting - just return:
		 */
		pr_debug("couldn't open %s\n", filename);
		return -1;
	}

352
	event->header.type = PERF_RECORD_MMAP2;
353
	t = rdclock();
354

355
	while (1) {
356 357 358 359
		char bf[BUFSIZ];
		char prot[5];
		char execname[PATH_MAX];
		char anonstr[] = "//anon";
360
		unsigned int ino;
361
		size_t size;
362
		ssize_t n;
363

364 365 366
		if (fgets(bf, sizeof(bf), fp) == NULL)
			break;

367 368 369 370 371
		if ((rdclock() - t) > timeout) {
			pr_warning("Reading %s time out. "
				   "You may want to increase "
				   "the time limit by --proc-map-timeout\n",
				   filename);
372 373 374 375
			truncation = true;
			goto out;
		}

376 377 378
		/* ensure null termination since stack will be reused. */
		strcpy(execname, "");

379
		/* 00400000-0040c000 r-xp 00000000 fd:01 41038  /bin/cat */
380
		n = sscanf(bf, "%"PRIx64"-%"PRIx64" %s %"PRIx64" %x:%x %u %[^\n]\n",
381 382 383 384 385
		       &event->mmap2.start, &event->mmap2.len, prot,
		       &event->mmap2.pgoff, &event->mmap2.maj,
		       &event->mmap2.min,
		       &ino, execname);

386 387 388
		/*
 		 * Anon maps don't have the execname.
 		 */
389
		if (n < 7)
390
			continue;
391 392 393

		event->mmap2.ino = (u64)ino;

394 395 396
		/*
		 * Just like the kernel, see __perf_event_mmap in kernel/perf_event.c
		 */
397 398 399 400
		if (machine__is_host(machine))
			event->header.misc = PERF_RECORD_MISC_USER;
		else
			event->header.misc = PERF_RECORD_MISC_GUEST_USER;
401

402 403 404 405 406 407 408 409 410 411 412 413 414 415 416
		/* map protection and flags bits */
		event->mmap2.prot = 0;
		event->mmap2.flags = 0;
		if (prot[0] == 'r')
			event->mmap2.prot |= PROT_READ;
		if (prot[1] == 'w')
			event->mmap2.prot |= PROT_WRITE;
		if (prot[2] == 'x')
			event->mmap2.prot |= PROT_EXEC;

		if (prot[3] == 's')
			event->mmap2.flags |= MAP_SHARED;
		else
			event->mmap2.flags |= MAP_PRIVATE;

417 418 419 420 421 422
		if (prot[2] != 'x') {
			if (!mmap_data || prot[0] != 'r')
				continue;

			event->header.misc |= PERF_RECORD_MISC_MMAP_DATA;
		}
423

424 425 426 427
out:
		if (truncation)
			event->header.misc |= PERF_RECORD_MISC_PROC_MAP_PARSE_TIMEOUT;

428 429
		if (!strcmp(execname, ""))
			strcpy(execname, anonstr);
430

431 432
		if (hugetlbfs_mnt_len &&
		    !strncmp(execname, hugetlbfs_mnt, hugetlbfs_mnt_len)) {
433 434 435
			strcpy(execname, anonstr);
			event->mmap2.flags |= MAP_HUGETLB;
		}
436 437

		size = strlen(execname) + 1;
438
		memcpy(event->mmap2.filename, execname, size);
439
		size = PERF_ALIGN(size, sizeof(u64));
440 441 442 443 444 445 446
		event->mmap2.len -= event->mmap.start;
		event->mmap2.header.size = (sizeof(event->mmap2) -
					(sizeof(event->mmap2.filename) - size));
		memset(event->mmap2.filename + size, 0, machine->id_hdr_size);
		event->mmap2.header.size += machine->id_hdr_size;
		event->mmap2.pid = tgid;
		event->mmap2.tid = pid;
447

448
		if (perf_tool__process_synth_event(tool, event, machine, process) != 0) {
449 450
			rc = -1;
			break;
451
		}
452 453 454

		if (truncation)
			break;
455 456 457
	}

	fclose(fp);
458
	return rc;
459 460
}

461
int perf_event__synthesize_modules(struct perf_tool *tool,
462
				   perf_event__handler_t process,
463
				   struct machine *machine)
464
{
465
	int rc = 0;
466
	struct map *pos;
467
	struct map_groups *kmaps = &machine->kmaps;
468
	struct maps *maps = &kmaps->maps[MAP__FUNCTION];
469
	union perf_event *event = zalloc((sizeof(event->mmap) +
470
					  machine->id_hdr_size));
471 472 473 474 475 476 477
	if (event == NULL) {
		pr_debug("Not enough memory synthesizing mmap event "
			 "for kernel modules\n");
		return -1;
	}

	event->header.type = PERF_RECORD_MMAP;
478

479 480 481 482
	/*
	 * kernel uses 0 for user space maps, see kernel/perf_event.c
	 * __perf_event_mmap
	 */
483
	if (machine__is_host(machine))
484
		event->header.misc = PERF_RECORD_MISC_KERNEL;
485
	else
486
		event->header.misc = PERF_RECORD_MISC_GUEST_KERNEL;
487

488
	for (pos = maps__first(maps); pos; pos = map__next(pos)) {
489 490
		size_t size;

491
		if (__map__is_kernel(pos))
492 493
			continue;

494
		size = PERF_ALIGN(pos->dso->long_name_len + 1, sizeof(u64));
495 496 497
		event->mmap.header.type = PERF_RECORD_MMAP;
		event->mmap.header.size = (sizeof(event->mmap) -
				        (sizeof(event->mmap.filename) - size));
498 499
		memset(event->mmap.filename + size, 0, machine->id_hdr_size);
		event->mmap.header.size += machine->id_hdr_size;
500 501 502 503 504
		event->mmap.start = pos->start;
		event->mmap.len   = pos->end - pos->start;
		event->mmap.pid   = machine->pid;

		memcpy(event->mmap.filename, pos->dso->long_name,
505
		       pos->dso->long_name_len + 1);
506
		if (perf_tool__process_synth_event(tool, event, machine, process) != 0) {
507 508 509
			rc = -1;
			break;
		}
510 511
	}

512
	free(event);
513
	return rc;
514 515
}

516 517
static int __event__synthesize_thread(union perf_event *comm_event,
				      union perf_event *mmap_event,
518
				      union perf_event *fork_event,
519
				      union perf_event *namespaces_event,
520
				      pid_t pid, int full,
521
				      perf_event__handler_t process,
522
				      struct perf_tool *tool,
523 524 525
				      struct machine *machine,
				      bool mmap_data,
				      unsigned int proc_map_timeout)
526
{
527 528
	char filename[PATH_MAX];
	DIR *tasks;
529
	struct dirent *dirent;
530
	pid_t tgid, ppid;
531
	int rc = 0;
532 533 534 535 536 537 538 539 540

	/* special case: only send one comm event using passed in pid */
	if (!full) {
		tgid = perf_event__synthesize_comm(tool, comm_event, pid,
						   process, machine);

		if (tgid == -1)
			return -1;

541 542 543 544 545
		if (perf_event__synthesize_namespaces(tool, namespaces_event, pid,
						      tgid, process, machine) < 0)
			return -1;


546
		return perf_event__synthesize_mmap_events(tool, mmap_event, pid, tgid,
547 548
							  process, machine, mmap_data,
							  proc_map_timeout);
549 550 551 552 553 554 555 556 557 558 559 560 561 562
	}

	if (machine__is_default_guest(machine))
		return 0;

	snprintf(filename, sizeof(filename), "%s/proc/%d/task",
		 machine->root_dir, pid);

	tasks = opendir(filename);
	if (tasks == NULL) {
		pr_debug("couldn't open %s\n", filename);
		return 0;
	}

563
	while ((dirent = readdir(tasks)) != NULL) {
564 565 566
		char *end;
		pid_t _pid;

567
		_pid = strtol(dirent->d_name, &end, 10);
568 569 570
		if (*end)
			continue;

571
		rc = -1;
572 573
		if (perf_event__prepare_comm(comm_event, _pid, machine,
					     &tgid, &ppid) != 0)
574
			break;
575

576
		if (perf_event__synthesize_fork(tool, fork_event, _pid, tgid,
577
						ppid, process, machine) < 0)
578
			break;
579 580 581 582 583

		if (perf_event__synthesize_namespaces(tool, namespaces_event, _pid,
						      tgid, process, machine) < 0)
			break;

584 585 586
		/*
		 * Send the prepared comm event
		 */
587
		if (perf_tool__process_synth_event(tool, comm_event, machine, process) != 0)
588
			break;
589

590
		rc = 0;
591 592 593
		if (_pid == pid) {
			/* process the parent's maps too */
			rc = perf_event__synthesize_mmap_events(tool, mmap_event, pid, tgid,
594
						process, machine, mmap_data, proc_map_timeout);
595 596
			if (rc)
				break;
597
		}
598 599 600
	}

	closedir(tasks);
601
	return rc;
602 603
}

604
int perf_event__synthesize_thread_map(struct perf_tool *tool,
605
				      struct thread_map *threads,
606
				      perf_event__handler_t process,
607
				      struct machine *machine,
608 609
				      bool mmap_data,
				      unsigned int proc_map_timeout)
610
{
611
	union perf_event *comm_event, *mmap_event, *fork_event;
612
	union perf_event *namespaces_event;
613
	int err = -1, thread, j;
614

615
	comm_event = malloc(sizeof(comm_event->comm) + machine->id_hdr_size);
616 617 618
	if (comm_event == NULL)
		goto out;

619
	mmap_event = malloc(sizeof(mmap_event->mmap2) + machine->id_hdr_size);
620 621 622
	if (mmap_event == NULL)
		goto out_free_comm;

623 624 625 626
	fork_event = malloc(sizeof(fork_event->fork) + machine->id_hdr_size);
	if (fork_event == NULL)
		goto out_free_mmap;

627 628 629 630 631 632
	namespaces_event = malloc(sizeof(namespaces_event->namespaces) +
				  (NR_NAMESPACES * sizeof(struct perf_ns_link_info)) +
				  machine->id_hdr_size);
	if (namespaces_event == NULL)
		goto out_free_fork;

633 634 635
	err = 0;
	for (thread = 0; thread < threads->nr; ++thread) {
		if (__event__synthesize_thread(comm_event, mmap_event,
636
					       fork_event, namespaces_event,
637
					       thread_map__pid(threads, thread), 0,
638
					       process, tool, machine,
639
					       mmap_data, proc_map_timeout)) {
640 641 642
			err = -1;
			break;
		}
643 644 645 646 647

		/*
		 * comm.pid is set to thread group id by
		 * perf_event__synthesize_comm
		 */
648
		if ((int) comm_event->comm.pid != thread_map__pid(threads, thread)) {
649 650 651 652
			bool need_leader = true;

			/* is thread group leader in thread_map? */
			for (j = 0; j < threads->nr; ++j) {
653
				if ((int) comm_event->comm.pid == thread_map__pid(threads, j)) {
654 655 656 657 658 659 660
					need_leader = false;
					break;
				}
			}

			/* if not, generate events for it */
			if (need_leader &&
661
			    __event__synthesize_thread(comm_event, mmap_event,
662
						       fork_event, namespaces_event,
663 664
						       comm_event->comm.pid, 0,
						       process, tool, machine,
665
						       mmap_data, proc_map_timeout)) {
666 667 668 669
				err = -1;
				break;
			}
		}
670
	}
671 672
	free(namespaces_event);
out_free_fork:
673 674
	free(fork_event);
out_free_mmap:
675 676 677 678 679 680 681
	free(mmap_event);
out_free_comm:
	free(comm_event);
out:
	return err;
}

682 683 684 685 686 687 688 689
static int __perf_event__synthesize_threads(struct perf_tool *tool,
					    perf_event__handler_t process,
					    struct machine *machine,
					    bool mmap_data,
					    unsigned int proc_map_timeout,
					    struct dirent **dirent,
					    int start,
					    int num)
690
{
691
	union perf_event *comm_event, *mmap_event, *fork_event;
692
	union perf_event *namespaces_event;
693
	int err = -1;
694 695
	char *end;
	pid_t pid;
696
	int i;
697

698
	comm_event = malloc(sizeof(comm_event->comm) + machine->id_hdr_size);
699 700 701
	if (comm_event == NULL)
		goto out;

702
	mmap_event = malloc(sizeof(mmap_event->mmap2) + machine->id_hdr_size);
703 704
	if (mmap_event == NULL)
		goto out_free_comm;
705

706 707 708 709
	fork_event = malloc(sizeof(fork_event->fork) + machine->id_hdr_size);
	if (fork_event == NULL)
		goto out_free_mmap;

710 711 712 713 714 715
	namespaces_event = malloc(sizeof(namespaces_event->namespaces) +
				  (NR_NAMESPACES * sizeof(struct perf_ns_link_info)) +
				  machine->id_hdr_size);
	if (namespaces_event == NULL)
		goto out_free_fork;

716
	for (i = start; i < start + num; i++) {
717
		if (!isdigit(dirent[i]->d_name[0]))
718 719
			continue;

720 721
		pid = (pid_t)strtol(dirent[i]->d_name, &end, 10);
		/* only interested in proper numerical dirents */
722 723 724 725 726 727 728 729 730 731
		if (*end)
			continue;
		/*
		 * We may race with exiting thread, so don't stop just because
		 * one thread couldn't be synthesized.
		 */
		__event__synthesize_thread(comm_event, mmap_event, fork_event,
					   namespaces_event, pid, 1, process,
					   tool, machine, mmap_data,
					   proc_map_timeout);
732
	}
733
	err = 0;
734

735
	free(namespaces_event);
736 737
out_free_fork:
	free(fork_event);
738 739 740 741 742 743
out_free_mmap:
	free(mmap_event);
out_free_comm:
	free(comm_event);
out:
	return err;
744
}
745

746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793
struct synthesize_threads_arg {
	struct perf_tool *tool;
	perf_event__handler_t process;
	struct machine *machine;
	bool mmap_data;
	unsigned int proc_map_timeout;
	struct dirent **dirent;
	int num;
	int start;
};

static void *synthesize_threads_worker(void *arg)
{
	struct synthesize_threads_arg *args = arg;

	__perf_event__synthesize_threads(args->tool, args->process,
					 args->machine, args->mmap_data,
					 args->proc_map_timeout, args->dirent,
					 args->start, args->num);
	return NULL;
}

int perf_event__synthesize_threads(struct perf_tool *tool,
				   perf_event__handler_t process,
				   struct machine *machine,
				   bool mmap_data,
				   unsigned int proc_map_timeout,
				   unsigned int nr_threads_synthesize)
{
	struct synthesize_threads_arg *args = NULL;
	pthread_t *synthesize_threads = NULL;
	char proc_path[PATH_MAX];
	struct dirent **dirent;
	int num_per_thread;
	int m, n, i, j;
	int thread_nr;
	int base = 0;
	int err = -1;


	if (machine__is_default_guest(machine))
		return 0;

	snprintf(proc_path, sizeof(proc_path), "%s/proc", machine->root_dir);
	n = scandir(proc_path, &dirent, 0, alphasort);
	if (n < 0)
		return err;

794 795 796 797
	if (nr_threads_synthesize == UINT_MAX)
		thread_nr = sysconf(_SC_NPROCESSORS_ONLN);
	else
		thread_nr = nr_threads_synthesize;
798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857

	if (thread_nr <= 1) {
		err = __perf_event__synthesize_threads(tool, process,
						       machine, mmap_data,
						       proc_map_timeout,
						       dirent, base, n);
		goto free_dirent;
	}
	if (thread_nr > n)
		thread_nr = n;

	synthesize_threads = calloc(sizeof(pthread_t), thread_nr);
	if (synthesize_threads == NULL)
		goto free_dirent;

	args = calloc(sizeof(*args), thread_nr);
	if (args == NULL)
		goto free_threads;

	num_per_thread = n / thread_nr;
	m = n % thread_nr;
	for (i = 0; i < thread_nr; i++) {
		args[i].tool = tool;
		args[i].process = process;
		args[i].machine = machine;
		args[i].mmap_data = mmap_data;
		args[i].proc_map_timeout = proc_map_timeout;
		args[i].dirent = dirent;
	}
	for (i = 0; i < m; i++) {
		args[i].num = num_per_thread + 1;
		args[i].start = i * args[i].num;
	}
	if (i != 0)
		base = args[i-1].start + args[i-1].num;
	for (j = i; j < thread_nr; j++) {
		args[j].num = num_per_thread;
		args[j].start = base + (j - i) * args[i].num;
	}

	for (i = 0; i < thread_nr; i++) {
		if (pthread_create(&synthesize_threads[i], NULL,
				   synthesize_threads_worker, &args[i]))
			goto out_join;
	}
	err = 0;
out_join:
	for (i = 0; i < thread_nr; i++)
		pthread_join(synthesize_threads[i], NULL);
	free(args);
free_threads:
	free(synthesize_threads);
free_dirent:
	for (i = 0; i < n; i++)
		free(dirent[i]);
	free(dirent);

	return err;
}

858 859 860 861 862
struct process_symbol_args {
	const char *name;
	u64	   start;
};

863
static int find_symbol_cb(void *arg, const char *name, char type,
864
			  u64 start)
865 866 867
{
	struct process_symbol_args *args = arg;

868 869 870 871 872 873
	/*
	 * Must be a function or at least an alias, as in PARISC64, where "_text" is
	 * an 'A' to the same address as "_stext".
	 */
	if (!(symbol_type__is_a(type, MAP__FUNCTION) ||
	      type == 'A') || strcmp(name, args->name))
874 875 876 877 878 879
		return 0;

	args->start = start;
	return 1;
}

880 881
int kallsyms__get_function_start(const char *kallsyms_filename,
				 const char *symbol_name, u64 *addr)
882 883 884 885
{
	struct process_symbol_args args = { .name = symbol_name, };

	if (kallsyms__parse(kallsyms_filename, &args, find_symbol_cb) <= 0)
886
		return -1;
887

888 889
	*addr = args.start;
	return 0;
890 891
}

892
int perf_event__synthesize_kernel_mmap(struct perf_tool *tool,
893
				       perf_event__handler_t process,
894
				       struct machine *machine)
895 896
{
	size_t size;
897
	struct map *map = machine__kernel_map(machine);
898
	struct kmap *kmap;
899
	int err;
900 901
	union perf_event *event;

902 903
	if (symbol_conf.kptr_restrict)
		return -1;
904
	if (map == NULL)
905 906
		return -1;

907 908 909 910 911
	/*
	 * We should get this from /sys/kernel/sections/.text, but till that is
	 * available use this, and after it is use this as a fallback for older
	 * kernels.
	 */
912
	event = zalloc((sizeof(event->mmap) + machine->id_hdr_size));
913 914 915 916 917
	if (event == NULL) {
		pr_debug("Not enough memory synthesizing mmap event "
			 "for kernel modules\n");
		return -1;
	}
918

919
	if (machine__is_host(machine)) {
920 921 922 923
		/*
		 * kernel uses PERF_RECORD_MISC_USER for user space maps,
		 * see kernel/perf_event.c __perf_event_mmap
		 */
924
		event->header.misc = PERF_RECORD_MISC_KERNEL;
925
	} else {
926
		event->header.misc = PERF_RECORD_MISC_GUEST_KERNEL;
927
	}
928

929
	kmap = map__kmap(map);
930
	size = snprintf(event->mmap.filename, sizeof(event->mmap.filename),
931
			"%s%s", machine->mmap_name, kmap->ref_reloc_sym->name) + 1;
932
	size = PERF_ALIGN(size, sizeof(u64));
933 934
	event->mmap.header.type = PERF_RECORD_MMAP;
	event->mmap.header.size = (sizeof(event->mmap) -
935
			(sizeof(event->mmap.filename) - size) + machine->id_hdr_size);
936
	event->mmap.pgoff = kmap->ref_reloc_sym->addr;
937 938 939 940
	event->mmap.start = map->start;
	event->mmap.len   = map->end - event->mmap.start;
	event->mmap.pid   = machine->pid;

941
	err = perf_tool__process_synth_event(tool, event, machine, process);
942 943 944
	free(event);

	return err;
945 946
}

947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982
int perf_event__synthesize_thread_map2(struct perf_tool *tool,
				      struct thread_map *threads,
				      perf_event__handler_t process,
				      struct machine *machine)
{
	union perf_event *event;
	int i, err, size;

	size  = sizeof(event->thread_map);
	size +=	threads->nr * sizeof(event->thread_map.entries[0]);

	event = zalloc(size);
	if (!event)
		return -ENOMEM;

	event->header.type = PERF_RECORD_THREAD_MAP;
	event->header.size = size;
	event->thread_map.nr = threads->nr;

	for (i = 0; i < threads->nr; i++) {
		struct thread_map_event_entry *entry = &event->thread_map.entries[i];
		char *comm = thread_map__comm(threads, i);

		if (!comm)
			comm = (char *) "";

		entry->pid = thread_map__pid(threads, i);
		strncpy((char *) &entry->comm, comm, sizeof(entry->comm));
	}

	err = process(tool, event, NULL, machine);

	free(event);
	return err;
}

983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 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 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113
static void synthesize_cpus(struct cpu_map_entries *cpus,
			    struct cpu_map *map)
{
	int i;

	cpus->nr = map->nr;

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

static void synthesize_mask(struct cpu_map_mask *mask,
			    struct cpu_map *map, int max)
{
	int i;

	mask->nr = BITS_TO_LONGS(max);
	mask->long_size = sizeof(long);

	for (i = 0; i < map->nr; i++)
		set_bit(map->map[i], mask->mask);
}

static size_t cpus_size(struct cpu_map *map)
{
	return sizeof(struct cpu_map_entries) + map->nr * sizeof(u16);
}

static size_t mask_size(struct cpu_map *map, int *max)
{
	int i;

	*max = 0;

	for (i = 0; i < map->nr; i++) {
		/* bit possition of the cpu is + 1 */
		int bit = map->map[i] + 1;

		if (bit > *max)
			*max = bit;
	}

	return sizeof(struct cpu_map_mask) + BITS_TO_LONGS(*max) * sizeof(long);
}

void *cpu_map_data__alloc(struct cpu_map *map, size_t *size, u16 *type, int *max)
{
	size_t size_cpus, size_mask;
	bool is_dummy = cpu_map__empty(map);

	/*
	 * Both array and mask data have variable size based
	 * on the number of cpus and their actual values.
	 * The size of the 'struct cpu_map_data' is:
	 *
	 *   array = size of 'struct cpu_map_entries' +
	 *           number of cpus * sizeof(u64)
	 *
	 *   mask  = size of 'struct cpu_map_mask' +
	 *           maximum cpu bit converted to size of longs
	 *
	 * and finaly + the size of 'struct cpu_map_data'.
	 */
	size_cpus = cpus_size(map);
	size_mask = mask_size(map, max);

	if (is_dummy || (size_cpus < size_mask)) {
		*size += size_cpus;
		*type  = PERF_CPU_MAP__CPUS;
	} else {
		*size += size_mask;
		*type  = PERF_CPU_MAP__MASK;
	}

	*size += sizeof(struct cpu_map_data);
	return zalloc(*size);
}

void cpu_map_data__synthesize(struct cpu_map_data *data, struct cpu_map *map,
			      u16 type, int max)
{
	data->type = type;

	switch (type) {
	case PERF_CPU_MAP__CPUS:
		synthesize_cpus((struct cpu_map_entries *) data->data, map);
		break;
	case PERF_CPU_MAP__MASK:
		synthesize_mask((struct cpu_map_mask *) data->data, map, max);
	default:
		break;
	};
}

static struct cpu_map_event* cpu_map_event__new(struct cpu_map *map)
{
	size_t size = sizeof(struct cpu_map_event);
	struct cpu_map_event *event;
	int max;
	u16 type;

	event = cpu_map_data__alloc(map, &size, &type, &max);
	if (!event)
		return NULL;

	event->header.type = PERF_RECORD_CPU_MAP;
	event->header.size = size;
	event->data.type   = type;

	cpu_map_data__synthesize(&event->data, map, type, max);
	return event;
}

int perf_event__synthesize_cpu_map(struct perf_tool *tool,
				   struct cpu_map *map,
				   perf_event__handler_t process,
				   struct machine *machine)
{
	struct cpu_map_event *event;
	int err;

	event = cpu_map_event__new(map);
	if (!event)
		return -ENOMEM;

	err = process(tool, (union perf_event *) event, NULL, machine);

	free(event);
	return err;
}

1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151
int perf_event__synthesize_stat_config(struct perf_tool *tool,
				       struct perf_stat_config *config,
				       perf_event__handler_t process,
				       struct machine *machine)
{
	struct stat_config_event *event;
	int size, i = 0, err;

	size  = sizeof(*event);
	size += (PERF_STAT_CONFIG_TERM__MAX * sizeof(event->data[0]));

	event = zalloc(size);
	if (!event)
		return -ENOMEM;

	event->header.type = PERF_RECORD_STAT_CONFIG;
	event->header.size = size;
	event->nr          = PERF_STAT_CONFIG_TERM__MAX;

#define ADD(__term, __val)					\
	event->data[i].tag = PERF_STAT_CONFIG_TERM__##__term;	\
	event->data[i].val = __val;				\
	i++;

	ADD(AGGR_MODE,	config->aggr_mode)
	ADD(INTERVAL,	config->interval)
	ADD(SCALE,	config->scale)

	WARN_ONCE(i != PERF_STAT_CONFIG_TERM__MAX,
		  "stat config terms unbalanced\n");
#undef ADD

	err = process(tool, (union perf_event *) event, NULL, machine);

	free(event);
	return err;
}

1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173
int perf_event__synthesize_stat(struct perf_tool *tool,
				u32 cpu, u32 thread, u64 id,
				struct perf_counts_values *count,
				perf_event__handler_t process,
				struct machine *machine)
{
	struct stat_event event;

	event.header.type = PERF_RECORD_STAT;
	event.header.size = sizeof(event);
	event.header.misc = 0;

	event.id        = id;
	event.cpu       = cpu;
	event.thread    = thread;
	event.val       = count->val;
	event.ena       = count->ena;
	event.run       = count->run;

	return process(tool, (union perf_event *) &event, NULL, machine);
}

1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190
int perf_event__synthesize_stat_round(struct perf_tool *tool,
				      u64 evtime, u64 type,
				      perf_event__handler_t process,
				      struct machine *machine)
{
	struct stat_round_event event;

	event.header.type = PERF_RECORD_STAT_ROUND;
	event.header.size = sizeof(event);
	event.header.misc = 0;

	event.time = evtime;
	event.type = type;

	return process(tool, (union perf_event *) &event, NULL, machine);
}

1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214
void perf_event__read_stat_config(struct perf_stat_config *config,
				  struct stat_config_event *event)
{
	unsigned i;

	for (i = 0; i < event->nr; i++) {

		switch (event->data[i].tag) {
#define CASE(__term, __val)					\
		case PERF_STAT_CONFIG_TERM__##__term:		\
			config->__val = event->data[i].val;	\
			break;

		CASE(AGGR_MODE, aggr_mode)
		CASE(SCALE,     scale)
		CASE(INTERVAL,  interval)
#undef CASE
		default:
			pr_warning("unknown stat config term %" PRIu64 "\n",
				   event->data[i].tag);
		}
	}
}

1215 1216
size_t perf_event__fprintf_comm(union perf_event *event, FILE *fp)
{
1217 1218 1219 1220 1221 1222 1223
	const char *s;

	if (event->header.misc & PERF_RECORD_MISC_COMM_EXEC)
		s = " exec";
	else
		s = "";

1224
	return fprintf(fp, "%s: %s:%d/%d\n", s, event->comm.comm, event->comm.pid, event->comm.tid);
1225 1226
}

1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253
size_t perf_event__fprintf_namespaces(union perf_event *event, FILE *fp)
{
	size_t ret = 0;
	struct perf_ns_link_info *ns_link_info;
	u32 nr_namespaces, idx;

	ns_link_info = event->namespaces.link_info;
	nr_namespaces = event->namespaces.nr_namespaces;

	ret += fprintf(fp, " %d/%d - nr_namespaces: %u\n\t\t[",
		       event->namespaces.pid,
		       event->namespaces.tid,
		       nr_namespaces);

	for (idx = 0; idx < nr_namespaces; idx++) {
		if (idx && (idx % 4 == 0))
			ret += fprintf(fp, "\n\t\t ");

		ret  += fprintf(fp, "%u/%s: %" PRIu64 "/%#" PRIx64 "%s", idx,
				perf_ns__name(idx), (u64)ns_link_info[idx].dev,
				(u64)ns_link_info[idx].ino,
				((idx + 1) != nr_namespaces) ? ", " : "]\n");
	}

	return ret;
}

1254
int perf_event__process_comm(struct perf_tool *tool __maybe_unused,
1255
			     union perf_event *event,
1256
			     struct perf_sample *sample,
1257
			     struct machine *machine)
1258
{
1259
	return machine__process_comm_event(machine, event, sample);
1260 1261
}

1262 1263 1264 1265 1266 1267 1268 1269
int perf_event__process_namespaces(struct perf_tool *tool __maybe_unused,
				   union perf_event *event,
				   struct perf_sample *sample,
				   struct machine *machine)
{
	return machine__process_namespaces_event(machine, event, sample);
}

1270
int perf_event__process_lost(struct perf_tool *tool __maybe_unused,
1271
			     union perf_event *event,
1272
			     struct perf_sample *sample,
1273
			     struct machine *machine)
1274
{
1275
	return machine__process_lost_event(machine, event, sample);
1276
}
1277

1278 1279 1280 1281 1282 1283 1284 1285
int perf_event__process_aux(struct perf_tool *tool __maybe_unused,
			    union perf_event *event,
			    struct perf_sample *sample __maybe_unused,
			    struct machine *machine)
{
	return machine__process_aux_event(machine, event);
}

1286 1287 1288 1289 1290 1291 1292 1293
int perf_event__process_itrace_start(struct perf_tool *tool __maybe_unused,
				     union perf_event *event,
				     struct perf_sample *sample __maybe_unused,
				     struct machine *machine)
{
	return machine__process_itrace_start_event(machine, event);
}

1294 1295 1296 1297 1298 1299 1300 1301
int perf_event__process_lost_samples(struct perf_tool *tool __maybe_unused,
				     union perf_event *event,
				     struct perf_sample *sample,
				     struct machine *machine)
{
	return machine__process_lost_samples_event(machine, event, sample);
}

1302 1303 1304 1305 1306 1307 1308 1309
int perf_event__process_switch(struct perf_tool *tool __maybe_unused,
			       union perf_event *event,
			       struct perf_sample *sample __maybe_unused,
			       struct machine *machine)
{
	return machine__process_switch_event(machine, event);
}

1310 1311
size_t perf_event__fprintf_mmap(union perf_event *event, FILE *fp)
{
1312
	return fprintf(fp, " %d/%d: [%#" PRIx64 "(%#" PRIx64 ") @ %#" PRIx64 "]: %c %s\n",
1313
		       event->mmap.pid, event->mmap.tid, event->mmap.start,
1314 1315 1316
		       event->mmap.len, event->mmap.pgoff,
		       (event->header.misc & PERF_RECORD_MISC_MMAP_DATA) ? 'r' : 'x',
		       event->mmap.filename);
1317 1318
}

1319 1320 1321
size_t perf_event__fprintf_mmap2(union perf_event *event, FILE *fp)
{
	return fprintf(fp, " %d/%d: [%#" PRIx64 "(%#" PRIx64 ") @ %#" PRIx64
1322
			   " %02x:%02x %"PRIu64" %"PRIu64"]: %c%c%c%c %s\n",
1323 1324 1325 1326
		       event->mmap2.pid, event->mmap2.tid, event->mmap2.start,
		       event->mmap2.len, event->mmap2.pgoff, event->mmap2.maj,
		       event->mmap2.min, event->mmap2.ino,
		       event->mmap2.ino_generation,
1327 1328 1329 1330
		       (event->mmap2.prot & PROT_READ) ? 'r' : '-',
		       (event->mmap2.prot & PROT_WRITE) ? 'w' : '-',
		       (event->mmap2.prot & PROT_EXEC) ? 'x' : '-',
		       (event->mmap2.flags & MAP_SHARED) ? 's' : 'p',
1331 1332 1333
		       event->mmap2.filename);
}

1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349
size_t perf_event__fprintf_thread_map(union perf_event *event, FILE *fp)
{
	struct thread_map *threads = thread_map__new_event(&event->thread_map);
	size_t ret;

	ret = fprintf(fp, " nr: ");

	if (threads)
		ret += thread_map__fprintf(threads, fp);
	else
		ret += fprintf(fp, "failed to get threads from event\n");

	thread_map__put(threads);
	return ret;
}

1350 1351 1352 1353 1354
size_t perf_event__fprintf_cpu_map(union perf_event *event, FILE *fp)
{
	struct cpu_map *cpus = cpu_map__new_data(&event->cpu_map.data);
	size_t ret;

1355
	ret = fprintf(fp, ": ");
1356 1357 1358 1359 1360 1361 1362 1363 1364 1365

	if (cpus)
		ret += cpu_map__fprintf(cpus, fp);
	else
		ret += fprintf(fp, "failed to get cpumap from event\n");

	cpu_map__put(cpus);
	return ret;
}

1366
int perf_event__process_mmap(struct perf_tool *tool __maybe_unused,
1367
			     union perf_event *event,
1368
			     struct perf_sample *sample,
1369
			     struct machine *machine)
1370
{
1371
	return machine__process_mmap_event(machine, event, sample);
1372 1373
}

1374 1375
int perf_event__process_mmap2(struct perf_tool *tool __maybe_unused,
			     union perf_event *event,
1376
			     struct perf_sample *sample,
1377 1378
			     struct machine *machine)
{
1379
	return machine__process_mmap2_event(machine, event, sample);
1380 1381
}

1382 1383 1384 1385 1386 1387 1388
size_t perf_event__fprintf_task(union perf_event *event, FILE *fp)
{
	return fprintf(fp, "(%d:%d):(%d:%d)\n",
		       event->fork.pid, event->fork.tid,
		       event->fork.ppid, event->fork.ptid);
}

1389
int perf_event__process_fork(struct perf_tool *tool __maybe_unused,
1390
			     union perf_event *event,
1391
			     struct perf_sample *sample,
1392
			     struct machine *machine)
1393
{
1394
	return machine__process_fork_event(machine, event, sample);
1395
}
1396

1397 1398
int perf_event__process_exit(struct perf_tool *tool __maybe_unused,
			     union perf_event *event,
1399
			     struct perf_sample *sample,
1400 1401
			     struct machine *machine)
{
1402
	return machine__process_exit_event(machine, event, sample);
1403 1404
}

1405 1406
size_t perf_event__fprintf_aux(union perf_event *event, FILE *fp)
{
1407
	return fprintf(fp, " offset: %#"PRIx64" size: %#"PRIx64" flags: %#"PRIx64" [%s%s%s]\n",
1408 1409 1410
		       event->aux.aux_offset, event->aux.aux_size,
		       event->aux.flags,
		       event->aux.flags & PERF_AUX_FLAG_TRUNCATED ? "T" : "",
1411 1412
		       event->aux.flags & PERF_AUX_FLAG_OVERWRITE ? "O" : "",
		       event->aux.flags & PERF_AUX_FLAG_PARTIAL   ? "P" : "");
1413 1414
}

1415 1416 1417 1418 1419 1420
size_t perf_event__fprintf_itrace_start(union perf_event *event, FILE *fp)
{
	return fprintf(fp, " pid: %u tid: %u\n",
		       event->itrace_start.pid, event->itrace_start.tid);
}

1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434
size_t perf_event__fprintf_switch(union perf_event *event, FILE *fp)
{
	bool out = event->header.misc & PERF_RECORD_MISC_SWITCH_OUT;
	const char *in_out = out ? "OUT" : "IN ";

	if (event->header.type == PERF_RECORD_SWITCH)
		return fprintf(fp, " %s\n", in_out);

	return fprintf(fp, " %s  %s pid/tid: %5u/%-5u\n",
		       in_out, out ? "next" : "prev",
		       event->context_switch.next_prev_pid,
		       event->context_switch.next_prev_tid);
}

1435 1436 1437 1438 1439
static size_t perf_event__fprintf_lost(union perf_event *event, FILE *fp)
{
	return fprintf(fp, " lost %" PRIu64 "\n", event->lost.lost);
}

1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454
size_t perf_event__fprintf(union perf_event *event, FILE *fp)
{
	size_t ret = fprintf(fp, "PERF_RECORD_%s",
			     perf_event__name(event->header.type));

	switch (event->header.type) {
	case PERF_RECORD_COMM:
		ret += perf_event__fprintf_comm(event, fp);
		break;
	case PERF_RECORD_FORK:
	case PERF_RECORD_EXIT:
		ret += perf_event__fprintf_task(event, fp);
		break;
	case PERF_RECORD_MMAP:
		ret += perf_event__fprintf_mmap(event, fp);
1455 1456 1457
		break;
	case PERF_RECORD_NAMESPACES:
		ret += perf_event__fprintf_namespaces(event, fp);
1458
		break;
1459 1460 1461
	case PERF_RECORD_MMAP2:
		ret += perf_event__fprintf_mmap2(event, fp);
		break;
1462 1463 1464
	case PERF_RECORD_AUX:
		ret += perf_event__fprintf_aux(event, fp);
		break;
1465 1466 1467
	case PERF_RECORD_ITRACE_START:
		ret += perf_event__fprintf_itrace_start(event, fp);
		break;
1468 1469 1470 1471
	case PERF_RECORD_SWITCH:
	case PERF_RECORD_SWITCH_CPU_WIDE:
		ret += perf_event__fprintf_switch(event, fp);
		break;
1472 1473 1474
	case PERF_RECORD_LOST:
		ret += perf_event__fprintf_lost(event, fp);
		break;
1475 1476 1477 1478 1479 1480 1481
	default:
		ret += fprintf(fp, "\n");
	}

	return ret;
}

1482 1483
int perf_event__process(struct perf_tool *tool __maybe_unused,
			union perf_event *event,
1484
			struct perf_sample *sample,
1485
			struct machine *machine)
1486
{
1487
	return machine__process_event(machine, event, sample);
1488 1489
}

1490
void thread__find_addr_map(struct thread *thread, u8 cpumode,
1491
			   enum map_type type, u64 addr,
1492
			   struct addr_location *al)
1493
{
1494
	struct map_groups *mg = thread->mg;
1495
	struct machine *machine = mg->machine;
1496
	bool load_map = false;
1497

1498
	al->machine = machine;
1499
	al->thread = thread;
1500
	al->addr = addr;
1501
	al->cpumode = cpumode;
1502
	al->filtered = 0;
1503

1504 1505 1506 1507 1508
	if (machine == NULL) {
		al->map = NULL;
		return;
	}

1509
	if (cpumode == PERF_RECORD_MISC_KERNEL && perf_host) {
1510
		al->level = 'k';
1511
		mg = &machine->kmaps;
1512
		load_map = true;
1513
	} else if (cpumode == PERF_RECORD_MISC_USER && perf_host) {
1514
		al->level = '.';
1515 1516
	} else if (cpumode == PERF_RECORD_MISC_GUEST_KERNEL && perf_guest) {
		al->level = 'g';
1517
		mg = &machine->kmaps;
1518
		load_map = true;
1519 1520
	} else if (cpumode == PERF_RECORD_MISC_GUEST_USER && perf_guest) {
		al->level = 'u';
1521
	} else {
1522
		al->level = 'H';
1523
		al->map = NULL;
1524 1525 1526 1527

		if ((cpumode == PERF_RECORD_MISC_GUEST_USER ||
			cpumode == PERF_RECORD_MISC_GUEST_KERNEL) &&
			!perf_guest)
1528
			al->filtered |= (1 << HIST_FILTER__GUEST);
1529 1530 1531
		if ((cpumode == PERF_RECORD_MISC_USER ||
			cpumode == PERF_RECORD_MISC_KERNEL) &&
			!perf_host)
1532
			al->filtered |= (1 << HIST_FILTER__HOST);
1533

1534 1535 1536
		return;
	}
try_again:
1537
	al->map = map_groups__find(mg, type, al->addr);
1538 1539 1540 1541 1542 1543 1544 1545 1546 1547
	if (al->map == NULL) {
		/*
		 * If this is outside of all known maps, and is a negative
		 * address, try to look it up in the kernel dso, as it might be
		 * a vsyscall or vdso (which executes in user-mode).
		 *
		 * XXX This is nasty, we should have a symbol list in the
		 * "[vdso]" dso, but for now lets use the old trick of looking
		 * in the whole kernel symbol list.
		 */
1548 1549 1550
		if (cpumode == PERF_RECORD_MISC_USER && machine &&
		    mg != &machine->kmaps &&
		    machine__kernel_ip(machine, al->addr)) {
1551
			mg = &machine->kmaps;
1552
			load_map = true;
1553 1554
			goto try_again;
		}
1555 1556 1557 1558 1559 1560
	} else {
		/*
		 * Kernel maps might be changed when loading symbols so loading
		 * must be done prior to using kernel maps.
		 */
		if (load_map)
1561
			map__load(al->map);
1562
		al->addr = al->map->map_ip(al->map, al->addr);
1563
	}
1564 1565
}

1566
void thread__find_addr_location(struct thread *thread,
1567
				u8 cpumode, enum map_type type, u64 addr,
1568
				struct addr_location *al)
1569
{
1570
	thread__find_addr_map(thread, cpumode, type, addr, al);
1571
	if (al->map != NULL)
1572
		al->sym = map__find_symbol(al->map, al->addr);
1573 1574
	else
		al->sym = NULL;
1575 1576
}

1577 1578 1579 1580
/*
 * Callers need to drop the reference to al->thread, obtained in
 * machine__findnew_thread()
 */
1581 1582
int machine__resolve(struct machine *machine, struct addr_location *al,
		     struct perf_sample *sample)
1583
{
1584
	struct thread *thread = machine__findnew_thread(machine, sample->pid,
1585
							sample->tid);
1586

1587 1588 1589
	if (thread == NULL)
		return -1;

1590
	dump_printf(" ... thread: %s:%d\n", thread__comm_str(thread), thread->tid);
1591
	thread__find_addr_map(thread, sample->cpumode, MAP__FUNCTION, sample->ip, al);
1592 1593 1594
	dump_printf(" ...... dso: %s\n",
		    al->map ? al->map->dso->long_name :
			al->level == 'H' ? "[hypervisor]" : "<not found>");
1595 1596 1597 1598

	if (thread__is_filtered(thread))
		al->filtered |= (1 << HIST_FILTER__THREAD);

1599
	al->sym = NULL;
1600
	al->cpu = sample->cpu;
1601
	al->socket = -1;
1602
	al->srcline = NULL;
1603 1604 1605 1606 1607 1608 1609

	if (al->cpu >= 0) {
		struct perf_env *env = machine->env;

		if (env && env->cpu)
			al->socket = env->cpu[al->cpu].socket_id;
	}
1610 1611

	if (al->map) {
1612 1613
		struct dso *dso = al->map->dso;

1614
		if (symbol_conf.dso_list &&
1615 1616 1617 1618
		    (!dso || !(strlist__has_entry(symbol_conf.dso_list,
						  dso->short_name) ||
			       (dso->short_name != dso->long_name &&
				strlist__has_entry(symbol_conf.dso_list,
1619 1620 1621
						   dso->long_name))))) {
			al->filtered |= (1 << HIST_FILTER__DSO);
		}
1622

1623
		al->sym = map__find_symbol(al->map, al->addr);
1624
	}
1625

1626 1627
	if (symbol_conf.sym_list &&
		(!al->sym || !strlist__has_entry(symbol_conf.sym_list,
1628 1629 1630
						al->sym->name))) {
		al->filtered |= (1 << HIST_FILTER__SYMBOL);
	}
1631 1632

	return 0;
1633
}
1634

1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645
/*
 * The preprocess_sample method will return with reference counts for the
 * in it, when done using (and perhaps getting ref counts if needing to
 * keep a pointer to one of those entries) it must be paired with
 * addr_location__put(), so that the refcounts can be decremented.
 */
void addr_location__put(struct addr_location *al)
{
	thread__zput(al->thread);
}

1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666
bool is_bts_event(struct perf_event_attr *attr)
{
	return attr->type == PERF_TYPE_HARDWARE &&
	       (attr->config & PERF_COUNT_HW_BRANCH_INSTRUCTIONS) &&
	       attr->sample_period == 1;
}

bool sample_addr_correlates_sym(struct perf_event_attr *attr)
{
	if (attr->type == PERF_TYPE_SOFTWARE &&
	    (attr->config == PERF_COUNT_SW_PAGE_FAULTS ||
	     attr->config == PERF_COUNT_SW_PAGE_FAULTS_MIN ||
	     attr->config == PERF_COUNT_SW_PAGE_FAULTS_MAJ))
		return true;

	if (is_bts_event(attr))
		return true;

	return false;
}

1667 1668
void thread__resolve(struct thread *thread, struct addr_location *al,
		     struct perf_sample *sample)
1669
{
1670
	thread__find_addr_map(thread, sample->cpumode, MAP__FUNCTION, sample->addr, al);
1671
	if (!al->map)
1672
		thread__find_addr_map(thread, sample->cpumode, MAP__VARIABLE,
1673 1674 1675 1676 1677 1678
				      sample->addr, al);

	al->cpu = sample->cpu;
	al->sym = NULL;

	if (al->map)
1679
		al->sym = map__find_symbol(al->map, al->addr);
1680
}