event.c 42.3 KB
Newer Older
1
#include <dirent.h>
2
#include <errno.h>
3
#include <fcntl.h>
4
#include <inttypes.h>
5
#include <linux/kernel.h>
6
#include <linux/types.h>
7 8 9
#include <sys/types.h>
#include <sys/stat.h>
#include <unistd.h>
10
#include <uapi/linux/mman.h> /* To get things like MAP_HUGETLB even on older libc headers */
11
#include <api/fs/fs.h>
12
#include <linux/perf_event.h>
13 14
#include "event.h"
#include "debug.h"
15
#include "hist.h"
16
#include "machine.h"
17
#include "sort.h"
18
#include "string2.h"
19
#include "strlist.h"
20
#include "thread.h"
21
#include "thread_map.h"
22
#include "sane_ctype.h"
23
#include "symbol/kallsyms.h"
24 25
#include "asm/bug.h"
#include "stat.h"
26

27
static const char *perf_event__names[] = {
28 29
	[0]					= "TOTAL",
	[PERF_RECORD_MMAP]			= "MMAP",
30
	[PERF_RECORD_MMAP2]			= "MMAP2",
31 32 33 34 35 36 37 38
	[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",
39
	[PERF_RECORD_AUX]			= "AUX",
40
	[PERF_RECORD_ITRACE_START]		= "ITRACE_START",
41
	[PERF_RECORD_LOST_SAMPLES]		= "LOST_SAMPLES",
42 43
	[PERF_RECORD_SWITCH]			= "SWITCH",
	[PERF_RECORD_SWITCH_CPU_WIDE]		= "SWITCH_CPU_WIDE",
44
	[PERF_RECORD_NAMESPACES]		= "NAMESPACES",
45 46 47 48 49
	[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
50
	[PERF_RECORD_ID_INDEX]			= "ID_INDEX",
51 52
	[PERF_RECORD_AUXTRACE_INFO]		= "AUXTRACE_INFO",
	[PERF_RECORD_AUXTRACE]			= "AUXTRACE",
53
	[PERF_RECORD_AUXTRACE_ERROR]		= "AUXTRACE_ERROR",
54
	[PERF_RECORD_THREAD_MAP]		= "THREAD_MAP",
55
	[PERF_RECORD_CPU_MAP]			= "CPU_MAP",
56
	[PERF_RECORD_STAT_CONFIG]		= "STAT_CONFIG",
J
Jiri Olsa 已提交
57
	[PERF_RECORD_STAT]			= "STAT",
58
	[PERF_RECORD_STAT_ROUND]		= "STAT_ROUND",
59
	[PERF_RECORD_EVENT_UPDATE]		= "EVENT_UPDATE",
60
	[PERF_RECORD_TIME_CONV]			= "TIME_CONV",
61
	[PERF_RECORD_HEADER_FEATURE]		= "FEATURE",
62 63
};

64 65 66 67 68 69 70 71 72 73
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",
};

74
const char *perf_event__name(unsigned int id)
75
{
76
	if (id >= ARRAY_SIZE(perf_event__names))
77
		return "INVALID";
78
	if (!perf_event__names[id])
79
		return "UNKNOWN";
80
	return perf_event__names[id];
81 82
}

83 84 85 86 87 88 89
static const char *perf_ns__name(unsigned int id)
{
	if (id >= ARRAY_SIZE(perf_ns__names))
		return "UNKNOWN";
	return perf_ns__names[id];
}

90 91 92 93 94 95
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 = {
96 97 98 99 100 101
	.pid	   = -1,
	.tid	   = -1,
	.time	   = -1,
	.stream_id = -1,
	.cpu	   = -1,
	.period	   = 1,
102 103 104 105
	.cpumode   = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK,
	};

	return process(tool, event, &synth_sample, machine);
106 107
};

108 109
/*
 * Assumes that the first 4095 bytes of /proc/pid/stat contains
110
 * the comm, tgid and ppid.
111
 */
112 113
static int perf_event__get_comm_ids(pid_t pid, char *comm, size_t len,
				    pid_t *tgid, pid_t *ppid)
114 115
{
	char filename[PATH_MAX];
116 117
	char bf[4096];
	int fd;
118 119
	size_t size = 0;
	ssize_t n;
120
	char *name, *tgids, *ppids;
121 122 123

	*tgid = -1;
	*ppid = -1;
124 125 126

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

127 128
	fd = open(filename, O_RDONLY);
	if (fd < 0) {
129
		pr_debug("couldn't open %s\n", filename);
130
		return -1;
131 132
	}

133 134 135
	n = read(fd, bf, sizeof(bf) - 1);
	close(fd);
	if (n <= 0) {
136
		pr_warning("Couldn't get COMM, tigd and ppid for pid %d\n",
137 138
			   pid);
		return -1;
139
	}
140
	bf[n] = '\0';
141

142 143
	name = strstr(bf, "Name:");
	tgids = strstr(bf, "Tgid:");
144
	ppids = strstr(bf, "PPid:");
145 146

	if (name) {
147 148
		char *nl;

149
		name += 5;  /* strlen("Name:") */
150 151 152 153 154 155
		name = ltrim(name);

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

156 157 158 159 160 161 162 163 164 165 166
		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:") */
167
		*tgid = atoi(tgids);
168 169 170
	} else {
		pr_debug("Tgid: string not found for pid %d\n", pid);
	}
171

172 173 174 175 176 177 178 179
	if (ppids) {
		ppids += 5;  /* strlen("PPid:") */
		*ppid = atoi(ppids);
	} else {
		pr_debug("PPid: string not found for pid %d\n", pid);
	}

	return 0;
180 181
}

182 183 184
static int perf_event__prepare_comm(union perf_event *event, pid_t pid,
				    struct machine *machine,
				    pid_t *tgid, pid_t *ppid)
185 186
{
	size_t size;
187 188

	*ppid = -1;
189 190 191

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

192 193 194 195 196 197 198 199 200
	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;
	}
201

202 203
	if (*tgid < 0)
		return -1;
204

205
	event->comm.pid = *tgid;
206
	event->comm.header.type = PERF_RECORD_COMM;
207 208

	size = strlen(event->comm.comm) + 1;
209
	size = PERF_ALIGN(size, sizeof(u64));
210
	memset(event->comm.comm + size, 0, machine->id_hdr_size);
211 212
	event->comm.header.size = (sizeof(event->comm) -
				(sizeof(event->comm.comm) - size) +
213
				machine->id_hdr_size);
214
	event->comm.tid = pid;
215 216

	return 0;
217 218
}

219
pid_t perf_event__synthesize_comm(struct perf_tool *tool,
220 221 222 223
					 union perf_event *event, pid_t pid,
					 perf_event__handler_t process,
					 struct machine *machine)
{
224
	pid_t tgid, ppid;
225

226 227
	if (perf_event__prepare_comm(event, pid, machine, &tgid, &ppid) != 0)
		return -1;
228

229
	if (perf_tool__process_synth_event(tool, event, machine, process) != 0)
230
		return -1;
231

232
	return tgid;
233 234
}

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
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;
}

287
static int perf_event__synthesize_fork(struct perf_tool *tool,
288 289 290
				       union perf_event *event,
				       pid_t pid, pid_t tgid, pid_t ppid,
				       perf_event__handler_t process,
291 292 293 294
				       struct machine *machine)
{
	memset(&event->fork, 0, sizeof(event->fork) + machine->id_hdr_size);

295 296 297 298 299 300 301 302 303 304 305 306
	/*
	 * 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;
	}
307 308 309 310 311 312
	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);

313
	if (perf_tool__process_synth_event(tool, event, machine, process) != 0)
314 315 316 317 318
		return -1;

	return 0;
}

319 320 321 322 323
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,
324 325
				       bool mmap_data,
				       unsigned int proc_map_timeout)
326 327 328
{
	char filename[PATH_MAX];
	FILE *fp;
329 330
	unsigned long long t;
	bool truncation = false;
331
	unsigned long long timeout = proc_map_timeout * 1000000ULL;
332
	int rc = 0;
333 334
	const char *hugetlbfs_mnt = hugetlbfs__mountpoint();
	int hugetlbfs_mnt_len = hugetlbfs_mnt ? strlen(hugetlbfs_mnt) : 0;
335

336 337 338
	if (machine__is_default_guest(machine))
		return 0;

339 340
	snprintf(filename, sizeof(filename), "%s/proc/%d/task/%d/maps",
		 machine->root_dir, pid, pid);
341 342 343 344 345 346 347 348 349 350

	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;
	}

351
	event->header.type = PERF_RECORD_MMAP2;
352
	t = rdclock();
353

354
	while (1) {
355 356 357 358
		char bf[BUFSIZ];
		char prot[5];
		char execname[PATH_MAX];
		char anonstr[] = "//anon";
359
		unsigned int ino;
360
		size_t size;
361
		ssize_t n;
362

363 364 365
		if (fgets(bf, sizeof(bf), fp) == NULL)
			break;

366 367 368 369 370
		if ((rdclock() - t) > timeout) {
			pr_warning("Reading %s time out. "
				   "You may want to increase "
				   "the time limit by --proc-map-timeout\n",
				   filename);
371 372 373 374
			truncation = true;
			goto out;
		}

375 376 377
		/* ensure null termination since stack will be reused. */
		strcpy(execname, "");

378
		/* 00400000-0040c000 r-xp 00000000 fd:01 41038  /bin/cat */
379
		n = sscanf(bf, "%"PRIx64"-%"PRIx64" %s %"PRIx64" %x:%x %u %[^\n]\n",
380 381 382 383 384
		       &event->mmap2.start, &event->mmap2.len, prot,
		       &event->mmap2.pgoff, &event->mmap2.maj,
		       &event->mmap2.min,
		       &ino, execname);

385 386 387
		/*
 		 * Anon maps don't have the execname.
 		 */
388
		if (n < 7)
389
			continue;
390 391 392

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

393 394 395
		/*
		 * Just like the kernel, see __perf_event_mmap in kernel/perf_event.c
		 */
396 397 398 399
		if (machine__is_host(machine))
			event->header.misc = PERF_RECORD_MISC_USER;
		else
			event->header.misc = PERF_RECORD_MISC_GUEST_USER;
400

401 402 403 404 405 406 407 408 409 410 411 412 413 414 415
		/* 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;

416 417 418 419 420 421
		if (prot[2] != 'x') {
			if (!mmap_data || prot[0] != 'r')
				continue;

			event->header.misc |= PERF_RECORD_MISC_MMAP_DATA;
		}
422

423 424 425 426
out:
		if (truncation)
			event->header.misc |= PERF_RECORD_MISC_PROC_MAP_PARSE_TIMEOUT;

427 428
		if (!strcmp(execname, ""))
			strcpy(execname, anonstr);
429

430 431
		if (hugetlbfs_mnt_len &&
		    !strncmp(execname, hugetlbfs_mnt, hugetlbfs_mnt_len)) {
432 433 434
			strcpy(execname, anonstr);
			event->mmap2.flags |= MAP_HUGETLB;
		}
435 436

		size = strlen(execname) + 1;
437
		memcpy(event->mmap2.filename, execname, size);
438
		size = PERF_ALIGN(size, sizeof(u64));
439 440 441 442 443 444 445
		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;
446

447
		if (perf_tool__process_synth_event(tool, event, machine, process) != 0) {
448 449
			rc = -1;
			break;
450
		}
451 452 453

		if (truncation)
			break;
454 455 456
	}

	fclose(fp);
457
	return rc;
458 459
}

460
int perf_event__synthesize_modules(struct perf_tool *tool,
461
				   perf_event__handler_t process,
462
				   struct machine *machine)
463
{
464
	int rc = 0;
465
	struct map *pos;
466
	struct map_groups *kmaps = &machine->kmaps;
467
	struct maps *maps = &kmaps->maps[MAP__FUNCTION];
468
	union perf_event *event = zalloc((sizeof(event->mmap) +
469
					  machine->id_hdr_size));
470 471 472 473 474 475 476
	if (event == NULL) {
		pr_debug("Not enough memory synthesizing mmap event "
			 "for kernel modules\n");
		return -1;
	}

	event->header.type = PERF_RECORD_MMAP;
477

478 479 480 481
	/*
	 * kernel uses 0 for user space maps, see kernel/perf_event.c
	 * __perf_event_mmap
	 */
482
	if (machine__is_host(machine))
483
		event->header.misc = PERF_RECORD_MISC_KERNEL;
484
	else
485
		event->header.misc = PERF_RECORD_MISC_GUEST_KERNEL;
486

487
	for (pos = maps__first(maps); pos; pos = map__next(pos)) {
488 489
		size_t size;

490
		if (__map__is_kernel(pos))
491 492
			continue;

493
		size = PERF_ALIGN(pos->dso->long_name_len + 1, sizeof(u64));
494 495 496
		event->mmap.header.type = PERF_RECORD_MMAP;
		event->mmap.header.size = (sizeof(event->mmap) -
				        (sizeof(event->mmap.filename) - size));
497 498
		memset(event->mmap.filename + size, 0, machine->id_hdr_size);
		event->mmap.header.size += machine->id_hdr_size;
499 500 501 502 503
		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,
504
		       pos->dso->long_name_len + 1);
505
		if (perf_tool__process_synth_event(tool, event, machine, process) != 0) {
506 507 508
			rc = -1;
			break;
		}
509 510
	}

511
	free(event);
512
	return rc;
513 514
}

515 516
static int __event__synthesize_thread(union perf_event *comm_event,
				      union perf_event *mmap_event,
517
				      union perf_event *fork_event,
518
				      union perf_event *namespaces_event,
519
				      pid_t pid, int full,
520
				      perf_event__handler_t process,
521
				      struct perf_tool *tool,
522 523 524
				      struct machine *machine,
				      bool mmap_data,
				      unsigned int proc_map_timeout)
525
{
526 527
	char filename[PATH_MAX];
	DIR *tasks;
528
	struct dirent *dirent;
529
	pid_t tgid, ppid;
530
	int rc = 0;
531 532 533 534 535 536 537 538 539

	/* 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;

540 541 542 543 544
		if (perf_event__synthesize_namespaces(tool, namespaces_event, pid,
						      tgid, process, machine) < 0)
			return -1;


545
		return perf_event__synthesize_mmap_events(tool, mmap_event, pid, tgid,
546 547
							  process, machine, mmap_data,
							  proc_map_timeout);
548 549 550 551 552 553 554 555 556 557 558 559 560 561
	}

	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;
	}

562
	while ((dirent = readdir(tasks)) != NULL) {
563 564 565
		char *end;
		pid_t _pid;

566
		_pid = strtol(dirent->d_name, &end, 10);
567 568 569
		if (*end)
			continue;

570
		rc = -1;
571 572
		if (perf_event__prepare_comm(comm_event, _pid, machine,
					     &tgid, &ppid) != 0)
573
			break;
574

575
		if (perf_event__synthesize_fork(tool, fork_event, _pid, tgid,
576
						ppid, process, machine) < 0)
577
			break;
578 579 580 581 582

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

583 584 585
		/*
		 * Send the prepared comm event
		 */
586
		if (perf_tool__process_synth_event(tool, comm_event, machine, process) != 0)
587
			break;
588

589
		rc = 0;
590 591 592
		if (_pid == pid) {
			/* process the parent's maps too */
			rc = perf_event__synthesize_mmap_events(tool, mmap_event, pid, tgid,
593
						process, machine, mmap_data, proc_map_timeout);
594 595
			if (rc)
				break;
596
		}
597 598 599
	}

	closedir(tasks);
600
	return rc;
601 602
}

603
int perf_event__synthesize_thread_map(struct perf_tool *tool,
604
				      struct thread_map *threads,
605
				      perf_event__handler_t process,
606
				      struct machine *machine,
607 608
				      bool mmap_data,
				      unsigned int proc_map_timeout)
609
{
610
	union perf_event *comm_event, *mmap_event, *fork_event;
611
	union perf_event *namespaces_event;
612
	int err = -1, thread, j;
613

614
	comm_event = malloc(sizeof(comm_event->comm) + machine->id_hdr_size);
615 616 617
	if (comm_event == NULL)
		goto out;

618
	mmap_event = malloc(sizeof(mmap_event->mmap2) + machine->id_hdr_size);
619 620 621
	if (mmap_event == NULL)
		goto out_free_comm;

622 623 624 625
	fork_event = malloc(sizeof(fork_event->fork) + machine->id_hdr_size);
	if (fork_event == NULL)
		goto out_free_mmap;

626 627 628 629 630 631
	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;

632 633 634
	err = 0;
	for (thread = 0; thread < threads->nr; ++thread) {
		if (__event__synthesize_thread(comm_event, mmap_event,
635
					       fork_event, namespaces_event,
636
					       thread_map__pid(threads, thread), 0,
637
					       process, tool, machine,
638
					       mmap_data, proc_map_timeout)) {
639 640 641
			err = -1;
			break;
		}
642 643 644 645 646

		/*
		 * comm.pid is set to thread group id by
		 * perf_event__synthesize_comm
		 */
647
		if ((int) comm_event->comm.pid != thread_map__pid(threads, thread)) {
648 649 650 651
			bool need_leader = true;

			/* is thread group leader in thread_map? */
			for (j = 0; j < threads->nr; ++j) {
652
				if ((int) comm_event->comm.pid == thread_map__pid(threads, j)) {
653 654 655 656 657 658 659
					need_leader = false;
					break;
				}
			}

			/* if not, generate events for it */
			if (need_leader &&
660
			    __event__synthesize_thread(comm_event, mmap_event,
661
						       fork_event, namespaces_event,
662 663
						       comm_event->comm.pid, 0,
						       process, tool, machine,
664
						       mmap_data, proc_map_timeout)) {
665 666 667 668
				err = -1;
				break;
			}
		}
669
	}
670 671
	free(namespaces_event);
out_free_fork:
672 673
	free(fork_event);
out_free_mmap:
674 675 676 677 678 679 680
	free(mmap_event);
out_free_comm:
	free(comm_event);
out:
	return err;
}

681 682 683 684 685 686 687 688
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)
689
{
690
	union perf_event *comm_event, *mmap_event, *fork_event;
691
	union perf_event *namespaces_event;
692
	int err = -1;
693 694
	char *end;
	pid_t pid;
695
	int i;
696

697
	comm_event = malloc(sizeof(comm_event->comm) + machine->id_hdr_size);
698 699 700
	if (comm_event == NULL)
		goto out;

701
	mmap_event = malloc(sizeof(mmap_event->mmap2) + machine->id_hdr_size);
702 703
	if (mmap_event == NULL)
		goto out_free_comm;
704

705 706 707 708
	fork_event = malloc(sizeof(fork_event->fork) + machine->id_hdr_size);
	if (fork_event == NULL)
		goto out_free_mmap;

709 710 711 712 713 714
	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;

715
	for (i = start; i < start + num; i++) {
716
		if (!isdigit(dirent[i]->d_name[0]))
717 718
			continue;

719 720
		pid = (pid_t)strtol(dirent[i]->d_name, &end, 10);
		/* only interested in proper numerical dirents */
721 722 723 724 725 726 727 728 729 730
		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);
731
	}
732
	err = 0;
733

734
	free(namespaces_event);
735 736
out_free_fork:
	free(fork_event);
737 738 739 740 741 742
out_free_mmap:
	free(mmap_event);
out_free_comm:
	free(comm_event);
out:
	return err;
743
}
744

745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853
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;

	thread_nr = nr_threads_synthesize;

	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;
}

854 855 856 857 858
struct process_symbol_args {
	const char *name;
	u64	   start;
};

859
static int find_symbol_cb(void *arg, const char *name, char type,
860
			  u64 start)
861 862 863
{
	struct process_symbol_args *args = arg;

864 865 866 867 868 869
	/*
	 * 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))
870 871 872 873 874 875
		return 0;

	args->start = start;
	return 1;
}

876 877
int kallsyms__get_function_start(const char *kallsyms_filename,
				 const char *symbol_name, u64 *addr)
878 879 880 881
{
	struct process_symbol_args args = { .name = symbol_name, };

	if (kallsyms__parse(kallsyms_filename, &args, find_symbol_cb) <= 0)
882
		return -1;
883

884 885
	*addr = args.start;
	return 0;
886 887
}

888
int perf_event__synthesize_kernel_mmap(struct perf_tool *tool,
889
				       perf_event__handler_t process,
890
				       struct machine *machine)
891 892
{
	size_t size;
893
	const char *mmap_name;
894
	char name_buff[PATH_MAX];
895
	struct map *map = machine__kernel_map(machine);
896
	struct kmap *kmap;
897
	int err;
898 899
	union perf_event *event;

900 901
	if (symbol_conf.kptr_restrict)
		return -1;
902
	if (map == NULL)
903 904
		return -1;

905 906 907 908 909
	/*
	 * 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.
	 */
910
	event = zalloc((sizeof(event->mmap) + machine->id_hdr_size));
911 912 913 914 915
	if (event == NULL) {
		pr_debug("Not enough memory synthesizing mmap event "
			 "for kernel modules\n");
		return -1;
	}
916

917
	mmap_name = machine__mmap_name(machine, name_buff, sizeof(name_buff));
918
	if (machine__is_host(machine)) {
919 920 921 922
		/*
		 * kernel uses PERF_RECORD_MISC_USER for user space maps,
		 * see kernel/perf_event.c __perf_event_mmap
		 */
923
		event->header.misc = PERF_RECORD_MISC_KERNEL;
924
	} else {
925
		event->header.misc = PERF_RECORD_MISC_GUEST_KERNEL;
926
	}
927

928
	kmap = map__kmap(map);
929
	size = snprintf(event->mmap.filename, sizeof(event->mmap.filename),
930
			"%s%s", mmap_name, kmap->ref_reloc_sym->name) + 1;
931
	size = PERF_ALIGN(size, sizeof(u64));
932 933
	event->mmap.header.type = PERF_RECORD_MMAP;
	event->mmap.header.size = (sizeof(event->mmap) -
934
			(sizeof(event->mmap.filename) - size) + machine->id_hdr_size);
935
	event->mmap.pgoff = kmap->ref_reloc_sym->addr;
936 937 938 939
	event->mmap.start = map->start;
	event->mmap.len   = map->end - event->mmap.start;
	event->mmap.pid   = machine->pid;

940
	err = perf_tool__process_synth_event(tool, event, machine, process);
941 942 943
	free(event);

	return err;
944 945
}

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
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;
}

982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 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
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;
}

1113 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
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;
}

1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172
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);
}

1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189
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);
}

1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213
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);
		}
	}
}

1214 1215
size_t perf_event__fprintf_comm(union perf_event *event, FILE *fp)
{
1216 1217 1218 1219 1220 1221 1222
	const char *s;

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

1223
	return fprintf(fp, "%s: %s:%d/%d\n", s, event->comm.comm, event->comm.pid, event->comm.tid);
1224 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
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;
}

1253
int perf_event__process_comm(struct perf_tool *tool __maybe_unused,
1254
			     union perf_event *event,
1255
			     struct perf_sample *sample,
1256
			     struct machine *machine)
1257
{
1258
	return machine__process_comm_event(machine, event, sample);
1259 1260
}

1261 1262 1263 1264 1265 1266 1267 1268
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);
}

1269
int perf_event__process_lost(struct perf_tool *tool __maybe_unused,
1270
			     union perf_event *event,
1271
			     struct perf_sample *sample,
1272
			     struct machine *machine)
1273
{
1274
	return machine__process_lost_event(machine, event, sample);
1275
}
1276

1277 1278 1279 1280 1281 1282 1283 1284
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);
}

1285 1286 1287 1288 1289 1290 1291 1292
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);
}

1293 1294 1295 1296 1297 1298 1299 1300
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);
}

1301 1302 1303 1304 1305 1306 1307 1308
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);
}

1309 1310
size_t perf_event__fprintf_mmap(union perf_event *event, FILE *fp)
{
1311
	return fprintf(fp, " %d/%d: [%#" PRIx64 "(%#" PRIx64 ") @ %#" PRIx64 "]: %c %s\n",
1312
		       event->mmap.pid, event->mmap.tid, event->mmap.start,
1313 1314 1315
		       event->mmap.len, event->mmap.pgoff,
		       (event->header.misc & PERF_RECORD_MISC_MMAP_DATA) ? 'r' : 'x',
		       event->mmap.filename);
1316 1317
}

1318 1319 1320
size_t perf_event__fprintf_mmap2(union perf_event *event, FILE *fp)
{
	return fprintf(fp, " %d/%d: [%#" PRIx64 "(%#" PRIx64 ") @ %#" PRIx64
1321
			   " %02x:%02x %"PRIu64" %"PRIu64"]: %c%c%c%c %s\n",
1322 1323 1324 1325
		       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,
1326 1327 1328 1329
		       (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',
1330 1331 1332
		       event->mmap2.filename);
}

1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348
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;
}

1349 1350 1351 1352 1353
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;

1354
	ret = fprintf(fp, ": ");
1355 1356 1357 1358 1359 1360 1361 1362 1363 1364

	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;
}

1365
int perf_event__process_mmap(struct perf_tool *tool __maybe_unused,
1366
			     union perf_event *event,
1367
			     struct perf_sample *sample,
1368
			     struct machine *machine)
1369
{
1370
	return machine__process_mmap_event(machine, event, sample);
1371 1372
}

1373 1374
int perf_event__process_mmap2(struct perf_tool *tool __maybe_unused,
			     union perf_event *event,
1375
			     struct perf_sample *sample,
1376 1377
			     struct machine *machine)
{
1378
	return machine__process_mmap2_event(machine, event, sample);
1379 1380
}

1381 1382 1383 1384 1385 1386 1387
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);
}

1388
int perf_event__process_fork(struct perf_tool *tool __maybe_unused,
1389
			     union perf_event *event,
1390
			     struct perf_sample *sample,
1391
			     struct machine *machine)
1392
{
1393
	return machine__process_fork_event(machine, event, sample);
1394
}
1395

1396 1397
int perf_event__process_exit(struct perf_tool *tool __maybe_unused,
			     union perf_event *event,
1398
			     struct perf_sample *sample,
1399 1400
			     struct machine *machine)
{
1401
	return machine__process_exit_event(machine, event, sample);
1402 1403
}

1404 1405
size_t perf_event__fprintf_aux(union perf_event *event, FILE *fp)
{
1406
	return fprintf(fp, " offset: %#"PRIx64" size: %#"PRIx64" flags: %#"PRIx64" [%s%s%s]\n",
1407 1408 1409
		       event->aux.aux_offset, event->aux.aux_size,
		       event->aux.flags,
		       event->aux.flags & PERF_AUX_FLAG_TRUNCATED ? "T" : "",
1410 1411
		       event->aux.flags & PERF_AUX_FLAG_OVERWRITE ? "O" : "",
		       event->aux.flags & PERF_AUX_FLAG_PARTIAL   ? "P" : "");
1412 1413
}

1414 1415 1416 1417 1418 1419
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);
}

1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433
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);
}

1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448
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);
1449 1450 1451
		break;
	case PERF_RECORD_NAMESPACES:
		ret += perf_event__fprintf_namespaces(event, fp);
1452
		break;
1453 1454 1455
	case PERF_RECORD_MMAP2:
		ret += perf_event__fprintf_mmap2(event, fp);
		break;
1456 1457 1458
	case PERF_RECORD_AUX:
		ret += perf_event__fprintf_aux(event, fp);
		break;
1459 1460 1461
	case PERF_RECORD_ITRACE_START:
		ret += perf_event__fprintf_itrace_start(event, fp);
		break;
1462 1463 1464 1465
	case PERF_RECORD_SWITCH:
	case PERF_RECORD_SWITCH_CPU_WIDE:
		ret += perf_event__fprintf_switch(event, fp);
		break;
1466 1467 1468 1469 1470 1471 1472
	default:
		ret += fprintf(fp, "\n");
	}

	return ret;
}

1473 1474
int perf_event__process(struct perf_tool *tool __maybe_unused,
			union perf_event *event,
1475
			struct perf_sample *sample,
1476
			struct machine *machine)
1477
{
1478
	return machine__process_event(machine, event, sample);
1479 1480
}

1481
void thread__find_addr_map(struct thread *thread, u8 cpumode,
1482
			   enum map_type type, u64 addr,
1483
			   struct addr_location *al)
1484
{
1485
	struct map_groups *mg = thread->mg;
1486
	struct machine *machine = mg->machine;
1487
	bool load_map = false;
1488

1489
	al->machine = machine;
1490
	al->thread = thread;
1491
	al->addr = addr;
1492
	al->cpumode = cpumode;
1493
	al->filtered = 0;
1494

1495 1496 1497 1498 1499
	if (machine == NULL) {
		al->map = NULL;
		return;
	}

1500
	if (cpumode == PERF_RECORD_MISC_KERNEL && perf_host) {
1501
		al->level = 'k';
1502
		mg = &machine->kmaps;
1503
		load_map = true;
1504
	} else if (cpumode == PERF_RECORD_MISC_USER && perf_host) {
1505
		al->level = '.';
1506 1507
	} else if (cpumode == PERF_RECORD_MISC_GUEST_KERNEL && perf_guest) {
		al->level = 'g';
1508
		mg = &machine->kmaps;
1509
		load_map = true;
1510 1511
	} else if (cpumode == PERF_RECORD_MISC_GUEST_USER && perf_guest) {
		al->level = 'u';
1512
	} else {
1513
		al->level = 'H';
1514
		al->map = NULL;
1515 1516 1517 1518

		if ((cpumode == PERF_RECORD_MISC_GUEST_USER ||
			cpumode == PERF_RECORD_MISC_GUEST_KERNEL) &&
			!perf_guest)
1519
			al->filtered |= (1 << HIST_FILTER__GUEST);
1520 1521 1522
		if ((cpumode == PERF_RECORD_MISC_USER ||
			cpumode == PERF_RECORD_MISC_KERNEL) &&
			!perf_host)
1523
			al->filtered |= (1 << HIST_FILTER__HOST);
1524

1525 1526 1527
		return;
	}
try_again:
1528
	al->map = map_groups__find(mg, type, al->addr);
1529 1530 1531 1532 1533 1534 1535 1536 1537 1538
	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.
		 */
1539 1540 1541
		if (cpumode == PERF_RECORD_MISC_USER && machine &&
		    mg != &machine->kmaps &&
		    machine__kernel_ip(machine, al->addr)) {
1542
			mg = &machine->kmaps;
1543
			load_map = true;
1544 1545
			goto try_again;
		}
1546 1547 1548 1549 1550 1551
	} else {
		/*
		 * Kernel maps might be changed when loading symbols so loading
		 * must be done prior to using kernel maps.
		 */
		if (load_map)
1552
			map__load(al->map);
1553
		al->addr = al->map->map_ip(al->map, al->addr);
1554
	}
1555 1556
}

1557
void thread__find_addr_location(struct thread *thread,
1558
				u8 cpumode, enum map_type type, u64 addr,
1559
				struct addr_location *al)
1560
{
1561
	thread__find_addr_map(thread, cpumode, type, addr, al);
1562
	if (al->map != NULL)
1563
		al->sym = map__find_symbol(al->map, al->addr);
1564 1565
	else
		al->sym = NULL;
1566 1567
}

1568 1569 1570 1571
/*
 * Callers need to drop the reference to al->thread, obtained in
 * machine__findnew_thread()
 */
1572 1573
int machine__resolve(struct machine *machine, struct addr_location *al,
		     struct perf_sample *sample)
1574
{
1575
	struct thread *thread = machine__findnew_thread(machine, sample->pid,
1576
							sample->tid);
1577

1578 1579 1580
	if (thread == NULL)
		return -1;

1581
	dump_printf(" ... thread: %s:%d\n", thread__comm_str(thread), thread->tid);
1582
	/*
1583
	 * Have we already created the kernel maps for this machine?
1584 1585 1586 1587 1588
	 *
	 * This should have happened earlier, when we processed the kernel MMAP
	 * events, but for older perf.data files there was no such thing, so do
	 * it now.
	 */
1589
	if (sample->cpumode == PERF_RECORD_MISC_KERNEL &&
1590
	    machine__kernel_map(machine) == NULL)
1591
		machine__create_kernel_maps(machine);
1592

1593
	thread__find_addr_map(thread, sample->cpumode, MAP__FUNCTION, sample->ip, al);
1594 1595 1596
	dump_printf(" ...... dso: %s\n",
		    al->map ? al->map->dso->long_name :
			al->level == 'H' ? "[hypervisor]" : "<not found>");
1597 1598 1599 1600

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

1601
	al->sym = NULL;
1602
	al->cpu = sample->cpu;
1603 1604 1605 1606 1607 1608 1609 1610
	al->socket = -1;

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

		if (env && env->cpu)
			al->socket = env->cpu[al->cpu].socket_id;
	}
1611 1612

	if (al->map) {
1613 1614
		struct dso *dso = al->map->dso;

1615
		if (symbol_conf.dso_list &&
1616 1617 1618 1619
		    (!dso || !(strlist__has_entry(symbol_conf.dso_list,
						  dso->short_name) ||
			       (dso->short_name != dso->long_name &&
				strlist__has_entry(symbol_conf.dso_list,
1620 1621 1622
						   dso->long_name))))) {
			al->filtered |= (1 << HIST_FILTER__DSO);
		}
1623

1624
		al->sym = map__find_symbol(al->map, al->addr);
1625
	}
1626

1627 1628
	if (symbol_conf.sym_list &&
		(!al->sym || !strlist__has_entry(symbol_conf.sym_list,
1629 1630 1631
						al->sym->name))) {
		al->filtered |= (1 << HIST_FILTER__SYMBOL);
	}
1632 1633

	return 0;
1634
}
1635

1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646
/*
 * 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);
}

1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667
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;
}

1668 1669
void thread__resolve(struct thread *thread, struct addr_location *al,
		     struct perf_sample *sample)
1670
{
1671
	thread__find_addr_map(thread, sample->cpumode, MAP__FUNCTION, sample->addr, al);
1672
	if (!al->map)
1673
		thread__find_addr_map(thread, sample->cpumode, MAP__VARIABLE,
1674 1675 1676 1677 1678 1679
				      sample->addr, al);

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

	if (al->map)
1680
		al->sym = map__find_symbol(al->map, al->addr);
1681
}