evsel.c 8.6 KB
Newer Older
1 2 3 4 5 6 7 8 9
/*
 * Copyright (C) 2011, Red Hat Inc, Arnaldo Carvalho de Melo <acme@redhat.com>
 *
 * Parts came from builtin-{top,stat,record}.c, see those files for further
 * copyright notes.
 *
 * Released under the GPL v2. (and only v2, not any later version)
 */

10
#include "evsel.h"
11
#include "evlist.h"
12
#include "util.h"
13
#include "cpumap.h"
14
#include "thread_map.h"
15

16 17
#define FD(e, x, y) (*(int *)xyarray__entry(e->fd, x, y))

18 19 20 21 22 23 24 25
void perf_evsel__init(struct perf_evsel *evsel,
		      struct perf_event_attr *attr, int idx)
{
	evsel->idx	   = idx;
	evsel->attr	   = *attr;
	INIT_LIST_HEAD(&evsel->node);
}

26
struct perf_evsel *perf_evsel__new(struct perf_event_attr *attr, int idx)
27 28 29
{
	struct perf_evsel *evsel = zalloc(sizeof(*evsel));

30 31
	if (evsel != NULL)
		perf_evsel__init(evsel, attr, idx);
32 33 34 35 36 37

	return evsel;
}

int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads)
{
38
	int cpu, thread;
39
	evsel->fd = xyarray__new(ncpus, nthreads, sizeof(int));
40 41 42 43 44 45 46 47 48

	if (evsel->fd) {
		for (cpu = 0; cpu < ncpus; cpu++) {
			for (thread = 0; thread < nthreads; thread++) {
				FD(evsel, cpu, thread) = -1;
			}
		}
	}

49 50 51
	return evsel->fd != NULL ? 0 : -ENOMEM;
}

52 53
int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads)
{
54 55 56 57 58 59 60 61 62 63 64 65
	evsel->sample_id = xyarray__new(ncpus, nthreads, sizeof(struct perf_sample_id));
	if (evsel->sample_id == NULL)
		return -ENOMEM;

	evsel->id = zalloc(ncpus * nthreads * sizeof(u64));
	if (evsel->id == NULL) {
		xyarray__delete(evsel->sample_id);
		evsel->sample_id = NULL;
		return -ENOMEM;
	}

	return 0;
66 67
}

68 69 70 71 72 73 74
int perf_evsel__alloc_counts(struct perf_evsel *evsel, int ncpus)
{
	evsel->counts = zalloc((sizeof(*evsel->counts) +
				(ncpus * sizeof(struct perf_counts_values))));
	return evsel->counts != NULL ? 0 : -ENOMEM;
}

75 76 77 78 79 80
void perf_evsel__free_fd(struct perf_evsel *evsel)
{
	xyarray__delete(evsel->fd);
	evsel->fd = NULL;
}

81 82
void perf_evsel__free_id(struct perf_evsel *evsel)
{
83 84 85
	xyarray__delete(evsel->sample_id);
	evsel->sample_id = NULL;
	free(evsel->id);
86 87 88
	evsel->id = NULL;
}

89 90 91 92 93 94 95 96 97 98 99
void perf_evsel__close_fd(struct perf_evsel *evsel, int ncpus, int nthreads)
{
	int cpu, thread;

	for (cpu = 0; cpu < ncpus; cpu++)
		for (thread = 0; thread < nthreads; ++thread) {
			close(FD(evsel, cpu, thread));
			FD(evsel, cpu, thread) = -1;
		}
}

100
void perf_evsel__exit(struct perf_evsel *evsel)
101 102 103
{
	assert(list_empty(&evsel->node));
	xyarray__delete(evsel->fd);
104 105
	xyarray__delete(evsel->sample_id);
	free(evsel->id);
106 107 108 109 110
}

void perf_evsel__delete(struct perf_evsel *evsel)
{
	perf_evsel__exit(evsel);
S
Stephane Eranian 已提交
111
	close_cgroup(evsel->cgrp);
112
	free(evsel->name);
113 114
	free(evsel);
}
115 116 117 118 119 120 121 122 123 124

int __perf_evsel__read_on_cpu(struct perf_evsel *evsel,
			      int cpu, int thread, bool scale)
{
	struct perf_counts_values count;
	size_t nv = scale ? 3 : 1;

	if (FD(evsel, cpu, thread) < 0)
		return -EINVAL;

125 126 127
	if (evsel->counts == NULL && perf_evsel__alloc_counts(evsel, cpu + 1) < 0)
		return -ENOMEM;

128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149
	if (readn(FD(evsel, cpu, thread), &count, nv * sizeof(u64)) < 0)
		return -errno;

	if (scale) {
		if (count.run == 0)
			count.val = 0;
		else if (count.run < count.ena)
			count.val = (u64)((double)count.val * count.ena / count.run + 0.5);
	} else
		count.ena = count.run = 0;

	evsel->counts->cpu[cpu] = count;
	return 0;
}

int __perf_evsel__read(struct perf_evsel *evsel,
		       int ncpus, int nthreads, bool scale)
{
	size_t nv = scale ? 3 : 1;
	int cpu, thread;
	struct perf_counts_values *aggr = &evsel->counts->aggr, count;

150
	aggr->val = aggr->ena = aggr->run = 0;
151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185

	for (cpu = 0; cpu < ncpus; cpu++) {
		for (thread = 0; thread < nthreads; thread++) {
			if (FD(evsel, cpu, thread) < 0)
				continue;

			if (readn(FD(evsel, cpu, thread),
				  &count, nv * sizeof(u64)) < 0)
				return -errno;

			aggr->val += count.val;
			if (scale) {
				aggr->ena += count.ena;
				aggr->run += count.run;
			}
		}
	}

	evsel->counts->scaled = 0;
	if (scale) {
		if (aggr->run == 0) {
			evsel->counts->scaled = -1;
			aggr->val = 0;
			return 0;
		}

		if (aggr->run < aggr->ena) {
			evsel->counts->scaled = 1;
			aggr->val = (u64)((double)aggr->val * aggr->ena / aggr->run + 0.5);
		}
	} else
		aggr->ena = aggr->run = 0;

	return 0;
}
186

187
static int __perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus,
188
			      struct thread_map *threads, bool group)
189
{
190
	int cpu, thread;
S
Stephane Eranian 已提交
191 192
	unsigned long flags = 0;
	int pid = -1;
193

194 195
	if (evsel->fd == NULL &&
	    perf_evsel__alloc_fd(evsel, cpus->nr, threads->nr) < 0)
196 197
		return -1;

S
Stephane Eranian 已提交
198 199 200 201 202
	if (evsel->cgrp) {
		flags = PERF_FLAG_PID_CGROUP;
		pid = evsel->cgrp->fd;
	}

203
	for (cpu = 0; cpu < cpus->nr; cpu++) {
204
		int group_fd = -1;
205

206
		for (thread = 0; thread < threads->nr; thread++) {
S
Stephane Eranian 已提交
207 208 209 210

			if (!evsel->cgrp)
				pid = threads->map[thread];

211
			FD(evsel, cpu, thread) = sys_perf_event_open(&evsel->attr,
S
Stephane Eranian 已提交
212
								     pid,
213
								     cpus->map[cpu],
S
Stephane Eranian 已提交
214
								     group_fd, flags);
215 216
			if (FD(evsel, cpu, thread) < 0)
				goto out_close;
217 218 219

			if (group && group_fd == -1)
				group_fd = FD(evsel, cpu, thread);
220
		}
221 222 223 224 225
	}

	return 0;

out_close:
226 227 228 229 230 231 232
	do {
		while (--thread >= 0) {
			close(FD(evsel, cpu, thread));
			FD(evsel, cpu, thread) = -1;
		}
		thread = threads->nr;
	} while (--cpu >= 0);
233 234 235
	return -1;
}

236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251
static struct {
	struct cpu_map map;
	int cpus[1];
} empty_cpu_map = {
	.map.nr	= 1,
	.cpus	= { -1, },
};

static struct {
	struct thread_map map;
	int threads[1];
} empty_thread_map = {
	.map.nr	 = 1,
	.threads = { -1, },
};

252
int perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus,
253
		     struct thread_map *threads, bool group)
254
{
255 256 257
	if (cpus == NULL) {
		/* Work around old compiler warnings about strict aliasing */
		cpus = &empty_cpu_map.map;
258 259
	}

260 261
	if (threads == NULL)
		threads = &empty_thread_map.map;
262

263
	return __perf_evsel__open(evsel, cpus, threads, group);
264 265
}

266
int perf_evsel__open_per_cpu(struct perf_evsel *evsel,
267
			     struct cpu_map *cpus, bool group)
268
{
269
	return __perf_evsel__open(evsel, cpus, &empty_thread_map.map, group);
270
}
271

272
int perf_evsel__open_per_thread(struct perf_evsel *evsel,
273
				struct thread_map *threads, bool group)
274
{
275
	return __perf_evsel__open(evsel, &empty_cpu_map.map, threads, group);
276
}
277

278 279
static int perf_event__parse_id_sample(const union perf_event *event, u64 type,
				       struct perf_sample *sample)
280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315
{
	const u64 *array = event->sample.array;

	array += ((event->header.size -
		   sizeof(event->header)) / sizeof(u64)) - 1;

	if (type & PERF_SAMPLE_CPU) {
		u32 *p = (u32 *)array;
		sample->cpu = *p;
		array--;
	}

	if (type & PERF_SAMPLE_STREAM_ID) {
		sample->stream_id = *array;
		array--;
	}

	if (type & PERF_SAMPLE_ID) {
		sample->id = *array;
		array--;
	}

	if (type & PERF_SAMPLE_TIME) {
		sample->time = *array;
		array--;
	}

	if (type & PERF_SAMPLE_TID) {
		u32 *p = (u32 *)array;
		sample->pid = p[0];
		sample->tid = p[1];
	}

	return 0;
}

316 317 318 319 320 321 322 323 324 325 326
static bool sample_overlap(const union perf_event *event,
			   const void *offset, u64 size)
{
	const void *base = event;

	if (offset + size > base + event->header.size)
		return true;

	return false;
}

327
int perf_event__parse_sample(const union perf_event *event, u64 type,
328 329
			     int sample_size, bool sample_id_all,
			     struct perf_sample *data)
330 331 332 333 334 335 336 337 338
{
	const u64 *array;

	data->cpu = data->pid = data->tid = -1;
	data->stream_id = data->id = data->time = -1ULL;

	if (event->header.type != PERF_RECORD_SAMPLE) {
		if (!sample_id_all)
			return 0;
339
		return perf_event__parse_id_sample(event, type, data);
340 341 342 343
	}

	array = event->sample.array;

344 345 346
	if (sample_size + sizeof(event->header) > event->header.size)
		return -EFAULT;

347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396
	if (type & PERF_SAMPLE_IP) {
		data->ip = event->ip.ip;
		array++;
	}

	if (type & PERF_SAMPLE_TID) {
		u32 *p = (u32 *)array;
		data->pid = p[0];
		data->tid = p[1];
		array++;
	}

	if (type & PERF_SAMPLE_TIME) {
		data->time = *array;
		array++;
	}

	if (type & PERF_SAMPLE_ADDR) {
		data->addr = *array;
		array++;
	}

	data->id = -1ULL;
	if (type & PERF_SAMPLE_ID) {
		data->id = *array;
		array++;
	}

	if (type & PERF_SAMPLE_STREAM_ID) {
		data->stream_id = *array;
		array++;
	}

	if (type & PERF_SAMPLE_CPU) {
		u32 *p = (u32 *)array;
		data->cpu = *p;
		array++;
	}

	if (type & PERF_SAMPLE_PERIOD) {
		data->period = *array;
		array++;
	}

	if (type & PERF_SAMPLE_READ) {
		fprintf(stderr, "PERF_SAMPLE_READ is unsuported for now\n");
		return -1;
	}

	if (type & PERF_SAMPLE_CALLCHAIN) {
397 398 399
		if (sample_overlap(event, array, sizeof(data->callchain->nr)))
			return -EFAULT;

400
		data->callchain = (struct ip_callchain *)array;
401 402 403 404

		if (sample_overlap(event, array, data->callchain->nr))
			return -EFAULT;

405 406 407 408 409
		array += 1 + data->callchain->nr;
	}

	if (type & PERF_SAMPLE_RAW) {
		u32 *p = (u32 *)array;
410 411 412 413

		if (sample_overlap(event, array, sizeof(u32)))
			return -EFAULT;

414 415
		data->raw_size = *p;
		p++;
416 417 418 419

		if (sample_overlap(event, p, data->raw_size))
			return -EFAULT;

420 421 422 423 424
		data->raw_data = p;
	}

	return 0;
}