evsel.c 7.5 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 38 39 40 41

	return evsel;
}

int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads)
{
	evsel->fd = xyarray__new(ncpus, nthreads, sizeof(int));
	return evsel->fd != NULL ? 0 : -ENOMEM;
}

42 43 44 45 46 47
int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads)
{
	evsel->id = xyarray__new(ncpus, nthreads, sizeof(struct perf_sample_id));
	return evsel->id != NULL ? 0 : -ENOMEM;
}

48 49 50 51 52 53 54
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;
}

55 56 57 58 59 60
void perf_evsel__free_fd(struct perf_evsel *evsel)
{
	xyarray__delete(evsel->fd);
	evsel->fd = NULL;
}

61 62 63 64 65 66
void perf_evsel__free_id(struct perf_evsel *evsel)
{
	xyarray__delete(evsel->id);
	evsel->id = NULL;
}

67 68 69 70 71 72 73 74 75 76 77
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;
		}
}

78
void perf_evsel__exit(struct perf_evsel *evsel)
79 80 81
{
	assert(list_empty(&evsel->node));
	xyarray__delete(evsel->fd);
82
	xyarray__delete(evsel->id);
83 84 85 86 87
}

void perf_evsel__delete(struct perf_evsel *evsel)
{
	perf_evsel__exit(evsel);
88 89
	free(evsel);
}
90 91 92 93 94 95 96 97 98 99

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;

100 101 102
	if (evsel->counts == NULL && perf_evsel__alloc_counts(evsel, cpu + 1) < 0)
		return -ENOMEM;

103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124
	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;

125
	aggr->val = aggr->ena = aggr->run = 0;
126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160

	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;
}
161

162
static int __perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus,
163
			      struct thread_map *threads, bool group, bool inherit)
164
{
165
	int cpu, thread;
166

167 168
	if (evsel->fd == NULL &&
	    perf_evsel__alloc_fd(evsel, cpus->nr, threads->nr) < 0)
169 170
		return -1;

171
	for (cpu = 0; cpu < cpus->nr; cpu++) {
172 173
		int group_fd = -1;

174 175
		evsel->attr.inherit = (cpus->map[cpu] < 0) && inherit;

176 177 178
		for (thread = 0; thread < threads->nr; thread++) {
			FD(evsel, cpu, thread) = sys_perf_event_open(&evsel->attr,
								     threads->map[thread],
179 180
								     cpus->map[cpu],
								     group_fd, 0);
181 182
			if (FD(evsel, cpu, thread) < 0)
				goto out_close;
183 184 185

			if (group && group_fd == -1)
				group_fd = FD(evsel, cpu, thread);
186
		}
187 188 189 190 191
	}

	return 0;

out_close:
192 193 194 195 196 197 198
	do {
		while (--thread >= 0) {
			close(FD(evsel, cpu, thread));
			FD(evsel, cpu, thread) = -1;
		}
		thread = threads->nr;
	} while (--cpu >= 0);
199 200 201
	return -1;
}

202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217
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, },
};

218
int perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus,
219
		     struct thread_map *threads, bool group, bool inherit)
220
{
221 222 223
	if (cpus == NULL) {
		/* Work around old compiler warnings about strict aliasing */
		cpus = &empty_cpu_map.map;
224 225
	}

226 227
	if (threads == NULL)
		threads = &empty_thread_map.map;
228

229
	return __perf_evsel__open(evsel, cpus, threads, group, inherit);
230 231
}

232
int perf_evsel__open_per_cpu(struct perf_evsel *evsel,
233
			     struct cpu_map *cpus, bool group, bool inherit)
234
{
235
	return __perf_evsel__open(evsel, cpus, &empty_thread_map.map, group, inherit);
236
}
237

238
int perf_evsel__open_per_thread(struct perf_evsel *evsel,
239
				struct thread_map *threads, bool group, bool inherit)
240
{
241
	return __perf_evsel__open(evsel, &empty_cpu_map.map, threads, group, inherit);
242
}
243

244 245
static int perf_event__parse_id_sample(const union perf_event *event, u64 type,
				       struct perf_sample *sample)
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
{
	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;
}

282 283
int perf_event__parse_sample(const union perf_event *event, u64 type,
			     bool sample_id_all, struct perf_sample *data)
284 285 286 287 288 289 290 291 292
{
	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;
293
		return perf_event__parse_id_sample(event, type, data);
294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
	}

	array = event->sample.array;

	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) {
		data->callchain = (struct ip_callchain *)array;
		array += 1 + data->callchain->nr;
	}

	if (type & PERF_SAMPLE_RAW) {
		u32 *p = (u32 *)array;
		data->raw_size = *p;
		p++;
		data->raw_data = p;
	}

	return 0;
}