SAFE public projects git trees. - safe/jmp/linux-2.6/blob - tools/perf/builtin-record.c

   1 /*
   2  * builtin-record.c
   3  *
   4  * Builtin record command: Record the profile of a workload
   5  * (or a CPU, or a PID) into the perf.data output file - for
   6  * later analysis via perf report.
   7  */
   8 #include "builtin.h"
   9
  10 #include "perf.h"
  11
  12 #include "util/util.h"
  13 #include "util/parse-options.h"
  14 #include "util/parse-events.h"
  15 #include "util/string.h"
  16
  17 #include "util/header.h"
  18 #include "util/event.h"
  19 #include "util/debug.h"
  20 #include "util/symbol.h"
  21
  22 #include <unistd.h>
  23 #include <sched.h>
  24
  25 static int                      fd[MAX_NR_CPUS][MAX_COUNTERS];
  26
  27 static long                     default_interval                =      0;
  28
  29 static int                      nr_cpus                         =      0;
  30 static unsigned int             page_size;
  31 static unsigned int             mmap_pages                      =    128;
  32 static int                      freq                            =   1000;
  33 static int                      output;
  34 static const char               *output_name                    = "perf.data";
  35 static int                      group                           =      0;
  36 static unsigned int             realtime_prio                   =      0;
  37 static int                      raw_samples                     =      0;
  38 static int                      system_wide                     =      0;
  39 static int                      profile_cpu                     =     -1;
  40 static pid_t                    target_pid                      =     -1;
  41 static pid_t                    child_pid                       =     -1;
  42 static int                      inherit                         =      1;
  43 static int                      force                           =      0;
  44 static int                      append_file                     =      0;
  45 static int                      call_graph                      =      0;
  46 static int                      inherit_stat                    =      0;
  47 static int                      no_samples                      =      0;
  48 static int                      sample_address                  =      0;
  49 static int                      multiplex                       =      0;
  50 static int                      multiplex_fd                    =     -1;
  51
  52 static long                     samples                         =      0;
  53 static struct timeval           last_read;
  54 static struct timeval           this_read;
  55
  56 static u64                      bytes_written                   =      0;
  57
  58 static struct pollfd            event_array[MAX_NR_CPUS * MAX_COUNTERS];
  59
  60 static int                      nr_poll                         =      0;
  61 static int                      nr_cpu                          =      0;
  62
  63 static int                      file_new                        =      1;
  64
  65 struct perf_header              *header                         =   NULL;
  66
  67 struct mmap_data {
  68         int                     counter;
  69         void                    *base;
  70         unsigned int            mask;
  71         unsigned int            prev;
  72 };
  73
  74 static struct mmap_data         mmap_array[MAX_NR_CPUS][MAX_COUNTERS];
  75
  76 static unsigned long mmap_read_head(struct mmap_data *md)
  77 {
  78         struct perf_event_mmap_page *pc = md->base;
  79         long head;
  80
  81         head = pc->data_head;
  82         rmb();
  83
  84         return head;
  85 }
  86
  87 static void mmap_write_tail(struct mmap_data *md, unsigned long tail)
  88 {
  89         struct perf_event_mmap_page *pc = md->base;
  90
  91         /*
  92          * ensure all reads are done before we write the tail out.
  93          */
  94         /* mb(); */
  95         pc->data_tail = tail;
  96 }
  97
  98 static void write_output(void *buf, size_t size)
  99 {
 100         while (size) {
 101                 int ret = write(output, buf, size);
 102
 103                 if (ret < 0)
 104                         die("failed to write");
 105
 106                 size -= ret;
 107                 buf += ret;
 108
 109                 bytes_written += ret;
 110         }
 111 }
 112
 113 static void write_event(event_t *buf, size_t size)
 114 {
 115         /*
 116         * Add it to the list of DSOs, so that when we finish this
 117          * record session we can pick the available build-ids.
 118          */
 119         if (buf->header.type == PERF_RECORD_MMAP)
 120                 dsos__findnew(buf->mmap.filename);
 121
 122         write_output(buf, size);
 123 }
 124
 125 static int process_synthesized_event(event_t *event)
 126 {
 127         write_event(event, event->header.size);
 128         return 0;
 129 }
 130
 131 static void mmap_read(struct mmap_data *md)
 132 {
 133         unsigned int head = mmap_read_head(md);
 134         unsigned int old = md->prev;
 135         unsigned char *data = md->base + page_size;
 136         unsigned long size;
 137         void *buf;
 138         int diff;
 139
 140         gettimeofday(&this_read, NULL);
 141
 142         /*
 143          * If we're further behind than half the buffer, there's a chance
 144          * the writer will bite our tail and mess up the samples under us.
 145          *
 146          * If we somehow ended up ahead of the head, we got messed up.
 147          *
 148          * In either case, truncate and restart at head.
 149          */
 150         diff = head - old;
 151         if (diff < 0) {
 152                 struct timeval iv;
 153                 unsigned long msecs;
 154
 155                 timersub(&this_read, &last_read, &iv);
 156                 msecs = iv.tv_sec*1000 + iv.tv_usec/1000;
 157
 158                 fprintf(stderr, "WARNING: failed to keep up with mmap data."
 159                                 "  Last read %lu msecs ago.\n", msecs);
 160
 161                 /*
 162                  * head points to a known good entry, start there.
 163                  */
 164                 old = head;
 165         }
 166
 167         last_read = this_read;
 168
 169         if (old != head)
 170                 samples++;
 171
 172         size = head - old;
 173
 174         if ((old & md->mask) + size != (head & md->mask)) {
 175                 buf = &data[old & md->mask];
 176                 size = md->mask + 1 - (old & md->mask);
 177                 old += size;
 178
 179                 write_event(buf, size);
 180         }
 181
 182         buf = &data[old & md->mask];
 183         size = head - old;
 184         old += size;
 185
 186         write_event(buf, size);
 187
 188         md->prev = old;
 189         mmap_write_tail(md, old);
 190 }
 191
 192 static volatile int done = 0;
 193 static volatile int signr = -1;
 194
 195 static void sig_handler(int sig)
 196 {
 197         done = 1;
 198         signr = sig;
 199 }
 200
 201 static void sig_atexit(void)
 202 {
 203         if (child_pid != -1)
 204                 kill(child_pid, SIGTERM);
 205
 206         if (signr == -1)
 207                 return;
 208
 209         signal(signr, SIG_DFL);
 210         kill(getpid(), signr);
 211 }
 212
 213 static int group_fd;
 214
 215 static struct perf_header_attr *get_header_attr(struct perf_event_attr *a, int nr)
 216 {
 217         struct perf_header_attr *h_attr;
 218
 219         if (nr < header->attrs) {
 220                 h_attr = header->attr[nr];
 221         } else {
 222                 h_attr = perf_header_attr__new(a);
 223                 if (h_attr != NULL)
 224                         if (perf_header__add_attr(header, h_attr) < 0) {
 225                                 perf_header_attr__delete(h_attr);
 226                                 h_attr = NULL;
 227                         }
 228         }
 229
 230         return h_attr;
 231 }
 232
 233 static void create_counter(int counter, int cpu, pid_t pid)
 234 {
 235         char *filter = filters[counter];
 236         struct perf_event_attr *attr = attrs + counter;
 237         struct perf_header_attr *h_attr;
 238         int track = !counter; /* only the first counter needs these */
 239         int ret;
 240         struct {
 241                 u64 count;
 242                 u64 time_enabled;
 243                 u64 time_running;
 244                 u64 id;
 245         } read_data;
 246
 247         attr->read_format       = PERF_FORMAT_TOTAL_TIME_ENABLED |
 248                                   PERF_FORMAT_TOTAL_TIME_RUNNING |
 249                                   PERF_FORMAT_ID;
 250
 251         attr->sample_type       |= PERF_SAMPLE_IP | PERF_SAMPLE_TID;
 252
 253         if (freq) {
 254                 attr->sample_type       |= PERF_SAMPLE_PERIOD;
 255                 attr->freq              = 1;
 256                 attr->sample_freq       = freq;
 257         }
 258
 259         if (no_samples)
 260                 attr->sample_freq = 0;
 261
 262         if (inherit_stat)
 263                 attr->inherit_stat = 1;
 264
 265         if (sample_address)
 266                 attr->sample_type       |= PERF_SAMPLE_ADDR;
 267
 268         if (call_graph)
 269                 attr->sample_type       |= PERF_SAMPLE_CALLCHAIN;
 270
 271         if (raw_samples) {
 272                 attr->sample_type       |= PERF_SAMPLE_TIME;
 273                 attr->sample_type       |= PERF_SAMPLE_RAW;
 274                 attr->sample_type       |= PERF_SAMPLE_CPU;
 275         }
 276
 277         attr->mmap              = track;
 278         attr->comm              = track;
 279         attr->inherit           = (cpu < 0) && inherit;
 280         attr->disabled          = 1;
 281
 282 try_again:
 283         fd[nr_cpu][counter] = sys_perf_event_open(attr, pid, cpu, group_fd, 0);
 284
 285         if (fd[nr_cpu][counter] < 0) {
 286                 int err = errno;
 287
 288                 if (err == EPERM || err == EACCES)
 289                         die("Permission error - are you root?\n");
 290                 else if (err ==  ENODEV && profile_cpu != -1)
 291                         die("No such device - did you specify an out-of-range profile CPU?\n");
 292
 293                 /*
 294                  * If it's cycles then fall back to hrtimer
 295                  * based cpu-clock-tick sw counter, which
 296                  * is always available even if no PMU support:
 297                  */
 298                 if (attr->type == PERF_TYPE_HARDWARE
 299                         && attr->config == PERF_COUNT_HW_CPU_CYCLES) {
 300
 301                         if (verbose)
 302                                 warning(" ... trying to fall back to cpu-clock-ticks\n");
 303                         attr->type = PERF_TYPE_SOFTWARE;
 304                         attr->config = PERF_COUNT_SW_CPU_CLOCK;
 305                         goto try_again;
 306                 }
 307                 printf("\n");
 308                 error("perfcounter syscall returned with %d (%s)\n",
 309                         fd[nr_cpu][counter], strerror(err));
 310
 311 #if defined(__i386__) || defined(__x86_64__)
 312                 if (attr->type == PERF_TYPE_HARDWARE && err == EOPNOTSUPP)
 313                         die("No hardware sampling interrupt available. No APIC? If so then you can boot the kernel with the \"lapic\" boot parameter to force-enable it.\n");
 314 #endif
 315
 316                 die("No CONFIG_PERF_EVENTS=y kernel support configured?\n");
 317                 exit(-1);
 318         }
 319
 320         h_attr = get_header_attr(attr, counter);
 321         if (h_attr == NULL)
 322                 die("nomem\n");
 323
 324         if (!file_new) {
 325                 if (memcmp(&h_attr->attr, attr, sizeof(*attr))) {
 326                         fprintf(stderr, "incompatible append\n");
 327                         exit(-1);
 328                 }
 329         }
 330
 331         if (read(fd[nr_cpu][counter], &read_data, sizeof(read_data)) == -1) {
 332                 perror("Unable to read perf file descriptor\n");
 333                 exit(-1);
 334         }
 335
 336         if (perf_header_attr__add_id(h_attr, read_data.id) < 0) {
 337                 pr_warning("Not enough memory to add id\n");
 338                 exit(-1);
 339         }
 340
 341         assert(fd[nr_cpu][counter] >= 0);
 342         fcntl(fd[nr_cpu][counter], F_SETFL, O_NONBLOCK);
 343
 344         /*
 345          * First counter acts as the group leader:
 346          */
 347         if (group && group_fd == -1)
 348                 group_fd = fd[nr_cpu][counter];
 349         if (multiplex && multiplex_fd == -1)
 350                 multiplex_fd = fd[nr_cpu][counter];
 351
 352         if (multiplex && fd[nr_cpu][counter] != multiplex_fd) {
 353
 354                 ret = ioctl(fd[nr_cpu][counter], PERF_EVENT_IOC_SET_OUTPUT, multiplex_fd);
 355                 assert(ret != -1);
 356         } else {
 357                 event_array[nr_poll].fd = fd[nr_cpu][counter];
 358                 event_array[nr_poll].events = POLLIN;
 359                 nr_poll++;
 360
 361                 mmap_array[nr_cpu][counter].counter = counter;
 362                 mmap_array[nr_cpu][counter].prev = 0;
 363                 mmap_array[nr_cpu][counter].mask = mmap_pages*page_size - 1;
 364                 mmap_array[nr_cpu][counter].base = mmap(NULL, (mmap_pages+1)*page_size,
 365                                 PROT_READ|PROT_WRITE, MAP_SHARED, fd[nr_cpu][counter], 0);
 366                 if (mmap_array[nr_cpu][counter].base == MAP_FAILED) {
 367                         error("failed to mmap with %d (%s)\n", errno, strerror(errno));
 368                         exit(-1);
 369                 }
 370         }
 371
 372         if (filter != NULL) {
 373                 ret = ioctl(fd[nr_cpu][counter],
 374                             PERF_EVENT_IOC_SET_FILTER, filter);
 375                 if (ret) {
 376                         error("failed to set filter with %d (%s)\n", errno,
 377                               strerror(errno));
 378                         exit(-1);
 379                 }
 380         }
 381
 382         ioctl(fd[nr_cpu][counter], PERF_EVENT_IOC_ENABLE);
 383 }
 384
 385 static void open_counters(int cpu, pid_t pid)
 386 {
 387         int counter;
 388
 389         group_fd = -1;
 390         for (counter = 0; counter < nr_counters; counter++)
 391                 create_counter(counter, cpu, pid);
 392
 393         nr_cpu++;
 394 }
 395
 396 static void atexit_header(void)
 397 {
 398         header->data_size += bytes_written;
 399
 400         perf_header__write(header, output, true);
 401 }
 402
 403 static int __cmd_record(int argc, const char **argv)
 404 {
 405         int i, counter;
 406         struct stat st;
 407         pid_t pid = 0;
 408         int flags;
 409         int err;
 410         unsigned long waking = 0;
 411
 412         page_size = sysconf(_SC_PAGE_SIZE);
 413         nr_cpus = sysconf(_SC_NPROCESSORS_ONLN);
 414         assert(nr_cpus <= MAX_NR_CPUS);
 415         assert(nr_cpus >= 0);
 416
 417         atexit(sig_atexit);
 418         signal(SIGCHLD, sig_handler);
 419         signal(SIGINT, sig_handler);
 420
 421         if (!stat(output_name, &st) && st.st_size) {
 422                 if (!force && !append_file) {
 423                         fprintf(stderr, "Error, output file %s exists, use -A to append or -f to overwrite.\n",
 424                                         output_name);
 425                         exit(-1);
 426                 }
 427         } else {
 428                 append_file = 0;
 429         }
 430
 431         flags = O_CREAT|O_RDWR;
 432         if (append_file)
 433                 file_new = 0;
 434         else
 435                 flags |= O_TRUNC;
 436
 437         output = open(output_name, flags, S_IRUSR|S_IWUSR);
 438         if (output < 0) {
 439                 perror("failed to create output file");
 440                 exit(-1);
 441         }
 442
 443         header = perf_header__new();
 444         if (header == NULL) {
 445                 pr_err("Not enough memory for reading perf file header\n");
 446                 return -1;
 447         }
 448
 449         if (!file_new) {
 450                 err = perf_header__read(header, output);
 451                 if (err < 0)
 452                         return err;
 453         }
 454
 455         if (raw_samples) {
 456                 perf_header__set_feat(header, HEADER_TRACE_INFO);
 457         } else {
 458                 for (i = 0; i < nr_counters; i++) {
 459                         if (attrs[i].sample_type & PERF_SAMPLE_RAW) {
 460                                 perf_header__set_feat(header, HEADER_TRACE_INFO);
 461                                 break;
 462                         }
 463                 }
 464         }
 465
 466         atexit(atexit_header);
 467
 468         if (!system_wide) {
 469                 pid = target_pid;
 470                 if (pid == -1)
 471                         pid = getpid();
 472
 473                 open_counters(profile_cpu, pid);
 474         } else {
 475                 if (profile_cpu != -1) {
 476                         open_counters(profile_cpu, target_pid);
 477                 } else {
 478                         for (i = 0; i < nr_cpus; i++)
 479                                 open_counters(i, target_pid);
 480                 }
 481         }
 482
 483         if (file_new) {
 484                 err = perf_header__write(header, output, false);
 485                 if (err < 0)
 486                         return err;
 487         }
 488
 489         if (!system_wide)
 490                 event__synthesize_thread(pid, process_synthesized_event);
 491         else
 492                 event__synthesize_threads(process_synthesized_event);
 493
 494         if (target_pid == -1 && argc) {
 495                 pid = fork();
 496                 if (pid < 0)
 497                         die("failed to fork");
 498
 499                 if (!pid) {
 500                         if (execvp(argv[0], (char **)argv)) {
 501                                 perror(argv[0]);
 502                                 exit(-1);
 503                         }
 504                 } else {
 505                         /*
 506                          * Wait a bit for the execv'ed child to appear
 507                          * and be updated in /proc
 508                          * FIXME: Do you know a less heuristical solution?
 509                          */
 510                         usleep(1000);
 511                         event__synthesize_thread(pid,
 512                                                  process_synthesized_event);
 513                 }
 514
 515                 child_pid = pid;
 516         }
 517
 518         if (realtime_prio) {
 519                 struct sched_param param;
 520
 521                 param.sched_priority = realtime_prio;
 522                 if (sched_setscheduler(0, SCHED_FIFO, &param)) {
 523                         pr_err("Could not set realtime priority.\n");
 524                         exit(-1);
 525                 }
 526         }
 527
 528         for (;;) {
 529                 int hits = samples;
 530
 531                 for (i = 0; i < nr_cpu; i++) {
 532                         for (counter = 0; counter < nr_counters; counter++) {
 533                                 if (mmap_array[i][counter].base)
 534                                         mmap_read(&mmap_array[i][counter]);
 535                         }
 536                 }
 537
 538                 if (hits == samples) {
 539                         if (done)
 540                                 break;
 541                         err = poll(event_array, nr_poll, -1);
 542                         waking++;
 543                 }
 544
 545                 if (done) {
 546                         for (i = 0; i < nr_cpu; i++) {
 547                                 for (counter = 0; counter < nr_counters; counter++)
 548                                         ioctl(fd[i][counter], PERF_EVENT_IOC_DISABLE);
 549                         }
 550                 }
 551         }
 552
 553         fprintf(stderr, "[ perf record: Woken up %ld times to write data ]\n", waking);
 554
 555         /*
 556          * Approximate RIP event size: 24 bytes.
 557          */
 558         fprintf(stderr,
 559                 "[ perf record: Captured and wrote %.3f MB %s (~%lld samples) ]\n",
 560                 (double)bytes_written / 1024.0 / 1024.0,
 561                 output_name,
 562                 bytes_written / 24);
 563
 564         return 0;
 565 }
 566
 567 static const char * const record_usage[] = {
 568         "perf record [<options>] [<command>]",
 569         "perf record [<options>] -- <command> [<options>]",
 570         NULL
 571 };
 572
 573 static const struct option options[] = {
 574         OPT_CALLBACK('e', "event", NULL, "event",
 575                      "event selector. use 'perf list' to list available events",
 576                      parse_events),
 577         OPT_CALLBACK(0, "filter", NULL, "filter",
 578                      "event filter", parse_filter),
 579         OPT_INTEGER('p', "pid", &target_pid,
 580                     "record events on existing pid"),
 581         OPT_INTEGER('r', "realtime", &realtime_prio,
 582                     "collect data with this RT SCHED_FIFO priority"),
 583         OPT_BOOLEAN('R', "raw-samples", &raw_samples,
 584                     "collect raw sample records from all opened counters"),
 585         OPT_BOOLEAN('a', "all-cpus", &system_wide,
 586                             "system-wide collection from all CPUs"),
 587         OPT_BOOLEAN('A', "append", &append_file,
 588                             "append to the output file to do incremental profiling"),
 589         OPT_INTEGER('C', "profile_cpu", &profile_cpu,
 590                             "CPU to profile on"),
 591         OPT_BOOLEAN('f', "force", &force,
 592                         "overwrite existing data file"),
 593         OPT_LONG('c', "count", &default_interval,
 594                     "event period to sample"),
 595         OPT_STRING('o', "output", &output_name, "file",
 596                     "output file name"),
 597         OPT_BOOLEAN('i', "inherit", &inherit,
 598                     "child tasks inherit counters"),
 599         OPT_INTEGER('F', "freq", &freq,
 600                     "profile at this frequency"),
 601         OPT_INTEGER('m', "mmap-pages", &mmap_pages,
 602                     "number of mmap data pages"),
 603         OPT_BOOLEAN('g', "call-graph", &call_graph,
 604                     "do call-graph (stack chain/backtrace) recording"),
 605         OPT_BOOLEAN('v', "verbose", &verbose,
 606                     "be more verbose (show counter open errors, etc)"),
 607         OPT_BOOLEAN('s', "stat", &inherit_stat,
 608                     "per thread counts"),
 609         OPT_BOOLEAN('d', "data", &sample_address,
 610                     "Sample addresses"),
 611         OPT_BOOLEAN('n', "no-samples", &no_samples,
 612                     "don't sample"),
 613         OPT_BOOLEAN('M', "multiplex", &multiplex,
 614                     "multiplex counter output in a single channel"),
 615         OPT_END()
 616 };
 617
 618 int cmd_record(int argc, const char **argv, const char *prefix __used)
 619 {
 620         int counter;
 621
 622         symbol__init(0);
 623
 624         argc = parse_options(argc, argv, options, record_usage,
 625                 PARSE_OPT_STOP_AT_NON_OPTION);
 626         if (!argc && target_pid == -1 && !system_wide)
 627                 usage_with_options(record_usage, options);
 628
 629         if (!nr_counters) {
 630                 nr_counters     = 1;
 631                 attrs[0].type   = PERF_TYPE_HARDWARE;
 632                 attrs[0].config = PERF_COUNT_HW_CPU_CYCLES;
 633         }
 634
 635         /*
 636          * User specified count overrides default frequency.
 637          */
 638         if (default_interval)
 639                 freq = 0;
 640         else if (freq) {
 641                 default_interval = freq;
 642         } else {
 643                 fprintf(stderr, "frequency and count are zero, aborting\n");
 644                 exit(EXIT_FAILURE);
 645         }
 646
 647         for (counter = 0; counter < nr_counters; counter++) {
 648                 if (attrs[counter].sample_period)
 649                         continue;
 650
 651                 attrs[counter].sample_period = default_interval;
 652         }
 653
 654         return __cmd_record(argc, argv);
 655 }