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 #define _FILE_OFFSET_BITS 64
   9
  10 #include "builtin.h"
  11
  12 #include "perf.h"
  13
  14 #include "util/build-id.h"
  15 #include "util/util.h"
  16 #include "util/parse-options.h"
  17 #include "util/parse-events.h"
  18 #include "util/string.h"
  19
  20 #include "util/header.h"
  21 #include "util/event.h"
  22 #include "util/debug.h"
  23 #include "util/session.h"
  24 #include "util/symbol.h"
  25 #include "util/cpumap.h"
  26
  27 #include <unistd.h>
  28 #include <sched.h>
  29
  30 static int                      fd[MAX_NR_CPUS][MAX_COUNTERS];
  31
  32 static long                     default_interval                =      0;
  33
  34 static int                      nr_cpus                         =      0;
  35 static unsigned int             page_size;
  36 static unsigned int             mmap_pages                      =    128;
  37 static int                      freq                            =   1000;
  38 static int                      output;
  39 static const char               *output_name                    = "perf.data";
  40 static int                      group                           =      0;
  41 static unsigned int             realtime_prio                   =      0;
  42 static int                      raw_samples                     =      0;
  43 static int                      system_wide                     =      0;
  44 static int                      profile_cpu                     =     -1;
  45 static pid_t                    target_pid                      =     -1;
  46 static pid_t                    child_pid                       =     -1;
  47 static int                      inherit                         =      1;
  48 static int                      force                           =      0;
  49 static int                      append_file                     =      0;
  50 static int                      call_graph                      =      0;
  51 static int                      inherit_stat                    =      0;
  52 static int                      no_samples                      =      0;
  53 static int                      sample_address                  =      0;
  54 static int                      multiplex                       =      0;
  55 static int                      multiplex_fd                    =     -1;
  56
  57 static long                     samples                         =      0;
  58 static struct timeval           last_read;
  59 static struct timeval           this_read;
  60
  61 static u64                      bytes_written                   =      0;
  62
  63 static struct pollfd            event_array[MAX_NR_CPUS * MAX_COUNTERS];
  64
  65 static int                      nr_poll                         =      0;
  66 static int                      nr_cpu                          =      0;
  67
  68 static int                      file_new                        =      1;
  69 static off_t                    post_processing_offset;
  70
  71 static struct perf_session      *session;
  72
  73 struct mmap_data {
  74         int                     counter;
  75         void                    *base;
  76         unsigned int            mask;
  77         unsigned int            prev;
  78 };
  79
  80 static struct mmap_data         mmap_array[MAX_NR_CPUS][MAX_COUNTERS];
  81
  82 static unsigned long mmap_read_head(struct mmap_data *md)
  83 {
  84         struct perf_event_mmap_page *pc = md->base;
  85         long head;
  86
  87         head = pc->data_head;
  88         rmb();
  89
  90         return head;
  91 }
  92
  93 static void mmap_write_tail(struct mmap_data *md, unsigned long tail)
  94 {
  95         struct perf_event_mmap_page *pc = md->base;
  96
  97         /*
  98          * ensure all reads are done before we write the tail out.
  99          */
 100         /* mb(); */
 101         pc->data_tail = tail;
 102 }
 103
 104 static void write_output(void *buf, size_t size)
 105 {
 106         while (size) {
 107                 int ret = write(output, buf, size);
 108
 109                 if (ret < 0)
 110                         die("failed to write");
 111
 112                 size -= ret;
 113                 buf += ret;
 114
 115                 bytes_written += ret;
 116         }
 117 }
 118
 119 static int process_synthesized_event(event_t *event,
 120                                      struct perf_session *self __used)
 121 {
 122         write_output(event, event->header.size);
 123         return 0;
 124 }
 125
 126 static void mmap_read(struct mmap_data *md)
 127 {
 128         unsigned int head = mmap_read_head(md);
 129         unsigned int old = md->prev;
 130         unsigned char *data = md->base + page_size;
 131         unsigned long size;
 132         void *buf;
 133         int diff;
 134
 135         gettimeofday(&this_read, NULL);
 136
 137         /*
 138          * If we're further behind than half the buffer, there's a chance
 139          * the writer will bite our tail and mess up the samples under us.
 140          *
 141          * If we somehow ended up ahead of the head, we got messed up.
 142          *
 143          * In either case, truncate and restart at head.
 144          */
 145         diff = head - old;
 146         if (diff < 0) {
 147                 struct timeval iv;
 148                 unsigned long msecs;
 149
 150                 timersub(&this_read, &last_read, &iv);
 151                 msecs = iv.tv_sec*1000 + iv.tv_usec/1000;
 152
 153                 fprintf(stderr, "WARNING: failed to keep up with mmap data."
 154                                 "  Last read %lu msecs ago.\n", msecs);
 155
 156                 /*
 157                  * head points to a known good entry, start there.
 158                  */
 159                 old = head;
 160         }
 161
 162         last_read = this_read;
 163
 164         if (old != head)
 165                 samples++;
 166
 167         size = head - old;
 168
 169         if ((old & md->mask) + size != (head & md->mask)) {
 170                 buf = &data[old & md->mask];
 171                 size = md->mask + 1 - (old & md->mask);
 172                 old += size;
 173
 174                 write_output(buf, size);
 175         }
 176
 177         buf = &data[old & md->mask];
 178         size = head - old;
 179         old += size;
 180
 181         write_output(buf, size);
 182
 183         md->prev = old;
 184         mmap_write_tail(md, old);
 185 }
 186
 187 static volatile int done = 0;
 188 static volatile int signr = -1;
 189
 190 static void sig_handler(int sig)
 191 {
 192         done = 1;
 193         signr = sig;
 194 }
 195
 196 static void sig_atexit(void)
 197 {
 198         if (child_pid != -1)
 199                 kill(child_pid, SIGTERM);
 200
 201         if (signr == -1)
 202                 return;
 203
 204         signal(signr, SIG_DFL);
 205         kill(getpid(), signr);
 206 }
 207
 208 static int group_fd;
 209
 210 static struct perf_header_attr *get_header_attr(struct perf_event_attr *a, int nr)
 211 {
 212         struct perf_header_attr *h_attr;
 213
 214         if (nr < session->header.attrs) {
 215                 h_attr = session->header.attr[nr];
 216         } else {
 217                 h_attr = perf_header_attr__new(a);
 218                 if (h_attr != NULL)
 219                         if (perf_header__add_attr(&session->header, h_attr) < 0) {
 220                                 perf_header_attr__delete(h_attr);
 221                                 h_attr = NULL;
 222                         }
 223         }
 224
 225         return h_attr;
 226 }
 227
 228 static void create_counter(int counter, int cpu, pid_t pid)
 229 {
 230         char *filter = filters[counter];
 231         struct perf_event_attr *attr = attrs + counter;
 232         struct perf_header_attr *h_attr;
 233         int track = !counter; /* only the first counter needs these */
 234         int ret;
 235         struct {
 236                 u64 count;
 237                 u64 time_enabled;
 238                 u64 time_running;
 239                 u64 id;
 240         } read_data;
 241
 242         attr->read_format       = PERF_FORMAT_TOTAL_TIME_ENABLED |
 243                                   PERF_FORMAT_TOTAL_TIME_RUNNING |
 244                                   PERF_FORMAT_ID;
 245
 246         attr->sample_type       |= PERF_SAMPLE_IP | PERF_SAMPLE_TID;
 247
 248         if (nr_counters > 1)
 249                 attr->sample_type |= PERF_SAMPLE_ID;
 250
 251         if (freq) {
 252                 attr->sample_type       |= PERF_SAMPLE_PERIOD;
 253                 attr->freq              = 1;
 254                 attr->sample_freq       = freq;
 255         }
 256
 257         if (no_samples)
 258                 attr->sample_freq = 0;
 259
 260         if (inherit_stat)
 261                 attr->inherit_stat = 1;
 262
 263         if (sample_address)
 264                 attr->sample_type       |= PERF_SAMPLE_ADDR;
 265
 266         if (call_graph)
 267                 attr->sample_type       |= PERF_SAMPLE_CALLCHAIN;
 268
 269         if (raw_samples) {
 270                 attr->sample_type       |= PERF_SAMPLE_TIME;
 271                 attr->sample_type       |= PERF_SAMPLE_RAW;
 272                 attr->sample_type       |= PERF_SAMPLE_CPU;
 273         }
 274
 275         attr->mmap              = track;
 276         attr->comm              = track;
 277         attr->inherit           = inherit;
 278         attr->disabled          = 1;
 279
 280 try_again:
 281         fd[nr_cpu][counter] = sys_perf_event_open(attr, pid, cpu, group_fd, 0);
 282
 283         if (fd[nr_cpu][counter] < 0) {
 284                 int err = errno;
 285
 286                 if (err == EPERM || err == EACCES)
 287                         die("Permission error - are you root?\n");
 288                 else if (err ==  ENODEV && profile_cpu != -1)
 289                         die("No such device - did you specify an out-of-range profile CPU?\n");
 290
 291                 /*
 292                  * If it's cycles then fall back to hrtimer
 293                  * based cpu-clock-tick sw counter, which
 294                  * is always available even if no PMU support:
 295                  */
 296                 if (attr->type == PERF_TYPE_HARDWARE
 297                         && attr->config == PERF_COUNT_HW_CPU_CYCLES) {
 298
 299                         if (verbose)
 300                                 warning(" ... trying to fall back to cpu-clock-ticks\n");
 301                         attr->type = PERF_TYPE_SOFTWARE;
 302                         attr->config = PERF_COUNT_SW_CPU_CLOCK;
 303                         goto try_again;
 304                 }
 305                 printf("\n");
 306                 error("perfcounter syscall returned with %d (%s)\n",
 307                         fd[nr_cpu][counter], strerror(err));
 308
 309 #if defined(__i386__) || defined(__x86_64__)
 310                 if (attr->type == PERF_TYPE_HARDWARE && err == EOPNOTSUPP)
 311                         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");
 312 #endif
 313
 314                 die("No CONFIG_PERF_EVENTS=y kernel support configured?\n");
 315                 exit(-1);
 316         }
 317
 318         h_attr = get_header_attr(attr, counter);
 319         if (h_attr == NULL)
 320                 die("nomem\n");
 321
 322         if (!file_new) {
 323                 if (memcmp(&h_attr->attr, attr, sizeof(*attr))) {
 324                         fprintf(stderr, "incompatible append\n");
 325                         exit(-1);
 326                 }
 327         }
 328
 329         if (read(fd[nr_cpu][counter], &read_data, sizeof(read_data)) == -1) {
 330                 perror("Unable to read perf file descriptor\n");
 331                 exit(-1);
 332         }
 333
 334         if (perf_header_attr__add_id(h_attr, read_data.id) < 0) {
 335                 pr_warning("Not enough memory to add id\n");
 336                 exit(-1);
 337         }
 338
 339         assert(fd[nr_cpu][counter] >= 0);
 340         fcntl(fd[nr_cpu][counter], F_SETFL, O_NONBLOCK);
 341
 342         /*
 343          * First counter acts as the group leader:
 344          */
 345         if (group && group_fd == -1)
 346                 group_fd = fd[nr_cpu][counter];
 347         if (multiplex && multiplex_fd == -1)
 348                 multiplex_fd = fd[nr_cpu][counter];
 349
 350         if (multiplex && fd[nr_cpu][counter] != multiplex_fd) {
 351
 352                 ret = ioctl(fd[nr_cpu][counter], PERF_EVENT_IOC_SET_OUTPUT, multiplex_fd);
 353                 assert(ret != -1);
 354         } else {
 355                 event_array[nr_poll].fd = fd[nr_cpu][counter];
 356                 event_array[nr_poll].events = POLLIN;
 357                 nr_poll++;
 358
 359                 mmap_array[nr_cpu][counter].counter = counter;
 360                 mmap_array[nr_cpu][counter].prev = 0;
 361                 mmap_array[nr_cpu][counter].mask = mmap_pages*page_size - 1;
 362                 mmap_array[nr_cpu][counter].base = mmap(NULL, (mmap_pages+1)*page_size,
 363                                 PROT_READ|PROT_WRITE, MAP_SHARED, fd[nr_cpu][counter], 0);
 364                 if (mmap_array[nr_cpu][counter].base == MAP_FAILED) {
 365                         error("failed to mmap with %d (%s)\n", errno, strerror(errno));
 366                         exit(-1);
 367                 }
 368         }
 369
 370         if (filter != NULL) {
 371                 ret = ioctl(fd[nr_cpu][counter],
 372                             PERF_EVENT_IOC_SET_FILTER, filter);
 373                 if (ret) {
 374                         error("failed to set filter with %d (%s)\n", errno,
 375                               strerror(errno));
 376                         exit(-1);
 377                 }
 378         }
 379
 380         ioctl(fd[nr_cpu][counter], PERF_EVENT_IOC_ENABLE);
 381 }
 382
 383 static void open_counters(int cpu, pid_t pid)
 384 {
 385         int counter;
 386
 387         group_fd = -1;
 388         for (counter = 0; counter < nr_counters; counter++)
 389                 create_counter(counter, cpu, pid);
 390
 391         nr_cpu++;
 392 }
 393
 394 static int process_buildids(void)
 395 {
 396         u64 size = lseek(output, 0, SEEK_CUR);
 397
 398         if (size == 0)
 399                 return 0;
 400
 401         session->fd = output;
 402         return __perf_session__process_events(session, post_processing_offset,
 403                                               size - post_processing_offset,
 404                                               size, &build_id__mark_dso_hit_ops);
 405 }
 406
 407 static void atexit_header(void)
 408 {
 409         session->header.data_size += bytes_written;
 410
 411         process_buildids();
 412         perf_header__write(&session->header, output, true);
 413 }
 414
 415 static int __cmd_record(int argc, const char **argv)
 416 {
 417         int i, counter;
 418         struct stat st;
 419         pid_t pid = 0;
 420         int flags;
 421         int err;
 422         unsigned long waking = 0;
 423         int child_ready_pipe[2], go_pipe[2];
 424         const bool forks = target_pid == -1 && argc > 0;
 425         char buf;
 426
 427         page_size = sysconf(_SC_PAGE_SIZE);
 428
 429         atexit(sig_atexit);
 430         signal(SIGCHLD, sig_handler);
 431         signal(SIGINT, sig_handler);
 432
 433         if (forks && (pipe(child_ready_pipe) < 0 || pipe(go_pipe) < 0)) {
 434                 perror("failed to create pipes");
 435                 exit(-1);
 436         }
 437
 438         if (!stat(output_name, &st) && st.st_size) {
 439                 if (!force) {
 440                         if (!append_file) {
 441                                 pr_err("Error, output file %s exists, use -A "
 442                                        "to append or -f to overwrite.\n",
 443                                        output_name);
 444                                 exit(-1);
 445                         }
 446                 } else {
 447                         char oldname[PATH_MAX];
 448                         snprintf(oldname, sizeof(oldname), "%s.old",
 449                                  output_name);
 450                         unlink(oldname);
 451                         rename(output_name, oldname);
 452                 }
 453         } else {
 454                 append_file = 0;
 455         }
 456
 457         flags = O_CREAT|O_RDWR;
 458         if (append_file)
 459                 file_new = 0;
 460         else
 461                 flags |= O_TRUNC;
 462
 463         output = open(output_name, flags, S_IRUSR|S_IWUSR);
 464         if (output < 0) {
 465                 perror("failed to create output file");
 466                 exit(-1);
 467         }
 468
 469         session = perf_session__new(output_name, O_WRONLY, force);
 470         if (session == NULL) {
 471                 pr_err("Not enough memory for reading perf file header\n");
 472                 return -1;
 473         }
 474
 475         if (!file_new) {
 476                 err = perf_header__read(&session->header, output);
 477                 if (err < 0)
 478                         return err;
 479         }
 480
 481         if (raw_samples) {
 482                 perf_header__set_feat(&session->header, HEADER_TRACE_INFO);
 483         } else {
 484                 for (i = 0; i < nr_counters; i++) {
 485                         if (attrs[i].sample_type & PERF_SAMPLE_RAW) {
 486                                 perf_header__set_feat(&session->header, HEADER_TRACE_INFO);
 487                                 break;
 488                         }
 489                 }
 490         }
 491
 492         atexit(atexit_header);
 493
 494         if (forks) {
 495                 pid = fork();
 496                 if (pid < 0) {
 497                         perror("failed to fork");
 498                         exit(-1);
 499                 }
 500
 501                 if (!pid) {
 502                         close(child_ready_pipe[0]);
 503                         close(go_pipe[1]);
 504                         fcntl(go_pipe[0], F_SETFD, FD_CLOEXEC);
 505
 506                         /*
 507                          * Do a dummy execvp to get the PLT entry resolved,
 508                          * so we avoid the resolver overhead on the real
 509                          * execvp call.
 510                          */
 511                         execvp("", (char **)argv);
 512
 513                         /*
 514                          * Tell the parent we're ready to go
 515                          */
 516                         close(child_ready_pipe[1]);
 517
 518                         /*
 519                          * Wait until the parent tells us to go.
 520                          */
 521                         if (read(go_pipe[0], &buf, 1) == -1)
 522                                 perror("unable to read pipe");
 523
 524                         execvp(argv[0], (char **)argv);
 525
 526                         perror(argv[0]);
 527                         exit(-1);
 528                 }
 529
 530                 child_pid = pid;
 531
 532                 if (!system_wide)
 533                         target_pid = pid;
 534
 535                 close(child_ready_pipe[1]);
 536                 close(go_pipe[0]);
 537                 /*
 538                  * wait for child to settle
 539                  */
 540                 if (read(child_ready_pipe[0], &buf, 1) == -1) {
 541                         perror("unable to read pipe");
 542                         exit(-1);
 543                 }
 544                 close(child_ready_pipe[0]);
 545         }
 546
 547
 548         if ((!system_wide && !inherit) || profile_cpu != -1) {
 549                 open_counters(profile_cpu, target_pid);
 550         } else {
 551                 nr_cpus = read_cpu_map();
 552                 for (i = 0; i < nr_cpus; i++)
 553                         open_counters(cpumap[i], target_pid);
 554         }
 555
 556         if (file_new) {
 557                 err = perf_header__write(&session->header, output, false);
 558                 if (err < 0)
 559                         return err;
 560         }
 561
 562         post_processing_offset = lseek(output, 0, SEEK_CUR);
 563
 564         err = event__synthesize_kernel_mmap(process_synthesized_event,
 565                                             session, "_text");
 566         if (err < 0)
 567                 err = event__synthesize_kernel_mmap(process_synthesized_event,
 568                                                     session, "_stext");
 569         if (err < 0) {
 570                 pr_err("Couldn't record kernel reference relocation symbol.\n");
 571                 return err;
 572         }
 573
 574         err = event__synthesize_modules(process_synthesized_event, session);
 575         if (err < 0) {
 576                 pr_err("Couldn't record kernel reference relocation symbol.\n");
 577                 return err;
 578         }
 579
 580         if (!system_wide && profile_cpu == -1)
 581                 event__synthesize_thread(target_pid, process_synthesized_event,
 582                                          session);
 583         else
 584                 event__synthesize_threads(process_synthesized_event, session);
 585
 586         if (realtime_prio) {
 587                 struct sched_param param;
 588
 589                 param.sched_priority = realtime_prio;
 590                 if (sched_setscheduler(0, SCHED_FIFO, &param)) {
 591                         pr_err("Could not set realtime priority.\n");
 592                         exit(-1);
 593                 }
 594         }
 595
 596         /*
 597          * Let the child rip
 598          */
 599         if (forks)
 600                 close(go_pipe[1]);
 601
 602         for (;;) {
 603                 int hits = samples;
 604
 605                 for (i = 0; i < nr_cpu; i++) {
 606                         for (counter = 0; counter < nr_counters; counter++) {
 607                                 if (mmap_array[i][counter].base)
 608                                         mmap_read(&mmap_array[i][counter]);
 609                         }
 610                 }
 611
 612                 if (hits == samples) {
 613                         if (done)
 614                                 break;
 615                         err = poll(event_array, nr_poll, -1);
 616                         waking++;
 617                 }
 618
 619                 if (done) {
 620                         for (i = 0; i < nr_cpu; i++) {
 621                                 for (counter = 0; counter < nr_counters; counter++)
 622                                         ioctl(fd[i][counter], PERF_EVENT_IOC_DISABLE);
 623                         }
 624                 }
 625         }
 626
 627         fprintf(stderr, "[ perf record: Woken up %ld times to write data ]\n", waking);
 628
 629         /*
 630          * Approximate RIP event size: 24 bytes.
 631          */
 632         fprintf(stderr,
 633                 "[ perf record: Captured and wrote %.3f MB %s (~%lld samples) ]\n",
 634                 (double)bytes_written / 1024.0 / 1024.0,
 635                 output_name,
 636                 bytes_written / 24);
 637
 638         return 0;
 639 }
 640
 641 static const char * const record_usage[] = {
 642         "perf record [<options>] [<command>]",
 643         "perf record [<options>] -- <command> [<options>]",
 644         NULL
 645 };
 646
 647 static const struct option options[] = {
 648         OPT_CALLBACK('e', "event", NULL, "event",
 649                      "event selector. use 'perf list' to list available events",
 650                      parse_events),
 651         OPT_CALLBACK(0, "filter", NULL, "filter",
 652                      "event filter", parse_filter),
 653         OPT_INTEGER('p', "pid", &target_pid,
 654                     "record events on existing pid"),
 655         OPT_INTEGER('r', "realtime", &realtime_prio,
 656                     "collect data with this RT SCHED_FIFO priority"),
 657         OPT_BOOLEAN('R', "raw-samples", &raw_samples,
 658                     "collect raw sample records from all opened counters"),
 659         OPT_BOOLEAN('a', "all-cpus", &system_wide,
 660                             "system-wide collection from all CPUs"),
 661         OPT_BOOLEAN('A', "append", &append_file,
 662                             "append to the output file to do incremental profiling"),
 663         OPT_INTEGER('C', "profile_cpu", &profile_cpu,
 664                             "CPU to profile on"),
 665         OPT_BOOLEAN('f', "force", &force,
 666                         "overwrite existing data file"),
 667         OPT_LONG('c', "count", &default_interval,
 668                     "event period to sample"),
 669         OPT_STRING('o', "output", &output_name, "file",
 670                     "output file name"),
 671         OPT_BOOLEAN('i', "inherit", &inherit,
 672                     "child tasks inherit counters"),
 673         OPT_INTEGER('F', "freq", &freq,
 674                     "profile at this frequency"),
 675         OPT_INTEGER('m', "mmap-pages", &mmap_pages,
 676                     "number of mmap data pages"),
 677         OPT_BOOLEAN('g', "call-graph", &call_graph,
 678                     "do call-graph (stack chain/backtrace) recording"),
 679         OPT_BOOLEAN('v', "verbose", &verbose,
 680                     "be more verbose (show counter open errors, etc)"),
 681         OPT_BOOLEAN('s', "stat", &inherit_stat,
 682                     "per thread counts"),
 683         OPT_BOOLEAN('d', "data", &sample_address,
 684                     "Sample addresses"),
 685         OPT_BOOLEAN('n', "no-samples", &no_samples,
 686                     "don't sample"),
 687         OPT_BOOLEAN('M', "multiplex", &multiplex,
 688                     "multiplex counter output in a single channel"),
 689         OPT_END()
 690 };
 691
 692 int cmd_record(int argc, const char **argv, const char *prefix __used)
 693 {
 694         int counter;
 695
 696         argc = parse_options(argc, argv, options, record_usage,
 697                             PARSE_OPT_STOP_AT_NON_OPTION);
 698         if (!argc && target_pid == -1 && !system_wide && profile_cpu == -1)
 699                 usage_with_options(record_usage, options);
 700
 701         symbol__init();
 702
 703         if (!nr_counters) {
 704                 nr_counters     = 1;
 705                 attrs[0].type   = PERF_TYPE_HARDWARE;
 706                 attrs[0].config = PERF_COUNT_HW_CPU_CYCLES;
 707         }
 708
 709         /*
 710          * User specified count overrides default frequency.
 711          */
 712         if (default_interval)
 713                 freq = 0;
 714         else if (freq) {
 715                 default_interval = freq;
 716         } else {
 717                 fprintf(stderr, "frequency and count are zero, aborting\n");
 718                 exit(EXIT_FAILURE);
 719         }
 720
 721         for (counter = 0; counter < nr_counters; counter++) {
 722                 if (attrs[counter].sample_period)
 723                         continue;
 724
 725                 attrs[counter].sample_period = default_interval;
 726         }
 727
 728         return __cmd_record(argc, argv);
 729 }