lttd/lttd.c

   1 /* lttd
   2  *
   3  * Linux Trace Toolkit Daemon
   4  *
   5  * This is a simple daemon that reads a few relay+debugfs channels and save
   6  * them in a trace.
   7  *
   8  * CPU hot-plugging is supported using inotify.
   9  *
  10  * Copyright 2005 -
  11  *      Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
  12  */
  13
  14 #ifdef HAVE_CONFIG_H
  15 #include <config.h>
  16 #endif
  17
  18 #define _REENTRANT
  19 #define _GNU_SOURCE
  20 #include <features.h>
  21 #include <stdio.h>
  22 #include <unistd.h>
  23 #include <errno.h>
  24 #include <sys/types.h>
  25 #include <sys/stat.h>
  26 #include <stdlib.h>
  27 #include <dirent.h>
  28 #include <string.h>
  29 #include <fcntl.h>
  30 #include <sys/poll.h>
  31 #include <sys/mman.h>
  32 #include <signal.h>
  33 #include <pthread.h>
  34 #include <sys/syscall.h>
  35 #include <unistd.h>
  36 #include <asm/ioctls.h>
  37
  38 #include <linux/version.h>
  39
  40 /* Relayfs IOCTL */
  41 #include <asm/ioctl.h>
  42 #include <asm/types.h>
  43
  44 /* Get the next sub buffer that can be read. */
  45 #define RELAY_GET_SUBBUF        _IOR(0xF5, 0x00,__u32)
  46 /* Release the oldest reserved (by "get") sub buffer. */
  47 #define RELAY_PUT_SUBBUF        _IOW(0xF5, 0x01,__u32)
  48 /* returns the number of sub buffers in the per cpu channel. */
  49 #define RELAY_GET_N_SUBBUFS     _IOR(0xF5, 0x02,__u32)
  50 /* returns the size of the sub buffers. */
  51 #define RELAY_GET_SUBBUF_SIZE   _IOR(0xF5, 0x03,__u32)
  52
  53 #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,14)
  54 #include <sys/inotify.h>
  55 #if 0   /* should now be provided by libc. */
  56 /* From the inotify-tools 2.6 package */
  57 static inline int inotify_init (void)
  58 {
  59         return syscall (__NR_inotify_init);
  60 }
  61
  62 static inline int inotify_add_watch (int fd, const char *name, __u32 mask)
  63 {
  64         return syscall (__NR_inotify_add_watch, fd, name, mask);
  65 }
  66
  67 static inline int inotify_rm_watch (int fd, __u32 wd)
  68 {
  69         return syscall (__NR_inotify_rm_watch, fd, wd);
  70 }
  71 #endif //0
  72 #define HAS_INOTIFY
  73 #else
  74 static inline int inotify_init (void)
  75 {
  76         return -1;
  77 }
  78
  79 static inline int inotify_add_watch (int fd, const char *name, __u32 mask)
  80 {
  81         return 0;
  82 }
  83
  84 static inline int inotify_rm_watch (int fd, __u32 wd)
  85 {
  86         return 0;
  87 }
  88 #undef HAS_INOTIFY
  89 #endif
  90
  91 enum {
  92         GET_SUBBUF,
  93         PUT_SUBBUF,
  94         GET_N_BUBBUFS,
  95         GET_SUBBUF_SIZE
  96 };
  97
  98 struct fd_pair {
  99         int channel;
 100         int trace;
 101         unsigned int n_subbufs;
 102         unsigned int subbuf_size;
 103         void *mmap;
 104         pthread_mutex_t mutex;
 105 };
 106
 107 struct channel_trace_fd {
 108         struct fd_pair *pair;
 109         int num_pairs;
 110 };
 111
 112 struct inotify_watch {
 113         int wd;
 114         char path_channel[PATH_MAX];
 115         char path_trace[PATH_MAX];
 116 };
 117
 118 struct inotify_watch_array {
 119         struct inotify_watch *elem;
 120         int num;
 121 };
 122
 123 static __thread int thread_pipe[2];
 124
 125 struct channel_trace_fd fd_pairs = { NULL, 0 };
 126 int inotify_fd = -1;
 127 struct inotify_watch_array inotify_watch_array = { NULL, 0 };
 128
 129 /* protects fd_pairs and inotify_watch_array */
 130 pthread_rwlock_t fd_pairs_lock = PTHREAD_RWLOCK_INITIALIZER;
 131
 132
 133 static char             *trace_name = NULL;
 134 static char             *channel_name = NULL;
 135 static int              daemon_mode = 0;
 136 static int              append_mode = 0;
 137 static unsigned long    num_threads = 1;
 138 volatile static int     quit_program = 0;       /* For signal handler */
 139 static int              dump_flight_only = 0;
 140 static int              dump_normal_only = 0;
 141 static int              verbose_mode = 0;
 142
 143 #define printf_verbose(fmt, args...) \
 144   do {                               \
 145     if (verbose_mode)                \
 146       printf(fmt, ##args);           \
 147   } while (0)
 148
 149 /* Args :
 150  *
 151  * -t directory         Directory name of the trace to write to. Will be created.
 152  * -c directory         Root directory of the debugfs trace channels.
 153  * -d                   Run in background (daemon).
 154  * -a                   Trace append mode.
 155  * -s                   Send SIGUSR1 to parent when ready for IO.
 156  */
 157 void show_arguments(void)
 158 {
 159         printf("Please use the following arguments :\n");
 160         printf("\n");
 161         printf("-t directory  Directory name of the trace to write to.\n"
 162                                  "              It will be created.\n");
 163         printf("-c directory  Root directory of the debugfs trace channels.\n");
 164         printf("-d            Run in background (daemon).\n");
 165         printf("-a            Append to an possibly existing trace.\n");
 166         printf("-N            Number of threads to start.\n");
 167         printf("-f            Dump only flight recorder channels.\n");
 168         printf("-n            Dump only normal channels.\n");
 169         printf("-v            Verbose mode.\n");
 170         printf("\n");
 171 }
 172
 173
 174 /* parse_arguments
 175  *
 176  * Parses the command line arguments.
 177  *
 178  * Returns 1 if the arguments were correct, but doesn't ask for program
 179  * continuation. Returns -1 if the arguments are incorrect, or 0 if OK.
 180  */
 181 int parse_arguments(int argc, char **argv)
 182 {
 183         int ret = 0;
 184         int argn = 1;
 185
 186         if(argc == 2) {
 187                 if(strcmp(argv[1], "-h") == 0) {
 188                         return 1;
 189                 }
 190         }
 191
 192         while(argn < argc) {
 193
 194                 switch(argv[argn][0]) {
 195                         case '-':
 196                                 switch(argv[argn][1]) {
 197                                         case 't':
 198                                                 if(argn+1 < argc) {
 199                                                         trace_name = argv[argn+1];
 200                                                         argn++;
 201                                                 }
 202                                                 break;
 203                                         case 'c':
 204                                                 if(argn+1 < argc) {
 205                                                         channel_name = argv[argn+1];
 206                                                         argn++;
 207                                                 }
 208                                                 break;
 209                                         case 'd':
 210                                                 daemon_mode = 1;
 211                                                 break;
 212                                         case 'a':
 213                                                 append_mode = 1;
 214                                                 break;
 215                                         case 'N':
 216                                                 if(argn+1 < argc) {
 217                                                         num_threads = strtoul(argv[argn+1], NULL, 0);
 218                                                         argn++;
 219                                                 }
 220                                                 break;
 221                                         case 'f':
 222                                                 dump_flight_only = 1;
 223                                                 break;
 224                                         case 'n':
 225                                                 dump_normal_only = 1;
 226                                                 break;
 227                                         case 'v':
 228                                                 verbose_mode = 1;
 229                                                 break;
 230                                         default:
 231                                                 printf("Invalid argument '%s'.\n", argv[argn]);
 232                                                 printf("\n");
 233                                                 ret = -1;
 234                                 }
 235                                 break;
 236                         default:
 237                                 printf("Invalid argument '%s'.\n", argv[argn]);
 238                                 printf("\n");
 239                                 ret = -1;
 240                 }
 241                 argn++;
 242         }
 243
 244         if(trace_name == NULL) {
 245                 printf("Please specify a trace name.\n");
 246                 printf("\n");
 247                 ret = -1;
 248         }
 249
 250         if(channel_name == NULL) {
 251                 printf("Please specify a channel name.\n");
 252                 printf("\n");
 253                 ret = -1;
 254         }
 255
 256         return ret;
 257 }
 258
 259 void show_info(void)
 260 {
 261         printf("Linux Trace Toolkit Trace Daemon " VERSION "\n");
 262         printf("\n");
 263         printf("Reading from debugfs directory : %s\n", channel_name);
 264         printf("Writing to trace directory : %s\n", trace_name);
 265         printf("\n");
 266 }
 267
 268
 269 /* signal handling */
 270
 271 static void handler(int signo)
 272 {
 273         printf("Signal %d received : exiting cleanly\n", signo);
 274         quit_program = 1;
 275 }
 276
 277
 278 int open_buffer_file(char *filename, char *path_channel, char *path_trace,
 279         struct channel_trace_fd *fd_pairs)
 280 {
 281         int open_ret = 0;
 282         int ret = 0;
 283         struct stat stat_buf;
 284
 285         if(strncmp(filename, "flight-", sizeof("flight-")-1) != 0) {
 286                 if(dump_flight_only) {
 287                         printf_verbose("Skipping normal channel %s\n",
 288                                 path_channel);
 289                         return 0;
 290                 }
 291         } else {
 292                 if(dump_normal_only) {
 293                         printf_verbose("Skipping flight channel %s\n",
 294                                 path_channel);
 295                         return 0;
 296                 }
 297         }
 298         printf_verbose("Opening file.\n");
 299
 300         fd_pairs->pair = realloc(fd_pairs->pair,
 301                         ++fd_pairs->num_pairs * sizeof(struct fd_pair));
 302
 303         /* Open the channel in read mode */
 304         fd_pairs->pair[fd_pairs->num_pairs-1].channel =
 305                 open(path_channel, O_RDONLY | O_NONBLOCK);
 306         if(fd_pairs->pair[fd_pairs->num_pairs-1].channel == -1) {
 307                 perror(path_channel);
 308                 fd_pairs->num_pairs--;
 309                 return 0;       /* continue */
 310         }
 311         /* Open the trace in write mode, only append if append_mode */
 312         ret = stat(path_trace, &stat_buf);
 313         if(ret == 0) {
 314                 if(append_mode) {
 315                         printf_verbose("Appending to file %s as requested\n",
 316                                 path_trace);
 317
 318                         fd_pairs->pair[fd_pairs->num_pairs-1].trace =
 319                                 open(path_trace, O_WRONLY,
 320                                                 S_IRWXU|S_IRWXG|S_IRWXO);
 321                         if(fd_pairs->pair[fd_pairs->num_pairs-1].trace == -1) {
 322                                 perror(path_trace);
 323                         }
 324                         ret = lseek(fd_pairs->pair[fd_pairs->num_pairs-1].trace,
 325                                     0, SEEK_END);
 326                         if (ret < 0) {
 327                                 perror(path_trace);
 328                         }
 329                 } else {
 330                         printf("File %s exists, cannot open. Try append mode.\n", path_trace);
 331                         open_ret = -1;
 332                         goto end;
 333                 }
 334         } else {
 335                 if(errno == ENOENT) {
 336                         fd_pairs->pair[fd_pairs->num_pairs-1].trace =
 337                                 open(path_trace, O_WRONLY|O_CREAT|O_EXCL,
 338                                                 S_IRWXU|S_IRWXG|S_IRWXO);
 339                         if(fd_pairs->pair[fd_pairs->num_pairs-1].trace == -1) {
 340                                 perror(path_trace);
 341                         }
 342                 }
 343         }
 344 end:
 345         return open_ret;
 346 }
 347
 348 int open_channel_trace_pairs(char *subchannel_name, char *subtrace_name,
 349                 struct channel_trace_fd *fd_pairs, int *inotify_fd,
 350                 struct inotify_watch_array *iwatch_array)
 351 {
 352         DIR *channel_dir = opendir(subchannel_name);
 353         struct dirent *entry;
 354         struct stat stat_buf;
 355         int ret;
 356         char path_channel[PATH_MAX];
 357         int path_channel_len;
 358         char *path_channel_ptr;
 359         char path_trace[PATH_MAX];
 360         int path_trace_len;
 361         char *path_trace_ptr;
 362         int open_ret = 0;
 363
 364         if(channel_dir == NULL) {
 365                 perror(subchannel_name);
 366                 open_ret = ENOENT;
 367                 goto end;
 368         }
 369
 370         printf_verbose("Creating trace subdirectory %s\n", subtrace_name);
 371         ret = mkdir(subtrace_name, S_IRWXU|S_IRWXG|S_IRWXO);
 372         if(ret == -1) {
 373                 if(errno != EEXIST) {
 374                         perror(subtrace_name);
 375                         open_ret = -1;
 376                         goto end;
 377                 }
 378         }
 379
 380         strncpy(path_channel, subchannel_name, PATH_MAX-1);
 381         path_channel_len = strlen(path_channel);
 382         path_channel[path_channel_len] = '/';
 383         path_channel_len++;
 384         path_channel_ptr = path_channel + path_channel_len;
 385
 386         strncpy(path_trace, subtrace_name, PATH_MAX-1);
 387         path_trace_len = strlen(path_trace);
 388         path_trace[path_trace_len] = '/';
 389         path_trace_len++;
 390         path_trace_ptr = path_trace + path_trace_len;
 391
 392 #ifdef HAS_INOTIFY
 393         iwatch_array->elem = realloc(iwatch_array->elem,
 394                 ++iwatch_array->num * sizeof(struct inotify_watch));
 395
 396         printf_verbose("Adding inotify for channel %s\n", path_channel);
 397         iwatch_array->elem[iwatch_array->num-1].wd = inotify_add_watch(*inotify_fd, path_channel, IN_CREATE);
 398         strcpy(iwatch_array->elem[iwatch_array->num-1].path_channel, path_channel);
 399         strcpy(iwatch_array->elem[iwatch_array->num-1].path_trace, path_trace);
 400         printf_verbose("Added inotify for channel %s, wd %u\n",
 401                 iwatch_array->elem[iwatch_array->num-1].path_channel,
 402                 iwatch_array->elem[iwatch_array->num-1].wd);
 403 #endif
 404
 405         while((entry = readdir(channel_dir)) != NULL) {
 406
 407                 if(entry->d_name[0] == '.') continue;
 408
 409                 strncpy(path_channel_ptr, entry->d_name, PATH_MAX - path_channel_len);
 410                 strncpy(path_trace_ptr, entry->d_name, PATH_MAX - path_trace_len);
 411
 412                 ret = stat(path_channel, &stat_buf);
 413                 if(ret == -1) {
 414                         perror(path_channel);
 415                         continue;
 416                 }
 417
 418                 printf_verbose("Channel file : %s\n", path_channel);
 419
 420                 if(S_ISDIR(stat_buf.st_mode)) {
 421
 422                         printf_verbose("Entering channel subdirectory...\n");
 423                         ret = open_channel_trace_pairs(path_channel, path_trace, fd_pairs,
 424                                 inotify_fd, iwatch_array);
 425                         if(ret < 0) continue;
 426                 } else if(S_ISREG(stat_buf.st_mode)) {
 427                         open_ret = open_buffer_file(entry->d_name, path_channel, path_trace,
 428                                 fd_pairs);
 429                         if(open_ret)
 430                                 goto end;
 431                 }
 432         }
 433
 434 end:
 435         closedir(channel_dir);
 436
 437         return open_ret;
 438 }
 439
 440
 441 int read_subbuffer(struct fd_pair *pair)
 442 {
 443         unsigned int consumed_old;
 444         int err;
 445         long ret;
 446         unsigned long len;
 447         off_t offset;
 448
 449
 450         err = ioctl(pair->channel, RELAY_GET_SUBBUF, &consumed_old);
 451         printf_verbose("cookie : %u\n", consumed_old);
 452         if(err != 0) {
 453                 ret = errno;
 454                 perror("Reserving sub buffer failed (everything is normal, it is due to concurrency)");
 455                 goto get_error;
 456         }
 457 #if 0
 458         err = TEMP_FAILURE_RETRY(write(pair->trace,
 459                                 pair->mmap
 460                                         + (consumed_old & ((pair->n_subbufs * pair->subbuf_size)-1)),
 461                                 pair->subbuf_size));
 462
 463         if(err < 0) {
 464                 ret = errno;
 465                 perror("Error in writing to file");
 466                 goto write_error;
 467         }
 468 #endif //0
 469         len = pair->subbuf_size;
 470         offset = 0;
 471         while (len > 0) {
 472                 printf_verbose("splice chan to pipe offset %lu\n",
 473                         (unsigned long)offset);
 474                 ret = splice(pair->channel, &offset, thread_pipe[1], NULL,
 475                         len, SPLICE_F_MOVE | SPLICE_F_MORE);
 476                 printf_verbose("splice chan to pipe ret %ld\n", ret);
 477                 if (ret < 0) {
 478                         perror("Error in relay splice");
 479                         goto write_error;
 480                 }
 481                 ret = splice(thread_pipe[0], NULL, pair->trace, NULL,
 482                         ret, SPLICE_F_MOVE | SPLICE_F_MORE);
 483                 printf_verbose("splice pipe to file %ld\n", ret);
 484                 if (ret < 0) {
 485                         perror("Error in file splice");
 486                         goto write_error;
 487                 }
 488                 len -= ret;
 489         }
 490
 491 #if 0
 492         err = fsync(pair->trace);
 493         if(err < 0) {
 494                 ret = errno;
 495                 perror("Error in writing to file");
 496                 goto write_error;
 497         }
 498 #endif //0
 499 write_error:
 500         ret = 0;
 501         err = ioctl(pair->channel, RELAY_PUT_SUBBUF, &consumed_old);
 502         if(err != 0) {
 503                 ret = errno;
 504                 if(errno == EFAULT) {
 505                         perror("Error in unreserving sub buffer\n");
 506                 } else if(errno == EIO) {
 507                         perror("Reader has been pushed by the writer, last subbuffer corrupted.");
 508                         /* FIXME : we may delete the last written buffer if we wish. */
 509                 }
 510                 goto get_error;
 511         }
 512
 513 get_error:
 514         return ret;
 515 }
 516
 517
 518 int map_channels(struct channel_trace_fd *fd_pairs,
 519         int idx_begin, int idx_end)
 520 {
 521         int i,j;
 522         int ret=0;
 523
 524         if(fd_pairs->num_pairs <= 0) {
 525                 printf("No channel to read\n");
 526                 goto end;
 527         }
 528
 529         /* Get the subbuf sizes and number */
 530
 531         for(i=idx_begin;i<idx_end;i++) {
 532                 struct fd_pair *pair = &fd_pairs->pair[i];
 533
 534                 ret = ioctl(pair->channel, RELAY_GET_N_SUBBUFS,
 535                                                         &pair->n_subbufs);
 536                 if(ret != 0) {
 537                         perror("Error in getting the number of subbuffers");
 538                         goto end;
 539                 }
 540                 ret = ioctl(pair->channel, RELAY_GET_SUBBUF_SIZE,
 541                                                         &pair->subbuf_size);
 542                 if(ret != 0) {
 543                         perror("Error in getting the size of the subbuffers");
 544                         goto end;
 545                 }
 546                 ret = pthread_mutex_init(&pair->mutex, NULL);   /* Fast mutex */
 547                 if(ret != 0) {
 548                         perror("Error in mutex init");
 549                         goto end;
 550                 }
 551         }
 552
 553 #if 0
 554         /* Mmap each FD */
 555         for(i=idx_begin;i<idx_end;i++) {
 556                 struct fd_pair *pair = &fd_pairs->pair[i];
 557
 558                 pair->mmap = mmap(0, pair->subbuf_size * pair->n_subbufs, PROT_READ,
 559                                 MAP_SHARED, pair->channel, 0);
 560                 if(pair->mmap == MAP_FAILED) {
 561                         perror("Mmap error");
 562                         goto munmap;
 563                 }
 564         }
 565
 566         goto end; /* success */
 567
 568         /* Error handling */
 569         /* munmap only the successfully mmapped indexes */
 570 munmap:
 571                 /* Munmap each FD */
 572         for(j=idx_begin;j<i;j++) {
 573                 struct fd_pair *pair = &fd_pairs->pair[j];
 574                 int err_ret;
 575
 576                 err_ret = munmap(pair->mmap, pair->subbuf_size * pair->n_subbufs);
 577                 if(err_ret != 0) {
 578                         perror("Error in munmap");
 579                 }
 580                 ret |= err_ret;
 581         }
 582
 583 #endif //0
 584 end:
 585         return ret;
 586 }
 587
 588 int unmap_channels(struct channel_trace_fd *fd_pairs)
 589 {
 590         int j;
 591         int ret=0;
 592
 593         /* Munmap each FD */
 594         for(j=0;j<fd_pairs->num_pairs;j++) {
 595                 struct fd_pair *pair = &fd_pairs->pair[j];
 596                 int err_ret;
 597
 598 #if 0
 599                 err_ret = munmap(pair->mmap, pair->subbuf_size * pair->n_subbufs);
 600                 if(err_ret != 0) {
 601                         perror("Error in munmap");
 602                 }
 603                 ret |= err_ret;
 604 #endif //0
 605                 err_ret = pthread_mutex_destroy(&pair->mutex);
 606                 if(err_ret != 0) {
 607                         perror("Error in mutex destroy");
 608                 }
 609                 ret |= err_ret;
 610         }
 611
 612         return ret;
 613 }
 614
 615 #ifdef HAS_INOTIFY
 616 /* Inotify event arrived.
 617  *
 618  * Only support add file for now.
 619  */
 620
 621 int read_inotify(int inotify_fd,
 622         struct channel_trace_fd *fd_pairs,
 623         struct inotify_watch_array *iwatch_array)
 624 {
 625         char buf[sizeof(struct inotify_event) + PATH_MAX];
 626         char path_channel[PATH_MAX];
 627         char path_trace[PATH_MAX];
 628         ssize_t len;
 629         struct inotify_event *ievent;
 630         size_t offset;
 631         unsigned int i;
 632         int ret;
 633         int old_num;
 634
 635         offset = 0;
 636         len = read(inotify_fd, buf, sizeof(struct inotify_event) + PATH_MAX);
 637         if(len < 0) {
 638
 639                 if(errno == EAGAIN)
 640                         return 0;  /* another thread got the data before us */
 641
 642                 printf("Error in read from inotify FD %s.\n", strerror(len));
 643                 return -1;
 644         }
 645         while(offset < len) {
 646                 ievent = (struct inotify_event *)&(buf[offset]);
 647                 for(i=0; i<iwatch_array->num; i++) {
 648                         if(iwatch_array->elem[i].wd == ievent->wd &&
 649                                 ievent->mask == IN_CREATE) {
 650                                 printf_verbose(
 651                                         "inotify wd %u event mask : %u for %s%s\n",
 652                                         ievent->wd, ievent->mask,
 653                                         iwatch_array->elem[i].path_channel,
 654                                         ievent->name);
 655                                 old_num = fd_pairs->num_pairs;
 656                                 strcpy(path_channel, iwatch_array->elem[i].path_channel);
 657                                 strcat(path_channel, ievent->name);
 658                                 strcpy(path_trace, iwatch_array->elem[i].path_trace);
 659                                 strcat(path_trace, ievent->name);
 660                                 if(ret = open_buffer_file(ievent->name, path_channel,
 661                                         path_trace, fd_pairs)) {
 662                                         printf("Error opening buffer file\n");
 663                                         return -1;
 664                                 }
 665                                 if(ret = map_channels(fd_pairs, old_num, fd_pairs->num_pairs)) {
 666                                         printf("Error mapping channel\n");
 667                                         return -1;
 668                                 }
 669
 670                         }
 671                 }
 672                 offset += sizeof(*ievent) + ievent->len;
 673         }
 674 }
 675 #endif //HAS_INOTIFY
 676
 677 /* read_channels
 678  *
 679  * Thread worker.
 680  *
 681  * Read the debugfs channels and write them in the paired tracefiles.
 682  *
 683  * @fd_pairs : paired channels and trace files.
 684  *
 685  * returns 0 on success, -1 on error.
 686  *
 687  * Note that the high priority polled channels are consumed first. We then poll
 688  * again to see if these channels are still in priority. Only when no
 689  * high priority channel is left, we start reading low priority channels.
 690  *
 691  * Note that a channel is considered high priority when the buffer is almost
 692  * full.
 693  */
 694
 695 int read_channels(unsigned long thread_num, struct channel_trace_fd *fd_pairs,
 696         int inotify_fd, struct inotify_watch_array *iwatch_array)
 697 {
 698         struct pollfd *pollfd = NULL;
 699         int num_pollfd;
 700         int i,j;
 701         int num_rdy, num_hup;
 702         int high_prio;
 703         int ret = 0;
 704         int inotify_fds;
 705         unsigned int old_num;
 706
 707 #ifdef HAS_INOTIFY
 708         inotify_fds = 1;
 709 #else
 710         inotify_fds = 0;
 711 #endif
 712
 713         pthread_rwlock_rdlock(&fd_pairs_lock);
 714
 715         /* Start polling the FD. Keep one fd for inotify */
 716         pollfd = malloc((inotify_fds + fd_pairs->num_pairs) * sizeof(struct pollfd));
 717
 718 #ifdef HAS_INOTIFY
 719         pollfd[0].fd = inotify_fd;
 720         pollfd[0].events = POLLIN|POLLPRI;
 721 #endif
 722
 723         for(i=0;i<fd_pairs->num_pairs;i++) {
 724                 pollfd[inotify_fds+i].fd = fd_pairs->pair[i].channel;
 725                 pollfd[inotify_fds+i].events = POLLIN|POLLPRI;
 726         }
 727         num_pollfd = inotify_fds + fd_pairs->num_pairs;
 728
 729
 730         pthread_rwlock_unlock(&fd_pairs_lock);
 731
 732         while(1) {
 733                 high_prio = 0;
 734                 num_hup = 0;
 735 #ifdef DEBUG
 736                 printf("Press a key for next poll...\n");
 737                 char buf[1];
 738                 read(STDIN_FILENO, &buf, 1);
 739                 printf("Next poll (polling %d fd) :\n", num_pollfd);
 740 #endif //DEBUG
 741
 742                 /* Have we received a signal ? */
 743                 if(quit_program) break;
 744
 745                 num_rdy = poll(pollfd, num_pollfd, -1);
 746
 747                 if(num_rdy == -1) {
 748                         perror("Poll error");
 749                         goto free_fd;
 750                 }
 751
 752                 printf_verbose("Data received\n");
 753 #ifdef HAS_INOTIFY
 754                 switch(pollfd[0].revents) {
 755                         case POLLERR:
 756                                 printf_verbose(
 757                                         "Error returned in polling inotify fd %d.\n",
 758                                         pollfd[0].fd);
 759                                 break;
 760                         case POLLHUP:
 761                                 printf_verbose(
 762                                         "Polling inotify fd %d tells it has hung up.\n",
 763                                         pollfd[0].fd);
 764                                 break;
 765                         case POLLNVAL:
 766                                 printf_verbose(
 767                                         "Polling inotify fd %d tells fd is not open.\n",
 768                                         pollfd[0].fd);
 769                                 break;
 770                         case POLLPRI:
 771                         case POLLIN:
 772                                 printf_verbose(
 773                                         "Polling inotify fd %d : data ready.\n",
 774                                         pollfd[0].fd);
 775
 776                                 pthread_rwlock_wrlock(&fd_pairs_lock);
 777                                 read_inotify(inotify_fd, fd_pairs, iwatch_array);
 778                                 pthread_rwlock_unlock(&fd_pairs_lock);
 779
 780                         break;
 781                 }
 782 #endif
 783
 784                 for(i=inotify_fds;i<num_pollfd;i++) {
 785                         switch(pollfd[i].revents) {
 786                                 case POLLERR:
 787                                         printf_verbose(
 788                                                 "Error returned in polling fd %d.\n",
 789                                                 pollfd[i].fd);
 790                                         num_hup++;
 791                                         break;
 792                                 case POLLHUP:
 793                                         printf_verbose(
 794                                                 "Polling fd %d tells it has hung up.\n",
 795                                                 pollfd[i].fd);
 796                                         num_hup++;
 797                                         break;
 798                                 case POLLNVAL:
 799                                         printf_verbose(
 800                                                 "Polling fd %d tells fd is not open.\n",
 801                                                 pollfd[i].fd);
 802                                         num_hup++;
 803                                         break;
 804                                 case POLLPRI:
 805                                         pthread_rwlock_rdlock(&fd_pairs_lock);
 806                                         if(pthread_mutex_trylock(&fd_pairs->pair[i-inotify_fds].mutex) == 0) {
 807                                                 printf_verbose(
 808                                                         "Urgent read on fd %d\n",
 809                                                         pollfd[i].fd);
 810                                                 /* Take care of high priority channels first. */
 811                                                 high_prio = 1;
 812                                                 /* it's ok to have an unavailable subbuffer */
 813                                                 ret = read_subbuffer(&fd_pairs->pair[i-inotify_fds]);
 814                                                 if(ret == EAGAIN) ret = 0;
 815
 816                                                 ret = pthread_mutex_unlock(&fd_pairs->pair[i-inotify_fds].mutex);
 817                                                 if(ret)
 818                                                         printf("Error in mutex unlock : %s\n", strerror(ret));
 819                                         }
 820                                         pthread_rwlock_unlock(&fd_pairs_lock);
 821                                         break;
 822                         }
 823                 }
 824                 /* If every buffer FD has hung up, we end the read loop here */
 825                 if(num_hup == num_pollfd - inotify_fds) break;
 826
 827                 if(!high_prio) {
 828                         for(i=inotify_fds;i<num_pollfd;i++) {
 829                                 switch(pollfd[i].revents) {
 830                                         case POLLIN:
 831                                                 pthread_rwlock_rdlock(&fd_pairs_lock);
 832                                                 if(pthread_mutex_trylock(&fd_pairs->pair[i-inotify_fds].mutex) == 0) {
 833                                                         /* Take care of low priority channels. */
 834                                                         printf_verbose(
 835                                                                 "Normal read on fd %d\n",
 836                                                                 pollfd[i].fd);
 837                                                         /* it's ok to have an unavailable subbuffer */
 838                                                         ret = read_subbuffer(&fd_pairs->pair[i-inotify_fds]);
 839                                                         if(ret == EAGAIN) ret = 0;
 840
 841                                                         ret = pthread_mutex_unlock(&fd_pairs->pair[i-inotify_fds].mutex);
 842                                                         if(ret)
 843                                                                 printf("Error in mutex unlock : %s\n", strerror(ret));
 844                                                 }
 845                                                 pthread_rwlock_unlock(&fd_pairs_lock);
 846                                                 break;
 847                                 }
 848                         }
 849                 }
 850
 851                 /* Update pollfd array if an entry was added to fd_pairs */
 852                 pthread_rwlock_rdlock(&fd_pairs_lock);
 853                 if((inotify_fds + fd_pairs->num_pairs) != num_pollfd) {
 854                         pollfd = realloc(pollfd,
 855                                         (inotify_fds + fd_pairs->num_pairs) * sizeof(struct pollfd));
 856                         for(i=num_pollfd-inotify_fds;i<fd_pairs->num_pairs;i++) {
 857                                 pollfd[inotify_fds+i].fd = fd_pairs->pair[i].channel;
 858                                 pollfd[inotify_fds+i].events = POLLIN|POLLPRI;
 859                         }
 860                         num_pollfd = fd_pairs->num_pairs + inotify_fds;
 861                 }
 862                 pthread_rwlock_unlock(&fd_pairs_lock);
 863
 864                 /* NB: If the fd_pairs structure is updated by another thread from this
 865                  *     point forward, the current thread will wait in the poll without
 866                  *     monitoring the new channel. However, this thread will add the
 867                  *     new channel on next poll (and this should not take too much time
 868                  *     on a loaded system).
 869                  *
 870                  *     This event is quite unlikely and can only occur if a CPU is
 871                  *     hot-plugged while multple lttd threads are running.
 872                  */
 873         }
 874
 875 free_fd:
 876         free(pollfd);
 877
 878 end:
 879         return ret;
 880 }
 881
 882
 883 void close_channel_trace_pairs(struct channel_trace_fd *fd_pairs, int inotify_fd,
 884         struct inotify_watch_array *iwatch_array)
 885 {
 886         int i;
 887         int ret;
 888
 889         for(i=0;i<fd_pairs->num_pairs;i++) {
 890                 ret = close(fd_pairs->pair[i].channel);
 891                 if(ret == -1) perror("Close error on channel");
 892                 ret = close(fd_pairs->pair[i].trace);
 893                 if(ret == -1) perror("Close error on trace");
 894         }
 895         free(fd_pairs->pair);
 896         free(iwatch_array->elem);
 897 }
 898
 899 /* Thread worker */
 900 void * thread_main(void *arg)
 901 {
 902         long ret;
 903         unsigned long thread_num = (unsigned long)arg;
 904
 905         ret = pipe(thread_pipe);
 906         if (ret < 0) {
 907                 perror("Error creating pipe");
 908                 return (void*)ret;
 909         }
 910         ret = read_channels(thread_num, &fd_pairs, inotify_fd, &inotify_watch_array);
 911         close(thread_pipe[0]);  /* close read end */
 912         close(thread_pipe[1]);  /* close write end */
 913         return (void*)ret;
 914 }
 915
 916
 917 int channels_init()
 918 {
 919         int ret = 0;
 920
 921         inotify_fd = inotify_init();
 922         fcntl(inotify_fd, F_SETFL, O_NONBLOCK);
 923
 924         if(ret = open_channel_trace_pairs(channel_name, trace_name, &fd_pairs,
 925                         &inotify_fd, &inotify_watch_array))
 926                 goto close_channel;
 927         if (fd_pairs.num_pairs == 0) {
 928                 printf("No channel available for reading, exiting\n");
 929                 ret = -ENOENT;
 930                 goto close_channel;
 931         }
 932         if(ret = map_channels(&fd_pairs, 0, fd_pairs.num_pairs))
 933                 goto close_channel;
 934         return 0;
 935
 936 close_channel:
 937         close_channel_trace_pairs(&fd_pairs, inotify_fd, &inotify_watch_array);
 938         if(inotify_fd >= 0)
 939                 close(inotify_fd);
 940         return ret;
 941 }
 942
 943
 944 int main(int argc, char ** argv)
 945 {
 946         int ret = 0;
 947         struct sigaction act;
 948         pthread_t *tids;
 949         unsigned long i;
 950         void *tret;
 951
 952         ret = parse_arguments(argc, argv);
 953
 954         if(ret != 0) show_arguments();
 955         if(ret < 0) return EINVAL;
 956         if(ret > 0) return 0;
 957
 958         show_info();
 959
 960         /* Connect the signal handlers */
 961         act.sa_handler = handler;
 962         act.sa_flags = 0;
 963         sigemptyset(&(act.sa_mask));
 964         sigaddset(&(act.sa_mask), SIGTERM);
 965         sigaddset(&(act.sa_mask), SIGQUIT);
 966         sigaddset(&(act.sa_mask), SIGINT);
 967         sigaction(SIGTERM, &act, NULL);
 968         sigaction(SIGQUIT, &act, NULL);
 969         sigaction(SIGINT, &act, NULL);
 970
 971         if(ret = channels_init())
 972                 return ret;
 973
 974         if(daemon_mode) {
 975                 ret = daemon(0, 0);
 976
 977                 if(ret == -1) {
 978                         perror("An error occured while daemonizing.");
 979                         exit(-1);
 980                 }
 981         }
 982
 983         tids = malloc(sizeof(pthread_t) * num_threads);
 984         for(i=0; i<num_threads; i++) {
 985
 986                 ret = pthread_create(&tids[i], NULL, thread_main, (void*)i);
 987                 if(ret) {
 988                         perror("Error creating thread");
 989                         break;
 990                 }
 991         }
 992
 993         for(i=0; i<num_threads; i++) {
 994                 ret = pthread_join(tids[i], &tret);
 995                 if(ret) {
 996                         perror("Error joining thread");
 997                         break;
 998                 }
 999                 if((long)tret != 0) {
1000                         printf("Error %s occured in thread %u\n",
1001                                 strerror((long)tret), i);
1002                 }
1003         }
1004
1005         free(tids);
1006         ret = unmap_channels(&fd_pairs);
1007         close_channel_trace_pairs(&fd_pairs, inotify_fd, &inotify_watch_array);
1008         if(inotify_fd >= 0)
1009                 close(inotify_fd);
1010
1011         return ret;
1012 }