blktrace.c

   1 /*
   2  * blktrace support code for fio
   3  */
   4 #include <stdio.h>
   5 #include <stdlib.h>
   6 #include <sys/ioctl.h>
   7 #include <linux/fs.h>
   8
   9 #include "flist.h"
  10 #include "fio.h"
  11 #include "blktrace.h"
  12 #include "blktrace_api.h"
  13 #include "oslib/linux-dev-lookup.h"
  14
  15 #define TRACE_FIFO_SIZE 8192
  16
  17 /*
  18  * fifo refill frontend, to avoid reading data in trace sized bites
  19  */
  20 static int refill_fifo(struct thread_data *td, struct fifo *fifo, int fd)
  21 {
  22         char buf[TRACE_FIFO_SIZE];
  23         unsigned int total;
  24         int ret;
  25
  26         total = sizeof(buf);
  27         if (total > fifo_room(fifo))
  28                 total = fifo_room(fifo);
  29
  30         ret = read(fd, buf, total);
  31         if (ret < 0) {
  32                 td_verror(td, errno, "read blktrace file");
  33                 return -1;
  34         }
  35
  36         if (ret > 0)
  37                 ret = fifo_put(fifo, buf, ret);
  38
  39         dprint(FD_BLKTRACE, "refill: filled %d bytes\n", ret);
  40         return ret;
  41 }
  42
  43 /*
  44  * Retrieve 'len' bytes from the fifo, refilling if necessary.
  45  */
  46 static int trace_fifo_get(struct thread_data *td, struct fifo *fifo, int fd,
  47                           void *buf, unsigned int len)
  48 {
  49         if (fifo_len(fifo) < len) {
  50                 int ret = refill_fifo(td, fifo, fd);
  51
  52                 if (ret < 0)
  53                         return ret;
  54         }
  55
  56         return fifo_get(fifo, buf, len);
  57 }
  58
  59 /*
  60  * Just discard the pdu by seeking past it.
  61  */
  62 static int discard_pdu(struct thread_data *td, struct fifo *fifo, int fd,
  63                        struct blk_io_trace *t)
  64 {
  65         if (t->pdu_len == 0)
  66                 return 0;
  67
  68         dprint(FD_BLKTRACE, "discard pdu len %u\n", t->pdu_len);
  69         return trace_fifo_get(td, fifo, fd, NULL, t->pdu_len);
  70 }
  71
  72 /*
  73  * Check if this is a blktrace binary data file. We read a single trace
  74  * into memory and check for the magic signature.
  75  */
  76 int is_blktrace(const char *filename, int *need_swap)
  77 {
  78         struct blk_io_trace t;
  79         int fd, ret;
  80
  81         fd = open(filename, O_RDONLY);
  82         if (fd < 0)
  83                 return 0;
  84
  85         ret = read(fd, &t, sizeof(t));
  86         close(fd);
  87
  88         if (ret < 0) {
  89                 perror("read blktrace");
  90                 return 0;
  91         } else if (ret != sizeof(t)) {
  92                 log_err("fio: short read on blktrace file\n");
  93                 return 0;
  94         }
  95
  96         if ((t.magic & 0xffffff00) == BLK_IO_TRACE_MAGIC) {
  97                 *need_swap = 0;
  98                 return 1;
  99         }
 100
 101         /*
 102          * Maybe it needs to be endian swapped...
 103          */
 104         t.magic = fio_swap32(t.magic);
 105         if ((t.magic & 0xffffff00) == BLK_IO_TRACE_MAGIC) {
 106                 *need_swap = 1;
 107                 return 1;
 108         }
 109
 110         return 0;
 111 }
 112
 113 #define FMINORBITS      20
 114 #define FMINORMASK      ((1U << FMINORBITS) - 1)
 115 #define FMAJOR(dev)     ((unsigned int) ((dev) >> FMINORBITS))
 116 #define FMINOR(dev)     ((unsigned int) ((dev) & FMINORMASK))
 117
 118 static void trace_add_open_close_event(struct thread_data *td, int fileno, enum file_log_act action)
 119 {
 120         struct io_piece *ipo;
 121
 122         ipo = calloc(1, sizeof(*ipo));
 123         init_ipo(ipo);
 124
 125         ipo->ddir = DDIR_INVAL;
 126         ipo->fileno = fileno;
 127         ipo->file_action = action;
 128         flist_add_tail(&ipo->list, &td->io_log_list);
 129 }
 130
 131 static int get_dev_blocksize(const char *dev, unsigned int *bs)
 132 {
 133         int fd;
 134
 135         fd = open(dev, O_RDONLY);
 136         if (fd < 0)
 137                 return 1;
 138
 139         if (ioctl(fd, BLKSSZGET, bs) < 0) {
 140                 close(fd);
 141                 return 1;
 142         }
 143
 144         close(fd);
 145         return 0;
 146 }
 147
 148 static int trace_add_file(struct thread_data *td, __u32 device,
 149                           unsigned int *bs)
 150 {
 151         static unsigned int last_maj, last_min, last_fileno, last_bs;
 152         unsigned int maj = FMAJOR(device);
 153         unsigned int min = FMINOR(device);
 154         struct fio_file *f;
 155         unsigned int i;
 156         char dev[256];
 157
 158         if (last_maj == maj && last_min == min) {
 159                 *bs = last_bs;
 160                 return last_fileno;
 161         }
 162
 163         last_maj = maj;
 164         last_min = min;
 165
 166         /*
 167          * check for this file in our list
 168          */
 169         for_each_file(td, f, i) {
 170                 if (f->major == maj && f->minor == min) {
 171                         last_fileno = f->fileno;
 172                         last_bs = f->bs;
 173                         goto out;
 174                 }
 175         }
 176
 177         strcpy(dev, "/dev");
 178         if (blktrace_lookup_device(td->o.replay_redirect, dev, maj, min)) {
 179                 unsigned int this_bs;
 180                 int fileno;
 181
 182                 if (td->o.replay_redirect)
 183                         dprint(FD_BLKTRACE, "device lookup: %d/%d\n overridden"
 184                                         " with: %s\n", maj, min,
 185                                         td->o.replay_redirect);
 186                 else
 187                         dprint(FD_BLKTRACE, "device lookup: %d/%d\n", maj, min);
 188
 189                 dprint(FD_BLKTRACE, "add devices %s\n", dev);
 190                 fileno = add_file_exclusive(td, dev);
 191
 192                 if (get_dev_blocksize(dev, &this_bs))
 193                         this_bs = 512;
 194
 195                 td->o.open_files++;
 196                 td->files[fileno]->major = maj;
 197                 td->files[fileno]->minor = min;
 198                 td->files[fileno]->bs = this_bs;
 199                 trace_add_open_close_event(td, fileno, FIO_LOG_OPEN_FILE);
 200
 201                 last_fileno = fileno;
 202                 last_bs = this_bs;
 203         }
 204
 205 out:
 206         *bs = last_bs;
 207         return last_fileno;
 208 }
 209
 210 static void t_bytes_align(struct thread_options *o, struct blk_io_trace *t)
 211 {
 212         if (!o->replay_align)
 213                 return;
 214
 215         t->bytes = (t->bytes + o->replay_align - 1) & ~(o->replay_align - 1);
 216 }
 217
 218 /*
 219  * Store blk_io_trace data in an ipo for later retrieval.
 220  */
 221 static void store_ipo(struct thread_data *td, unsigned long long offset,
 222                       unsigned int bytes, int rw, unsigned long long ttime,
 223                       int fileno, unsigned int bs)
 224 {
 225         struct io_piece *ipo;
 226
 227         ipo = calloc(1, sizeof(*ipo));
 228         init_ipo(ipo);
 229
 230         ipo->offset = offset * bs;
 231         if (td->o.replay_scale)
 232                 ipo->offset = ipo->offset / td->o.replay_scale;
 233         ipo_bytes_align(td->o.replay_align, ipo);
 234         ipo->len = bytes;
 235         ipo->delay = ttime / 1000;
 236         if (rw)
 237                 ipo->ddir = DDIR_WRITE;
 238         else
 239                 ipo->ddir = DDIR_READ;
 240         ipo->fileno = fileno;
 241
 242         dprint(FD_BLKTRACE, "store ddir=%d, off=%llu, len=%lu, delay=%lu\n",
 243                                                         ipo->ddir, ipo->offset,
 244                                                         ipo->len, ipo->delay);
 245         queue_io_piece(td, ipo);
 246 }
 247
 248 static void handle_trace_notify(struct blk_io_trace *t)
 249 {
 250         switch (t->action) {
 251         case BLK_TN_PROCESS:
 252                 dprint(FD_BLKTRACE, "got process notify: %x, %d\n",
 253                                 t->action, t->pid);
 254                 break;
 255         case BLK_TN_TIMESTAMP:
 256                 dprint(FD_BLKTRACE, "got timestamp notify: %x, %d\n",
 257                                 t->action, t->pid);
 258                 break;
 259         case BLK_TN_MESSAGE:
 260                 break;
 261         default:
 262                 dprint(FD_BLKTRACE, "unknown trace act %x\n", t->action);
 263                 break;
 264         }
 265 }
 266
 267 static void handle_trace_discard(struct thread_data *td,
 268                                  struct blk_io_trace *t,
 269                                  unsigned long long ttime,
 270                                  unsigned long *ios, unsigned int *rw_bs)
 271 {
 272         struct io_piece *ipo;
 273         unsigned int bs;
 274         int fileno;
 275
 276         if (td->o.replay_skip & (1u << DDIR_TRIM))
 277                 return;
 278
 279         ipo = calloc(1, sizeof(*ipo));
 280         init_ipo(ipo);
 281         fileno = trace_add_file(td, t->device, &bs);
 282
 283         ios[DDIR_TRIM]++;
 284         if (t->bytes > rw_bs[DDIR_TRIM])
 285                 rw_bs[DDIR_TRIM] = t->bytes;
 286
 287         td->o.size += t->bytes;
 288
 289         INIT_FLIST_HEAD(&ipo->list);
 290
 291         ipo->offset = t->sector * bs;
 292         if (td->o.replay_scale)
 293                 ipo->offset = ipo->offset / td->o.replay_scale;
 294         ipo_bytes_align(td->o.replay_align, ipo);
 295         ipo->len = t->bytes;
 296         ipo->delay = ttime / 1000;
 297         ipo->ddir = DDIR_TRIM;
 298         ipo->fileno = fileno;
 299
 300         dprint(FD_BLKTRACE, "store discard, off=%llu, len=%lu, delay=%lu\n",
 301                                                         ipo->offset, ipo->len,
 302                                                         ipo->delay);
 303         queue_io_piece(td, ipo);
 304 }
 305
 306 static void handle_trace_fs(struct thread_data *td, struct blk_io_trace *t,
 307                             unsigned long long ttime, unsigned long *ios,
 308                             unsigned int *rw_bs)
 309 {
 310         unsigned int bs;
 311         int rw;
 312         int fileno;
 313
 314         fileno = trace_add_file(td, t->device, &bs);
 315
 316         rw = (t->action & BLK_TC_ACT(BLK_TC_WRITE)) != 0;
 317
 318         if (rw) {
 319                 if (td->o.replay_skip & (1u << DDIR_WRITE))
 320                         return;
 321         } else {
 322                 if (td->o.replay_skip & (1u << DDIR_READ))
 323                         return;
 324         }
 325
 326         assert(t->bytes);
 327
 328         if (t->bytes > rw_bs[rw])
 329                 rw_bs[rw] = t->bytes;
 330
 331         ios[rw]++;
 332         td->o.size += t->bytes;
 333         store_ipo(td, t->sector, t->bytes, rw, ttime, fileno, bs);
 334 }
 335
 336 static void handle_trace_flush(struct thread_data *td, struct blk_io_trace *t,
 337                                unsigned long long ttime)
 338 {
 339         struct io_piece *ipo;
 340         unsigned int bs;
 341         int fileno;
 342
 343         if (td->o.replay_skip & (1u << DDIR_SYNC))
 344                 return;
 345
 346         ipo = calloc(1, sizeof(*ipo));
 347         init_ipo(ipo);
 348         fileno = trace_add_file(td, t->device, &bs);
 349
 350         ipo->delay = ttime / 1000;
 351         ipo->ddir = DDIR_SYNC;
 352         ipo->fileno = fileno;
 353
 354         dprint(FD_BLKTRACE, "store flush delay=%lu\n", ipo->delay);
 355         queue_io_piece(td, ipo);
 356 }
 357
 358 /*
 359  * We only care for queue traces, most of the others are side effects
 360  * due to internal workings of the block layer.
 361  */
 362 static void handle_trace(struct thread_data *td, struct blk_io_trace *t,
 363                          unsigned long *ios, unsigned int *bs)
 364 {
 365         static unsigned long long last_ttime;
 366         unsigned long long delay = 0;
 367
 368         if ((t->action & 0xffff) != __BLK_TA_QUEUE)
 369                 return;
 370
 371         if (!(t->action & BLK_TC_ACT(BLK_TC_NOTIFY))) {
 372                 if (!last_ttime || td->o.no_stall)
 373                         delay = 0;
 374                 else if (td->o.replay_time_scale == 100)
 375                         delay = t->time - last_ttime;
 376                 else {
 377                         double tmp = t->time - last_ttime;
 378                         double scale;
 379
 380                         scale = (double) 100.0 / (double) td->o.replay_time_scale;
 381                         tmp *= scale;
 382                         delay = tmp;
 383                 }
 384                 last_ttime = t->time;
 385         }
 386
 387         t_bytes_align(&td->o, t);
 388
 389         if (t->action & BLK_TC_ACT(BLK_TC_NOTIFY))
 390                 handle_trace_notify(t);
 391         else if (t->action & BLK_TC_ACT(BLK_TC_DISCARD))
 392                 handle_trace_discard(td, t, delay, ios, bs);
 393         else if (t->action & BLK_TC_ACT(BLK_TC_FLUSH))
 394                 handle_trace_flush(td, t, delay);
 395         else
 396                 handle_trace_fs(td, t, delay, ios, bs);
 397 }
 398
 399 static void byteswap_trace(struct blk_io_trace *t)
 400 {
 401         t->magic = fio_swap32(t->magic);
 402         t->sequence = fio_swap32(t->sequence);
 403         t->time = fio_swap64(t->time);
 404         t->sector = fio_swap64(t->sector);
 405         t->bytes = fio_swap32(t->bytes);
 406         t->action = fio_swap32(t->action);
 407         t->pid = fio_swap32(t->pid);
 408         t->device = fio_swap32(t->device);
 409         t->cpu = fio_swap32(t->cpu);
 410         t->error = fio_swap16(t->error);
 411         t->pdu_len = fio_swap16(t->pdu_len);
 412 }
 413
 414 static int t_is_write(struct blk_io_trace *t)
 415 {
 416         return (t->action & BLK_TC_ACT(BLK_TC_WRITE | BLK_TC_DISCARD)) != 0;
 417 }
 418
 419 static enum fio_ddir t_get_ddir(struct blk_io_trace *t)
 420 {
 421         if (t->action & BLK_TC_ACT(BLK_TC_READ))
 422                 return DDIR_READ;
 423         else if (t->action & BLK_TC_ACT(BLK_TC_WRITE))
 424                 return DDIR_WRITE;
 425         else if (t->action & BLK_TC_ACT(BLK_TC_DISCARD))
 426                 return DDIR_TRIM;
 427
 428         return DDIR_INVAL;
 429 }
 430
 431 static void depth_inc(struct blk_io_trace *t, int *depth)
 432 {
 433         enum fio_ddir ddir;
 434
 435         ddir = t_get_ddir(t);
 436         if (ddir != DDIR_INVAL)
 437                 depth[ddir]++;
 438 }
 439
 440 static void depth_dec(struct blk_io_trace *t, int *depth)
 441 {
 442         enum fio_ddir ddir;
 443
 444         ddir = t_get_ddir(t);
 445         if (ddir != DDIR_INVAL)
 446                 depth[ddir]--;
 447 }
 448
 449 static void depth_end(struct blk_io_trace *t, int *this_depth, int *depth)
 450 {
 451         enum fio_ddir ddir = DDIR_INVAL;
 452
 453         ddir = t_get_ddir(t);
 454         if (ddir != DDIR_INVAL) {
 455                 depth[ddir] = max(depth[ddir], this_depth[ddir]);
 456                 this_depth[ddir] = 0;
 457         }
 458 }
 459
 460 /*
 461  * Load a blktrace file by reading all the blk_io_trace entries, and storing
 462  * them as io_pieces like the fio text version would do.
 463  */
 464 int load_blktrace(struct thread_data *td, const char *filename, int need_swap)
 465 {
 466         struct blk_io_trace t;
 467         unsigned long ios[DDIR_RWDIR_CNT], skipped_writes;
 468         unsigned int rw_bs[DDIR_RWDIR_CNT];
 469         struct fifo *fifo;
 470         int fd, i, old_state;
 471         struct fio_file *f;
 472         int this_depth[DDIR_RWDIR_CNT], depth[DDIR_RWDIR_CNT], max_depth;
 473
 474         fd = open(filename, O_RDONLY);
 475         if (fd < 0) {
 476                 td_verror(td, errno, "open blktrace file");
 477                 return 1;
 478         }
 479
 480         fifo = fifo_alloc(TRACE_FIFO_SIZE);
 481
 482         old_state = td_bump_runstate(td, TD_SETTING_UP);
 483
 484         td->o.size = 0;
 485
 486         for (i = 0; i < DDIR_RWDIR_CNT; i++) {
 487                 ios[i] = 0;
 488                 rw_bs[i] = 0;
 489                 this_depth[i] = 0;
 490                 depth[i] = 0;
 491         }
 492
 493         skipped_writes = 0;
 494         do {
 495                 int ret = trace_fifo_get(td, fifo, fd, &t, sizeof(t));
 496
 497                 if (ret < 0)
 498                         goto err;
 499                 else if (!ret)
 500                         break;
 501                 else if (ret < (int) sizeof(t)) {
 502                         log_err("fio: short fifo get\n");
 503                         break;
 504                 }
 505
 506                 if (need_swap)
 507                         byteswap_trace(&t);
 508
 509                 if ((t.magic & 0xffffff00) != BLK_IO_TRACE_MAGIC) {
 510                         log_err("fio: bad magic in blktrace data: %x\n",
 511                                                                 t.magic);
 512                         goto err;
 513                 }
 514                 if ((t.magic & 0xff) != BLK_IO_TRACE_VERSION) {
 515                         log_err("fio: bad blktrace version %d\n",
 516                                                                 t.magic & 0xff);
 517                         goto err;
 518                 }
 519                 ret = discard_pdu(td, fifo, fd, &t);
 520                 if (ret < 0) {
 521                         td_verror(td, ret, "blktrace lseek");
 522                         goto err;
 523                 } else if (t.pdu_len != ret) {
 524                         log_err("fio: discarded %d of %d\n", ret, t.pdu_len);
 525                         goto err;
 526                 }
 527                 if ((t.action & BLK_TC_ACT(BLK_TC_NOTIFY)) == 0) {
 528                         if ((t.action & 0xffff) == __BLK_TA_QUEUE)
 529                                 depth_inc(&t, this_depth);
 530                         else if (((t.action & 0xffff) == __BLK_TA_BACKMERGE) ||
 531                                 ((t.action & 0xffff) == __BLK_TA_FRONTMERGE))
 532                                 depth_dec(&t, this_depth);
 533                         else if ((t.action & 0xffff) == __BLK_TA_COMPLETE)
 534                                 depth_end(&t, this_depth, depth);
 535
 536                         if (t_is_write(&t) && read_only) {
 537                                 skipped_writes++;
 538                                 continue;
 539                         }
 540                 }
 541
 542                 handle_trace(td, &t, ios, rw_bs);
 543         } while (1);
 544
 545         for_each_file(td, f, i)
 546                 trace_add_open_close_event(td, f->fileno, FIO_LOG_CLOSE_FILE);
 547
 548         fifo_free(fifo);
 549         close(fd);
 550
 551         td_restore_runstate(td, old_state);
 552
 553         if (!td->files_index) {
 554                 log_err("fio: did not find replay device(s)\n");
 555                 return 1;
 556         }
 557
 558         /*
 559          * For stacked devices, we don't always get a COMPLETE event so
 560          * the depth grows to insane values. Limit it to something sane(r).
 561          */
 562         max_depth = 0;
 563         for (i = 0; i < DDIR_RWDIR_CNT; i++) {
 564                 if (depth[i] > 1024)
 565                         depth[i] = 1024;
 566                 else if (!depth[i] && ios[i])
 567                         depth[i] = 1;
 568                 max_depth = max(depth[i], max_depth);
 569         }
 570
 571         if (skipped_writes)
 572                 log_err("fio: %s skips replay of %lu writes due to read-only\n",
 573                                                 td->o.name, skipped_writes);
 574
 575         if (!ios[DDIR_READ] && !ios[DDIR_WRITE]) {
 576                 log_err("fio: found no ios in blktrace data\n");
 577                 return 1;
 578         } else if (ios[DDIR_READ] && !ios[DDIR_WRITE]) {
 579                 td->o.td_ddir = TD_DDIR_READ;
 580                 td->o.max_bs[DDIR_READ] = rw_bs[DDIR_READ];
 581         } else if (!ios[DDIR_READ] && ios[DDIR_WRITE]) {
 582                 td->o.td_ddir = TD_DDIR_WRITE;
 583                 td->o.max_bs[DDIR_WRITE] = rw_bs[DDIR_WRITE];
 584         } else {
 585                 td->o.td_ddir = TD_DDIR_RW;
 586                 td->o.max_bs[DDIR_READ] = rw_bs[DDIR_READ];
 587                 td->o.max_bs[DDIR_WRITE] = rw_bs[DDIR_WRITE];
 588                 td->o.max_bs[DDIR_TRIM] = rw_bs[DDIR_TRIM];
 589         }
 590
 591         /*
 592          * We need to do direct/raw ios to the device, to avoid getting
 593          * read-ahead in our way. But only do so if the minimum block size
 594          * is a multiple of 4k, otherwise we don't know if it's safe to do so.
 595          */
 596         if (!fio_option_is_set(&td->o, odirect) && !(td_min_bs(td) & 4095))
 597                 td->o.odirect = 1;
 598
 599         /*
 600          * If depth wasn't manually set, use probed depth
 601          */
 602         if (!fio_option_is_set(&td->o, iodepth))
 603                 td->o.iodepth = td->o.iodepth_low = max_depth;
 604
 605         return 0;
 606 err:
 607         close(fd);
 608         fifo_free(fifo);
 609         return 1;
 610 }