| 1 | /* |
| 2 | * Small tool to check for dedupable blocks in a file or device. Basically |
| 3 | * just scans the filename for extents of the given size, checksums them, |
| 4 | * and orders them up. |
| 5 | */ |
| 6 | #include <fcntl.h> |
| 7 | #include <inttypes.h> |
| 8 | #include <stdio.h> |
| 9 | #include <string.h> |
| 10 | #include <unistd.h> |
| 11 | #include <sys/stat.h> |
| 12 | |
| 13 | #include "../fio.h" |
| 14 | #include "../flist.h" |
| 15 | #include "../log.h" |
| 16 | #include "../fio_sem.h" |
| 17 | #include "../smalloc.h" |
| 18 | #include "../minmax.h" |
| 19 | #include "../crc/md5.h" |
| 20 | #include "../os/os.h" |
| 21 | #include "../gettime.h" |
| 22 | #include "../fio_time.h" |
| 23 | #include "../lib/rbtree.h" |
| 24 | |
| 25 | #include "../lib/bloom.h" |
| 26 | #include "debug.h" |
| 27 | #include "zlib.h" |
| 28 | |
| 29 | struct zlib_ctrl { |
| 30 | z_stream stream; |
| 31 | unsigned char *buf_in; |
| 32 | unsigned char *buf_out; |
| 33 | }; |
| 34 | |
| 35 | struct worker_thread { |
| 36 | struct zlib_ctrl zc; |
| 37 | pthread_t thread; |
| 38 | uint64_t cur_offset; |
| 39 | uint64_t size; |
| 40 | unsigned long long unique_capacity; |
| 41 | unsigned long items; |
| 42 | unsigned long dupes; |
| 43 | int err; |
| 44 | int fd; |
| 45 | volatile int done; |
| 46 | }; |
| 47 | |
| 48 | struct extent { |
| 49 | struct flist_head list; |
| 50 | uint64_t offset; |
| 51 | }; |
| 52 | |
| 53 | struct chunk { |
| 54 | struct fio_rb_node rb_node; |
| 55 | uint64_t count; |
| 56 | uint32_t hash[MD5_HASH_WORDS]; |
| 57 | struct flist_head extent_list[0]; |
| 58 | }; |
| 59 | |
| 60 | struct item { |
| 61 | uint64_t offset; |
| 62 | uint32_t hash[MD5_HASH_WORDS]; |
| 63 | }; |
| 64 | |
| 65 | static struct rb_root rb_root; |
| 66 | static struct bloom *bloom; |
| 67 | static struct fio_sem *rb_lock; |
| 68 | |
| 69 | static unsigned int blocksize = 4096; |
| 70 | static unsigned int num_threads; |
| 71 | static unsigned int chunk_size = 1048576; |
| 72 | static unsigned int dump_output; |
| 73 | static unsigned int odirect; |
| 74 | static unsigned int collision_check; |
| 75 | static unsigned int print_progress = 1; |
| 76 | static unsigned int use_bloom = 1; |
| 77 | static unsigned int compression = 0; |
| 78 | |
| 79 | static uint64_t total_size; |
| 80 | static uint64_t cur_offset; |
| 81 | static struct fio_sem *size_lock; |
| 82 | |
| 83 | static struct fio_file file; |
| 84 | |
| 85 | static uint64_t get_size(struct fio_file *f, struct stat *sb) |
| 86 | { |
| 87 | uint64_t ret; |
| 88 | |
| 89 | if (S_ISBLK(sb->st_mode)) { |
| 90 | unsigned long long bytes = 0; |
| 91 | |
| 92 | if (blockdev_size(f, &bytes)) { |
| 93 | log_err("dedupe: failed getting bdev size\n"); |
| 94 | return 0; |
| 95 | } |
| 96 | ret = bytes; |
| 97 | } else { |
| 98 | ret = sb->st_size; |
| 99 | } |
| 100 | |
| 101 | return (ret & ~((uint64_t)blocksize - 1)); |
| 102 | } |
| 103 | |
| 104 | static int get_work(uint64_t *offset, uint64_t *size) |
| 105 | { |
| 106 | uint64_t this_chunk; |
| 107 | int ret = 1; |
| 108 | |
| 109 | fio_sem_down(size_lock); |
| 110 | |
| 111 | if (cur_offset < total_size) { |
| 112 | *offset = cur_offset; |
| 113 | this_chunk = min((uint64_t)chunk_size, total_size - cur_offset); |
| 114 | *size = this_chunk; |
| 115 | cur_offset += this_chunk; |
| 116 | ret = 0; |
| 117 | } |
| 118 | |
| 119 | fio_sem_up(size_lock); |
| 120 | return ret; |
| 121 | } |
| 122 | |
| 123 | static int __read_block(int fd, void *buf, off_t offset, size_t count) |
| 124 | { |
| 125 | ssize_t ret; |
| 126 | |
| 127 | ret = pread(fd, buf, count, offset); |
| 128 | if (ret < 0) { |
| 129 | perror("pread"); |
| 130 | return 1; |
| 131 | } else if (!ret) { |
| 132 | return 1; |
| 133 | } else if (ret != count) { |
| 134 | log_err("dedupe: short read on block\n"); |
| 135 | return 1; |
| 136 | } |
| 137 | |
| 138 | return 0; |
| 139 | } |
| 140 | |
| 141 | static int read_block(int fd, void *buf, off_t offset) |
| 142 | { |
| 143 | return __read_block(fd, buf, offset, blocksize); |
| 144 | } |
| 145 | |
| 146 | static int account_unique_capacity(uint64_t offset, uint64_t *unique_capacity, |
| 147 | struct zlib_ctrl *zc) |
| 148 | { |
| 149 | z_stream *stream = &zc->stream; |
| 150 | unsigned int compressed_len; |
| 151 | int ret; |
| 152 | |
| 153 | if (read_block(file.fd, zc->buf_in, offset)) |
| 154 | return 1; |
| 155 | |
| 156 | stream->next_in = zc->buf_in; |
| 157 | stream->avail_in = blocksize; |
| 158 | stream->avail_out = deflateBound(stream, blocksize); |
| 159 | stream->next_out = zc->buf_out; |
| 160 | |
| 161 | ret = deflate(stream, Z_FINISH); |
| 162 | if (ret == Z_STREAM_ERROR) |
| 163 | return 1; |
| 164 | compressed_len = blocksize - stream->avail_out; |
| 165 | |
| 166 | if (dump_output) |
| 167 | printf("offset 0x%lx compressed to %d blocksize %d ratio %.2f \n", |
| 168 | (unsigned long) offset, compressed_len, blocksize, |
| 169 | (float)compressed_len / (float)blocksize); |
| 170 | |
| 171 | *unique_capacity += compressed_len; |
| 172 | deflateReset(stream); |
| 173 | return 0; |
| 174 | } |
| 175 | |
| 176 | static void add_item(struct chunk *c, struct item *i) |
| 177 | { |
| 178 | /* |
| 179 | * Save some memory and don't add extent items, if we don't |
| 180 | * use them. |
| 181 | */ |
| 182 | if (dump_output || collision_check) { |
| 183 | struct extent *e; |
| 184 | |
| 185 | e = malloc(sizeof(*e)); |
| 186 | e->offset = i->offset; |
| 187 | flist_add_tail(&e->list, &c->extent_list[0]); |
| 188 | } |
| 189 | |
| 190 | c->count++; |
| 191 | } |
| 192 | |
| 193 | static int col_check(struct chunk *c, struct item *i) |
| 194 | { |
| 195 | struct extent *e; |
| 196 | char *cbuf, *ibuf; |
| 197 | int ret = 1; |
| 198 | |
| 199 | cbuf = fio_memalign(blocksize, blocksize, false); |
| 200 | ibuf = fio_memalign(blocksize, blocksize, false); |
| 201 | |
| 202 | e = flist_entry(c->extent_list[0].next, struct extent, list); |
| 203 | if (read_block(file.fd, cbuf, e->offset)) |
| 204 | goto out; |
| 205 | |
| 206 | if (read_block(file.fd, ibuf, i->offset)) |
| 207 | goto out; |
| 208 | |
| 209 | ret = memcmp(ibuf, cbuf, blocksize); |
| 210 | out: |
| 211 | fio_memfree(cbuf, blocksize, false); |
| 212 | fio_memfree(ibuf, blocksize, false); |
| 213 | return ret; |
| 214 | } |
| 215 | |
| 216 | static struct chunk *alloc_chunk(void) |
| 217 | { |
| 218 | struct chunk *c; |
| 219 | |
| 220 | if (collision_check || dump_output) { |
| 221 | c = malloc(sizeof(struct chunk) + sizeof(struct flist_head)); |
| 222 | INIT_FLIST_HEAD(&c->extent_list[0]); |
| 223 | } else { |
| 224 | c = malloc(sizeof(struct chunk)); |
| 225 | } |
| 226 | |
| 227 | return c; |
| 228 | } |
| 229 | |
| 230 | static int insert_chunk(struct item *i, uint64_t *unique_capacity, |
| 231 | struct zlib_ctrl *zc) |
| 232 | { |
| 233 | struct fio_rb_node **p, *parent; |
| 234 | struct chunk *c; |
| 235 | int ret, diff; |
| 236 | |
| 237 | p = &rb_root.rb_node; |
| 238 | parent = NULL; |
| 239 | while (*p) { |
| 240 | parent = *p; |
| 241 | |
| 242 | c = rb_entry(parent, struct chunk, rb_node); |
| 243 | diff = memcmp(i->hash, c->hash, sizeof(i->hash)); |
| 244 | if (diff < 0) { |
| 245 | p = &(*p)->rb_left; |
| 246 | } else if (diff > 0) { |
| 247 | p = &(*p)->rb_right; |
| 248 | } else { |
| 249 | if (!collision_check) |
| 250 | goto add; |
| 251 | |
| 252 | fio_sem_up(rb_lock); |
| 253 | ret = col_check(c, i); |
| 254 | fio_sem_down(rb_lock); |
| 255 | |
| 256 | if (!ret) |
| 257 | goto add; |
| 258 | |
| 259 | p = &(*p)->rb_right; |
| 260 | } |
| 261 | } |
| 262 | |
| 263 | c = alloc_chunk(); |
| 264 | RB_CLEAR_NODE(&c->rb_node); |
| 265 | c->count = 0; |
| 266 | memcpy(c->hash, i->hash, sizeof(i->hash)); |
| 267 | rb_link_node(&c->rb_node, parent, p); |
| 268 | rb_insert_color(&c->rb_node, &rb_root); |
| 269 | if (compression) { |
| 270 | ret = account_unique_capacity(i->offset, unique_capacity, zc); |
| 271 | if (ret) |
| 272 | return ret; |
| 273 | } |
| 274 | add: |
| 275 | add_item(c, i); |
| 276 | return 0; |
| 277 | } |
| 278 | |
| 279 | static int insert_chunks(struct item *items, unsigned int nitems, |
| 280 | uint64_t *ndupes, uint64_t *unique_capacity, |
| 281 | struct zlib_ctrl *zc) |
| 282 | { |
| 283 | int i, ret = 0; |
| 284 | |
| 285 | fio_sem_down(rb_lock); |
| 286 | |
| 287 | for (i = 0; i < nitems; i++) { |
| 288 | if (bloom) { |
| 289 | unsigned int s; |
| 290 | int r; |
| 291 | |
| 292 | s = sizeof(items[i].hash) / sizeof(uint32_t); |
| 293 | r = bloom_set(bloom, items[i].hash, s); |
| 294 | *ndupes += r; |
| 295 | } else { |
| 296 | ret = insert_chunk(&items[i], unique_capacity, zc); |
| 297 | if (ret) |
| 298 | break; |
| 299 | } |
| 300 | } |
| 301 | |
| 302 | fio_sem_up(rb_lock); |
| 303 | return ret; |
| 304 | } |
| 305 | |
| 306 | static void crc_buf(void *buf, uint32_t *hash) |
| 307 | { |
| 308 | struct fio_md5_ctx ctx = { .hash = hash }; |
| 309 | |
| 310 | fio_md5_init(&ctx); |
| 311 | fio_md5_update(&ctx, buf, blocksize); |
| 312 | fio_md5_final(&ctx); |
| 313 | } |
| 314 | |
| 315 | static unsigned int read_blocks(int fd, void *buf, off_t offset, size_t size) |
| 316 | { |
| 317 | if (__read_block(fd, buf, offset, size)) |
| 318 | return 0; |
| 319 | |
| 320 | return size / blocksize; |
| 321 | } |
| 322 | |
| 323 | static int do_work(struct worker_thread *thread, void *buf) |
| 324 | { |
| 325 | unsigned int nblocks, i; |
| 326 | off_t offset; |
| 327 | int nitems = 0; |
| 328 | uint64_t ndupes = 0; |
| 329 | uint64_t unique_capacity = 0; |
| 330 | struct item *items; |
| 331 | int ret; |
| 332 | |
| 333 | offset = thread->cur_offset; |
| 334 | |
| 335 | nblocks = read_blocks(thread->fd, buf, offset, |
| 336 | min(thread->size, (uint64_t) chunk_size)); |
| 337 | if (!nblocks) |
| 338 | return 1; |
| 339 | |
| 340 | items = malloc(sizeof(*items) * nblocks); |
| 341 | |
| 342 | for (i = 0; i < nblocks; i++) { |
| 343 | void *thisptr = buf + (i * blocksize); |
| 344 | |
| 345 | items[i].offset = offset; |
| 346 | crc_buf(thisptr, items[i].hash); |
| 347 | offset += blocksize; |
| 348 | nitems++; |
| 349 | } |
| 350 | |
| 351 | ret = insert_chunks(items, nitems, &ndupes, &unique_capacity, &thread->zc); |
| 352 | |
| 353 | free(items); |
| 354 | if (!ret) { |
| 355 | thread->items += nitems; |
| 356 | thread->dupes += ndupes; |
| 357 | thread->unique_capacity += unique_capacity; |
| 358 | return 0; |
| 359 | } |
| 360 | |
| 361 | return ret; |
| 362 | } |
| 363 | |
| 364 | static void thread_init_zlib_control(struct worker_thread *thread) |
| 365 | { |
| 366 | size_t sz; |
| 367 | |
| 368 | z_stream *stream = &thread->zc.stream; |
| 369 | stream->zalloc = Z_NULL; |
| 370 | stream->zfree = Z_NULL; |
| 371 | stream->opaque = Z_NULL; |
| 372 | |
| 373 | if (deflateInit(stream, Z_DEFAULT_COMPRESSION) != Z_OK) |
| 374 | return; |
| 375 | |
| 376 | thread->zc.buf_in = fio_memalign(blocksize, blocksize, false); |
| 377 | sz = deflateBound(stream, blocksize); |
| 378 | thread->zc.buf_out = fio_memalign(blocksize, sz, false); |
| 379 | } |
| 380 | |
| 381 | static void *thread_fn(void *data) |
| 382 | { |
| 383 | struct worker_thread *thread = data; |
| 384 | void *buf; |
| 385 | |
| 386 | buf = fio_memalign(blocksize, chunk_size, false); |
| 387 | thread_init_zlib_control(thread); |
| 388 | |
| 389 | do { |
| 390 | if (get_work(&thread->cur_offset, &thread->size)) { |
| 391 | thread->err = 1; |
| 392 | break; |
| 393 | } |
| 394 | if (do_work(thread, buf)) { |
| 395 | thread->err = 1; |
| 396 | break; |
| 397 | } |
| 398 | } while (1); |
| 399 | |
| 400 | thread->done = 1; |
| 401 | fio_memfree(buf, chunk_size, false); |
| 402 | return NULL; |
| 403 | } |
| 404 | |
| 405 | static void show_progress(struct worker_thread *threads, unsigned long total) |
| 406 | { |
| 407 | unsigned long last_nitems = 0; |
| 408 | struct timespec last_tv; |
| 409 | |
| 410 | fio_gettime(&last_tv, NULL); |
| 411 | |
| 412 | while (print_progress) { |
| 413 | unsigned long this_items; |
| 414 | unsigned long nitems = 0; |
| 415 | uint64_t tdiff; |
| 416 | float perc; |
| 417 | int some_done = 0; |
| 418 | int i; |
| 419 | |
| 420 | for (i = 0; i < num_threads; i++) { |
| 421 | nitems += threads[i].items; |
| 422 | some_done = threads[i].done; |
| 423 | if (some_done) |
| 424 | break; |
| 425 | } |
| 426 | |
| 427 | if (some_done) |
| 428 | break; |
| 429 | |
| 430 | perc = (float) nitems / (float) total; |
| 431 | perc *= 100.0; |
| 432 | this_items = nitems - last_nitems; |
| 433 | this_items *= blocksize; |
| 434 | tdiff = mtime_since_now(&last_tv); |
| 435 | if (tdiff) { |
| 436 | this_items = (this_items * 1000) / (tdiff * 1024); |
| 437 | printf("%3.2f%% done (%luKiB/sec)\r", perc, this_items); |
| 438 | last_nitems = nitems; |
| 439 | fio_gettime(&last_tv, NULL); |
| 440 | } else { |
| 441 | printf("%3.2f%% done\r", perc); |
| 442 | } |
| 443 | fflush(stdout); |
| 444 | usleep(250000); |
| 445 | }; |
| 446 | } |
| 447 | |
| 448 | static int run_dedupe_threads(struct fio_file *f, uint64_t dev_size, |
| 449 | uint64_t *nextents, uint64_t *nchunks, |
| 450 | uint64_t *unique_capacity) |
| 451 | { |
| 452 | struct worker_thread *threads; |
| 453 | unsigned long nitems, total_items; |
| 454 | int i, err = 0; |
| 455 | |
| 456 | total_size = dev_size; |
| 457 | total_items = dev_size / blocksize; |
| 458 | cur_offset = 0; |
| 459 | size_lock = fio_sem_init(FIO_SEM_UNLOCKED); |
| 460 | |
| 461 | threads = malloc(num_threads * sizeof(struct worker_thread)); |
| 462 | for (i = 0; i < num_threads; i++) { |
| 463 | memset(&threads[i], 0, sizeof(struct worker_thread)); |
| 464 | threads[i].fd = f->fd; |
| 465 | |
| 466 | err = pthread_create(&threads[i].thread, NULL, thread_fn, &threads[i]); |
| 467 | if (err) { |
| 468 | log_err("fio: thread startup failed\n"); |
| 469 | break; |
| 470 | } |
| 471 | } |
| 472 | |
| 473 | show_progress(threads, total_items); |
| 474 | |
| 475 | nitems = 0; |
| 476 | *nextents = 0; |
| 477 | *nchunks = 1; |
| 478 | *unique_capacity = 0; |
| 479 | for (i = 0; i < num_threads; i++) { |
| 480 | void *ret; |
| 481 | pthread_join(threads[i].thread, &ret); |
| 482 | nitems += threads[i].items; |
| 483 | *nchunks += threads[i].dupes; |
| 484 | *unique_capacity += threads[i].unique_capacity; |
| 485 | } |
| 486 | |
| 487 | printf("Threads(%u): %lu items processed\n", num_threads, nitems); |
| 488 | |
| 489 | *nextents = nitems; |
| 490 | *nchunks = nitems - *nchunks; |
| 491 | |
| 492 | fio_sem_remove(size_lock); |
| 493 | free(threads); |
| 494 | return err; |
| 495 | } |
| 496 | |
| 497 | static int dedupe_check(const char *filename, uint64_t *nextents, |
| 498 | uint64_t *nchunks, uint64_t *unique_capacity) |
| 499 | { |
| 500 | uint64_t dev_size; |
| 501 | struct stat sb; |
| 502 | int flags; |
| 503 | |
| 504 | flags = O_RDONLY; |
| 505 | if (odirect) |
| 506 | flags |= OS_O_DIRECT; |
| 507 | |
| 508 | memset(&file, 0, sizeof(file)); |
| 509 | file.file_name = strdup(filename); |
| 510 | |
| 511 | file.fd = open(filename, flags); |
| 512 | if (file.fd == -1) { |
| 513 | perror("open"); |
| 514 | goto err; |
| 515 | } |
| 516 | |
| 517 | if (fstat(file.fd, &sb) < 0) { |
| 518 | perror("fstat"); |
| 519 | goto err; |
| 520 | } |
| 521 | |
| 522 | dev_size = get_size(&file, &sb); |
| 523 | if (!dev_size) |
| 524 | goto err; |
| 525 | |
| 526 | if (use_bloom) { |
| 527 | uint64_t bloom_entries; |
| 528 | |
| 529 | bloom_entries = 8 * (dev_size / blocksize); |
| 530 | bloom = bloom_new(bloom_entries); |
| 531 | } |
| 532 | |
| 533 | printf("Will check <%s>, size <%llu>, using %u threads\n", filename, |
| 534 | (unsigned long long) dev_size, num_threads); |
| 535 | |
| 536 | return run_dedupe_threads(&file, dev_size, nextents, nchunks, |
| 537 | unique_capacity); |
| 538 | err: |
| 539 | if (file.fd != -1) |
| 540 | close(file.fd); |
| 541 | free(file.file_name); |
| 542 | return 1; |
| 543 | } |
| 544 | |
| 545 | static void show_chunk(struct chunk *c) |
| 546 | { |
| 547 | struct flist_head *n; |
| 548 | struct extent *e; |
| 549 | |
| 550 | printf("c hash %8x %8x %8x %8x, count %lu\n", c->hash[0], c->hash[1], |
| 551 | c->hash[2], c->hash[3], (unsigned long) c->count); |
| 552 | flist_for_each(n, &c->extent_list[0]) { |
| 553 | e = flist_entry(n, struct extent, list); |
| 554 | printf("\toffset %llu\n", (unsigned long long) e->offset); |
| 555 | } |
| 556 | } |
| 557 | |
| 558 | static const char *capacity_unit[] = {"b","KB", "MB", "GB", "TB", "PB", "EB"}; |
| 559 | |
| 560 | static uint64_t bytes_to_human_readable_unit(uint64_t n, const char **unit_out) |
| 561 | { |
| 562 | uint8_t i = 0; |
| 563 | |
| 564 | while (n >= 1024) { |
| 565 | i++; |
| 566 | n /= 1024; |
| 567 | } |
| 568 | |
| 569 | *unit_out = capacity_unit[i]; |
| 570 | return n; |
| 571 | } |
| 572 | |
| 573 | static void show_stat(uint64_t nextents, uint64_t nchunks, uint64_t ndupextents, |
| 574 | uint64_t unique_capacity) |
| 575 | { |
| 576 | double perc, ratio; |
| 577 | const char *unit; |
| 578 | uint64_t uc_human; |
| 579 | |
| 580 | printf("Extents=%lu, Unique extents=%lu", (unsigned long) nextents, |
| 581 | (unsigned long) nchunks); |
| 582 | if (!bloom) |
| 583 | printf(" Duplicated extents=%lu", (unsigned long) ndupextents); |
| 584 | printf("\n"); |
| 585 | |
| 586 | if (nchunks) { |
| 587 | ratio = (double) nextents / (double) nchunks; |
| 588 | printf("De-dupe ratio: 1:%3.2f\n", ratio - 1.0); |
| 589 | } else { |
| 590 | printf("De-dupe ratio: 1:infinite\n"); |
| 591 | } |
| 592 | |
| 593 | if (ndupextents) { |
| 594 | printf("De-dupe working set at least: %3.2f%%\n", |
| 595 | 100.0 * (double) ndupextents / (double) nextents); |
| 596 | } |
| 597 | |
| 598 | perc = 1.00 - ((double) nchunks / (double) nextents); |
| 599 | perc *= 100.0; |
| 600 | printf("Fio setting: dedupe_percentage=%u\n", (int) (perc + 0.50)); |
| 601 | |
| 602 | |
| 603 | if (compression) { |
| 604 | uc_human = bytes_to_human_readable_unit(unique_capacity, &unit); |
| 605 | printf("Unique capacity %lu%s\n", (unsigned long) uc_human, unit); |
| 606 | } |
| 607 | } |
| 608 | |
| 609 | static void iter_rb_tree(uint64_t *nextents, uint64_t *nchunks, uint64_t *ndupextents) |
| 610 | { |
| 611 | struct fio_rb_node *n; |
| 612 | *nchunks = *nextents = *ndupextents = 0; |
| 613 | |
| 614 | n = rb_first(&rb_root); |
| 615 | if (!n) |
| 616 | return; |
| 617 | |
| 618 | do { |
| 619 | struct chunk *c; |
| 620 | |
| 621 | c = rb_entry(n, struct chunk, rb_node); |
| 622 | (*nchunks)++; |
| 623 | *nextents += c->count; |
| 624 | *ndupextents += (c->count > 1); |
| 625 | |
| 626 | if (dump_output) |
| 627 | show_chunk(c); |
| 628 | |
| 629 | } while ((n = rb_next(n)) != NULL); |
| 630 | } |
| 631 | |
| 632 | static int usage(char *argv[]) |
| 633 | { |
| 634 | log_err("Check for dedupable blocks on a device/file\n\n"); |
| 635 | log_err("%s: [options] <device or file>\n", argv[0]); |
| 636 | log_err("\t-b\tChunk size to use\n"); |
| 637 | log_err("\t-t\tNumber of threads to use\n"); |
| 638 | log_err("\t-d\tFull extent/chunk debug output\n"); |
| 639 | log_err("\t-o\tUse O_DIRECT\n"); |
| 640 | log_err("\t-c\tFull collision check\n"); |
| 641 | log_err("\t-B\tUse probabilistic bloom filter\n"); |
| 642 | log_err("\t-p\tPrint progress indicator\n"); |
| 643 | log_err("\t-C\tCalculate compressible size\n"); |
| 644 | return 1; |
| 645 | } |
| 646 | |
| 647 | int main(int argc, char *argv[]) |
| 648 | { |
| 649 | uint64_t nextents = 0, nchunks = 0, ndupextents = 0, unique_capacity; |
| 650 | int c, ret; |
| 651 | |
| 652 | arch_init(argv); |
| 653 | debug_init(); |
| 654 | |
| 655 | while ((c = getopt(argc, argv, "b:t:d:o:c:p:B:C:")) != -1) { |
| 656 | switch (c) { |
| 657 | case 'b': |
| 658 | blocksize = atoi(optarg); |
| 659 | break; |
| 660 | case 't': |
| 661 | num_threads = atoi(optarg); |
| 662 | break; |
| 663 | case 'd': |
| 664 | dump_output = atoi(optarg); |
| 665 | break; |
| 666 | case 'o': |
| 667 | odirect = atoi(optarg); |
| 668 | break; |
| 669 | case 'c': |
| 670 | collision_check = atoi(optarg); |
| 671 | break; |
| 672 | case 'p': |
| 673 | print_progress = atoi(optarg); |
| 674 | break; |
| 675 | case 'B': |
| 676 | use_bloom = atoi(optarg); |
| 677 | break; |
| 678 | case 'C': |
| 679 | compression = atoi(optarg); |
| 680 | break; |
| 681 | case '?': |
| 682 | default: |
| 683 | return usage(argv); |
| 684 | } |
| 685 | } |
| 686 | |
| 687 | if (collision_check || dump_output || compression) |
| 688 | use_bloom = 0; |
| 689 | |
| 690 | if (!num_threads) |
| 691 | num_threads = cpus_configured(); |
| 692 | |
| 693 | if (argc == optind) |
| 694 | return usage(argv); |
| 695 | |
| 696 | sinit(); |
| 697 | |
| 698 | rb_root = RB_ROOT; |
| 699 | rb_lock = fio_sem_init(FIO_SEM_UNLOCKED); |
| 700 | |
| 701 | ret = dedupe_check(argv[optind], &nextents, &nchunks, &unique_capacity); |
| 702 | |
| 703 | if (!ret) { |
| 704 | if (!bloom) |
| 705 | iter_rb_tree(&nextents, &nchunks, &ndupextents); |
| 706 | |
| 707 | show_stat(nextents, nchunks, ndupextents, unique_capacity); |
| 708 | } |
| 709 | |
| 710 | fio_sem_remove(rb_lock); |
| 711 | if (bloom) |
| 712 | bloom_free(bloom); |
| 713 | scleanup(); |
| 714 | return ret; |
| 715 | } |