pack-bitmap.c

   1 #include "cache.h"
   2 #include "commit.h"
   3 #include "tag.h"
   4 #include "diff.h"
   5 #include "revision.h"
   6 #include "progress.h"
   7 #include "list-objects.h"
   8 #include "pack.h"
   9 #include "pack-bitmap.h"
  10 #include "pack-revindex.h"
  11 #include "pack-objects.h"
  12 #include "packfile.h"
  13 #include "repository.h"
  14 #include "object-store.h"
  15
  16 /*
  17  * An entry on the bitmap index, representing the bitmap for a given
  18  * commit.
  19  */
  20 struct stored_bitmap {
  21         unsigned char sha1[20];
  22         struct ewah_bitmap *root;
  23         struct stored_bitmap *xor;
  24         int flags;
  25 };
  26
  27 /*
  28  * The active bitmap index for a repository. By design, repositories only have
  29  * a single bitmap index available (the index for the biggest packfile in
  30  * the repository), since bitmap indexes need full closure.
  31  *
  32  * If there is more than one bitmap index available (e.g. because of alternates),
  33  * the active bitmap index is the largest one.
  34  */
  35 struct bitmap_index {
  36         /* Packfile to which this bitmap index belongs to */
  37         struct packed_git *pack;
  38
  39         /*
  40          * Mark the first `reuse_objects` in the packfile as reused:
  41          * they will be sent as-is without using them for repacking
  42          * calculations
  43          */
  44         uint32_t reuse_objects;
  45
  46         /* mmapped buffer of the whole bitmap index */
  47         unsigned char *map;
  48         size_t map_size; /* size of the mmaped buffer */
  49         size_t map_pos; /* current position when loading the index */
  50
  51         /*
  52          * Type indexes.
  53          *
  54          * Each bitmap marks which objects in the packfile  are of the given
  55          * type. This provides type information when yielding the objects from
  56          * the packfile during a walk, which allows for better delta bases.
  57          */
  58         struct ewah_bitmap *commits;
  59         struct ewah_bitmap *trees;
  60         struct ewah_bitmap *blobs;
  61         struct ewah_bitmap *tags;
  62
  63         /* Map from SHA1 -> `stored_bitmap` for all the bitmapped commits */
  64         khash_sha1 *bitmaps;
  65
  66         /* Number of bitmapped commits */
  67         uint32_t entry_count;
  68
  69         /* If not NULL, this is a name-hash cache pointing into map. */
  70         uint32_t *hashes;
  71
  72         /*
  73          * Extended index.
  74          *
  75          * When trying to perform bitmap operations with objects that are not
  76          * packed in `pack`, these objects are added to this "fake index" and
  77          * are assumed to appear at the end of the packfile for all operations
  78          */
  79         struct eindex {
  80                 struct object **objects;
  81                 uint32_t *hashes;
  82                 uint32_t count, alloc;
  83                 khash_sha1_pos *positions;
  84         } ext_index;
  85
  86         /* Bitmap result of the last performed walk */
  87         struct bitmap *result;
  88
  89         /* "have" bitmap from the last performed walk */
  90         struct bitmap *haves;
  91
  92         /* Version of the bitmap index */
  93         unsigned int version;
  94 };
  95
  96 static struct ewah_bitmap *lookup_stored_bitmap(struct stored_bitmap *st)
  97 {
  98         struct ewah_bitmap *parent;
  99         struct ewah_bitmap *composed;
 100
 101         if (st->xor == NULL)
 102                 return st->root;
 103
 104         composed = ewah_pool_new();
 105         parent = lookup_stored_bitmap(st->xor);
 106         ewah_xor(st->root, parent, composed);
 107
 108         ewah_pool_free(st->root);
 109         st->root = composed;
 110         st->xor = NULL;
 111
 112         return composed;
 113 }
 114
 115 /*
 116  * Read a bitmap from the current read position on the mmaped
 117  * index, and increase the read position accordingly
 118  */
 119 static struct ewah_bitmap *read_bitmap_1(struct bitmap_index *index)
 120 {
 121         struct ewah_bitmap *b = ewah_pool_new();
 122
 123         ssize_t bitmap_size = ewah_read_mmap(b,
 124                 index->map + index->map_pos,
 125                 index->map_size - index->map_pos);
 126
 127         if (bitmap_size < 0) {
 128                 error("Failed to load bitmap index (corrupted?)");
 129                 ewah_pool_free(b);
 130                 return NULL;
 131         }
 132
 133         index->map_pos += bitmap_size;
 134         return b;
 135 }
 136
 137 static int load_bitmap_header(struct bitmap_index *index)
 138 {
 139         struct bitmap_disk_header *header = (void *)index->map;
 140
 141         if (index->map_size < sizeof(*header) + 20)
 142                 return error("Corrupted bitmap index (missing header data)");
 143
 144         if (memcmp(header->magic, BITMAP_IDX_SIGNATURE, sizeof(BITMAP_IDX_SIGNATURE)) != 0)
 145                 return error("Corrupted bitmap index file (wrong header)");
 146
 147         index->version = ntohs(header->version);
 148         if (index->version != 1)
 149                 return error("Unsupported version for bitmap index file (%d)", index->version);
 150
 151         /* Parse known bitmap format options */
 152         {
 153                 uint32_t flags = ntohs(header->options);
 154
 155                 if ((flags & BITMAP_OPT_FULL_DAG) == 0)
 156                         return error("Unsupported options for bitmap index file "
 157                                 "(Git requires BITMAP_OPT_FULL_DAG)");
 158
 159                 if (flags & BITMAP_OPT_HASH_CACHE) {
 160                         unsigned char *end = index->map + index->map_size - 20;
 161                         index->hashes = ((uint32_t *)end) - index->pack->num_objects;
 162                 }
 163         }
 164
 165         index->entry_count = ntohl(header->entry_count);
 166         index->map_pos += sizeof(*header);
 167         return 0;
 168 }
 169
 170 static struct stored_bitmap *store_bitmap(struct bitmap_index *index,
 171                                           struct ewah_bitmap *root,
 172                                           const unsigned char *sha1,
 173                                           struct stored_bitmap *xor_with,
 174                                           int flags)
 175 {
 176         struct stored_bitmap *stored;
 177         khiter_t hash_pos;
 178         int ret;
 179
 180         stored = xmalloc(sizeof(struct stored_bitmap));
 181         stored->root = root;
 182         stored->xor = xor_with;
 183         stored->flags = flags;
 184         hashcpy(stored->sha1, sha1);
 185
 186         hash_pos = kh_put_sha1(index->bitmaps, stored->sha1, &ret);
 187
 188         /* a 0 return code means the insertion succeeded with no changes,
 189          * because the SHA1 already existed on the map. this is bad, there
 190          * shouldn't be duplicated commits in the index */
 191         if (ret == 0) {
 192                 error("Duplicate entry in bitmap index: %s", sha1_to_hex(sha1));
 193                 return NULL;
 194         }
 195
 196         kh_value(index->bitmaps, hash_pos) = stored;
 197         return stored;
 198 }
 199
 200 static inline uint32_t read_be32(const unsigned char *buffer, size_t *pos)
 201 {
 202         uint32_t result = get_be32(buffer + *pos);
 203         (*pos) += sizeof(result);
 204         return result;
 205 }
 206
 207 static inline uint8_t read_u8(const unsigned char *buffer, size_t *pos)
 208 {
 209         return buffer[(*pos)++];
 210 }
 211
 212 #define MAX_XOR_OFFSET 160
 213
 214 static int load_bitmap_entries_v1(struct bitmap_index *index)
 215 {
 216         uint32_t i;
 217         struct stored_bitmap *recent_bitmaps[MAX_XOR_OFFSET] = { NULL };
 218
 219         for (i = 0; i < index->entry_count; ++i) {
 220                 int xor_offset, flags;
 221                 struct ewah_bitmap *bitmap = NULL;
 222                 struct stored_bitmap *xor_bitmap = NULL;
 223                 uint32_t commit_idx_pos;
 224                 const unsigned char *sha1;
 225
 226                 commit_idx_pos = read_be32(index->map, &index->map_pos);
 227                 xor_offset = read_u8(index->map, &index->map_pos);
 228                 flags = read_u8(index->map, &index->map_pos);
 229
 230                 sha1 = nth_packed_object_sha1(index->pack, commit_idx_pos);
 231
 232                 bitmap = read_bitmap_1(index);
 233                 if (!bitmap)
 234                         return -1;
 235
 236                 if (xor_offset > MAX_XOR_OFFSET || xor_offset > i)
 237                         return error("Corrupted bitmap pack index");
 238
 239                 if (xor_offset > 0) {
 240                         xor_bitmap = recent_bitmaps[(i - xor_offset) % MAX_XOR_OFFSET];
 241
 242                         if (xor_bitmap == NULL)
 243                                 return error("Invalid XOR offset in bitmap pack index");
 244                 }
 245
 246                 recent_bitmaps[i % MAX_XOR_OFFSET] = store_bitmap(
 247                         index, bitmap, sha1, xor_bitmap, flags);
 248         }
 249
 250         return 0;
 251 }
 252
 253 static char *pack_bitmap_filename(struct packed_git *p)
 254 {
 255         size_t len;
 256
 257         if (!strip_suffix(p->pack_name, ".pack", &len))
 258                 BUG("pack_name does not end in .pack");
 259         return xstrfmt("%.*s.bitmap", (int)len, p->pack_name);
 260 }
 261
 262 static int open_pack_bitmap_1(struct bitmap_index *bitmap_git, struct packed_git *packfile)
 263 {
 264         int fd;
 265         struct stat st;
 266         char *idx_name;
 267
 268         if (open_pack_index(packfile))
 269                 return -1;
 270
 271         idx_name = pack_bitmap_filename(packfile);
 272         fd = git_open(idx_name);
 273         free(idx_name);
 274
 275         if (fd < 0)
 276                 return -1;
 277
 278         if (fstat(fd, &st)) {
 279                 close(fd);
 280                 return -1;
 281         }
 282
 283         if (bitmap_git->pack) {
 284                 warning("ignoring extra bitmap file: %s", packfile->pack_name);
 285                 close(fd);
 286                 return -1;
 287         }
 288
 289         bitmap_git->pack = packfile;
 290         bitmap_git->map_size = xsize_t(st.st_size);
 291         bitmap_git->map = xmmap(NULL, bitmap_git->map_size, PROT_READ, MAP_PRIVATE, fd, 0);
 292         bitmap_git->map_pos = 0;
 293         close(fd);
 294
 295         if (load_bitmap_header(bitmap_git) < 0) {
 296                 munmap(bitmap_git->map, bitmap_git->map_size);
 297                 bitmap_git->map = NULL;
 298                 bitmap_git->map_size = 0;
 299                 return -1;
 300         }
 301
 302         return 0;
 303 }
 304
 305 static int load_pack_bitmap(struct bitmap_index *bitmap_git)
 306 {
 307         assert(bitmap_git->map);
 308
 309         bitmap_git->bitmaps = kh_init_sha1();
 310         bitmap_git->ext_index.positions = kh_init_sha1_pos();
 311         load_pack_revindex(bitmap_git->pack);
 312
 313         if (!(bitmap_git->commits = read_bitmap_1(bitmap_git)) ||
 314                 !(bitmap_git->trees = read_bitmap_1(bitmap_git)) ||
 315                 !(bitmap_git->blobs = read_bitmap_1(bitmap_git)) ||
 316                 !(bitmap_git->tags = read_bitmap_1(bitmap_git)))
 317                 goto failed;
 318
 319         if (load_bitmap_entries_v1(bitmap_git) < 0)
 320                 goto failed;
 321
 322         return 0;
 323
 324 failed:
 325         munmap(bitmap_git->map, bitmap_git->map_size);
 326         bitmap_git->map = NULL;
 327         bitmap_git->map_size = 0;
 328         return -1;
 329 }
 330
 331 static int open_pack_bitmap(struct repository *r,
 332                             struct bitmap_index *bitmap_git)
 333 {
 334         struct packed_git *p;
 335         int ret = -1;
 336
 337         assert(!bitmap_git->map);
 338
 339         for (p = get_all_packs(r); p; p = p->next) {
 340                 if (open_pack_bitmap_1(bitmap_git, p) == 0)
 341                         ret = 0;
 342         }
 343
 344         return ret;
 345 }
 346
 347 struct bitmap_index *prepare_bitmap_git(struct repository *r)
 348 {
 349         struct bitmap_index *bitmap_git = xcalloc(1, sizeof(*bitmap_git));
 350
 351         if (!open_pack_bitmap(r, bitmap_git) && !load_pack_bitmap(bitmap_git))
 352                 return bitmap_git;
 353
 354         free_bitmap_index(bitmap_git);
 355         return NULL;
 356 }
 357
 358 struct include_data {
 359         struct bitmap_index *bitmap_git;
 360         struct bitmap *base;
 361         struct bitmap *seen;
 362 };
 363
 364 static inline int bitmap_position_extended(struct bitmap_index *bitmap_git,
 365                                            const unsigned char *sha1)
 366 {
 367         khash_sha1_pos *positions = bitmap_git->ext_index.positions;
 368         khiter_t pos = kh_get_sha1_pos(positions, sha1);
 369
 370         if (pos < kh_end(positions)) {
 371                 int bitmap_pos = kh_value(positions, pos);
 372                 return bitmap_pos + bitmap_git->pack->num_objects;
 373         }
 374
 375         return -1;
 376 }
 377
 378 static inline int bitmap_position_packfile(struct bitmap_index *bitmap_git,
 379                                            const unsigned char *sha1)
 380 {
 381         off_t offset = find_pack_entry_one(sha1, bitmap_git->pack);
 382         if (!offset)
 383                 return -1;
 384
 385         return find_revindex_position(bitmap_git->pack, offset);
 386 }
 387
 388 static int bitmap_position(struct bitmap_index *bitmap_git,
 389                            const unsigned char *sha1)
 390 {
 391         int pos = bitmap_position_packfile(bitmap_git, sha1);
 392         return (pos >= 0) ? pos : bitmap_position_extended(bitmap_git, sha1);
 393 }
 394
 395 static int ext_index_add_object(struct bitmap_index *bitmap_git,
 396                                 struct object *object, const char *name)
 397 {
 398         struct eindex *eindex = &bitmap_git->ext_index;
 399
 400         khiter_t hash_pos;
 401         int hash_ret;
 402         int bitmap_pos;
 403
 404         hash_pos = kh_put_sha1_pos(eindex->positions, object->oid.hash, &hash_ret);
 405         if (hash_ret > 0) {
 406                 if (eindex->count >= eindex->alloc) {
 407                         eindex->alloc = (eindex->alloc + 16) * 3 / 2;
 408                         REALLOC_ARRAY(eindex->objects, eindex->alloc);
 409                         REALLOC_ARRAY(eindex->hashes, eindex->alloc);
 410                 }
 411
 412                 bitmap_pos = eindex->count;
 413                 eindex->objects[eindex->count] = object;
 414                 eindex->hashes[eindex->count] = pack_name_hash(name);
 415                 kh_value(eindex->positions, hash_pos) = bitmap_pos;
 416                 eindex->count++;
 417         } else {
 418                 bitmap_pos = kh_value(eindex->positions, hash_pos);
 419         }
 420
 421         return bitmap_pos + bitmap_git->pack->num_objects;
 422 }
 423
 424 struct bitmap_show_data {
 425         struct bitmap_index *bitmap_git;
 426         struct bitmap *base;
 427 };
 428
 429 static void show_object(struct object *object, const char *name, void *data_)
 430 {
 431         struct bitmap_show_data *data = data_;
 432         int bitmap_pos;
 433
 434         bitmap_pos = bitmap_position(data->bitmap_git, object->oid.hash);
 435
 436         if (bitmap_pos < 0)
 437                 bitmap_pos = ext_index_add_object(data->bitmap_git, object,
 438                                                   name);
 439
 440         bitmap_set(data->base, bitmap_pos);
 441 }
 442
 443 static void show_commit(struct commit *commit, void *data)
 444 {
 445 }
 446
 447 static int add_to_include_set(struct bitmap_index *bitmap_git,
 448                               struct include_data *data,
 449                               const unsigned char *sha1,
 450                               int bitmap_pos)
 451 {
 452         khiter_t hash_pos;
 453
 454         if (data->seen && bitmap_get(data->seen, bitmap_pos))
 455                 return 0;
 456
 457         if (bitmap_get(data->base, bitmap_pos))
 458                 return 0;
 459
 460         hash_pos = kh_get_sha1(bitmap_git->bitmaps, sha1);
 461         if (hash_pos < kh_end(bitmap_git->bitmaps)) {
 462                 struct stored_bitmap *st = kh_value(bitmap_git->bitmaps, hash_pos);
 463                 bitmap_or_ewah(data->base, lookup_stored_bitmap(st));
 464                 return 0;
 465         }
 466
 467         bitmap_set(data->base, bitmap_pos);
 468         return 1;
 469 }
 470
 471 static int should_include(struct commit *commit, void *_data)
 472 {
 473         struct include_data *data = _data;
 474         int bitmap_pos;
 475
 476         bitmap_pos = bitmap_position(data->bitmap_git, commit->object.oid.hash);
 477         if (bitmap_pos < 0)
 478                 bitmap_pos = ext_index_add_object(data->bitmap_git,
 479                                                   (struct object *)commit,
 480                                                   NULL);
 481
 482         if (!add_to_include_set(data->bitmap_git, data, commit->object.oid.hash,
 483                                 bitmap_pos)) {
 484                 struct commit_list *parent = commit->parents;
 485
 486                 while (parent) {
 487                         parent->item->object.flags |= SEEN;
 488                         parent = parent->next;
 489                 }
 490
 491                 return 0;
 492         }
 493
 494         return 1;
 495 }
 496
 497 static struct bitmap *find_objects(struct bitmap_index *bitmap_git,
 498                                    struct rev_info *revs,
 499                                    struct object_list *roots,
 500                                    struct bitmap *seen)
 501 {
 502         struct bitmap *base = NULL;
 503         int needs_walk = 0;
 504
 505         struct object_list *not_mapped = NULL;
 506
 507         /*
 508          * Go through all the roots for the walk. The ones that have bitmaps
 509          * on the bitmap index will be `or`ed together to form an initial
 510          * global reachability analysis.
 511          *
 512          * The ones without bitmaps in the index will be stored in the
 513          * `not_mapped_list` for further processing.
 514          */
 515         while (roots) {
 516                 struct object *object = roots->item;
 517                 roots = roots->next;
 518
 519                 if (object->type == OBJ_COMMIT) {
 520                         khiter_t pos = kh_get_sha1(bitmap_git->bitmaps, object->oid.hash);
 521
 522                         if (pos < kh_end(bitmap_git->bitmaps)) {
 523                                 struct stored_bitmap *st = kh_value(bitmap_git->bitmaps, pos);
 524                                 struct ewah_bitmap *or_with = lookup_stored_bitmap(st);
 525
 526                                 if (base == NULL)
 527                                         base = ewah_to_bitmap(or_with);
 528                                 else
 529                                         bitmap_or_ewah(base, or_with);
 530
 531                                 object->flags |= SEEN;
 532                                 continue;
 533                         }
 534                 }
 535
 536                 object_list_insert(object, &not_mapped);
 537         }
 538
 539         /*
 540          * Best case scenario: We found bitmaps for all the roots,
 541          * so the resulting `or` bitmap has the full reachability analysis
 542          */
 543         if (not_mapped == NULL)
 544                 return base;
 545
 546         roots = not_mapped;
 547
 548         /*
 549          * Let's iterate through all the roots that don't have bitmaps to
 550          * check if we can determine them to be reachable from the existing
 551          * global bitmap.
 552          *
 553          * If we cannot find them in the existing global bitmap, we'll need
 554          * to push them to an actual walk and run it until we can confirm
 555          * they are reachable
 556          */
 557         while (roots) {
 558                 struct object *object = roots->item;
 559                 int pos;
 560
 561                 roots = roots->next;
 562                 pos = bitmap_position(bitmap_git, object->oid.hash);
 563
 564                 if (pos < 0 || base == NULL || !bitmap_get(base, pos)) {
 565                         object->flags &= ~UNINTERESTING;
 566                         add_pending_object(revs, object, "");
 567                         needs_walk = 1;
 568                 } else {
 569                         object->flags |= SEEN;
 570                 }
 571         }
 572
 573         if (needs_walk) {
 574                 struct include_data incdata;
 575                 struct bitmap_show_data show_data;
 576
 577                 if (base == NULL)
 578                         base = bitmap_new();
 579
 580                 incdata.bitmap_git = bitmap_git;
 581                 incdata.base = base;
 582                 incdata.seen = seen;
 583
 584                 revs->include_check = should_include;
 585                 revs->include_check_data = &incdata;
 586
 587                 if (prepare_revision_walk(revs))
 588                         die("revision walk setup failed");
 589
 590                 show_data.bitmap_git = bitmap_git;
 591                 show_data.base = base;
 592
 593                 traverse_commit_list(revs, show_commit, show_object,
 594                                      &show_data);
 595         }
 596
 597         return base;
 598 }
 599
 600 static void show_extended_objects(struct bitmap_index *bitmap_git,
 601                                   show_reachable_fn show_reach)
 602 {
 603         struct bitmap *objects = bitmap_git->result;
 604         struct eindex *eindex = &bitmap_git->ext_index;
 605         uint32_t i;
 606
 607         for (i = 0; i < eindex->count; ++i) {
 608                 struct object *obj;
 609
 610                 if (!bitmap_get(objects, bitmap_git->pack->num_objects + i))
 611                         continue;
 612
 613                 obj = eindex->objects[i];
 614                 show_reach(&obj->oid, obj->type, 0, eindex->hashes[i], NULL, 0);
 615         }
 616 }
 617
 618 static void show_objects_for_type(
 619         struct bitmap_index *bitmap_git,
 620         struct ewah_bitmap *type_filter,
 621         enum object_type object_type,
 622         show_reachable_fn show_reach)
 623 {
 624         size_t pos = 0, i = 0;
 625         uint32_t offset;
 626
 627         struct ewah_iterator it;
 628         eword_t filter;
 629
 630         struct bitmap *objects = bitmap_git->result;
 631
 632         if (bitmap_git->reuse_objects == bitmap_git->pack->num_objects)
 633                 return;
 634
 635         ewah_iterator_init(&it, type_filter);
 636
 637         while (i < objects->word_alloc && ewah_iterator_next(&filter, &it)) {
 638                 eword_t word = objects->words[i] & filter;
 639
 640                 for (offset = 0; offset < BITS_IN_EWORD; ++offset) {
 641                         struct object_id oid;
 642                         struct revindex_entry *entry;
 643                         uint32_t hash = 0;
 644
 645                         if ((word >> offset) == 0)
 646                                 break;
 647
 648                         offset += ewah_bit_ctz64(word >> offset);
 649
 650                         if (pos + offset < bitmap_git->reuse_objects)
 651                                 continue;
 652
 653                         entry = &bitmap_git->pack->revindex[pos + offset];
 654                         nth_packed_object_oid(&oid, bitmap_git->pack, entry->nr);
 655
 656                         if (bitmap_git->hashes)
 657                                 hash = get_be32(bitmap_git->hashes + entry->nr);
 658
 659                         show_reach(&oid, object_type, 0, hash, bitmap_git->pack, entry->offset);
 660                 }
 661
 662                 pos += BITS_IN_EWORD;
 663                 i++;
 664         }
 665 }
 666
 667 static int in_bitmapped_pack(struct bitmap_index *bitmap_git,
 668                              struct object_list *roots)
 669 {
 670         while (roots) {
 671                 struct object *object = roots->item;
 672                 roots = roots->next;
 673
 674                 if (find_pack_entry_one(object->oid.hash, bitmap_git->pack) > 0)
 675                         return 1;
 676         }
 677
 678         return 0;
 679 }
 680
 681 struct bitmap_index *prepare_bitmap_walk(struct rev_info *revs)
 682 {
 683         unsigned int i;
 684
 685         struct object_list *wants = NULL;
 686         struct object_list *haves = NULL;
 687
 688         struct bitmap *wants_bitmap = NULL;
 689         struct bitmap *haves_bitmap = NULL;
 690
 691         struct bitmap_index *bitmap_git = xcalloc(1, sizeof(*bitmap_git));
 692         /* try to open a bitmapped pack, but don't parse it yet
 693          * because we may not need to use it */
 694         if (open_pack_bitmap(revs->repo, bitmap_git) < 0)
 695                 goto cleanup;
 696
 697         for (i = 0; i < revs->pending.nr; ++i) {
 698                 struct object *object = revs->pending.objects[i].item;
 699
 700                 if (object->type == OBJ_NONE)
 701                         parse_object_or_die(&object->oid, NULL);
 702
 703                 while (object->type == OBJ_TAG) {
 704                         struct tag *tag = (struct tag *) object;
 705
 706                         if (object->flags & UNINTERESTING)
 707                                 object_list_insert(object, &haves);
 708                         else
 709                                 object_list_insert(object, &wants);
 710
 711                         if (!tag->tagged)
 712                                 die("bad tag");
 713                         object = parse_object_or_die(&tag->tagged->oid, NULL);
 714                 }
 715
 716                 if (object->flags & UNINTERESTING)
 717                         object_list_insert(object, &haves);
 718                 else
 719                         object_list_insert(object, &wants);
 720         }
 721
 722         /*
 723          * if we have a HAVES list, but none of those haves is contained
 724          * in the packfile that has a bitmap, we don't have anything to
 725          * optimize here
 726          */
 727         if (haves && !in_bitmapped_pack(bitmap_git, haves))
 728                 goto cleanup;
 729
 730         /* if we don't want anything, we're done here */
 731         if (!wants)
 732                 goto cleanup;
 733
 734         /*
 735          * now we're going to use bitmaps, so load the actual bitmap entries
 736          * from disk. this is the point of no return; after this the rev_list
 737          * becomes invalidated and we must perform the revwalk through bitmaps
 738          */
 739         if (load_pack_bitmap(bitmap_git) < 0)
 740                 goto cleanup;
 741
 742         object_array_clear(&revs->pending);
 743
 744         if (haves) {
 745                 revs->ignore_missing_links = 1;
 746                 haves_bitmap = find_objects(bitmap_git, revs, haves, NULL);
 747                 reset_revision_walk();
 748                 revs->ignore_missing_links = 0;
 749
 750                 if (haves_bitmap == NULL)
 751                         BUG("failed to perform bitmap walk");
 752         }
 753
 754         wants_bitmap = find_objects(bitmap_git, revs, wants, haves_bitmap);
 755
 756         if (!wants_bitmap)
 757                 BUG("failed to perform bitmap walk");
 758
 759         if (haves_bitmap)
 760                 bitmap_and_not(wants_bitmap, haves_bitmap);
 761
 762         bitmap_git->result = wants_bitmap;
 763         bitmap_git->haves = haves_bitmap;
 764
 765         return bitmap_git;
 766
 767 cleanup:
 768         free_bitmap_index(bitmap_git);
 769         return NULL;
 770 }
 771
 772 int reuse_partial_packfile_from_bitmap(struct bitmap_index *bitmap_git,
 773                                        struct packed_git **packfile,
 774                                        uint32_t *entries,
 775                                        off_t *up_to)
 776 {
 777         /*
 778          * Reuse the packfile content if we need more than
 779          * 90% of its objects
 780          */
 781         static const double REUSE_PERCENT = 0.9;
 782
 783         struct bitmap *result = bitmap_git->result;
 784         uint32_t reuse_threshold;
 785         uint32_t i, reuse_objects = 0;
 786
 787         assert(result);
 788
 789         for (i = 0; i < result->word_alloc; ++i) {
 790                 if (result->words[i] != (eword_t)~0) {
 791                         reuse_objects += ewah_bit_ctz64(~result->words[i]);
 792                         break;
 793                 }
 794
 795                 reuse_objects += BITS_IN_EWORD;
 796         }
 797
 798 #ifdef GIT_BITMAP_DEBUG
 799         {
 800                 const unsigned char *sha1;
 801                 struct revindex_entry *entry;
 802
 803                 entry = &bitmap_git->reverse_index->revindex[reuse_objects];
 804                 sha1 = nth_packed_object_sha1(bitmap_git->pack, entry->nr);
 805
 806                 fprintf(stderr, "Failed to reuse at %d (%016llx)\n",
 807                         reuse_objects, result->words[i]);
 808                 fprintf(stderr, " %s\n", sha1_to_hex(sha1));
 809         }
 810 #endif
 811
 812         if (!reuse_objects)
 813                 return -1;
 814
 815         if (reuse_objects >= bitmap_git->pack->num_objects) {
 816                 bitmap_git->reuse_objects = *entries = bitmap_git->pack->num_objects;
 817                 *up_to = -1; /* reuse the full pack */
 818                 *packfile = bitmap_git->pack;
 819                 return 0;
 820         }
 821
 822         reuse_threshold = bitmap_popcount(bitmap_git->result) * REUSE_PERCENT;
 823
 824         if (reuse_objects < reuse_threshold)
 825                 return -1;
 826
 827         bitmap_git->reuse_objects = *entries = reuse_objects;
 828         *up_to = bitmap_git->pack->revindex[reuse_objects].offset;
 829         *packfile = bitmap_git->pack;
 830
 831         return 0;
 832 }
 833
 834 void traverse_bitmap_commit_list(struct bitmap_index *bitmap_git,
 835                                  show_reachable_fn show_reachable)
 836 {
 837         assert(bitmap_git->result);
 838
 839         show_objects_for_type(bitmap_git, bitmap_git->commits,
 840                 OBJ_COMMIT, show_reachable);
 841         show_objects_for_type(bitmap_git, bitmap_git->trees,
 842                 OBJ_TREE, show_reachable);
 843         show_objects_for_type(bitmap_git, bitmap_git->blobs,
 844                 OBJ_BLOB, show_reachable);
 845         show_objects_for_type(bitmap_git, bitmap_git->tags,
 846                 OBJ_TAG, show_reachable);
 847
 848         show_extended_objects(bitmap_git, show_reachable);
 849 }
 850
 851 static uint32_t count_object_type(struct bitmap_index *bitmap_git,
 852                                   enum object_type type)
 853 {
 854         struct bitmap *objects = bitmap_git->result;
 855         struct eindex *eindex = &bitmap_git->ext_index;
 856
 857         uint32_t i = 0, count = 0;
 858         struct ewah_iterator it;
 859         eword_t filter;
 860
 861         switch (type) {
 862         case OBJ_COMMIT:
 863                 ewah_iterator_init(&it, bitmap_git->commits);
 864                 break;
 865
 866         case OBJ_TREE:
 867                 ewah_iterator_init(&it, bitmap_git->trees);
 868                 break;
 869
 870         case OBJ_BLOB:
 871                 ewah_iterator_init(&it, bitmap_git->blobs);
 872                 break;
 873
 874         case OBJ_TAG:
 875                 ewah_iterator_init(&it, bitmap_git->tags);
 876                 break;
 877
 878         default:
 879                 return 0;
 880         }
 881
 882         while (i < objects->word_alloc && ewah_iterator_next(&filter, &it)) {
 883                 eword_t word = objects->words[i++] & filter;
 884                 count += ewah_bit_popcount64(word);
 885         }
 886
 887         for (i = 0; i < eindex->count; ++i) {
 888                 if (eindex->objects[i]->type == type &&
 889                         bitmap_get(objects, bitmap_git->pack->num_objects + i))
 890                         count++;
 891         }
 892
 893         return count;
 894 }
 895
 896 void count_bitmap_commit_list(struct bitmap_index *bitmap_git,
 897                               uint32_t *commits, uint32_t *trees,
 898                               uint32_t *blobs, uint32_t *tags)
 899 {
 900         assert(bitmap_git->result);
 901
 902         if (commits)
 903                 *commits = count_object_type(bitmap_git, OBJ_COMMIT);
 904
 905         if (trees)
 906                 *trees = count_object_type(bitmap_git, OBJ_TREE);
 907
 908         if (blobs)
 909                 *blobs = count_object_type(bitmap_git, OBJ_BLOB);
 910
 911         if (tags)
 912                 *tags = count_object_type(bitmap_git, OBJ_TAG);
 913 }
 914
 915 struct bitmap_test_data {
 916         struct bitmap_index *bitmap_git;
 917         struct bitmap *base;
 918         struct progress *prg;
 919         size_t seen;
 920 };
 921
 922 static void test_show_object(struct object *object, const char *name,
 923                              void *data)
 924 {
 925         struct bitmap_test_data *tdata = data;
 926         int bitmap_pos;
 927
 928         bitmap_pos = bitmap_position(tdata->bitmap_git, object->oid.hash);
 929         if (bitmap_pos < 0)
 930                 die("Object not in bitmap: %s\n", oid_to_hex(&object->oid));
 931
 932         bitmap_set(tdata->base, bitmap_pos);
 933         display_progress(tdata->prg, ++tdata->seen);
 934 }
 935
 936 static void test_show_commit(struct commit *commit, void *data)
 937 {
 938         struct bitmap_test_data *tdata = data;
 939         int bitmap_pos;
 940
 941         bitmap_pos = bitmap_position(tdata->bitmap_git,
 942                                      commit->object.oid.hash);
 943         if (bitmap_pos < 0)
 944                 die("Object not in bitmap: %s\n", oid_to_hex(&commit->object.oid));
 945
 946         bitmap_set(tdata->base, bitmap_pos);
 947         display_progress(tdata->prg, ++tdata->seen);
 948 }
 949
 950 void test_bitmap_walk(struct rev_info *revs)
 951 {
 952         struct object *root;
 953         struct bitmap *result = NULL;
 954         khiter_t pos;
 955         size_t result_popcnt;
 956         struct bitmap_test_data tdata;
 957         struct bitmap_index *bitmap_git;
 958
 959         if (!(bitmap_git = prepare_bitmap_git(revs->repo)))
 960                 die("failed to load bitmap indexes");
 961
 962         if (revs->pending.nr != 1)
 963                 die("you must specify exactly one commit to test");
 964
 965         fprintf(stderr, "Bitmap v%d test (%d entries loaded)\n",
 966                 bitmap_git->version, bitmap_git->entry_count);
 967
 968         root = revs->pending.objects[0].item;
 969         pos = kh_get_sha1(bitmap_git->bitmaps, root->oid.hash);
 970
 971         if (pos < kh_end(bitmap_git->bitmaps)) {
 972                 struct stored_bitmap *st = kh_value(bitmap_git->bitmaps, pos);
 973                 struct ewah_bitmap *bm = lookup_stored_bitmap(st);
 974
 975                 fprintf(stderr, "Found bitmap for %s. %d bits / %08x checksum\n",
 976                         oid_to_hex(&root->oid), (int)bm->bit_size, ewah_checksum(bm));
 977
 978                 result = ewah_to_bitmap(bm);
 979         }
 980
 981         if (result == NULL)
 982                 die("Commit %s doesn't have an indexed bitmap", oid_to_hex(&root->oid));
 983
 984         revs->tag_objects = 1;
 985         revs->tree_objects = 1;
 986         revs->blob_objects = 1;
 987
 988         result_popcnt = bitmap_popcount(result);
 989
 990         if (prepare_revision_walk(revs))
 991                 die("revision walk setup failed");
 992
 993         tdata.bitmap_git = bitmap_git;
 994         tdata.base = bitmap_new();
 995         tdata.prg = start_progress("Verifying bitmap entries", result_popcnt);
 996         tdata.seen = 0;
 997
 998         traverse_commit_list(revs, &test_show_commit, &test_show_object, &tdata);
 999
1000         stop_progress(&tdata.prg);
1001
1002         if (bitmap_equals(result, tdata.base))
1003                 fprintf(stderr, "OK!\n");
1004         else
1005                 fprintf(stderr, "Mismatch!\n");
1006
1007         free_bitmap_index(bitmap_git);
1008 }
1009
1010 static int rebuild_bitmap(uint32_t *reposition,
1011                           struct ewah_bitmap *source,
1012                           struct bitmap *dest)
1013 {
1014         uint32_t pos = 0;
1015         struct ewah_iterator it;
1016         eword_t word;
1017
1018         ewah_iterator_init(&it, source);
1019
1020         while (ewah_iterator_next(&word, &it)) {
1021                 uint32_t offset, bit_pos;
1022
1023                 for (offset = 0; offset < BITS_IN_EWORD; ++offset) {
1024                         if ((word >> offset) == 0)
1025                                 break;
1026
1027                         offset += ewah_bit_ctz64(word >> offset);
1028
1029                         bit_pos = reposition[pos + offset];
1030                         if (bit_pos > 0)
1031                                 bitmap_set(dest, bit_pos - 1);
1032                         else /* can't reuse, we don't have the object */
1033                                 return -1;
1034                 }
1035
1036                 pos += BITS_IN_EWORD;
1037         }
1038         return 0;
1039 }
1040
1041 int rebuild_existing_bitmaps(struct bitmap_index *bitmap_git,
1042                              struct packing_data *mapping,
1043                              khash_sha1 *reused_bitmaps,
1044                              int show_progress)
1045 {
1046         uint32_t i, num_objects;
1047         uint32_t *reposition;
1048         struct bitmap *rebuild;
1049         struct stored_bitmap *stored;
1050         struct progress *progress = NULL;
1051
1052         khiter_t hash_pos;
1053         int hash_ret;
1054
1055         num_objects = bitmap_git->pack->num_objects;
1056         reposition = xcalloc(num_objects, sizeof(uint32_t));
1057
1058         for (i = 0; i < num_objects; ++i) {
1059                 const unsigned char *sha1;
1060                 struct revindex_entry *entry;
1061                 struct object_entry *oe;
1062
1063                 entry = &bitmap_git->pack->revindex[i];
1064                 sha1 = nth_packed_object_sha1(bitmap_git->pack, entry->nr);
1065                 oe = packlist_find(mapping, sha1, NULL);
1066
1067                 if (oe)
1068                         reposition[i] = oe_in_pack_pos(mapping, oe) + 1;
1069         }
1070
1071         rebuild = bitmap_new();
1072         i = 0;
1073
1074         if (show_progress)
1075                 progress = start_progress("Reusing bitmaps", 0);
1076
1077         kh_foreach_value(bitmap_git->bitmaps, stored, {
1078                 if (stored->flags & BITMAP_FLAG_REUSE) {
1079                         if (!rebuild_bitmap(reposition,
1080                                             lookup_stored_bitmap(stored),
1081                                             rebuild)) {
1082                                 hash_pos = kh_put_sha1(reused_bitmaps,
1083                                                        stored->sha1,
1084                                                        &hash_ret);
1085                                 kh_value(reused_bitmaps, hash_pos) =
1086                                         bitmap_to_ewah(rebuild);
1087                         }
1088                         bitmap_reset(rebuild);
1089                         display_progress(progress, ++i);
1090                 }
1091         });
1092
1093         stop_progress(&progress);
1094
1095         free(reposition);
1096         bitmap_free(rebuild);
1097         return 0;
1098 }
1099
1100 void free_bitmap_index(struct bitmap_index *b)
1101 {
1102         if (!b)
1103                 return;
1104
1105         if (b->map)
1106                 munmap(b->map, b->map_size);
1107         ewah_pool_free(b->commits);
1108         ewah_pool_free(b->trees);
1109         ewah_pool_free(b->blobs);
1110         ewah_pool_free(b->tags);
1111         kh_destroy_sha1(b->bitmaps);
1112         free(b->ext_index.objects);
1113         free(b->ext_index.hashes);
1114         bitmap_free(b->result);
1115         bitmap_free(b->haves);
1116         free(b);
1117 }
1118
1119 int bitmap_has_sha1_in_uninteresting(struct bitmap_index *bitmap_git,
1120                                      const unsigned char *sha1)
1121 {
1122         int pos;
1123
1124         if (!bitmap_git)
1125                 return 0; /* no bitmap loaded */
1126         if (!bitmap_git->haves)
1127                 return 0; /* walk had no "haves" */
1128
1129         pos = bitmap_position_packfile(bitmap_git, sha1);
1130         if (pos < 0)
1131                 return 0;
1132
1133         return bitmap_get(bitmap_git->haves, pos);
1134 }