7 #include "list-objects.h"
9 #include "pack-bitmap.h"
10 #include "pack-revindex.h"
11 #include "pack-objects.h"
13 #include "repository.h"
14 #include "object-store.h"
15 #include "list-objects-filter-options.h"
18 * An entry on the bitmap index, representing the bitmap for a given
21 struct stored_bitmap {
23 struct ewah_bitmap *root;
24 struct stored_bitmap *xor;
29 * The active bitmap index for a repository. By design, repositories only have
30 * a single bitmap index available (the index for the biggest packfile in
31 * the repository), since bitmap indexes need full closure.
33 * If there is more than one bitmap index available (e.g. because of alternates),
34 * the active bitmap index is the largest one.
37 /* Packfile to which this bitmap index belongs to */
38 struct packed_git *pack;
41 * Mark the first `reuse_objects` in the packfile as reused:
42 * they will be sent as-is without using them for repacking
45 uint32_t reuse_objects;
47 /* mmapped buffer of the whole bitmap index */
49 size_t map_size; /* size of the mmaped buffer */
50 size_t map_pos; /* current position when loading the index */
55 * Each bitmap marks which objects in the packfile are of the given
56 * type. This provides type information when yielding the objects from
57 * the packfile during a walk, which allows for better delta bases.
59 struct ewah_bitmap *commits;
60 struct ewah_bitmap *trees;
61 struct ewah_bitmap *blobs;
62 struct ewah_bitmap *tags;
64 /* Map from object ID -> `stored_bitmap` for all the bitmapped commits */
65 kh_oid_map_t *bitmaps;
67 /* Number of bitmapped commits */
70 /* If not NULL, this is a name-hash cache pointing into map. */
76 * When trying to perform bitmap operations with objects that are not
77 * packed in `pack`, these objects are added to this "fake index" and
78 * are assumed to appear at the end of the packfile for all operations
81 struct object **objects;
83 uint32_t count, alloc;
84 kh_oid_pos_t *positions;
87 /* Bitmap result of the last performed walk */
88 struct bitmap *result;
90 /* "have" bitmap from the last performed walk */
93 /* Version of the bitmap index */
97 static struct ewah_bitmap *lookup_stored_bitmap(struct stored_bitmap *st)
99 struct ewah_bitmap *parent;
100 struct ewah_bitmap *composed;
105 composed = ewah_pool_new();
106 parent = lookup_stored_bitmap(st->xor);
107 ewah_xor(st->root, parent, composed);
109 ewah_pool_free(st->root);
117 * Read a bitmap from the current read position on the mmaped
118 * index, and increase the read position accordingly
120 static struct ewah_bitmap *read_bitmap_1(struct bitmap_index *index)
122 struct ewah_bitmap *b = ewah_pool_new();
124 ssize_t bitmap_size = ewah_read_mmap(b,
125 index->map + index->map_pos,
126 index->map_size - index->map_pos);
128 if (bitmap_size < 0) {
129 error("Failed to load bitmap index (corrupted?)");
134 index->map_pos += bitmap_size;
138 static int load_bitmap_header(struct bitmap_index *index)
140 struct bitmap_disk_header *header = (void *)index->map;
142 if (index->map_size < sizeof(*header) + the_hash_algo->rawsz)
143 return error("Corrupted bitmap index (missing header data)");
145 if (memcmp(header->magic, BITMAP_IDX_SIGNATURE, sizeof(BITMAP_IDX_SIGNATURE)) != 0)
146 return error("Corrupted bitmap index file (wrong header)");
148 index->version = ntohs(header->version);
149 if (index->version != 1)
150 return error("Unsupported version for bitmap index file (%d)", index->version);
152 /* Parse known bitmap format options */
154 uint32_t flags = ntohs(header->options);
156 if ((flags & BITMAP_OPT_FULL_DAG) == 0)
157 return error("Unsupported options for bitmap index file "
158 "(Git requires BITMAP_OPT_FULL_DAG)");
160 if (flags & BITMAP_OPT_HASH_CACHE) {
161 unsigned char *end = index->map + index->map_size - the_hash_algo->rawsz;
162 index->hashes = ((uint32_t *)end) - index->pack->num_objects;
166 index->entry_count = ntohl(header->entry_count);
167 index->map_pos += sizeof(*header) - GIT_MAX_RAWSZ + the_hash_algo->rawsz;
171 static struct stored_bitmap *store_bitmap(struct bitmap_index *index,
172 struct ewah_bitmap *root,
173 const struct object_id *oid,
174 struct stored_bitmap *xor_with,
177 struct stored_bitmap *stored;
181 stored = xmalloc(sizeof(struct stored_bitmap));
183 stored->xor = xor_with;
184 stored->flags = flags;
185 oidcpy(&stored->oid, oid);
187 hash_pos = kh_put_oid_map(index->bitmaps, stored->oid, &ret);
189 /* a 0 return code means the insertion succeeded with no changes,
190 * because the SHA1 already existed on the map. this is bad, there
191 * shouldn't be duplicated commits in the index */
193 error("Duplicate entry in bitmap index: %s", oid_to_hex(oid));
197 kh_value(index->bitmaps, hash_pos) = stored;
201 static inline uint32_t read_be32(const unsigned char *buffer, size_t *pos)
203 uint32_t result = get_be32(buffer + *pos);
204 (*pos) += sizeof(result);
208 static inline uint8_t read_u8(const unsigned char *buffer, size_t *pos)
210 return buffer[(*pos)++];
213 #define MAX_XOR_OFFSET 160
215 static int load_bitmap_entries_v1(struct bitmap_index *index)
218 struct stored_bitmap *recent_bitmaps[MAX_XOR_OFFSET] = { NULL };
220 for (i = 0; i < index->entry_count; ++i) {
221 int xor_offset, flags;
222 struct ewah_bitmap *bitmap = NULL;
223 struct stored_bitmap *xor_bitmap = NULL;
224 uint32_t commit_idx_pos;
225 struct object_id oid;
227 commit_idx_pos = read_be32(index->map, &index->map_pos);
228 xor_offset = read_u8(index->map, &index->map_pos);
229 flags = read_u8(index->map, &index->map_pos);
231 nth_packed_object_id(&oid, index->pack, commit_idx_pos);
233 bitmap = read_bitmap_1(index);
237 if (xor_offset > MAX_XOR_OFFSET || xor_offset > i)
238 return error("Corrupted bitmap pack index");
240 if (xor_offset > 0) {
241 xor_bitmap = recent_bitmaps[(i - xor_offset) % MAX_XOR_OFFSET];
243 if (xor_bitmap == NULL)
244 return error("Invalid XOR offset in bitmap pack index");
247 recent_bitmaps[i % MAX_XOR_OFFSET] = store_bitmap(
248 index, bitmap, &oid, xor_bitmap, flags);
254 static char *pack_bitmap_filename(struct packed_git *p)
258 if (!strip_suffix(p->pack_name, ".pack", &len))
259 BUG("pack_name does not end in .pack");
260 return xstrfmt("%.*s.bitmap", (int)len, p->pack_name);
263 static int open_pack_bitmap_1(struct bitmap_index *bitmap_git, struct packed_git *packfile)
269 if (open_pack_index(packfile))
272 idx_name = pack_bitmap_filename(packfile);
273 fd = git_open(idx_name);
279 if (fstat(fd, &st)) {
284 if (bitmap_git->pack) {
285 warning("ignoring extra bitmap file: %s", packfile->pack_name);
290 bitmap_git->pack = packfile;
291 bitmap_git->map_size = xsize_t(st.st_size);
292 bitmap_git->map = xmmap(NULL, bitmap_git->map_size, PROT_READ, MAP_PRIVATE, fd, 0);
293 bitmap_git->map_pos = 0;
296 if (load_bitmap_header(bitmap_git) < 0) {
297 munmap(bitmap_git->map, bitmap_git->map_size);
298 bitmap_git->map = NULL;
299 bitmap_git->map_size = 0;
306 static int load_pack_bitmap(struct bitmap_index *bitmap_git)
308 assert(bitmap_git->map);
310 bitmap_git->bitmaps = kh_init_oid_map();
311 bitmap_git->ext_index.positions = kh_init_oid_pos();
312 if (load_pack_revindex(bitmap_git->pack))
315 if (!(bitmap_git->commits = read_bitmap_1(bitmap_git)) ||
316 !(bitmap_git->trees = read_bitmap_1(bitmap_git)) ||
317 !(bitmap_git->blobs = read_bitmap_1(bitmap_git)) ||
318 !(bitmap_git->tags = read_bitmap_1(bitmap_git)))
321 if (load_bitmap_entries_v1(bitmap_git) < 0)
327 munmap(bitmap_git->map, bitmap_git->map_size);
328 bitmap_git->map = NULL;
329 bitmap_git->map_size = 0;
331 kh_destroy_oid_map(bitmap_git->bitmaps);
332 bitmap_git->bitmaps = NULL;
334 kh_destroy_oid_pos(bitmap_git->ext_index.positions);
335 bitmap_git->ext_index.positions = NULL;
340 static int open_pack_bitmap(struct repository *r,
341 struct bitmap_index *bitmap_git)
343 struct packed_git *p;
346 assert(!bitmap_git->map);
348 for (p = get_all_packs(r); p; p = p->next) {
349 if (open_pack_bitmap_1(bitmap_git, p) == 0)
356 struct bitmap_index *prepare_bitmap_git(struct repository *r)
358 struct bitmap_index *bitmap_git = xcalloc(1, sizeof(*bitmap_git));
360 if (!open_pack_bitmap(r, bitmap_git) && !load_pack_bitmap(bitmap_git))
363 free_bitmap_index(bitmap_git);
367 struct include_data {
368 struct bitmap_index *bitmap_git;
373 static inline int bitmap_position_extended(struct bitmap_index *bitmap_git,
374 const struct object_id *oid)
376 kh_oid_pos_t *positions = bitmap_git->ext_index.positions;
377 khiter_t pos = kh_get_oid_pos(positions, *oid);
379 if (pos < kh_end(positions)) {
380 int bitmap_pos = kh_value(positions, pos);
381 return bitmap_pos + bitmap_git->pack->num_objects;
387 static inline int bitmap_position_packfile(struct bitmap_index *bitmap_git,
388 const struct object_id *oid)
390 off_t offset = find_pack_entry_one(oid->hash, bitmap_git->pack);
394 return find_revindex_position(bitmap_git->pack, offset);
397 static int bitmap_position(struct bitmap_index *bitmap_git,
398 const struct object_id *oid)
400 int pos = bitmap_position_packfile(bitmap_git, oid);
401 return (pos >= 0) ? pos : bitmap_position_extended(bitmap_git, oid);
404 static int ext_index_add_object(struct bitmap_index *bitmap_git,
405 struct object *object, const char *name)
407 struct eindex *eindex = &bitmap_git->ext_index;
413 hash_pos = kh_put_oid_pos(eindex->positions, object->oid, &hash_ret);
415 if (eindex->count >= eindex->alloc) {
416 eindex->alloc = (eindex->alloc + 16) * 3 / 2;
417 REALLOC_ARRAY(eindex->objects, eindex->alloc);
418 REALLOC_ARRAY(eindex->hashes, eindex->alloc);
421 bitmap_pos = eindex->count;
422 eindex->objects[eindex->count] = object;
423 eindex->hashes[eindex->count] = pack_name_hash(name);
424 kh_value(eindex->positions, hash_pos) = bitmap_pos;
427 bitmap_pos = kh_value(eindex->positions, hash_pos);
430 return bitmap_pos + bitmap_git->pack->num_objects;
433 struct bitmap_show_data {
434 struct bitmap_index *bitmap_git;
438 static void show_object(struct object *object, const char *name, void *data_)
440 struct bitmap_show_data *data = data_;
443 bitmap_pos = bitmap_position(data->bitmap_git, &object->oid);
446 bitmap_pos = ext_index_add_object(data->bitmap_git, object,
449 bitmap_set(data->base, bitmap_pos);
452 static void show_commit(struct commit *commit, void *data)
456 static int add_to_include_set(struct bitmap_index *bitmap_git,
457 struct include_data *data,
458 const struct object_id *oid,
463 if (data->seen && bitmap_get(data->seen, bitmap_pos))
466 if (bitmap_get(data->base, bitmap_pos))
469 hash_pos = kh_get_oid_map(bitmap_git->bitmaps, *oid);
470 if (hash_pos < kh_end(bitmap_git->bitmaps)) {
471 struct stored_bitmap *st = kh_value(bitmap_git->bitmaps, hash_pos);
472 bitmap_or_ewah(data->base, lookup_stored_bitmap(st));
476 bitmap_set(data->base, bitmap_pos);
480 static int should_include(struct commit *commit, void *_data)
482 struct include_data *data = _data;
485 bitmap_pos = bitmap_position(data->bitmap_git, &commit->object.oid);
487 bitmap_pos = ext_index_add_object(data->bitmap_git,
488 (struct object *)commit,
491 if (!add_to_include_set(data->bitmap_git, data, &commit->object.oid,
493 struct commit_list *parent = commit->parents;
496 parent->item->object.flags |= SEEN;
497 parent = parent->next;
506 static struct bitmap *find_objects(struct bitmap_index *bitmap_git,
507 struct rev_info *revs,
508 struct object_list *roots,
511 struct bitmap *base = NULL;
514 struct object_list *not_mapped = NULL;
517 * Go through all the roots for the walk. The ones that have bitmaps
518 * on the bitmap index will be `or`ed together to form an initial
519 * global reachability analysis.
521 * The ones without bitmaps in the index will be stored in the
522 * `not_mapped_list` for further processing.
525 struct object *object = roots->item;
528 if (object->type == OBJ_COMMIT) {
529 khiter_t pos = kh_get_oid_map(bitmap_git->bitmaps, object->oid);
531 if (pos < kh_end(bitmap_git->bitmaps)) {
532 struct stored_bitmap *st = kh_value(bitmap_git->bitmaps, pos);
533 struct ewah_bitmap *or_with = lookup_stored_bitmap(st);
536 base = ewah_to_bitmap(or_with);
538 bitmap_or_ewah(base, or_with);
540 object->flags |= SEEN;
545 object_list_insert(object, ¬_mapped);
549 * Best case scenario: We found bitmaps for all the roots,
550 * so the resulting `or` bitmap has the full reachability analysis
552 if (not_mapped == NULL)
558 * Let's iterate through all the roots that don't have bitmaps to
559 * check if we can determine them to be reachable from the existing
562 * If we cannot find them in the existing global bitmap, we'll need
563 * to push them to an actual walk and run it until we can confirm
567 struct object *object = roots->item;
571 pos = bitmap_position(bitmap_git, &object->oid);
573 if (pos < 0 || base == NULL || !bitmap_get(base, pos)) {
574 object->flags &= ~UNINTERESTING;
575 add_pending_object(revs, object, "");
578 object->flags |= SEEN;
583 struct include_data incdata;
584 struct bitmap_show_data show_data;
589 incdata.bitmap_git = bitmap_git;
593 revs->include_check = should_include;
594 revs->include_check_data = &incdata;
596 if (prepare_revision_walk(revs))
597 die("revision walk setup failed");
599 show_data.bitmap_git = bitmap_git;
600 show_data.base = base;
602 traverse_commit_list(revs, show_commit, show_object,
609 static void show_extended_objects(struct bitmap_index *bitmap_git,
610 struct rev_info *revs,
611 show_reachable_fn show_reach)
613 struct bitmap *objects = bitmap_git->result;
614 struct eindex *eindex = &bitmap_git->ext_index;
617 for (i = 0; i < eindex->count; ++i) {
620 if (!bitmap_get(objects, bitmap_git->pack->num_objects + i))
623 obj = eindex->objects[i];
624 if ((obj->type == OBJ_BLOB && !revs->blob_objects) ||
625 (obj->type == OBJ_TREE && !revs->tree_objects) ||
626 (obj->type == OBJ_TAG && !revs->tag_objects))
629 show_reach(&obj->oid, obj->type, 0, eindex->hashes[i], NULL, 0);
633 static void init_type_iterator(struct ewah_iterator *it,
634 struct bitmap_index *bitmap_git,
635 enum object_type type)
639 ewah_iterator_init(it, bitmap_git->commits);
643 ewah_iterator_init(it, bitmap_git->trees);
647 ewah_iterator_init(it, bitmap_git->blobs);
651 ewah_iterator_init(it, bitmap_git->tags);
655 BUG("object type %d not stored by bitmap type index", type);
660 static void show_objects_for_type(
661 struct bitmap_index *bitmap_git,
662 enum object_type object_type,
663 show_reachable_fn show_reach)
668 struct ewah_iterator it;
671 struct bitmap *objects = bitmap_git->result;
673 init_type_iterator(&it, bitmap_git, object_type);
675 for (i = 0; i < objects->word_alloc &&
676 ewah_iterator_next(&filter, &it); i++) {
677 eword_t word = objects->words[i] & filter;
678 size_t pos = (i * BITS_IN_EWORD);
683 for (offset = 0; offset < BITS_IN_EWORD; ++offset) {
684 struct object_id oid;
685 struct revindex_entry *entry;
688 if ((word >> offset) == 0)
691 offset += ewah_bit_ctz64(word >> offset);
693 entry = &bitmap_git->pack->revindex[pos + offset];
694 nth_packed_object_id(&oid, bitmap_git->pack, entry->nr);
696 if (bitmap_git->hashes)
697 hash = get_be32(bitmap_git->hashes + entry->nr);
699 show_reach(&oid, object_type, 0, hash, bitmap_git->pack, entry->offset);
704 static int in_bitmapped_pack(struct bitmap_index *bitmap_git,
705 struct object_list *roots)
708 struct object *object = roots->item;
711 if (find_pack_entry_one(object->oid.hash, bitmap_git->pack) > 0)
718 static struct bitmap *find_tip_objects(struct bitmap_index *bitmap_git,
719 struct object_list *tip_objects,
720 enum object_type type)
722 struct bitmap *result = bitmap_new();
723 struct object_list *p;
725 for (p = tip_objects; p; p = p->next) {
728 if (p->item->type != type)
731 pos = bitmap_position(bitmap_git, &p->item->oid);
735 bitmap_set(result, pos);
741 static void filter_bitmap_exclude_type(struct bitmap_index *bitmap_git,
742 struct object_list *tip_objects,
743 struct bitmap *to_filter,
744 enum object_type type)
746 struct eindex *eindex = &bitmap_git->ext_index;
748 struct ewah_iterator it;
752 if (type != OBJ_BLOB && type != OBJ_TREE)
753 BUG("filter_bitmap_exclude_type: unsupported type '%d'", type);
756 * The non-bitmap version of this filter never removes
757 * objects which the other side specifically asked for,
758 * so we must match that behavior.
760 tips = find_tip_objects(bitmap_git, tip_objects, type);
763 * We can use the blob type-bitmap to work in whole words
764 * for the objects that are actually in the bitmapped packfile.
766 for (i = 0, init_type_iterator(&it, bitmap_git, type);
767 i < to_filter->word_alloc && ewah_iterator_next(&mask, &it);
769 if (i < tips->word_alloc)
770 mask &= ~tips->words[i];
771 to_filter->words[i] &= ~mask;
775 * Clear any blobs that weren't in the packfile (and so would not have
776 * been caught by the loop above. We'll have to check them
779 for (i = 0; i < eindex->count; i++) {
780 uint32_t pos = i + bitmap_git->pack->num_objects;
781 if (eindex->objects[i]->type == type &&
782 bitmap_get(to_filter, pos) &&
783 !bitmap_get(tips, pos))
784 bitmap_unset(to_filter, pos);
790 static void filter_bitmap_blob_none(struct bitmap_index *bitmap_git,
791 struct object_list *tip_objects,
792 struct bitmap *to_filter)
794 filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter,
798 static unsigned long get_size_by_pos(struct bitmap_index *bitmap_git,
801 struct packed_git *pack = bitmap_git->pack;
803 struct object_info oi = OBJECT_INFO_INIT;
807 if (pos < pack->num_objects) {
808 struct revindex_entry *entry = &pack->revindex[pos];
809 if (packed_object_info(the_repository, pack,
810 entry->offset, &oi) < 0) {
811 struct object_id oid;
812 nth_packed_object_id(&oid, pack, entry->nr);
813 die(_("unable to get size of %s"), oid_to_hex(&oid));
816 struct eindex *eindex = &bitmap_git->ext_index;
817 struct object *obj = eindex->objects[pos - pack->num_objects];
818 if (oid_object_info_extended(the_repository, &obj->oid, &oi, 0) < 0)
819 die(_("unable to get size of %s"), oid_to_hex(&obj->oid));
825 static void filter_bitmap_blob_limit(struct bitmap_index *bitmap_git,
826 struct object_list *tip_objects,
827 struct bitmap *to_filter,
830 struct eindex *eindex = &bitmap_git->ext_index;
832 struct ewah_iterator it;
836 tips = find_tip_objects(bitmap_git, tip_objects, OBJ_BLOB);
838 for (i = 0, init_type_iterator(&it, bitmap_git, OBJ_BLOB);
839 i < to_filter->word_alloc && ewah_iterator_next(&mask, &it);
841 eword_t word = to_filter->words[i] & mask;
844 for (offset = 0; offset < BITS_IN_EWORD; offset++) {
847 if ((word >> offset) == 0)
849 offset += ewah_bit_ctz64(word >> offset);
850 pos = i * BITS_IN_EWORD + offset;
852 if (!bitmap_get(tips, pos) &&
853 get_size_by_pos(bitmap_git, pos) >= limit)
854 bitmap_unset(to_filter, pos);
858 for (i = 0; i < eindex->count; i++) {
859 uint32_t pos = i + bitmap_git->pack->num_objects;
860 if (eindex->objects[i]->type == OBJ_BLOB &&
861 bitmap_get(to_filter, pos) &&
862 !bitmap_get(tips, pos) &&
863 get_size_by_pos(bitmap_git, pos) >= limit)
864 bitmap_unset(to_filter, pos);
870 static void filter_bitmap_tree_depth(struct bitmap_index *bitmap_git,
871 struct object_list *tip_objects,
872 struct bitmap *to_filter,
876 BUG("filter_bitmap_tree_depth given non-zero limit");
878 filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter,
880 filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter,
884 static int filter_bitmap(struct bitmap_index *bitmap_git,
885 struct object_list *tip_objects,
886 struct bitmap *to_filter,
887 struct list_objects_filter_options *filter)
889 if (!filter || filter->choice == LOFC_DISABLED)
892 if (filter->choice == LOFC_BLOB_NONE) {
894 filter_bitmap_blob_none(bitmap_git, tip_objects,
899 if (filter->choice == LOFC_BLOB_LIMIT) {
901 filter_bitmap_blob_limit(bitmap_git, tip_objects,
903 filter->blob_limit_value);
907 if (filter->choice == LOFC_TREE_DEPTH &&
908 filter->tree_exclude_depth == 0) {
910 filter_bitmap_tree_depth(bitmap_git, tip_objects,
912 filter->tree_exclude_depth);
916 /* filter choice not handled */
920 static int can_filter_bitmap(struct list_objects_filter_options *filter)
922 return !filter_bitmap(NULL, NULL, NULL, filter);
925 struct bitmap_index *prepare_bitmap_walk(struct rev_info *revs,
926 struct list_objects_filter_options *filter)
930 struct object_list *wants = NULL;
931 struct object_list *haves = NULL;
933 struct bitmap *wants_bitmap = NULL;
934 struct bitmap *haves_bitmap = NULL;
936 struct bitmap_index *bitmap_git;
939 * We can't do pathspec limiting with bitmaps, because we don't know
940 * which commits are associated with which object changes (let alone
941 * even which objects are associated with which paths).
946 if (!can_filter_bitmap(filter))
949 /* try to open a bitmapped pack, but don't parse it yet
950 * because we may not need to use it */
951 bitmap_git = xcalloc(1, sizeof(*bitmap_git));
952 if (open_pack_bitmap(revs->repo, bitmap_git) < 0)
955 for (i = 0; i < revs->pending.nr; ++i) {
956 struct object *object = revs->pending.objects[i].item;
958 if (object->type == OBJ_NONE)
959 parse_object_or_die(&object->oid, NULL);
961 while (object->type == OBJ_TAG) {
962 struct tag *tag = (struct tag *) object;
964 if (object->flags & UNINTERESTING)
965 object_list_insert(object, &haves);
967 object_list_insert(object, &wants);
969 object = parse_object_or_die(get_tagged_oid(tag), NULL);
972 if (object->flags & UNINTERESTING)
973 object_list_insert(object, &haves);
975 object_list_insert(object, &wants);
979 * if we have a HAVES list, but none of those haves is contained
980 * in the packfile that has a bitmap, we don't have anything to
983 if (haves && !in_bitmapped_pack(bitmap_git, haves))
986 /* if we don't want anything, we're done here */
991 * now we're going to use bitmaps, so load the actual bitmap entries
992 * from disk. this is the point of no return; after this the rev_list
993 * becomes invalidated and we must perform the revwalk through bitmaps
995 if (load_pack_bitmap(bitmap_git) < 0)
998 object_array_clear(&revs->pending);
1001 revs->ignore_missing_links = 1;
1002 haves_bitmap = find_objects(bitmap_git, revs, haves, NULL);
1003 reset_revision_walk();
1004 revs->ignore_missing_links = 0;
1006 if (haves_bitmap == NULL)
1007 BUG("failed to perform bitmap walk");
1010 wants_bitmap = find_objects(bitmap_git, revs, wants, haves_bitmap);
1013 BUG("failed to perform bitmap walk");
1016 bitmap_and_not(wants_bitmap, haves_bitmap);
1018 filter_bitmap(bitmap_git, wants, wants_bitmap, filter);
1020 bitmap_git->result = wants_bitmap;
1021 bitmap_git->haves = haves_bitmap;
1023 object_list_free(&wants);
1024 object_list_free(&haves);
1029 free_bitmap_index(bitmap_git);
1030 object_list_free(&wants);
1031 object_list_free(&haves);
1035 static void try_partial_reuse(struct bitmap_index *bitmap_git,
1037 struct bitmap *reuse,
1038 struct pack_window **w_curs)
1040 struct revindex_entry *revidx;
1042 enum object_type type;
1045 if (pos >= bitmap_git->pack->num_objects)
1046 return; /* not actually in the pack */
1048 revidx = &bitmap_git->pack->revindex[pos];
1049 offset = revidx->offset;
1050 type = unpack_object_header(bitmap_git->pack, w_curs, &offset, &size);
1052 return; /* broken packfile, punt */
1054 if (type == OBJ_REF_DELTA || type == OBJ_OFS_DELTA) {
1059 * Find the position of the base object so we can look it up
1060 * in our bitmaps. If we can't come up with an offset, or if
1061 * that offset is not in the revidx, the pack is corrupt.
1062 * There's nothing we can do, so just punt on this object,
1063 * and the normal slow path will complain about it in
1066 base_offset = get_delta_base(bitmap_git->pack, w_curs,
1067 &offset, type, revidx->offset);
1070 base_pos = find_revindex_position(bitmap_git->pack, base_offset);
1075 * We assume delta dependencies always point backwards. This
1076 * lets us do a single pass, and is basically always true
1077 * due to the way OFS_DELTAs work. You would not typically
1078 * find REF_DELTA in a bitmapped pack, since we only bitmap
1079 * packs we write fresh, and OFS_DELTA is the default). But
1080 * let's double check to make sure the pack wasn't written with
1083 if (base_pos >= pos)
1087 * And finally, if we're not sending the base as part of our
1088 * reuse chunk, then don't send this object either. The base
1089 * would come after us, along with other objects not
1090 * necessarily in the pack, which means we'd need to convert
1091 * to REF_DELTA on the fly. Better to just let the normal
1092 * object_entry code path handle it.
1094 if (!bitmap_get(reuse, base_pos))
1099 * If we got here, then the object is OK to reuse. Mark it.
1101 bitmap_set(reuse, pos);
1104 int reuse_partial_packfile_from_bitmap(struct bitmap_index *bitmap_git,
1105 struct packed_git **packfile_out,
1107 struct bitmap **reuse_out)
1109 struct bitmap *result = bitmap_git->result;
1110 struct bitmap *reuse;
1111 struct pack_window *w_curs = NULL;
1117 while (i < result->word_alloc && result->words[i] == (eword_t)~0)
1120 /* Don't mark objects not in the packfile */
1121 if (i > bitmap_git->pack->num_objects / BITS_IN_EWORD)
1122 i = bitmap_git->pack->num_objects / BITS_IN_EWORD;
1124 reuse = bitmap_word_alloc(i);
1125 memset(reuse->words, 0xFF, i * sizeof(eword_t));
1127 for (; i < result->word_alloc; ++i) {
1128 eword_t word = result->words[i];
1129 size_t pos = (i * BITS_IN_EWORD);
1131 for (offset = 0; offset < BITS_IN_EWORD; ++offset) {
1132 if ((word >> offset) == 0)
1135 offset += ewah_bit_ctz64(word >> offset);
1136 try_partial_reuse(bitmap_git, pos + offset, reuse, &w_curs);
1140 unuse_pack(&w_curs);
1142 *entries = bitmap_popcount(reuse);
1149 * Drop any reused objects from the result, since they will not
1150 * need to be handled separately.
1152 bitmap_and_not(result, reuse);
1153 *packfile_out = bitmap_git->pack;
1158 int bitmap_walk_contains(struct bitmap_index *bitmap_git,
1159 struct bitmap *bitmap, const struct object_id *oid)
1166 idx = bitmap_position(bitmap_git, oid);
1167 return idx >= 0 && bitmap_get(bitmap, idx);
1170 void traverse_bitmap_commit_list(struct bitmap_index *bitmap_git,
1171 struct rev_info *revs,
1172 show_reachable_fn show_reachable)
1174 assert(bitmap_git->result);
1176 show_objects_for_type(bitmap_git, OBJ_COMMIT, show_reachable);
1177 if (revs->tree_objects)
1178 show_objects_for_type(bitmap_git, OBJ_TREE, show_reachable);
1179 if (revs->blob_objects)
1180 show_objects_for_type(bitmap_git, OBJ_BLOB, show_reachable);
1181 if (revs->tag_objects)
1182 show_objects_for_type(bitmap_git, OBJ_TAG, show_reachable);
1184 show_extended_objects(bitmap_git, revs, show_reachable);
1187 static uint32_t count_object_type(struct bitmap_index *bitmap_git,
1188 enum object_type type)
1190 struct bitmap *objects = bitmap_git->result;
1191 struct eindex *eindex = &bitmap_git->ext_index;
1193 uint32_t i = 0, count = 0;
1194 struct ewah_iterator it;
1197 init_type_iterator(&it, bitmap_git, type);
1199 while (i < objects->word_alloc && ewah_iterator_next(&filter, &it)) {
1200 eword_t word = objects->words[i++] & filter;
1201 count += ewah_bit_popcount64(word);
1204 for (i = 0; i < eindex->count; ++i) {
1205 if (eindex->objects[i]->type == type &&
1206 bitmap_get(objects, bitmap_git->pack->num_objects + i))
1213 void count_bitmap_commit_list(struct bitmap_index *bitmap_git,
1214 uint32_t *commits, uint32_t *trees,
1215 uint32_t *blobs, uint32_t *tags)
1217 assert(bitmap_git->result);
1220 *commits = count_object_type(bitmap_git, OBJ_COMMIT);
1223 *trees = count_object_type(bitmap_git, OBJ_TREE);
1226 *blobs = count_object_type(bitmap_git, OBJ_BLOB);
1229 *tags = count_object_type(bitmap_git, OBJ_TAG);
1232 struct bitmap_test_data {
1233 struct bitmap_index *bitmap_git;
1234 struct bitmap *base;
1235 struct progress *prg;
1239 static void test_show_object(struct object *object, const char *name,
1242 struct bitmap_test_data *tdata = data;
1245 bitmap_pos = bitmap_position(tdata->bitmap_git, &object->oid);
1247 die("Object not in bitmap: %s\n", oid_to_hex(&object->oid));
1249 bitmap_set(tdata->base, bitmap_pos);
1250 display_progress(tdata->prg, ++tdata->seen);
1253 static void test_show_commit(struct commit *commit, void *data)
1255 struct bitmap_test_data *tdata = data;
1258 bitmap_pos = bitmap_position(tdata->bitmap_git,
1259 &commit->object.oid);
1261 die("Object not in bitmap: %s\n", oid_to_hex(&commit->object.oid));
1263 bitmap_set(tdata->base, bitmap_pos);
1264 display_progress(tdata->prg, ++tdata->seen);
1267 void test_bitmap_walk(struct rev_info *revs)
1269 struct object *root;
1270 struct bitmap *result = NULL;
1272 size_t result_popcnt;
1273 struct bitmap_test_data tdata;
1274 struct bitmap_index *bitmap_git;
1276 if (!(bitmap_git = prepare_bitmap_git(revs->repo)))
1277 die("failed to load bitmap indexes");
1279 if (revs->pending.nr != 1)
1280 die("you must specify exactly one commit to test");
1282 fprintf(stderr, "Bitmap v%d test (%d entries loaded)\n",
1283 bitmap_git->version, bitmap_git->entry_count);
1285 root = revs->pending.objects[0].item;
1286 pos = kh_get_oid_map(bitmap_git->bitmaps, root->oid);
1288 if (pos < kh_end(bitmap_git->bitmaps)) {
1289 struct stored_bitmap *st = kh_value(bitmap_git->bitmaps, pos);
1290 struct ewah_bitmap *bm = lookup_stored_bitmap(st);
1292 fprintf(stderr, "Found bitmap for %s. %d bits / %08x checksum\n",
1293 oid_to_hex(&root->oid), (int)bm->bit_size, ewah_checksum(bm));
1295 result = ewah_to_bitmap(bm);
1299 die("Commit %s doesn't have an indexed bitmap", oid_to_hex(&root->oid));
1301 revs->tag_objects = 1;
1302 revs->tree_objects = 1;
1303 revs->blob_objects = 1;
1305 result_popcnt = bitmap_popcount(result);
1307 if (prepare_revision_walk(revs))
1308 die("revision walk setup failed");
1310 tdata.bitmap_git = bitmap_git;
1311 tdata.base = bitmap_new();
1312 tdata.prg = start_progress("Verifying bitmap entries", result_popcnt);
1315 traverse_commit_list(revs, &test_show_commit, &test_show_object, &tdata);
1317 stop_progress(&tdata.prg);
1319 if (bitmap_equals(result, tdata.base))
1320 fprintf(stderr, "OK!\n");
1322 fprintf(stderr, "Mismatch!\n");
1324 free_bitmap_index(bitmap_git);
1327 static int rebuild_bitmap(uint32_t *reposition,
1328 struct ewah_bitmap *source,
1329 struct bitmap *dest)
1332 struct ewah_iterator it;
1335 ewah_iterator_init(&it, source);
1337 while (ewah_iterator_next(&word, &it)) {
1338 uint32_t offset, bit_pos;
1340 for (offset = 0; offset < BITS_IN_EWORD; ++offset) {
1341 if ((word >> offset) == 0)
1344 offset += ewah_bit_ctz64(word >> offset);
1346 bit_pos = reposition[pos + offset];
1348 bitmap_set(dest, bit_pos - 1);
1349 else /* can't reuse, we don't have the object */
1353 pos += BITS_IN_EWORD;
1358 int rebuild_existing_bitmaps(struct bitmap_index *bitmap_git,
1359 struct packing_data *mapping,
1360 kh_oid_map_t *reused_bitmaps,
1363 uint32_t i, num_objects;
1364 uint32_t *reposition;
1365 struct bitmap *rebuild;
1366 struct stored_bitmap *stored;
1367 struct progress *progress = NULL;
1372 num_objects = bitmap_git->pack->num_objects;
1373 reposition = xcalloc(num_objects, sizeof(uint32_t));
1375 for (i = 0; i < num_objects; ++i) {
1376 struct object_id oid;
1377 struct revindex_entry *entry;
1378 struct object_entry *oe;
1380 entry = &bitmap_git->pack->revindex[i];
1381 nth_packed_object_id(&oid, bitmap_git->pack, entry->nr);
1382 oe = packlist_find(mapping, &oid);
1385 reposition[i] = oe_in_pack_pos(mapping, oe) + 1;
1388 rebuild = bitmap_new();
1392 progress = start_progress("Reusing bitmaps", 0);
1394 kh_foreach_value(bitmap_git->bitmaps, stored, {
1395 if (stored->flags & BITMAP_FLAG_REUSE) {
1396 if (!rebuild_bitmap(reposition,
1397 lookup_stored_bitmap(stored),
1399 hash_pos = kh_put_oid_map(reused_bitmaps,
1402 kh_value(reused_bitmaps, hash_pos) =
1403 bitmap_to_ewah(rebuild);
1405 bitmap_reset(rebuild);
1406 display_progress(progress, ++i);
1410 stop_progress(&progress);
1413 bitmap_free(rebuild);
1417 void free_bitmap_index(struct bitmap_index *b)
1423 munmap(b->map, b->map_size);
1424 ewah_pool_free(b->commits);
1425 ewah_pool_free(b->trees);
1426 ewah_pool_free(b->blobs);
1427 ewah_pool_free(b->tags);
1428 kh_destroy_oid_map(b->bitmaps);
1429 free(b->ext_index.objects);
1430 free(b->ext_index.hashes);
1431 bitmap_free(b->result);
1432 bitmap_free(b->haves);
1436 int bitmap_has_oid_in_uninteresting(struct bitmap_index *bitmap_git,
1437 const struct object_id *oid)
1439 return bitmap_git &&
1440 bitmap_walk_contains(bitmap_git, bitmap_git->haves, oid);