?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
upload-pack: skip parse-object re-hashing of "want" objects
[git/debian.git] / pack-bitmap.c
blob9a208abc1fdd920926a1b8ee22e0a8bf5371743e
1 #include "cache.h"
2 #include "commit.h"
3 #include "strbuf.h"
4 #include "tag.h"
5 #include "diff.h"
6 #include "revision.h"
7 #include "progress.h"
8 #include "list-objects.h"
9 #include "pack.h"
10 #include "pack-bitmap.h"
11 #include "pack-revindex.h"
12 #include "pack-objects.h"
13 #include "packfile.h"
14 #include "repository.h"
15 #include "object-store.h"
16 #include "list-objects-filter-options.h"
17 #include "midx.h"
18 #include "config.h"
19
20 /*
21 * An entry on the bitmap index, representing the bitmap for a given
22 * commit.
23 */
24 struct stored_bitmap {
25 struct object_id oid;
26 struct ewah_bitmap *root;
27 struct stored_bitmap *xor;
28 int flags;
29 };
30
31 /*
32 * The active bitmap index for a repository. By design, repositories only have
33 * a single bitmap index available (the index for the biggest packfile in
34 * the repository), since bitmap indexes need full closure.
35 *
36 * If there is more than one bitmap index available (e.g. because of alternates),
37 * the active bitmap index is the largest one.
38 */
39 struct bitmap_index {
40 /*
41 * The pack or multi-pack index (MIDX) that this bitmap index belongs
42 * to.
43 *
44 * Exactly one of these must be non-NULL; this specifies the object
45 * order used to interpret this bitmap.
46 */
47 struct packed_git *pack;
48 struct multi_pack_index *midx;
49
50 /*
51 * Mark the first `reuse_objects` in the packfile as reused:
52 * they will be sent as-is without using them for repacking
53 * calculations
54 */
55 uint32_t reuse_objects;
56
57 /* mmapped buffer of the whole bitmap index */
58 unsigned char *map;
59 size_t map_size; /* size of the mmaped buffer */
60 size_t map_pos; /* current position when loading the index */
61
62 /*
63 * Type indexes.
64 *
65 * Each bitmap marks which objects in the packfile are of the given
66 * type. This provides type information when yielding the objects from
67 * the packfile during a walk, which allows for better delta bases.
68 */
69 struct ewah_bitmap *commits;
70 struct ewah_bitmap *trees;
71 struct ewah_bitmap *blobs;
72 struct ewah_bitmap *tags;
73
74 /* Map from object ID -> `stored_bitmap` for all the bitmapped commits */
75 kh_oid_map_t *bitmaps;
76
77 /* Number of bitmapped commits */
78 uint32_t entry_count;
79
80 /* If not NULL, this is a name-hash cache pointing into map. */
81 uint32_t *hashes;
82
83 /* The checksum of the packfile or MIDX; points into map. */
84 const unsigned char *checksum;
85
86 /*
87 * If not NULL, this point into the commit table extension
88 * (within the memory mapped region `map`).
89 */
90 unsigned char *table_lookup;
91
92 /*
93 * Extended index.
94 *
95 * When trying to perform bitmap operations with objects that are not
96 * packed in `pack`, these objects are added to this "fake index" and
97 * are assumed to appear at the end of the packfile for all operations
98 */
99 struct eindex {
100 struct object **objects;
101 uint32_t *hashes;
102 uint32_t count, alloc;
103 kh_oid_pos_t *positions;
104 } ext_index;
105
106 /* Bitmap result of the last performed walk */
107 struct bitmap *result;
108
109 /* "have" bitmap from the last performed walk */
110 struct bitmap *haves;
111
112 /* Version of the bitmap index */
113 unsigned int version;
114 };
115
116 static struct ewah_bitmap *lookup_stored_bitmap(struct stored_bitmap *st)
117 {
118 struct ewah_bitmap *parent;
119 struct ewah_bitmap *composed;
120
121 if (!st->xor)
122 return st->root;
123
124 composed = ewah_pool_new();
125 parent = lookup_stored_bitmap(st->xor);
126 ewah_xor(st->root, parent, composed);
127
128 ewah_pool_free(st->root);
129 st->root = composed;
130 st->xor = NULL;
131
132 return composed;
133 }
134
135 /*
136 * Read a bitmap from the current read position on the mmaped
137 * index, and increase the read position accordingly
138 */
139 static struct ewah_bitmap *read_bitmap_1(struct bitmap_index *index)
140 {
141 struct ewah_bitmap *b = ewah_pool_new();
142
143 ssize_t bitmap_size = ewah_read_mmap(b,
144 index->map + index->map_pos,
145 index->map_size - index->map_pos);
146
147 if (bitmap_size < 0) {
148 error(_("failed to load bitmap index (corrupted?)"));
149 ewah_pool_free(b);
150 return NULL;
151 }
152
153 index->map_pos += bitmap_size;
154 return b;
155 }
156
157 static uint32_t bitmap_num_objects(struct bitmap_index *index)
158 {
159 if (index->midx)
160 return index->midx->num_objects;
161 return index->pack->num_objects;
162 }
163
164 static int load_bitmap_header(struct bitmap_index *index)
165 {
166 struct bitmap_disk_header *header = (void *)index->map;
167 size_t header_size = sizeof(*header) - GIT_MAX_RAWSZ + the_hash_algo->rawsz;
168
169 if (index->map_size < header_size + the_hash_algo->rawsz)
170 return error(_("corrupted bitmap index (too small)"));
171
172 if (memcmp(header->magic, BITMAP_IDX_SIGNATURE, sizeof(BITMAP_IDX_SIGNATURE)) != 0)
173 return error(_("corrupted bitmap index file (wrong header)"));
174
175 index->version = ntohs(header->version);
176 if (index->version != 1)
177 return error(_("unsupported version '%d' for bitmap index file"), index->version);
178
179 /* Parse known bitmap format options */
180 {
181 uint32_t flags = ntohs(header->options);
182 size_t cache_size = st_mult(bitmap_num_objects(index), sizeof(uint32_t));
183 unsigned char *index_end = index->map + index->map_size - the_hash_algo->rawsz;
184
185 if ((flags & BITMAP_OPT_FULL_DAG) == 0)
186 BUG("unsupported options for bitmap index file "
187 "(Git requires BITMAP_OPT_FULL_DAG)");
188
189 if (flags & BITMAP_OPT_HASH_CACHE) {
190 if (cache_size > index_end - index->map - header_size)
191 return error(_("corrupted bitmap index file (too short to fit hash cache)"));
192 index->hashes = (void *)(index_end - cache_size);
193 index_end -= cache_size;
194 }
195
196 if (flags & BITMAP_OPT_LOOKUP_TABLE) {
197 size_t table_size = st_mult(ntohl(header->entry_count),
198 BITMAP_LOOKUP_TABLE_TRIPLET_WIDTH);
199 if (table_size > index_end - index->map - header_size)
200 return error(_("corrupted bitmap index file (too short to fit lookup table)"));
201 if (git_env_bool("GIT_TEST_READ_COMMIT_TABLE", 1))
202 index->table_lookup = (void *)(index_end - table_size);
203 index_end -= table_size;
204 }
205 }
206
207 index->entry_count = ntohl(header->entry_count);
208 index->checksum = header->checksum;
209 index->map_pos += header_size;
210 return 0;
211 }
212
213 static struct stored_bitmap *store_bitmap(struct bitmap_index *index,
214 struct ewah_bitmap *root,
215 const struct object_id *oid,
216 struct stored_bitmap *xor_with,
217 int flags)
218 {
219 struct stored_bitmap *stored;
220 khiter_t hash_pos;
221 int ret;
222
223 stored = xmalloc(sizeof(struct stored_bitmap));
224 stored->root = root;
225 stored->xor = xor_with;
226 stored->flags = flags;
227 oidcpy(&stored->oid, oid);
228
229 hash_pos = kh_put_oid_map(index->bitmaps, stored->oid, &ret);
230
231 /*
232 * A 0 return code means the insertion succeeded with no changes,
233 * because the SHA1 already existed on the map. This is bad, there
234 * shouldn't be duplicated commits in the index.
235 */
236 if (ret == 0) {
237 error(_("duplicate entry in bitmap index: '%s'"), oid_to_hex(oid));
238 return NULL;
239 }
240
241 kh_value(index->bitmaps, hash_pos) = stored;
242 return stored;
243 }
244
245 static inline uint32_t read_be32(const unsigned char *buffer, size_t *pos)
246 {
247 uint32_t result = get_be32(buffer + *pos);
248 (*pos) += sizeof(result);
249 return result;
250 }
251
252 static inline uint8_t read_u8(const unsigned char *buffer, size_t *pos)
253 {
254 return buffer[(*pos)++];
255 }
256
257 #define MAX_XOR_OFFSET 160
258
259 static int nth_bitmap_object_oid(struct bitmap_index *index,
260 struct object_id *oid,
261 uint32_t n)
262 {
263 if (index->midx)
264 return nth_midxed_object_oid(oid, index->midx, n) ? 0 : -1;
265 return nth_packed_object_id(oid, index->pack, n);
266 }
267
268 static int load_bitmap_entries_v1(struct bitmap_index *index)
269 {
270 uint32_t i;
271 struct stored_bitmap *recent_bitmaps[MAX_XOR_OFFSET] = { NULL };
272
273 for (i = 0; i < index->entry_count; ++i) {
274 int xor_offset, flags;
275 struct ewah_bitmap *bitmap = NULL;
276 struct stored_bitmap *xor_bitmap = NULL;
277 uint32_t commit_idx_pos;
278 struct object_id oid;
279
280 if (index->map_size - index->map_pos < 6)
281 return error(_("corrupt ewah bitmap: truncated header for entry %d"), i);
282
283 commit_idx_pos = read_be32(index->map, &index->map_pos);
284 xor_offset = read_u8(index->map, &index->map_pos);
285 flags = read_u8(index->map, &index->map_pos);
286
287 if (nth_bitmap_object_oid(index, &oid, commit_idx_pos) < 0)
288 return error(_("corrupt ewah bitmap: commit index %u out of range"),
289 (unsigned)commit_idx_pos);
290
291 bitmap = read_bitmap_1(index);
292 if (!bitmap)
293 return -1;
294
295 if (xor_offset > MAX_XOR_OFFSET || xor_offset > i)
296 return error(_("corrupted bitmap pack index"));
297
298 if (xor_offset > 0) {
299 xor_bitmap = recent_bitmaps[(i - xor_offset) % MAX_XOR_OFFSET];
300
301 if (!xor_bitmap)
302 return error(_("invalid XOR offset in bitmap pack index"));
303 }
304
305 recent_bitmaps[i % MAX_XOR_OFFSET] = store_bitmap(
306 index, bitmap, &oid, xor_bitmap, flags);
307 }
308
309 return 0;
310 }
311
312 char *midx_bitmap_filename(struct multi_pack_index *midx)
313 {
314 struct strbuf buf = STRBUF_INIT;
315
316 get_midx_filename(&buf, midx->object_dir);
317 strbuf_addf(&buf, "-%s.bitmap", hash_to_hex(get_midx_checksum(midx)));
318
319 return strbuf_detach(&buf, NULL);
320 }
321
322 char *pack_bitmap_filename(struct packed_git *p)
323 {
324 size_t len;
325
326 if (!strip_suffix(p->pack_name, ".pack", &len))
327 BUG("pack_name does not end in .pack");
328 return xstrfmt("%.*s.bitmap", (int)len, p->pack_name);
329 }
330
331 static int open_midx_bitmap_1(struct bitmap_index *bitmap_git,
332 struct multi_pack_index *midx)
333 {
334 struct stat st;
335 char *bitmap_name = midx_bitmap_filename(midx);
336 int fd = git_open(bitmap_name);
337 uint32_t i;
338 struct packed_git *preferred;
339
340 if (fd < 0) {
341 if (errno != ENOENT)
342 warning_errno("cannot open '%s'", bitmap_name);
343 free(bitmap_name);
344 return -1;
345 }
346 free(bitmap_name);
347
348 if (fstat(fd, &st)) {
349 error_errno(_("cannot fstat bitmap file"));
350 close(fd);
351 return -1;
352 }
353
354 if (bitmap_git->pack || bitmap_git->midx) {
355 struct strbuf buf = STRBUF_INIT;
356 get_midx_filename(&buf, midx->object_dir);
357 /* ignore extra bitmap file; we can only handle one */
358 warning(_("ignoring extra bitmap file: '%s'"), buf.buf);
359 close(fd);
360 strbuf_release(&buf);
361 return -1;
362 }
363
364 bitmap_git->midx = midx;
365 bitmap_git->map_size = xsize_t(st.st_size);
366 bitmap_git->map_pos = 0;
367 bitmap_git->map = xmmap(NULL, bitmap_git->map_size, PROT_READ,
368 MAP_PRIVATE, fd, 0);
369 close(fd);
370
371 if (load_bitmap_header(bitmap_git) < 0)
372 goto cleanup;
373
374 if (!hasheq(get_midx_checksum(bitmap_git->midx), bitmap_git->checksum)) {
375 error(_("checksum doesn't match in MIDX and bitmap"));
376 goto cleanup;
377 }
378
379 if (load_midx_revindex(bitmap_git->midx) < 0) {
380 warning(_("multi-pack bitmap is missing required reverse index"));
381 goto cleanup;
382 }
383
384 for (i = 0; i < bitmap_git->midx->num_packs; i++) {
385 if (prepare_midx_pack(the_repository, bitmap_git->midx, i))
386 die(_("could not open pack %s"),
387 bitmap_git->midx->pack_names[i]);
388 }
389
390 preferred = bitmap_git->midx->packs[midx_preferred_pack(bitmap_git)];
391 if (!is_pack_valid(preferred)) {
392 warning(_("preferred pack (%s) is invalid"),
393 preferred->pack_name);
394 goto cleanup;
395 }
396
397 return 0;
398
399 cleanup:
400 munmap(bitmap_git->map, bitmap_git->map_size);
401 bitmap_git->map_size = 0;
402 bitmap_git->map_pos = 0;
403 bitmap_git->map = NULL;
404 bitmap_git->midx = NULL;
405 return -1;
406 }
407
408 static int open_pack_bitmap_1(struct bitmap_index *bitmap_git, struct packed_git *packfile)
409 {
410 int fd;
411 struct stat st;
412 char *bitmap_name;
413
414 if (open_pack_index(packfile))
415 return -1;
416
417 bitmap_name = pack_bitmap_filename(packfile);
418 fd = git_open(bitmap_name);
419
420 if (fd < 0) {
421 if (errno != ENOENT)
422 warning_errno("cannot open '%s'", bitmap_name);
423 free(bitmap_name);
424 return -1;
425 }
426 free(bitmap_name);
427
428 if (fstat(fd, &st)) {
429 error_errno(_("cannot fstat bitmap file"));
430 close(fd);
431 return -1;
432 }
433
434 if (bitmap_git->pack || bitmap_git->midx) {
435 /* ignore extra bitmap file; we can only handle one */
436 warning(_("ignoring extra bitmap file: '%s'"), packfile->pack_name);
437 close(fd);
438 return -1;
439 }
440
441 if (!is_pack_valid(packfile)) {
442 close(fd);
443 return -1;
444 }
445
446 bitmap_git->pack = packfile;
447 bitmap_git->map_size = xsize_t(st.st_size);
448 bitmap_git->map = xmmap(NULL, bitmap_git->map_size, PROT_READ, MAP_PRIVATE, fd, 0);
449 bitmap_git->map_pos = 0;
450 close(fd);
451
452 if (load_bitmap_header(bitmap_git) < 0) {
453 munmap(bitmap_git->map, bitmap_git->map_size);
454 bitmap_git->map = NULL;
455 bitmap_git->map_size = 0;
456 bitmap_git->map_pos = 0;
457 bitmap_git->pack = NULL;
458 return -1;
459 }
460
461 return 0;
462 }
463
464 static int load_reverse_index(struct bitmap_index *bitmap_git)
465 {
466 if (bitmap_is_midx(bitmap_git)) {
467 uint32_t i;
468 int ret;
469
470 /*
471 * The multi-pack-index's .rev file is already loaded via
472 * open_pack_bitmap_1().
473 *
474 * But we still need to open the individual pack .rev files,
475 * since we will need to make use of them in pack-objects.
476 */
477 for (i = 0; i < bitmap_git->midx->num_packs; i++) {
478 ret = load_pack_revindex(bitmap_git->midx->packs[i]);
479 if (ret)
480 return ret;
481 }
482 return 0;
483 }
484 return load_pack_revindex(bitmap_git->pack);
485 }
486
487 static int load_bitmap(struct bitmap_index *bitmap_git)
488 {
489 assert(bitmap_git->map);
490
491 bitmap_git->bitmaps = kh_init_oid_map();
492 bitmap_git->ext_index.positions = kh_init_oid_pos();
493
494 if (load_reverse_index(bitmap_git))
495 goto failed;
496
497 if (!(bitmap_git->commits = read_bitmap_1(bitmap_git)) ||
498 !(bitmap_git->trees = read_bitmap_1(bitmap_git)) ||
499 !(bitmap_git->blobs = read_bitmap_1(bitmap_git)) ||
500 !(bitmap_git->tags = read_bitmap_1(bitmap_git)))
501 goto failed;
502
503 if (!bitmap_git->table_lookup && load_bitmap_entries_v1(bitmap_git) < 0)
504 goto failed;
505
506 return 0;
507
508 failed:
509 munmap(bitmap_git->map, bitmap_git->map_size);
510 bitmap_git->map = NULL;
511 bitmap_git->map_size = 0;
512
513 kh_destroy_oid_map(bitmap_git->bitmaps);
514 bitmap_git->bitmaps = NULL;
515
516 kh_destroy_oid_pos(bitmap_git->ext_index.positions);
517 bitmap_git->ext_index.positions = NULL;
518
519 return -1;
520 }
521
522 static int open_pack_bitmap(struct repository *r,
523 struct bitmap_index *bitmap_git)
524 {
525 struct packed_git *p;
526 int ret = -1;
527
528 assert(!bitmap_git->map);
529
530 for (p = get_all_packs(r); p; p = p->next) {
531 if (open_pack_bitmap_1(bitmap_git, p) == 0)
532 ret = 0;
533 }
534
535 return ret;
536 }
537
538 static int open_midx_bitmap(struct repository *r,
539 struct bitmap_index *bitmap_git)
540 {
541 int ret = -1;
542 struct multi_pack_index *midx;
543
544 assert(!bitmap_git->map);
545
546 for (midx = get_multi_pack_index(r); midx; midx = midx->next) {
547 if (!open_midx_bitmap_1(bitmap_git, midx))
548 ret = 0;
549 }
550 return ret;
551 }
552
553 static int open_bitmap(struct repository *r,
554 struct bitmap_index *bitmap_git)
555 {
556 assert(!bitmap_git->map);
557
558 if (!open_midx_bitmap(r, bitmap_git))
559 return 0;
560 return open_pack_bitmap(r, bitmap_git);
561 }
562
563 struct bitmap_index *prepare_bitmap_git(struct repository *r)
564 {
565 struct bitmap_index *bitmap_git = xcalloc(1, sizeof(*bitmap_git));
566
567 if (!open_bitmap(r, bitmap_git) && !load_bitmap(bitmap_git))
568 return bitmap_git;
569
570 free_bitmap_index(bitmap_git);
571 return NULL;
572 }
573
574 struct bitmap_index *prepare_midx_bitmap_git(struct multi_pack_index *midx)
575 {
576 struct bitmap_index *bitmap_git = xcalloc(1, sizeof(*bitmap_git));
577
578 if (!open_midx_bitmap_1(bitmap_git, midx) && !load_bitmap(bitmap_git))
579 return bitmap_git;
580
581 free_bitmap_index(bitmap_git);
582 return NULL;
583 }
584
585 struct include_data {
586 struct bitmap_index *bitmap_git;
587 struct bitmap *base;
588 struct bitmap *seen;
589 };
590
591 struct bitmap_lookup_table_triplet {
592 uint32_t commit_pos;
593 uint64_t offset;
594 uint32_t xor_row;
595 };
596
597 struct bitmap_lookup_table_xor_item {
598 struct object_id oid;
599 uint64_t offset;
600 };
601
602 /*
603 * Given a `triplet` struct pointer and pointer `p`, this
604 * function reads the triplet beginning at `p` into the struct.
605 * Note that this function assumes that there is enough memory
606 * left for filling the `triplet` struct from `p`.
607 */
608 static int bitmap_lookup_table_get_triplet_by_pointer(struct bitmap_lookup_table_triplet *triplet,
609 const unsigned char *p)
610 {
611 if (!triplet)
612 return -1;
613
614 triplet->commit_pos = get_be32(p);
615 p += sizeof(uint32_t);
616 triplet->offset = get_be64(p);
617 p += sizeof(uint64_t);
618 triplet->xor_row = get_be32(p);
619 return 0;
620 }
621
622 /*
623 * This function gets the raw triplet from `row`'th row in the
624 * lookup table and fills that data to the `triplet`.
625 */
626 static int bitmap_lookup_table_get_triplet(struct bitmap_index *bitmap_git,
627 uint32_t pos,
628 struct bitmap_lookup_table_triplet *triplet)
629 {
630 unsigned char *p = NULL;
631 if (pos >= bitmap_git->entry_count)
632 return error(_("corrupt bitmap lookup table: triplet position out of index"));
633
634 p = bitmap_git->table_lookup + st_mult(pos, BITMAP_LOOKUP_TABLE_TRIPLET_WIDTH);
635
636 return bitmap_lookup_table_get_triplet_by_pointer(triplet, p);
637 }
638
639 /*
640 * Searches for a matching triplet. `commit_pos` is a pointer
641 * to the wanted commit position value. `table_entry` points to
642 * a triplet in lookup table. The first 4 bytes of each
643 * triplet (pointed by `table_entry`) are compared with `*commit_pos`.
644 */
645 static int triplet_cmp(const void *commit_pos, const void *table_entry)
646 {
647
648 uint32_t a = *(uint32_t *)commit_pos;
649 uint32_t b = get_be32(table_entry);
650 if (a > b)
651 return 1;
652 else if (a < b)
653 return -1;
654
655 return 0;
656 }
657
658 static uint32_t bitmap_bsearch_pos(struct bitmap_index *bitmap_git,
659 struct object_id *oid,
660 uint32_t *result)
661 {
662 int found;
663
664 if (bitmap_is_midx(bitmap_git))
665 found = bsearch_midx(oid, bitmap_git->midx, result);
666 else
667 found = bsearch_pack(oid, bitmap_git->pack, result);
668
669 return found;
670 }
671
672 /*
673 * `bsearch_triplet_by_pos` function searches for the raw triplet
674 * having commit position same as `commit_pos` and fills `triplet`
675 * object from the raw triplet. Returns 1 on success and 0 on
676 * failure.
677 */
678 static int bitmap_bsearch_triplet_by_pos(uint32_t commit_pos,
679 struct bitmap_index *bitmap_git,
680 struct bitmap_lookup_table_triplet *triplet)
681 {
682 unsigned char *p = bsearch(&commit_pos, bitmap_git->table_lookup, bitmap_git->entry_count,
683 BITMAP_LOOKUP_TABLE_TRIPLET_WIDTH, triplet_cmp);
684
685 if (!p)
686 return -1;
687
688 return bitmap_lookup_table_get_triplet_by_pointer(triplet, p);
689 }
690
691 static struct stored_bitmap *lazy_bitmap_for_commit(struct bitmap_index *bitmap_git,
692 struct commit *commit)
693 {
694 uint32_t commit_pos, xor_row;
695 uint64_t offset;
696 int flags;
697 struct bitmap_lookup_table_triplet triplet;
698 struct object_id *oid = &commit->object.oid;
699 struct ewah_bitmap *bitmap;
700 struct stored_bitmap *xor_bitmap = NULL;
701 const int bitmap_header_size = 6;
702 static struct bitmap_lookup_table_xor_item *xor_items = NULL;
703 static size_t xor_items_nr = 0, xor_items_alloc = 0;
704 static int is_corrupt = 0;
705 int xor_flags;
706 khiter_t hash_pos;
707 struct bitmap_lookup_table_xor_item *xor_item;
708
709 if (is_corrupt)
710 return NULL;
711
712 if (!bitmap_bsearch_pos(bitmap_git, oid, &commit_pos))
713 return NULL;
714
715 if (bitmap_bsearch_triplet_by_pos(commit_pos, bitmap_git, &triplet) < 0)
716 return NULL;
717
718 xor_items_nr = 0;
719 offset = triplet.offset;
720 xor_row = triplet.xor_row;
721
722 while (xor_row != 0xffffffff) {
723 ALLOC_GROW(xor_items, xor_items_nr + 1, xor_items_alloc);
724
725 if (xor_items_nr + 1 >= bitmap_git->entry_count) {
726 error(_("corrupt bitmap lookup table: xor chain exceed entry count"));
727 goto corrupt;
728 }
729
730 if (bitmap_lookup_table_get_triplet(bitmap_git, xor_row, &triplet) < 0)
731 goto corrupt;
732
733 xor_item = &xor_items[xor_items_nr];
734 xor_item->offset = triplet.offset;
735
736 if (nth_bitmap_object_oid(bitmap_git, &xor_item->oid, triplet.commit_pos) < 0) {
737 error(_("corrupt bitmap lookup table: commit index %u out of range"),
738 triplet.commit_pos);
739 goto corrupt;
740 }
741
742 hash_pos = kh_get_oid_map(bitmap_git->bitmaps, xor_item->oid);
743
744 /*
745 * If desired bitmap is already stored, we don't need
746 * to iterate further. Because we know that bitmaps
747 * that are needed to be parsed to parse this bitmap
748 * has already been stored. So, assign this stored bitmap
749 * to the xor_bitmap.
750 */
751 if (hash_pos < kh_end(bitmap_git->bitmaps) &&
752 (xor_bitmap = kh_value(bitmap_git->bitmaps, hash_pos)))
753 break;
754 xor_items_nr++;
755 xor_row = triplet.xor_row;
756 }
757
758 while (xor_items_nr) {
759 xor_item = &xor_items[xor_items_nr - 1];
760 bitmap_git->map_pos = xor_item->offset;
761 if (bitmap_git->map_size - bitmap_git->map_pos < bitmap_header_size) {
762 error(_("corrupt ewah bitmap: truncated header for bitmap of commit \"%s\""),
763 oid_to_hex(&xor_item->oid));
764 goto corrupt;
765 }
766
767 bitmap_git->map_pos += sizeof(uint32_t) + sizeof(uint8_t);
768 xor_flags = read_u8(bitmap_git->map, &bitmap_git->map_pos);
769 bitmap = read_bitmap_1(bitmap_git);
770
771 if (!bitmap)
772 goto corrupt;
773
774 xor_bitmap = store_bitmap(bitmap_git, bitmap, &xor_item->oid, xor_bitmap, xor_flags);
775 xor_items_nr--;
776 }
777
778 bitmap_git->map_pos = offset;
779 if (bitmap_git->map_size - bitmap_git->map_pos < bitmap_header_size) {
780 error(_("corrupt ewah bitmap: truncated header for bitmap of commit \"%s\""),
781 oid_to_hex(oid));
782 goto corrupt;
783 }
784
785 /*
786 * Don't bother reading the commit's index position or its xor
787 * offset:
788 *
789 * - The commit's index position is irrelevant to us, since
790 * load_bitmap_entries_v1 only uses it to learn the object
791 * id which is used to compute the hashmap's key. We already
792 * have an object id, so no need to look it up again.
793 *
794 * - The xor_offset is unusable for us, since it specifies how
795 * many entries previous to ours we should look at. This
796 * makes sense when reading the bitmaps sequentially (as in
797 * load_bitmap_entries_v1()), since we can keep track of
798 * each bitmap as we read them.
799 *
800 * But it can't work for us, since the bitmap's don't have a
801 * fixed size. So we learn the position of the xor'd bitmap
802 * from the commit table (and resolve it to a bitmap in the
803 * above if-statement).
804 *
805 * Instead, we can skip ahead and immediately read the flags and
806 * ewah bitmap.
807 */
808 bitmap_git->map_pos += sizeof(uint32_t) + sizeof(uint8_t);
809 flags = read_u8(bitmap_git->map, &bitmap_git->map_pos);
810 bitmap = read_bitmap_1(bitmap_git);
811
812 if (!bitmap)
813 goto corrupt;
814
815 return store_bitmap(bitmap_git, bitmap, oid, xor_bitmap, flags);
816
817 corrupt:
818 free(xor_items);
819 is_corrupt = 1;
820 return NULL;
821 }
822
823 struct ewah_bitmap *bitmap_for_commit(struct bitmap_index *bitmap_git,
824 struct commit *commit)
825 {
826 khiter_t hash_pos = kh_get_oid_map(bitmap_git->bitmaps,
827 commit->object.oid);
828 if (hash_pos >= kh_end(bitmap_git->bitmaps)) {
829 struct stored_bitmap *bitmap = NULL;
830 if (!bitmap_git->table_lookup)
831 return NULL;
832
833 trace2_region_enter("pack-bitmap", "reading_lookup_table", the_repository);
834 /* NEEDSWORK: cache misses aren't recorded */
835 bitmap = lazy_bitmap_for_commit(bitmap_git, commit);
836 trace2_region_leave("pack-bitmap", "reading_lookup_table", the_repository);
837 if (!bitmap)
838 return NULL;
839 return lookup_stored_bitmap(bitmap);
840 }
841 return lookup_stored_bitmap(kh_value(bitmap_git->bitmaps, hash_pos));
842 }
843
844 static inline int bitmap_position_extended(struct bitmap_index *bitmap_git,
845 const struct object_id *oid)
846 {
847 kh_oid_pos_t *positions = bitmap_git->ext_index.positions;
848 khiter_t pos = kh_get_oid_pos(positions, *oid);
849
850 if (pos < kh_end(positions)) {
851 int bitmap_pos = kh_value(positions, pos);
852 return bitmap_pos + bitmap_num_objects(bitmap_git);
853 }
854
855 return -1;
856 }
857
858 static inline int bitmap_position_packfile(struct bitmap_index *bitmap_git,
859 const struct object_id *oid)
860 {
861 uint32_t pos;
862 off_t offset = find_pack_entry_one(oid->hash, bitmap_git->pack);
863 if (!offset)
864 return -1;
865
866 if (offset_to_pack_pos(bitmap_git->pack, offset, &pos) < 0)
867 return -1;
868 return pos;
869 }
870
871 static int bitmap_position_midx(struct bitmap_index *bitmap_git,
872 const struct object_id *oid)
873 {
874 uint32_t want, got;
875 if (!bsearch_midx(oid, bitmap_git->midx, &want))
876 return -1;
877
878 if (midx_to_pack_pos(bitmap_git->midx, want, &got) < 0)
879 return -1;
880 return got;
881 }
882
883 static int bitmap_position(struct bitmap_index *bitmap_git,
884 const struct object_id *oid)
885 {
886 int pos;
887 if (bitmap_is_midx(bitmap_git))
888 pos = bitmap_position_midx(bitmap_git, oid);
889 else
890 pos = bitmap_position_packfile(bitmap_git, oid);
891 return (pos >= 0) ? pos : bitmap_position_extended(bitmap_git, oid);
892 }
893
894 static int ext_index_add_object(struct bitmap_index *bitmap_git,
895 struct object *object, const char *name)
896 {
897 struct eindex *eindex = &bitmap_git->ext_index;
898
899 khiter_t hash_pos;
900 int hash_ret;
901 int bitmap_pos;
902
903 hash_pos = kh_put_oid_pos(eindex->positions, object->oid, &hash_ret);
904 if (hash_ret > 0) {
905 if (eindex->count >= eindex->alloc) {
906 eindex->alloc = (eindex->alloc + 16) * 3 / 2;
907 REALLOC_ARRAY(eindex->objects, eindex->alloc);
908 REALLOC_ARRAY(eindex->hashes, eindex->alloc);
909 }
910
911 bitmap_pos = eindex->count;
912 eindex->objects[eindex->count] = object;
913 eindex->hashes[eindex->count] = pack_name_hash(name);
914 kh_value(eindex->positions, hash_pos) = bitmap_pos;
915 eindex->count++;
916 } else {
917 bitmap_pos = kh_value(eindex->positions, hash_pos);
918 }
919
920 return bitmap_pos + bitmap_num_objects(bitmap_git);
921 }
922
923 struct bitmap_show_data {
924 struct bitmap_index *bitmap_git;
925 struct bitmap *base;
926 };
927
928 static void show_object(struct object *object, const char *name, void *data_)
929 {
930 struct bitmap_show_data *data = data_;
931 int bitmap_pos;
932
933 bitmap_pos = bitmap_position(data->bitmap_git, &object->oid);
934
935 if (bitmap_pos < 0)
936 bitmap_pos = ext_index_add_object(data->bitmap_git, object,
937 name);
938
939 bitmap_set(data->base, bitmap_pos);
940 }
941
942 static void show_commit(struct commit *commit, void *data)
943 {
944 }
945
946 static int add_to_include_set(struct bitmap_index *bitmap_git,
947 struct include_data *data,
948 struct commit *commit,
949 int bitmap_pos)
950 {
951 struct ewah_bitmap *partial;
952
953 if (data->seen && bitmap_get(data->seen, bitmap_pos))
954 return 0;
955
956 if (bitmap_get(data->base, bitmap_pos))
957 return 0;
958
959 partial = bitmap_for_commit(bitmap_git, commit);
960 if (partial) {
961 bitmap_or_ewah(data->base, partial);
962 return 0;
963 }
964
965 bitmap_set(data->base, bitmap_pos);
966 return 1;
967 }
968
969 static int should_include(struct commit *commit, void *_data)
970 {
971 struct include_data *data = _data;
972 int bitmap_pos;
973
974 bitmap_pos = bitmap_position(data->bitmap_git, &commit->object.oid);
975 if (bitmap_pos < 0)
976 bitmap_pos = ext_index_add_object(data->bitmap_git,
977 (struct object *)commit,
978 NULL);
979
980 if (!add_to_include_set(data->bitmap_git, data, commit, bitmap_pos)) {
981 struct commit_list *parent = commit->parents;
982
983 while (parent) {
984 parent->item->object.flags |= SEEN;
985 parent = parent->next;
986 }
987
988 return 0;
989 }
990
991 return 1;
992 }
993
994 static int should_include_obj(struct object *obj, void *_data)
995 {
996 struct include_data *data = _data;
997 int bitmap_pos;
998
999 bitmap_pos = bitmap_position(data->bitmap_git, &obj->oid);
1000 if (bitmap_pos < 0)
1001 return 1;
1002 if ((data->seen && bitmap_get(data->seen, bitmap_pos)) ||
1003 bitmap_get(data->base, bitmap_pos)) {
1004 obj->flags |= SEEN;
1005 return 0;
1006 }
1007 return 1;
1008 }
1009
1010 static int add_commit_to_bitmap(struct bitmap_index *bitmap_git,
1011 struct bitmap **base,
1012 struct commit *commit)
1013 {
1014 struct ewah_bitmap *or_with = bitmap_for_commit(bitmap_git, commit);
1015
1016 if (!or_with)
1017 return 0;
1018
1019 if (!*base)
1020 *base = ewah_to_bitmap(or_with);
1021 else
1022 bitmap_or_ewah(*base, or_with);
1023
1024 return 1;
1025 }
1026
1027 static struct bitmap *find_objects(struct bitmap_index *bitmap_git,
1028 struct rev_info *revs,
1029 struct object_list *roots,
1030 struct bitmap *seen)
1031 {
1032 struct bitmap *base = NULL;
1033 int needs_walk = 0;
1034
1035 struct object_list *not_mapped = NULL;
1036
1037 /*
1038 * Go through all the roots for the walk. The ones that have bitmaps
1039 * on the bitmap index will be `or`ed together to form an initial
1040 * global reachability analysis.
1041 *
1042 * The ones without bitmaps in the index will be stored in the
1043 * `not_mapped_list` for further processing.
1044 */
1045 while (roots) {
1046 struct object *object = roots->item;
1047 roots = roots->next;
1048
1049 if (object->type == OBJ_COMMIT &&
1050 add_commit_to_bitmap(bitmap_git, &base, (struct commit *)object)) {
1051 object->flags |= SEEN;
1052 continue;
1053 }
1054
1055 object_list_insert(object, &not_mapped);
1056 }
1057
1058 /*
1059 * Best case scenario: We found bitmaps for all the roots,
1060 * so the resulting `or` bitmap has the full reachability analysis
1061 */
1062 if (!not_mapped)
1063 return base;
1064
1065 roots = not_mapped;
1066
1067 /*
1068 * Let's iterate through all the roots that don't have bitmaps to
1069 * check if we can determine them to be reachable from the existing
1070 * global bitmap.
1071 *
1072 * If we cannot find them in the existing global bitmap, we'll need
1073 * to push them to an actual walk and run it until we can confirm
1074 * they are reachable
1075 */
1076 while (roots) {
1077 struct object *object = roots->item;
1078 int pos;
1079
1080 roots = roots->next;
1081 pos = bitmap_position(bitmap_git, &object->oid);
1082
1083 if (pos < 0 || base == NULL || !bitmap_get(base, pos)) {
1084 object->flags &= ~UNINTERESTING;
1085 add_pending_object(revs, object, "");
1086 needs_walk = 1;
1087 } else {
1088 object->flags |= SEEN;
1089 }
1090 }
1091
1092 if (needs_walk) {
1093 struct include_data incdata;
1094 struct bitmap_show_data show_data;
1095
1096 if (!base)
1097 base = bitmap_new();
1098
1099 incdata.bitmap_git = bitmap_git;
1100 incdata.base = base;
1101 incdata.seen = seen;
1102
1103 revs->include_check = should_include;
1104 revs->include_check_obj = should_include_obj;
1105 revs->include_check_data = &incdata;
1106
1107 if (prepare_revision_walk(revs))
1108 die(_("revision walk setup failed"));
1109
1110 show_data.bitmap_git = bitmap_git;
1111 show_data.base = base;
1112
1113 traverse_commit_list(revs,
1114 show_commit, show_object,
1115 &show_data);
1116
1117 revs->include_check = NULL;
1118 revs->include_check_obj = NULL;
1119 revs->include_check_data = NULL;
1120 }
1121
1122 return base;
1123 }
1124
1125 static void show_extended_objects(struct bitmap_index *bitmap_git,
1126 struct rev_info *revs,
1127 show_reachable_fn show_reach)
1128 {
1129 struct bitmap *objects = bitmap_git->result;
1130 struct eindex *eindex = &bitmap_git->ext_index;
1131 uint32_t i;
1132
1133 for (i = 0; i < eindex->count; ++i) {
1134 struct object *obj;
1135
1136 if (!bitmap_get(objects, bitmap_num_objects(bitmap_git) + i))
1137 continue;
1138
1139 obj = eindex->objects[i];
1140 if ((obj->type == OBJ_BLOB && !revs->blob_objects) ||
1141 (obj->type == OBJ_TREE && !revs->tree_objects) ||
1142 (obj->type == OBJ_TAG && !revs->tag_objects))
1143 continue;
1144
1145 show_reach(&obj->oid, obj->type, 0, eindex->hashes[i], NULL, 0);
1146 }
1147 }
1148
1149 static void init_type_iterator(struct ewah_iterator *it,
1150 struct bitmap_index *bitmap_git,
1151 enum object_type type)
1152 {
1153 switch (type) {
1154 case OBJ_COMMIT:
1155 ewah_iterator_init(it, bitmap_git->commits);
1156 break;
1157
1158 case OBJ_TREE:
1159 ewah_iterator_init(it, bitmap_git->trees);
1160 break;
1161
1162 case OBJ_BLOB:
1163 ewah_iterator_init(it, bitmap_git->blobs);
1164 break;
1165
1166 case OBJ_TAG:
1167 ewah_iterator_init(it, bitmap_git->tags);
1168 break;
1169
1170 default:
1171 BUG("object type %d not stored by bitmap type index", type);
1172 break;
1173 }
1174 }
1175
1176 static void show_objects_for_type(
1177 struct bitmap_index *bitmap_git,
1178 enum object_type object_type,
1179 show_reachable_fn show_reach)
1180 {
1181 size_t i = 0;
1182 uint32_t offset;
1183
1184 struct ewah_iterator it;
1185 eword_t filter;
1186
1187 struct bitmap *objects = bitmap_git->result;
1188
1189 init_type_iterator(&it, bitmap_git, object_type);
1190
1191 for (i = 0; i < objects->word_alloc &&
1192 ewah_iterator_next(&filter, &it); i++) {
1193 eword_t word = objects->words[i] & filter;
1194 size_t pos = (i * BITS_IN_EWORD);
1195
1196 if (!word)
1197 continue;
1198
1199 for (offset = 0; offset < BITS_IN_EWORD; ++offset) {
1200 struct packed_git *pack;
1201 struct object_id oid;
1202 uint32_t hash = 0, index_pos;
1203 off_t ofs;
1204
1205 if ((word >> offset) == 0)
1206 break;
1207
1208 offset += ewah_bit_ctz64(word >> offset);
1209
1210 if (bitmap_is_midx(bitmap_git)) {
1211 struct multi_pack_index *m = bitmap_git->midx;
1212 uint32_t pack_id;
1213
1214 index_pos = pack_pos_to_midx(m, pos + offset);
1215 ofs = nth_midxed_offset(m, index_pos);
1216 nth_midxed_object_oid(&oid, m, index_pos);
1217
1218 pack_id = nth_midxed_pack_int_id(m, index_pos);
1219 pack = bitmap_git->midx->packs[pack_id];
1220 } else {
1221 index_pos = pack_pos_to_index(bitmap_git->pack, pos + offset);
1222 ofs = pack_pos_to_offset(bitmap_git->pack, pos + offset);
1223 nth_bitmap_object_oid(bitmap_git, &oid, index_pos);
1224
1225 pack = bitmap_git->pack;
1226 }
1227
1228 if (bitmap_git->hashes)
1229 hash = get_be32(bitmap_git->hashes + index_pos);
1230
1231 show_reach(&oid, object_type, 0, hash, pack, ofs);
1232 }
1233 }
1234 }
1235
1236 static int in_bitmapped_pack(struct bitmap_index *bitmap_git,
1237 struct object_list *roots)
1238 {
1239 while (roots) {
1240 struct object *object = roots->item;
1241 roots = roots->next;
1242
1243 if (bitmap_is_midx(bitmap_git)) {
1244 if (bsearch_midx(&object->oid, bitmap_git->midx, NULL))
1245 return 1;
1246 } else {
1247 if (find_pack_entry_one(object->oid.hash, bitmap_git->pack) > 0)
1248 return 1;
1249 }
1250 }
1251
1252 return 0;
1253 }
1254
1255 static struct bitmap *find_tip_objects(struct bitmap_index *bitmap_git,
1256 struct object_list *tip_objects,
1257 enum object_type type)
1258 {
1259 struct bitmap *result = bitmap_new();
1260 struct object_list *p;
1261
1262 for (p = tip_objects; p; p = p->next) {
1263 int pos;
1264
1265 if (p->item->type != type)
1266 continue;
1267
1268 pos = bitmap_position(bitmap_git, &p->item->oid);
1269 if (pos < 0)
1270 continue;
1271
1272 bitmap_set(result, pos);
1273 }
1274
1275 return result;
1276 }
1277
1278 static void filter_bitmap_exclude_type(struct bitmap_index *bitmap_git,
1279 struct object_list *tip_objects,
1280 struct bitmap *to_filter,
1281 enum object_type type)
1282 {
1283 struct eindex *eindex = &bitmap_git->ext_index;
1284 struct bitmap *tips;
1285 struct ewah_iterator it;
1286 eword_t mask;
1287 uint32_t i;
1288
1289 /*
1290 * The non-bitmap version of this filter never removes
1291 * objects which the other side specifically asked for,
1292 * so we must match that behavior.
1293 */
1294 tips = find_tip_objects(bitmap_git, tip_objects, type);
1295
1296 /*
1297 * We can use the type-level bitmap for 'type' to work in whole
1298 * words for the objects that are actually in the bitmapped
1299 * packfile.
1300 */
1301 for (i = 0, init_type_iterator(&it, bitmap_git, type);
1302 i < to_filter->word_alloc && ewah_iterator_next(&mask, &it);
1303 i++) {
1304 if (i < tips->word_alloc)
1305 mask &= ~tips->words[i];
1306 to_filter->words[i] &= ~mask;
1307 }
1308
1309 /*
1310 * Clear any objects that weren't in the packfile (and so would
1311 * not have been caught by the loop above. We'll have to check
1312 * them individually.
1313 */
1314 for (i = 0; i < eindex->count; i++) {
1315 uint32_t pos = i + bitmap_num_objects(bitmap_git);
1316 if (eindex->objects[i]->type == type &&
1317 bitmap_get(to_filter, pos) &&
1318 !bitmap_get(tips, pos))
1319 bitmap_unset(to_filter, pos);
1320 }
1321
1322 bitmap_free(tips);
1323 }
1324
1325 static void filter_bitmap_blob_none(struct bitmap_index *bitmap_git,
1326 struct object_list *tip_objects,
1327 struct bitmap *to_filter)
1328 {
1329 filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter,
1330 OBJ_BLOB);
1331 }
1332
1333 static unsigned long get_size_by_pos(struct bitmap_index *bitmap_git,
1334 uint32_t pos)
1335 {
1336 unsigned long size;
1337 struct object_info oi = OBJECT_INFO_INIT;
1338
1339 oi.sizep = &size;
1340
1341 if (pos < bitmap_num_objects(bitmap_git)) {
1342 struct packed_git *pack;
1343 off_t ofs;
1344
1345 if (bitmap_is_midx(bitmap_git)) {
1346 uint32_t midx_pos = pack_pos_to_midx(bitmap_git->midx, pos);
1347 uint32_t pack_id = nth_midxed_pack_int_id(bitmap_git->midx, midx_pos);
1348
1349 pack = bitmap_git->midx->packs[pack_id];
1350 ofs = nth_midxed_offset(bitmap_git->midx, midx_pos);
1351 } else {
1352 pack = bitmap_git->pack;
1353 ofs = pack_pos_to_offset(pack, pos);
1354 }
1355
1356 if (packed_object_info(the_repository, pack, ofs, &oi) < 0) {
1357 struct object_id oid;
1358 nth_bitmap_object_oid(bitmap_git, &oid,
1359 pack_pos_to_index(pack, pos));
1360 die(_("unable to get size of %s"), oid_to_hex(&oid));
1361 }
1362 } else {
1363 struct eindex *eindex = &bitmap_git->ext_index;
1364 struct object *obj = eindex->objects[pos - bitmap_num_objects(bitmap_git)];
1365 if (oid_object_info_extended(the_repository, &obj->oid, &oi, 0) < 0)
1366 die(_("unable to get size of %s"), oid_to_hex(&obj->oid));
1367 }
1368
1369 return size;
1370 }
1371
1372 static void filter_bitmap_blob_limit(struct bitmap_index *bitmap_git,
1373 struct object_list *tip_objects,
1374 struct bitmap *to_filter,
1375 unsigned long limit)
1376 {
1377 struct eindex *eindex = &bitmap_git->ext_index;
1378 struct bitmap *tips;
1379 struct ewah_iterator it;
1380 eword_t mask;
1381 uint32_t i;
1382
1383 tips = find_tip_objects(bitmap_git, tip_objects, OBJ_BLOB);
1384
1385 for (i = 0, init_type_iterator(&it, bitmap_git, OBJ_BLOB);
1386 i < to_filter->word_alloc && ewah_iterator_next(&mask, &it);
1387 i++) {
1388 eword_t word = to_filter->words[i] & mask;
1389 unsigned offset;
1390
1391 for (offset = 0; offset < BITS_IN_EWORD; offset++) {
1392 uint32_t pos;
1393
1394 if ((word >> offset) == 0)
1395 break;
1396 offset += ewah_bit_ctz64(word >> offset);
1397 pos = i * BITS_IN_EWORD + offset;
1398
1399 if (!bitmap_get(tips, pos) &&
1400 get_size_by_pos(bitmap_git, pos) >= limit)
1401 bitmap_unset(to_filter, pos);
1402 }
1403 }
1404
1405 for (i = 0; i < eindex->count; i++) {
1406 uint32_t pos = i + bitmap_num_objects(bitmap_git);
1407 if (eindex->objects[i]->type == OBJ_BLOB &&
1408 bitmap_get(to_filter, pos) &&
1409 !bitmap_get(tips, pos) &&
1410 get_size_by_pos(bitmap_git, pos) >= limit)
1411 bitmap_unset(to_filter, pos);
1412 }
1413
1414 bitmap_free(tips);
1415 }
1416
1417 static void filter_bitmap_tree_depth(struct bitmap_index *bitmap_git,
1418 struct object_list *tip_objects,
1419 struct bitmap *to_filter,
1420 unsigned long limit)
1421 {
1422 if (limit)
1423 BUG("filter_bitmap_tree_depth given non-zero limit");
1424
1425 filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter,
1426 OBJ_TREE);
1427 filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter,
1428 OBJ_BLOB);
1429 }
1430
1431 static void filter_bitmap_object_type(struct bitmap_index *bitmap_git,
1432 struct object_list *tip_objects,
1433 struct bitmap *to_filter,
1434 enum object_type object_type)
1435 {
1436 if (object_type < OBJ_COMMIT || object_type > OBJ_TAG)
1437 BUG("filter_bitmap_object_type given invalid object");
1438
1439 if (object_type != OBJ_TAG)
1440 filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, OBJ_TAG);
1441 if (object_type != OBJ_COMMIT)
1442 filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, OBJ_COMMIT);
1443 if (object_type != OBJ_TREE)
1444 filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, OBJ_TREE);
1445 if (object_type != OBJ_BLOB)
1446 filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, OBJ_BLOB);
1447 }
1448
1449 static int filter_bitmap(struct bitmap_index *bitmap_git,
1450 struct object_list *tip_objects,
1451 struct bitmap *to_filter,
1452 struct list_objects_filter_options *filter)
1453 {
1454 if (!filter || filter->choice == LOFC_DISABLED)
1455 return 0;
1456
1457 if (filter->choice == LOFC_BLOB_NONE) {
1458 if (bitmap_git)
1459 filter_bitmap_blob_none(bitmap_git, tip_objects,
1460 to_filter);
1461 return 0;
1462 }
1463
1464 if (filter->choice == LOFC_BLOB_LIMIT) {
1465 if (bitmap_git)
1466 filter_bitmap_blob_limit(bitmap_git, tip_objects,
1467 to_filter,
1468 filter->blob_limit_value);
1469 return 0;
1470 }
1471
1472 if (filter->choice == LOFC_TREE_DEPTH &&
1473 filter->tree_exclude_depth == 0) {
1474 if (bitmap_git)
1475 filter_bitmap_tree_depth(bitmap_git, tip_objects,
1476 to_filter,
1477 filter->tree_exclude_depth);
1478 return 0;
1479 }
1480
1481 if (filter->choice == LOFC_OBJECT_TYPE) {
1482 if (bitmap_git)
1483 filter_bitmap_object_type(bitmap_git, tip_objects,
1484 to_filter,
1485 filter->object_type);
1486 return 0;
1487 }
1488
1489 if (filter->choice == LOFC_COMBINE) {
1490 int i;
1491 for (i = 0; i < filter->sub_nr; i++) {
1492 if (filter_bitmap(bitmap_git, tip_objects, to_filter,
1493 &filter->sub[i]) < 0)
1494 return -1;
1495 }
1496 return 0;
1497 }
1498
1499 /* filter choice not handled */
1500 return -1;
1501 }
1502
1503 static int can_filter_bitmap(struct list_objects_filter_options *filter)
1504 {
1505 return !filter_bitmap(NULL, NULL, NULL, filter);
1506 }
1507
1508 struct bitmap_index *prepare_bitmap_walk(struct rev_info *revs,
1509 int filter_provided_objects)
1510 {
1511 unsigned int i;
1512
1513 struct object_list *wants = NULL;
1514 struct object_list *haves = NULL;
1515
1516 struct bitmap *wants_bitmap = NULL;
1517 struct bitmap *haves_bitmap = NULL;
1518
1519 struct bitmap_index *bitmap_git;
1520
1521 /*
1522 * We can't do pathspec limiting with bitmaps, because we don't know
1523 * which commits are associated with which object changes (let alone
1524 * even which objects are associated with which paths).
1525 */
1526 if (revs->prune)
1527 return NULL;
1528
1529 if (!can_filter_bitmap(&revs->filter))
1530 return NULL;
1531
1532 /* try to open a bitmapped pack, but don't parse it yet
1533 * because we may not need to use it */
1534 CALLOC_ARRAY(bitmap_git, 1);
1535 if (open_bitmap(revs->repo, bitmap_git) < 0)
1536 goto cleanup;
1537
1538 for (i = 0; i < revs->pending.nr; ++i) {
1539 struct object *object = revs->pending.objects[i].item;
1540
1541 if (object->type == OBJ_NONE)
1542 parse_object_or_die(&object->oid, NULL);
1543
1544 while (object->type == OBJ_TAG) {
1545 struct tag *tag = (struct tag *) object;
1546
1547 if (object->flags & UNINTERESTING)
1548 object_list_insert(object, &haves);
1549 else
1550 object_list_insert(object, &wants);
1551
1552 object = parse_object_or_die(get_tagged_oid(tag), NULL);
1553 object->flags |= (tag->object.flags & UNINTERESTING);
1554 }
1555
1556 if (object->flags & UNINTERESTING)
1557 object_list_insert(object, &haves);
1558 else
1559 object_list_insert(object, &wants);
1560 }
1561
1562 /*
1563 * if we have a HAVES list, but none of those haves is contained
1564 * in the packfile that has a bitmap, we don't have anything to
1565 * optimize here
1566 */
1567 if (haves && !in_bitmapped_pack(bitmap_git, haves))
1568 goto cleanup;
1569
1570 /* if we don't want anything, we're done here */
1571 if (!wants)
1572 goto cleanup;
1573
1574 /*
1575 * now we're going to use bitmaps, so load the actual bitmap entries
1576 * from disk. this is the point of no return; after this the rev_list
1577 * becomes invalidated and we must perform the revwalk through bitmaps
1578 */
1579 if (load_bitmap(bitmap_git) < 0)
1580 goto cleanup;
1581
1582 object_array_clear(&revs->pending);
1583
1584 if (haves) {
1585 revs->ignore_missing_links = 1;
1586 haves_bitmap = find_objects(bitmap_git, revs, haves, NULL);
1587 reset_revision_walk();
1588 revs->ignore_missing_links = 0;
1589
1590 if (!haves_bitmap)
1591 BUG("failed to perform bitmap walk");
1592 }
1593
1594 wants_bitmap = find_objects(bitmap_git, revs, wants, haves_bitmap);
1595
1596 if (!wants_bitmap)
1597 BUG("failed to perform bitmap walk");
1598
1599 if (haves_bitmap)
1600 bitmap_and_not(wants_bitmap, haves_bitmap);
1601
1602 filter_bitmap(bitmap_git,
1603 (revs->filter.choice && filter_provided_objects) ? NULL : wants,
1604 wants_bitmap,
1605 &revs->filter);
1606
1607 bitmap_git->result = wants_bitmap;
1608 bitmap_git->haves = haves_bitmap;
1609
1610 object_list_free(&wants);
1611 object_list_free(&haves);
1612
1613 return bitmap_git;
1614
1615 cleanup:
1616 free_bitmap_index(bitmap_git);
1617 object_list_free(&wants);
1618 object_list_free(&haves);
1619 return NULL;
1620 }
1621
1622 /*
1623 * -1 means "stop trying further objects"; 0 means we may or may not have
1624 * reused, but you can keep feeding bits.
1625 */
1626 static int try_partial_reuse(struct packed_git *pack,
1627 size_t pos,
1628 struct bitmap *reuse,
1629 struct pack_window **w_curs)
1630 {
1631 off_t offset, delta_obj_offset;
1632 enum object_type type;
1633 unsigned long size;
1634
1635 /*
1636 * try_partial_reuse() is called either on (a) objects in the
1637 * bitmapped pack (in the case of a single-pack bitmap) or (b)
1638 * objects in the preferred pack of a multi-pack bitmap.
1639 * Importantly, the latter can pretend as if only a single pack
1640 * exists because:
1641 *
1642 * - The first pack->num_objects bits of a MIDX bitmap are
1643 * reserved for the preferred pack, and
1644 *
1645 * - Ties due to duplicate objects are always resolved in
1646 * favor of the preferred pack.
1647 *
1648 * Therefore we do not need to ever ask the MIDX for its copy of
1649 * an object by OID, since it will always select it from the
1650 * preferred pack. Likewise, the selected copy of the base
1651 * object for any deltas will reside in the same pack.
1652 *
1653 * This means that we can reuse pos when looking up the bit in
1654 * the reuse bitmap, too, since bits corresponding to the
1655 * preferred pack precede all bits from other packs.
1656 */
1657
1658 if (pos >= pack->num_objects)
1659 return -1; /* not actually in the pack or MIDX preferred pack */
1660
1661 offset = delta_obj_offset = pack_pos_to_offset(pack, pos);
1662 type = unpack_object_header(pack, w_curs, &offset, &size);
1663 if (type < 0)
1664 return -1; /* broken packfile, punt */
1665
1666 if (type == OBJ_REF_DELTA || type == OBJ_OFS_DELTA) {
1667 off_t base_offset;
1668 uint32_t base_pos;
1669
1670 /*
1671 * Find the position of the base object so we can look it up
1672 * in our bitmaps. If we can't come up with an offset, or if
1673 * that offset is not in the revidx, the pack is corrupt.
1674 * There's nothing we can do, so just punt on this object,
1675 * and the normal slow path will complain about it in
1676 * more detail.
1677 */
1678 base_offset = get_delta_base(pack, w_curs, &offset, type,
1679 delta_obj_offset);
1680 if (!base_offset)
1681 return 0;
1682 if (offset_to_pack_pos(pack, base_offset, &base_pos) < 0)
1683 return 0;
1684
1685 /*
1686 * We assume delta dependencies always point backwards. This
1687 * lets us do a single pass, and is basically always true
1688 * due to the way OFS_DELTAs work. You would not typically
1689 * find REF_DELTA in a bitmapped pack, since we only bitmap
1690 * packs we write fresh, and OFS_DELTA is the default). But
1691 * let's double check to make sure the pack wasn't written with
1692 * odd parameters.
1693 */
1694 if (base_pos >= pos)
1695 return 0;
1696
1697 /*
1698 * And finally, if we're not sending the base as part of our
1699 * reuse chunk, then don't send this object either. The base
1700 * would come after us, along with other objects not
1701 * necessarily in the pack, which means we'd need to convert
1702 * to REF_DELTA on the fly. Better to just let the normal
1703 * object_entry code path handle it.
1704 */
1705 if (!bitmap_get(reuse, base_pos))
1706 return 0;
1707 }
1708
1709 /*
1710 * If we got here, then the object is OK to reuse. Mark it.
1711 */
1712 bitmap_set(reuse, pos);
1713 return 0;
1714 }
1715
1716 uint32_t midx_preferred_pack(struct bitmap_index *bitmap_git)
1717 {
1718 struct multi_pack_index *m = bitmap_git->midx;
1719 if (!m)
1720 BUG("midx_preferred_pack: requires non-empty MIDX");
1721 return nth_midxed_pack_int_id(m, pack_pos_to_midx(bitmap_git->midx, 0));
1722 }
1723
1724 int reuse_partial_packfile_from_bitmap(struct bitmap_index *bitmap_git,
1725 struct packed_git **packfile_out,
1726 uint32_t *entries,
1727 struct bitmap **reuse_out)
1728 {
1729 struct packed_git *pack;
1730 struct bitmap *result = bitmap_git->result;
1731 struct bitmap *reuse;
1732 struct pack_window *w_curs = NULL;
1733 size_t i = 0;
1734 uint32_t offset;
1735 uint32_t objects_nr;
1736
1737 assert(result);
1738
1739 load_reverse_index(bitmap_git);
1740
1741 if (bitmap_is_midx(bitmap_git))
1742 pack = bitmap_git->midx->packs[midx_preferred_pack(bitmap_git)];
1743 else
1744 pack = bitmap_git->pack;
1745 objects_nr = pack->num_objects;
1746
1747 while (i < result->word_alloc && result->words[i] == (eword_t)~0)
1748 i++;
1749
1750 /*
1751 * Don't mark objects not in the packfile or preferred pack. This bitmap
1752 * marks objects eligible for reuse, but the pack-reuse code only
1753 * understands how to reuse a single pack. Since the preferred pack is
1754 * guaranteed to have all bases for its deltas (in a multi-pack bitmap),
1755 * we use it instead of another pack. In single-pack bitmaps, the choice
1756 * is made for us.
1757 */
1758 if (i > objects_nr / BITS_IN_EWORD)
1759 i = objects_nr / BITS_IN_EWORD;
1760
1761 reuse = bitmap_word_alloc(i);
1762 memset(reuse->words, 0xFF, i * sizeof(eword_t));
1763
1764 for (; i < result->word_alloc; ++i) {
1765 eword_t word = result->words[i];
1766 size_t pos = (i * BITS_IN_EWORD);
1767
1768 for (offset = 0; offset < BITS_IN_EWORD; ++offset) {
1769 if ((word >> offset) == 0)
1770 break;
1771
1772 offset += ewah_bit_ctz64(word >> offset);
1773 if (try_partial_reuse(pack, pos + offset,
1774 reuse, &w_curs) < 0) {
1775 /*
1776 * try_partial_reuse indicated we couldn't reuse
1777 * any bits, so there is no point in trying more
1778 * bits in the current word, or any other words
1779 * in result.
1780 *
1781 * Jump out of both loops to avoid future
1782 * unnecessary calls to try_partial_reuse.
1783 */
1784 goto done;
1785 }
1786 }
1787 }
1788
1789 done:
1790 unuse_pack(&w_curs);
1791
1792 *entries = bitmap_popcount(reuse);
1793 if (!*entries) {
1794 bitmap_free(reuse);
1795 return -1;
1796 }
1797
1798 /*
1799 * Drop any reused objects from the result, since they will not
1800 * need to be handled separately.
1801 */
1802 bitmap_and_not(result, reuse);
1803 *packfile_out = pack;
1804 *reuse_out = reuse;
1805 return 0;
1806 }
1807
1808 int bitmap_walk_contains(struct bitmap_index *bitmap_git,
1809 struct bitmap *bitmap, const struct object_id *oid)
1810 {
1811 int idx;
1812
1813 if (!bitmap)
1814 return 0;
1815
1816 idx = bitmap_position(bitmap_git, oid);
1817 return idx >= 0 && bitmap_get(bitmap, idx);
1818 }
1819
1820 void traverse_bitmap_commit_list(struct bitmap_index *bitmap_git,
1821 struct rev_info *revs,
1822 show_reachable_fn show_reachable)
1823 {
1824 assert(bitmap_git->result);
1825
1826 show_objects_for_type(bitmap_git, OBJ_COMMIT, show_reachable);
1827 if (revs->tree_objects)
1828 show_objects_for_type(bitmap_git, OBJ_TREE, show_reachable);
1829 if (revs->blob_objects)
1830 show_objects_for_type(bitmap_git, OBJ_BLOB, show_reachable);
1831 if (revs->tag_objects)
1832 show_objects_for_type(bitmap_git, OBJ_TAG, show_reachable);
1833
1834 show_extended_objects(bitmap_git, revs, show_reachable);
1835 }
1836
1837 static uint32_t count_object_type(struct bitmap_index *bitmap_git,
1838 enum object_type type)
1839 {
1840 struct bitmap *objects = bitmap_git->result;
1841 struct eindex *eindex = &bitmap_git->ext_index;
1842
1843 uint32_t i = 0, count = 0;
1844 struct ewah_iterator it;
1845 eword_t filter;
1846
1847 init_type_iterator(&it, bitmap_git, type);
1848
1849 while (i < objects->word_alloc && ewah_iterator_next(&filter, &it)) {
1850 eword_t word = objects->words[i++] & filter;
1851 count += ewah_bit_popcount64(word);
1852 }
1853
1854 for (i = 0; i < eindex->count; ++i) {
1855 if (eindex->objects[i]->type == type &&
1856 bitmap_get(objects, bitmap_num_objects(bitmap_git) + i))
1857 count++;
1858 }
1859
1860 return count;
1861 }
1862
1863 void count_bitmap_commit_list(struct bitmap_index *bitmap_git,
1864 uint32_t *commits, uint32_t *trees,
1865 uint32_t *blobs, uint32_t *tags)
1866 {
1867 assert(bitmap_git->result);
1868
1869 if (commits)
1870 *commits = count_object_type(bitmap_git, OBJ_COMMIT);
1871
1872 if (trees)
1873 *trees = count_object_type(bitmap_git, OBJ_TREE);
1874
1875 if (blobs)
1876 *blobs = count_object_type(bitmap_git, OBJ_BLOB);
1877
1878 if (tags)
1879 *tags = count_object_type(bitmap_git, OBJ_TAG);
1880 }
1881
1882 struct bitmap_test_data {
1883 struct bitmap_index *bitmap_git;
1884 struct bitmap *base;
1885 struct bitmap *commits;
1886 struct bitmap *trees;
1887 struct bitmap *blobs;
1888 struct bitmap *tags;
1889 struct progress *prg;
1890 size_t seen;
1891 };
1892
1893 static void test_bitmap_type(struct bitmap_test_data *tdata,
1894 struct object *obj, int pos)
1895 {
1896 enum object_type bitmap_type = OBJ_NONE;
1897 int bitmaps_nr = 0;
1898
1899 if (bitmap_get(tdata->commits, pos)) {
1900 bitmap_type = OBJ_COMMIT;
1901 bitmaps_nr++;
1902 }
1903 if (bitmap_get(tdata->trees, pos)) {
1904 bitmap_type = OBJ_TREE;
1905 bitmaps_nr++;
1906 }
1907 if (bitmap_get(tdata->blobs, pos)) {
1908 bitmap_type = OBJ_BLOB;
1909 bitmaps_nr++;
1910 }
1911 if (bitmap_get(tdata->tags, pos)) {
1912 bitmap_type = OBJ_TAG;
1913 bitmaps_nr++;
1914 }
1915
1916 if (bitmap_type == OBJ_NONE)
1917 die(_("object '%s' not found in type bitmaps"),
1918 oid_to_hex(&obj->oid));
1919
1920 if (bitmaps_nr > 1)
1921 die(_("object '%s' does not have a unique type"),
1922 oid_to_hex(&obj->oid));
1923
1924 if (bitmap_type != obj->type)
1925 die(_("object '%s': real type '%s', expected: '%s'"),
1926 oid_to_hex(&obj->oid),
1927 type_name(obj->type),
1928 type_name(bitmap_type));
1929 }
1930
1931 static void test_show_object(struct object *object, const char *name,
1932 void *data)
1933 {
1934 struct bitmap_test_data *tdata = data;
1935 int bitmap_pos;
1936
1937 bitmap_pos = bitmap_position(tdata->bitmap_git, &object->oid);
1938 if (bitmap_pos < 0)
1939 die(_("object not in bitmap: '%s'"), oid_to_hex(&object->oid));
1940 test_bitmap_type(tdata, object, bitmap_pos);
1941
1942 bitmap_set(tdata->base, bitmap_pos);
1943 display_progress(tdata->prg, ++tdata->seen);
1944 }
1945
1946 static void test_show_commit(struct commit *commit, void *data)
1947 {
1948 struct bitmap_test_data *tdata = data;
1949 int bitmap_pos;
1950
1951 bitmap_pos = bitmap_position(tdata->bitmap_git,
1952 &commit->object.oid);
1953 if (bitmap_pos < 0)
1954 die(_("object not in bitmap: '%s'"), oid_to_hex(&commit->object.oid));
1955 test_bitmap_type(tdata, &commit->object, bitmap_pos);
1956
1957 bitmap_set(tdata->base, bitmap_pos);
1958 display_progress(tdata->prg, ++tdata->seen);
1959 }
1960
1961 void test_bitmap_walk(struct rev_info *revs)
1962 {
1963 struct object *root;
1964 struct bitmap *result = NULL;
1965 size_t result_popcnt;
1966 struct bitmap_test_data tdata;
1967 struct bitmap_index *bitmap_git;
1968 struct ewah_bitmap *bm;
1969
1970 if (!(bitmap_git = prepare_bitmap_git(revs->repo)))
1971 die(_("failed to load bitmap indexes"));
1972
1973 if (revs->pending.nr != 1)
1974 die(_("you must specify exactly one commit to test"));
1975
1976 fprintf_ln(stderr, "Bitmap v%d test (%d entries%s)",
1977 bitmap_git->version,
1978 bitmap_git->entry_count,
1979 bitmap_git->table_lookup ? "" : " loaded");
1980
1981 root = revs->pending.objects[0].item;
1982 bm = bitmap_for_commit(bitmap_git, (struct commit *)root);
1983
1984 if (bm) {
1985 fprintf_ln(stderr, "Found bitmap for '%s'. %d bits / %08x checksum",
1986 oid_to_hex(&root->oid), (int)bm->bit_size, ewah_checksum(bm));
1987
1988 result = ewah_to_bitmap(bm);
1989 }
1990
1991 if (!result)
1992 die(_("commit '%s' doesn't have an indexed bitmap"), oid_to_hex(&root->oid));
1993
1994 revs->tag_objects = 1;
1995 revs->tree_objects = 1;
1996 revs->blob_objects = 1;
1997
1998 result_popcnt = bitmap_popcount(result);
1999
2000 if (prepare_revision_walk(revs))
2001 die(_("revision walk setup failed"));
2002
2003 tdata.bitmap_git = bitmap_git;
2004 tdata.base = bitmap_new();
2005 tdata.commits = ewah_to_bitmap(bitmap_git->commits);
2006 tdata.trees = ewah_to_bitmap(bitmap_git->trees);
2007 tdata.blobs = ewah_to_bitmap(bitmap_git->blobs);
2008 tdata.tags = ewah_to_bitmap(bitmap_git->tags);
2009 tdata.prg = start_progress("Verifying bitmap entries", result_popcnt);
2010 tdata.seen = 0;
2011
2012 traverse_commit_list(revs, &test_show_commit, &test_show_object, &tdata);
2013
2014 stop_progress(&tdata.prg);
2015
2016 if (bitmap_equals(result, tdata.base))
2017 fprintf_ln(stderr, "OK!");
2018 else
2019 die(_("mismatch in bitmap results"));
2020
2021 bitmap_free(result);
2022 bitmap_free(tdata.base);
2023 bitmap_free(tdata.commits);
2024 bitmap_free(tdata.trees);
2025 bitmap_free(tdata.blobs);
2026 bitmap_free(tdata.tags);
2027 free_bitmap_index(bitmap_git);
2028 }
2029
2030 int test_bitmap_commits(struct repository *r)
2031 {
2032 struct object_id oid;
2033 MAYBE_UNUSED void *value;
2034 struct bitmap_index *bitmap_git = prepare_bitmap_git(r);
2035
2036 if (!bitmap_git)
2037 die(_("failed to load bitmap indexes"));
2038
2039 /*
2040 * As this function is only used to print bitmap selected
2041 * commits, we don't have to read the commit table.
2042 */
2043 if (bitmap_git->table_lookup) {
2044 if (load_bitmap_entries_v1(bitmap_git) < 0)
2045 die(_("failed to load bitmap indexes"));
2046 }
2047
2048 kh_foreach(bitmap_git->bitmaps, oid, value, {
2049 printf_ln("%s", oid_to_hex(&oid));
2050 });
2051
2052 free_bitmap_index(bitmap_git);
2053
2054 return 0;
2055 }
2056
2057 int test_bitmap_hashes(struct repository *r)
2058 {
2059 struct bitmap_index *bitmap_git = prepare_bitmap_git(r);
2060 struct object_id oid;
2061 uint32_t i, index_pos;
2062
2063 if (!bitmap_git || !bitmap_git->hashes)
2064 goto cleanup;
2065
2066 for (i = 0; i < bitmap_num_objects(bitmap_git); i++) {
2067 if (bitmap_is_midx(bitmap_git))
2068 index_pos = pack_pos_to_midx(bitmap_git->midx, i);
2069 else
2070 index_pos = pack_pos_to_index(bitmap_git->pack, i);
2071
2072 nth_bitmap_object_oid(bitmap_git, &oid, index_pos);
2073
2074 printf_ln("%s %"PRIu32"",
2075 oid_to_hex(&oid), get_be32(bitmap_git->hashes + index_pos));
2076 }
2077
2078 cleanup:
2079 free_bitmap_index(bitmap_git);
2080
2081 return 0;
2082 }
2083
2084 int rebuild_bitmap(const uint32_t *reposition,
2085 struct ewah_bitmap *source,
2086 struct bitmap *dest)
2087 {
2088 uint32_t pos = 0;
2089 struct ewah_iterator it;
2090 eword_t word;
2091
2092 ewah_iterator_init(&it, source);
2093
2094 while (ewah_iterator_next(&word, &it)) {
2095 uint32_t offset, bit_pos;
2096
2097 for (offset = 0; offset < BITS_IN_EWORD; ++offset) {
2098 if ((word >> offset) == 0)
2099 break;
2100
2101 offset += ewah_bit_ctz64(word >> offset);
2102
2103 bit_pos = reposition[pos + offset];
2104 if (bit_pos > 0)
2105 bitmap_set(dest, bit_pos - 1);
2106 else /* can't reuse, we don't have the object */
2107 return -1;
2108 }
2109
2110 pos += BITS_IN_EWORD;
2111 }
2112 return 0;
2113 }
2114
2115 uint32_t *create_bitmap_mapping(struct bitmap_index *bitmap_git,
2116 struct packing_data *mapping)
2117 {
2118 uint32_t i, num_objects;
2119 uint32_t *reposition;
2120
2121 if (!bitmap_is_midx(bitmap_git))
2122 load_reverse_index(bitmap_git);
2123 else if (load_midx_revindex(bitmap_git->midx) < 0)
2124 BUG("rebuild_existing_bitmaps: missing required rev-cache "
2125 "extension");
2126
2127 num_objects = bitmap_num_objects(bitmap_git);
2128 CALLOC_ARRAY(reposition, num_objects);
2129
2130 for (i = 0; i < num_objects; ++i) {
2131 struct object_id oid;
2132 struct object_entry *oe;
2133 uint32_t index_pos;
2134
2135 if (bitmap_is_midx(bitmap_git))
2136 index_pos = pack_pos_to_midx(bitmap_git->midx, i);
2137 else
2138 index_pos = pack_pos_to_index(bitmap_git->pack, i);
2139 nth_bitmap_object_oid(bitmap_git, &oid, index_pos);
2140 oe = packlist_find(mapping, &oid);
2141
2142 if (oe) {
2143 reposition[i] = oe_in_pack_pos(mapping, oe) + 1;
2144 if (bitmap_git->hashes && !oe->hash)
2145 oe->hash = get_be32(bitmap_git->hashes + index_pos);
2146 }
2147 }
2148
2149 return reposition;
2150 }
2151
2152 void free_bitmap_index(struct bitmap_index *b)
2153 {
2154 if (!b)
2155 return;
2156
2157 if (b->map)
2158 munmap(b->map, b->map_size);
2159 ewah_pool_free(b->commits);
2160 ewah_pool_free(b->trees);
2161 ewah_pool_free(b->blobs);
2162 ewah_pool_free(b->tags);
2163 if (b->bitmaps) {
2164 struct stored_bitmap *sb;
2165 kh_foreach_value(b->bitmaps, sb, {
2166 ewah_pool_free(sb->root);
2167 free(sb);
2168 });
2169 }
2170 kh_destroy_oid_map(b->bitmaps);
2171 free(b->ext_index.objects);
2172 free(b->ext_index.hashes);
2173 kh_destroy_oid_pos(b->ext_index.positions);
2174 bitmap_free(b->result);
2175 bitmap_free(b->haves);
2176 if (bitmap_is_midx(b)) {
2177 /*
2178 * Multi-pack bitmaps need to have resources associated with
2179 * their on-disk reverse indexes unmapped so that stale .rev and
2180 * .bitmap files can be removed.
2181 *
2182 * Unlike pack-based bitmaps, multi-pack bitmaps can be read and
2183 * written in the same 'git multi-pack-index write --bitmap'
2184 * process. Close resources so they can be removed safely on
2185 * platforms like Windows.
2186 */
2187 close_midx_revindex(b->midx);
2188 }
2189 free(b);
2190 }
2191
2192 int bitmap_has_oid_in_uninteresting(struct bitmap_index *bitmap_git,
2193 const struct object_id *oid)
2194 {
2195 return bitmap_git &&
2196 bitmap_walk_contains(bitmap_git, bitmap_git->haves, oid);
2197 }
2198
2199 static off_t get_disk_usage_for_type(struct bitmap_index *bitmap_git,
2200 enum object_type object_type)
2201 {
2202 struct bitmap *result = bitmap_git->result;
2203 off_t total = 0;
2204 struct ewah_iterator it;
2205 eword_t filter;
2206 size_t i;
2207
2208 init_type_iterator(&it, bitmap_git, object_type);
2209 for (i = 0; i < result->word_alloc &&
2210 ewah_iterator_next(&filter, &it); i++) {
2211 eword_t word = result->words[i] & filter;
2212 size_t base = (i * BITS_IN_EWORD);
2213 unsigned offset;
2214
2215 if (!word)
2216 continue;
2217
2218 for (offset = 0; offset < BITS_IN_EWORD; offset++) {
2219 if ((word >> offset) == 0)
2220 break;
2221
2222 offset += ewah_bit_ctz64(word >> offset);
2223
2224 if (bitmap_is_midx(bitmap_git)) {
2225 uint32_t pack_pos;
2226 uint32_t midx_pos = pack_pos_to_midx(bitmap_git->midx, base + offset);
2227 off_t offset = nth_midxed_offset(bitmap_git->midx, midx_pos);
2228
2229 uint32_t pack_id = nth_midxed_pack_int_id(bitmap_git->midx, midx_pos);
2230 struct packed_git *pack = bitmap_git->midx->packs[pack_id];
2231
2232 if (offset_to_pack_pos(pack, offset, &pack_pos) < 0) {
2233 struct object_id oid;
2234 nth_midxed_object_oid(&oid, bitmap_git->midx, midx_pos);
2235
2236 die(_("could not find '%s' in pack '%s' at offset %"PRIuMAX),
2237 oid_to_hex(&oid),
2238 pack->pack_name,
2239 (uintmax_t)offset);
2240 }
2241
2242 total += pack_pos_to_offset(pack, pack_pos + 1) - offset;
2243 } else {
2244 size_t pos = base + offset;
2245 total += pack_pos_to_offset(bitmap_git->pack, pos + 1) -
2246 pack_pos_to_offset(bitmap_git->pack, pos);
2247 }
2248 }
2249 }
2250
2251 return total;
2252 }
2253
2254 static off_t get_disk_usage_for_extended(struct bitmap_index *bitmap_git)
2255 {
2256 struct bitmap *result = bitmap_git->result;
2257 struct eindex *eindex = &bitmap_git->ext_index;
2258 off_t total = 0;
2259 struct object_info oi = OBJECT_INFO_INIT;
2260 off_t object_size;
2261 size_t i;
2262
2263 oi.disk_sizep = &object_size;
2264
2265 for (i = 0; i < eindex->count; i++) {
2266 struct object *obj = eindex->objects[i];
2267
2268 if (!bitmap_get(result, bitmap_num_objects(bitmap_git) + i))
2269 continue;
2270
2271 if (oid_object_info_extended(the_repository, &obj->oid, &oi, 0) < 0)
2272 die(_("unable to get disk usage of '%s'"),
2273 oid_to_hex(&obj->oid));
2274
2275 total += object_size;
2276 }
2277 return total;
2278 }
2279
2280 off_t get_disk_usage_from_bitmap(struct bitmap_index *bitmap_git,
2281 struct rev_info *revs)
2282 {
2283 off_t total = 0;
2284
2285 total += get_disk_usage_for_type(bitmap_git, OBJ_COMMIT);
2286 if (revs->tree_objects)
2287 total += get_disk_usage_for_type(bitmap_git, OBJ_TREE);
2288 if (revs->blob_objects)
2289 total += get_disk_usage_for_type(bitmap_git, OBJ_BLOB);
2290 if (revs->tag_objects)
2291 total += get_disk_usage_for_type(bitmap_git, OBJ_TAG);
2292
2293 total += get_disk_usage_for_extended(bitmap_git);
2294
2295 return total;
2296 }
2297
2298 int bitmap_is_midx(struct bitmap_index *bitmap_git)
2299 {
2300 return !!bitmap_git->midx;
2301 }
2302
2303 const struct string_list *bitmap_preferred_tips(struct repository *r)
2304 {
2305 return repo_config_get_value_multi(r, "pack.preferbitmaptips");
2306 }
2307
2308 int bitmap_is_preferred_refname(struct repository *r, const char *refname)
2309 {
2310 const struct string_list *preferred_tips = bitmap_preferred_tips(r);
2311 struct string_list_item *item;
2312
2313 if (!preferred_tips)
2314 return 0;
2315
2316 for_each_string_list_item(item, preferred_tips) {
2317 if (starts_with(refname, item->string))
2318 return 1;
2319 }
2320
2321 return 0;
2322 }
git for Debian