pack-bitmap.c · leroysheep/MovieWebApp

bisect

bisect jch maint master next seen test todo

c 3,420 lines 87.9 KB

Raw

1	#define DISABLE_SIGN_COMPARE_WARNINGS
2
3	#include "git-compat-util.h"
4	#include "commit.h"
5	#include "gettext.h"
6	#include "hex.h"
7	#include "strbuf.h"
8	#include "tag.h"
9	#include "diff.h"
10	#include "revision.h"
11	#include "progress.h"
12	#include "list-objects.h"
13	#include "pack.h"
14	#include "pack-bitmap.h"
15	#include "pack-revindex.h"
16	#include "pack-objects.h"
17	#include "packfile.h"
18	#include "repository.h"
19	#include "trace2.h"
20	#include "odb.h"
21	#include "list-objects-filter-options.h"
22	#include "midx.h"
23	#include "config.h"
24	#include "pseudo-merge.h"
25
26	/*
27	* An entry on the bitmap index, representing the bitmap for a given
28	* commit.
29	*/
30	struct stored_bitmap {
31	struct object_id oid;
32	struct ewah_bitmap *root;
33	struct stored_bitmap *xor;
34	size_t map_pos;
35	int flags;
36	};
37
38	/*
39	* The active bitmap index for a repository. By design, repositories only have
40	* a single bitmap index available (the index for the biggest packfile in
41	* the repository), since bitmap indexes need full closure.
42	*
43	* If there is more than one bitmap index available (e.g. because of alternates),
44	* the active bitmap index is the largest one.
45	*/
46	struct bitmap_index {
47	/*
48	* The pack or multi-pack index (MIDX) that this bitmap index belongs
49	* to.
50	*
51	* Exactly one of these must be non-NULL; this specifies the object
52	* order used to interpret this bitmap.
53	*/
54	struct packed_git *pack;
55	struct multi_pack_index *midx;
56
57	/*
58	* If using a multi-pack index chain, 'base' points to the
59	* bitmap index corresponding to this bitmap's midx->base_midx.
60	*
61	* base_nr indicates how many layers precede this one, and is
62	* zero when base is NULL.
63	*/
64	struct bitmap_index *base;
65	uint32_t base_nr;
66
67	/* mmapped buffer of the whole bitmap index */
68	unsigned char *map;
69	size_t map_size; /* size of the mmaped buffer */
70	size_t map_pos; /* current position when loading the index */
71
72	/*
73	* Type indexes.
74	*
75	* Each bitmap marks which objects in the packfile are of the given
76	* type. This provides type information when yielding the objects from
77	* the packfile during a walk, which allows for better delta bases.
78	*/
79	struct ewah_bitmap *commits;
80	struct ewah_bitmap *trees;
81	struct ewah_bitmap *blobs;
82	struct ewah_bitmap *tags;
83
84	/*
85	* Type index arrays when this bitmap is associated with an
86	* incremental multi-pack index chain.
87	*
88	* If n is the number of unique layers in the MIDX chain, then
89	* commits_all[n-1] is this structs 'commits' field,
90	* commits_all[n-2] is the commits field of this bitmap's
91	* 'base', and so on.
92	*
93	* When associated either with a non-incremental MIDX or a
94	* single packfile, these arrays each contain a single element.
95	*/
96	struct ewah_bitmap **commits_all;
97	struct ewah_bitmap **trees_all;
98	struct ewah_bitmap **blobs_all;
99	struct ewah_bitmap **tags_all;
100
101	/* Map from object ID -> `stored_bitmap` for all the bitmapped commits */
102	kh_oid_map_t *bitmaps;
103
104	/* Number of bitmapped commits */
105	uint32_t entry_count;
106
107	/* If not NULL, this is a name-hash cache pointing into map. */
108	uint32_t *hashes;
109
110	/* The checksum of the packfile or MIDX; points into map. */
111	const unsigned char *checksum;
112
113	/*
114	* If not NULL, this point into the commit table extension
115	* (within the memory mapped region `map`).
116	*/
117	unsigned char *table_lookup;
118
119	/* This contains the pseudo-merge cache within 'map' (if found). */
120	struct pseudo_merge_map pseudo_merges;
121
122	/*
123	* Extended index.
124	*
125	* When trying to perform bitmap operations with objects that are not
126	* packed in `pack`, these objects are added to this "fake index" and
127	* are assumed to appear at the end of the packfile for all operations
128	*/
129	struct eindex {
130	struct object **objects;
131	uint32_t *hashes;
132	uint32_t count, alloc;
133	kh_oid_pos_t *positions;
134	} ext_index;
135
136	/* Bitmap result of the last performed walk */
137	struct bitmap *result;
138
139	/* "have" bitmap from the last performed walk */
140	struct bitmap *haves;
141
142	/* Version of the bitmap index */
143	unsigned int version;
144	};
145
146	static int pseudo_merges_satisfied_nr;
147	static int pseudo_merges_cascades_nr;
148	static int existing_bitmaps_hits_nr;
149	static int existing_bitmaps_misses_nr;
150	static int roots_with_bitmaps_nr;
151	static int roots_without_bitmaps_nr;
152
153	static struct ewah_bitmap lookup_stored_bitmap(struct stored_bitmap st)
154	{
155	struct ewah_bitmap *parent;
156	struct ewah_bitmap *composed;
157
158	if (!st->xor)
159	return st->root;
160
161	composed = ewah_pool_new();
162	parent = lookup_stored_bitmap(st->xor);
163	ewah_xor(st->root, parent, composed);
164
165	ewah_pool_free(st->root);
166	st->root = composed;
167	st->xor = NULL;
168
169	return composed;
170	}
171
172	struct ewah_bitmap read_bitmap(const unsigned char map,
173	size_t map_size, size_t *map_pos)
174	{
175	struct ewah_bitmap *b = ewah_pool_new();
176
177	ssize_t bitmap_size = ewah_read_mmap(b, map + *map_pos,
178	map_size - *map_pos);
179
180	if (bitmap_size < 0) {
181	error(_("failed to load bitmap index (corrupted?)"));
182	ewah_pool_free(b);
183	return NULL;
184	}
185
186	*map_pos += bitmap_size;
187
188	return b;
189	}
190
191	/*
192	* Read a bitmap from the current read position on the mmaped
193	* index, and increase the read position accordingly
194	*/
195	static struct ewah_bitmap read_bitmap_1(struct bitmap_index index)
196	{
197	return read_bitmap(index->map, index->map_size, &index->map_pos);
198	}
199
200	static uint32_t bitmap_num_objects_total(struct bitmap_index *index)
201	{
202	if (index->midx) {
203	struct multi_pack_index *m = index->midx;
204	return m->num_objects + m->num_objects_in_base;
205	}
206	return index->pack->num_objects;
207	}
208
209	static uint32_t bitmap_num_objects(struct bitmap_index *index)
210	{
211	if (index->midx)
212	return index->midx->num_objects;
213	return index->pack->num_objects;
214	}
215
216	static uint32_t bitmap_name_hash(struct bitmap_index *index, uint32_t pos)
217	{
218	if (bitmap_is_midx(index)) {
219	while (index && pos < index->midx->num_objects_in_base) {
220	ASSERT(bitmap_is_midx(index));
221	index = index->base;
222	}
223
224	if (!index)
225	BUG("NULL base bitmap for object position: %"PRIu32, pos);
226
227	pos -= index->midx->num_objects_in_base;
228	if (pos >= index->midx->num_objects)
229	BUG("out-of-bounds midx bitmap object at %"PRIu32, pos);
230	}
231
232	if (!index->hashes)
233	return 0;
234
235	return get_be32(index->hashes + pos);
236	}
237
238	static struct repository bitmap_repo(struct bitmap_index bitmap_git)
239	{
240	if (bitmap_is_midx(bitmap_git))
241	return bitmap_git->midx->source->odb->repo;
242	return bitmap_git->pack->repo;
243	}
244
245	static int load_bitmap_header(struct bitmap_index *index)
246	{
247	struct bitmap_disk_header header = (void )index->map;
248	const struct git_hash_algo *hash_algo = bitmap_repo(index)->hash_algo;
249
250	size_t header_size = sizeof(*header) - GIT_MAX_RAWSZ + hash_algo->rawsz;
251
252	if (index->map_size < header_size + hash_algo->rawsz)
253	return error(_("corrupted bitmap index (too small)"));
254
255	if (memcmp(header->magic, BITMAP_IDX_SIGNATURE, sizeof(BITMAP_IDX_SIGNATURE)) != 0)
256	return error(_("corrupted bitmap index file (wrong header)"));
257
258	index->version = ntohs(header->version);
259	if (index->version != 1)
260	return error(_("unsupported version '%d' for bitmap index file"), index->version);
261
262	/* Parse known bitmap format options */
263	{
264	uint32_t flags = ntohs(header->options);
265	size_t cache_size = st_mult(bitmap_num_objects(index), sizeof(uint32_t));
266	unsigned char *index_end = index->map + index->map_size - hash_algo->rawsz;
267
268	if ((flags & BITMAP_OPT_FULL_DAG) == 0)
269	BUG("unsupported options for bitmap index file "
270	"(Git requires BITMAP_OPT_FULL_DAG)");
271
272	if (flags & BITMAP_OPT_HASH_CACHE) {
273	if (cache_size > index_end - index->map - header_size)
274	return error(_("corrupted bitmap index file (too short to fit hash cache)"));
275	index->hashes = (void *)(index_end - cache_size);
276	index_end -= cache_size;
277	}
278
279	if (flags & BITMAP_OPT_LOOKUP_TABLE) {
280	size_t table_size = st_mult(ntohl(header->entry_count),
281	BITMAP_LOOKUP_TABLE_TRIPLET_WIDTH);
282	if (table_size > index_end - index->map - header_size)
283	return error(_("corrupted bitmap index file (too short to fit lookup table)"));
284	if (git_env_bool("GIT_TEST_READ_COMMIT_TABLE", 1))
285	index->table_lookup = (void *)(index_end - table_size);
286	index_end -= table_size;
287	}
288
289	if (flags & BITMAP_OPT_PSEUDO_MERGES) {
290	unsigned char *pseudo_merge_ofs;
291	size_t table_size;
292	uint32_t i;
293
294	if (sizeof(table_size) > index_end - index->map - header_size)
295	return error(_("corrupted bitmap index file (too short to fit pseudo-merge table header)"));
296
297	table_size = get_be64(index_end - 8);
298	if (table_size > index_end - index->map - header_size)
299	return error(_("corrupted bitmap index file (too short to fit pseudo-merge table)"));
300
301	if (git_env_bool("GIT_TEST_USE_PSEUDO_MERGES", 1)) {
302	const unsigned char *ext = (index_end - table_size);
303
304	index->pseudo_merges.map = index->map;
305	index->pseudo_merges.map_size = index->map_size;
306	index->pseudo_merges.commits = ext + get_be64(index_end - 16);
307	index->pseudo_merges.commits_nr = get_be32(index_end - 20);
308	index->pseudo_merges.nr = get_be32(index_end - 24);
309
310	if (st_add(st_mult(index->pseudo_merges.nr,
311	sizeof(uint64_t)),
312	24) > table_size)
313	return error(_("corrupted bitmap index file, pseudo-merge table too short"));
314
315	CALLOC_ARRAY(index->pseudo_merges.v,
316	index->pseudo_merges.nr);
317
318	pseudo_merge_ofs = index_end - 24 -
319	(index->pseudo_merges.nr * sizeof(uint64_t));
320	for (i = 0; i < index->pseudo_merges.nr; i++) {
321	index->pseudo_merges.v[i].at = get_be64(pseudo_merge_ofs);
322	pseudo_merge_ofs += sizeof(uint64_t);
323	}
324	}
325
326	index_end -= table_size;
327	}
328	}
329
330	index->entry_count = ntohl(header->entry_count);
331	index->checksum = header->checksum;
332	index->map_pos += header_size;
333	return 0;
334	}
335
336	static struct stored_bitmap store_bitmap(struct bitmap_index index,
337	struct ewah_bitmap *root,
338	const struct object_id *oid,
339	struct stored_bitmap *xor_with,
340	int flags, size_t map_pos)
341	{
342	struct stored_bitmap *stored;
343	khiter_t hash_pos;
344	int ret;
345
346	stored = xmalloc(sizeof(struct stored_bitmap));
347	stored->map_pos = map_pos;
348	stored->root = root;
349	stored->xor = xor_with;
350	stored->flags = flags;
351	oidcpy(&stored->oid, oid);
352
353	hash_pos = kh_put_oid_map(index->bitmaps, stored->oid, &ret);
354
355	/*
356	* A 0 return code means the insertion succeeded with no changes,
357	* because the SHA1 already existed on the map. This is bad, there
358	* shouldn't be duplicated commits in the index.
359	*/
360	if (ret == 0) {
361	error(_("duplicate entry in bitmap index: '%s'"), oid_to_hex(oid));
362	return NULL;
363	}
364
365	kh_value(index->bitmaps, hash_pos) = stored;
366	return stored;
367	}
368
369	static inline uint32_t read_be32(const unsigned char buffer, size_t pos)
370	{
371	uint32_t result = get_be32(buffer + *pos);
372	(*pos) += sizeof(result);
373	return result;
374	}
375
376	static inline uint8_t read_u8(const unsigned char buffer, size_t pos)
377	{
378	return buffer[(*pos)++];
379	}
380
381	#define MAX_XOR_OFFSET 160
382
383	static int nth_bitmap_object_oid(struct bitmap_index *index,
384	struct object_id *oid,
385	uint32_t n)
386	{
387	if (index->midx)
388	return nth_midxed_object_oid(oid, index->midx, n) ? 0 : -1;
389	return nth_packed_object_id(oid, index->pack, n);
390	}
391
392	static int load_bitmap_entries_v1(struct bitmap_index *index)
393	{
394	uint32_t i;
395	struct stored_bitmap *recent_bitmaps[MAX_XOR_OFFSET] = { NULL };
396
397	for (i = 0; i < index->entry_count; ++i) {
398	int xor_offset, flags;
399	struct ewah_bitmap *bitmap = NULL;
400	struct stored_bitmap *xor_bitmap = NULL;
401	uint32_t commit_idx_pos;
402	struct object_id oid;
403	size_t entry_map_pos;
404
405	if (index->map_size - index->map_pos < 6)
406	return error(_("corrupt ewah bitmap: truncated header for entry %d"), i);
407
408	entry_map_pos = index->map_pos;
409	commit_idx_pos = read_be32(index->map, &index->map_pos);
410	xor_offset = read_u8(index->map, &index->map_pos);
411	flags = read_u8(index->map, &index->map_pos);
412
413	if (nth_bitmap_object_oid(index, &oid, commit_idx_pos) < 0)
414	return error(_("corrupt ewah bitmap: commit index %u out of range"),
415	(unsigned)commit_idx_pos);
416
417	if (xor_offset > MAX_XOR_OFFSET \|\| xor_offset > i)
418	return error(_("corrupted bitmap pack index"));
419
420	if (xor_offset > 0) {
421	xor_bitmap = recent_bitmaps[(i - xor_offset) % MAX_XOR_OFFSET];
422
423	if (!xor_bitmap)
424	return error(_("invalid XOR offset in bitmap pack index"));
425	}
426
427	bitmap = read_bitmap_1(index);
428	if (!bitmap)
429	return -1;
430
431	recent_bitmaps[i % MAX_XOR_OFFSET] =
432	store_bitmap(index, bitmap, &oid, xor_bitmap, flags,
433	entry_map_pos);
434	}
435
436	return 0;
437	}
438
439	char midx_bitmap_filename(struct multi_pack_index midx)
440	{
441	struct strbuf buf = STRBUF_INIT;
442	if (midx->has_chain)
443	get_split_midx_filename_ext(midx->source, &buf,
444	midx_get_checksum_hash(midx),
445	MIDX_EXT_BITMAP);
446	else
447	get_midx_filename_ext(midx->source, &buf,
448	midx_get_checksum_hash(midx),
449	MIDX_EXT_BITMAP);
450
451	return strbuf_detach(&buf, NULL);
452	}
453
454	char pack_bitmap_filename(struct packed_git p)
455	{
456	size_t len;
457
458	if (!strip_suffix(p->pack_name, ".pack", &len))
459	BUG("pack_name does not end in .pack");
460	return xstrfmt("%.*s.bitmap", (int)len, p->pack_name);
461	}
462
463	static int open_midx_bitmap_1(struct bitmap_index *bitmap_git,
464	struct multi_pack_index *midx)
465	{
466	struct stat st;
467	char *bitmap_name = midx_bitmap_filename(midx);
468	int fd = git_open(bitmap_name);
469	uint32_t i;
470
471	if (fd < 0) {
472	if (errno != ENOENT)
473	warning_errno("cannot open '%s'", bitmap_name);
474	free(bitmap_name);
475	return -1;
476	}
477	free(bitmap_name);
478
479	if (fstat(fd, &st)) {
480	error_errno(_("cannot fstat bitmap file"));
481	close(fd);
482	return -1;
483	}
484
485	if (bitmap_git->pack \|\| bitmap_git->midx) {
486	struct strbuf buf = STRBUF_INIT;
487	get_midx_filename(midx->source, &buf);
488	trace2_data_string("bitmap", bitmap_repo(bitmap_git),
489	"ignoring extra midx bitmap file", buf.buf);
490	close(fd);
491	strbuf_release(&buf);
492	return -1;
493	}
494
495	bitmap_git->midx = midx;
496	bitmap_git->map_size = xsize_t(st.st_size);
497	bitmap_git->map_pos = 0;
498	bitmap_git->map = xmmap(NULL, bitmap_git->map_size, PROT_READ,
499	MAP_PRIVATE, fd, 0);
500	close(fd);
501
502	if (load_bitmap_header(bitmap_git) < 0)
503	goto cleanup;
504
505	if (!hasheq(midx_get_checksum_hash(bitmap_git->midx), bitmap_git->checksum,
506	bitmap_repo(bitmap_git)->hash_algo)) {
507	error(_("checksum doesn't match in MIDX and bitmap"));
508	goto cleanup;
509	}
510
511	if (load_midx_revindex(bitmap_git->midx)) {
512	warning(_("multi-pack bitmap is missing required reverse index"));
513	goto cleanup;
514	}
515
516	for (i = 0; i < bitmap_git->midx->num_packs + bitmap_git->midx->num_packs_in_base; i++) {
517	if (prepare_midx_pack(bitmap_git->midx, i)) {
518	warning(_("could not open pack %s"),
519	bitmap_git->midx->pack_names[i]);
520	goto cleanup;
521	}
522	}
523
524	if (midx->base_midx) {
525	bitmap_git->base = prepare_midx_bitmap_git(midx->base_midx);
526	bitmap_git->base_nr = bitmap_git->base->base_nr + 1;
527	} else {
528	bitmap_git->base_nr = 0;
529	}
530
531	return 0;
532
533	cleanup:
534	munmap(bitmap_git->map, bitmap_git->map_size);
535	bitmap_git->map_size = 0;
536	bitmap_git->map_pos = 0;
537	bitmap_git->map = NULL;
538	bitmap_git->midx = NULL;
539	return -1;
540	}
541
542	static int open_pack_bitmap_1(struct bitmap_index bitmap_git, struct packed_git packfile)
543	{
544	int fd;
545	struct stat st;
546	char *bitmap_name;
547
548	bitmap_name = pack_bitmap_filename(packfile);
549	fd = git_open(bitmap_name);
550
551	if (fd < 0) {
552	if (errno != ENOENT)
553	warning_errno("cannot open '%s'", bitmap_name);
554	free(bitmap_name);
555	return -1;
556	}
557	free(bitmap_name);
558
559	if (fstat(fd, &st)) {
560	error_errno(_("cannot fstat bitmap file"));
561	close(fd);
562	return -1;
563	}
564
565	if (bitmap_git->pack \|\| bitmap_git->midx) {
566	trace2_data_string("bitmap", bitmap_repo(bitmap_git),
567	"ignoring extra bitmap file",
568	packfile->pack_name);
569	close(fd);
570	return -1;
571	}
572
573	if (!is_pack_valid(packfile)) {
574	close(fd);
575	return -1;
576	}
577
578	bitmap_git->pack = packfile;
579	bitmap_git->map_size = xsize_t(st.st_size);
580	bitmap_git->map = xmmap(NULL, bitmap_git->map_size, PROT_READ, MAP_PRIVATE, fd, 0);
581	bitmap_git->map_pos = 0;
582	bitmap_git->base_nr = 0;
583	close(fd);
584
585	if (load_bitmap_header(bitmap_git) < 0) {
586	munmap(bitmap_git->map, bitmap_git->map_size);
587	bitmap_git->map = NULL;
588	bitmap_git->map_size = 0;
589	bitmap_git->map_pos = 0;
590	bitmap_git->pack = NULL;
591	return -1;
592	}
593
594	trace2_data_string("bitmap", bitmap_repo(bitmap_git),
595	"opened bitmap file", packfile->pack_name);
596	return 0;
597	}
598
599	static int load_reverse_index(struct repository r, struct bitmap_index bitmap_git)
600	{
601	if (bitmap_is_midx(bitmap_git)) {
602	struct multi_pack_index *m;
603
604	/*
605	* The multi-pack-index's .rev file is already loaded via
606	* open_pack_bitmap_1().
607	*
608	* But we still need to open the individual pack .rev files,
609	* since we will need to make use of them in pack-objects.
610	*/
611	for (m = bitmap_git->midx; m; m = m->base_midx) {
612	uint32_t i;
613	int ret;
614
615	for (i = 0; i < m->num_packs; i++) {
616	ret = load_pack_revindex(r, m->packs[i]);
617	if (ret)
618	return ret;
619	}
620	}
621	return 0;
622	}
623	return load_pack_revindex(r, bitmap_git->pack);
624	}
625
626	static void load_all_type_bitmaps(struct bitmap_index *bitmap_git)
627	{
628	struct bitmap_index *curr = bitmap_git;
629	size_t i = bitmap_git->base_nr;
630
631	ALLOC_ARRAY(bitmap_git->commits_all, bitmap_git->base_nr + 1);
632	ALLOC_ARRAY(bitmap_git->trees_all, bitmap_git->base_nr + 1);
633	ALLOC_ARRAY(bitmap_git->blobs_all, bitmap_git->base_nr + 1);
634	ALLOC_ARRAY(bitmap_git->tags_all, bitmap_git->base_nr + 1);
635
636	while (curr) {
637	bitmap_git->commits_all[i] = curr->commits;
638	bitmap_git->trees_all[i] = curr->trees;
639	bitmap_git->blobs_all[i] = curr->blobs;
640	bitmap_git->tags_all[i] = curr->tags;
641
642	curr = curr->base;
643	if (curr && !i)
644	BUG("unexpected number of bitmap layers, expected %"PRIu32,
645	bitmap_git->base_nr + 1);
646	i -= 1;
647	}
648	}
649
650	static int load_bitmap(struct repository r, struct bitmap_index bitmap_git,
651	int recursing)
652	{
653	assert(bitmap_git->map);
654
655	bitmap_git->bitmaps = kh_init_oid_map();
656	bitmap_git->ext_index.positions = kh_init_oid_pos();
657
658	if (load_reverse_index(r, bitmap_git))
659	return -1;
660
661	if (!(bitmap_git->commits = read_bitmap_1(bitmap_git)) \|\|
662	!(bitmap_git->trees = read_bitmap_1(bitmap_git)) \|\|
663	!(bitmap_git->blobs = read_bitmap_1(bitmap_git)) \|\|
664	!(bitmap_git->tags = read_bitmap_1(bitmap_git)))
665	return -1;
666
667	if (!bitmap_git->table_lookup && load_bitmap_entries_v1(bitmap_git) < 0)
668	return -1;
669
670	if (bitmap_git->base) {
671	if (!bitmap_is_midx(bitmap_git))
672	BUG("non-MIDX bitmap has non-NULL base bitmap index");
673	if (load_bitmap(r, bitmap_git->base, 1) < 0)
674	return -1;
675	}
676
677	if (!recursing)
678	load_all_type_bitmaps(bitmap_git);
679
680	return 0;
681	}
682
683	static int open_pack_bitmap(struct repository *r,
684	struct bitmap_index *bitmap_git)
685	{
686	struct packed_git *p;
687	int ret = -1;
688
689	repo_for_each_pack(r, p) {
690	if (open_pack_bitmap_1(bitmap_git, p) == 0) {
691	ret = 0;
692	/*
693	* The only reason to keep looking is to report
694	* duplicates.
695	*/
696	if (!trace2_is_enabled())
697	break;
698	}
699	}
700
701	return ret;
702	}
703
704	static int open_midx_bitmap(struct repository *r,
705	struct bitmap_index *bitmap_git)
706	{
707	struct odb_source *source;
708	int ret = -1;
709
710	assert(!bitmap_git->map);
711
712	odb_prepare_alternates(r->objects);
713	for (source = r->objects->sources; source; source = source->next) {
714	struct multi_pack_index *midx = get_multi_pack_index(source);
715	if (midx && !open_midx_bitmap_1(bitmap_git, midx))
716	ret = 0;
717	}
718	return ret;
719	}
720
721	static int open_bitmap(struct repository *r,
722	struct bitmap_index *bitmap_git)
723	{
724	int found;
725
726	assert(!bitmap_git->map);
727
728	found = !open_midx_bitmap(r, bitmap_git);
729
730	/*
731	* these will all be skipped if we opened a midx bitmap; but run it
732	* anyway if tracing is enabled to report the duplicates
733	*/
734	if (!found \|\| trace2_is_enabled())
735	found \|= !open_pack_bitmap(r, bitmap_git);
736
737	return found ? 0 : -1;
738	}
739
740	struct bitmap_index prepare_bitmap_git(struct repository r)
741	{
742	struct bitmap_index bitmap_git = xcalloc(1, sizeof(bitmap_git));
743
744	if (!open_bitmap(r, bitmap_git) && !load_bitmap(r, bitmap_git, 0))
745	return bitmap_git;
746
747	free_bitmap_index(bitmap_git);
748	return NULL;
749	}
750
751	struct bitmap_index prepare_midx_bitmap_git(struct multi_pack_index midx)
752	{
753	struct bitmap_index bitmap_git = xcalloc(1, sizeof(bitmap_git));
754
755	if (!open_midx_bitmap_1(bitmap_git, midx))
756	return bitmap_git;
757
758	free_bitmap_index(bitmap_git);
759	return NULL;
760	}
761
762	int bitmap_index_contains_pack(struct bitmap_index bitmap, struct packed_git pack)
763	{
764	for (; bitmap; bitmap = bitmap->base) {
765	if (bitmap_is_midx(bitmap)) {
766	for (size_t i = 0; i < bitmap->midx->num_packs; i++)
767	if (bitmap->midx->packs[i] == pack)
768	return 1;
769	} else if (bitmap->pack == pack) {
770	return 1;
771	}
772	}
773
774	return 0;
775	}
776
777	struct include_data {
778	struct bitmap_index *bitmap_git;
779	struct bitmap *base;
780	struct bitmap *seen;
781	};
782
783	struct bitmap_lookup_table_triplet {
784	uint32_t commit_pos;
785	uint64_t offset;
786	uint32_t xor_row;
787	};
788
789	struct bitmap_lookup_table_xor_item {
790	struct object_id oid;
791	uint64_t offset;
792	};
793
794	/*
795	* Given a `triplet` struct pointer and pointer `p`, this
796	* function reads the triplet beginning at `p` into the struct.
797	* Note that this function assumes that there is enough memory
798	* left for filling the `triplet` struct from `p`.
799	*/
800	static int bitmap_lookup_table_get_triplet_by_pointer(struct bitmap_lookup_table_triplet *triplet,
801	const unsigned char *p)
802	{
803	if (!triplet)
804	return -1;
805
806	triplet->commit_pos = get_be32(p);
807	p += sizeof(uint32_t);
808	triplet->offset = get_be64(p);
809	p += sizeof(uint64_t);
810	triplet->xor_row = get_be32(p);
811	return 0;
812	}
813
814	/*
815	* This function gets the raw triplet from `row`'th row in the
816	* lookup table and fills that data to the `triplet`.
817	*/
818	static int bitmap_lookup_table_get_triplet(struct bitmap_index *bitmap_git,
819	uint32_t pos,
820	struct bitmap_lookup_table_triplet *triplet)
821	{
822	unsigned char *p = NULL;
823	if (pos >= bitmap_git->entry_count)
824	return error(_("corrupt bitmap lookup table: triplet position out of index"));
825
826	p = bitmap_git->table_lookup + st_mult(pos, BITMAP_LOOKUP_TABLE_TRIPLET_WIDTH);
827
828	return bitmap_lookup_table_get_triplet_by_pointer(triplet, p);
829	}
830
831	/*
832	* Searches for a matching triplet. `commit_pos` is a pointer
833	* to the wanted commit position value. `table_entry` points to
834	* a triplet in lookup table. The first 4 bytes of each
835	* triplet (pointed by `table_entry`) are compared with `*commit_pos`.
836	*/
837	static int triplet_cmp(const void commit_pos, const void table_entry)
838	{
839
840	uint32_t a = (uint32_t )commit_pos;
841	uint32_t b = get_be32(table_entry);
842	if (a > b)
843	return 1;
844	else if (a < b)
845	return -1;
846
847	return 0;
848	}
849
850	static uint32_t bitmap_bsearch_pos(struct bitmap_index *bitmap_git,
851	struct object_id *oid,
852	uint32_t *result)
853	{
854	int found;
855
856	if (bitmap_is_midx(bitmap_git))
857	found = bsearch_midx(oid, bitmap_git->midx, result);
858	else
859	found = bsearch_pack(oid, bitmap_git->pack, result);
860
861	return found;
862	}
863
864	/*
865	* `bsearch_triplet_by_pos` function searches for the raw triplet
866	* having commit position same as `commit_pos` and fills `triplet`
867	* object from the raw triplet. Returns 1 on success and 0 on
868	* failure.
869	*/
870	static int bitmap_bsearch_triplet_by_pos(uint32_t commit_pos,
871	struct bitmap_index *bitmap_git,
872	struct bitmap_lookup_table_triplet *triplet)
873	{
874	unsigned char *p = bsearch(&commit_pos, bitmap_git->table_lookup, bitmap_git->entry_count,
875	BITMAP_LOOKUP_TABLE_TRIPLET_WIDTH, triplet_cmp);
876
877	if (!p)
878	return -1;
879
880	return bitmap_lookup_table_get_triplet_by_pointer(triplet, p);
881	}
882
883	static struct stored_bitmap lazy_bitmap_for_commit(struct bitmap_index bitmap_git,
884	struct commit *commit)
885	{
886	uint32_t commit_pos, xor_row;
887	uint64_t offset;
888	int flags;
889	struct bitmap_lookup_table_triplet triplet;
890	struct object_id *oid = &commit->object.oid;
891	struct ewah_bitmap *bitmap;
892	struct stored_bitmap *xor_bitmap = NULL;
893	const int bitmap_header_size = 6;
894	static struct bitmap_lookup_table_xor_item *xor_items = NULL;
895	static size_t xor_items_nr = 0, xor_items_alloc = 0;
896	static int is_corrupt = 0;
897	int xor_flags;
898	khiter_t hash_pos;
899	struct bitmap_lookup_table_xor_item *xor_item;
900	size_t entry_map_pos;
901
902	if (is_corrupt)
903	return NULL;
904
905	if (!bitmap_bsearch_pos(bitmap_git, oid, &commit_pos))
906	return NULL;
907
908	if (bitmap_bsearch_triplet_by_pos(commit_pos, bitmap_git, &triplet) < 0)
909	return NULL;
910
911	xor_items_nr = 0;
912	offset = triplet.offset;
913	xor_row = triplet.xor_row;
914
915	while (xor_row != 0xffffffff) {
916	ALLOC_GROW(xor_items, xor_items_nr + 1, xor_items_alloc);
917
918	if (xor_items_nr + 1 >= bitmap_git->entry_count) {
919	error(_("corrupt bitmap lookup table: xor chain exceeds entry count"));
920	goto corrupt;
921	}
922
923	if (bitmap_lookup_table_get_triplet(bitmap_git, xor_row, &triplet) < 0)
924	goto corrupt;
925
926	xor_item = &xor_items[xor_items_nr];
927	xor_item->offset = triplet.offset;
928
929	if (nth_bitmap_object_oid(bitmap_git, &xor_item->oid, triplet.commit_pos) < 0) {
930	error(_("corrupt bitmap lookup table: commit index %u out of range"),
931	triplet.commit_pos);
932	goto corrupt;
933	}
934
935	hash_pos = kh_get_oid_map(bitmap_git->bitmaps, xor_item->oid);
936
937	/*
938	* If desired bitmap is already stored, we don't need
939	* to iterate further. Because we know that bitmaps
940	* that are needed to be parsed to parse this bitmap
941	* has already been stored. So, assign this stored bitmap
942	* to the xor_bitmap.
943	*/
944	if (hash_pos < kh_end(bitmap_git->bitmaps) &&
945	(xor_bitmap = kh_value(bitmap_git->bitmaps, hash_pos)))
946	break;
947	xor_items_nr++;
948	xor_row = triplet.xor_row;
949	}
950
951	while (xor_items_nr) {
952	xor_item = &xor_items[xor_items_nr - 1];
953	bitmap_git->map_pos = xor_item->offset;
954	if (bitmap_git->map_size - bitmap_git->map_pos < bitmap_header_size) {
955	error(_("corrupt ewah bitmap: truncated header for bitmap of commit \"%s\""),
956	oid_to_hex(&xor_item->oid));
957	goto corrupt;
958	}
959
960	entry_map_pos = bitmap_git->map_pos;
961	bitmap_git->map_pos += sizeof(uint32_t) + sizeof(uint8_t);
962	xor_flags = read_u8(bitmap_git->map, &bitmap_git->map_pos);
963	bitmap = read_bitmap_1(bitmap_git);
964
965	if (!bitmap)
966	goto corrupt;
967
968	xor_bitmap = store_bitmap(bitmap_git, bitmap, &xor_item->oid,
969	xor_bitmap, xor_flags, entry_map_pos);
970	xor_items_nr--;
971	}
972
973	bitmap_git->map_pos = offset;
974	if (bitmap_git->map_size - bitmap_git->map_pos < bitmap_header_size) {
975	error(_("corrupt ewah bitmap: truncated header for bitmap of commit \"%s\""),
976	oid_to_hex(oid));
977	goto corrupt;
978	}
979
980	/*
981	* Don't bother reading the commit's index position or its xor
982	* offset:
983	*
984	* - The commit's index position is irrelevant to us, since
985	* load_bitmap_entries_v1 only uses it to learn the object
986	* id which is used to compute the hashmap's key. We already
987	* have an object id, so no need to look it up again.
988	*
989	* - The xor_offset is unusable for us, since it specifies how
990	* many entries previous to ours we should look at. This
991	* makes sense when reading the bitmaps sequentially (as in
992	* load_bitmap_entries_v1()), since we can keep track of
993	* each bitmap as we read them.
994	*
995	* But it can't work for us, since the bitmap's don't have a
996	* fixed size. So we learn the position of the xor'd bitmap
997	* from the commit table (and resolve it to a bitmap in the
998	* above if-statement).
999	*
1000	* Instead, we can skip ahead and immediately read the flags and
1001	* ewah bitmap.
1002	*/
1003	entry_map_pos = bitmap_git->map_pos;
1004	bitmap_git->map_pos += sizeof(uint32_t) + sizeof(uint8_t);
1005	flags = read_u8(bitmap_git->map, &bitmap_git->map_pos);
1006	bitmap = read_bitmap_1(bitmap_git);
1007
1008	if (!bitmap)
1009	goto corrupt;
1010
1011	return store_bitmap(bitmap_git, bitmap, oid, xor_bitmap, flags,
1012	entry_map_pos);
1013
1014	corrupt:
1015	free(xor_items);
1016	is_corrupt = 1;
1017	return NULL;
1018	}
1019
1020	static struct ewah_bitmap find_bitmap_for_commit(struct bitmap_index bitmap_git,
1021	struct commit *commit,
1022	struct bitmap_index **found)
1023	{
1024	khiter_t hash_pos;
1025	if (!bitmap_git)
1026	return NULL;
1027
1028	hash_pos = kh_get_oid_map(bitmap_git->bitmaps, commit->object.oid);
1029	if (hash_pos >= kh_end(bitmap_git->bitmaps)) {
1030	struct stored_bitmap *bitmap = NULL;
1031	if (!bitmap_git->table_lookup)
1032	return find_bitmap_for_commit(bitmap_git->base, commit,
1033	found);
1034
1035	/* this is a fairly hot codepath - no trace2_region please */
1036	/* NEEDSWORK: cache misses aren't recorded */
1037	bitmap = lazy_bitmap_for_commit(bitmap_git, commit);
1038	if (!bitmap)
1039	return find_bitmap_for_commit(bitmap_git->base, commit,
1040	found);
1041	if (found)
1042	*found = bitmap_git;
1043	return lookup_stored_bitmap(bitmap);
1044	}
1045	if (found)
1046	*found = bitmap_git;
1047	return lookup_stored_bitmap(kh_value(bitmap_git->bitmaps, hash_pos));
1048	}
1049
1050	struct ewah_bitmap bitmap_for_commit(struct bitmap_index bitmap_git,
1051	struct commit *commit)
1052	{
1053	return find_bitmap_for_commit(bitmap_git, commit, NULL);
1054	}
1055
1056	static inline int bitmap_position_extended(struct bitmap_index *bitmap_git,
1057	const struct object_id *oid)
1058	{
1059	kh_oid_pos_t *positions = bitmap_git->ext_index.positions;
1060	khiter_t pos = kh_get_oid_pos(positions, *oid);
1061
1062	if (pos < kh_end(positions)) {
1063	int bitmap_pos = kh_value(positions, pos);
1064	return bitmap_pos + bitmap_num_objects_total(bitmap_git);
1065	}
1066
1067	return -1;
1068	}
1069
1070	static inline int bitmap_position_packfile(struct bitmap_index *bitmap_git,
1071	const struct object_id *oid)
1072	{
1073	uint32_t pos;
1074	off_t offset = find_pack_entry_one(oid, bitmap_git->pack);
1075	if (!offset)
1076	return -1;
1077
1078	if (offset_to_pack_pos(bitmap_git->pack, offset, &pos) < 0)
1079	return -1;
1080	return pos;
1081	}
1082
1083	static int bitmap_position_midx(struct bitmap_index *bitmap_git,
1084	const struct object_id *oid)
1085	{
1086	uint32_t want, got;
1087	if (!bsearch_midx(oid, bitmap_git->midx, &want))
1088	return -1;
1089
1090	if (midx_to_pack_pos(bitmap_git->midx, want, &got) < 0)
1091	return -1;
1092	return got;
1093	}
1094
1095	static int bitmap_position(struct bitmap_index *bitmap_git,
1096	const struct object_id *oid)
1097	{
1098	int pos;
1099	if (bitmap_is_midx(bitmap_git))
1100	pos = bitmap_position_midx(bitmap_git, oid);
1101	else
1102	pos = bitmap_position_packfile(bitmap_git, oid);
1103	return (pos >= 0) ? pos : bitmap_position_extended(bitmap_git, oid);
1104	}
1105
1106	static int ext_index_add_object(struct bitmap_index *bitmap_git,
1107	struct object object, const char name)
1108	{
1109	struct eindex *eindex = &bitmap_git->ext_index;
1110
1111	khiter_t hash_pos;
1112	int hash_ret;
1113	int bitmap_pos;
1114
1115	hash_pos = kh_put_oid_pos(eindex->positions, object->oid, &hash_ret);
1116	if (hash_ret > 0) {
1117	if (eindex->count >= eindex->alloc) {
1118	eindex->alloc = (eindex->alloc + 16) * 3 / 2;
1119	REALLOC_ARRAY(eindex->objects, eindex->alloc);
1120	REALLOC_ARRAY(eindex->hashes, eindex->alloc);
1121	}
1122
1123	bitmap_pos = eindex->count;
1124	eindex->objects[eindex->count] = object;
1125	eindex->hashes[eindex->count] = pack_name_hash(name);
1126	kh_value(eindex->positions, hash_pos) = bitmap_pos;
1127	eindex->count++;
1128	} else {
1129	bitmap_pos = kh_value(eindex->positions, hash_pos);
1130	}
1131
1132	return bitmap_pos + bitmap_num_objects_total(bitmap_git);
1133	}
1134
1135	struct bitmap_show_data {
1136	struct bitmap_index *bitmap_git;
1137	struct bitmap *base;
1138	};
1139
1140	static void show_object(struct object object, const char name, void *data_)
1141	{
1142	struct bitmap_show_data *data = data_;
1143	int bitmap_pos;
1144
1145	bitmap_pos = bitmap_position(data->bitmap_git, &object->oid);
1146
1147	if (bitmap_pos < 0)
1148	bitmap_pos = ext_index_add_object(data->bitmap_git, object,
1149	name);
1150
1151	bitmap_set(data->base, bitmap_pos);
1152	}
1153
1154	static void show_commit(struct commit *commit UNUSED,
1155	void *data UNUSED)
1156	{
1157	}
1158
1159	static unsigned apply_pseudo_merges_for_commit_1(struct bitmap_index *bitmap_git,
1160	struct bitmap *result,
1161	struct commit *commit,
1162	uint32_t commit_pos)
1163	{
1164	struct bitmap_index *curr = bitmap_git;
1165	int ret = 0;
1166
1167	while (curr) {
1168	ret += apply_pseudo_merges_for_commit(&curr->pseudo_merges,
1169	result, commit,
1170	commit_pos);
1171	curr = curr->base;
1172	}
1173
1174	if (ret)
1175	pseudo_merges_satisfied_nr += ret;
1176
1177	return ret;
1178	}
1179
1180	static int add_to_include_set(struct bitmap_index *bitmap_git,
1181	struct include_data *data,
1182	struct commit *commit,
1183	int bitmap_pos)
1184	{
1185	struct ewah_bitmap *partial;
1186
1187	if (data->seen && bitmap_get(data->seen, bitmap_pos))
1188	return 0;
1189
1190	if (bitmap_get(data->base, bitmap_pos))
1191	return 0;
1192
1193	partial = bitmap_for_commit(bitmap_git, commit);
1194	if (partial) {
1195	existing_bitmaps_hits_nr++;
1196
1197	bitmap_or_ewah(data->base, partial);
1198	return 0;
1199	}
1200
1201	existing_bitmaps_misses_nr++;
1202
1203	bitmap_set(data->base, bitmap_pos);
1204	if (apply_pseudo_merges_for_commit_1(bitmap_git, data->base, commit,
1205	bitmap_pos))
1206	return 0;
1207
1208	return 1;
1209	}
1210
1211	static int should_include(struct commit commit, void _data)
1212	{
1213	struct include_data *data = _data;
1214	int bitmap_pos;
1215
1216	bitmap_pos = bitmap_position(data->bitmap_git, &commit->object.oid);
1217	if (bitmap_pos < 0)
1218	bitmap_pos = ext_index_add_object(data->bitmap_git,
1219	(struct object *)commit,
1220	NULL);
1221
1222	if (!add_to_include_set(data->bitmap_git, data, commit, bitmap_pos)) {
1223	struct commit_list *parent = commit->parents;
1224
1225	while (parent) {
1226	parent->item->object.flags \|= SEEN;
1227	parent = parent->next;
1228	}
1229
1230	return 0;
1231	}
1232
1233	return 1;
1234	}
1235
1236	static int should_include_obj(struct object obj, void _data)
1237	{
1238	struct include_data *data = _data;
1239	int bitmap_pos;
1240
1241	bitmap_pos = bitmap_position(data->bitmap_git, &obj->oid);
1242	if (bitmap_pos < 0)
1243	return 1;
1244	if ((data->seen && bitmap_get(data->seen, bitmap_pos)) \|\|
1245	bitmap_get(data->base, bitmap_pos)) {
1246	obj->flags \|= SEEN;
1247	return 0;
1248	}
1249	return 1;
1250	}
1251
1252	static int add_commit_to_bitmap(struct bitmap_index *bitmap_git,
1253	struct bitmap **base,
1254	struct commit *commit)
1255	{
1256	struct ewah_bitmap *or_with = bitmap_for_commit(bitmap_git, commit);
1257
1258	if (!or_with) {
1259	existing_bitmaps_misses_nr++;
1260	return 0;
1261	}
1262
1263	existing_bitmaps_hits_nr++;
1264
1265	if (!*base)
1266	*base = ewah_to_bitmap(or_with);
1267	else
1268	bitmap_or_ewah(*base, or_with);
1269
1270	return 1;
1271	}
1272
1273	static struct bitmap fill_in_bitmap(struct bitmap_index bitmap_git,
1274	struct rev_info *revs,
1275	struct bitmap *base,
1276	struct bitmap *seen)
1277	{
1278	struct include_data incdata;
1279	struct bitmap_show_data show_data;
1280
1281	if (!base)
1282	base = bitmap_new();
1283
1284	incdata.bitmap_git = bitmap_git;
1285	incdata.base = base;
1286	incdata.seen = seen;
1287
1288	revs->include_check = should_include;
1289	revs->include_check_obj = should_include_obj;
1290	revs->include_check_data = &incdata;
1291
1292	if (prepare_revision_walk(revs))
1293	die(_("revision walk setup failed"));
1294
1295	show_data.bitmap_git = bitmap_git;
1296	show_data.base = base;
1297
1298	traverse_commit_list(revs, show_commit, show_object, &show_data);
1299
1300	revs->include_check = NULL;
1301	revs->include_check_obj = NULL;
1302	revs->include_check_data = NULL;
1303
1304	return base;
1305	}
1306
1307	struct bitmap_boundary_cb {
1308	struct bitmap_index *bitmap_git;
1309	struct bitmap *base;
1310
1311	struct object_array boundary;
1312	};
1313
1314	static void show_boundary_commit(struct commit commit, void _data)
1315	{
1316	struct bitmap_boundary_cb *data = _data;
1317
1318	if (commit->object.flags & BOUNDARY)
1319	add_object_array(&commit->object, "", &data->boundary);
1320
1321	if (commit->object.flags & UNINTERESTING) {
1322	if (bitmap_walk_contains(data->bitmap_git, data->base,
1323	&commit->object.oid))
1324	return;
1325
1326	add_commit_to_bitmap(data->bitmap_git, &data->base, commit);
1327	}
1328	}
1329
1330	static void show_boundary_object(struct object *object UNUSED,
1331	const char *name UNUSED,
1332	void *data UNUSED)
1333	{
1334	BUG("should not be called");
1335	}
1336
1337	static unsigned cascade_pseudo_merges_1(struct bitmap_index *bitmap_git,
1338	struct bitmap *result,
1339	struct bitmap *roots)
1340	{
1341	int ret = cascade_pseudo_merges(&bitmap_git->pseudo_merges,
1342	result, roots);
1343	if (ret) {
1344	pseudo_merges_cascades_nr++;
1345	pseudo_merges_satisfied_nr += ret;
1346	}
1347
1348	return ret;
1349	}
1350
1351	static struct bitmap find_boundary_objects(struct bitmap_index bitmap_git,
1352	struct rev_info *revs,
1353	struct object_list *roots)
1354	{
1355	struct bitmap_boundary_cb cb;
1356	struct object_list *root;
1357	struct repository *repo;
1358	unsigned int i;
1359	unsigned int tmp_blobs, tmp_trees, tmp_tags;
1360	int any_missing = 0;
1361	int existing_bitmaps = 0;
1362
1363	cb.bitmap_git = bitmap_git;
1364	cb.base = bitmap_new();
1365	object_array_init(&cb.boundary);
1366
1367	repo = bitmap_repo(bitmap_git);
1368
1369	revs->ignore_missing_links = 1;
1370
1371	if (bitmap_git->pseudo_merges.nr) {
1372	struct bitmap *roots_bitmap = bitmap_new();
1373	struct object_list *objects = NULL;
1374
1375	for (objects = roots; objects; objects = objects->next) {
1376	struct object *object = objects->item;
1377	int pos;
1378
1379	pos = bitmap_position(bitmap_git, &object->oid);
1380	if (pos < 0)
1381	continue;
1382
1383	bitmap_set(roots_bitmap, pos);
1384	}
1385
1386	cascade_pseudo_merges_1(bitmap_git, cb.base, roots_bitmap);
1387	bitmap_free(roots_bitmap);
1388	}
1389
1390	/*
1391	* OR in any existing reachability bitmaps among `roots` into
1392	* `cb.base`.
1393	*/
1394	for (root = roots; root; root = root->next) {
1395	struct object *object = root->item;
1396	if (object->type != OBJ_COMMIT \|\|
1397	bitmap_walk_contains(bitmap_git, cb.base, &object->oid))
1398	continue;
1399
1400	if (add_commit_to_bitmap(bitmap_git, &cb.base,
1401	(struct commit *)object)) {
1402	existing_bitmaps = 1;
1403	continue;
1404	}
1405
1406	any_missing = 1;
1407	}
1408
1409	if (!any_missing)
1410	goto cleanup;
1411
1412	if (existing_bitmaps)
1413	cascade_pseudo_merges_1(bitmap_git, cb.base, NULL);
1414
1415	tmp_blobs = revs->blob_objects;
1416	tmp_trees = revs->tree_objects;
1417	tmp_tags = revs->tag_objects;
1418	revs->blob_objects = 0;
1419	revs->tree_objects = 0;
1420	revs->tag_objects = 0;
1421
1422	/*
1423	* We didn't have complete coverage of the roots. First setup a
1424	* revision walk to (a) OR in any bitmaps that are UNINTERESTING
1425	* between the tips and boundary, and (b) record the boundary.
1426	*/
1427	trace2_region_enter("pack-bitmap", "boundary-prepare", repo);
1428	if (prepare_revision_walk(revs))
1429	die("revision walk setup failed");
1430	trace2_region_leave("pack-bitmap", "boundary-prepare", repo);
1431
1432	trace2_region_enter("pack-bitmap", "boundary-traverse", repo);
1433	revs->boundary = 1;
1434	traverse_commit_list_filtered(revs,
1435	show_boundary_commit,
1436	show_boundary_object,
1437	&cb, NULL);
1438	revs->boundary = 0;
1439	trace2_region_leave("pack-bitmap", "boundary-traverse", repo);
1440
1441	revs->blob_objects = tmp_blobs;
1442	revs->tree_objects = tmp_trees;
1443	revs->tag_objects = tmp_tags;
1444
1445	reset_revision_walk();
1446	clear_object_flags(repo, UNINTERESTING);
1447
1448	/*
1449	* Then add the boundary commit(s) as fill-in traversal tips.
1450	*/
1451	trace2_region_enter("pack-bitmap", "boundary-fill-in", repo);
1452	for (i = 0; i < cb.boundary.nr; i++) {
1453	struct object *obj = cb.boundary.objects[i].item;
1454	if (bitmap_walk_contains(bitmap_git, cb.base, &obj->oid))
1455	obj->flags \|= SEEN;
1456	else
1457	add_pending_object(revs, obj, "");
1458	}
1459	if (revs->pending.nr)
1460	cb.base = fill_in_bitmap(bitmap_git, revs, cb.base, NULL);
1461	trace2_region_leave("pack-bitmap", "boundary-fill-in", repo);
1462
1463	cleanup:
1464	object_array_clear(&cb.boundary);
1465	revs->ignore_missing_links = 0;
1466
1467	return cb.base;
1468	}
1469
1470	struct ewah_bitmap pseudo_merge_bitmap_for_commit(struct bitmap_index bitmap_git,
1471	struct commit *commit)
1472	{
1473	struct commit_list *p;
1474	struct bitmap *parents;
1475	struct pseudo_merge *match = NULL;
1476
1477	if (!bitmap_git->pseudo_merges.nr)
1478	return NULL;
1479
1480	parents = bitmap_new();
1481
1482	for (p = commit->parents; p; p = p->next) {
1483	int pos = bitmap_position(bitmap_git, &p->item->object.oid);
1484	if (pos < 0 \|\| pos >= bitmap_num_objects(bitmap_git))
1485	goto done;
1486
1487	/*
1488	* Use bitmap-relative positions instead of offsetting
1489	* by bitmap_git->num_objects_in_base because we use
1490	* this to find a match in pseudo_merge_for_parents(),
1491	* and pseudo-merge groups cannot span multiple bitmap
1492	* layers.
1493	*/
1494	bitmap_set(parents, pos);
1495	}
1496
1497	match = pseudo_merge_for_parents(&bitmap_git->pseudo_merges, parents);
1498
1499	done:
1500	bitmap_free(parents);
1501	if (match)
1502	return pseudo_merge_bitmap(&bitmap_git->pseudo_merges, match);
1503
1504	return NULL;
1505	}
1506
1507	static void unsatisfy_all_pseudo_merges(struct bitmap_index *bitmap_git)
1508	{
1509	uint32_t i;
1510	for (i = 0; i < bitmap_git->pseudo_merges.nr; i++)
1511	bitmap_git->pseudo_merges.v[i].satisfied = 0;
1512	}
1513
1514	static struct bitmap find_objects(struct bitmap_index bitmap_git,
1515	struct rev_info *revs,
1516	struct object_list *roots,
1517	struct bitmap *seen)
1518	{
1519	struct bitmap *base = NULL;
1520	int needs_walk = 0;
1521	unsigned existing_bitmaps = 0;
1522
1523	struct object_list *not_mapped = NULL;
1524
1525	unsatisfy_all_pseudo_merges(bitmap_git);
1526
1527	if (bitmap_git->pseudo_merges.nr) {
1528	struct bitmap *roots_bitmap = bitmap_new();
1529	struct object_list *objects = NULL;
1530
1531	for (objects = roots; objects; objects = objects->next) {
1532	struct object *object = objects->item;
1533	int pos;
1534
1535	pos = bitmap_position(bitmap_git, &object->oid);
1536	if (pos < 0)
1537	continue;
1538
1539	bitmap_set(roots_bitmap, pos);
1540	}
1541
1542	base = bitmap_new();
1543	cascade_pseudo_merges_1(bitmap_git, base, roots_bitmap);
1544	bitmap_free(roots_bitmap);
1545	}
1546
1547	/*
1548	* Go through all the roots for the walk. The ones that have bitmaps
1549	* on the bitmap index will be `or`ed together to form an initial
1550	* global reachability analysis.
1551	*
1552	* The ones without bitmaps in the index will be stored in the
1553	* `not_mapped_list` for further processing.
1554	*/
1555	while (roots) {
1556	struct object *object = roots->item;
1557
1558	roots = roots->next;
1559
1560	if (base) {
1561	int pos = bitmap_position(bitmap_git, &object->oid);
1562	if (pos > 0 && bitmap_get(base, pos)) {
1563	object->flags \|= SEEN;
1564	continue;
1565	}
1566	}
1567
1568	if (object->type == OBJ_COMMIT &&
1569	add_commit_to_bitmap(bitmap_git, &base, (struct commit *)object)) {
1570	object->flags \|= SEEN;
1571	existing_bitmaps = 1;
1572	continue;
1573	}
1574
1575	object_list_insert(object, &not_mapped);
1576	}
1577
1578	/*
1579	* Best case scenario: We found bitmaps for all the roots,
1580	* so the resulting `or` bitmap has the full reachability analysis
1581	*/
1582	if (!not_mapped)
1583	return base;
1584
1585	roots = not_mapped;
1586
1587	if (existing_bitmaps)
1588	cascade_pseudo_merges_1(bitmap_git, base, NULL);
1589
1590	/*
1591	* Let's iterate through all the roots that don't have bitmaps to
1592	* check if we can determine them to be reachable from the existing
1593	* global bitmap.
1594	*
1595	* If we cannot find them in the existing global bitmap, we'll need
1596	* to push them to an actual walk and run it until we can confirm
1597	* they are reachable
1598	*/
1599	while (roots) {
1600	struct object *object = roots->item;
1601	int pos;
1602
1603	roots = roots->next;
1604	pos = bitmap_position(bitmap_git, &object->oid);
1605
1606	if (pos < 0 \|\| base == NULL \|\| !bitmap_get(base, pos)) {
1607	object->flags &= ~UNINTERESTING;
1608	add_pending_object(revs, object, "");
1609	needs_walk = 1;
1610
1611	roots_without_bitmaps_nr++;
1612	} else {
1613	object->flags \|= SEEN;
1614
1615	roots_with_bitmaps_nr++;
1616	}
1617	}
1618
1619	if (needs_walk) {
1620	/*
1621	* This fill-in traversal may walk over some objects
1622	* again, since we have already traversed in order to
1623	* find the boundary.
1624	*
1625	* But this extra walk should be extremely cheap, since
1626	* all commit objects are loaded into memory, and
1627	* because we skip walking to parents that are
1628	* UNINTERESTING, since it will be marked in the haves
1629	* bitmap already (or it has an on-disk bitmap, since
1630	* OR-ing it in covers all of its ancestors).
1631	*/
1632	base = fill_in_bitmap(bitmap_git, revs, base, seen);
1633	}
1634
1635	object_list_free(&not_mapped);
1636
1637	return base;
1638	}
1639
1640	static void show_extended_objects(struct bitmap_index *bitmap_git,
1641	struct rev_info *revs,
1642	show_reachable_fn show_reach)
1643	{
1644	struct bitmap *objects = bitmap_git->result;
1645	struct eindex *eindex = &bitmap_git->ext_index;
1646	uint32_t i;
1647
1648	for (i = 0; i < eindex->count; ++i) {
1649	struct object *obj;
1650
1651	if (!bitmap_get(objects,
1652	st_add(bitmap_num_objects_total(bitmap_git),
1653	i)))
1654	continue;
1655
1656	obj = eindex->objects[i];
1657	if ((obj->type == OBJ_BLOB && !revs->blob_objects) \|\|
1658	(obj->type == OBJ_TREE && !revs->tree_objects) \|\|
1659	(obj->type == OBJ_TAG && !revs->tag_objects))
1660	continue;
1661
1662	show_reach(&obj->oid, obj->type, 0, eindex->hashes[i], NULL, 0, NULL);
1663	}
1664	}
1665
1666	static void init_type_iterator(struct ewah_or_iterator *it,
1667	struct bitmap_index *bitmap_git,
1668	enum object_type type)
1669	{
1670	switch (type) {
1671	case OBJ_COMMIT:
1672	ewah_or_iterator_init(it, bitmap_git->commits_all,
1673	bitmap_git->base_nr + 1);
1674	break;
1675
1676	case OBJ_TREE:
1677	ewah_or_iterator_init(it, bitmap_git->trees_all,
1678	bitmap_git->base_nr + 1);
1679	break;
1680
1681	case OBJ_BLOB:
1682	ewah_or_iterator_init(it, bitmap_git->blobs_all,
1683	bitmap_git->base_nr + 1);
1684	break;
1685
1686	case OBJ_TAG:
1687	ewah_or_iterator_init(it, bitmap_git->tags_all,
1688	bitmap_git->base_nr + 1);
1689	break;
1690
1691	default:
1692	BUG("object type %d not stored by bitmap type index", type);
1693	break;
1694	}
1695	}
1696
1697	static void show_objects_for_type(
1698	struct bitmap_index *bitmap_git,
1699	struct bitmap *objects,
1700	enum object_type object_type,
1701	show_reachable_fn show_reach,
1702	void *payload)
1703	{
1704	size_t i = 0;
1705	uint32_t offset;
1706
1707	struct ewah_or_iterator it;
1708	eword_t filter;
1709
1710	init_type_iterator(&it, bitmap_git, object_type);
1711
1712	for (i = 0; i < objects->word_alloc &&
1713	ewah_or_iterator_next(&filter, &it); i++) {
1714	eword_t word = objects->words[i] & filter;
1715	size_t pos = (i * BITS_IN_EWORD);
1716
1717	if (!word)
1718	continue;
1719
1720	for (offset = 0; offset < BITS_IN_EWORD; ++offset) {
1721	struct packed_git *pack;
1722	struct object_id oid;
1723	uint32_t hash = 0, index_pos;
1724	off_t ofs;
1725
1726	if ((word >> offset) == 0)
1727	break;
1728
1729	offset += ewah_bit_ctz64(word >> offset);
1730
1731	if (bitmap_is_midx(bitmap_git)) {
1732	struct multi_pack_index *m = bitmap_git->midx;
1733	uint32_t pack_id;
1734
1735	index_pos = pack_pos_to_midx(m, pos + offset);
1736	ofs = nth_midxed_offset(m, index_pos);
1737	nth_midxed_object_oid(&oid, m, index_pos);
1738
1739	pack_id = nth_midxed_pack_int_id(m, index_pos);
1740	pack = nth_midxed_pack(bitmap_git->midx, pack_id);
1741	} else {
1742	index_pos = pack_pos_to_index(bitmap_git->pack, pos + offset);
1743	ofs = pack_pos_to_offset(bitmap_git->pack, pos + offset);
1744	nth_bitmap_object_oid(bitmap_git, &oid, index_pos);
1745
1746	pack = bitmap_git->pack;
1747	}
1748
1749	hash = bitmap_name_hash(bitmap_git, index_pos);
1750
1751	show_reach(&oid, object_type, 0, hash, pack, ofs, payload);
1752	}
1753	}
1754
1755	ewah_or_iterator_release(&it);
1756	}
1757
1758	static int in_bitmapped_pack(struct bitmap_index *bitmap_git,
1759	struct object_list *roots)
1760	{
1761	while (roots) {
1762	struct object *object = roots->item;
1763	roots = roots->next;
1764
1765	if (bitmap_is_midx(bitmap_git)) {
1766	if (bsearch_midx(&object->oid, bitmap_git->midx, NULL))
1767	return 1;
1768	} else {
1769	if (find_pack_entry_one(&object->oid, bitmap_git->pack) > 0)
1770	return 1;
1771	}
1772	}
1773
1774	return 0;
1775	}
1776
1777	static struct bitmap find_tip_objects(struct bitmap_index bitmap_git,
1778	struct object_list *tip_objects,
1779	enum object_type type)
1780	{
1781	struct bitmap *result = bitmap_new();
1782	struct object_list *p;
1783
1784	for (p = tip_objects; p; p = p->next) {
1785	int pos;
1786
1787	if (p->item->type != type)
1788	continue;
1789
1790	pos = bitmap_position(bitmap_git, &p->item->oid);
1791	if (pos < 0)
1792	continue;
1793
1794	bitmap_set(result, pos);
1795	}
1796
1797	return result;
1798	}
1799
1800	static void filter_bitmap_exclude_type(struct bitmap_index *bitmap_git,
1801	struct object_list *tip_objects,
1802	struct bitmap *to_filter,
1803	enum object_type type)
1804	{
1805	struct eindex *eindex = &bitmap_git->ext_index;
1806	struct bitmap *tips;
1807	struct ewah_or_iterator it;
1808	eword_t mask;
1809	uint32_t i;
1810
1811	/*
1812	* The non-bitmap version of this filter never removes
1813	* objects which the other side specifically asked for,
1814	* so we must match that behavior.
1815	*/
1816	tips = find_tip_objects(bitmap_git, tip_objects, type);
1817
1818	/*
1819	* We can use the type-level bitmap for 'type' to work in whole
1820	* words for the objects that are actually in the bitmapped
1821	* packfile.
1822	*/
1823	for (i = 0, init_type_iterator(&it, bitmap_git, type);
1824	i < to_filter->word_alloc && ewah_or_iterator_next(&mask, &it);
1825	i++) {
1826	if (i < tips->word_alloc)
1827	mask &= ~tips->words[i];
1828	to_filter->words[i] &= ~mask;
1829	}
1830
1831	/*
1832	* Clear any objects that weren't in the packfile (and so would
1833	* not have been caught by the loop above. We'll have to check
1834	* them individually.
1835	*/
1836	for (i = 0; i < eindex->count; i++) {
1837	size_t pos = st_add(i, bitmap_num_objects_total(bitmap_git));
1838	if (eindex->objects[i]->type == type &&
1839	bitmap_get(to_filter, pos) &&
1840	!bitmap_get(tips, pos))
1841	bitmap_unset(to_filter, pos);
1842	}
1843
1844	ewah_or_iterator_release(&it);
1845	bitmap_free(tips);
1846	}
1847
1848	static void filter_bitmap_blob_none(struct bitmap_index *bitmap_git,
1849	struct object_list *tip_objects,
1850	struct bitmap *to_filter)
1851	{
1852	filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter,
1853	OBJ_BLOB);
1854	}
1855
1856	static unsigned long get_size_by_pos(struct bitmap_index *bitmap_git,
1857	uint32_t pos)
1858	{
1859	unsigned long size;
1860	struct object_info oi = OBJECT_INFO_INIT;
1861
1862	oi.sizep = &size;
1863
1864	if (pos < bitmap_num_objects_total(bitmap_git)) {
1865	struct packed_git *pack;
1866	off_t ofs;
1867
1868	if (bitmap_is_midx(bitmap_git)) {
1869	uint32_t midx_pos = pack_pos_to_midx(bitmap_git->midx, pos);
1870	uint32_t pack_id = nth_midxed_pack_int_id(bitmap_git->midx, midx_pos);
1871
1872	pack = nth_midxed_pack(bitmap_git->midx, pack_id);
1873	ofs = nth_midxed_offset(bitmap_git->midx, midx_pos);
1874	} else {
1875	pack = bitmap_git->pack;
1876	ofs = pack_pos_to_offset(pack, pos);
1877	}
1878
1879	if (packed_object_info(pack, ofs, &oi) < 0) {
1880	struct object_id oid;
1881	nth_bitmap_object_oid(bitmap_git, &oid,
1882	pack_pos_to_index(pack, pos));
1883	die(_("unable to get size of %s"), oid_to_hex(&oid));
1884	}
1885	} else {
1886	size_t eindex_pos = pos - bitmap_num_objects_total(bitmap_git);
1887	struct eindex *eindex = &bitmap_git->ext_index;
1888	struct object *obj = eindex->objects[eindex_pos];
1889	if (odb_read_object_info_extended(bitmap_repo(bitmap_git)->objects, &obj->oid,
1890	&oi, 0) < 0)
1891	die(_("unable to get size of %s"), oid_to_hex(&obj->oid));
1892	}
1893
1894	return size;
1895	}
1896
1897	static void filter_bitmap_blob_limit(struct bitmap_index *bitmap_git,
1898	struct object_list *tip_objects,
1899	struct bitmap *to_filter,
1900	unsigned long limit)
1901	{
1902	struct eindex *eindex = &bitmap_git->ext_index;
1903	struct bitmap *tips;
1904	struct ewah_or_iterator it;
1905	eword_t mask;
1906	uint32_t i;
1907
1908	tips = find_tip_objects(bitmap_git, tip_objects, OBJ_BLOB);
1909
1910	for (i = 0, init_type_iterator(&it, bitmap_git, OBJ_BLOB);
1911	i < to_filter->word_alloc && ewah_or_iterator_next(&mask, &it);
1912	i++) {
1913	eword_t word = to_filter->words[i] & mask;
1914	unsigned offset;
1915
1916	for (offset = 0; offset < BITS_IN_EWORD; offset++) {
1917	uint32_t pos;
1918
1919	if ((word >> offset) == 0)
1920	break;
1921	offset += ewah_bit_ctz64(word >> offset);
1922	pos = i * BITS_IN_EWORD + offset;
1923
1924	if (!bitmap_get(tips, pos) &&
1925	get_size_by_pos(bitmap_git, pos) >= limit)
1926	bitmap_unset(to_filter, pos);
1927	}
1928	}
1929
1930	for (i = 0; i < eindex->count; i++) {
1931	size_t pos = st_add(i, bitmap_num_objects(bitmap_git));
1932	if (eindex->objects[i]->type == OBJ_BLOB &&
1933	bitmap_get(to_filter, pos) &&
1934	!bitmap_get(tips, pos) &&
1935	get_size_by_pos(bitmap_git, pos) >= limit)
1936	bitmap_unset(to_filter, pos);
1937	}
1938
1939	ewah_or_iterator_release(&it);
1940	bitmap_free(tips);
1941	}
1942
1943	static void filter_bitmap_tree_depth(struct bitmap_index *bitmap_git,
1944	struct object_list *tip_objects,
1945	struct bitmap *to_filter,
1946	unsigned long limit)
1947	{
1948	if (limit)
1949	BUG("filter_bitmap_tree_depth given non-zero limit");
1950
1951	filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter,
1952	OBJ_TREE);
1953	filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter,
1954	OBJ_BLOB);
1955	}
1956
1957	static void filter_bitmap_object_type(struct bitmap_index *bitmap_git,
1958	struct object_list *tip_objects,
1959	struct bitmap *to_filter,
1960	enum object_type object_type)
1961	{
1962	if (object_type < OBJ_COMMIT \|\| object_type > OBJ_TAG)
1963	BUG("filter_bitmap_object_type given invalid object");
1964
1965	if (object_type != OBJ_TAG)
1966	filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, OBJ_TAG);
1967	if (object_type != OBJ_COMMIT)
1968	filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, OBJ_COMMIT);
1969	if (object_type != OBJ_TREE)
1970	filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, OBJ_TREE);
1971	if (object_type != OBJ_BLOB)
1972	filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, OBJ_BLOB);
1973	}
1974
1975	static int filter_bitmap(struct bitmap_index *bitmap_git,
1976	struct object_list *tip_objects,
1977	struct bitmap *to_filter,
1978	struct list_objects_filter_options *filter)
1979	{
1980	if (!filter \|\| filter->choice == LOFC_DISABLED)
1981	return 0;
1982
1983	if (filter->choice == LOFC_BLOB_NONE) {
1984	if (bitmap_git)
1985	filter_bitmap_blob_none(bitmap_git, tip_objects,
1986	to_filter);
1987	return 0;
1988	}
1989
1990	if (filter->choice == LOFC_BLOB_LIMIT) {
1991	if (bitmap_git)
1992	filter_bitmap_blob_limit(bitmap_git, tip_objects,
1993	to_filter,
1994	filter->blob_limit_value);
1995	return 0;
1996	}
1997
1998	if (filter->choice == LOFC_TREE_DEPTH &&
1999	filter->tree_exclude_depth == 0) {
2000	if (bitmap_git)
2001	filter_bitmap_tree_depth(bitmap_git, tip_objects,
2002	to_filter,
2003	filter->tree_exclude_depth);
2004	return 0;
2005	}
2006
2007	if (filter->choice == LOFC_OBJECT_TYPE) {
2008	if (bitmap_git)
2009	filter_bitmap_object_type(bitmap_git, tip_objects,
2010	to_filter,
2011	filter->object_type);
2012	return 0;
2013	}
2014
2015	if (filter->choice == LOFC_COMBINE) {
2016	int i;
2017	for (i = 0; i < filter->sub_nr; i++) {
2018	if (filter_bitmap(bitmap_git, tip_objects, to_filter,
2019	&filter->sub[i]) < 0)
2020	return -1;
2021	}
2022	return 0;
2023	}
2024
2025	/* filter choice not handled */
2026	return -1;
2027	}
2028
2029	static int can_filter_bitmap(struct list_objects_filter_options *filter)
2030	{
2031	return !filter_bitmap(NULL, NULL, NULL, filter);
2032	}
2033
2034
2035	static void filter_packed_objects_from_bitmap(struct bitmap_index *bitmap_git,
2036	struct bitmap *result)
2037	{
2038	struct eindex *eindex = &bitmap_git->ext_index;
2039	uint32_t objects_nr;
2040	size_t i, pos;
2041
2042	objects_nr = bitmap_num_objects_total(bitmap_git);
2043	pos = objects_nr / BITS_IN_EWORD;
2044
2045	if (pos > result->word_alloc)
2046	pos = result->word_alloc;
2047
2048	memset(result->words, 0x00, sizeof(eword_t) * pos);
2049	for (i = pos * BITS_IN_EWORD; i < objects_nr; i++)
2050	bitmap_unset(result, i);
2051
2052	for (i = 0; i < eindex->count; ++i) {
2053	if (has_object_pack(bitmap_repo(bitmap_git),
2054	&eindex->objects[i]->oid))
2055	bitmap_unset(result, objects_nr + i);
2056	}
2057	}
2058
2059	int for_each_bitmapped_object(struct bitmap_index *bitmap_git,
2060	struct list_objects_filter_options *filter,
2061	show_reachable_fn show_reach,
2062	void *payload)
2063	{
2064	struct bitmap *filtered_bitmap = NULL;
2065	uint32_t objects_nr;
2066	size_t full_word_count;
2067	int ret;
2068
2069	if (!can_filter_bitmap(filter)) {
2070	ret = -1;
2071	goto out;
2072	}
2073
2074	objects_nr = bitmap_num_objects(bitmap_git);
2075	full_word_count = objects_nr / BITS_IN_EWORD;
2076
2077	/* We start from the all-1 bitmap and then filter down from there. */
2078	filtered_bitmap = bitmap_word_alloc(full_word_count + !!(objects_nr % BITS_IN_EWORD));
2079	memset(filtered_bitmap->words, 0xff, full_word_count * sizeof(*filtered_bitmap->words));
2080	for (size_t i = full_word_count * BITS_IN_EWORD; i < objects_nr; i++)
2081	bitmap_set(filtered_bitmap, i);
2082
2083	if (filter_bitmap(bitmap_git, NULL, filtered_bitmap, filter) < 0) {
2084	ret = -1;
2085	goto out;
2086	}
2087
2088	show_objects_for_type(bitmap_git, filtered_bitmap,
2089	OBJ_COMMIT, show_reach, payload);
2090	show_objects_for_type(bitmap_git, filtered_bitmap,
2091	OBJ_TREE, show_reach, payload);
2092	show_objects_for_type(bitmap_git, filtered_bitmap,
2093	OBJ_BLOB, show_reach, payload);
2094	show_objects_for_type(bitmap_git, filtered_bitmap,
2095	OBJ_TAG, show_reach, payload);
2096
2097	ret = 0;
2098	out:
2099	bitmap_free(filtered_bitmap);
2100	return ret;
2101	}
2102
2103	struct bitmap_index prepare_bitmap_walk(struct rev_info revs,
2104	int filter_provided_objects)
2105	{
2106	unsigned int i;
2107	int use_boundary_traversal;
2108
2109	struct object_list *wants = NULL;
2110	struct object_list *haves = NULL;
2111
2112	struct bitmap *wants_bitmap = NULL;
2113	struct bitmap *haves_bitmap = NULL;
2114
2115	struct bitmap_index *bitmap_git;
2116	struct repository *repo;
2117
2118	/*
2119	* We can't do pathspec limiting with bitmaps, because we don't know
2120	* which commits are associated with which object changes (let alone
2121	* even which objects are associated with which paths).
2122	*/
2123	if (revs->prune)
2124	return NULL;
2125
2126	if (!can_filter_bitmap(&revs->filter))
2127	return NULL;
2128
2129	/* try to open a bitmapped pack, but don't parse it yet
2130	* because we may not need to use it */
2131	CALLOC_ARRAY(bitmap_git, 1);
2132	if (open_bitmap(revs->repo, bitmap_git) < 0)
2133	goto cleanup;
2134
2135	for (i = 0; i < revs->pending.nr; ++i) {
2136	struct object *object = revs->pending.objects[i].item;
2137
2138	if (object->type == OBJ_NONE)
2139	parse_object_or_die(revs->repo, &object->oid, NULL);
2140
2141	while (object->type == OBJ_TAG) {
2142	struct tag tag = (struct tag ) object;
2143
2144	if (object->flags & UNINTERESTING)
2145	object_list_insert(object, &haves);
2146	else
2147	object_list_insert(object, &wants);
2148
2149	object = parse_object_or_die(revs->repo, get_tagged_oid(tag), NULL);
2150	object->flags \|= (tag->object.flags & UNINTERESTING);
2151	}
2152
2153	if (object->flags & UNINTERESTING)
2154	object_list_insert(object, &haves);
2155	else
2156	object_list_insert(object, &wants);
2157	}
2158
2159	use_boundary_traversal = git_env_bool(GIT_TEST_PACK_USE_BITMAP_BOUNDARY_TRAVERSAL, -1);
2160	if (use_boundary_traversal < 0) {
2161	prepare_repo_settings(revs->repo);
2162	use_boundary_traversal = revs->repo->settings.pack_use_bitmap_boundary_traversal;
2163	}
2164
2165	if (!use_boundary_traversal) {
2166	/*
2167	* if we have a HAVES list, but none of those haves is contained
2168	* in the packfile that has a bitmap, we don't have anything to
2169	* optimize here
2170	*/
2171	if (haves && !in_bitmapped_pack(bitmap_git, haves))
2172	goto cleanup;
2173	}
2174
2175	/* if we don't want anything, we're done here */
2176	if (!wants)
2177	goto cleanup;
2178
2179	/*
2180	* now we're going to use bitmaps, so load the actual bitmap entries
2181	* from disk. this is the point of no return; after this the rev_list
2182	* becomes invalidated and we must perform the revwalk through bitmaps
2183	*/
2184	if (load_bitmap(revs->repo, bitmap_git, 0) < 0)
2185	goto cleanup;
2186
2187	if (!use_boundary_traversal)
2188	object_array_clear(&revs->pending);
2189
2190	repo = bitmap_repo(bitmap_git);
2191
2192	if (haves) {
2193	if (use_boundary_traversal) {
2194	trace2_region_enter("pack-bitmap", "haves/boundary", repo);
2195	haves_bitmap = find_boundary_objects(bitmap_git, revs, haves);
2196	trace2_region_leave("pack-bitmap", "haves/boundary", repo);
2197	} else {
2198	trace2_region_enter("pack-bitmap", "haves/classic", repo);
2199	revs->ignore_missing_links = 1;
2200	haves_bitmap = find_objects(bitmap_git, revs, haves, NULL);
2201	reset_revision_walk();
2202	revs->ignore_missing_links = 0;
2203	trace2_region_leave("pack-bitmap", "haves/classic", repo);
2204	}
2205
2206	if (!haves_bitmap)
2207	BUG("failed to perform bitmap walk");
2208	}
2209
2210	if (use_boundary_traversal) {
2211	object_array_clear(&revs->pending);
2212	reset_revision_walk();
2213	}
2214
2215	wants_bitmap = find_objects(bitmap_git, revs, wants, haves_bitmap);
2216
2217	if (!wants_bitmap)
2218	BUG("failed to perform bitmap walk");
2219
2220	if (haves_bitmap)
2221	bitmap_and_not(wants_bitmap, haves_bitmap);
2222
2223	filter_bitmap(bitmap_git,
2224	(revs->filter.choice && filter_provided_objects) ? NULL : wants,
2225	wants_bitmap,
2226	&revs->filter);
2227
2228	if (revs->unpacked)
2229	filter_packed_objects_from_bitmap(bitmap_git, wants_bitmap);
2230
2231	bitmap_git->result = wants_bitmap;
2232	bitmap_git->haves = haves_bitmap;
2233
2234	object_list_free(&wants);
2235	object_list_free(&haves);
2236
2237	trace2_data_intmax("bitmap", repo, "pseudo_merges_satisfied",
2238	pseudo_merges_satisfied_nr);
2239	trace2_data_intmax("bitmap", repo, "pseudo_merges_cascades",
2240	pseudo_merges_cascades_nr);
2241	trace2_data_intmax("bitmap", repo, "bitmap/hits",
2242	existing_bitmaps_hits_nr);
2243	trace2_data_intmax("bitmap", repo, "bitmap/misses",
2244	existing_bitmaps_misses_nr);
2245	trace2_data_intmax("bitmap", repo, "bitmap/roots_with_bitmap",
2246	roots_with_bitmaps_nr);
2247	trace2_data_intmax("bitmap", repo, "bitmap/roots_without_bitmap",
2248	roots_without_bitmaps_nr);
2249
2250	return bitmap_git;
2251
2252	cleanup:
2253	free_bitmap_index(bitmap_git);
2254	object_list_free(&wants);
2255	object_list_free(&haves);
2256	return NULL;
2257	}
2258
2259	/*
2260	* -1 means "stop trying further objects"; 0 means we may or may not have
2261	* reused, but you can keep feeding bits.
2262	*/
2263	static int try_partial_reuse(struct bitmap_index *bitmap_git,
2264	struct bitmapped_pack *pack,
2265	size_t bitmap_pos,
2266	uint32_t pack_pos,
2267	off_t offset,
2268	struct bitmap *reuse,
2269	struct pack_window **w_curs)
2270	{
2271	off_t delta_obj_offset;
2272	enum object_type type;
2273	size_t size;
2274
2275	if (pack_pos >= pack->p->num_objects)
2276	return -1; /* not actually in the pack */
2277
2278	delta_obj_offset = offset;
2279	type = unpack_object_header(pack->p, w_curs, &offset, &size);
2280	if (type < 0)
2281	return -1; /* broken packfile, punt */
2282
2283	if (type == OBJ_REF_DELTA \|\| type == OBJ_OFS_DELTA) {
2284	off_t base_offset;
2285	uint32_t base_pos;
2286	uint32_t base_bitmap_pos;
2287
2288	/*
2289	* Find the position of the base object so we can look it up
2290	* in our bitmaps. If we can't come up with an offset, or if
2291	* that offset is not in the revidx, the pack is corrupt.
2292	* There's nothing we can do, so just punt on this object,
2293	* and the normal slow path will complain about it in
2294	* more detail.
2295	*/
2296	base_offset = get_delta_base(pack->p, w_curs, &offset, type,
2297	delta_obj_offset);
2298	if (!base_offset)
2299	return 0;
2300
2301	offset_to_pack_pos(pack->p, base_offset, &base_pos);
2302
2303	if (bitmap_is_midx(bitmap_git)) {
2304	/*
2305	* Cross-pack deltas are rejected for now, but could
2306	* theoretically be supported in the future.
2307	*
2308	* We would need to ensure that we're sending both
2309	* halves of the delta/base pair, regardless of whether
2310	* or not the two cross a pack boundary. If they do,
2311	* then we must convert the delta to an REF_DELTA to
2312	* refer back to the base in the other pack.
2313	* */
2314	if (midx_pair_to_pack_pos(bitmap_git->midx,
2315	pack->pack_int_id,
2316	base_offset,
2317	&base_bitmap_pos) < 0) {
2318	return 0;
2319	}
2320	} else {
2321	if (offset_to_pack_pos(pack->p, base_offset,
2322	&base_pos) < 0)
2323	return 0;
2324	/*
2325	* We assume delta dependencies always point backwards.
2326	* This lets us do a single pass, and is basically
2327	* always true due to the way OFS_DELTAs work. You would
2328	* not typically find REF_DELTA in a bitmapped pack,
2329	* since we only bitmap packs we write fresh, and
2330	* OFS_DELTA is the default). But let's double check to
2331	* make sure the pack wasn't written with odd
2332	* parameters.
2333	*/
2334	if (base_pos >= pack_pos)
2335	return 0;
2336	base_bitmap_pos = pack->bitmap_pos + base_pos;
2337	}
2338
2339	/*
2340	* And finally, if we're not sending the base as part of our
2341	* reuse chunk, then don't send this object either. The base
2342	* would come after us, along with other objects not
2343	* necessarily in the pack, which means we'd need to convert
2344	* to REF_DELTA on the fly. Better to just let the normal
2345	* object_entry code path handle it.
2346	*/
2347	if (!bitmap_get(reuse, base_bitmap_pos))
2348	return 0;
2349	}
2350
2351	/*
2352	* If we got here, then the object is OK to reuse. Mark it.
2353	*/
2354	bitmap_set(reuse, bitmap_pos);
2355	return 0;
2356	}
2357
2358	static void reuse_partial_packfile_from_bitmap_1(struct bitmap_index *bitmap_git,
2359	struct bitmapped_pack *pack,
2360	struct bitmap *reuse)
2361	{
2362	struct bitmap *result = bitmap_git->result;
2363	struct pack_window *w_curs = NULL;
2364	size_t pos = pack->bitmap_pos / BITS_IN_EWORD;
2365
2366	if (!pack->bitmap_pos) {
2367	/*
2368	* If we're processing the first (in the case of a MIDX, the
2369	* preferred pack) or the only (in the case of single-pack
2370	* bitmaps) pack, then we can reuse whole words at a time.
2371	*
2372	* This is because we know that any deltas in this range must
2373	* have their bases chosen from the same pack, since:
2374	*
2375	* - In the single pack case, there is no other pack to choose
2376	* them from.
2377	*
2378	* - In the MIDX case, the first pack is the preferred pack, so
2379	* all ties are broken in favor of that pack (i.e. the one
2380	* we're currently processing). So any duplicate bases will be
2381	* resolved in favor of the pack we're processing.
2382	*/
2383	while (pos < result->word_alloc &&
2384	pos < pack->bitmap_nr / BITS_IN_EWORD &&
2385	result->words[pos] == (eword_t)~0)
2386	pos++;
2387	memset(reuse->words, 0xFF, pos * sizeof(eword_t));
2388	}
2389
2390	for (; pos < result->word_alloc; pos++) {
2391	eword_t word = result->words[pos];
2392	size_t offset;
2393
2394	for (offset = 0; offset < BITS_IN_EWORD; offset++) {
2395	size_t bit_pos;
2396	uint32_t pack_pos;
2397	off_t ofs;
2398
2399	if (word >> offset == 0)
2400	break;
2401
2402	offset += ewah_bit_ctz64(word >> offset);
2403
2404	bit_pos = pos * BITS_IN_EWORD + offset;
2405	if (bit_pos < pack->bitmap_pos)
2406	continue;
2407	if (bit_pos >= pack->bitmap_pos + pack->bitmap_nr)
2408	goto done;
2409
2410	if (bitmap_is_midx(bitmap_git)) {
2411	uint32_t midx_pos;
2412
2413	midx_pos = pack_pos_to_midx(bitmap_git->midx, bit_pos);
2414	ofs = nth_midxed_offset(bitmap_git->midx, midx_pos);
2415
2416	if (offset_to_pack_pos(pack->p, ofs, &pack_pos) < 0)
2417	BUG("could not find object in pack %s "
2418	"at offset %"PRIuMAX" in MIDX",
2419	pack_basename(pack->p), (uintmax_t)ofs);
2420	} else {
2421	pack_pos = cast_size_t_to_uint32_t(st_sub(bit_pos, pack->bitmap_pos));
2422	if (pack_pos >= pack->p->num_objects)
2423	BUG("advanced beyond the end of pack %s (%"PRIuMAX" > %"PRIu32")",
2424	pack_basename(pack->p), (uintmax_t)pack_pos,
2425	pack->p->num_objects);
2426
2427	ofs = pack_pos_to_offset(pack->p, pack_pos);
2428	}
2429
2430	if (try_partial_reuse(bitmap_git, pack, bit_pos,
2431	pack_pos, ofs, reuse, &w_curs) < 0) {
2432	/*
2433	* try_partial_reuse indicated we couldn't reuse
2434	* any bits, so there is no point in trying more
2435	* bits in the current word, or any other words
2436	* in result.
2437	*
2438	* Jump out of both loops to avoid future
2439	* unnecessary calls to try_partial_reuse.
2440	*/
2441	goto done;
2442	}
2443	}
2444	}
2445
2446	done:
2447	unuse_pack(&w_curs);
2448	}
2449
2450	static int bitmapped_pack_cmp(const void va, const void vb)
2451	{
2452	const struct bitmapped_pack *a = va;
2453	const struct bitmapped_pack *b = vb;
2454
2455	if (a->bitmap_pos < b->bitmap_pos)
2456	return -1;
2457	if (a->bitmap_pos > b->bitmap_pos)
2458	return 1;
2459	return 0;
2460	}
2461
2462	void reuse_partial_packfile_from_bitmap(struct bitmap_index *bitmap_git,
2463	struct bitmapped_pack **packs_out,
2464	size_t *packs_nr_out,
2465	struct bitmap **reuse_out,
2466	int multi_pack_reuse)
2467	{
2468	struct repository *r = bitmap_repo(bitmap_git);
2469	struct bitmapped_pack *packs = NULL;
2470	struct bitmap *result = bitmap_git->result;
2471	struct bitmap *reuse;
2472	size_t i;
2473	size_t packs_nr = 0, packs_alloc = 0;
2474	size_t word_alloc;
2475	uint32_t objects_nr = 0;
2476
2477	assert(result);
2478
2479	load_reverse_index(r, bitmap_git);
2480
2481	if (!bitmap_is_midx(bitmap_git) \|\| !bitmap_git->midx->chunk_bitmapped_packs)
2482	multi_pack_reuse = 0;
2483
2484	if (multi_pack_reuse) {
2485	struct multi_pack_index *m = bitmap_git->midx;
2486	for (i = 0; i < m->num_packs + m->num_packs_in_base; i++) {
2487	struct bitmapped_pack pack;
2488	if (nth_bitmapped_pack(bitmap_git->midx, &pack, i) < 0) {
2489	warning(_("unable to load pack: '%s', disabling pack-reuse"),
2490	bitmap_git->midx->pack_names[i]);
2491	free(packs);
2492	return;
2493	}
2494
2495	if (!pack.bitmap_nr)
2496	continue;
2497
2498	if (is_pack_valid(pack.p)) {
2499	ALLOC_GROW(packs, packs_nr + 1, packs_alloc);
2500	memcpy(&packs[packs_nr++], &pack, sizeof(pack));
2501	}
2502
2503	objects_nr += pack.p->num_objects;
2504	}
2505
2506	QSORT(packs, packs_nr, bitmapped_pack_cmp);
2507	} else {
2508	struct packed_git *pack;
2509	uint32_t pack_int_id;
2510
2511	if (bitmap_is_midx(bitmap_git)) {
2512	struct multi_pack_index *m = bitmap_git->midx;
2513	uint32_t preferred_pack_pos;
2514
2515	while (m->base_midx)
2516	m = m->base_midx;
2517
2518	if (midx_preferred_pack(m, &preferred_pack_pos) < 0) {
2519	warning(_("unable to compute preferred pack, disabling pack-reuse"));
2520	return;
2521	}
2522
2523	pack = nth_midxed_pack(m, preferred_pack_pos);
2524	pack_int_id = preferred_pack_pos;
2525	} else {
2526	pack = bitmap_git->pack;
2527	/*
2528	* Any value for 'pack_int_id' will do here. When we
2529	* process the pack via try_partial_reuse(), we won't
2530	* use the `pack_int_id` field since we have a non-MIDX
2531	* bitmap.
2532	*
2533	* Use '-1' as a sentinel value to make it clear
2534	* that we do not expect to read this field.
2535	*/
2536	pack_int_id = -1;
2537	}
2538
2539	if (is_pack_valid(pack)) {
2540	ALLOC_GROW(packs, packs_nr + 1, packs_alloc);
2541	packs[packs_nr].p = pack;
2542	packs[packs_nr].pack_int_id = pack_int_id;
2543	packs[packs_nr].bitmap_nr = pack->num_objects;
2544	packs[packs_nr].bitmap_pos = 0;
2545	packs[packs_nr].from_midx = bitmap_git->midx;
2546	packs_nr++;
2547	}
2548
2549	objects_nr = pack->num_objects;
2550	}
2551
2552	if (!packs_nr)
2553	return;
2554
2555	word_alloc = objects_nr / BITS_IN_EWORD;
2556	if (objects_nr % BITS_IN_EWORD)
2557	word_alloc++;
2558	reuse = bitmap_word_alloc(word_alloc);
2559
2560	for (i = 0; i < packs_nr; i++)
2561	reuse_partial_packfile_from_bitmap_1(bitmap_git, &packs[i], reuse);
2562
2563	if (bitmap_is_empty(reuse)) {
2564	free(packs);
2565	bitmap_free(reuse);
2566	return;
2567	}
2568
2569	/*
2570	* Drop any reused objects from the result, since they will not
2571	* need to be handled separately.
2572	*/
2573	bitmap_and_not(result, reuse);
2574	*packs_out = packs;
2575	*packs_nr_out = packs_nr;
2576	*reuse_out = reuse;
2577	}
2578
2579	int bitmap_walk_contains(struct bitmap_index *bitmap_git,
2580	struct bitmap bitmap, const struct object_id oid)
2581	{
2582	int idx;
2583
2584	if (!bitmap)
2585	return 0;
2586
2587	idx = bitmap_position(bitmap_git, oid);
2588	return idx >= 0 && bitmap_get(bitmap, idx);
2589	}
2590
2591	void traverse_bitmap_commit_list(struct bitmap_index *bitmap_git,
2592	struct rev_info *revs,
2593	show_reachable_fn show_reachable)
2594	{
2595	assert(bitmap_git->result);
2596
2597	show_objects_for_type(bitmap_git, bitmap_git->result,
2598	OBJ_COMMIT, show_reachable, NULL);
2599	if (revs->tree_objects)
2600	show_objects_for_type(bitmap_git, bitmap_git->result,
2601	OBJ_TREE, show_reachable, NULL);
2602	if (revs->blob_objects)
2603	show_objects_for_type(bitmap_git, bitmap_git->result,
2604	OBJ_BLOB, show_reachable, NULL);
2605	if (revs->tag_objects)
2606	show_objects_for_type(bitmap_git, bitmap_git->result,
2607	OBJ_TAG, show_reachable, NULL);
2608
2609	show_extended_objects(bitmap_git, revs, show_reachable);
2610	}
2611
2612	static uint32_t count_object_type(struct bitmap_index *bitmap_git,
2613	enum object_type type)
2614	{
2615	struct bitmap *objects = bitmap_git->result;
2616	struct eindex *eindex = &bitmap_git->ext_index;
2617
2618	uint32_t i = 0, count = 0;
2619	struct ewah_or_iterator it;
2620	eword_t filter;
2621
2622	init_type_iterator(&it, bitmap_git, type);
2623
2624	while (i < objects->word_alloc && ewah_or_iterator_next(&filter, &it)) {
2625	eword_t word = objects->words[i++] & filter;
2626	count += ewah_bit_popcount64(word);
2627	}
2628
2629	for (i = 0; i < eindex->count; ++i) {
2630	if (eindex->objects[i]->type == type &&
2631	bitmap_get(objects,
2632	st_add(bitmap_num_objects_total(bitmap_git), i)))
2633	count++;
2634	}
2635
2636	ewah_or_iterator_release(&it);
2637
2638	return count;
2639	}
2640
2641	void count_bitmap_commit_list(struct bitmap_index *bitmap_git,
2642	uint32_t commits, uint32_t trees,
2643	uint32_t blobs, uint32_t tags)
2644	{
2645	assert(bitmap_git->result);
2646
2647	if (commits)
2648	*commits = count_object_type(bitmap_git, OBJ_COMMIT);
2649
2650	if (trees)
2651	*trees = count_object_type(bitmap_git, OBJ_TREE);
2652
2653	if (blobs)
2654	*blobs = count_object_type(bitmap_git, OBJ_BLOB);
2655
2656	if (tags)
2657	*tags = count_object_type(bitmap_git, OBJ_TAG);
2658	}
2659
2660	struct bitmap_test_data {
2661	struct bitmap_index *bitmap_git;
2662	struct bitmap *base;
2663	struct bitmap *commits;
2664	struct bitmap *trees;
2665	struct bitmap *blobs;
2666	struct bitmap *tags;
2667	struct progress *prg;
2668	size_t seen;
2669
2670	struct bitmap_test_data *base_tdata;
2671	};
2672
2673	static void test_bitmap_type(struct bitmap_test_data *tdata,
2674	struct object *obj, int pos)
2675	{
2676	enum object_type bitmap_type = OBJ_NONE;
2677	int bitmaps_nr = 0;
2678
2679	if (bitmap_is_midx(tdata->bitmap_git)) {
2680	while (pos < tdata->bitmap_git->midx->num_objects_in_base)
2681	tdata = tdata->base_tdata;
2682	}
2683
2684	if (bitmap_get(tdata->commits, pos)) {
2685	bitmap_type = OBJ_COMMIT;
2686	bitmaps_nr++;
2687	}
2688	if (bitmap_get(tdata->trees, pos)) {
2689	bitmap_type = OBJ_TREE;
2690	bitmaps_nr++;
2691	}
2692	if (bitmap_get(tdata->blobs, pos)) {
2693	bitmap_type = OBJ_BLOB;
2694	bitmaps_nr++;
2695	}
2696	if (bitmap_get(tdata->tags, pos)) {
2697	bitmap_type = OBJ_TAG;
2698	bitmaps_nr++;
2699	}
2700
2701	if (bitmap_type == OBJ_NONE)
2702	die(_("object '%s' not found in type bitmaps"),
2703	oid_to_hex(&obj->oid));
2704
2705	if (bitmaps_nr > 1)
2706	die(_("object '%s' does not have a unique type"),
2707	oid_to_hex(&obj->oid));
2708
2709	if (bitmap_type != obj->type)
2710	die(_("object '%s': real type '%s', expected: '%s'"),
2711	oid_to_hex(&obj->oid),
2712	type_name(obj->type),
2713	type_name(bitmap_type));
2714	}
2715
2716	static void test_show_object(struct object *object,
2717	const char *name UNUSED,
2718	void *data)
2719	{
2720	struct bitmap_test_data *tdata = data;
2721	int bitmap_pos;
2722
2723	bitmap_pos = bitmap_position(tdata->bitmap_git, &object->oid);
2724	if (bitmap_pos < 0)
2725	die(_("object not in bitmap: '%s'"), oid_to_hex(&object->oid));
2726	test_bitmap_type(tdata, object, bitmap_pos);
2727
2728	bitmap_set(tdata->base, bitmap_pos);
2729	display_progress(tdata->prg, ++tdata->seen);
2730	}
2731
2732	static void test_show_commit(struct commit commit, void data)
2733	{
2734	struct bitmap_test_data *tdata = data;
2735	int bitmap_pos;
2736
2737	bitmap_pos = bitmap_position(tdata->bitmap_git,
2738	&commit->object.oid);
2739	if (bitmap_pos < 0)
2740	die(_("object not in bitmap: '%s'"), oid_to_hex(&commit->object.oid));
2741	test_bitmap_type(tdata, &commit->object, bitmap_pos);
2742
2743	bitmap_set(tdata->base, bitmap_pos);
2744	display_progress(tdata->prg, ++tdata->seen);
2745	}
2746
2747	static uint32_t bitmap_total_entry_count(struct bitmap_index *bitmap_git)
2748	{
2749	uint32_t total = 0;
2750	do {
2751	total = st_add(total, bitmap_git->entry_count);
2752	bitmap_git = bitmap_git->base;
2753	} while (bitmap_git);
2754
2755	return total;
2756	}
2757
2758	static void bitmap_test_data_prepare(struct bitmap_test_data *tdata,
2759	struct bitmap_index *bitmap_git)
2760	{
2761	memset(tdata, 0, sizeof(struct bitmap_test_data));
2762
2763	tdata->bitmap_git = bitmap_git;
2764	tdata->base = bitmap_new();
2765	tdata->commits = ewah_to_bitmap(bitmap_git->commits);
2766	tdata->trees = ewah_to_bitmap(bitmap_git->trees);
2767	tdata->blobs = ewah_to_bitmap(bitmap_git->blobs);
2768	tdata->tags = ewah_to_bitmap(bitmap_git->tags);
2769
2770	if (bitmap_git->base) {
2771	tdata->base_tdata = xmalloc(sizeof(struct bitmap_test_data));
2772	bitmap_test_data_prepare(tdata->base_tdata, bitmap_git->base);
2773	}
2774	}
2775
2776	static void bitmap_test_data_release(struct bitmap_test_data *tdata)
2777	{
2778	if (!tdata)
2779	return;
2780
2781	bitmap_test_data_release(tdata->base_tdata);
2782	free(tdata->base_tdata);
2783
2784	bitmap_free(tdata->base);
2785	bitmap_free(tdata->commits);
2786	bitmap_free(tdata->trees);
2787	bitmap_free(tdata->blobs);
2788	bitmap_free(tdata->tags);
2789	}
2790
2791	void test_bitmap_walk(struct rev_info *revs)
2792	{
2793	struct object *root;
2794	struct bitmap *result = NULL;
2795	size_t result_popcnt;
2796	struct bitmap_test_data tdata;
2797	struct bitmap_index bitmap_git, found;
2798	struct ewah_bitmap *bm;
2799
2800	if (!(bitmap_git = prepare_bitmap_git(revs->repo)))
2801	die(_("failed to load bitmap indexes"));
2802
2803	if (revs->pending.nr != 1)
2804	die(_("you must specify exactly one commit to test"));
2805
2806	fprintf_ln(stderr, "Bitmap v%d test (%d entries%s, %d total)",
2807	bitmap_git->version,
2808	bitmap_git->entry_count,
2809	bitmap_git->table_lookup ? "" : " loaded",
2810	bitmap_total_entry_count(bitmap_git));
2811
2812	root = revs->pending.objects[0].item;
2813	bm = find_bitmap_for_commit(bitmap_git, (struct commit *)root, &found);
2814
2815	if (bm) {
2816	fprintf_ln(stderr, "Found bitmap for '%s'. %d bits / %08x checksum",
2817	oid_to_hex(&root->oid),
2818	(int)bm->bit_size, ewah_checksum(bm));
2819
2820	if (bitmap_is_midx(found))
2821	fprintf_ln(stderr, "Located via MIDX '%s'.",
2822	midx_get_checksum_hex(found->midx));
2823	else
2824	fprintf_ln(stderr, "Located via pack '%s'.",
2825	hash_to_hex_algop(found->pack->hash,
2826	revs->repo->hash_algo));
2827
2828	result = ewah_to_bitmap(bm);
2829	}
2830
2831	if (!result)
2832	die(_("commit '%s' doesn't have an indexed bitmap"), oid_to_hex(&root->oid));
2833
2834	revs->tag_objects = 1;
2835	revs->tree_objects = 1;
2836	revs->blob_objects = 1;
2837
2838	result_popcnt = bitmap_popcount(result);
2839
2840	if (prepare_revision_walk(revs))
2841	die(_("revision walk setup failed"));
2842
2843	bitmap_test_data_prepare(&tdata, bitmap_git);
2844	tdata.prg = start_progress(revs->repo,
2845	"Verifying bitmap entries",
2846	result_popcnt);
2847
2848	traverse_commit_list(revs, &test_show_commit, &test_show_object, &tdata);
2849
2850	stop_progress(&tdata.prg);
2851
2852	if (bitmap_equals(result, tdata.base))
2853	fprintf_ln(stderr, "OK!");
2854	else
2855	die(_("mismatch in bitmap results"));
2856
2857	bitmap_free(result);
2858	bitmap_test_data_release(&tdata);
2859	free_bitmap_index(bitmap_git);
2860	}
2861
2862	int test_bitmap_commits(struct repository *r)
2863	{
2864	struct object_id oid;
2865	MAYBE_UNUSED void *value;
2866	struct bitmap_index *bitmap_git = prepare_bitmap_git(r);
2867
2868	if (!bitmap_git)
2869	die(_("failed to load bitmap indexes"));
2870
2871	/*
2872	* Since this function needs to print the bitmapped
2873	* commits, bypass the commit lookup table (if one exists)
2874	* by forcing the bitmap to eagerly load its entries.
2875	*/
2876	if (bitmap_git->table_lookup) {
2877	if (load_bitmap_entries_v1(bitmap_git) < 0)
2878	die(_("failed to load bitmap indexes"));
2879	}
2880
2881	kh_foreach(bitmap_git->bitmaps, oid, value, {
2882	printf_ln("%s", oid_to_hex(&oid));
2883	});
2884
2885	free_bitmap_index(bitmap_git);
2886
2887	return 0;
2888	}
2889
2890	int test_bitmap_commits_with_offset(struct repository *r)
2891	{
2892	struct object_id oid;
2893	struct stored_bitmap *stored;
2894	struct bitmap_index *bitmap_git;
2895	size_t commit_idx_pos_map_pos, xor_offset_map_pos, flag_map_pos,
2896	ewah_bitmap_map_pos;
2897
2898	bitmap_git = prepare_bitmap_git(r);
2899	if (!bitmap_git)
2900	die(_("failed to load bitmap indexes"));
2901
2902	/*
2903	* Since this function needs to know the position of each individual
2904	* bitmap, bypass the commit lookup table (if one exists) by forcing
2905	* the bitmap to eagerly load its entries.
2906	*/
2907	if (bitmap_git->table_lookup) {
2908	if (load_bitmap_entries_v1(bitmap_git) < 0)
2909	die(_("failed to load bitmap indexes"));
2910	}
2911
2912	kh_foreach (bitmap_git->bitmaps, oid, stored, {
2913	commit_idx_pos_map_pos = stored->map_pos;
2914	xor_offset_map_pos = stored->map_pos + sizeof(uint32_t);
2915	flag_map_pos = xor_offset_map_pos + sizeof(uint8_t);
2916	ewah_bitmap_map_pos = flag_map_pos + sizeof(uint8_t);
2917
2918	printf_ln("%s %"PRIuMAX" %"PRIuMAX" %"PRIuMAX" %"PRIuMAX,
2919	oid_to_hex(&oid),
2920	(uintmax_t)commit_idx_pos_map_pos,
2921	(uintmax_t)xor_offset_map_pos,
2922	(uintmax_t)flag_map_pos,
2923	(uintmax_t)ewah_bitmap_map_pos);
2924	})
2925	;
2926
2927	free_bitmap_index(bitmap_git);
2928
2929	return 0;
2930	}
2931
2932	int test_bitmap_hashes(struct repository *r)
2933	{
2934	struct bitmap_index *bitmap_git = prepare_bitmap_git(r);
2935	struct object_id oid;
2936	uint32_t i, index_pos;
2937
2938	if (!bitmap_git \|\| !bitmap_git->hashes)
2939	goto cleanup;
2940
2941	for (i = 0; i < bitmap_num_objects(bitmap_git); i++) {
2942	if (bitmap_is_midx(bitmap_git))
2943	index_pos = pack_pos_to_midx(bitmap_git->midx, i);
2944	else
2945	index_pos = pack_pos_to_index(bitmap_git->pack, i);
2946
2947	nth_bitmap_object_oid(bitmap_git, &oid, index_pos);
2948
2949	printf_ln("%s %"PRIu32"",
2950	oid_to_hex(&oid), get_be32(bitmap_git->hashes + index_pos));
2951	}
2952
2953	cleanup:
2954	free_bitmap_index(bitmap_git);
2955
2956	return 0;
2957	}
2958
2959	static void bit_pos_to_object_id(struct bitmap_index *bitmap_git,
2960	uint32_t bit_pos,
2961	struct object_id *oid)
2962	{
2963	uint32_t index_pos;
2964
2965	if (bitmap_is_midx(bitmap_git))
2966	index_pos = pack_pos_to_midx(bitmap_git->midx, bit_pos);
2967	else
2968	index_pos = pack_pos_to_index(bitmap_git->pack, bit_pos);
2969
2970	nth_bitmap_object_oid(bitmap_git, oid, index_pos);
2971	}
2972
2973	int test_bitmap_pseudo_merges(struct repository *r)
2974	{
2975	struct bitmap_index *bitmap_git;
2976	uint32_t i;
2977
2978	bitmap_git = prepare_bitmap_git(r);
2979	if (!bitmap_git \|\| !bitmap_git->pseudo_merges.nr)
2980	goto cleanup;
2981
2982	for (i = 0; i < bitmap_git->pseudo_merges.nr; i++) {
2983	struct pseudo_merge *merge;
2984	struct ewah_bitmap commits_bitmap, merge_bitmap;
2985
2986	merge = use_pseudo_merge(&bitmap_git->pseudo_merges,
2987	&bitmap_git->pseudo_merges.v[i]);
2988	commits_bitmap = merge->commits;
2989	merge_bitmap = pseudo_merge_bitmap(&bitmap_git->pseudo_merges,
2990	merge);
2991
2992	printf("at=%"PRIuMAX", commits=%"PRIuMAX", objects=%"PRIuMAX"\n",
2993	(uintmax_t)merge->at,
2994	(uintmax_t)ewah_bitmap_popcount(commits_bitmap),
2995	(uintmax_t)ewah_bitmap_popcount(merge_bitmap));
2996	}
2997
2998	cleanup:
2999	free_bitmap_index(bitmap_git);
3000	return 0;
3001	}
3002
3003	static void dump_ewah_object_ids(struct bitmap_index *bitmap_git,
3004	struct ewah_bitmap *bitmap)
3005
3006	{
3007	struct ewah_iterator it;
3008	eword_t word;
3009	uint32_t pos = 0;
3010
3011	ewah_iterator_init(&it, bitmap);
3012
3013	while (ewah_iterator_next(&word, &it)) {
3014	struct object_id oid;
3015	uint32_t offset;
3016
3017	for (offset = 0; offset < BITS_IN_EWORD; offset++) {
3018	if (!(word >> offset))
3019	break;
3020
3021	offset += ewah_bit_ctz64(word >> offset);
3022
3023	bit_pos_to_object_id(bitmap_git, pos + offset, &oid);
3024	printf("%s\n", oid_to_hex(&oid));
3025	}
3026	pos += BITS_IN_EWORD;
3027	}
3028	}
3029
3030	int test_bitmap_pseudo_merge_commits(struct repository *r, uint32_t n)
3031	{
3032	struct bitmap_index *bitmap_git;
3033	struct pseudo_merge *merge;
3034	int ret = 0;
3035
3036	bitmap_git = prepare_bitmap_git(r);
3037	if (!bitmap_git \|\| !bitmap_git->pseudo_merges.nr)
3038	goto cleanup;
3039
3040	if (n >= bitmap_git->pseudo_merges.nr) {
3041	ret = error(_("pseudo-merge index out of range "
3042	"(%"PRIu32" >= %"PRIuMAX")"),
3043	n, (uintmax_t)bitmap_git->pseudo_merges.nr);
3044	goto cleanup;
3045	}
3046
3047	merge = use_pseudo_merge(&bitmap_git->pseudo_merges,
3048	&bitmap_git->pseudo_merges.v[n]);
3049	dump_ewah_object_ids(bitmap_git, merge->commits);
3050
3051	cleanup:
3052	free_bitmap_index(bitmap_git);
3053	return ret;
3054	}
3055
3056	int test_bitmap_pseudo_merge_objects(struct repository *r, uint32_t n)
3057	{
3058	struct bitmap_index *bitmap_git;
3059	struct pseudo_merge *merge;
3060	int ret = 0;
3061
3062	bitmap_git = prepare_bitmap_git(r);
3063	if (!bitmap_git \|\| !bitmap_git->pseudo_merges.nr)
3064	goto cleanup;
3065
3066	if (n >= bitmap_git->pseudo_merges.nr) {
3067	ret = error(_("pseudo-merge index out of range "
3068	"(%"PRIu32" >= %"PRIuMAX")"),
3069	n, (uintmax_t)bitmap_git->pseudo_merges.nr);
3070	goto cleanup;
3071	}
3072
3073	merge = use_pseudo_merge(&bitmap_git->pseudo_merges,
3074	&bitmap_git->pseudo_merges.v[n]);
3075
3076	dump_ewah_object_ids(bitmap_git,
3077	pseudo_merge_bitmap(&bitmap_git->pseudo_merges,
3078	merge));
3079
3080	cleanup:
3081	free_bitmap_index(bitmap_git);
3082	return ret;
3083	}
3084
3085	int rebuild_bitmap(const uint32_t *reposition,
3086	struct ewah_bitmap *source,
3087	struct bitmap *dest)
3088	{
3089	uint32_t pos = 0;
3090	struct ewah_iterator it;
3091	eword_t word;
3092
3093	ewah_iterator_init(&it, source);
3094
3095	while (ewah_iterator_next(&word, &it)) {
3096	uint32_t offset, bit_pos;
3097
3098	for (offset = 0; offset < BITS_IN_EWORD; ++offset) {
3099	if ((word >> offset) == 0)
3100	break;
3101
3102	offset += ewah_bit_ctz64(word >> offset);
3103
3104	bit_pos = reposition[pos + offset];
3105	if (bit_pos > 0)
3106	bitmap_set(dest, bit_pos - 1);
3107	else /* can't reuse, we don't have the object */
3108	return -1;
3109	}
3110
3111	pos += BITS_IN_EWORD;
3112	}
3113	return 0;
3114	}
3115
3116	uint32_t create_bitmap_mapping(struct bitmap_index bitmap_git,
3117	struct packing_data *mapping)
3118	{
3119	struct repository *r = bitmap_repo(bitmap_git);
3120	uint32_t i, num_objects;
3121	uint32_t *reposition;
3122
3123	if (!bitmap_is_midx(bitmap_git))
3124	load_reverse_index(r, bitmap_git);
3125	else if (load_midx_revindex(bitmap_git->midx))
3126	BUG("rebuild_existing_bitmaps: missing required rev-cache "
3127	"extension");
3128
3129	num_objects = bitmap_num_objects_total(bitmap_git);
3130	CALLOC_ARRAY(reposition, num_objects);
3131
3132	for (i = 0; i < num_objects; ++i) {
3133	struct object_id oid;
3134	struct object_entry *oe;
3135	uint32_t index_pos;
3136
3137	if (bitmap_is_midx(bitmap_git))
3138	index_pos = pack_pos_to_midx(bitmap_git->midx, i);
3139	else
3140	index_pos = pack_pos_to_index(bitmap_git->pack, i);
3141	nth_bitmap_object_oid(bitmap_git, &oid, index_pos);
3142	oe = packlist_find(mapping, &oid);
3143
3144	if (oe) {
3145	reposition[i] = oe_in_pack_pos(mapping, oe) + 1;
3146	if (!oe->hash)
3147	oe->hash = bitmap_name_hash(bitmap_git, index_pos);
3148	}
3149	}
3150
3151	return reposition;
3152	}
3153
3154	void free_bitmap_index(struct bitmap_index *b)
3155	{
3156	if (!b)
3157	return;
3158
3159	if (b->map)
3160	munmap(b->map, b->map_size);
3161	ewah_pool_free(b->commits);
3162	ewah_pool_free(b->trees);
3163	ewah_pool_free(b->blobs);
3164	ewah_pool_free(b->tags);
3165	free(b->commits_all);
3166	free(b->trees_all);
3167	free(b->blobs_all);
3168	free(b->tags_all);
3169	if (b->bitmaps) {
3170	struct stored_bitmap *sb;
3171	kh_foreach_value(b->bitmaps, sb, {
3172	ewah_pool_free(sb->root);
3173	free(sb);
3174	});
3175	}
3176	kh_destroy_oid_map(b->bitmaps);
3177	free(b->ext_index.objects);
3178	free(b->ext_index.hashes);
3179	kh_destroy_oid_pos(b->ext_index.positions);
3180	bitmap_free(b->result);
3181	bitmap_free(b->haves);
3182	if (bitmap_is_midx(b)) {
3183	/*
3184	* Multi-pack bitmaps need to have resources associated with
3185	* their on-disk reverse indexes unmapped so that stale .rev and
3186	* .bitmap files can be removed.
3187	*
3188	* Unlike pack-based bitmaps, multi-pack bitmaps can be read and
3189	* written in the same 'git multi-pack-index write --bitmap'
3190	* process. Close resources so they can be removed safely on
3191	* platforms like Windows.
3192	*/
3193	close_midx_revindex(b->midx);
3194	}
3195	free_pseudo_merge_map(&b->pseudo_merges);
3196	free_bitmap_index(b->base);
3197	free(b);
3198	}
3199
3200	int bitmap_has_oid_in_uninteresting(struct bitmap_index *bitmap_git,
3201	const struct object_id *oid)
3202	{
3203	return bitmap_git &&
3204	bitmap_walk_contains(bitmap_git, bitmap_git->haves, oid);
3205	}
3206
3207	static off_t get_disk_usage_for_type(struct bitmap_index *bitmap_git,
3208	enum object_type object_type)
3209	{
3210	struct bitmap *result = bitmap_git->result;
3211	off_t total = 0;
3212	struct ewah_or_iterator it;
3213	eword_t filter;
3214	size_t i;
3215
3216	init_type_iterator(&it, bitmap_git, object_type);
3217	for (i = 0; i < result->word_alloc &&
3218	ewah_or_iterator_next(&filter, &it); i++) {
3219	eword_t word = result->words[i] & filter;
3220	size_t base = (i * BITS_IN_EWORD);
3221	unsigned offset;
3222
3223	if (!word)
3224	continue;
3225
3226	for (offset = 0; offset < BITS_IN_EWORD; offset++) {
3227	if ((word >> offset) == 0)
3228	break;
3229
3230	offset += ewah_bit_ctz64(word >> offset);
3231
3232	if (bitmap_is_midx(bitmap_git)) {
3233	uint32_t pack_pos;
3234	uint32_t midx_pos = pack_pos_to_midx(bitmap_git->midx, base + offset);
3235	off_t offset = nth_midxed_offset(bitmap_git->midx, midx_pos);
3236
3237	uint32_t pack_id = nth_midxed_pack_int_id(bitmap_git->midx, midx_pos);
3238	struct packed_git *pack = nth_midxed_pack(bitmap_git->midx, pack_id);
3239
3240	if (offset_to_pack_pos(pack, offset, &pack_pos) < 0) {
3241	struct object_id oid;
3242	nth_midxed_object_oid(&oid, bitmap_git->midx, midx_pos);
3243
3244	die(_("could not find '%s' in pack '%s' at offset %"PRIuMAX),
3245	oid_to_hex(&oid),
3246	pack->pack_name,
3247	(uintmax_t)offset);
3248	}
3249
3250	total += pack_pos_to_offset(pack, pack_pos + 1) - offset;
3251	} else {
3252	size_t pos = base + offset;
3253	total += pack_pos_to_offset(bitmap_git->pack, pos + 1) -
3254	pack_pos_to_offset(bitmap_git->pack, pos);
3255	}
3256	}
3257	}
3258
3259	ewah_or_iterator_release(&it);
3260
3261	return total;
3262	}
3263
3264	static off_t get_disk_usage_for_extended(struct bitmap_index *bitmap_git)
3265	{
3266	struct bitmap *result = bitmap_git->result;
3267	struct eindex *eindex = &bitmap_git->ext_index;
3268	off_t total = 0;
3269	struct object_info oi = OBJECT_INFO_INIT;
3270	off_t object_size;
3271	size_t i;
3272
3273	oi.disk_sizep = &object_size;
3274
3275	for (i = 0; i < eindex->count; i++) {
3276	struct object *obj = eindex->objects[i];
3277
3278	if (!bitmap_get(result,
3279	st_add(bitmap_num_objects_total(bitmap_git),
3280	i)))
3281	continue;
3282
3283	if (odb_read_object_info_extended(bitmap_repo(bitmap_git)->objects,
3284	&obj->oid, &oi, 0) < 0)
3285	die(_("unable to get disk usage of '%s'"),
3286	oid_to_hex(&obj->oid));
3287
3288	total += object_size;
3289	}
3290	return total;
3291	}
3292
3293	off_t get_disk_usage_from_bitmap(struct bitmap_index *bitmap_git,
3294	struct rev_info *revs)
3295	{
3296	off_t total = 0;
3297
3298	total += get_disk_usage_for_type(bitmap_git, OBJ_COMMIT);
3299	if (revs->tree_objects)
3300	total += get_disk_usage_for_type(bitmap_git, OBJ_TREE);
3301	if (revs->blob_objects)
3302	total += get_disk_usage_for_type(bitmap_git, OBJ_BLOB);
3303	if (revs->tag_objects)
3304	total += get_disk_usage_for_type(bitmap_git, OBJ_TAG);
3305
3306	total += get_disk_usage_for_extended(bitmap_git);
3307
3308	return total;
3309	}
3310
3311	int bitmap_is_midx(struct bitmap_index *bitmap_git)
3312	{
3313	return !!bitmap_git->midx;
3314	}
3315
3316	static const struct string_list bitmap_preferred_tips(struct repository r)
3317	{
3318	const struct string_list *dest;
3319
3320	if (!repo_config_get_string_multi(r, "pack.preferbitmaptips", &dest))
3321	return dest;
3322	return NULL;
3323	}
3324
3325	void for_each_preferred_bitmap_tip(struct repository *repo,
3326	refs_for_each_cb cb, void *cb_data)
3327	{
3328	struct refs_for_each_ref_options opts = { 0 };
3329	struct string_list_item *item;
3330	const struct string_list *preferred_tips;
3331	struct strbuf buf = STRBUF_INIT;
3332
3333	preferred_tips = bitmap_preferred_tips(repo);
3334	if (!preferred_tips)
3335	return;
3336
3337	for_each_string_list_item(item, preferred_tips) {
3338	opts.prefix = item->string;
3339
3340	if (!ends_with(opts.prefix, "/")) {
3341	strbuf_reset(&buf);
3342	strbuf_addf(&buf, "%s/", opts.prefix);
3343	opts.prefix = buf.buf;
3344	}
3345
3346	refs_for_each_ref_ext(get_main_ref_store(repo),
3347	cb, cb_data, &opts);
3348	}
3349
3350	strbuf_release(&buf);
3351	}
3352
3353	int bitmap_is_preferred_refname(struct repository r, const char refname)
3354	{
3355	const struct string_list *preferred_tips = bitmap_preferred_tips(r);
3356	struct string_list_item *item;
3357
3358	if (!preferred_tips)
3359	return 0;
3360
3361	for_each_string_list_item(item, preferred_tips) {
3362	if (starts_with(refname, item->string))
3363	return 1;
3364	}
3365
3366	return 0;
3367	}
3368
3369	static int verify_bitmap_file(const struct git_hash_algo *algop,
3370	const char *name)
3371	{
3372	struct stat st;
3373	unsigned char *data;
3374	int fd = git_open(name);
3375	int res = 0;
3376
3377	/* It is OK to not have the file. */
3378	if (fd < 0 \|\| fstat(fd, &st)) {
3379	if (fd >= 0)
3380	close(fd);
3381	return 0;
3382	}
3383
3384	data = xmmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
3385	close(fd);
3386	if (!hashfile_checksum_valid(algop, data, st.st_size))
3387	res = error(_("bitmap file '%s' has invalid checksum"),
3388	name);
3389
3390	munmap(data, st.st_size);
3391	return res;
3392	}
3393
3394	int verify_bitmap_files(struct repository *r)
3395	{
3396	struct odb_source *source;
3397	struct packed_git *p;
3398	int res = 0;
3399
3400	odb_prepare_alternates(r->objects);
3401	for (source = r->objects->sources; source; source = source->next) {
3402	struct multi_pack_index *m = get_multi_pack_index(source);
3403	char *midx_bitmap_name;
3404
3405	if (!m)
3406	continue;
3407
3408	midx_bitmap_name = midx_bitmap_filename(m);
3409	res \|= verify_bitmap_file(r->hash_algo, midx_bitmap_name);
3410	free(midx_bitmap_name);
3411	}
3412
3413	repo_for_each_pack(r, p) {
3414	char *pack_bitmap_name = pack_bitmap_filename(p);
3415	res \|= verify_bitmap_file(r->hash_algo, pack_bitmap_name);
3416	free(pack_bitmap_name);
3417	}
3418
3419	return res;
3420	}