7 #include "list-objects.h"
9 #include "pack-bitmap.h"
10 #include "pack-revindex.h"
11 #include "pack-objects.h"
13 #include "repository.h"
14 #include "object-store.h"
17 * An entry on the bitmap index, representing the bitmap for a given
20 struct stored_bitmap
{
21 unsigned char sha1
[20];
22 struct ewah_bitmap
*root
;
23 struct stored_bitmap
*xor;
28 * The active bitmap index for a repository. By design, repositories only have
29 * a single bitmap index available (the index for the biggest packfile in
30 * the repository), since bitmap indexes need full closure.
32 * If there is more than one bitmap index available (e.g. because of alternates),
33 * the active bitmap index is the largest one.
36 /* Packfile to which this bitmap index belongs to */
37 struct packed_git
*pack
;
40 * Mark the first `reuse_objects` in the packfile as reused:
41 * they will be sent as-is without using them for repacking
44 uint32_t reuse_objects
;
46 /* mmapped buffer of the whole bitmap index */
48 size_t map_size
; /* size of the mmaped buffer */
49 size_t map_pos
; /* current position when loading the index */
54 * Each bitmap marks which objects in the packfile are of the given
55 * type. This provides type information when yielding the objects from
56 * the packfile during a walk, which allows for better delta bases.
58 struct ewah_bitmap
*commits
;
59 struct ewah_bitmap
*trees
;
60 struct ewah_bitmap
*blobs
;
61 struct ewah_bitmap
*tags
;
63 /* Map from SHA1 -> `stored_bitmap` for all the bitmapped commits */
66 /* Number of bitmapped commits */
69 /* Name-hash cache (or NULL if not present). */
75 * When trying to perform bitmap operations with objects that are not
76 * packed in `pack`, these objects are added to this "fake index" and
77 * are assumed to appear at the end of the packfile for all operations
80 struct object
**objects
;
82 uint32_t count
, alloc
;
83 khash_sha1_pos
*positions
;
86 /* Bitmap result of the last performed walk */
87 struct bitmap
*result
;
89 /* Version of the bitmap index */
95 static struct ewah_bitmap
*lookup_stored_bitmap(struct stored_bitmap
*st
)
97 struct ewah_bitmap
*parent
;
98 struct ewah_bitmap
*composed
;
103 composed
= ewah_pool_new();
104 parent
= lookup_stored_bitmap(st
->xor);
105 ewah_xor(st
->root
, parent
, composed
);
107 ewah_pool_free(st
->root
);
115 * Read a bitmap from the current read position on the mmaped
116 * index, and increase the read position accordingly
118 static struct ewah_bitmap
*read_bitmap_1(struct bitmap_index
*index
)
120 struct ewah_bitmap
*b
= ewah_pool_new();
122 ssize_t bitmap_size
= ewah_read_mmap(b
,
123 index
->map
+ index
->map_pos
,
124 index
->map_size
- index
->map_pos
);
126 if (bitmap_size
< 0) {
127 error("Failed to load bitmap index (corrupted?)");
132 index
->map_pos
+= bitmap_size
;
136 static int load_bitmap_header(struct bitmap_index
*index
)
138 struct bitmap_disk_header
*header
= (void *)index
->map
;
140 if (index
->map_size
< sizeof(*header
) + 20)
141 return error("Corrupted bitmap index (missing header data)");
143 if (memcmp(header
->magic
, BITMAP_IDX_SIGNATURE
, sizeof(BITMAP_IDX_SIGNATURE
)) != 0)
144 return error("Corrupted bitmap index file (wrong header)");
146 index
->version
= ntohs(header
->version
);
147 if (index
->version
!= 1)
148 return error("Unsupported version for bitmap index file (%d)", index
->version
);
150 /* Parse known bitmap format options */
152 uint32_t flags
= ntohs(header
->options
);
154 if ((flags
& BITMAP_OPT_FULL_DAG
) == 0)
155 return error("Unsupported options for bitmap index file "
156 "(Git requires BITMAP_OPT_FULL_DAG)");
158 if (flags
& BITMAP_OPT_HASH_CACHE
) {
159 unsigned char *end
= index
->map
+ index
->map_size
- 20;
160 index
->hashes
= ((uint32_t *)end
) - index
->pack
->num_objects
;
164 index
->entry_count
= ntohl(header
->entry_count
);
165 index
->map_pos
+= sizeof(*header
);
169 static struct stored_bitmap
*store_bitmap(struct bitmap_index
*index
,
170 struct ewah_bitmap
*root
,
171 const unsigned char *sha1
,
172 struct stored_bitmap
*xor_with
,
175 struct stored_bitmap
*stored
;
179 stored
= xmalloc(sizeof(struct stored_bitmap
));
181 stored
->xor = xor_with
;
182 stored
->flags
= flags
;
183 hashcpy(stored
->sha1
, sha1
);
185 hash_pos
= kh_put_sha1(index
->bitmaps
, stored
->sha1
, &ret
);
187 /* a 0 return code means the insertion succeeded with no changes,
188 * because the SHA1 already existed on the map. this is bad, there
189 * shouldn't be duplicated commits in the index */
191 error("Duplicate entry in bitmap index: %s", sha1_to_hex(sha1
));
195 kh_value(index
->bitmaps
, hash_pos
) = stored
;
199 static inline uint32_t read_be32(const unsigned char *buffer
, size_t *pos
)
201 uint32_t result
= get_be32(buffer
+ *pos
);
202 (*pos
) += sizeof(result
);
206 static inline uint8_t read_u8(const unsigned char *buffer
, size_t *pos
)
208 return buffer
[(*pos
)++];
211 #define MAX_XOR_OFFSET 160
213 static int load_bitmap_entries_v1(struct bitmap_index
*index
)
216 struct stored_bitmap
*recent_bitmaps
[MAX_XOR_OFFSET
] = { NULL
};
218 for (i
= 0; i
< index
->entry_count
; ++i
) {
219 int xor_offset
, flags
;
220 struct ewah_bitmap
*bitmap
= NULL
;
221 struct stored_bitmap
*xor_bitmap
= NULL
;
222 uint32_t commit_idx_pos
;
223 const unsigned char *sha1
;
225 commit_idx_pos
= read_be32(index
->map
, &index
->map_pos
);
226 xor_offset
= read_u8(index
->map
, &index
->map_pos
);
227 flags
= read_u8(index
->map
, &index
->map_pos
);
229 sha1
= nth_packed_object_sha1(index
->pack
, commit_idx_pos
);
231 bitmap
= read_bitmap_1(index
);
235 if (xor_offset
> MAX_XOR_OFFSET
|| xor_offset
> i
)
236 return error("Corrupted bitmap pack index");
238 if (xor_offset
> 0) {
239 xor_bitmap
= recent_bitmaps
[(i
- xor_offset
) % MAX_XOR_OFFSET
];
241 if (xor_bitmap
== NULL
)
242 return error("Invalid XOR offset in bitmap pack index");
245 recent_bitmaps
[i
% MAX_XOR_OFFSET
] = store_bitmap(
246 index
, bitmap
, sha1
, xor_bitmap
, flags
);
252 static char *pack_bitmap_filename(struct packed_git
*p
)
256 if (!strip_suffix(p
->pack_name
, ".pack", &len
))
257 BUG("pack_name does not end in .pack");
258 return xstrfmt("%.*s.bitmap", (int)len
, p
->pack_name
);
261 static int open_pack_bitmap_1(struct bitmap_index
*bitmap_git
, struct packed_git
*packfile
)
267 if (open_pack_index(packfile
))
270 idx_name
= pack_bitmap_filename(packfile
);
271 fd
= git_open(idx_name
);
277 if (fstat(fd
, &st
)) {
282 if (bitmap_git
->pack
) {
283 warning("ignoring extra bitmap file: %s", packfile
->pack_name
);
288 bitmap_git
->pack
= packfile
;
289 bitmap_git
->map_size
= xsize_t(st
.st_size
);
290 bitmap_git
->map
= xmmap(NULL
, bitmap_git
->map_size
, PROT_READ
, MAP_PRIVATE
, fd
, 0);
291 bitmap_git
->map_pos
= 0;
294 if (load_bitmap_header(bitmap_git
) < 0) {
295 munmap(bitmap_git
->map
, bitmap_git
->map_size
);
296 bitmap_git
->map
= NULL
;
297 bitmap_git
->map_size
= 0;
304 static int load_pack_bitmap(struct bitmap_index
*bitmap_git
)
306 assert(bitmap_git
->map
&& !bitmap_git
->loaded
);
308 bitmap_git
->bitmaps
= kh_init_sha1();
309 bitmap_git
->ext_index
.positions
= kh_init_sha1_pos();
310 load_pack_revindex(bitmap_git
->pack
);
312 if (!(bitmap_git
->commits
= read_bitmap_1(bitmap_git
)) ||
313 !(bitmap_git
->trees
= read_bitmap_1(bitmap_git
)) ||
314 !(bitmap_git
->blobs
= read_bitmap_1(bitmap_git
)) ||
315 !(bitmap_git
->tags
= read_bitmap_1(bitmap_git
)))
318 if (load_bitmap_entries_v1(bitmap_git
) < 0)
321 bitmap_git
->loaded
= 1;
325 munmap(bitmap_git
->map
, bitmap_git
->map_size
);
326 bitmap_git
->map
= NULL
;
327 bitmap_git
->map_size
= 0;
331 static int open_pack_bitmap(struct bitmap_index
*bitmap_git
)
333 struct packed_git
*p
;
336 assert(!bitmap_git
->map
&& !bitmap_git
->loaded
);
338 for (p
= get_packed_git(the_repository
); p
; p
= p
->next
) {
339 if (open_pack_bitmap_1(bitmap_git
, p
) == 0)
346 struct bitmap_index
*prepare_bitmap_git(void)
348 struct bitmap_index
*bitmap_git
= xcalloc(1, sizeof(*bitmap_git
));
350 if (!open_pack_bitmap(bitmap_git
) && !load_pack_bitmap(bitmap_git
))
356 struct include_data
{
357 struct bitmap_index
*bitmap_git
;
362 static inline int bitmap_position_extended(struct bitmap_index
*bitmap_git
,
363 const unsigned char *sha1
)
365 khash_sha1_pos
*positions
= bitmap_git
->ext_index
.positions
;
366 khiter_t pos
= kh_get_sha1_pos(positions
, sha1
);
368 if (pos
< kh_end(positions
)) {
369 int bitmap_pos
= kh_value(positions
, pos
);
370 return bitmap_pos
+ bitmap_git
->pack
->num_objects
;
376 static inline int bitmap_position_packfile(struct bitmap_index
*bitmap_git
,
377 const unsigned char *sha1
)
379 off_t offset
= find_pack_entry_one(sha1
, bitmap_git
->pack
);
383 return find_revindex_position(bitmap_git
->pack
, offset
);
386 static int bitmap_position(struct bitmap_index
*bitmap_git
,
387 const unsigned char *sha1
)
389 int pos
= bitmap_position_packfile(bitmap_git
, sha1
);
390 return (pos
>= 0) ? pos
: bitmap_position_extended(bitmap_git
, sha1
);
393 static int ext_index_add_object(struct bitmap_index
*bitmap_git
,
394 struct object
*object
, const char *name
)
396 struct eindex
*eindex
= &bitmap_git
->ext_index
;
402 hash_pos
= kh_put_sha1_pos(eindex
->positions
, object
->oid
.hash
, &hash_ret
);
404 if (eindex
->count
>= eindex
->alloc
) {
405 eindex
->alloc
= (eindex
->alloc
+ 16) * 3 / 2;
406 REALLOC_ARRAY(eindex
->objects
, eindex
->alloc
);
407 REALLOC_ARRAY(eindex
->hashes
, eindex
->alloc
);
410 bitmap_pos
= eindex
->count
;
411 eindex
->objects
[eindex
->count
] = object
;
412 eindex
->hashes
[eindex
->count
] = pack_name_hash(name
);
413 kh_value(eindex
->positions
, hash_pos
) = bitmap_pos
;
416 bitmap_pos
= kh_value(eindex
->positions
, hash_pos
);
419 return bitmap_pos
+ bitmap_git
->pack
->num_objects
;
422 struct bitmap_show_data
{
423 struct bitmap_index
*bitmap_git
;
427 static void show_object(struct object
*object
, const char *name
, void *data_
)
429 struct bitmap_show_data
*data
= data_
;
432 bitmap_pos
= bitmap_position(data
->bitmap_git
, object
->oid
.hash
);
435 bitmap_pos
= ext_index_add_object(data
->bitmap_git
, object
,
438 bitmap_set(data
->base
, bitmap_pos
);
441 static void show_commit(struct commit
*commit
, void *data
)
445 static int add_to_include_set(struct bitmap_index
*bitmap_git
,
446 struct include_data
*data
,
447 const unsigned char *sha1
,
452 if (data
->seen
&& bitmap_get(data
->seen
, bitmap_pos
))
455 if (bitmap_get(data
->base
, bitmap_pos
))
458 hash_pos
= kh_get_sha1(bitmap_git
->bitmaps
, sha1
);
459 if (hash_pos
< kh_end(bitmap_git
->bitmaps
)) {
460 struct stored_bitmap
*st
= kh_value(bitmap_git
->bitmaps
, hash_pos
);
461 bitmap_or_ewah(data
->base
, lookup_stored_bitmap(st
));
465 bitmap_set(data
->base
, bitmap_pos
);
469 static int should_include(struct commit
*commit
, void *_data
)
471 struct include_data
*data
= _data
;
474 bitmap_pos
= bitmap_position(data
->bitmap_git
, commit
->object
.oid
.hash
);
476 bitmap_pos
= ext_index_add_object(data
->bitmap_git
,
477 (struct object
*)commit
,
480 if (!add_to_include_set(data
->bitmap_git
, data
, commit
->object
.oid
.hash
,
482 struct commit_list
*parent
= commit
->parents
;
485 parent
->item
->object
.flags
|= SEEN
;
486 parent
= parent
->next
;
495 static struct bitmap
*find_objects(struct bitmap_index
*bitmap_git
,
496 struct rev_info
*revs
,
497 struct object_list
*roots
,
500 struct bitmap
*base
= NULL
;
503 struct object_list
*not_mapped
= NULL
;
506 * Go through all the roots for the walk. The ones that have bitmaps
507 * on the bitmap index will be `or`ed together to form an initial
508 * global reachability analysis.
510 * The ones without bitmaps in the index will be stored in the
511 * `not_mapped_list` for further processing.
514 struct object
*object
= roots
->item
;
517 if (object
->type
== OBJ_COMMIT
) {
518 khiter_t pos
= kh_get_sha1(bitmap_git
->bitmaps
, object
->oid
.hash
);
520 if (pos
< kh_end(bitmap_git
->bitmaps
)) {
521 struct stored_bitmap
*st
= kh_value(bitmap_git
->bitmaps
, pos
);
522 struct ewah_bitmap
*or_with
= lookup_stored_bitmap(st
);
525 base
= ewah_to_bitmap(or_with
);
527 bitmap_or_ewah(base
, or_with
);
529 object
->flags
|= SEEN
;
534 object_list_insert(object
, ¬_mapped
);
538 * Best case scenario: We found bitmaps for all the roots,
539 * so the resulting `or` bitmap has the full reachability analysis
541 if (not_mapped
== NULL
)
547 * Let's iterate through all the roots that don't have bitmaps to
548 * check if we can determine them to be reachable from the existing
551 * If we cannot find them in the existing global bitmap, we'll need
552 * to push them to an actual walk and run it until we can confirm
556 struct object
*object
= roots
->item
;
560 pos
= bitmap_position(bitmap_git
, object
->oid
.hash
);
562 if (pos
< 0 || base
== NULL
|| !bitmap_get(base
, pos
)) {
563 object
->flags
&= ~UNINTERESTING
;
564 add_pending_object(revs
, object
, "");
567 object
->flags
|= SEEN
;
572 struct include_data incdata
;
573 struct bitmap_show_data show_data
;
578 incdata
.bitmap_git
= bitmap_git
;
582 revs
->include_check
= should_include
;
583 revs
->include_check_data
= &incdata
;
585 if (prepare_revision_walk(revs
))
586 die("revision walk setup failed");
588 show_data
.bitmap_git
= bitmap_git
;
589 show_data
.base
= base
;
591 traverse_commit_list(revs
, show_commit
, show_object
,
598 static void show_extended_objects(struct bitmap_index
*bitmap_git
,
599 show_reachable_fn show_reach
)
601 struct bitmap
*objects
= bitmap_git
->result
;
602 struct eindex
*eindex
= &bitmap_git
->ext_index
;
605 for (i
= 0; i
< eindex
->count
; ++i
) {
608 if (!bitmap_get(objects
, bitmap_git
->pack
->num_objects
+ i
))
611 obj
= eindex
->objects
[i
];
612 show_reach(&obj
->oid
, obj
->type
, 0, eindex
->hashes
[i
], NULL
, 0);
616 static void show_objects_for_type(
617 struct bitmap_index
*bitmap_git
,
618 struct ewah_bitmap
*type_filter
,
619 enum object_type object_type
,
620 show_reachable_fn show_reach
)
622 size_t pos
= 0, i
= 0;
625 struct ewah_iterator it
;
628 struct bitmap
*objects
= bitmap_git
->result
;
630 if (bitmap_git
->reuse_objects
== bitmap_git
->pack
->num_objects
)
633 ewah_iterator_init(&it
, type_filter
);
635 while (i
< objects
->word_alloc
&& ewah_iterator_next(&filter
, &it
)) {
636 eword_t word
= objects
->words
[i
] & filter
;
638 for (offset
= 0; offset
< BITS_IN_EWORD
; ++offset
) {
639 struct object_id oid
;
640 struct revindex_entry
*entry
;
643 if ((word
>> offset
) == 0)
646 offset
+= ewah_bit_ctz64(word
>> offset
);
648 if (pos
+ offset
< bitmap_git
->reuse_objects
)
651 entry
= &bitmap_git
->pack
->revindex
[pos
+ offset
];
652 nth_packed_object_oid(&oid
, bitmap_git
->pack
, entry
->nr
);
654 if (bitmap_git
->hashes
)
655 hash
= get_be32(bitmap_git
->hashes
+ entry
->nr
);
657 show_reach(&oid
, object_type
, 0, hash
, bitmap_git
->pack
, entry
->offset
);
660 pos
+= BITS_IN_EWORD
;
665 static int in_bitmapped_pack(struct bitmap_index
*bitmap_git
,
666 struct object_list
*roots
)
669 struct object
*object
= roots
->item
;
672 if (find_pack_entry_one(object
->oid
.hash
, bitmap_git
->pack
) > 0)
679 struct bitmap_index
*prepare_bitmap_walk(struct rev_info
*revs
)
683 struct object_list
*wants
= NULL
;
684 struct object_list
*haves
= NULL
;
686 struct bitmap
*wants_bitmap
= NULL
;
687 struct bitmap
*haves_bitmap
= NULL
;
689 struct bitmap_index
*bitmap_git
= xcalloc(1, sizeof(*bitmap_git
));
690 /* try to open a bitmapped pack, but don't parse it yet
691 * because we may not need to use it */
692 if (open_pack_bitmap(bitmap_git
) < 0)
695 for (i
= 0; i
< revs
->pending
.nr
; ++i
) {
696 struct object
*object
= revs
->pending
.objects
[i
].item
;
698 if (object
->type
== OBJ_NONE
)
699 parse_object_or_die(&object
->oid
, NULL
);
701 while (object
->type
== OBJ_TAG
) {
702 struct tag
*tag
= (struct tag
*) object
;
704 if (object
->flags
& UNINTERESTING
)
705 object_list_insert(object
, &haves
);
707 object_list_insert(object
, &wants
);
711 object
= parse_object_or_die(&tag
->tagged
->oid
, NULL
);
714 if (object
->flags
& UNINTERESTING
)
715 object_list_insert(object
, &haves
);
717 object_list_insert(object
, &wants
);
721 * if we have a HAVES list, but none of those haves is contained
722 * in the packfile that has a bitmap, we don't have anything to
725 if (haves
&& !in_bitmapped_pack(bitmap_git
, haves
))
728 /* if we don't want anything, we're done here */
733 * now we're going to use bitmaps, so load the actual bitmap entries
734 * from disk. this is the point of no return; after this the rev_list
735 * becomes invalidated and we must perform the revwalk through bitmaps
737 if (!bitmap_git
->loaded
&& load_pack_bitmap(bitmap_git
) < 0)
740 object_array_clear(&revs
->pending
);
743 revs
->ignore_missing_links
= 1;
744 haves_bitmap
= find_objects(bitmap_git
, revs
, haves
, NULL
);
745 reset_revision_walk();
746 revs
->ignore_missing_links
= 0;
748 if (haves_bitmap
== NULL
)
749 BUG("failed to perform bitmap walk");
752 wants_bitmap
= find_objects(bitmap_git
, revs
, wants
, haves_bitmap
);
755 BUG("failed to perform bitmap walk");
758 bitmap_and_not(wants_bitmap
, haves_bitmap
);
760 bitmap_git
->result
= wants_bitmap
;
762 bitmap_free(haves_bitmap
);
766 int reuse_partial_packfile_from_bitmap(struct bitmap_index
*bitmap_git
,
767 struct packed_git
**packfile
,
772 * Reuse the packfile content if we need more than
775 static const double REUSE_PERCENT
= 0.9;
777 struct bitmap
*result
= bitmap_git
->result
;
778 uint32_t reuse_threshold
;
779 uint32_t i
, reuse_objects
= 0;
783 for (i
= 0; i
< result
->word_alloc
; ++i
) {
784 if (result
->words
[i
] != (eword_t
)~0) {
785 reuse_objects
+= ewah_bit_ctz64(~result
->words
[i
]);
789 reuse_objects
+= BITS_IN_EWORD
;
792 #ifdef GIT_BITMAP_DEBUG
794 const unsigned char *sha1
;
795 struct revindex_entry
*entry
;
797 entry
= &bitmap_git
->reverse_index
->revindex
[reuse_objects
];
798 sha1
= nth_packed_object_sha1(bitmap_git
->pack
, entry
->nr
);
800 fprintf(stderr
, "Failed to reuse at %d (%016llx)\n",
801 reuse_objects
, result
->words
[i
]);
802 fprintf(stderr
, " %s\n", sha1_to_hex(sha1
));
809 if (reuse_objects
>= bitmap_git
->pack
->num_objects
) {
810 bitmap_git
->reuse_objects
= *entries
= bitmap_git
->pack
->num_objects
;
811 *up_to
= -1; /* reuse the full pack */
812 *packfile
= bitmap_git
->pack
;
816 reuse_threshold
= bitmap_popcount(bitmap_git
->result
) * REUSE_PERCENT
;
818 if (reuse_objects
< reuse_threshold
)
821 bitmap_git
->reuse_objects
= *entries
= reuse_objects
;
822 *up_to
= bitmap_git
->pack
->revindex
[reuse_objects
].offset
;
823 *packfile
= bitmap_git
->pack
;
828 void traverse_bitmap_commit_list(struct bitmap_index
*bitmap_git
,
829 show_reachable_fn show_reachable
)
831 assert(bitmap_git
->result
);
833 show_objects_for_type(bitmap_git
, bitmap_git
->commits
,
834 OBJ_COMMIT
, show_reachable
);
835 show_objects_for_type(bitmap_git
, bitmap_git
->trees
,
836 OBJ_TREE
, show_reachable
);
837 show_objects_for_type(bitmap_git
, bitmap_git
->blobs
,
838 OBJ_BLOB
, show_reachable
);
839 show_objects_for_type(bitmap_git
, bitmap_git
->tags
,
840 OBJ_TAG
, show_reachable
);
842 show_extended_objects(bitmap_git
, show_reachable
);
844 bitmap_free(bitmap_git
->result
);
845 bitmap_git
->result
= NULL
;
848 static uint32_t count_object_type(struct bitmap_index
*bitmap_git
,
849 enum object_type type
)
851 struct bitmap
*objects
= bitmap_git
->result
;
852 struct eindex
*eindex
= &bitmap_git
->ext_index
;
854 uint32_t i
= 0, count
= 0;
855 struct ewah_iterator it
;
860 ewah_iterator_init(&it
, bitmap_git
->commits
);
864 ewah_iterator_init(&it
, bitmap_git
->trees
);
868 ewah_iterator_init(&it
, bitmap_git
->blobs
);
872 ewah_iterator_init(&it
, bitmap_git
->tags
);
879 while (i
< objects
->word_alloc
&& ewah_iterator_next(&filter
, &it
)) {
880 eword_t word
= objects
->words
[i
++] & filter
;
881 count
+= ewah_bit_popcount64(word
);
884 for (i
= 0; i
< eindex
->count
; ++i
) {
885 if (eindex
->objects
[i
]->type
== type
&&
886 bitmap_get(objects
, bitmap_git
->pack
->num_objects
+ i
))
893 void count_bitmap_commit_list(struct bitmap_index
*bitmap_git
,
894 uint32_t *commits
, uint32_t *trees
,
895 uint32_t *blobs
, uint32_t *tags
)
897 assert(bitmap_git
->result
);
900 *commits
= count_object_type(bitmap_git
, OBJ_COMMIT
);
903 *trees
= count_object_type(bitmap_git
, OBJ_TREE
);
906 *blobs
= count_object_type(bitmap_git
, OBJ_BLOB
);
909 *tags
= count_object_type(bitmap_git
, OBJ_TAG
);
912 struct bitmap_test_data
{
913 struct bitmap_index
*bitmap_git
;
915 struct progress
*prg
;
919 static void test_show_object(struct object
*object
, const char *name
,
922 struct bitmap_test_data
*tdata
= data
;
925 bitmap_pos
= bitmap_position(tdata
->bitmap_git
, object
->oid
.hash
);
927 die("Object not in bitmap: %s\n", oid_to_hex(&object
->oid
));
929 bitmap_set(tdata
->base
, bitmap_pos
);
930 display_progress(tdata
->prg
, ++tdata
->seen
);
933 static void test_show_commit(struct commit
*commit
, void *data
)
935 struct bitmap_test_data
*tdata
= data
;
938 bitmap_pos
= bitmap_position(tdata
->bitmap_git
,
939 commit
->object
.oid
.hash
);
941 die("Object not in bitmap: %s\n", oid_to_hex(&commit
->object
.oid
));
943 bitmap_set(tdata
->base
, bitmap_pos
);
944 display_progress(tdata
->prg
, ++tdata
->seen
);
947 void test_bitmap_walk(struct rev_info
*revs
)
950 struct bitmap
*result
= NULL
;
952 size_t result_popcnt
;
953 struct bitmap_test_data tdata
;
954 struct bitmap_index
*bitmap_git
;
956 if (!(bitmap_git
= prepare_bitmap_git()))
957 die("failed to load bitmap indexes");
959 if (revs
->pending
.nr
!= 1)
960 die("you must specify exactly one commit to test");
962 fprintf(stderr
, "Bitmap v%d test (%d entries loaded)\n",
963 bitmap_git
->version
, bitmap_git
->entry_count
);
965 root
= revs
->pending
.objects
[0].item
;
966 pos
= kh_get_sha1(bitmap_git
->bitmaps
, root
->oid
.hash
);
968 if (pos
< kh_end(bitmap_git
->bitmaps
)) {
969 struct stored_bitmap
*st
= kh_value(bitmap_git
->bitmaps
, pos
);
970 struct ewah_bitmap
*bm
= lookup_stored_bitmap(st
);
972 fprintf(stderr
, "Found bitmap for %s. %d bits / %08x checksum\n",
973 oid_to_hex(&root
->oid
), (int)bm
->bit_size
, ewah_checksum(bm
));
975 result
= ewah_to_bitmap(bm
);
979 die("Commit %s doesn't have an indexed bitmap", oid_to_hex(&root
->oid
));
981 revs
->tag_objects
= 1;
982 revs
->tree_objects
= 1;
983 revs
->blob_objects
= 1;
985 result_popcnt
= bitmap_popcount(result
);
987 if (prepare_revision_walk(revs
))
988 die("revision walk setup failed");
990 tdata
.bitmap_git
= bitmap_git
;
991 tdata
.base
= bitmap_new();
992 tdata
.prg
= start_progress("Verifying bitmap entries", result_popcnt
);
995 traverse_commit_list(revs
, &test_show_commit
, &test_show_object
, &tdata
);
997 stop_progress(&tdata
.prg
);
999 if (bitmap_equals(result
, tdata
.base
))
1000 fprintf(stderr
, "OK!\n");
1002 fprintf(stderr
, "Mismatch!\n");
1004 bitmap_free(result
);
1007 static int rebuild_bitmap(uint32_t *reposition
,
1008 struct ewah_bitmap
*source
,
1009 struct bitmap
*dest
)
1012 struct ewah_iterator it
;
1015 ewah_iterator_init(&it
, source
);
1017 while (ewah_iterator_next(&word
, &it
)) {
1018 uint32_t offset
, bit_pos
;
1020 for (offset
= 0; offset
< BITS_IN_EWORD
; ++offset
) {
1021 if ((word
>> offset
) == 0)
1024 offset
+= ewah_bit_ctz64(word
>> offset
);
1026 bit_pos
= reposition
[pos
+ offset
];
1028 bitmap_set(dest
, bit_pos
- 1);
1029 else /* can't reuse, we don't have the object */
1033 pos
+= BITS_IN_EWORD
;
1038 int rebuild_existing_bitmaps(struct bitmap_index
*bitmap_git
,
1039 struct packing_data
*mapping
,
1040 khash_sha1
*reused_bitmaps
,
1043 uint32_t i
, num_objects
;
1044 uint32_t *reposition
;
1045 struct bitmap
*rebuild
;
1046 struct stored_bitmap
*stored
;
1047 struct progress
*progress
= NULL
;
1052 num_objects
= bitmap_git
->pack
->num_objects
;
1053 reposition
= xcalloc(num_objects
, sizeof(uint32_t));
1055 for (i
= 0; i
< num_objects
; ++i
) {
1056 const unsigned char *sha1
;
1057 struct revindex_entry
*entry
;
1058 struct object_entry
*oe
;
1060 entry
= &bitmap_git
->pack
->revindex
[i
];
1061 sha1
= nth_packed_object_sha1(bitmap_git
->pack
, entry
->nr
);
1062 oe
= packlist_find(mapping
, sha1
, NULL
);
1065 reposition
[i
] = oe_in_pack_pos(mapping
, oe
) + 1;
1068 rebuild
= bitmap_new();
1072 progress
= start_progress("Reusing bitmaps", 0);
1074 kh_foreach_value(bitmap_git
->bitmaps
, stored
, {
1075 if (stored
->flags
& BITMAP_FLAG_REUSE
) {
1076 if (!rebuild_bitmap(reposition
,
1077 lookup_stored_bitmap(stored
),
1079 hash_pos
= kh_put_sha1(reused_bitmaps
,
1082 kh_value(reused_bitmaps
, hash_pos
) =
1083 bitmap_to_ewah(rebuild
);
1085 bitmap_reset(rebuild
);
1086 display_progress(progress
, ++i
);
1090 stop_progress(&progress
);
1093 bitmap_free(rebuild
);