7 #include "list-objects.h"
9 #include "pack-bitmap.h"
10 #include "pack-revindex.h"
11 #include "pack-objects.h"
13 #include "repository.h"
14 #include "object-store.h"
15 #include "list-objects-filter-options.h"
18 * An entry on the bitmap index, representing the bitmap for a given
21 struct stored_bitmap
{
23 struct ewah_bitmap
*root
;
24 struct stored_bitmap
*xor;
29 * The active bitmap index for a repository. By design, repositories only have
30 * a single bitmap index available (the index for the biggest packfile in
31 * the repository), since bitmap indexes need full closure.
33 * If there is more than one bitmap index available (e.g. because of alternates),
34 * the active bitmap index is the largest one.
37 /* Packfile to which this bitmap index belongs to */
38 struct packed_git
*pack
;
41 * Mark the first `reuse_objects` in the packfile as reused:
42 * they will be sent as-is without using them for repacking
45 uint32_t reuse_objects
;
47 /* mmapped buffer of the whole bitmap index */
49 size_t map_size
; /* size of the mmaped buffer */
50 size_t map_pos
; /* current position when loading the index */
55 * Each bitmap marks which objects in the packfile are of the given
56 * type. This provides type information when yielding the objects from
57 * the packfile during a walk, which allows for better delta bases.
59 struct ewah_bitmap
*commits
;
60 struct ewah_bitmap
*trees
;
61 struct ewah_bitmap
*blobs
;
62 struct ewah_bitmap
*tags
;
64 /* Map from object ID -> `stored_bitmap` for all the bitmapped commits */
65 kh_oid_map_t
*bitmaps
;
67 /* Number of bitmapped commits */
70 /* If not NULL, this is a name-hash cache pointing into map. */
76 * When trying to perform bitmap operations with objects that are not
77 * packed in `pack`, these objects are added to this "fake index" and
78 * are assumed to appear at the end of the packfile for all operations
81 struct object
**objects
;
83 uint32_t count
, alloc
;
84 kh_oid_pos_t
*positions
;
87 /* Bitmap result of the last performed walk */
88 struct bitmap
*result
;
90 /* "have" bitmap from the last performed walk */
93 /* Version of the bitmap index */
97 static struct ewah_bitmap
*lookup_stored_bitmap(struct stored_bitmap
*st
)
99 struct ewah_bitmap
*parent
;
100 struct ewah_bitmap
*composed
;
105 composed
= ewah_pool_new();
106 parent
= lookup_stored_bitmap(st
->xor);
107 ewah_xor(st
->root
, parent
, composed
);
109 ewah_pool_free(st
->root
);
117 * Read a bitmap from the current read position on the mmaped
118 * index, and increase the read position accordingly
120 static struct ewah_bitmap
*read_bitmap_1(struct bitmap_index
*index
)
122 struct ewah_bitmap
*b
= ewah_pool_new();
124 ssize_t bitmap_size
= ewah_read_mmap(b
,
125 index
->map
+ index
->map_pos
,
126 index
->map_size
- index
->map_pos
);
128 if (bitmap_size
< 0) {
129 error("Failed to load bitmap index (corrupted?)");
134 index
->map_pos
+= bitmap_size
;
138 static int load_bitmap_header(struct bitmap_index
*index
)
140 struct bitmap_disk_header
*header
= (void *)index
->map
;
141 size_t header_size
= sizeof(*header
) - GIT_MAX_RAWSZ
+ the_hash_algo
->rawsz
;
143 if (index
->map_size
< header_size
+ the_hash_algo
->rawsz
)
144 return error("Corrupted bitmap index (too small)");
146 if (memcmp(header
->magic
, BITMAP_IDX_SIGNATURE
, sizeof(BITMAP_IDX_SIGNATURE
)) != 0)
147 return error("Corrupted bitmap index file (wrong header)");
149 index
->version
= ntohs(header
->version
);
150 if (index
->version
!= 1)
151 return error("Unsupported version for bitmap index file (%d)", index
->version
);
153 /* Parse known bitmap format options */
155 uint32_t flags
= ntohs(header
->options
);
156 size_t cache_size
= st_mult(index
->pack
->num_objects
, sizeof(uint32_t));
157 unsigned char *index_end
= index
->map
+ index
->map_size
- the_hash_algo
->rawsz
;
159 if ((flags
& BITMAP_OPT_FULL_DAG
) == 0)
160 return error("Unsupported options for bitmap index file "
161 "(Git requires BITMAP_OPT_FULL_DAG)");
163 if (flags
& BITMAP_OPT_HASH_CACHE
) {
164 if (cache_size
> index_end
- index
->map
- header_size
)
165 return error("corrupted bitmap index file (too short to fit hash cache)");
166 index
->hashes
= (void *)(index_end
- cache_size
);
167 index_end
-= cache_size
;
171 index
->entry_count
= ntohl(header
->entry_count
);
172 index
->map_pos
+= header_size
;
176 static struct stored_bitmap
*store_bitmap(struct bitmap_index
*index
,
177 struct ewah_bitmap
*root
,
178 const struct object_id
*oid
,
179 struct stored_bitmap
*xor_with
,
182 struct stored_bitmap
*stored
;
186 stored
= xmalloc(sizeof(struct stored_bitmap
));
188 stored
->xor = xor_with
;
189 stored
->flags
= flags
;
190 oidcpy(&stored
->oid
, oid
);
192 hash_pos
= kh_put_oid_map(index
->bitmaps
, stored
->oid
, &ret
);
194 /* a 0 return code means the insertion succeeded with no changes,
195 * because the SHA1 already existed on the map. this is bad, there
196 * shouldn't be duplicated commits in the index */
198 error("Duplicate entry in bitmap index: %s", oid_to_hex(oid
));
202 kh_value(index
->bitmaps
, hash_pos
) = stored
;
206 static inline uint32_t read_be32(const unsigned char *buffer
, size_t *pos
)
208 uint32_t result
= get_be32(buffer
+ *pos
);
209 (*pos
) += sizeof(result
);
213 static inline uint8_t read_u8(const unsigned char *buffer
, size_t *pos
)
215 return buffer
[(*pos
)++];
218 #define MAX_XOR_OFFSET 160
220 static int load_bitmap_entries_v1(struct bitmap_index
*index
)
223 struct stored_bitmap
*recent_bitmaps
[MAX_XOR_OFFSET
] = { NULL
};
225 for (i
= 0; i
< index
->entry_count
; ++i
) {
226 int xor_offset
, flags
;
227 struct ewah_bitmap
*bitmap
= NULL
;
228 struct stored_bitmap
*xor_bitmap
= NULL
;
229 uint32_t commit_idx_pos
;
230 struct object_id oid
;
232 if (index
->map_size
- index
->map_pos
< 6)
233 return error("corrupt ewah bitmap: truncated header for entry %d", i
);
235 commit_idx_pos
= read_be32(index
->map
, &index
->map_pos
);
236 xor_offset
= read_u8(index
->map
, &index
->map_pos
);
237 flags
= read_u8(index
->map
, &index
->map_pos
);
239 if (nth_packed_object_id(&oid
, index
->pack
, commit_idx_pos
) < 0)
240 return error("corrupt ewah bitmap: commit index %u out of range",
241 (unsigned)commit_idx_pos
);
243 bitmap
= read_bitmap_1(index
);
247 if (xor_offset
> MAX_XOR_OFFSET
|| xor_offset
> i
)
248 return error("Corrupted bitmap pack index");
250 if (xor_offset
> 0) {
251 xor_bitmap
= recent_bitmaps
[(i
- xor_offset
) % MAX_XOR_OFFSET
];
253 if (xor_bitmap
== NULL
)
254 return error("Invalid XOR offset in bitmap pack index");
257 recent_bitmaps
[i
% MAX_XOR_OFFSET
] = store_bitmap(
258 index
, bitmap
, &oid
, xor_bitmap
, flags
);
264 static char *pack_bitmap_filename(struct packed_git
*p
)
268 if (!strip_suffix(p
->pack_name
, ".pack", &len
))
269 BUG("pack_name does not end in .pack");
270 return xstrfmt("%.*s.bitmap", (int)len
, p
->pack_name
);
273 static int open_pack_bitmap_1(struct bitmap_index
*bitmap_git
, struct packed_git
*packfile
)
279 if (open_pack_index(packfile
))
282 idx_name
= pack_bitmap_filename(packfile
);
283 fd
= git_open(idx_name
);
289 if (fstat(fd
, &st
)) {
294 if (bitmap_git
->pack
) {
295 warning("ignoring extra bitmap file: %s", packfile
->pack_name
);
300 bitmap_git
->pack
= packfile
;
301 bitmap_git
->map_size
= xsize_t(st
.st_size
);
302 bitmap_git
->map
= xmmap(NULL
, bitmap_git
->map_size
, PROT_READ
, MAP_PRIVATE
, fd
, 0);
303 bitmap_git
->map_pos
= 0;
306 if (load_bitmap_header(bitmap_git
) < 0) {
307 munmap(bitmap_git
->map
, bitmap_git
->map_size
);
308 bitmap_git
->map
= NULL
;
309 bitmap_git
->map_size
= 0;
316 static int load_pack_bitmap(struct bitmap_index
*bitmap_git
)
318 assert(bitmap_git
->map
);
320 bitmap_git
->bitmaps
= kh_init_oid_map();
321 bitmap_git
->ext_index
.positions
= kh_init_oid_pos();
322 if (load_pack_revindex(bitmap_git
->pack
))
325 if (!(bitmap_git
->commits
= read_bitmap_1(bitmap_git
)) ||
326 !(bitmap_git
->trees
= read_bitmap_1(bitmap_git
)) ||
327 !(bitmap_git
->blobs
= read_bitmap_1(bitmap_git
)) ||
328 !(bitmap_git
->tags
= read_bitmap_1(bitmap_git
)))
331 if (load_bitmap_entries_v1(bitmap_git
) < 0)
337 munmap(bitmap_git
->map
, bitmap_git
->map_size
);
338 bitmap_git
->map
= NULL
;
339 bitmap_git
->map_size
= 0;
341 kh_destroy_oid_map(bitmap_git
->bitmaps
);
342 bitmap_git
->bitmaps
= NULL
;
344 kh_destroy_oid_pos(bitmap_git
->ext_index
.positions
);
345 bitmap_git
->ext_index
.positions
= NULL
;
350 static int open_pack_bitmap(struct repository
*r
,
351 struct bitmap_index
*bitmap_git
)
353 struct packed_git
*p
;
356 assert(!bitmap_git
->map
);
358 for (p
= get_all_packs(r
); p
; p
= p
->next
) {
359 if (open_pack_bitmap_1(bitmap_git
, p
) == 0)
366 struct bitmap_index
*prepare_bitmap_git(struct repository
*r
)
368 struct bitmap_index
*bitmap_git
= xcalloc(1, sizeof(*bitmap_git
));
370 if (!open_pack_bitmap(r
, bitmap_git
) && !load_pack_bitmap(bitmap_git
))
373 free_bitmap_index(bitmap_git
);
377 struct include_data
{
378 struct bitmap_index
*bitmap_git
;
383 struct ewah_bitmap
*bitmap_for_commit(struct bitmap_index
*bitmap_git
,
384 struct commit
*commit
)
386 khiter_t hash_pos
= kh_get_oid_map(bitmap_git
->bitmaps
,
388 if (hash_pos
>= kh_end(bitmap_git
->bitmaps
))
390 return lookup_stored_bitmap(kh_value(bitmap_git
->bitmaps
, hash_pos
));
393 static inline int bitmap_position_extended(struct bitmap_index
*bitmap_git
,
394 const struct object_id
*oid
)
396 kh_oid_pos_t
*positions
= bitmap_git
->ext_index
.positions
;
397 khiter_t pos
= kh_get_oid_pos(positions
, *oid
);
399 if (pos
< kh_end(positions
)) {
400 int bitmap_pos
= kh_value(positions
, pos
);
401 return bitmap_pos
+ bitmap_git
->pack
->num_objects
;
407 static inline int bitmap_position_packfile(struct bitmap_index
*bitmap_git
,
408 const struct object_id
*oid
)
410 off_t offset
= find_pack_entry_one(oid
->hash
, bitmap_git
->pack
);
414 return find_revindex_position(bitmap_git
->pack
, offset
);
417 static int bitmap_position(struct bitmap_index
*bitmap_git
,
418 const struct object_id
*oid
)
420 int pos
= bitmap_position_packfile(bitmap_git
, oid
);
421 return (pos
>= 0) ? pos
: bitmap_position_extended(bitmap_git
, oid
);
424 static int ext_index_add_object(struct bitmap_index
*bitmap_git
,
425 struct object
*object
, const char *name
)
427 struct eindex
*eindex
= &bitmap_git
->ext_index
;
433 hash_pos
= kh_put_oid_pos(eindex
->positions
, object
->oid
, &hash_ret
);
435 if (eindex
->count
>= eindex
->alloc
) {
436 eindex
->alloc
= (eindex
->alloc
+ 16) * 3 / 2;
437 REALLOC_ARRAY(eindex
->objects
, eindex
->alloc
);
438 REALLOC_ARRAY(eindex
->hashes
, eindex
->alloc
);
441 bitmap_pos
= eindex
->count
;
442 eindex
->objects
[eindex
->count
] = object
;
443 eindex
->hashes
[eindex
->count
] = pack_name_hash(name
);
444 kh_value(eindex
->positions
, hash_pos
) = bitmap_pos
;
447 bitmap_pos
= kh_value(eindex
->positions
, hash_pos
);
450 return bitmap_pos
+ bitmap_git
->pack
->num_objects
;
453 struct bitmap_show_data
{
454 struct bitmap_index
*bitmap_git
;
458 static void show_object(struct object
*object
, const char *name
, void *data_
)
460 struct bitmap_show_data
*data
= data_
;
463 bitmap_pos
= bitmap_position(data
->bitmap_git
, &object
->oid
);
466 bitmap_pos
= ext_index_add_object(data
->bitmap_git
, object
,
469 bitmap_set(data
->base
, bitmap_pos
);
472 static void show_commit(struct commit
*commit
, void *data
)
476 static int add_to_include_set(struct bitmap_index
*bitmap_git
,
477 struct include_data
*data
,
478 struct commit
*commit
,
481 struct ewah_bitmap
*partial
;
483 if (data
->seen
&& bitmap_get(data
->seen
, bitmap_pos
))
486 if (bitmap_get(data
->base
, bitmap_pos
))
489 partial
= bitmap_for_commit(bitmap_git
, commit
);
491 bitmap_or_ewah(data
->base
, partial
);
495 bitmap_set(data
->base
, bitmap_pos
);
499 static int should_include(struct commit
*commit
, void *_data
)
501 struct include_data
*data
= _data
;
504 bitmap_pos
= bitmap_position(data
->bitmap_git
, &commit
->object
.oid
);
506 bitmap_pos
= ext_index_add_object(data
->bitmap_git
,
507 (struct object
*)commit
,
510 if (!add_to_include_set(data
->bitmap_git
, data
, commit
, bitmap_pos
)) {
511 struct commit_list
*parent
= commit
->parents
;
514 parent
->item
->object
.flags
|= SEEN
;
515 parent
= parent
->next
;
524 static int add_commit_to_bitmap(struct bitmap_index
*bitmap_git
,
525 struct bitmap
**base
,
526 struct commit
*commit
)
528 struct ewah_bitmap
*or_with
= bitmap_for_commit(bitmap_git
, commit
);
534 *base
= ewah_to_bitmap(or_with
);
536 bitmap_or_ewah(*base
, or_with
);
541 static struct bitmap
*find_objects(struct bitmap_index
*bitmap_git
,
542 struct rev_info
*revs
,
543 struct object_list
*roots
,
545 struct list_objects_filter_options
*filter
)
547 struct bitmap
*base
= NULL
;
550 struct object_list
*not_mapped
= NULL
;
553 * Go through all the roots for the walk. The ones that have bitmaps
554 * on the bitmap index will be `or`ed together to form an initial
555 * global reachability analysis.
557 * The ones without bitmaps in the index will be stored in the
558 * `not_mapped_list` for further processing.
561 struct object
*object
= roots
->item
;
564 if (object
->type
== OBJ_COMMIT
&&
565 add_commit_to_bitmap(bitmap_git
, &base
, (struct commit
*)object
)) {
566 object
->flags
|= SEEN
;
570 object_list_insert(object
, ¬_mapped
);
574 * Best case scenario: We found bitmaps for all the roots,
575 * so the resulting `or` bitmap has the full reachability analysis
577 if (not_mapped
== NULL
)
583 * Let's iterate through all the roots that don't have bitmaps to
584 * check if we can determine them to be reachable from the existing
587 * If we cannot find them in the existing global bitmap, we'll need
588 * to push them to an actual walk and run it until we can confirm
592 struct object
*object
= roots
->item
;
596 pos
= bitmap_position(bitmap_git
, &object
->oid
);
598 if (pos
< 0 || base
== NULL
|| !bitmap_get(base
, pos
)) {
599 object
->flags
&= ~UNINTERESTING
;
600 add_pending_object(revs
, object
, "");
603 object
->flags
|= SEEN
;
608 struct include_data incdata
;
609 struct bitmap_show_data show_data
;
614 incdata
.bitmap_git
= bitmap_git
;
618 revs
->include_check
= should_include
;
619 revs
->include_check_data
= &incdata
;
621 if (prepare_revision_walk(revs
))
622 die("revision walk setup failed");
624 show_data
.bitmap_git
= bitmap_git
;
625 show_data
.base
= base
;
627 traverse_commit_list_filtered(filter
, revs
,
628 show_commit
, show_object
,
635 static void show_extended_objects(struct bitmap_index
*bitmap_git
,
636 struct rev_info
*revs
,
637 show_reachable_fn show_reach
)
639 struct bitmap
*objects
= bitmap_git
->result
;
640 struct eindex
*eindex
= &bitmap_git
->ext_index
;
643 for (i
= 0; i
< eindex
->count
; ++i
) {
646 if (!bitmap_get(objects
, bitmap_git
->pack
->num_objects
+ i
))
649 obj
= eindex
->objects
[i
];
650 if ((obj
->type
== OBJ_BLOB
&& !revs
->blob_objects
) ||
651 (obj
->type
== OBJ_TREE
&& !revs
->tree_objects
) ||
652 (obj
->type
== OBJ_TAG
&& !revs
->tag_objects
))
655 show_reach(&obj
->oid
, obj
->type
, 0, eindex
->hashes
[i
], NULL
, 0);
659 static void init_type_iterator(struct ewah_iterator
*it
,
660 struct bitmap_index
*bitmap_git
,
661 enum object_type type
)
665 ewah_iterator_init(it
, bitmap_git
->commits
);
669 ewah_iterator_init(it
, bitmap_git
->trees
);
673 ewah_iterator_init(it
, bitmap_git
->blobs
);
677 ewah_iterator_init(it
, bitmap_git
->tags
);
681 BUG("object type %d not stored by bitmap type index", type
);
686 static void show_objects_for_type(
687 struct bitmap_index
*bitmap_git
,
688 enum object_type object_type
,
689 show_reachable_fn show_reach
)
694 struct ewah_iterator it
;
697 struct bitmap
*objects
= bitmap_git
->result
;
699 init_type_iterator(&it
, bitmap_git
, object_type
);
701 for (i
= 0; i
< objects
->word_alloc
&&
702 ewah_iterator_next(&filter
, &it
); i
++) {
703 eword_t word
= objects
->words
[i
] & filter
;
704 size_t pos
= (i
* BITS_IN_EWORD
);
709 for (offset
= 0; offset
< BITS_IN_EWORD
; ++offset
) {
710 struct object_id oid
;
711 struct revindex_entry
*entry
;
714 if ((word
>> offset
) == 0)
717 offset
+= ewah_bit_ctz64(word
>> offset
);
719 entry
= &bitmap_git
->pack
->revindex
[pos
+ offset
];
720 nth_packed_object_id(&oid
, bitmap_git
->pack
, entry
->nr
);
722 if (bitmap_git
->hashes
)
723 hash
= get_be32(bitmap_git
->hashes
+ entry
->nr
);
725 show_reach(&oid
, object_type
, 0, hash
, bitmap_git
->pack
, entry
->offset
);
730 static int in_bitmapped_pack(struct bitmap_index
*bitmap_git
,
731 struct object_list
*roots
)
734 struct object
*object
= roots
->item
;
737 if (find_pack_entry_one(object
->oid
.hash
, bitmap_git
->pack
) > 0)
744 static struct bitmap
*find_tip_objects(struct bitmap_index
*bitmap_git
,
745 struct object_list
*tip_objects
,
746 enum object_type type
)
748 struct bitmap
*result
= bitmap_new();
749 struct object_list
*p
;
751 for (p
= tip_objects
; p
; p
= p
->next
) {
754 if (p
->item
->type
!= type
)
757 pos
= bitmap_position(bitmap_git
, &p
->item
->oid
);
761 bitmap_set(result
, pos
);
767 static void filter_bitmap_exclude_type(struct bitmap_index
*bitmap_git
,
768 struct object_list
*tip_objects
,
769 struct bitmap
*to_filter
,
770 enum object_type type
)
772 struct eindex
*eindex
= &bitmap_git
->ext_index
;
774 struct ewah_iterator it
;
778 if (type
!= OBJ_BLOB
&& type
!= OBJ_TREE
)
779 BUG("filter_bitmap_exclude_type: unsupported type '%d'", type
);
782 * The non-bitmap version of this filter never removes
783 * objects which the other side specifically asked for,
784 * so we must match that behavior.
786 tips
= find_tip_objects(bitmap_git
, tip_objects
, type
);
789 * We can use the blob type-bitmap to work in whole words
790 * for the objects that are actually in the bitmapped packfile.
792 for (i
= 0, init_type_iterator(&it
, bitmap_git
, type
);
793 i
< to_filter
->word_alloc
&& ewah_iterator_next(&mask
, &it
);
795 if (i
< tips
->word_alloc
)
796 mask
&= ~tips
->words
[i
];
797 to_filter
->words
[i
] &= ~mask
;
801 * Clear any blobs that weren't in the packfile (and so would not have
802 * been caught by the loop above. We'll have to check them
805 for (i
= 0; i
< eindex
->count
; i
++) {
806 uint32_t pos
= i
+ bitmap_git
->pack
->num_objects
;
807 if (eindex
->objects
[i
]->type
== type
&&
808 bitmap_get(to_filter
, pos
) &&
809 !bitmap_get(tips
, pos
))
810 bitmap_unset(to_filter
, pos
);
816 static void filter_bitmap_blob_none(struct bitmap_index
*bitmap_git
,
817 struct object_list
*tip_objects
,
818 struct bitmap
*to_filter
)
820 filter_bitmap_exclude_type(bitmap_git
, tip_objects
, to_filter
,
824 static unsigned long get_size_by_pos(struct bitmap_index
*bitmap_git
,
827 struct packed_git
*pack
= bitmap_git
->pack
;
829 struct object_info oi
= OBJECT_INFO_INIT
;
833 if (pos
< pack
->num_objects
) {
834 struct revindex_entry
*entry
= &pack
->revindex
[pos
];
835 if (packed_object_info(the_repository
, pack
,
836 entry
->offset
, &oi
) < 0) {
837 struct object_id oid
;
838 nth_packed_object_id(&oid
, pack
, entry
->nr
);
839 die(_("unable to get size of %s"), oid_to_hex(&oid
));
842 struct eindex
*eindex
= &bitmap_git
->ext_index
;
843 struct object
*obj
= eindex
->objects
[pos
- pack
->num_objects
];
844 if (oid_object_info_extended(the_repository
, &obj
->oid
, &oi
, 0) < 0)
845 die(_("unable to get size of %s"), oid_to_hex(&obj
->oid
));
851 static void filter_bitmap_blob_limit(struct bitmap_index
*bitmap_git
,
852 struct object_list
*tip_objects
,
853 struct bitmap
*to_filter
,
856 struct eindex
*eindex
= &bitmap_git
->ext_index
;
858 struct ewah_iterator it
;
862 tips
= find_tip_objects(bitmap_git
, tip_objects
, OBJ_BLOB
);
864 for (i
= 0, init_type_iterator(&it
, bitmap_git
, OBJ_BLOB
);
865 i
< to_filter
->word_alloc
&& ewah_iterator_next(&mask
, &it
);
867 eword_t word
= to_filter
->words
[i
] & mask
;
870 for (offset
= 0; offset
< BITS_IN_EWORD
; offset
++) {
873 if ((word
>> offset
) == 0)
875 offset
+= ewah_bit_ctz64(word
>> offset
);
876 pos
= i
* BITS_IN_EWORD
+ offset
;
878 if (!bitmap_get(tips
, pos
) &&
879 get_size_by_pos(bitmap_git
, pos
) >= limit
)
880 bitmap_unset(to_filter
, pos
);
884 for (i
= 0; i
< eindex
->count
; i
++) {
885 uint32_t pos
= i
+ bitmap_git
->pack
->num_objects
;
886 if (eindex
->objects
[i
]->type
== OBJ_BLOB
&&
887 bitmap_get(to_filter
, pos
) &&
888 !bitmap_get(tips
, pos
) &&
889 get_size_by_pos(bitmap_git
, pos
) >= limit
)
890 bitmap_unset(to_filter
, pos
);
896 static void filter_bitmap_tree_depth(struct bitmap_index
*bitmap_git
,
897 struct object_list
*tip_objects
,
898 struct bitmap
*to_filter
,
902 BUG("filter_bitmap_tree_depth given non-zero limit");
904 filter_bitmap_exclude_type(bitmap_git
, tip_objects
, to_filter
,
906 filter_bitmap_exclude_type(bitmap_git
, tip_objects
, to_filter
,
910 static int filter_bitmap(struct bitmap_index
*bitmap_git
,
911 struct object_list
*tip_objects
,
912 struct bitmap
*to_filter
,
913 struct list_objects_filter_options
*filter
)
915 if (!filter
|| filter
->choice
== LOFC_DISABLED
)
918 if (filter
->choice
== LOFC_BLOB_NONE
) {
920 filter_bitmap_blob_none(bitmap_git
, tip_objects
,
925 if (filter
->choice
== LOFC_BLOB_LIMIT
) {
927 filter_bitmap_blob_limit(bitmap_git
, tip_objects
,
929 filter
->blob_limit_value
);
933 if (filter
->choice
== LOFC_TREE_DEPTH
&&
934 filter
->tree_exclude_depth
== 0) {
936 filter_bitmap_tree_depth(bitmap_git
, tip_objects
,
938 filter
->tree_exclude_depth
);
942 /* filter choice not handled */
946 static int can_filter_bitmap(struct list_objects_filter_options
*filter
)
948 return !filter_bitmap(NULL
, NULL
, NULL
, filter
);
951 struct bitmap_index
*prepare_bitmap_walk(struct rev_info
*revs
,
952 struct list_objects_filter_options
*filter
)
956 struct object_list
*wants
= NULL
;
957 struct object_list
*haves
= NULL
;
959 struct bitmap
*wants_bitmap
= NULL
;
960 struct bitmap
*haves_bitmap
= NULL
;
962 struct bitmap_index
*bitmap_git
;
965 * We can't do pathspec limiting with bitmaps, because we don't know
966 * which commits are associated with which object changes (let alone
967 * even which objects are associated with which paths).
972 if (!can_filter_bitmap(filter
))
975 /* try to open a bitmapped pack, but don't parse it yet
976 * because we may not need to use it */
977 bitmap_git
= xcalloc(1, sizeof(*bitmap_git
));
978 if (open_pack_bitmap(revs
->repo
, bitmap_git
) < 0)
981 for (i
= 0; i
< revs
->pending
.nr
; ++i
) {
982 struct object
*object
= revs
->pending
.objects
[i
].item
;
984 if (object
->type
== OBJ_NONE
)
985 parse_object_or_die(&object
->oid
, NULL
);
987 while (object
->type
== OBJ_TAG
) {
988 struct tag
*tag
= (struct tag
*) object
;
990 if (object
->flags
& UNINTERESTING
)
991 object_list_insert(object
, &haves
);
993 object_list_insert(object
, &wants
);
995 object
= parse_object_or_die(get_tagged_oid(tag
), NULL
);
998 if (object
->flags
& UNINTERESTING
)
999 object_list_insert(object
, &haves
);
1001 object_list_insert(object
, &wants
);
1005 * if we have a HAVES list, but none of those haves is contained
1006 * in the packfile that has a bitmap, we don't have anything to
1009 if (haves
&& !in_bitmapped_pack(bitmap_git
, haves
))
1012 /* if we don't want anything, we're done here */
1017 * now we're going to use bitmaps, so load the actual bitmap entries
1018 * from disk. this is the point of no return; after this the rev_list
1019 * becomes invalidated and we must perform the revwalk through bitmaps
1021 if (load_pack_bitmap(bitmap_git
) < 0)
1024 object_array_clear(&revs
->pending
);
1027 revs
->ignore_missing_links
= 1;
1028 haves_bitmap
= find_objects(bitmap_git
, revs
, haves
, NULL
,
1030 reset_revision_walk();
1031 revs
->ignore_missing_links
= 0;
1033 if (haves_bitmap
== NULL
)
1034 BUG("failed to perform bitmap walk");
1037 wants_bitmap
= find_objects(bitmap_git
, revs
, wants
, haves_bitmap
,
1041 BUG("failed to perform bitmap walk");
1044 bitmap_and_not(wants_bitmap
, haves_bitmap
);
1046 filter_bitmap(bitmap_git
, wants
, wants_bitmap
, filter
);
1048 bitmap_git
->result
= wants_bitmap
;
1049 bitmap_git
->haves
= haves_bitmap
;
1051 object_list_free(&wants
);
1052 object_list_free(&haves
);
1057 free_bitmap_index(bitmap_git
);
1058 object_list_free(&wants
);
1059 object_list_free(&haves
);
1063 static void try_partial_reuse(struct bitmap_index
*bitmap_git
,
1065 struct bitmap
*reuse
,
1066 struct pack_window
**w_curs
)
1068 struct revindex_entry
*revidx
;
1070 enum object_type type
;
1073 if (pos
>= bitmap_git
->pack
->num_objects
)
1074 return; /* not actually in the pack */
1076 revidx
= &bitmap_git
->pack
->revindex
[pos
];
1077 offset
= revidx
->offset
;
1078 type
= unpack_object_header(bitmap_git
->pack
, w_curs
, &offset
, &size
);
1080 return; /* broken packfile, punt */
1082 if (type
== OBJ_REF_DELTA
|| type
== OBJ_OFS_DELTA
) {
1087 * Find the position of the base object so we can look it up
1088 * in our bitmaps. If we can't come up with an offset, or if
1089 * that offset is not in the revidx, the pack is corrupt.
1090 * There's nothing we can do, so just punt on this object,
1091 * and the normal slow path will complain about it in
1094 base_offset
= get_delta_base(bitmap_git
->pack
, w_curs
,
1095 &offset
, type
, revidx
->offset
);
1098 base_pos
= find_revindex_position(bitmap_git
->pack
, base_offset
);
1103 * We assume delta dependencies always point backwards. This
1104 * lets us do a single pass, and is basically always true
1105 * due to the way OFS_DELTAs work. You would not typically
1106 * find REF_DELTA in a bitmapped pack, since we only bitmap
1107 * packs we write fresh, and OFS_DELTA is the default). But
1108 * let's double check to make sure the pack wasn't written with
1111 if (base_pos
>= pos
)
1115 * And finally, if we're not sending the base as part of our
1116 * reuse chunk, then don't send this object either. The base
1117 * would come after us, along with other objects not
1118 * necessarily in the pack, which means we'd need to convert
1119 * to REF_DELTA on the fly. Better to just let the normal
1120 * object_entry code path handle it.
1122 if (!bitmap_get(reuse
, base_pos
))
1127 * If we got here, then the object is OK to reuse. Mark it.
1129 bitmap_set(reuse
, pos
);
1132 int reuse_partial_packfile_from_bitmap(struct bitmap_index
*bitmap_git
,
1133 struct packed_git
**packfile_out
,
1135 struct bitmap
**reuse_out
)
1137 struct bitmap
*result
= bitmap_git
->result
;
1138 struct bitmap
*reuse
;
1139 struct pack_window
*w_curs
= NULL
;
1145 while (i
< result
->word_alloc
&& result
->words
[i
] == (eword_t
)~0)
1148 /* Don't mark objects not in the packfile */
1149 if (i
> bitmap_git
->pack
->num_objects
/ BITS_IN_EWORD
)
1150 i
= bitmap_git
->pack
->num_objects
/ BITS_IN_EWORD
;
1152 reuse
= bitmap_word_alloc(i
);
1153 memset(reuse
->words
, 0xFF, i
* sizeof(eword_t
));
1155 for (; i
< result
->word_alloc
; ++i
) {
1156 eword_t word
= result
->words
[i
];
1157 size_t pos
= (i
* BITS_IN_EWORD
);
1159 for (offset
= 0; offset
< BITS_IN_EWORD
; ++offset
) {
1160 if ((word
>> offset
) == 0)
1163 offset
+= ewah_bit_ctz64(word
>> offset
);
1164 try_partial_reuse(bitmap_git
, pos
+ offset
, reuse
, &w_curs
);
1168 unuse_pack(&w_curs
);
1170 *entries
= bitmap_popcount(reuse
);
1177 * Drop any reused objects from the result, since they will not
1178 * need to be handled separately.
1180 bitmap_and_not(result
, reuse
);
1181 *packfile_out
= bitmap_git
->pack
;
1186 int bitmap_walk_contains(struct bitmap_index
*bitmap_git
,
1187 struct bitmap
*bitmap
, const struct object_id
*oid
)
1194 idx
= bitmap_position(bitmap_git
, oid
);
1195 return idx
>= 0 && bitmap_get(bitmap
, idx
);
1198 void traverse_bitmap_commit_list(struct bitmap_index
*bitmap_git
,
1199 struct rev_info
*revs
,
1200 show_reachable_fn show_reachable
)
1202 assert(bitmap_git
->result
);
1204 show_objects_for_type(bitmap_git
, OBJ_COMMIT
, show_reachable
);
1205 if (revs
->tree_objects
)
1206 show_objects_for_type(bitmap_git
, OBJ_TREE
, show_reachable
);
1207 if (revs
->blob_objects
)
1208 show_objects_for_type(bitmap_git
, OBJ_BLOB
, show_reachable
);
1209 if (revs
->tag_objects
)
1210 show_objects_for_type(bitmap_git
, OBJ_TAG
, show_reachable
);
1212 show_extended_objects(bitmap_git
, revs
, show_reachable
);
1215 static uint32_t count_object_type(struct bitmap_index
*bitmap_git
,
1216 enum object_type type
)
1218 struct bitmap
*objects
= bitmap_git
->result
;
1219 struct eindex
*eindex
= &bitmap_git
->ext_index
;
1221 uint32_t i
= 0, count
= 0;
1222 struct ewah_iterator it
;
1225 init_type_iterator(&it
, bitmap_git
, type
);
1227 while (i
< objects
->word_alloc
&& ewah_iterator_next(&filter
, &it
)) {
1228 eword_t word
= objects
->words
[i
++] & filter
;
1229 count
+= ewah_bit_popcount64(word
);
1232 for (i
= 0; i
< eindex
->count
; ++i
) {
1233 if (eindex
->objects
[i
]->type
== type
&&
1234 bitmap_get(objects
, bitmap_git
->pack
->num_objects
+ i
))
1241 void count_bitmap_commit_list(struct bitmap_index
*bitmap_git
,
1242 uint32_t *commits
, uint32_t *trees
,
1243 uint32_t *blobs
, uint32_t *tags
)
1245 assert(bitmap_git
->result
);
1248 *commits
= count_object_type(bitmap_git
, OBJ_COMMIT
);
1251 *trees
= count_object_type(bitmap_git
, OBJ_TREE
);
1254 *blobs
= count_object_type(bitmap_git
, OBJ_BLOB
);
1257 *tags
= count_object_type(bitmap_git
, OBJ_TAG
);
1260 struct bitmap_test_data
{
1261 struct bitmap_index
*bitmap_git
;
1262 struct bitmap
*base
;
1263 struct progress
*prg
;
1267 static void test_show_object(struct object
*object
, const char *name
,
1270 struct bitmap_test_data
*tdata
= data
;
1273 bitmap_pos
= bitmap_position(tdata
->bitmap_git
, &object
->oid
);
1275 die("Object not in bitmap: %s\n", oid_to_hex(&object
->oid
));
1277 bitmap_set(tdata
->base
, bitmap_pos
);
1278 display_progress(tdata
->prg
, ++tdata
->seen
);
1281 static void test_show_commit(struct commit
*commit
, void *data
)
1283 struct bitmap_test_data
*tdata
= data
;
1286 bitmap_pos
= bitmap_position(tdata
->bitmap_git
,
1287 &commit
->object
.oid
);
1289 die("Object not in bitmap: %s\n", oid_to_hex(&commit
->object
.oid
));
1291 bitmap_set(tdata
->base
, bitmap_pos
);
1292 display_progress(tdata
->prg
, ++tdata
->seen
);
1295 void test_bitmap_walk(struct rev_info
*revs
)
1297 struct object
*root
;
1298 struct bitmap
*result
= NULL
;
1299 size_t result_popcnt
;
1300 struct bitmap_test_data tdata
;
1301 struct bitmap_index
*bitmap_git
;
1302 struct ewah_bitmap
*bm
;
1304 if (!(bitmap_git
= prepare_bitmap_git(revs
->repo
)))
1305 die("failed to load bitmap indexes");
1307 if (revs
->pending
.nr
!= 1)
1308 die("you must specify exactly one commit to test");
1310 fprintf(stderr
, "Bitmap v%d test (%d entries loaded)\n",
1311 bitmap_git
->version
, bitmap_git
->entry_count
);
1313 root
= revs
->pending
.objects
[0].item
;
1314 bm
= bitmap_for_commit(bitmap_git
, (struct commit
*)root
);
1317 fprintf(stderr
, "Found bitmap for %s. %d bits / %08x checksum\n",
1318 oid_to_hex(&root
->oid
), (int)bm
->bit_size
, ewah_checksum(bm
));
1320 result
= ewah_to_bitmap(bm
);
1324 die("Commit %s doesn't have an indexed bitmap", oid_to_hex(&root
->oid
));
1326 revs
->tag_objects
= 1;
1327 revs
->tree_objects
= 1;
1328 revs
->blob_objects
= 1;
1330 result_popcnt
= bitmap_popcount(result
);
1332 if (prepare_revision_walk(revs
))
1333 die("revision walk setup failed");
1335 tdata
.bitmap_git
= bitmap_git
;
1336 tdata
.base
= bitmap_new();
1337 tdata
.prg
= start_progress("Verifying bitmap entries", result_popcnt
);
1340 traverse_commit_list(revs
, &test_show_commit
, &test_show_object
, &tdata
);
1342 stop_progress(&tdata
.prg
);
1344 if (bitmap_equals(result
, tdata
.base
))
1345 fprintf(stderr
, "OK!\n");
1347 die("mismatch in bitmap results");
1349 free_bitmap_index(bitmap_git
);
1352 int rebuild_bitmap(const uint32_t *reposition
,
1353 struct ewah_bitmap
*source
,
1354 struct bitmap
*dest
)
1357 struct ewah_iterator it
;
1360 ewah_iterator_init(&it
, source
);
1362 while (ewah_iterator_next(&word
, &it
)) {
1363 uint32_t offset
, bit_pos
;
1365 for (offset
= 0; offset
< BITS_IN_EWORD
; ++offset
) {
1366 if ((word
>> offset
) == 0)
1369 offset
+= ewah_bit_ctz64(word
>> offset
);
1371 bit_pos
= reposition
[pos
+ offset
];
1373 bitmap_set(dest
, bit_pos
- 1);
1374 else /* can't reuse, we don't have the object */
1378 pos
+= BITS_IN_EWORD
;
1383 uint32_t *create_bitmap_mapping(struct bitmap_index
*bitmap_git
,
1384 struct packing_data
*mapping
)
1386 uint32_t i
, num_objects
;
1387 uint32_t *reposition
;
1389 num_objects
= bitmap_git
->pack
->num_objects
;
1390 reposition
= xcalloc(num_objects
, sizeof(uint32_t));
1392 for (i
= 0; i
< num_objects
; ++i
) {
1393 struct object_id oid
;
1394 struct revindex_entry
*entry
;
1395 struct object_entry
*oe
;
1397 entry
= &bitmap_git
->pack
->revindex
[i
];
1398 nth_packed_object_id(&oid
, bitmap_git
->pack
, entry
->nr
);
1399 oe
= packlist_find(mapping
, &oid
);
1402 reposition
[i
] = oe_in_pack_pos(mapping
, oe
) + 1;
1408 void free_bitmap_index(struct bitmap_index
*b
)
1414 munmap(b
->map
, b
->map_size
);
1415 ewah_pool_free(b
->commits
);
1416 ewah_pool_free(b
->trees
);
1417 ewah_pool_free(b
->blobs
);
1418 ewah_pool_free(b
->tags
);
1419 kh_destroy_oid_map(b
->bitmaps
);
1420 free(b
->ext_index
.objects
);
1421 free(b
->ext_index
.hashes
);
1422 bitmap_free(b
->result
);
1423 bitmap_free(b
->haves
);
1427 int bitmap_has_oid_in_uninteresting(struct bitmap_index
*bitmap_git
,
1428 const struct object_id
*oid
)
1430 return bitmap_git
&&
1431 bitmap_walk_contains(bitmap_git
, bitmap_git
->haves
, oid
);