]>
Commit | Line | Data |
---|---|---|
fff42755 VM |
1 | #include "cache.h" |
2 | #include "commit.h" | |
baf20c39 | 3 | #include "strbuf.h" |
fff42755 VM |
4 | #include "tag.h" |
5 | #include "diff.h" | |
6 | #include "revision.h" | |
7 | #include "progress.h" | |
8 | #include "list-objects.h" | |
9 | #include "pack.h" | |
10 | #include "pack-bitmap.h" | |
11 | #include "pack-revindex.h" | |
12 | #include "pack-objects.h" | |
0317f455 | 13 | #include "packfile.h" |
a80d72db SB |
14 | #include "repository.h" |
15 | #include "object-store.h" | |
6663ae0a | 16 | #include "list-objects-filter-options.h" |
0f533c72 | 17 | #include "midx.h" |
3f267a11 | 18 | #include "config.h" |
fff42755 VM |
19 | |
20 | /* | |
21 | * An entry on the bitmap index, representing the bitmap for a given | |
22 | * commit. | |
23 | */ | |
24 | struct stored_bitmap { | |
53636539 | 25 | struct object_id oid; |
fff42755 VM |
26 | struct ewah_bitmap *root; |
27 | struct stored_bitmap *xor; | |
28 | int flags; | |
29 | }; | |
30 | ||
31 | /* | |
3ae5fa07 | 32 | * The active bitmap index for a repository. By design, repositories only have |
fff42755 VM |
33 | * a single bitmap index available (the index for the biggest packfile in |
34 | * the repository), since bitmap indexes need full closure. | |
35 | * | |
36 | * If there is more than one bitmap index available (e.g. because of alternates), | |
37 | * the active bitmap index is the largest one. | |
38 | */ | |
3ae5fa07 | 39 | struct bitmap_index { |
0f533c72 TB |
40 | /* |
41 | * The pack or multi-pack index (MIDX) that this bitmap index belongs | |
42 | * to. | |
43 | * | |
44 | * Exactly one of these must be non-NULL; this specifies the object | |
45 | * order used to interpret this bitmap. | |
46 | */ | |
fff42755 | 47 | struct packed_git *pack; |
0f533c72 | 48 | struct multi_pack_index *midx; |
fff42755 | 49 | |
fff42755 VM |
50 | /* |
51 | * Mark the first `reuse_objects` in the packfile as reused: | |
52 | * they will be sent as-is without using them for repacking | |
53 | * calculations | |
54 | */ | |
55 | uint32_t reuse_objects; | |
56 | ||
57 | /* mmapped buffer of the whole bitmap index */ | |
58 | unsigned char *map; | |
59 | size_t map_size; /* size of the mmaped buffer */ | |
60 | size_t map_pos; /* current position when loading the index */ | |
61 | ||
62 | /* | |
63 | * Type indexes. | |
64 | * | |
65 | * Each bitmap marks which objects in the packfile are of the given | |
66 | * type. This provides type information when yielding the objects from | |
67 | * the packfile during a walk, which allows for better delta bases. | |
68 | */ | |
69 | struct ewah_bitmap *commits; | |
70 | struct ewah_bitmap *trees; | |
71 | struct ewah_bitmap *blobs; | |
72 | struct ewah_bitmap *tags; | |
73 | ||
3c771448 | 74 | /* Map from object ID -> `stored_bitmap` for all the bitmapped commits */ |
75 | kh_oid_map_t *bitmaps; | |
fff42755 VM |
76 | |
77 | /* Number of bitmapped commits */ | |
78 | uint32_t entry_count; | |
79 | ||
f3c23db2 | 80 | /* If not NULL, this is a name-hash cache pointing into map. */ |
ae4f07fb VM |
81 | uint32_t *hashes; |
82 | ||
0f533c72 TB |
83 | /* The checksum of the packfile or MIDX; points into map. */ |
84 | const unsigned char *checksum; | |
85 | ||
fff42755 VM |
86 | /* |
87 | * Extended index. | |
88 | * | |
89 | * When trying to perform bitmap operations with objects that are not | |
90 | * packed in `pack`, these objects are added to this "fake index" and | |
91 | * are assumed to appear at the end of the packfile for all operations | |
92 | */ | |
93 | struct eindex { | |
94 | struct object **objects; | |
95 | uint32_t *hashes; | |
96 | uint32_t count, alloc; | |
3c771448 | 97 | kh_oid_pos_t *positions; |
fff42755 VM |
98 | } ext_index; |
99 | ||
100 | /* Bitmap result of the last performed walk */ | |
101 | struct bitmap *result; | |
102 | ||
30cdc33f JK |
103 | /* "have" bitmap from the last performed walk */ |
104 | struct bitmap *haves; | |
105 | ||
fff42755 VM |
106 | /* Version of the bitmap index */ |
107 | unsigned int version; | |
3ae5fa07 | 108 | }; |
fff42755 VM |
109 | |
110 | static struct ewah_bitmap *lookup_stored_bitmap(struct stored_bitmap *st) | |
111 | { | |
112 | struct ewah_bitmap *parent; | |
113 | struct ewah_bitmap *composed; | |
114 | ||
115 | if (st->xor == NULL) | |
116 | return st->root; | |
117 | ||
118 | composed = ewah_pool_new(); | |
119 | parent = lookup_stored_bitmap(st->xor); | |
120 | ewah_xor(st->root, parent, composed); | |
121 | ||
122 | ewah_pool_free(st->root); | |
123 | st->root = composed; | |
124 | st->xor = NULL; | |
125 | ||
126 | return composed; | |
127 | } | |
128 | ||
129 | /* | |
130 | * Read a bitmap from the current read position on the mmaped | |
131 | * index, and increase the read position accordingly | |
132 | */ | |
133 | static struct ewah_bitmap *read_bitmap_1(struct bitmap_index *index) | |
134 | { | |
135 | struct ewah_bitmap *b = ewah_pool_new(); | |
136 | ||
1140bf01 | 137 | ssize_t bitmap_size = ewah_read_mmap(b, |
fff42755 VM |
138 | index->map + index->map_pos, |
139 | index->map_size - index->map_pos); | |
140 | ||
141 | if (bitmap_size < 0) { | |
9975975d | 142 | error(_("failed to load bitmap index (corrupted?)")); |
fff42755 VM |
143 | ewah_pool_free(b); |
144 | return NULL; | |
145 | } | |
146 | ||
147 | index->map_pos += bitmap_size; | |
148 | return b; | |
149 | } | |
150 | ||
ed184620 TB |
151 | static uint32_t bitmap_num_objects(struct bitmap_index *index) |
152 | { | |
0f533c72 TB |
153 | if (index->midx) |
154 | return index->midx->num_objects; | |
ed184620 TB |
155 | return index->pack->num_objects; |
156 | } | |
157 | ||
fff42755 VM |
158 | static int load_bitmap_header(struct bitmap_index *index) |
159 | { | |
160 | struct bitmap_disk_header *header = (void *)index->map; | |
ca510902 | 161 | size_t header_size = sizeof(*header) - GIT_MAX_RAWSZ + the_hash_algo->rawsz; |
fff42755 | 162 | |
ca510902 | 163 | if (index->map_size < header_size + the_hash_algo->rawsz) |
9975975d | 164 | return error(_("corrupted bitmap index (too small)")); |
fff42755 VM |
165 | |
166 | if (memcmp(header->magic, BITMAP_IDX_SIGNATURE, sizeof(BITMAP_IDX_SIGNATURE)) != 0) | |
9975975d | 167 | return error(_("corrupted bitmap index file (wrong header)")); |
fff42755 VM |
168 | |
169 | index->version = ntohs(header->version); | |
170 | if (index->version != 1) | |
9975975d | 171 | return error(_("unsupported version '%d' for bitmap index file"), index->version); |
fff42755 VM |
172 | |
173 | /* Parse known bitmap format options */ | |
174 | { | |
175 | uint32_t flags = ntohs(header->options); | |
ed184620 | 176 | size_t cache_size = st_mult(bitmap_num_objects(index), sizeof(uint32_t)); |
ec6c7b43 | 177 | unsigned char *index_end = index->map + index->map_size - the_hash_algo->rawsz; |
fff42755 VM |
178 | |
179 | if ((flags & BITMAP_OPT_FULL_DAG) == 0) | |
baf20c39 | 180 | BUG("unsupported options for bitmap index file " |
fff42755 | 181 | "(Git requires BITMAP_OPT_FULL_DAG)"); |
ae4f07fb VM |
182 | |
183 | if (flags & BITMAP_OPT_HASH_CACHE) { | |
ec6c7b43 | 184 | if (cache_size > index_end - index->map - header_size) |
9975975d | 185 | return error(_("corrupted bitmap index file (too short to fit hash cache)")); |
ec6c7b43 JK |
186 | index->hashes = (void *)(index_end - cache_size); |
187 | index_end -= cache_size; | |
ae4f07fb | 188 | } |
fff42755 VM |
189 | } |
190 | ||
191 | index->entry_count = ntohl(header->entry_count); | |
0f533c72 | 192 | index->checksum = header->checksum; |
ca510902 | 193 | index->map_pos += header_size; |
fff42755 VM |
194 | return 0; |
195 | } | |
196 | ||
197 | static struct stored_bitmap *store_bitmap(struct bitmap_index *index, | |
198 | struct ewah_bitmap *root, | |
500e4f23 | 199 | const struct object_id *oid, |
fff42755 VM |
200 | struct stored_bitmap *xor_with, |
201 | int flags) | |
202 | { | |
203 | struct stored_bitmap *stored; | |
204 | khiter_t hash_pos; | |
205 | int ret; | |
206 | ||
207 | stored = xmalloc(sizeof(struct stored_bitmap)); | |
208 | stored->root = root; | |
209 | stored->xor = xor_with; | |
210 | stored->flags = flags; | |
500e4f23 | 211 | oidcpy(&stored->oid, oid); |
fff42755 | 212 | |
3c771448 | 213 | hash_pos = kh_put_oid_map(index->bitmaps, stored->oid, &ret); |
fff42755 VM |
214 | |
215 | /* a 0 return code means the insertion succeeded with no changes, | |
216 | * because the SHA1 already existed on the map. this is bad, there | |
217 | * shouldn't be duplicated commits in the index */ | |
218 | if (ret == 0) { | |
9975975d | 219 | error(_("duplicate entry in bitmap index: '%s'"), oid_to_hex(oid)); |
fff42755 VM |
220 | return NULL; |
221 | } | |
222 | ||
223 | kh_value(index->bitmaps, hash_pos) = stored; | |
224 | return stored; | |
225 | } | |
226 | ||
b5007211 KB |
227 | static inline uint32_t read_be32(const unsigned char *buffer, size_t *pos) |
228 | { | |
229 | uint32_t result = get_be32(buffer + *pos); | |
230 | (*pos) += sizeof(result); | |
231 | return result; | |
232 | } | |
233 | ||
234 | static inline uint8_t read_u8(const unsigned char *buffer, size_t *pos) | |
235 | { | |
236 | return buffer[(*pos)++]; | |
237 | } | |
238 | ||
599dc766 RS |
239 | #define MAX_XOR_OFFSET 160 |
240 | ||
6b4277e6 TB |
241 | static int nth_bitmap_object_oid(struct bitmap_index *index, |
242 | struct object_id *oid, | |
243 | uint32_t n) | |
244 | { | |
0f533c72 TB |
245 | if (index->midx) |
246 | return nth_midxed_object_oid(oid, index->midx, n) ? 0 : -1; | |
6b4277e6 TB |
247 | return nth_packed_object_id(oid, index->pack, n); |
248 | } | |
249 | ||
fff42755 VM |
250 | static int load_bitmap_entries_v1(struct bitmap_index *index) |
251 | { | |
fff42755 | 252 | uint32_t i; |
599dc766 | 253 | struct stored_bitmap *recent_bitmaps[MAX_XOR_OFFSET] = { NULL }; |
fff42755 VM |
254 | |
255 | for (i = 0; i < index->entry_count; ++i) { | |
256 | int xor_offset, flags; | |
257 | struct ewah_bitmap *bitmap = NULL; | |
258 | struct stored_bitmap *xor_bitmap = NULL; | |
259 | uint32_t commit_idx_pos; | |
500e4f23 | 260 | struct object_id oid; |
fff42755 | 261 | |
c6b0c391 | 262 | if (index->map_size - index->map_pos < 6) |
9975975d | 263 | return error(_("corrupt ewah bitmap: truncated header for entry %d"), i); |
c6b0c391 | 264 | |
b5007211 KB |
265 | commit_idx_pos = read_be32(index->map, &index->map_pos); |
266 | xor_offset = read_u8(index->map, &index->map_pos); | |
267 | flags = read_u8(index->map, &index->map_pos); | |
fff42755 | 268 | |
6b4277e6 | 269 | if (nth_bitmap_object_oid(index, &oid, commit_idx_pos) < 0) |
9975975d | 270 | return error(_("corrupt ewah bitmap: commit index %u out of range"), |
c6b0c391 | 271 | (unsigned)commit_idx_pos); |
fff42755 | 272 | |
fff42755 VM |
273 | bitmap = read_bitmap_1(index); |
274 | if (!bitmap) | |
275 | return -1; | |
276 | ||
277 | if (xor_offset > MAX_XOR_OFFSET || xor_offset > i) | |
9975975d | 278 | return error(_("corrupted bitmap pack index")); |
fff42755 VM |
279 | |
280 | if (xor_offset > 0) { | |
281 | xor_bitmap = recent_bitmaps[(i - xor_offset) % MAX_XOR_OFFSET]; | |
282 | ||
283 | if (xor_bitmap == NULL) | |
9975975d | 284 | return error(_("invalid XOR offset in bitmap pack index")); |
fff42755 VM |
285 | } |
286 | ||
287 | recent_bitmaps[i % MAX_XOR_OFFSET] = store_bitmap( | |
500e4f23 | 288 | index, bitmap, &oid, xor_bitmap, flags); |
fff42755 VM |
289 | } |
290 | ||
291 | return 0; | |
292 | } | |
293 | ||
0f533c72 TB |
294 | char *midx_bitmap_filename(struct multi_pack_index *midx) |
295 | { | |
60980aed TB |
296 | struct strbuf buf = STRBUF_INIT; |
297 | ||
298 | get_midx_filename(&buf, midx->object_dir); | |
299 | strbuf_addf(&buf, "-%s.bitmap", hash_to_hex(get_midx_checksum(midx))); | |
300 | ||
301 | return strbuf_detach(&buf, NULL); | |
0f533c72 TB |
302 | } |
303 | ||
304 | char *pack_bitmap_filename(struct packed_git *p) | |
cb468050 | 305 | { |
9ae97018 | 306 | size_t len; |
cb468050 | 307 | |
9ae97018 | 308 | if (!strip_suffix(p->pack_name, ".pack", &len)) |
033abf97 | 309 | BUG("pack_name does not end in .pack"); |
9ae97018 | 310 | return xstrfmt("%.*s.bitmap", (int)len, p->pack_name); |
cb468050 JH |
311 | } |
312 | ||
0f533c72 TB |
313 | static int open_midx_bitmap_1(struct bitmap_index *bitmap_git, |
314 | struct multi_pack_index *midx) | |
315 | { | |
316 | struct stat st; | |
349c26ff TL |
317 | char *bitmap_name = midx_bitmap_filename(midx); |
318 | int fd = git_open(bitmap_name); | |
0f533c72 | 319 | |
6411cc08 TL |
320 | if (fd < 0) { |
321 | if (errno != ENOENT) | |
322 | warning_errno("cannot open '%s'", bitmap_name); | |
323 | free(bitmap_name); | |
0f533c72 | 324 | return -1; |
6411cc08 TL |
325 | } |
326 | free(bitmap_name); | |
0f533c72 TB |
327 | |
328 | if (fstat(fd, &st)) { | |
329 | close(fd); | |
330 | return -1; | |
331 | } | |
332 | ||
333 | if (bitmap_git->pack || bitmap_git->midx) { | |
60980aed TB |
334 | struct strbuf buf = STRBUF_INIT; |
335 | get_midx_filename(&buf, midx->object_dir); | |
0f533c72 | 336 | /* ignore extra bitmap file; we can only handle one */ |
9975975d | 337 | warning(_("ignoring extra bitmap file: '%s'"), buf.buf); |
0f533c72 | 338 | close(fd); |
60980aed | 339 | strbuf_release(&buf); |
0f533c72 TB |
340 | return -1; |
341 | } | |
342 | ||
343 | bitmap_git->midx = midx; | |
344 | bitmap_git->map_size = xsize_t(st.st_size); | |
345 | bitmap_git->map_pos = 0; | |
346 | bitmap_git->map = xmmap(NULL, bitmap_git->map_size, PROT_READ, | |
347 | MAP_PRIVATE, fd, 0); | |
348 | close(fd); | |
349 | ||
350 | if (load_bitmap_header(bitmap_git) < 0) | |
351 | goto cleanup; | |
352 | ||
353 | if (!hasheq(get_midx_checksum(bitmap_git->midx), bitmap_git->checksum)) | |
354 | goto cleanup; | |
355 | ||
356 | if (load_midx_revindex(bitmap_git->midx) < 0) { | |
357 | warning(_("multi-pack bitmap is missing required reverse index")); | |
358 | goto cleanup; | |
359 | } | |
360 | return 0; | |
361 | ||
362 | cleanup: | |
363 | munmap(bitmap_git->map, bitmap_git->map_size); | |
364 | bitmap_git->map_size = 0; | |
f8b60cf9 | 365 | bitmap_git->map_pos = 0; |
0f533c72 | 366 | bitmap_git->map = NULL; |
f8b60cf9 | 367 | bitmap_git->midx = NULL; |
0f533c72 TB |
368 | return -1; |
369 | } | |
370 | ||
3ae5fa07 | 371 | static int open_pack_bitmap_1(struct bitmap_index *bitmap_git, struct packed_git *packfile) |
fff42755 VM |
372 | { |
373 | int fd; | |
374 | struct stat st; | |
349c26ff | 375 | char *bitmap_name; |
fff42755 VM |
376 | |
377 | if (open_pack_index(packfile)) | |
378 | return -1; | |
379 | ||
349c26ff TL |
380 | bitmap_name = pack_bitmap_filename(packfile); |
381 | fd = git_open(bitmap_name); | |
fff42755 | 382 | |
6411cc08 TL |
383 | if (fd < 0) { |
384 | if (errno != ENOENT) | |
385 | warning_errno("cannot open '%s'", bitmap_name); | |
386 | free(bitmap_name); | |
fff42755 | 387 | return -1; |
6411cc08 TL |
388 | } |
389 | free(bitmap_name); | |
fff42755 VM |
390 | |
391 | if (fstat(fd, &st)) { | |
392 | close(fd); | |
393 | return -1; | |
394 | } | |
395 | ||
0f533c72 TB |
396 | if (bitmap_git->pack || bitmap_git->midx) { |
397 | /* ignore extra bitmap file; we can only handle one */ | |
9975975d | 398 | warning(_("ignoring extra bitmap file: '%s'"), packfile->pack_name); |
fff42755 VM |
399 | close(fd); |
400 | return -1; | |
401 | } | |
402 | ||
dc1daacd JK |
403 | if (!is_pack_valid(packfile)) { |
404 | close(fd); | |
405 | return -1; | |
406 | } | |
407 | ||
3ae5fa07 JT |
408 | bitmap_git->pack = packfile; |
409 | bitmap_git->map_size = xsize_t(st.st_size); | |
410 | bitmap_git->map = xmmap(NULL, bitmap_git->map_size, PROT_READ, MAP_PRIVATE, fd, 0); | |
411 | bitmap_git->map_pos = 0; | |
fff42755 VM |
412 | close(fd); |
413 | ||
3ae5fa07 JT |
414 | if (load_bitmap_header(bitmap_git) < 0) { |
415 | munmap(bitmap_git->map, bitmap_git->map_size); | |
416 | bitmap_git->map = NULL; | |
417 | bitmap_git->map_size = 0; | |
f8b60cf9 TB |
418 | bitmap_git->map_pos = 0; |
419 | bitmap_git->pack = NULL; | |
fff42755 VM |
420 | return -1; |
421 | } | |
422 | ||
423 | return 0; | |
424 | } | |
425 | ||
0f533c72 TB |
426 | static int load_reverse_index(struct bitmap_index *bitmap_git) |
427 | { | |
428 | if (bitmap_is_midx(bitmap_git)) { | |
429 | uint32_t i; | |
430 | int ret; | |
431 | ||
432 | /* | |
433 | * The multi-pack-index's .rev file is already loaded via | |
434 | * open_pack_bitmap_1(). | |
435 | * | |
436 | * But we still need to open the individual pack .rev files, | |
437 | * since we will need to make use of them in pack-objects. | |
438 | */ | |
439 | for (i = 0; i < bitmap_git->midx->num_packs; i++) { | |
440 | if (prepare_midx_pack(the_repository, bitmap_git->midx, i)) | |
441 | die(_("load_reverse_index: could not open pack")); | |
442 | ret = load_pack_revindex(bitmap_git->midx->packs[i]); | |
443 | if (ret) | |
444 | return ret; | |
445 | } | |
446 | return 0; | |
447 | } | |
448 | return load_pack_revindex(bitmap_git->pack); | |
449 | } | |
450 | ||
451 | static int load_bitmap(struct bitmap_index *bitmap_git) | |
fff42755 | 452 | { |
199c86be | 453 | assert(bitmap_git->map); |
fff42755 | 454 | |
3c771448 | 455 | bitmap_git->bitmaps = kh_init_oid_map(); |
456 | bitmap_git->ext_index.positions = kh_init_oid_pos(); | |
0f533c72 TB |
457 | |
458 | if (load_reverse_index(bitmap_git)) | |
4828ce98 | 459 | goto failed; |
fff42755 | 460 | |
3ae5fa07 JT |
461 | if (!(bitmap_git->commits = read_bitmap_1(bitmap_git)) || |
462 | !(bitmap_git->trees = read_bitmap_1(bitmap_git)) || | |
463 | !(bitmap_git->blobs = read_bitmap_1(bitmap_git)) || | |
464 | !(bitmap_git->tags = read_bitmap_1(bitmap_git))) | |
fff42755 VM |
465 | goto failed; |
466 | ||
3ae5fa07 | 467 | if (load_bitmap_entries_v1(bitmap_git) < 0) |
fff42755 VM |
468 | goto failed; |
469 | ||
fff42755 VM |
470 | return 0; |
471 | ||
472 | failed: | |
3ae5fa07 JT |
473 | munmap(bitmap_git->map, bitmap_git->map_size); |
474 | bitmap_git->map = NULL; | |
475 | bitmap_git->map_size = 0; | |
bb514de3 JK |
476 | |
477 | kh_destroy_oid_map(bitmap_git->bitmaps); | |
478 | bitmap_git->bitmaps = NULL; | |
479 | ||
480 | kh_destroy_oid_pos(bitmap_git->ext_index.positions); | |
481 | bitmap_git->ext_index.positions = NULL; | |
482 | ||
fff42755 VM |
483 | return -1; |
484 | } | |
485 | ||
7c141127 NTND |
486 | static int open_pack_bitmap(struct repository *r, |
487 | struct bitmap_index *bitmap_git) | |
fff42755 VM |
488 | { |
489 | struct packed_git *p; | |
490 | int ret = -1; | |
491 | ||
199c86be | 492 | assert(!bitmap_git->map); |
fff42755 | 493 | |
7c141127 | 494 | for (p = get_all_packs(r); p; p = p->next) { |
3ae5fa07 | 495 | if (open_pack_bitmap_1(bitmap_git, p) == 0) |
fff42755 VM |
496 | ret = 0; |
497 | } | |
498 | ||
499 | return ret; | |
500 | } | |
501 | ||
0f533c72 TB |
502 | static int open_midx_bitmap(struct repository *r, |
503 | struct bitmap_index *bitmap_git) | |
504 | { | |
505 | struct multi_pack_index *midx; | |
506 | ||
507 | assert(!bitmap_git->map); | |
508 | ||
509 | for (midx = get_multi_pack_index(r); midx; midx = midx->next) { | |
510 | if (!open_midx_bitmap_1(bitmap_git, midx)) | |
511 | return 0; | |
512 | } | |
513 | return -1; | |
514 | } | |
515 | ||
516 | static int open_bitmap(struct repository *r, | |
517 | struct bitmap_index *bitmap_git) | |
518 | { | |
519 | assert(!bitmap_git->map); | |
520 | ||
521 | if (!open_midx_bitmap(r, bitmap_git)) | |
522 | return 0; | |
523 | return open_pack_bitmap(r, bitmap_git); | |
524 | } | |
525 | ||
7c141127 | 526 | struct bitmap_index *prepare_bitmap_git(struct repository *r) |
fff42755 | 527 | { |
3ae5fa07 | 528 | struct bitmap_index *bitmap_git = xcalloc(1, sizeof(*bitmap_git)); |
fff42755 | 529 | |
0f533c72 TB |
530 | if (!open_bitmap(r, bitmap_git) && !load_bitmap(bitmap_git)) |
531 | return bitmap_git; | |
532 | ||
533 | free_bitmap_index(bitmap_git); | |
534 | return NULL; | |
535 | } | |
536 | ||
bfbb60d3 | 537 | struct bitmap_index *prepare_midx_bitmap_git(struct multi_pack_index *midx) |
0f533c72 TB |
538 | { |
539 | struct bitmap_index *bitmap_git = xcalloc(1, sizeof(*bitmap_git)); | |
540 | ||
541 | if (!open_midx_bitmap_1(bitmap_git, midx) && !load_bitmap(bitmap_git)) | |
3ae5fa07 | 542 | return bitmap_git; |
fff42755 | 543 | |
f3c23db2 | 544 | free_bitmap_index(bitmap_git); |
3ae5fa07 | 545 | return NULL; |
fff42755 VM |
546 | } |
547 | ||
548 | struct include_data { | |
3ae5fa07 | 549 | struct bitmap_index *bitmap_git; |
fff42755 VM |
550 | struct bitmap *base; |
551 | struct bitmap *seen; | |
552 | }; | |
553 | ||
98c31f36 TB |
554 | struct ewah_bitmap *bitmap_for_commit(struct bitmap_index *bitmap_git, |
555 | struct commit *commit) | |
556 | { | |
557 | khiter_t hash_pos = kh_get_oid_map(bitmap_git->bitmaps, | |
558 | commit->object.oid); | |
559 | if (hash_pos >= kh_end(bitmap_git->bitmaps)) | |
560 | return NULL; | |
561 | return lookup_stored_bitmap(kh_value(bitmap_git->bitmaps, hash_pos)); | |
562 | } | |
563 | ||
3ae5fa07 | 564 | static inline int bitmap_position_extended(struct bitmap_index *bitmap_git, |
3c771448 | 565 | const struct object_id *oid) |
fff42755 | 566 | { |
4ed43d16 | 567 | kh_oid_pos_t *positions = bitmap_git->ext_index.positions; |
3c771448 | 568 | khiter_t pos = kh_get_oid_pos(positions, *oid); |
fff42755 VM |
569 | |
570 | if (pos < kh_end(positions)) { | |
571 | int bitmap_pos = kh_value(positions, pos); | |
ed184620 | 572 | return bitmap_pos + bitmap_num_objects(bitmap_git); |
fff42755 VM |
573 | } |
574 | ||
575 | return -1; | |
576 | } | |
577 | ||
3ae5fa07 | 578 | static inline int bitmap_position_packfile(struct bitmap_index *bitmap_git, |
3c771448 | 579 | const struct object_id *oid) |
fff42755 | 580 | { |
57665086 | 581 | uint32_t pos; |
3c771448 | 582 | off_t offset = find_pack_entry_one(oid->hash, bitmap_git->pack); |
fff42755 VM |
583 | if (!offset) |
584 | return -1; | |
585 | ||
57665086 TB |
586 | if (offset_to_pack_pos(bitmap_git->pack, offset, &pos) < 0) |
587 | return -1; | |
588 | return pos; | |
fff42755 VM |
589 | } |
590 | ||
0f533c72 TB |
591 | static int bitmap_position_midx(struct bitmap_index *bitmap_git, |
592 | const struct object_id *oid) | |
593 | { | |
594 | uint32_t want, got; | |
595 | if (!bsearch_midx(oid, bitmap_git->midx, &want)) | |
596 | return -1; | |
597 | ||
598 | if (midx_to_pack_pos(bitmap_git->midx, want, &got) < 0) | |
599 | return -1; | |
600 | return got; | |
601 | } | |
602 | ||
3ae5fa07 | 603 | static int bitmap_position(struct bitmap_index *bitmap_git, |
3c771448 | 604 | const struct object_id *oid) |
fff42755 | 605 | { |
0f533c72 TB |
606 | int pos; |
607 | if (bitmap_is_midx(bitmap_git)) | |
608 | pos = bitmap_position_midx(bitmap_git, oid); | |
609 | else | |
610 | pos = bitmap_position_packfile(bitmap_git, oid); | |
3c771448 | 611 | return (pos >= 0) ? pos : bitmap_position_extended(bitmap_git, oid); |
fff42755 VM |
612 | } |
613 | ||
3ae5fa07 JT |
614 | static int ext_index_add_object(struct bitmap_index *bitmap_git, |
615 | struct object *object, const char *name) | |
fff42755 | 616 | { |
3ae5fa07 | 617 | struct eindex *eindex = &bitmap_git->ext_index; |
fff42755 VM |
618 | |
619 | khiter_t hash_pos; | |
620 | int hash_ret; | |
621 | int bitmap_pos; | |
622 | ||
3c771448 | 623 | hash_pos = kh_put_oid_pos(eindex->positions, object->oid, &hash_ret); |
fff42755 VM |
624 | if (hash_ret > 0) { |
625 | if (eindex->count >= eindex->alloc) { | |
626 | eindex->alloc = (eindex->alloc + 16) * 3 / 2; | |
2756ca43 RS |
627 | REALLOC_ARRAY(eindex->objects, eindex->alloc); |
628 | REALLOC_ARRAY(eindex->hashes, eindex->alloc); | |
fff42755 VM |
629 | } |
630 | ||
631 | bitmap_pos = eindex->count; | |
632 | eindex->objects[eindex->count] = object; | |
633 | eindex->hashes[eindex->count] = pack_name_hash(name); | |
634 | kh_value(eindex->positions, hash_pos) = bitmap_pos; | |
635 | eindex->count++; | |
636 | } else { | |
637 | bitmap_pos = kh_value(eindex->positions, hash_pos); | |
638 | } | |
639 | ||
ed184620 | 640 | return bitmap_pos + bitmap_num_objects(bitmap_git); |
fff42755 VM |
641 | } |
642 | ||
3ae5fa07 JT |
643 | struct bitmap_show_data { |
644 | struct bitmap_index *bitmap_git; | |
645 | struct bitmap *base; | |
646 | }; | |
647 | ||
648 | static void show_object(struct object *object, const char *name, void *data_) | |
fff42755 | 649 | { |
3ae5fa07 | 650 | struct bitmap_show_data *data = data_; |
fff42755 VM |
651 | int bitmap_pos; |
652 | ||
3c771448 | 653 | bitmap_pos = bitmap_position(data->bitmap_git, &object->oid); |
fff42755 | 654 | |
de1e67d0 | 655 | if (bitmap_pos < 0) |
3ae5fa07 JT |
656 | bitmap_pos = ext_index_add_object(data->bitmap_git, object, |
657 | name); | |
fff42755 | 658 | |
3ae5fa07 | 659 | bitmap_set(data->base, bitmap_pos); |
fff42755 VM |
660 | } |
661 | ||
662 | static void show_commit(struct commit *commit, void *data) | |
663 | { | |
664 | } | |
665 | ||
3ae5fa07 JT |
666 | static int add_to_include_set(struct bitmap_index *bitmap_git, |
667 | struct include_data *data, | |
98c31f36 | 668 | struct commit *commit, |
fff42755 VM |
669 | int bitmap_pos) |
670 | { | |
98c31f36 | 671 | struct ewah_bitmap *partial; |
fff42755 VM |
672 | |
673 | if (data->seen && bitmap_get(data->seen, bitmap_pos)) | |
674 | return 0; | |
675 | ||
676 | if (bitmap_get(data->base, bitmap_pos)) | |
677 | return 0; | |
678 | ||
98c31f36 TB |
679 | partial = bitmap_for_commit(bitmap_git, commit); |
680 | if (partial) { | |
681 | bitmap_or_ewah(data->base, partial); | |
fff42755 VM |
682 | return 0; |
683 | } | |
684 | ||
685 | bitmap_set(data->base, bitmap_pos); | |
686 | return 1; | |
687 | } | |
688 | ||
689 | static int should_include(struct commit *commit, void *_data) | |
690 | { | |
691 | struct include_data *data = _data; | |
692 | int bitmap_pos; | |
693 | ||
3c771448 | 694 | bitmap_pos = bitmap_position(data->bitmap_git, &commit->object.oid); |
fff42755 | 695 | if (bitmap_pos < 0) |
3ae5fa07 JT |
696 | bitmap_pos = ext_index_add_object(data->bitmap_git, |
697 | (struct object *)commit, | |
698 | NULL); | |
fff42755 | 699 | |
98c31f36 | 700 | if (!add_to_include_set(data->bitmap_git, data, commit, bitmap_pos)) { |
fff42755 VM |
701 | struct commit_list *parent = commit->parents; |
702 | ||
703 | while (parent) { | |
704 | parent->item->object.flags |= SEEN; | |
705 | parent = parent->next; | |
706 | } | |
707 | ||
708 | return 0; | |
709 | } | |
710 | ||
711 | return 1; | |
712 | } | |
713 | ||
aa9ad6fe JK |
714 | static int should_include_obj(struct object *obj, void *_data) |
715 | { | |
716 | struct include_data *data = _data; | |
717 | int bitmap_pos; | |
718 | ||
719 | bitmap_pos = bitmap_position(data->bitmap_git, &obj->oid); | |
720 | if (bitmap_pos < 0) | |
721 | return 1; | |
722 | if ((data->seen && bitmap_get(data->seen, bitmap_pos)) || | |
723 | bitmap_get(data->base, bitmap_pos)) { | |
724 | obj->flags |= SEEN; | |
725 | return 0; | |
726 | } | |
727 | return 1; | |
728 | } | |
729 | ||
83578051 TB |
730 | static int add_commit_to_bitmap(struct bitmap_index *bitmap_git, |
731 | struct bitmap **base, | |
732 | struct commit *commit) | |
733 | { | |
734 | struct ewah_bitmap *or_with = bitmap_for_commit(bitmap_git, commit); | |
735 | ||
736 | if (!or_with) | |
737 | return 0; | |
738 | ||
739 | if (*base == NULL) | |
740 | *base = ewah_to_bitmap(or_with); | |
741 | else | |
742 | bitmap_or_ewah(*base, or_with); | |
743 | ||
744 | return 1; | |
745 | } | |
746 | ||
3ae5fa07 JT |
747 | static struct bitmap *find_objects(struct bitmap_index *bitmap_git, |
748 | struct rev_info *revs, | |
fff42755 | 749 | struct object_list *roots, |
09d4a79e | 750 | struct bitmap *seen) |
fff42755 VM |
751 | { |
752 | struct bitmap *base = NULL; | |
753 | int needs_walk = 0; | |
754 | ||
755 | struct object_list *not_mapped = NULL; | |
756 | ||
757 | /* | |
758 | * Go through all the roots for the walk. The ones that have bitmaps | |
759 | * on the bitmap index will be `or`ed together to form an initial | |
760 | * global reachability analysis. | |
761 | * | |
762 | * The ones without bitmaps in the index will be stored in the | |
763 | * `not_mapped_list` for further processing. | |
764 | */ | |
765 | while (roots) { | |
766 | struct object *object = roots->item; | |
767 | roots = roots->next; | |
768 | ||
83578051 TB |
769 | if (object->type == OBJ_COMMIT && |
770 | add_commit_to_bitmap(bitmap_git, &base, (struct commit *)object)) { | |
771 | object->flags |= SEEN; | |
772 | continue; | |
fff42755 VM |
773 | } |
774 | ||
775 | object_list_insert(object, ¬_mapped); | |
776 | } | |
777 | ||
778 | /* | |
779 | * Best case scenario: We found bitmaps for all the roots, | |
780 | * so the resulting `or` bitmap has the full reachability analysis | |
781 | */ | |
782 | if (not_mapped == NULL) | |
783 | return base; | |
784 | ||
785 | roots = not_mapped; | |
786 | ||
787 | /* | |
788 | * Let's iterate through all the roots that don't have bitmaps to | |
789 | * check if we can determine them to be reachable from the existing | |
790 | * global bitmap. | |
791 | * | |
792 | * If we cannot find them in the existing global bitmap, we'll need | |
793 | * to push them to an actual walk and run it until we can confirm | |
794 | * they are reachable | |
795 | */ | |
796 | while (roots) { | |
797 | struct object *object = roots->item; | |
798 | int pos; | |
799 | ||
800 | roots = roots->next; | |
3c771448 | 801 | pos = bitmap_position(bitmap_git, &object->oid); |
fff42755 VM |
802 | |
803 | if (pos < 0 || base == NULL || !bitmap_get(base, pos)) { | |
804 | object->flags &= ~UNINTERESTING; | |
805 | add_pending_object(revs, object, ""); | |
806 | needs_walk = 1; | |
807 | } else { | |
808 | object->flags |= SEEN; | |
809 | } | |
810 | } | |
811 | ||
812 | if (needs_walk) { | |
813 | struct include_data incdata; | |
3ae5fa07 | 814 | struct bitmap_show_data show_data; |
fff42755 VM |
815 | |
816 | if (base == NULL) | |
817 | base = bitmap_new(); | |
818 | ||
3ae5fa07 | 819 | incdata.bitmap_git = bitmap_git; |
fff42755 VM |
820 | incdata.base = base; |
821 | incdata.seen = seen; | |
822 | ||
823 | revs->include_check = should_include; | |
aa9ad6fe | 824 | revs->include_check_obj = should_include_obj; |
fff42755 VM |
825 | revs->include_check_data = &incdata; |
826 | ||
827 | if (prepare_revision_walk(revs)) | |
9975975d | 828 | die(_("revision walk setup failed")); |
fff42755 | 829 | |
3ae5fa07 JT |
830 | show_data.bitmap_git = bitmap_git; |
831 | show_data.base = base; | |
832 | ||
3e0370a8 DS |
833 | traverse_commit_list(revs, |
834 | show_commit, show_object, | |
835 | &show_data); | |
1e951c64 JK |
836 | |
837 | revs->include_check = NULL; | |
aa9ad6fe | 838 | revs->include_check_obj = NULL; |
1e951c64 | 839 | revs->include_check_data = NULL; |
fff42755 VM |
840 | } |
841 | ||
842 | return base; | |
843 | } | |
844 | ||
3ae5fa07 | 845 | static void show_extended_objects(struct bitmap_index *bitmap_git, |
4eb707eb | 846 | struct rev_info *revs, |
fff42755 VM |
847 | show_reachable_fn show_reach) |
848 | { | |
3ae5fa07 JT |
849 | struct bitmap *objects = bitmap_git->result; |
850 | struct eindex *eindex = &bitmap_git->ext_index; | |
fff42755 VM |
851 | uint32_t i; |
852 | ||
853 | for (i = 0; i < eindex->count; ++i) { | |
854 | struct object *obj; | |
855 | ||
ed184620 | 856 | if (!bitmap_get(objects, bitmap_num_objects(bitmap_git) + i)) |
fff42755 VM |
857 | continue; |
858 | ||
859 | obj = eindex->objects[i]; | |
4eb707eb JK |
860 | if ((obj->type == OBJ_BLOB && !revs->blob_objects) || |
861 | (obj->type == OBJ_TREE && !revs->tree_objects) || | |
862 | (obj->type == OBJ_TAG && !revs->tag_objects)) | |
863 | continue; | |
864 | ||
20664967 | 865 | show_reach(&obj->oid, obj->type, 0, eindex->hashes[i], NULL, 0); |
fff42755 VM |
866 | } |
867 | } | |
868 | ||
551cf8b6 JK |
869 | static void init_type_iterator(struct ewah_iterator *it, |
870 | struct bitmap_index *bitmap_git, | |
871 | enum object_type type) | |
872 | { | |
873 | switch (type) { | |
874 | case OBJ_COMMIT: | |
875 | ewah_iterator_init(it, bitmap_git->commits); | |
876 | break; | |
877 | ||
878 | case OBJ_TREE: | |
879 | ewah_iterator_init(it, bitmap_git->trees); | |
880 | break; | |
881 | ||
882 | case OBJ_BLOB: | |
883 | ewah_iterator_init(it, bitmap_git->blobs); | |
884 | break; | |
885 | ||
886 | case OBJ_TAG: | |
887 | ewah_iterator_init(it, bitmap_git->tags); | |
888 | break; | |
889 | ||
890 | default: | |
891 | BUG("object type %d not stored by bitmap type index", type); | |
892 | break; | |
893 | } | |
894 | } | |
895 | ||
fff42755 | 896 | static void show_objects_for_type( |
3ae5fa07 | 897 | struct bitmap_index *bitmap_git, |
fff42755 VM |
898 | enum object_type object_type, |
899 | show_reachable_fn show_reach) | |
900 | { | |
d2ea0310 | 901 | size_t i = 0; |
fff42755 VM |
902 | uint32_t offset; |
903 | ||
904 | struct ewah_iterator it; | |
905 | eword_t filter; | |
906 | ||
3ae5fa07 JT |
907 | struct bitmap *objects = bitmap_git->result; |
908 | ||
551cf8b6 | 909 | init_type_iterator(&it, bitmap_git, object_type); |
fff42755 | 910 | |
d2ea0310 JK |
911 | for (i = 0; i < objects->word_alloc && |
912 | ewah_iterator_next(&filter, &it); i++) { | |
fff42755 | 913 | eword_t word = objects->words[i] & filter; |
d2ea0310 JK |
914 | size_t pos = (i * BITS_IN_EWORD); |
915 | ||
916 | if (!word) | |
917 | continue; | |
fff42755 | 918 | |
34b935c0 | 919 | for (offset = 0; offset < BITS_IN_EWORD; ++offset) { |
0f533c72 | 920 | struct packed_git *pack; |
20664967 | 921 | struct object_id oid; |
cf98f2e8 TB |
922 | uint32_t hash = 0, index_pos; |
923 | off_t ofs; | |
fff42755 VM |
924 | |
925 | if ((word >> offset) == 0) | |
926 | break; | |
927 | ||
928 | offset += ewah_bit_ctz64(word >> offset); | |
929 | ||
0f533c72 TB |
930 | if (bitmap_is_midx(bitmap_git)) { |
931 | struct multi_pack_index *m = bitmap_git->midx; | |
932 | uint32_t pack_id; | |
933 | ||
934 | index_pos = pack_pos_to_midx(m, pos + offset); | |
935 | ofs = nth_midxed_offset(m, index_pos); | |
936 | nth_midxed_object_oid(&oid, m, index_pos); | |
937 | ||
938 | pack_id = nth_midxed_pack_int_id(m, index_pos); | |
939 | pack = bitmap_git->midx->packs[pack_id]; | |
940 | } else { | |
941 | index_pos = pack_pos_to_index(bitmap_git->pack, pos + offset); | |
942 | ofs = pack_pos_to_offset(bitmap_git->pack, pos + offset); | |
943 | nth_bitmap_object_oid(bitmap_git, &oid, index_pos); | |
944 | ||
945 | pack = bitmap_git->pack; | |
946 | } | |
fff42755 | 947 | |
3ae5fa07 | 948 | if (bitmap_git->hashes) |
cf98f2e8 | 949 | hash = get_be32(bitmap_git->hashes + index_pos); |
ae4f07fb | 950 | |
0f533c72 | 951 | show_reach(&oid, object_type, 0, hash, pack, ofs); |
fff42755 | 952 | } |
fff42755 VM |
953 | } |
954 | } | |
955 | ||
3ae5fa07 JT |
956 | static int in_bitmapped_pack(struct bitmap_index *bitmap_git, |
957 | struct object_list *roots) | |
fff42755 VM |
958 | { |
959 | while (roots) { | |
960 | struct object *object = roots->item; | |
961 | roots = roots->next; | |
962 | ||
0f533c72 TB |
963 | if (bitmap_is_midx(bitmap_git)) { |
964 | if (bsearch_midx(&object->oid, bitmap_git->midx, NULL)) | |
965 | return 1; | |
966 | } else { | |
967 | if (find_pack_entry_one(object->oid.hash, bitmap_git->pack) > 0) | |
968 | return 1; | |
969 | } | |
fff42755 VM |
970 | } |
971 | ||
972 | return 0; | |
973 | } | |
974 | ||
856e12c1 TB |
975 | static struct bitmap *find_tip_objects(struct bitmap_index *bitmap_git, |
976 | struct object_list *tip_objects, | |
977 | enum object_type type) | |
4f3bd560 JK |
978 | { |
979 | struct bitmap *result = bitmap_new(); | |
980 | struct object_list *p; | |
981 | ||
982 | for (p = tip_objects; p; p = p->next) { | |
983 | int pos; | |
984 | ||
856e12c1 | 985 | if (p->item->type != type) |
4f3bd560 JK |
986 | continue; |
987 | ||
988 | pos = bitmap_position(bitmap_git, &p->item->oid); | |
989 | if (pos < 0) | |
990 | continue; | |
991 | ||
992 | bitmap_set(result, pos); | |
993 | } | |
994 | ||
995 | return result; | |
996 | } | |
997 | ||
856e12c1 TB |
998 | static void filter_bitmap_exclude_type(struct bitmap_index *bitmap_git, |
999 | struct object_list *tip_objects, | |
1000 | struct bitmap *to_filter, | |
1001 | enum object_type type) | |
4f3bd560 JK |
1002 | { |
1003 | struct eindex *eindex = &bitmap_git->ext_index; | |
1004 | struct bitmap *tips; | |
1005 | struct ewah_iterator it; | |
1006 | eword_t mask; | |
1007 | uint32_t i; | |
1008 | ||
1009 | /* | |
1010 | * The non-bitmap version of this filter never removes | |
856e12c1 | 1011 | * objects which the other side specifically asked for, |
4f3bd560 JK |
1012 | * so we must match that behavior. |
1013 | */ | |
856e12c1 | 1014 | tips = find_tip_objects(bitmap_git, tip_objects, type); |
4f3bd560 JK |
1015 | |
1016 | /* | |
ddcb189d TB |
1017 | * We can use the type-level bitmap for 'type' to work in whole |
1018 | * words for the objects that are actually in the bitmapped | |
1019 | * packfile. | |
4f3bd560 | 1020 | */ |
856e12c1 | 1021 | for (i = 0, init_type_iterator(&it, bitmap_git, type); |
4f3bd560 JK |
1022 | i < to_filter->word_alloc && ewah_iterator_next(&mask, &it); |
1023 | i++) { | |
1024 | if (i < tips->word_alloc) | |
1025 | mask &= ~tips->words[i]; | |
1026 | to_filter->words[i] &= ~mask; | |
1027 | } | |
1028 | ||
1029 | /* | |
ddcb189d TB |
1030 | * Clear any objects that weren't in the packfile (and so would |
1031 | * not have been caught by the loop above. We'll have to check | |
1032 | * them individually. | |
4f3bd560 JK |
1033 | */ |
1034 | for (i = 0; i < eindex->count; i++) { | |
ed184620 | 1035 | uint32_t pos = i + bitmap_num_objects(bitmap_git); |
856e12c1 | 1036 | if (eindex->objects[i]->type == type && |
4f3bd560 JK |
1037 | bitmap_get(to_filter, pos) && |
1038 | !bitmap_get(tips, pos)) | |
1039 | bitmap_unset(to_filter, pos); | |
1040 | } | |
1041 | ||
1042 | bitmap_free(tips); | |
1043 | } | |
1044 | ||
856e12c1 TB |
1045 | static void filter_bitmap_blob_none(struct bitmap_index *bitmap_git, |
1046 | struct object_list *tip_objects, | |
1047 | struct bitmap *to_filter) | |
1048 | { | |
1049 | filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, | |
1050 | OBJ_BLOB); | |
1051 | } | |
1052 | ||
84243da1 JK |
1053 | static unsigned long get_size_by_pos(struct bitmap_index *bitmap_git, |
1054 | uint32_t pos) | |
1055 | { | |
84243da1 JK |
1056 | unsigned long size; |
1057 | struct object_info oi = OBJECT_INFO_INIT; | |
1058 | ||
1059 | oi.sizep = &size; | |
1060 | ||
ed184620 | 1061 | if (pos < bitmap_num_objects(bitmap_git)) { |
0f533c72 TB |
1062 | struct packed_git *pack; |
1063 | off_t ofs; | |
1064 | ||
1065 | if (bitmap_is_midx(bitmap_git)) { | |
1066 | uint32_t midx_pos = pack_pos_to_midx(bitmap_git->midx, pos); | |
1067 | uint32_t pack_id = nth_midxed_pack_int_id(bitmap_git->midx, midx_pos); | |
1068 | ||
1069 | pack = bitmap_git->midx->packs[pack_id]; | |
1070 | ofs = nth_midxed_offset(bitmap_git->midx, midx_pos); | |
1071 | } else { | |
1072 | pack = bitmap_git->pack; | |
1073 | ofs = pack_pos_to_offset(pack, pos); | |
1074 | } | |
1075 | ||
a78a9032 | 1076 | if (packed_object_info(the_repository, pack, ofs, &oi) < 0) { |
84243da1 | 1077 | struct object_id oid; |
6b4277e6 TB |
1078 | nth_bitmap_object_oid(bitmap_git, &oid, |
1079 | pack_pos_to_index(pack, pos)); | |
84243da1 JK |
1080 | die(_("unable to get size of %s"), oid_to_hex(&oid)); |
1081 | } | |
1082 | } else { | |
1083 | struct eindex *eindex = &bitmap_git->ext_index; | |
ed184620 | 1084 | struct object *obj = eindex->objects[pos - bitmap_num_objects(bitmap_git)]; |
84243da1 JK |
1085 | if (oid_object_info_extended(the_repository, &obj->oid, &oi, 0) < 0) |
1086 | die(_("unable to get size of %s"), oid_to_hex(&obj->oid)); | |
1087 | } | |
1088 | ||
1089 | return size; | |
1090 | } | |
1091 | ||
1092 | static void filter_bitmap_blob_limit(struct bitmap_index *bitmap_git, | |
1093 | struct object_list *tip_objects, | |
1094 | struct bitmap *to_filter, | |
1095 | unsigned long limit) | |
1096 | { | |
1097 | struct eindex *eindex = &bitmap_git->ext_index; | |
1098 | struct bitmap *tips; | |
1099 | struct ewah_iterator it; | |
1100 | eword_t mask; | |
1101 | uint32_t i; | |
1102 | ||
856e12c1 | 1103 | tips = find_tip_objects(bitmap_git, tip_objects, OBJ_BLOB); |
84243da1 JK |
1104 | |
1105 | for (i = 0, init_type_iterator(&it, bitmap_git, OBJ_BLOB); | |
1106 | i < to_filter->word_alloc && ewah_iterator_next(&mask, &it); | |
1107 | i++) { | |
1108 | eword_t word = to_filter->words[i] & mask; | |
1109 | unsigned offset; | |
1110 | ||
1111 | for (offset = 0; offset < BITS_IN_EWORD; offset++) { | |
1112 | uint32_t pos; | |
1113 | ||
1114 | if ((word >> offset) == 0) | |
1115 | break; | |
1116 | offset += ewah_bit_ctz64(word >> offset); | |
1117 | pos = i * BITS_IN_EWORD + offset; | |
1118 | ||
1119 | if (!bitmap_get(tips, pos) && | |
1120 | get_size_by_pos(bitmap_git, pos) >= limit) | |
1121 | bitmap_unset(to_filter, pos); | |
1122 | } | |
1123 | } | |
1124 | ||
1125 | for (i = 0; i < eindex->count; i++) { | |
ed184620 | 1126 | uint32_t pos = i + bitmap_num_objects(bitmap_git); |
84243da1 JK |
1127 | if (eindex->objects[i]->type == OBJ_BLOB && |
1128 | bitmap_get(to_filter, pos) && | |
1129 | !bitmap_get(tips, pos) && | |
1130 | get_size_by_pos(bitmap_git, pos) >= limit) | |
1131 | bitmap_unset(to_filter, pos); | |
1132 | } | |
1133 | ||
1134 | bitmap_free(tips); | |
1135 | } | |
1136 | ||
b0a8d482 TB |
1137 | static void filter_bitmap_tree_depth(struct bitmap_index *bitmap_git, |
1138 | struct object_list *tip_objects, | |
1139 | struct bitmap *to_filter, | |
1140 | unsigned long limit) | |
1141 | { | |
1142 | if (limit) | |
1143 | BUG("filter_bitmap_tree_depth given non-zero limit"); | |
1144 | ||
1145 | filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, | |
1146 | OBJ_TREE); | |
1147 | filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, | |
1148 | OBJ_BLOB); | |
1149 | } | |
1150 | ||
7ab6aafa PS |
1151 | static void filter_bitmap_object_type(struct bitmap_index *bitmap_git, |
1152 | struct object_list *tip_objects, | |
1153 | struct bitmap *to_filter, | |
1154 | enum object_type object_type) | |
1155 | { | |
1156 | if (object_type < OBJ_COMMIT || object_type > OBJ_TAG) | |
1157 | BUG("filter_bitmap_object_type given invalid object"); | |
1158 | ||
1159 | if (object_type != OBJ_TAG) | |
1160 | filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, OBJ_TAG); | |
1161 | if (object_type != OBJ_COMMIT) | |
1162 | filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, OBJ_COMMIT); | |
1163 | if (object_type != OBJ_TREE) | |
1164 | filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, OBJ_TREE); | |
1165 | if (object_type != OBJ_BLOB) | |
1166 | filter_bitmap_exclude_type(bitmap_git, tip_objects, to_filter, OBJ_BLOB); | |
1167 | } | |
1168 | ||
6663ae0a JK |
1169 | static int filter_bitmap(struct bitmap_index *bitmap_git, |
1170 | struct object_list *tip_objects, | |
1171 | struct bitmap *to_filter, | |
1172 | struct list_objects_filter_options *filter) | |
1173 | { | |
1174 | if (!filter || filter->choice == LOFC_DISABLED) | |
1175 | return 0; | |
1176 | ||
4f3bd560 JK |
1177 | if (filter->choice == LOFC_BLOB_NONE) { |
1178 | if (bitmap_git) | |
1179 | filter_bitmap_blob_none(bitmap_git, tip_objects, | |
1180 | to_filter); | |
1181 | return 0; | |
1182 | } | |
1183 | ||
84243da1 JK |
1184 | if (filter->choice == LOFC_BLOB_LIMIT) { |
1185 | if (bitmap_git) | |
1186 | filter_bitmap_blob_limit(bitmap_git, tip_objects, | |
1187 | to_filter, | |
1188 | filter->blob_limit_value); | |
1189 | return 0; | |
1190 | } | |
1191 | ||
b0a8d482 TB |
1192 | if (filter->choice == LOFC_TREE_DEPTH && |
1193 | filter->tree_exclude_depth == 0) { | |
1194 | if (bitmap_git) | |
1195 | filter_bitmap_tree_depth(bitmap_git, tip_objects, | |
1196 | to_filter, | |
1197 | filter->tree_exclude_depth); | |
1198 | return 0; | |
1199 | } | |
1200 | ||
7ab6aafa PS |
1201 | if (filter->choice == LOFC_OBJECT_TYPE) { |
1202 | if (bitmap_git) | |
1203 | filter_bitmap_object_type(bitmap_git, tip_objects, | |
1204 | to_filter, | |
1205 | filter->object_type); | |
1206 | return 0; | |
1207 | } | |
1208 | ||
169a15eb PS |
1209 | if (filter->choice == LOFC_COMBINE) { |
1210 | int i; | |
1211 | for (i = 0; i < filter->sub_nr; i++) { | |
1212 | if (filter_bitmap(bitmap_git, tip_objects, to_filter, | |
1213 | &filter->sub[i]) < 0) | |
1214 | return -1; | |
1215 | } | |
1216 | return 0; | |
1217 | } | |
1218 | ||
6663ae0a JK |
1219 | /* filter choice not handled */ |
1220 | return -1; | |
1221 | } | |
1222 | ||
1223 | static int can_filter_bitmap(struct list_objects_filter_options *filter) | |
1224 | { | |
1225 | return !filter_bitmap(NULL, NULL, NULL, filter); | |
1226 | } | |
1227 | ||
1228 | struct bitmap_index *prepare_bitmap_walk(struct rev_info *revs, | |
9cf68b27 | 1229 | int filter_provided_objects) |
fff42755 VM |
1230 | { |
1231 | unsigned int i; | |
fff42755 VM |
1232 | |
1233 | struct object_list *wants = NULL; | |
1234 | struct object_list *haves = NULL; | |
1235 | ||
1236 | struct bitmap *wants_bitmap = NULL; | |
1237 | struct bitmap *haves_bitmap = NULL; | |
1238 | ||
d90fe06e JK |
1239 | struct bitmap_index *bitmap_git; |
1240 | ||
1241 | /* | |
1242 | * We can't do pathspec limiting with bitmaps, because we don't know | |
1243 | * which commits are associated with which object changes (let alone | |
1244 | * even which objects are associated with which paths). | |
1245 | */ | |
1246 | if (revs->prune) | |
1247 | return NULL; | |
1248 | ||
09d4a79e | 1249 | if (!can_filter_bitmap(&revs->filter)) |
6663ae0a JK |
1250 | return NULL; |
1251 | ||
3ae5fa07 JT |
1252 | /* try to open a bitmapped pack, but don't parse it yet |
1253 | * because we may not need to use it */ | |
ca56dadb | 1254 | CALLOC_ARRAY(bitmap_git, 1); |
0f533c72 | 1255 | if (open_bitmap(revs->repo, bitmap_git) < 0) |
f3c23db2 | 1256 | goto cleanup; |
fff42755 | 1257 | |
4d01a7fa MÅ |
1258 | for (i = 0; i < revs->pending.nr; ++i) { |
1259 | struct object *object = revs->pending.objects[i].item; | |
fff42755 VM |
1260 | |
1261 | if (object->type == OBJ_NONE) | |
c251c83d | 1262 | parse_object_or_die(&object->oid, NULL); |
fff42755 VM |
1263 | |
1264 | while (object->type == OBJ_TAG) { | |
1265 | struct tag *tag = (struct tag *) object; | |
1266 | ||
1267 | if (object->flags & UNINTERESTING) | |
1268 | object_list_insert(object, &haves); | |
1269 | else | |
1270 | object_list_insert(object, &wants); | |
1271 | ||
dad3f060 | 1272 | object = parse_object_or_die(get_tagged_oid(tag), NULL); |
540cdc11 | 1273 | object->flags |= (tag->object.flags & UNINTERESTING); |
fff42755 VM |
1274 | } |
1275 | ||
1276 | if (object->flags & UNINTERESTING) | |
1277 | object_list_insert(object, &haves); | |
1278 | else | |
1279 | object_list_insert(object, &wants); | |
1280 | } | |
1281 | ||
1282 | /* | |
1283 | * if we have a HAVES list, but none of those haves is contained | |
1284 | * in the packfile that has a bitmap, we don't have anything to | |
1285 | * optimize here | |
1286 | */ | |
3ae5fa07 | 1287 | if (haves && !in_bitmapped_pack(bitmap_git, haves)) |
f3c23db2 | 1288 | goto cleanup; |
fff42755 VM |
1289 | |
1290 | /* if we don't want anything, we're done here */ | |
1291 | if (!wants) | |
f3c23db2 | 1292 | goto cleanup; |
fff42755 VM |
1293 | |
1294 | /* | |
1295 | * now we're going to use bitmaps, so load the actual bitmap entries | |
1296 | * from disk. this is the point of no return; after this the rev_list | |
1297 | * becomes invalidated and we must perform the revwalk through bitmaps | |
1298 | */ | |
0f533c72 | 1299 | if (load_bitmap(bitmap_git) < 0) |
f3c23db2 | 1300 | goto cleanup; |
fff42755 | 1301 | |
4d01a7fa | 1302 | object_array_clear(&revs->pending); |
fff42755 VM |
1303 | |
1304 | if (haves) { | |
2db1a43f | 1305 | revs->ignore_missing_links = 1; |
09d4a79e | 1306 | haves_bitmap = find_objects(bitmap_git, revs, haves, NULL); |
fff42755 | 1307 | reset_revision_walk(); |
2db1a43f | 1308 | revs->ignore_missing_links = 0; |
fff42755 VM |
1309 | |
1310 | if (haves_bitmap == NULL) | |
033abf97 | 1311 | BUG("failed to perform bitmap walk"); |
fff42755 VM |
1312 | } |
1313 | ||
09d4a79e | 1314 | wants_bitmap = find_objects(bitmap_git, revs, wants, haves_bitmap); |
fff42755 VM |
1315 | |
1316 | if (!wants_bitmap) | |
033abf97 | 1317 | BUG("failed to perform bitmap walk"); |
fff42755 VM |
1318 | |
1319 | if (haves_bitmap) | |
1320 | bitmap_and_not(wants_bitmap, haves_bitmap); | |
1321 | ||
09d4a79e DS |
1322 | filter_bitmap(bitmap_git, |
1323 | (revs->filter.choice && filter_provided_objects) ? NULL : wants, | |
1324 | wants_bitmap, | |
1325 | &revs->filter); | |
6663ae0a | 1326 | |
3ae5fa07 | 1327 | bitmap_git->result = wants_bitmap; |
30cdc33f | 1328 | bitmap_git->haves = haves_bitmap; |
fff42755 | 1329 | |
acac50dd JK |
1330 | object_list_free(&wants); |
1331 | object_list_free(&haves); | |
1332 | ||
3ae5fa07 | 1333 | return bitmap_git; |
f3c23db2 JT |
1334 | |
1335 | cleanup: | |
1336 | free_bitmap_index(bitmap_git); | |
acac50dd JK |
1337 | object_list_free(&wants); |
1338 | object_list_free(&haves); | |
f3c23db2 | 1339 | return NULL; |
fff42755 VM |
1340 | } |
1341 | ||
a5f9f24a TB |
1342 | /* |
1343 | * -1 means "stop trying further objects"; 0 means we may or may not have | |
1344 | * reused, but you can keep feeding bits. | |
1345 | */ | |
73cd7d94 | 1346 | static int try_partial_reuse(struct packed_git *pack, |
a5f9f24a TB |
1347 | size_t pos, |
1348 | struct bitmap *reuse, | |
1349 | struct pack_window **w_curs) | |
fff42755 | 1350 | { |
0f533c72 | 1351 | off_t offset, delta_obj_offset; |
bb514de3 JK |
1352 | enum object_type type; |
1353 | unsigned long size; | |
1354 | ||
0f533c72 TB |
1355 | /* |
1356 | * try_partial_reuse() is called either on (a) objects in the | |
1357 | * bitmapped pack (in the case of a single-pack bitmap) or (b) | |
1358 | * objects in the preferred pack of a multi-pack bitmap. | |
1359 | * Importantly, the latter can pretend as if only a single pack | |
1360 | * exists because: | |
1361 | * | |
1362 | * - The first pack->num_objects bits of a MIDX bitmap are | |
1363 | * reserved for the preferred pack, and | |
1364 | * | |
1365 | * - Ties due to duplicate objects are always resolved in | |
1366 | * favor of the preferred pack. | |
1367 | * | |
1368 | * Therefore we do not need to ever ask the MIDX for its copy of | |
1369 | * an object by OID, since it will always select it from the | |
1370 | * preferred pack. Likewise, the selected copy of the base | |
1371 | * object for any deltas will reside in the same pack. | |
1372 | * | |
1373 | * This means that we can reuse pos when looking up the bit in | |
1374 | * the reuse bitmap, too, since bits corresponding to the | |
1375 | * preferred pack precede all bits from other packs. | |
1376 | */ | |
1377 | ||
1378 | if (pos >= pack->num_objects) | |
1379 | return -1; /* not actually in the pack or MIDX preferred pack */ | |
bb514de3 | 1380 | |
0f533c72 TB |
1381 | offset = delta_obj_offset = pack_pos_to_offset(pack, pos); |
1382 | type = unpack_object_header(pack, w_curs, &offset, &size); | |
bb514de3 | 1383 | if (type < 0) |
a5f9f24a | 1384 | return -1; /* broken packfile, punt */ |
bb514de3 JK |
1385 | |
1386 | if (type == OBJ_REF_DELTA || type == OBJ_OFS_DELTA) { | |
1387 | off_t base_offset; | |
011f3fd5 | 1388 | uint32_t base_pos; |
bb514de3 JK |
1389 | |
1390 | /* | |
1391 | * Find the position of the base object so we can look it up | |
1392 | * in our bitmaps. If we can't come up with an offset, or if | |
1393 | * that offset is not in the revidx, the pack is corrupt. | |
1394 | * There's nothing we can do, so just punt on this object, | |
1395 | * and the normal slow path will complain about it in | |
1396 | * more detail. | |
1397 | */ | |
0f533c72 TB |
1398 | base_offset = get_delta_base(pack, w_curs, &offset, type, |
1399 | delta_obj_offset); | |
bb514de3 | 1400 | if (!base_offset) |
a5f9f24a | 1401 | return 0; |
0f533c72 | 1402 | if (offset_to_pack_pos(pack, base_offset, &base_pos) < 0) |
a5f9f24a | 1403 | return 0; |
bb514de3 JK |
1404 | |
1405 | /* | |
1406 | * We assume delta dependencies always point backwards. This | |
1407 | * lets us do a single pass, and is basically always true | |
1408 | * due to the way OFS_DELTAs work. You would not typically | |
1409 | * find REF_DELTA in a bitmapped pack, since we only bitmap | |
1410 | * packs we write fresh, and OFS_DELTA is the default). But | |
1411 | * let's double check to make sure the pack wasn't written with | |
1412 | * odd parameters. | |
1413 | */ | |
1414 | if (base_pos >= pos) | |
a5f9f24a | 1415 | return 0; |
bb514de3 JK |
1416 | |
1417 | /* | |
1418 | * And finally, if we're not sending the base as part of our | |
1419 | * reuse chunk, then don't send this object either. The base | |
1420 | * would come after us, along with other objects not | |
1421 | * necessarily in the pack, which means we'd need to convert | |
1422 | * to REF_DELTA on the fly. Better to just let the normal | |
1423 | * object_entry code path handle it. | |
1424 | */ | |
1425 | if (!bitmap_get(reuse, base_pos)) | |
a5f9f24a | 1426 | return 0; |
bb514de3 JK |
1427 | } |
1428 | ||
fff42755 | 1429 | /* |
bb514de3 | 1430 | * If we got here, then the object is OK to reuse. Mark it. |
fff42755 | 1431 | */ |
bb514de3 | 1432 | bitmap_set(reuse, pos); |
a5f9f24a | 1433 | return 0; |
bb514de3 | 1434 | } |
fff42755 | 1435 | |
6d08b9d4 | 1436 | uint32_t midx_preferred_pack(struct bitmap_index *bitmap_git) |
0f533c72 TB |
1437 | { |
1438 | struct multi_pack_index *m = bitmap_git->midx; | |
1439 | if (!m) | |
1440 | BUG("midx_preferred_pack: requires non-empty MIDX"); | |
1441 | return nth_midxed_pack_int_id(m, pack_pos_to_midx(bitmap_git->midx, 0)); | |
1442 | } | |
1443 | ||
bb514de3 JK |
1444 | int reuse_partial_packfile_from_bitmap(struct bitmap_index *bitmap_git, |
1445 | struct packed_git **packfile_out, | |
1446 | uint32_t *entries, | |
1447 | struct bitmap **reuse_out) | |
1448 | { | |
0f533c72 | 1449 | struct packed_git *pack; |
3ae5fa07 | 1450 | struct bitmap *result = bitmap_git->result; |
bb514de3 JK |
1451 | struct bitmap *reuse; |
1452 | struct pack_window *w_curs = NULL; | |
1453 | size_t i = 0; | |
1454 | uint32_t offset; | |
0f533c72 | 1455 | uint32_t objects_nr; |
fff42755 VM |
1456 | |
1457 | assert(result); | |
1458 | ||
0f533c72 TB |
1459 | load_reverse_index(bitmap_git); |
1460 | ||
1461 | if (bitmap_is_midx(bitmap_git)) | |
1462 | pack = bitmap_git->midx->packs[midx_preferred_pack(bitmap_git)]; | |
1463 | else | |
1464 | pack = bitmap_git->pack; | |
1465 | objects_nr = pack->num_objects; | |
1466 | ||
bb514de3 JK |
1467 | while (i < result->word_alloc && result->words[i] == (eword_t)~0) |
1468 | i++; | |
fff42755 | 1469 | |
0f533c72 TB |
1470 | /* |
1471 | * Don't mark objects not in the packfile or preferred pack. This bitmap | |
1472 | * marks objects eligible for reuse, but the pack-reuse code only | |
1473 | * understands how to reuse a single pack. Since the preferred pack is | |
1474 | * guaranteed to have all bases for its deltas (in a multi-pack bitmap), | |
1475 | * we use it instead of another pack. In single-pack bitmaps, the choice | |
1476 | * is made for us. | |
1477 | */ | |
ed184620 TB |
1478 | if (i > objects_nr / BITS_IN_EWORD) |
1479 | i = objects_nr / BITS_IN_EWORD; | |
fff42755 | 1480 | |
bb514de3 JK |
1481 | reuse = bitmap_word_alloc(i); |
1482 | memset(reuse->words, 0xFF, i * sizeof(eword_t)); | |
fff42755 | 1483 | |
bb514de3 JK |
1484 | for (; i < result->word_alloc; ++i) { |
1485 | eword_t word = result->words[i]; | |
1486 | size_t pos = (i * BITS_IN_EWORD); | |
fff42755 | 1487 | |
bb514de3 JK |
1488 | for (offset = 0; offset < BITS_IN_EWORD; ++offset) { |
1489 | if ((word >> offset) == 0) | |
1490 | break; | |
fff42755 | 1491 | |
bb514de3 | 1492 | offset += ewah_bit_ctz64(word >> offset); |
73cd7d94 | 1493 | if (try_partial_reuse(pack, pos + offset, |
0f533c72 | 1494 | reuse, &w_curs) < 0) { |
a5f9f24a TB |
1495 | /* |
1496 | * try_partial_reuse indicated we couldn't reuse | |
1497 | * any bits, so there is no point in trying more | |
1498 | * bits in the current word, or any other words | |
1499 | * in result. | |
1500 | * | |
1501 | * Jump out of both loops to avoid future | |
1502 | * unnecessary calls to try_partial_reuse. | |
1503 | */ | |
1504 | goto done; | |
1505 | } | |
bb514de3 | 1506 | } |
fff42755 | 1507 | } |
fff42755 | 1508 | |
a5f9f24a | 1509 | done: |
bb514de3 | 1510 | unuse_pack(&w_curs); |
fff42755 | 1511 | |
bb514de3 JK |
1512 | *entries = bitmap_popcount(reuse); |
1513 | if (!*entries) { | |
1514 | bitmap_free(reuse); | |
fff42755 | 1515 | return -1; |
fff42755 VM |
1516 | } |
1517 | ||
bb514de3 JK |
1518 | /* |
1519 | * Drop any reused objects from the result, since they will not | |
1520 | * need to be handled separately. | |
1521 | */ | |
1522 | bitmap_and_not(result, reuse); | |
0f533c72 | 1523 | *packfile_out = pack; |
bb514de3 | 1524 | *reuse_out = reuse; |
fff42755 VM |
1525 | return 0; |
1526 | } | |
fff42755 | 1527 | |
40d18ff8 JK |
1528 | int bitmap_walk_contains(struct bitmap_index *bitmap_git, |
1529 | struct bitmap *bitmap, const struct object_id *oid) | |
1530 | { | |
1531 | int idx; | |
fff42755 | 1532 | |
40d18ff8 JK |
1533 | if (!bitmap) |
1534 | return 0; | |
fff42755 | 1535 | |
40d18ff8 JK |
1536 | idx = bitmap_position(bitmap_git, oid); |
1537 | return idx >= 0 && bitmap_get(bitmap, idx); | |
fff42755 VM |
1538 | } |
1539 | ||
3ae5fa07 | 1540 | void traverse_bitmap_commit_list(struct bitmap_index *bitmap_git, |
4eb707eb | 1541 | struct rev_info *revs, |
3ae5fa07 | 1542 | show_reachable_fn show_reachable) |
fff42755 | 1543 | { |
3ae5fa07 | 1544 | assert(bitmap_git->result); |
fff42755 | 1545 | |
551cf8b6 | 1546 | show_objects_for_type(bitmap_git, OBJ_COMMIT, show_reachable); |
4eb707eb JK |
1547 | if (revs->tree_objects) |
1548 | show_objects_for_type(bitmap_git, OBJ_TREE, show_reachable); | |
1549 | if (revs->blob_objects) | |
1550 | show_objects_for_type(bitmap_git, OBJ_BLOB, show_reachable); | |
1551 | if (revs->tag_objects) | |
1552 | show_objects_for_type(bitmap_git, OBJ_TAG, show_reachable); | |
fff42755 | 1553 | |
4eb707eb | 1554 | show_extended_objects(bitmap_git, revs, show_reachable); |
fff42755 VM |
1555 | } |
1556 | ||
3ae5fa07 | 1557 | static uint32_t count_object_type(struct bitmap_index *bitmap_git, |
fff42755 VM |
1558 | enum object_type type) |
1559 | { | |
3ae5fa07 JT |
1560 | struct bitmap *objects = bitmap_git->result; |
1561 | struct eindex *eindex = &bitmap_git->ext_index; | |
fff42755 VM |
1562 | |
1563 | uint32_t i = 0, count = 0; | |
1564 | struct ewah_iterator it; | |
1565 | eword_t filter; | |
1566 | ||
551cf8b6 | 1567 | init_type_iterator(&it, bitmap_git, type); |
fff42755 VM |
1568 | |
1569 | while (i < objects->word_alloc && ewah_iterator_next(&filter, &it)) { | |
1570 | eword_t word = objects->words[i++] & filter; | |
1571 | count += ewah_bit_popcount64(word); | |
1572 | } | |
1573 | ||
1574 | for (i = 0; i < eindex->count; ++i) { | |
1575 | if (eindex->objects[i]->type == type && | |
ed184620 | 1576 | bitmap_get(objects, bitmap_num_objects(bitmap_git) + i)) |
fff42755 VM |
1577 | count++; |
1578 | } | |
1579 | ||
1580 | return count; | |
1581 | } | |
1582 | ||
3ae5fa07 JT |
1583 | void count_bitmap_commit_list(struct bitmap_index *bitmap_git, |
1584 | uint32_t *commits, uint32_t *trees, | |
fff42755 VM |
1585 | uint32_t *blobs, uint32_t *tags) |
1586 | { | |
3ae5fa07 | 1587 | assert(bitmap_git->result); |
fff42755 VM |
1588 | |
1589 | if (commits) | |
3ae5fa07 | 1590 | *commits = count_object_type(bitmap_git, OBJ_COMMIT); |
fff42755 VM |
1591 | |
1592 | if (trees) | |
3ae5fa07 | 1593 | *trees = count_object_type(bitmap_git, OBJ_TREE); |
fff42755 VM |
1594 | |
1595 | if (blobs) | |
3ae5fa07 | 1596 | *blobs = count_object_type(bitmap_git, OBJ_BLOB); |
fff42755 VM |
1597 | |
1598 | if (tags) | |
3ae5fa07 | 1599 | *tags = count_object_type(bitmap_git, OBJ_TAG); |
fff42755 VM |
1600 | } |
1601 | ||
1602 | struct bitmap_test_data { | |
3ae5fa07 | 1603 | struct bitmap_index *bitmap_git; |
fff42755 | 1604 | struct bitmap *base; |
fa95666a TB |
1605 | struct bitmap *commits; |
1606 | struct bitmap *trees; | |
1607 | struct bitmap *blobs; | |
1608 | struct bitmap *tags; | |
fff42755 VM |
1609 | struct progress *prg; |
1610 | size_t seen; | |
1611 | }; | |
1612 | ||
fa95666a TB |
1613 | static void test_bitmap_type(struct bitmap_test_data *tdata, |
1614 | struct object *obj, int pos) | |
1615 | { | |
1616 | enum object_type bitmap_type = OBJ_NONE; | |
1617 | int bitmaps_nr = 0; | |
1618 | ||
1619 | if (bitmap_get(tdata->commits, pos)) { | |
1620 | bitmap_type = OBJ_COMMIT; | |
1621 | bitmaps_nr++; | |
1622 | } | |
1623 | if (bitmap_get(tdata->trees, pos)) { | |
1624 | bitmap_type = OBJ_TREE; | |
1625 | bitmaps_nr++; | |
1626 | } | |
1627 | if (bitmap_get(tdata->blobs, pos)) { | |
1628 | bitmap_type = OBJ_BLOB; | |
1629 | bitmaps_nr++; | |
1630 | } | |
1631 | if (bitmap_get(tdata->tags, pos)) { | |
1632 | bitmap_type = OBJ_TAG; | |
1633 | bitmaps_nr++; | |
1634 | } | |
1635 | ||
1636 | if (bitmap_type == OBJ_NONE) | |
9975975d | 1637 | die(_("object '%s' not found in type bitmaps"), |
fa95666a TB |
1638 | oid_to_hex(&obj->oid)); |
1639 | ||
1640 | if (bitmaps_nr > 1) | |
9975975d | 1641 | die(_("object '%s' does not have a unique type"), |
fa95666a TB |
1642 | oid_to_hex(&obj->oid)); |
1643 | ||
1644 | if (bitmap_type != obj->type) | |
9975975d | 1645 | die(_("object '%s': real type '%s', expected: '%s'"), |
fa95666a TB |
1646 | oid_to_hex(&obj->oid), |
1647 | type_name(obj->type), | |
1648 | type_name(bitmap_type)); | |
1649 | } | |
1650 | ||
de1e67d0 JK |
1651 | static void test_show_object(struct object *object, const char *name, |
1652 | void *data) | |
fff42755 VM |
1653 | { |
1654 | struct bitmap_test_data *tdata = data; | |
1655 | int bitmap_pos; | |
1656 | ||
3c771448 | 1657 | bitmap_pos = bitmap_position(tdata->bitmap_git, &object->oid); |
fff42755 | 1658 | if (bitmap_pos < 0) |
9975975d | 1659 | die(_("object not in bitmap: '%s'"), oid_to_hex(&object->oid)); |
fa95666a | 1660 | test_bitmap_type(tdata, object, bitmap_pos); |
fff42755 VM |
1661 | |
1662 | bitmap_set(tdata->base, bitmap_pos); | |
1663 | display_progress(tdata->prg, ++tdata->seen); | |
1664 | } | |
1665 | ||
1666 | static void test_show_commit(struct commit *commit, void *data) | |
1667 | { | |
1668 | struct bitmap_test_data *tdata = data; | |
1669 | int bitmap_pos; | |
1670 | ||
3ae5fa07 | 1671 | bitmap_pos = bitmap_position(tdata->bitmap_git, |
3c771448 | 1672 | &commit->object.oid); |
fff42755 | 1673 | if (bitmap_pos < 0) |
9975975d | 1674 | die(_("object not in bitmap: '%s'"), oid_to_hex(&commit->object.oid)); |
fa95666a | 1675 | test_bitmap_type(tdata, &commit->object, bitmap_pos); |
fff42755 VM |
1676 | |
1677 | bitmap_set(tdata->base, bitmap_pos); | |
1678 | display_progress(tdata->prg, ++tdata->seen); | |
1679 | } | |
1680 | ||
1681 | void test_bitmap_walk(struct rev_info *revs) | |
1682 | { | |
1683 | struct object *root; | |
1684 | struct bitmap *result = NULL; | |
fff42755 VM |
1685 | size_t result_popcnt; |
1686 | struct bitmap_test_data tdata; | |
3ae5fa07 | 1687 | struct bitmap_index *bitmap_git; |
98c31f36 | 1688 | struct ewah_bitmap *bm; |
fff42755 | 1689 | |
7c141127 | 1690 | if (!(bitmap_git = prepare_bitmap_git(revs->repo))) |
9975975d | 1691 | die(_("failed to load bitmap indexes")); |
fff42755 VM |
1692 | |
1693 | if (revs->pending.nr != 1) | |
9975975d | 1694 | die(_("you must specify exactly one commit to test")); |
fff42755 | 1695 | |
baf20c39 | 1696 | fprintf_ln(stderr, "Bitmap v%d test (%d entries loaded)", |
3ae5fa07 | 1697 | bitmap_git->version, bitmap_git->entry_count); |
fff42755 VM |
1698 | |
1699 | root = revs->pending.objects[0].item; | |
98c31f36 | 1700 | bm = bitmap_for_commit(bitmap_git, (struct commit *)root); |
fff42755 | 1701 | |
98c31f36 | 1702 | if (bm) { |
baf20c39 | 1703 | fprintf_ln(stderr, "Found bitmap for '%s'. %d bits / %08x checksum", |
f2fd0760 | 1704 | oid_to_hex(&root->oid), (int)bm->bit_size, ewah_checksum(bm)); |
fff42755 VM |
1705 | |
1706 | result = ewah_to_bitmap(bm); | |
1707 | } | |
1708 | ||
1709 | if (result == NULL) | |
9975975d | 1710 | die(_("commit '%s' doesn't have an indexed bitmap"), oid_to_hex(&root->oid)); |
fff42755 VM |
1711 | |
1712 | revs->tag_objects = 1; | |
1713 | revs->tree_objects = 1; | |
1714 | revs->blob_objects = 1; | |
1715 | ||
1716 | result_popcnt = bitmap_popcount(result); | |
1717 | ||
1718 | if (prepare_revision_walk(revs)) | |
9975975d | 1719 | die(_("revision walk setup failed")); |
fff42755 | 1720 | |
3ae5fa07 | 1721 | tdata.bitmap_git = bitmap_git; |
fff42755 | 1722 | tdata.base = bitmap_new(); |
fa95666a TB |
1723 | tdata.commits = ewah_to_bitmap(bitmap_git->commits); |
1724 | tdata.trees = ewah_to_bitmap(bitmap_git->trees); | |
1725 | tdata.blobs = ewah_to_bitmap(bitmap_git->blobs); | |
1726 | tdata.tags = ewah_to_bitmap(bitmap_git->tags); | |
fff42755 VM |
1727 | tdata.prg = start_progress("Verifying bitmap entries", result_popcnt); |
1728 | tdata.seen = 0; | |
1729 | ||
1730 | traverse_commit_list(revs, &test_show_commit, &test_show_object, &tdata); | |
1731 | ||
1732 | stop_progress(&tdata.prg); | |
1733 | ||
1734 | if (bitmap_equals(result, tdata.base)) | |
baf20c39 | 1735 | fprintf_ln(stderr, "OK!"); |
fff42755 | 1736 | else |
9975975d | 1737 | die(_("mismatch in bitmap results")); |
f86a3747 | 1738 | |
02281511 TB |
1739 | bitmap_free(result); |
1740 | bitmap_free(tdata.base); | |
1741 | bitmap_free(tdata.commits); | |
1742 | bitmap_free(tdata.trees); | |
1743 | bitmap_free(tdata.blobs); | |
1744 | bitmap_free(tdata.tags); | |
f3c23db2 | 1745 | free_bitmap_index(bitmap_git); |
fff42755 | 1746 | } |
7cc8f971 | 1747 | |
dff5e49e TB |
1748 | int test_bitmap_commits(struct repository *r) |
1749 | { | |
1750 | struct bitmap_index *bitmap_git = prepare_bitmap_git(r); | |
1751 | struct object_id oid; | |
1752 | MAYBE_UNUSED void *value; | |
1753 | ||
1754 | if (!bitmap_git) | |
9975975d | 1755 | die(_("failed to load bitmap indexes")); |
dff5e49e TB |
1756 | |
1757 | kh_foreach(bitmap_git->bitmaps, oid, value, { | |
baf20c39 | 1758 | printf_ln("%s", oid_to_hex(&oid)); |
dff5e49e TB |
1759 | }); |
1760 | ||
1761 | free_bitmap_index(bitmap_git); | |
1762 | ||
1763 | return 0; | |
1764 | } | |
1765 | ||
a05f02b1 TB |
1766 | int test_bitmap_hashes(struct repository *r) |
1767 | { | |
1768 | struct bitmap_index *bitmap_git = prepare_bitmap_git(r); | |
1769 | struct object_id oid; | |
1770 | uint32_t i, index_pos; | |
1771 | ||
875da7f0 | 1772 | if (!bitmap_git || !bitmap_git->hashes) |
a05f02b1 TB |
1773 | goto cleanup; |
1774 | ||
1775 | for (i = 0; i < bitmap_num_objects(bitmap_git); i++) { | |
1776 | if (bitmap_is_midx(bitmap_git)) | |
1777 | index_pos = pack_pos_to_midx(bitmap_git->midx, i); | |
1778 | else | |
1779 | index_pos = pack_pos_to_index(bitmap_git->pack, i); | |
1780 | ||
1781 | nth_bitmap_object_oid(bitmap_git, &oid, index_pos); | |
1782 | ||
baf20c39 | 1783 | printf_ln("%s %"PRIu32"", |
a05f02b1 TB |
1784 | oid_to_hex(&oid), get_be32(bitmap_git->hashes + index_pos)); |
1785 | } | |
1786 | ||
1787 | cleanup: | |
1788 | free_bitmap_index(bitmap_git); | |
1789 | ||
1790 | return 0; | |
1791 | } | |
1792 | ||
449fa5ee JK |
1793 | int rebuild_bitmap(const uint32_t *reposition, |
1794 | struct ewah_bitmap *source, | |
1795 | struct bitmap *dest) | |
7cc8f971 VM |
1796 | { |
1797 | uint32_t pos = 0; | |
1798 | struct ewah_iterator it; | |
1799 | eword_t word; | |
1800 | ||
1801 | ewah_iterator_init(&it, source); | |
1802 | ||
1803 | while (ewah_iterator_next(&word, &it)) { | |
1804 | uint32_t offset, bit_pos; | |
1805 | ||
34b935c0 | 1806 | for (offset = 0; offset < BITS_IN_EWORD; ++offset) { |
7cc8f971 VM |
1807 | if ((word >> offset) == 0) |
1808 | break; | |
1809 | ||
1810 | offset += ewah_bit_ctz64(word >> offset); | |
1811 | ||
1812 | bit_pos = reposition[pos + offset]; | |
1813 | if (bit_pos > 0) | |
1814 | bitmap_set(dest, bit_pos - 1); | |
1815 | else /* can't reuse, we don't have the object */ | |
1816 | return -1; | |
1817 | } | |
1818 | ||
34b935c0 | 1819 | pos += BITS_IN_EWORD; |
7cc8f971 VM |
1820 | } |
1821 | return 0; | |
1822 | } | |
1823 | ||
449fa5ee JK |
1824 | uint32_t *create_bitmap_mapping(struct bitmap_index *bitmap_git, |
1825 | struct packing_data *mapping) | |
7cc8f971 VM |
1826 | { |
1827 | uint32_t i, num_objects; | |
1828 | uint32_t *reposition; | |
7cc8f971 | 1829 | |
0f533c72 TB |
1830 | if (!bitmap_is_midx(bitmap_git)) |
1831 | load_reverse_index(bitmap_git); | |
1832 | else if (load_midx_revindex(bitmap_git->midx) < 0) | |
1833 | BUG("rebuild_existing_bitmaps: missing required rev-cache " | |
1834 | "extension"); | |
1835 | ||
ed184620 | 1836 | num_objects = bitmap_num_objects(bitmap_git); |
ca56dadb | 1837 | CALLOC_ARRAY(reposition, num_objects); |
7cc8f971 VM |
1838 | |
1839 | for (i = 0; i < num_objects; ++i) { | |
3df28cae | 1840 | struct object_id oid; |
7cc8f971 | 1841 | struct object_entry *oe; |
8de300e1 | 1842 | uint32_t index_pos; |
7cc8f971 | 1843 | |
0f533c72 | 1844 | if (bitmap_is_midx(bitmap_git)) |
8de300e1 | 1845 | index_pos = pack_pos_to_midx(bitmap_git->midx, i); |
0f533c72 | 1846 | else |
8de300e1 TB |
1847 | index_pos = pack_pos_to_index(bitmap_git->pack, i); |
1848 | nth_bitmap_object_oid(bitmap_git, &oid, index_pos); | |
3a37876b | 1849 | oe = packlist_find(mapping, &oid); |
7cc8f971 | 1850 | |
8de300e1 | 1851 | if (oe) { |
06af3bba | 1852 | reposition[i] = oe_in_pack_pos(mapping, oe) + 1; |
8de300e1 TB |
1853 | if (bitmap_git->hashes && !oe->hash) |
1854 | oe->hash = get_be32(bitmap_git->hashes + index_pos); | |
1855 | } | |
7cc8f971 VM |
1856 | } |
1857 | ||
449fa5ee | 1858 | return reposition; |
7cc8f971 | 1859 | } |
f3c23db2 JT |
1860 | |
1861 | void free_bitmap_index(struct bitmap_index *b) | |
1862 | { | |
1863 | if (!b) | |
1864 | return; | |
1865 | ||
1866 | if (b->map) | |
1867 | munmap(b->map, b->map_size); | |
1868 | ewah_pool_free(b->commits); | |
1869 | ewah_pool_free(b->trees); | |
1870 | ewah_pool_free(b->blobs); | |
1871 | ewah_pool_free(b->tags); | |
655b8561 TB |
1872 | if (b->bitmaps) { |
1873 | struct stored_bitmap *sb; | |
1874 | kh_foreach_value(b->bitmaps, sb, { | |
1875 | ewah_pool_free(sb->root); | |
1876 | free(sb); | |
1877 | }); | |
1878 | } | |
3c771448 | 1879 | kh_destroy_oid_map(b->bitmaps); |
f3c23db2 JT |
1880 | free(b->ext_index.objects); |
1881 | free(b->ext_index.hashes); | |
655b8561 | 1882 | kh_destroy_oid_pos(b->ext_index.positions); |
f3c23db2 | 1883 | bitmap_free(b->result); |
30cdc33f | 1884 | bitmap_free(b->haves); |
0f533c72 TB |
1885 | if (bitmap_is_midx(b)) { |
1886 | /* | |
1887 | * Multi-pack bitmaps need to have resources associated with | |
1888 | * their on-disk reverse indexes unmapped so that stale .rev and | |
1889 | * .bitmap files can be removed. | |
1890 | * | |
1891 | * Unlike pack-based bitmaps, multi-pack bitmaps can be read and | |
1892 | * written in the same 'git multi-pack-index write --bitmap' | |
1893 | * process. Close resources so they can be removed safely on | |
1894 | * platforms like Windows. | |
1895 | */ | |
1896 | close_midx_revindex(b->midx); | |
1897 | } | |
f3c23db2 JT |
1898 | free(b); |
1899 | } | |
30cdc33f | 1900 | |
3c771448 | 1901 | int bitmap_has_oid_in_uninteresting(struct bitmap_index *bitmap_git, |
1902 | const struct object_id *oid) | |
30cdc33f | 1903 | { |
8ebf5296 JK |
1904 | return bitmap_git && |
1905 | bitmap_walk_contains(bitmap_git, bitmap_git->haves, oid); | |
30cdc33f | 1906 | } |
16950f83 JK |
1907 | |
1908 | static off_t get_disk_usage_for_type(struct bitmap_index *bitmap_git, | |
1909 | enum object_type object_type) | |
1910 | { | |
1911 | struct bitmap *result = bitmap_git->result; | |
16950f83 JK |
1912 | off_t total = 0; |
1913 | struct ewah_iterator it; | |
1914 | eword_t filter; | |
1915 | size_t i; | |
1916 | ||
1917 | init_type_iterator(&it, bitmap_git, object_type); | |
1918 | for (i = 0; i < result->word_alloc && | |
1919 | ewah_iterator_next(&filter, &it); i++) { | |
1920 | eword_t word = result->words[i] & filter; | |
1921 | size_t base = (i * BITS_IN_EWORD); | |
1922 | unsigned offset; | |
1923 | ||
1924 | if (!word) | |
1925 | continue; | |
1926 | ||
1927 | for (offset = 0; offset < BITS_IN_EWORD; offset++) { | |
16950f83 JK |
1928 | if ((word >> offset) == 0) |
1929 | break; | |
1930 | ||
1931 | offset += ewah_bit_ctz64(word >> offset); | |
0f533c72 TB |
1932 | |
1933 | if (bitmap_is_midx(bitmap_git)) { | |
1934 | uint32_t pack_pos; | |
1935 | uint32_t midx_pos = pack_pos_to_midx(bitmap_git->midx, base + offset); | |
1936 | off_t offset = nth_midxed_offset(bitmap_git->midx, midx_pos); | |
1937 | ||
1938 | uint32_t pack_id = nth_midxed_pack_int_id(bitmap_git->midx, midx_pos); | |
1939 | struct packed_git *pack = bitmap_git->midx->packs[pack_id]; | |
1940 | ||
1941 | if (offset_to_pack_pos(pack, offset, &pack_pos) < 0) { | |
1942 | struct object_id oid; | |
1943 | nth_midxed_object_oid(&oid, bitmap_git->midx, midx_pos); | |
1944 | ||
baf20c39 | 1945 | die(_("could not find '%s' in pack '%s' at offset %"PRIuMAX), |
0f533c72 TB |
1946 | oid_to_hex(&oid), |
1947 | pack->pack_name, | |
1948 | (uintmax_t)offset); | |
1949 | } | |
1950 | ||
1951 | total += pack_pos_to_offset(pack, pack_pos + 1) - offset; | |
1952 | } else { | |
1953 | size_t pos = base + offset; | |
1954 | total += pack_pos_to_offset(bitmap_git->pack, pos + 1) - | |
1955 | pack_pos_to_offset(bitmap_git->pack, pos); | |
1956 | } | |
16950f83 JK |
1957 | } |
1958 | } | |
1959 | ||
1960 | return total; | |
1961 | } | |
1962 | ||
1963 | static off_t get_disk_usage_for_extended(struct bitmap_index *bitmap_git) | |
1964 | { | |
1965 | struct bitmap *result = bitmap_git->result; | |
16950f83 JK |
1966 | struct eindex *eindex = &bitmap_git->ext_index; |
1967 | off_t total = 0; | |
1968 | struct object_info oi = OBJECT_INFO_INIT; | |
1969 | off_t object_size; | |
1970 | size_t i; | |
1971 | ||
1972 | oi.disk_sizep = &object_size; | |
1973 | ||
1974 | for (i = 0; i < eindex->count; i++) { | |
1975 | struct object *obj = eindex->objects[i]; | |
1976 | ||
ed184620 | 1977 | if (!bitmap_get(result, bitmap_num_objects(bitmap_git) + i)) |
16950f83 JK |
1978 | continue; |
1979 | ||
1980 | if (oid_object_info_extended(the_repository, &obj->oid, &oi, 0) < 0) | |
baf20c39 | 1981 | die(_("unable to get disk usage of '%s'"), |
16950f83 JK |
1982 | oid_to_hex(&obj->oid)); |
1983 | ||
1984 | total += object_size; | |
1985 | } | |
1986 | return total; | |
1987 | } | |
1988 | ||
1989 | off_t get_disk_usage_from_bitmap(struct bitmap_index *bitmap_git, | |
1990 | struct rev_info *revs) | |
1991 | { | |
1992 | off_t total = 0; | |
1993 | ||
1994 | total += get_disk_usage_for_type(bitmap_git, OBJ_COMMIT); | |
1995 | if (revs->tree_objects) | |
1996 | total += get_disk_usage_for_type(bitmap_git, OBJ_TREE); | |
1997 | if (revs->blob_objects) | |
1998 | total += get_disk_usage_for_type(bitmap_git, OBJ_BLOB); | |
1999 | if (revs->tag_objects) | |
2000 | total += get_disk_usage_for_type(bitmap_git, OBJ_TAG); | |
2001 | ||
2002 | total += get_disk_usage_for_extended(bitmap_git); | |
2003 | ||
2004 | return total; | |
2005 | } | |
3f267a11 | 2006 | |
0f533c72 TB |
2007 | int bitmap_is_midx(struct bitmap_index *bitmap_git) |
2008 | { | |
2009 | return !!bitmap_git->midx; | |
2010 | } | |
2011 | ||
3f267a11 TB |
2012 | const struct string_list *bitmap_preferred_tips(struct repository *r) |
2013 | { | |
2014 | return repo_config_get_value_multi(r, "pack.preferbitmaptips"); | |
2015 | } | |
711260fd TB |
2016 | |
2017 | int bitmap_is_preferred_refname(struct repository *r, const char *refname) | |
2018 | { | |
2019 | const struct string_list *preferred_tips = bitmap_preferred_tips(r); | |
2020 | struct string_list_item *item; | |
2021 | ||
2022 | if (!preferred_tips) | |
2023 | return 0; | |
2024 | ||
2025 | for_each_string_list_item(item, preferred_tips) { | |
2026 | if (starts_with(refname, item->string)) | |
2027 | return 1; | |
2028 | } | |
2029 | ||
2030 | return 0; | |
2031 | } |