]>
Commit | Line | Data |
---|---|---|
5e3f94df | 1 | #include "git-compat-util.h" |
0b027f6c | 2 | #include "abspath.h" |
c4d25228 | 3 | #include "config.h" |
fc59e748 | 4 | #include "csum-file.h" |
396f2570 | 5 | #include "dir.h" |
f394e093 | 6 | #include "gettext.h" |
41771fa4 | 7 | #include "hex.h" |
fc59e748 | 8 | #include "lockfile.h" |
396f2570 | 9 | #include "packfile.h" |
87bed179 | 10 | #include "object-file.h" |
a034e910 | 11 | #include "object-store-ll.h" |
bc626927 | 12 | #include "hash-lookup.h" |
a3407730 | 13 | #include "midx.h" |
144d7033 | 14 | #include "progress.h" |
d829223a | 15 | #include "trace2.h" |
ce1e4a10 | 16 | #include "run-command.h" |
18e449f8 | 17 | #include "repository.h" |
63a8f0e9 | 18 | #include "chunk-format.h" |
38ff7cab | 19 | #include "pack.h" |
c528e179 TB |
20 | #include "pack-bitmap.h" |
21 | #include "refs.h" | |
22 | #include "revision.h" | |
23 | #include "list-objects.h" | |
a3407730 | 24 | |
fc59e748 DS |
25 | #define MIDX_SIGNATURE 0x4d494458 /* "MIDX" */ |
26 | #define MIDX_VERSION 1 | |
4d80560c DS |
27 | #define MIDX_BYTE_FILE_VERSION 4 |
28 | #define MIDX_BYTE_HASH_VERSION 5 | |
29 | #define MIDX_BYTE_NUM_CHUNKS 6 | |
30 | #define MIDX_BYTE_NUM_PACKS 8 | |
fc59e748 | 31 | #define MIDX_HEADER_SIZE 12 |
aaa95dfa | 32 | #define MIDX_MIN_SIZE (MIDX_HEADER_SIZE + the_hash_algo->rawsz) |
fc59e748 | 33 | |
32f3c541 DS |
34 | #define MIDX_CHUNK_ALIGNMENT 4 |
35 | #define MIDX_CHUNKID_PACKNAMES 0x504e414d /* "PNAM" */ | |
d7cacf29 | 36 | #define MIDX_CHUNKID_OIDFANOUT 0x4f494446 /* "OIDF" */ |
0d5b3a5e | 37 | #define MIDX_CHUNKID_OIDLOOKUP 0x4f49444c /* "OIDL" */ |
662148c4 DS |
38 | #define MIDX_CHUNKID_OBJECTOFFSETS 0x4f4f4646 /* "OOFF" */ |
39 | #define MIDX_CHUNKID_LARGEOFFSETS 0x4c4f4646 /* "LOFF" */ | |
95e8383b | 40 | #define MIDX_CHUNKID_REVINDEX 0x52494458 /* "RIDX" */ |
d7cacf29 | 41 | #define MIDX_CHUNK_FANOUT_SIZE (sizeof(uint32_t) * 256) |
662148c4 DS |
42 | #define MIDX_CHUNK_OFFSET_WIDTH (2 * sizeof(uint32_t)) |
43 | #define MIDX_CHUNK_LARGE_OFFSET_WIDTH (sizeof(uint64_t)) | |
44 | #define MIDX_LARGE_OFFSET_NEEDED 0x80000000 | |
32f3c541 | 45 | |
19575c7c DS |
46 | #define PACK_EXPIRED UINT_MAX |
47 | ||
0f533c72 | 48 | const unsigned char *get_midx_checksum(struct multi_pack_index *m) |
f894081d TB |
49 | { |
50 | return m->data + m->data_len - the_hash_algo->rawsz; | |
51 | } | |
52 | ||
60980aed | 53 | void get_midx_filename(struct strbuf *out, const char *object_dir) |
fc59e748 | 54 | { |
60980aed | 55 | strbuf_addf(out, "%s/pack/multi-pack-index", object_dir); |
fc59e748 DS |
56 | } |
57 | ||
60980aed | 58 | void get_midx_rev_filename(struct strbuf *out, struct multi_pack_index *m) |
f894081d | 59 | { |
60980aed TB |
60 | get_midx_filename(out, m->object_dir); |
61 | strbuf_addf(out, "-%s.rev", hash_to_hex(get_midx_checksum(m))); | |
f894081d TB |
62 | } |
63 | ||
6ab3b8b8 DS |
64 | static int midx_read_oid_fanout(const unsigned char *chunk_start, |
65 | size_t chunk_size, void *data) | |
66 | { | |
67 | struct multi_pack_index *m = data; | |
68 | m->chunk_oid_fanout = (uint32_t *)chunk_start; | |
69 | ||
70 | if (chunk_size != 4 * 256) { | |
71 | error(_("multi-pack-index OID fanout is of the wrong size")); | |
72 | return 1; | |
73 | } | |
fc926567 JK |
74 | m->num_objects = ntohl(m->chunk_oid_fanout[255]); |
75 | return 0; | |
76 | } | |
77 | ||
78 | static int midx_read_oid_lookup(const unsigned char *chunk_start, | |
79 | size_t chunk_size, void *data) | |
80 | { | |
81 | struct multi_pack_index *m = data; | |
82 | m->chunk_oid_lookup = chunk_start; | |
83 | ||
84 | if (chunk_size != st_mult(m->hash_len, m->num_objects)) { | |
85 | error(_("multi-pack-index OID lookup chunk is the wrong size")); | |
86 | return 1; | |
87 | } | |
6ab3b8b8 DS |
88 | return 0; |
89 | } | |
90 | ||
0924869b JK |
91 | static int midx_read_object_offsets(const unsigned char *chunk_start, |
92 | size_t chunk_size, void *data) | |
93 | { | |
94 | struct multi_pack_index *m = data; | |
95 | m->chunk_object_offsets = chunk_start; | |
96 | ||
97 | if (chunk_size != st_mult(m->num_objects, MIDX_CHUNK_OFFSET_WIDTH)) { | |
98 | error(_("multi-pack-index object offset chunk is the wrong size")); | |
99 | return 1; | |
100 | } | |
101 | return 0; | |
102 | } | |
103 | ||
2cf489a3 | 104 | struct multi_pack_index *load_multi_pack_index(const char *object_dir, int local) |
4d80560c DS |
105 | { |
106 | struct multi_pack_index *m = NULL; | |
107 | int fd; | |
108 | struct stat st; | |
109 | size_t midx_size; | |
110 | void *midx_map = NULL; | |
111 | uint32_t hash_version; | |
60980aed | 112 | struct strbuf midx_name = STRBUF_INIT; |
32f3c541 | 113 | uint32_t i; |
3227565c | 114 | const char *cur_pack_name; |
6ab3b8b8 | 115 | struct chunkfile *cf = NULL; |
4d80560c | 116 | |
60980aed TB |
117 | get_midx_filename(&midx_name, object_dir); |
118 | ||
119 | fd = git_open(midx_name.buf); | |
4d80560c DS |
120 | |
121 | if (fd < 0) | |
122 | goto cleanup_fail; | |
123 | if (fstat(fd, &st)) { | |
60980aed | 124 | error_errno(_("failed to read %s"), midx_name.buf); |
4d80560c DS |
125 | goto cleanup_fail; |
126 | } | |
127 | ||
128 | midx_size = xsize_t(st.st_size); | |
129 | ||
130 | if (midx_size < MIDX_MIN_SIZE) { | |
60980aed | 131 | error(_("multi-pack-index file %s is too small"), midx_name.buf); |
4d80560c DS |
132 | goto cleanup_fail; |
133 | } | |
134 | ||
60980aed | 135 | strbuf_release(&midx_name); |
4d80560c DS |
136 | |
137 | midx_map = xmmap(NULL, midx_size, PROT_READ, MAP_PRIVATE, fd, 0); | |
6c7ff7cf | 138 | close(fd); |
4d80560c | 139 | |
577314ca | 140 | FLEX_ALLOC_STR(m, object_dir, object_dir); |
4d80560c DS |
141 | m->data = midx_map; |
142 | m->data_len = midx_size; | |
2cf489a3 | 143 | m->local = local; |
4d80560c DS |
144 | |
145 | m->signature = get_be32(m->data); | |
53ad0407 DS |
146 | if (m->signature != MIDX_SIGNATURE) |
147 | die(_("multi-pack-index signature 0x%08x does not match signature 0x%08x"), | |
4d80560c | 148 | m->signature, MIDX_SIGNATURE); |
4d80560c DS |
149 | |
150 | m->version = m->data[MIDX_BYTE_FILE_VERSION]; | |
53ad0407 DS |
151 | if (m->version != MIDX_VERSION) |
152 | die(_("multi-pack-index version %d not recognized"), | |
4d80560c | 153 | m->version); |
4d80560c DS |
154 | |
155 | hash_version = m->data[MIDX_BYTE_HASH_VERSION]; | |
d9fef9d9 | 156 | if (hash_version != oid_version(the_hash_algo)) { |
d9607542 | 157 | error(_("multi-pack-index hash version %u does not match version %u"), |
d9fef9d9 | 158 | hash_version, oid_version(the_hash_algo)); |
d9607542 DS |
159 | goto cleanup_fail; |
160 | } | |
aaa95dfa | 161 | m->hash_len = the_hash_algo->rawsz; |
4d80560c DS |
162 | |
163 | m->num_chunks = m->data[MIDX_BYTE_NUM_CHUNKS]; | |
164 | ||
165 | m->num_packs = get_be32(m->data + MIDX_BYTE_NUM_PACKS); | |
166 | ||
6ab3b8b8 | 167 | cf = init_chunkfile(NULL); |
32f3c541 | 168 | |
6ab3b8b8 | 169 | if (read_table_of_contents(cf, m->data, midx_size, |
c9b9fefc JK |
170 | MIDX_HEADER_SIZE, m->num_chunks, |
171 | MIDX_CHUNK_ALIGNMENT)) | |
6ab3b8b8 | 172 | goto cleanup_fail; |
32f3c541 | 173 | |
72a9a082 | 174 | if (pair_chunk(cf, MIDX_CHUNKID_PACKNAMES, &m->chunk_pack_names, &m->chunk_pack_names_len)) |
e3c96003 JK |
175 | die(_("multi-pack-index required pack-name chunk missing or corrupted")); |
176 | if (read_chunk(cf, MIDX_CHUNKID_OIDFANOUT, midx_read_oid_fanout, m)) | |
177 | die(_("multi-pack-index required OID fanout chunk missing or corrupted")); | |
fc926567 | 178 | if (read_chunk(cf, MIDX_CHUNKID_OIDLOOKUP, midx_read_oid_lookup, m)) |
e3c96003 | 179 | die(_("multi-pack-index required OID lookup chunk missing or corrupted")); |
0924869b | 180 | if (read_chunk(cf, MIDX_CHUNKID_OBJECTOFFSETS, midx_read_object_offsets, m)) |
e3c96003 | 181 | die(_("multi-pack-index required object offsets chunk missing or corrupted")); |
32f3c541 | 182 | |
2abd56e9 JK |
183 | pair_chunk(cf, MIDX_CHUNKID_LARGEOFFSETS, &m->chunk_large_offsets, |
184 | &m->chunk_large_offsets_len); | |
6ab3b8b8 | 185 | |
7f514b7a | 186 | if (git_env_bool("GIT_TEST_MIDX_READ_RIDX", 1)) |
c0fe9b2d JK |
187 | pair_chunk(cf, MIDX_CHUNKID_REVINDEX, &m->chunk_revindex, |
188 | &m->chunk_revindex_len); | |
7f514b7a | 189 | |
ca56dadb RS |
190 | CALLOC_ARRAY(m->pack_names, m->num_packs); |
191 | CALLOC_ARRAY(m->packs, m->num_packs); | |
3227565c DS |
192 | |
193 | cur_pack_name = (const char *)m->chunk_pack_names; | |
194 | for (i = 0; i < m->num_packs; i++) { | |
72a9a082 JK |
195 | const char *end; |
196 | size_t avail = m->chunk_pack_names_len - | |
197 | (cur_pack_name - (const char *)m->chunk_pack_names); | |
198 | ||
3227565c DS |
199 | m->pack_names[i] = cur_pack_name; |
200 | ||
72a9a082 JK |
201 | end = memchr(cur_pack_name, '\0', avail); |
202 | if (!end) | |
203 | die(_("multi-pack-index pack-name chunk is too short")); | |
204 | cur_pack_name = end + 1; | |
3227565c | 205 | |
8e72a3c3 DS |
206 | if (i && strcmp(m->pack_names[i], m->pack_names[i - 1]) <= 0) |
207 | die(_("multi-pack-index pack names out of order: '%s' before '%s'"), | |
3227565c DS |
208 | m->pack_names[i - 1], |
209 | m->pack_names[i]); | |
3227565c DS |
210 | } |
211 | ||
d829223a JH |
212 | trace2_data_intmax("midx", the_repository, "load/num_packs", m->num_packs); |
213 | trace2_data_intmax("midx", the_repository, "load/num_objects", m->num_objects); | |
214 | ||
692305ec | 215 | free_chunkfile(cf); |
4d80560c DS |
216 | return m; |
217 | ||
218 | cleanup_fail: | |
219 | free(m); | |
60980aed | 220 | strbuf_release(&midx_name); |
692305ec | 221 | free_chunkfile(cf); |
4d80560c DS |
222 | if (midx_map) |
223 | munmap(midx_map, midx_size); | |
224 | if (0 <= fd) | |
225 | close(fd); | |
226 | return NULL; | |
227 | } | |
228 | ||
1dcd9f20 | 229 | void close_midx(struct multi_pack_index *m) |
a40498a1 DS |
230 | { |
231 | uint32_t i; | |
1dcd9f20 DS |
232 | |
233 | if (!m) | |
234 | return; | |
235 | ||
9bb6c2e5 TB |
236 | close_midx(m->next); |
237 | ||
a40498a1 | 238 | munmap((unsigned char *)m->data, m->data_len); |
a40498a1 DS |
239 | |
240 | for (i = 0; i < m->num_packs; i++) { | |
af96fe33 DS |
241 | if (m->packs[i]) |
242 | m->packs[i]->multi_pack_index = 0; | |
a40498a1 DS |
243 | } |
244 | FREE_AND_NULL(m->packs); | |
245 | FREE_AND_NULL(m->pack_names); | |
9bb6c2e5 | 246 | free(m); |
a40498a1 DS |
247 | } |
248 | ||
64404a24 | 249 | int prepare_midx_pack(struct repository *r, struct multi_pack_index *m, uint32_t pack_int_id) |
3715a633 DS |
250 | { |
251 | struct strbuf pack_name = STRBUF_INIT; | |
af96fe33 | 252 | struct packed_git *p; |
3715a633 DS |
253 | |
254 | if (pack_int_id >= m->num_packs) | |
d355e46a | 255 | die(_("bad pack-int-id: %u (%u total packs)"), |
cc6af73c | 256 | pack_int_id, m->num_packs); |
3715a633 DS |
257 | |
258 | if (m->packs[pack_int_id]) | |
259 | return 0; | |
260 | ||
261 | strbuf_addf(&pack_name, "%s/pack/%s", m->object_dir, | |
262 | m->pack_names[pack_int_id]); | |
263 | ||
af96fe33 | 264 | p = add_packed_git(pack_name.buf, pack_name.len, m->local); |
3715a633 | 265 | strbuf_release(&pack_name); |
af96fe33 DS |
266 | |
267 | if (!p) | |
268 | return 1; | |
269 | ||
270 | p->multi_pack_index = 1; | |
271 | m->packs[pack_int_id] = p; | |
272 | install_packed_git(r, p); | |
273 | list_add_tail(&p->mru, &r->objects->packed_git_mru); | |
274 | ||
275 | return 0; | |
3715a633 DS |
276 | } |
277 | ||
278 | int bsearch_midx(const struct object_id *oid, struct multi_pack_index *m, uint32_t *result) | |
279 | { | |
280 | return bsearch_hash(oid->hash, m->chunk_oid_fanout, m->chunk_oid_lookup, | |
aaa95dfa | 281 | the_hash_algo->rawsz, result); |
3715a633 DS |
282 | } |
283 | ||
8aac67a1 DS |
284 | struct object_id *nth_midxed_object_oid(struct object_id *oid, |
285 | struct multi_pack_index *m, | |
286 | uint32_t n) | |
287 | { | |
288 | if (n >= m->num_objects) | |
289 | return NULL; | |
290 | ||
c2b24ede | 291 | oidread(oid, m->chunk_oid_lookup + st_mult(m->hash_len, n)); |
8aac67a1 DS |
292 | return oid; |
293 | } | |
294 | ||
62f2c1b5 | 295 | off_t nth_midxed_offset(struct multi_pack_index *m, uint32_t pos) |
3715a633 DS |
296 | { |
297 | const unsigned char *offset_data; | |
298 | uint32_t offset32; | |
299 | ||
329fac3a | 300 | offset_data = m->chunk_object_offsets + (off_t)pos * MIDX_CHUNK_OFFSET_WIDTH; |
3715a633 DS |
301 | offset32 = get_be32(offset_data + sizeof(uint32_t)); |
302 | ||
303 | if (m->chunk_large_offsets && offset32 & MIDX_LARGE_OFFSET_NEEDED) { | |
d8ac9ee1 | 304 | if (sizeof(off_t) < sizeof(uint64_t)) |
3715a633 DS |
305 | die(_("multi-pack-index stores a 64-bit offset, but off_t is too small")); |
306 | ||
307 | offset32 ^= MIDX_LARGE_OFFSET_NEEDED; | |
2abd56e9 JK |
308 | if (offset32 >= m->chunk_large_offsets_len / sizeof(uint64_t)) |
309 | die(_("multi-pack-index large offset out of bounds")); | |
310 | return get_be64(m->chunk_large_offsets + sizeof(uint64_t) * offset32); | |
3715a633 DS |
311 | } |
312 | ||
313 | return offset32; | |
314 | } | |
315 | ||
62f2c1b5 | 316 | uint32_t nth_midxed_pack_int_id(struct multi_pack_index *m, uint32_t pos) |
3715a633 | 317 | { |
329fac3a DS |
318 | return get_be32(m->chunk_object_offsets + |
319 | (off_t)pos * MIDX_CHUNK_OFFSET_WIDTH); | |
3715a633 DS |
320 | } |
321 | ||
a8437f3c | 322 | int fill_midx_entry(struct repository *r, |
893b5635 RS |
323 | const struct object_id *oid, |
324 | struct pack_entry *e, | |
325 | struct multi_pack_index *m) | |
3715a633 | 326 | { |
893b5635 | 327 | uint32_t pos; |
3715a633 DS |
328 | uint32_t pack_int_id; |
329 | struct packed_git *p; | |
330 | ||
893b5635 RS |
331 | if (!bsearch_midx(oid, m, &pos)) |
332 | return 0; | |
333 | ||
3715a633 DS |
334 | if (pos >= m->num_objects) |
335 | return 0; | |
336 | ||
337 | pack_int_id = nth_midxed_pack_int_id(m, pos); | |
338 | ||
64404a24 | 339 | if (prepare_midx_pack(r, m, pack_int_id)) |
506ec2fb | 340 | return 0; |
3715a633 DS |
341 | p = m->packs[pack_int_id]; |
342 | ||
343 | /* | |
344 | * We are about to tell the caller where they can locate the | |
345 | * requested object. We better make sure the packfile is | |
346 | * still here and can be accessed before supplying that | |
347 | * answer, as it may have been deleted since the MIDX was | |
348 | * loaded! | |
349 | */ | |
350 | if (!is_pack_valid(p)) | |
351 | return 0; | |
352 | ||
09ef6617 RS |
353 | if (oidset_size(&p->bad_objects) && |
354 | oidset_contains(&p->bad_objects, oid)) | |
355 | return 0; | |
c39b02ae | 356 | |
3715a633 DS |
357 | e->offset = nth_midxed_offset(m, pos); |
358 | e->p = p; | |
359 | ||
360 | return 1; | |
361 | } | |
362 | ||
013fd7ad JK |
363 | /* Match "foo.idx" against either "foo.pack" _or_ "foo.idx". */ |
364 | static int cmp_idx_or_pack_name(const char *idx_or_pack_name, | |
365 | const char *idx_name) | |
366 | { | |
367 | /* Skip past any initial matching prefix. */ | |
368 | while (*idx_name && *idx_name == *idx_or_pack_name) { | |
369 | idx_name++; | |
370 | idx_or_pack_name++; | |
371 | } | |
372 | ||
373 | /* | |
374 | * If we didn't match completely, we may have matched "pack-1234." and | |
375 | * be left with "idx" and "pack" respectively, which is also OK. We do | |
376 | * not have to check for "idx" and "idx", because that would have been | |
377 | * a complete match (and in that case these strcmps will be false, but | |
378 | * we'll correctly return 0 from the final strcmp() below. | |
379 | * | |
380 | * Technically this matches "fooidx" and "foopack", but we'd never have | |
381 | * such names in the first place. | |
382 | */ | |
383 | if (!strcmp(idx_name, "idx") && !strcmp(idx_or_pack_name, "pack")) | |
384 | return 0; | |
385 | ||
386 | /* | |
387 | * This not only checks for a complete match, but also orders based on | |
388 | * the first non-identical character, which means our ordering will | |
389 | * match a raw strcmp(). That makes it OK to use this to binary search | |
390 | * a naively-sorted list. | |
391 | */ | |
392 | return strcmp(idx_or_pack_name, idx_name); | |
393 | } | |
394 | ||
395 | int midx_contains_pack(struct multi_pack_index *m, const char *idx_or_pack_name) | |
a40498a1 DS |
396 | { |
397 | uint32_t first = 0, last = m->num_packs; | |
398 | ||
399 | while (first < last) { | |
400 | uint32_t mid = first + (last - first) / 2; | |
401 | const char *current; | |
402 | int cmp; | |
403 | ||
404 | current = m->pack_names[mid]; | |
013fd7ad | 405 | cmp = cmp_idx_or_pack_name(idx_or_pack_name, current); |
a40498a1 DS |
406 | if (!cmp) |
407 | return 1; | |
408 | if (cmp > 0) { | |
409 | first = mid + 1; | |
410 | continue; | |
411 | } | |
412 | last = mid; | |
413 | } | |
414 | ||
415 | return 0; | |
416 | } | |
417 | ||
2cf489a3 | 418 | int prepare_multi_pack_index_one(struct repository *r, const char *object_dir, int local) |
c4d25228 | 419 | { |
29e2016b | 420 | struct multi_pack_index *m; |
c4d25228 | 421 | struct multi_pack_index *m_search; |
c4d25228 | 422 | |
18e449f8 DS |
423 | prepare_repo_settings(r); |
424 | if (!r->settings.core_multi_pack_index) | |
c4d25228 DS |
425 | return 0; |
426 | ||
29e2016b | 427 | for (m_search = r->objects->multi_pack_index; m_search; m_search = m_search->next) |
c4d25228 DS |
428 | if (!strcmp(object_dir, m_search->object_dir)) |
429 | return 1; | |
430 | ||
29e2016b | 431 | m = load_multi_pack_index(object_dir, local); |
c4d25228 | 432 | |
29e2016b | 433 | if (m) { |
59552fb3 TB |
434 | struct multi_pack_index *mp = r->objects->multi_pack_index; |
435 | if (mp) { | |
436 | m->next = mp->next; | |
437 | mp->next = m; | |
438 | } else | |
439 | r->objects->multi_pack_index = m; | |
c4d25228 DS |
440 | return 1; |
441 | } | |
442 | ||
443 | return 0; | |
444 | } | |
445 | ||
fc59e748 DS |
446 | static size_t write_midx_header(struct hashfile *f, |
447 | unsigned char num_chunks, | |
448 | uint32_t num_packs) | |
449 | { | |
fc59e748 | 450 | hashwrite_be32(f, MIDX_SIGNATURE); |
014f1447 | 451 | hashwrite_u8(f, MIDX_VERSION); |
d9fef9d9 | 452 | hashwrite_u8(f, oid_version(the_hash_algo)); |
014f1447 RS |
453 | hashwrite_u8(f, num_chunks); |
454 | hashwrite_u8(f, 0); /* unused */ | |
fc59e748 DS |
455 | hashwrite_be32(f, num_packs); |
456 | ||
457 | return MIDX_HEADER_SIZE; | |
458 | } | |
459 | ||
d01bf2e6 DS |
460 | struct pack_info { |
461 | uint32_t orig_pack_int_id; | |
462 | char *pack_name; | |
463 | struct packed_git *p; | |
19575c7c | 464 | unsigned expired : 1; |
d01bf2e6 DS |
465 | }; |
466 | ||
467 | static int pack_info_compare(const void *_a, const void *_b) | |
468 | { | |
469 | struct pack_info *a = (struct pack_info *)_a; | |
470 | struct pack_info *b = (struct pack_info *)_b; | |
471 | return strcmp(a->pack_name, b->pack_name); | |
472 | } | |
473 | ||
9218c6a4 TB |
474 | static int idx_or_pack_name_cmp(const void *_va, const void *_vb) |
475 | { | |
476 | const char *pack_name = _va; | |
477 | const struct pack_info *compar = _vb; | |
478 | ||
479 | return cmp_idx_or_pack_name(pack_name, compar->pack_name); | |
480 | } | |
481 | ||
577dc496 | 482 | struct write_midx_context { |
d01bf2e6 | 483 | struct pack_info *info; |
cc381274 TB |
484 | size_t nr; |
485 | size_t alloc; | |
a40498a1 | 486 | struct multi_pack_index *m; |
840cef0c WB |
487 | struct progress *progress; |
488 | unsigned pack_paths_checked; | |
31bda9a2 DS |
489 | |
490 | struct pack_midx_entry *entries; | |
cc381274 | 491 | size_t entries_nr; |
7a3ada11 DS |
492 | |
493 | uint32_t *pack_perm; | |
38ff7cab | 494 | uint32_t *pack_order; |
7a3ada11 | 495 | unsigned large_offsets_needed:1; |
980f525c | 496 | uint32_t num_large_offsets; |
9218c6a4 TB |
497 | |
498 | int preferred_pack_idx; | |
56d863e9 TB |
499 | |
500 | struct string_list *to_include; | |
396f2570 DS |
501 | }; |
502 | ||
503 | static void add_pack_to_midx(const char *full_path, size_t full_path_len, | |
504 | const char *file_name, void *data) | |
505 | { | |
577dc496 | 506 | struct write_midx_context *ctx = data; |
396f2570 DS |
507 | |
508 | if (ends_with(file_name, ".idx")) { | |
577dc496 | 509 | display_progress(ctx->progress, ++ctx->pack_paths_checked); |
56d863e9 TB |
510 | /* |
511 | * Note that at most one of ctx->m and ctx->to_include are set, | |
512 | * so we are testing midx_contains_pack() and | |
513 | * string_list_has_string() independently (guarded by the | |
514 | * appropriate NULL checks). | |
515 | * | |
516 | * We could support passing to_include while reusing an existing | |
517 | * MIDX, but don't currently since the reuse process drags | |
518 | * forward all packs from an existing MIDX (without checking | |
519 | * whether or not they appear in the to_include list). | |
520 | * | |
521 | * If we added support for that, these next two conditional | |
522 | * should be performed independently (likely checking | |
523 | * to_include before the existing MIDX). | |
524 | */ | |
577dc496 | 525 | if (ctx->m && midx_contains_pack(ctx->m, file_name)) |
a40498a1 | 526 | return; |
56d863e9 TB |
527 | else if (ctx->to_include && |
528 | !string_list_has_string(ctx->to_include, file_name)) | |
529 | return; | |
a40498a1 | 530 | |
577dc496 | 531 | ALLOC_GROW(ctx->info, ctx->nr + 1, ctx->alloc); |
396f2570 | 532 | |
577dc496 DS |
533 | ctx->info[ctx->nr].p = add_packed_git(full_path, |
534 | full_path_len, | |
535 | 0); | |
fe1ed56f | 536 | |
577dc496 | 537 | if (!ctx->info[ctx->nr].p) { |
396f2570 DS |
538 | warning(_("failed to add packfile '%s'"), |
539 | full_path); | |
540 | return; | |
541 | } | |
542 | ||
577dc496 | 543 | if (open_pack_index(ctx->info[ctx->nr].p)) { |
fe1ed56f DS |
544 | warning(_("failed to open pack-index '%s'"), |
545 | full_path); | |
577dc496 DS |
546 | close_pack(ctx->info[ctx->nr].p); |
547 | FREE_AND_NULL(ctx->info[ctx->nr].p); | |
fe1ed56f DS |
548 | return; |
549 | } | |
550 | ||
577dc496 DS |
551 | ctx->info[ctx->nr].pack_name = xstrdup(file_name); |
552 | ctx->info[ctx->nr].orig_pack_int_id = ctx->nr; | |
553 | ctx->info[ctx->nr].expired = 0; | |
554 | ctx->nr++; | |
396f2570 DS |
555 | } |
556 | } | |
557 | ||
fe1ed56f DS |
558 | struct pack_midx_entry { |
559 | struct object_id oid; | |
560 | uint32_t pack_int_id; | |
561 | time_t pack_mtime; | |
562 | uint64_t offset; | |
9218c6a4 | 563 | unsigned preferred : 1; |
fe1ed56f DS |
564 | }; |
565 | ||
566 | static int midx_oid_compare(const void *_a, const void *_b) | |
567 | { | |
568 | const struct pack_midx_entry *a = (const struct pack_midx_entry *)_a; | |
569 | const struct pack_midx_entry *b = (const struct pack_midx_entry *)_b; | |
570 | int cmp = oidcmp(&a->oid, &b->oid); | |
571 | ||
572 | if (cmp) | |
573 | return cmp; | |
574 | ||
9218c6a4 TB |
575 | /* Sort objects in a preferred pack first when multiple copies exist. */ |
576 | if (a->preferred > b->preferred) | |
577 | return -1; | |
578 | if (a->preferred < b->preferred) | |
579 | return 1; | |
580 | ||
fe1ed56f DS |
581 | if (a->pack_mtime > b->pack_mtime) |
582 | return -1; | |
583 | else if (a->pack_mtime < b->pack_mtime) | |
584 | return 1; | |
585 | ||
586 | return a->pack_int_id - b->pack_int_id; | |
587 | } | |
588 | ||
a40498a1 | 589 | static int nth_midxed_pack_midx_entry(struct multi_pack_index *m, |
a40498a1 DS |
590 | struct pack_midx_entry *e, |
591 | uint32_t pos) | |
592 | { | |
593 | if (pos >= m->num_objects) | |
594 | return 1; | |
595 | ||
596 | nth_midxed_object_oid(&e->oid, m, pos); | |
d01bf2e6 | 597 | e->pack_int_id = nth_midxed_pack_int_id(m, pos); |
a40498a1 DS |
598 | e->offset = nth_midxed_offset(m, pos); |
599 | ||
600 | /* consider objects in midx to be from "old" packs */ | |
601 | e->pack_mtime = 0; | |
602 | return 0; | |
603 | } | |
604 | ||
fe1ed56f DS |
605 | static void fill_pack_entry(uint32_t pack_int_id, |
606 | struct packed_git *p, | |
607 | uint32_t cur_object, | |
9218c6a4 TB |
608 | struct pack_midx_entry *entry, |
609 | int preferred) | |
fe1ed56f | 610 | { |
0763671b | 611 | if (nth_packed_object_id(&entry->oid, p, cur_object) < 0) |
fe1ed56f DS |
612 | die(_("failed to locate object %d in packfile"), cur_object); |
613 | ||
614 | entry->pack_int_id = pack_int_id; | |
615 | entry->pack_mtime = p->mtime; | |
616 | ||
617 | entry->offset = nth_packed_object_offset(p, cur_object); | |
9218c6a4 | 618 | entry->preferred = !!preferred; |
fe1ed56f DS |
619 | } |
620 | ||
989d9cbd TB |
621 | struct midx_fanout { |
622 | struct pack_midx_entry *entries; | |
e6c71f23 | 623 | size_t nr, alloc; |
989d9cbd TB |
624 | }; |
625 | ||
e6c71f23 | 626 | static void midx_fanout_grow(struct midx_fanout *fanout, size_t nr) |
989d9cbd | 627 | { |
e6c71f23 TB |
628 | if (nr < fanout->nr) |
629 | BUG("negative growth in midx_fanout_grow() (%"PRIuMAX" < %"PRIuMAX")", | |
630 | (uintmax_t)nr, (uintmax_t)fanout->nr); | |
989d9cbd TB |
631 | ALLOC_GROW(fanout->entries, nr, fanout->alloc); |
632 | } | |
633 | ||
634 | static void midx_fanout_sort(struct midx_fanout *fanout) | |
635 | { | |
636 | QSORT(fanout->entries, fanout->nr, midx_oid_compare); | |
637 | } | |
638 | ||
852c5301 TB |
639 | static void midx_fanout_add_midx_fanout(struct midx_fanout *fanout, |
640 | struct multi_pack_index *m, | |
99e4d084 TB |
641 | uint32_t cur_fanout, |
642 | int preferred_pack) | |
852c5301 TB |
643 | { |
644 | uint32_t start = 0, end; | |
645 | uint32_t cur_object; | |
646 | ||
647 | if (cur_fanout) | |
648 | start = ntohl(m->chunk_oid_fanout[cur_fanout - 1]); | |
649 | end = ntohl(m->chunk_oid_fanout[cur_fanout]); | |
650 | ||
651 | for (cur_object = start; cur_object < end; cur_object++) { | |
99e4d084 TB |
652 | if ((preferred_pack > -1) && |
653 | (preferred_pack == nth_midxed_pack_int_id(m, cur_object))) { | |
654 | /* | |
655 | * Objects from preferred packs are added | |
656 | * separately. | |
657 | */ | |
658 | continue; | |
659 | } | |
660 | ||
852c5301 TB |
661 | midx_fanout_grow(fanout, fanout->nr + 1); |
662 | nth_midxed_pack_midx_entry(m, | |
663 | &fanout->entries[fanout->nr], | |
664 | cur_object); | |
cdf517be | 665 | fanout->entries[fanout->nr].preferred = 0; |
852c5301 TB |
666 | fanout->nr++; |
667 | } | |
668 | } | |
669 | ||
1d6f4c64 TB |
670 | static void midx_fanout_add_pack_fanout(struct midx_fanout *fanout, |
671 | struct pack_info *info, | |
672 | uint32_t cur_pack, | |
673 | int preferred, | |
674 | uint32_t cur_fanout) | |
675 | { | |
676 | struct packed_git *pack = info[cur_pack].p; | |
677 | uint32_t start = 0, end; | |
678 | uint32_t cur_object; | |
679 | ||
680 | if (cur_fanout) | |
681 | start = get_pack_fanout(pack, cur_fanout - 1); | |
682 | end = get_pack_fanout(pack, cur_fanout); | |
683 | ||
684 | for (cur_object = start; cur_object < end; cur_object++) { | |
685 | midx_fanout_grow(fanout, fanout->nr + 1); | |
686 | fill_pack_entry(cur_pack, | |
687 | info[cur_pack].p, | |
688 | cur_object, | |
689 | &fanout->entries[fanout->nr], | |
690 | preferred); | |
691 | fanout->nr++; | |
692 | } | |
693 | } | |
694 | ||
fe1ed56f DS |
695 | /* |
696 | * It is possible to artificially get into a state where there are many | |
697 | * duplicate copies of objects. That can create high memory pressure if | |
698 | * we are to create a list of all objects before de-duplication. To reduce | |
699 | * this memory pressure without a significant performance drop, automatically | |
700 | * group objects by the first byte of their object id. Use the IDX fanout | |
701 | * tables to group the data, copy to a local array, then sort. | |
702 | * | |
703 | * Copy only the de-duplicated entries (selected by most-recent modified time | |
704 | * of a packfile containing the object). | |
705 | */ | |
a40498a1 | 706 | static struct pack_midx_entry *get_sorted_entries(struct multi_pack_index *m, |
d01bf2e6 | 707 | struct pack_info *info, |
fe1ed56f | 708 | uint32_t nr_packs, |
cc381274 | 709 | size_t *nr_objects, |
9218c6a4 | 710 | int preferred_pack) |
fe1ed56f DS |
711 | { |
712 | uint32_t cur_fanout, cur_pack, cur_object; | |
cc381274 | 713 | size_t alloc_objects, total_objects = 0; |
989d9cbd | 714 | struct midx_fanout fanout = { 0 }; |
fe1ed56f | 715 | struct pack_midx_entry *deduplicated_entries = NULL; |
a40498a1 | 716 | uint32_t start_pack = m ? m->num_packs : 0; |
fe1ed56f | 717 | |
a40498a1 | 718 | for (cur_pack = start_pack; cur_pack < nr_packs; cur_pack++) |
cc381274 TB |
719 | total_objects = st_add(total_objects, |
720 | info[cur_pack].p->num_objects); | |
fe1ed56f DS |
721 | |
722 | /* | |
723 | * As we de-duplicate by fanout value, we expect the fanout | |
724 | * slices to be evenly distributed, with some noise. Hence, | |
725 | * allocate slightly more than one 256th. | |
726 | */ | |
989d9cbd | 727 | alloc_objects = fanout.alloc = total_objects > 3200 ? total_objects / 200 : 16; |
fe1ed56f | 728 | |
989d9cbd | 729 | ALLOC_ARRAY(fanout.entries, fanout.alloc); |
fe1ed56f DS |
730 | ALLOC_ARRAY(deduplicated_entries, alloc_objects); |
731 | *nr_objects = 0; | |
732 | ||
733 | for (cur_fanout = 0; cur_fanout < 256; cur_fanout++) { | |
989d9cbd | 734 | fanout.nr = 0; |
fe1ed56f | 735 | |
852c5301 | 736 | if (m) |
99e4d084 TB |
737 | midx_fanout_add_midx_fanout(&fanout, m, cur_fanout, |
738 | preferred_pack); | |
a40498a1 DS |
739 | |
740 | for (cur_pack = start_pack; cur_pack < nr_packs; cur_pack++) { | |
9218c6a4 | 741 | int preferred = cur_pack == preferred_pack; |
1d6f4c64 TB |
742 | midx_fanout_add_pack_fanout(&fanout, |
743 | info, cur_pack, | |
744 | preferred, cur_fanout); | |
fe1ed56f DS |
745 | } |
746 | ||
cdf517be TB |
747 | if (-1 < preferred_pack && preferred_pack < start_pack) |
748 | midx_fanout_add_pack_fanout(&fanout, info, | |
749 | preferred_pack, 1, | |
750 | cur_fanout); | |
751 | ||
989d9cbd | 752 | midx_fanout_sort(&fanout); |
fe1ed56f DS |
753 | |
754 | /* | |
755 | * The batch is now sorted by OID and then mtime (descending). | |
756 | * Take only the first duplicate. | |
757 | */ | |
989d9cbd TB |
758 | for (cur_object = 0; cur_object < fanout.nr; cur_object++) { |
759 | if (cur_object && oideq(&fanout.entries[cur_object - 1].oid, | |
760 | &fanout.entries[cur_object].oid)) | |
fe1ed56f DS |
761 | continue; |
762 | ||
cc381274 TB |
763 | ALLOC_GROW(deduplicated_entries, st_add(*nr_objects, 1), |
764 | alloc_objects); | |
fe1ed56f | 765 | memcpy(&deduplicated_entries[*nr_objects], |
989d9cbd | 766 | &fanout.entries[cur_object], |
fe1ed56f DS |
767 | sizeof(struct pack_midx_entry)); |
768 | (*nr_objects)++; | |
769 | } | |
770 | } | |
771 | ||
989d9cbd | 772 | free(fanout.entries); |
fe1ed56f DS |
773 | return deduplicated_entries; |
774 | } | |
775 | ||
0ccd713c | 776 | static int write_midx_pack_names(struct hashfile *f, void *data) |
32f3c541 | 777 | { |
b4d94142 | 778 | struct write_midx_context *ctx = data; |
32f3c541 DS |
779 | uint32_t i; |
780 | unsigned char padding[MIDX_CHUNK_ALIGNMENT]; | |
781 | size_t written = 0; | |
782 | ||
b4d94142 | 783 | for (i = 0; i < ctx->nr; i++) { |
19575c7c DS |
784 | size_t writelen; |
785 | ||
b4d94142 | 786 | if (ctx->info[i].expired) |
19575c7c | 787 | continue; |
32f3c541 | 788 | |
b4d94142 | 789 | if (i && strcmp(ctx->info[i].pack_name, ctx->info[i - 1].pack_name) <= 0) |
32f3c541 | 790 | BUG("incorrect pack-file order: %s before %s", |
b4d94142 DS |
791 | ctx->info[i - 1].pack_name, |
792 | ctx->info[i].pack_name); | |
32f3c541 | 793 | |
b4d94142 DS |
794 | writelen = strlen(ctx->info[i].pack_name) + 1; |
795 | hashwrite(f, ctx->info[i].pack_name, writelen); | |
32f3c541 DS |
796 | written += writelen; |
797 | } | |
798 | ||
799 | /* add padding to be aligned */ | |
800 | i = MIDX_CHUNK_ALIGNMENT - (written % MIDX_CHUNK_ALIGNMENT); | |
801 | if (i < MIDX_CHUNK_ALIGNMENT) { | |
802 | memset(padding, 0, sizeof(padding)); | |
803 | hashwrite(f, padding, i); | |
32f3c541 DS |
804 | } |
805 | ||
0ccd713c | 806 | return 0; |
32f3c541 DS |
807 | } |
808 | ||
0ccd713c DS |
809 | static int write_midx_oid_fanout(struct hashfile *f, |
810 | void *data) | |
d7cacf29 | 811 | { |
31bda9a2 DS |
812 | struct write_midx_context *ctx = data; |
813 | struct pack_midx_entry *list = ctx->entries; | |
814 | struct pack_midx_entry *last = ctx->entries + ctx->entries_nr; | |
d7cacf29 DS |
815 | uint32_t count = 0; |
816 | uint32_t i; | |
817 | ||
818 | /* | |
819 | * Write the first-level table (the list is sorted, | |
820 | * but we use a 256-entry lookup to be able to avoid | |
821 | * having to do eight extra binary search iterations). | |
822 | */ | |
823 | for (i = 0; i < 256; i++) { | |
824 | struct pack_midx_entry *next = list; | |
825 | ||
826 | while (next < last && next->oid.hash[0] == i) { | |
827 | count++; | |
828 | next++; | |
829 | } | |
830 | ||
831 | hashwrite_be32(f, count); | |
832 | list = next; | |
833 | } | |
834 | ||
0ccd713c | 835 | return 0; |
d7cacf29 DS |
836 | } |
837 | ||
0ccd713c DS |
838 | static int write_midx_oid_lookup(struct hashfile *f, |
839 | void *data) | |
0d5b3a5e | 840 | { |
31bda9a2 DS |
841 | struct write_midx_context *ctx = data; |
842 | unsigned char hash_len = the_hash_algo->rawsz; | |
843 | struct pack_midx_entry *list = ctx->entries; | |
0d5b3a5e | 844 | uint32_t i; |
0d5b3a5e | 845 | |
31bda9a2 | 846 | for (i = 0; i < ctx->entries_nr; i++) { |
0d5b3a5e DS |
847 | struct pack_midx_entry *obj = list++; |
848 | ||
31bda9a2 | 849 | if (i < ctx->entries_nr - 1) { |
0d5b3a5e DS |
850 | struct pack_midx_entry *next = list; |
851 | if (oidcmp(&obj->oid, &next->oid) >= 0) | |
852 | BUG("OIDs not in order: %s >= %s", | |
853 | oid_to_hex(&obj->oid), | |
854 | oid_to_hex(&next->oid)); | |
855 | } | |
856 | ||
857 | hashwrite(f, obj->oid.hash, (int)hash_len); | |
0d5b3a5e DS |
858 | } |
859 | ||
0ccd713c | 860 | return 0; |
0d5b3a5e DS |
861 | } |
862 | ||
0ccd713c DS |
863 | static int write_midx_object_offsets(struct hashfile *f, |
864 | void *data) | |
662148c4 | 865 | { |
7a3ada11 DS |
866 | struct write_midx_context *ctx = data; |
867 | struct pack_midx_entry *list = ctx->entries; | |
662148c4 | 868 | uint32_t i, nr_large_offset = 0; |
662148c4 | 869 | |
7a3ada11 | 870 | for (i = 0; i < ctx->entries_nr; i++) { |
662148c4 DS |
871 | struct pack_midx_entry *obj = list++; |
872 | ||
7a3ada11 | 873 | if (ctx->pack_perm[obj->pack_int_id] == PACK_EXPIRED) |
19575c7c DS |
874 | BUG("object %s is in an expired pack with int-id %d", |
875 | oid_to_hex(&obj->oid), | |
876 | obj->pack_int_id); | |
877 | ||
7a3ada11 | 878 | hashwrite_be32(f, ctx->pack_perm[obj->pack_int_id]); |
662148c4 | 879 | |
7a3ada11 | 880 | if (ctx->large_offsets_needed && obj->offset >> 31) |
662148c4 | 881 | hashwrite_be32(f, MIDX_LARGE_OFFSET_NEEDED | nr_large_offset++); |
7a3ada11 | 882 | else if (!ctx->large_offsets_needed && obj->offset >> 32) |
662148c4 DS |
883 | BUG("object %s requires a large offset (%"PRIx64") but the MIDX is not writing large offsets!", |
884 | oid_to_hex(&obj->oid), | |
885 | obj->offset); | |
886 | else | |
887 | hashwrite_be32(f, (uint32_t)obj->offset); | |
662148c4 DS |
888 | } |
889 | ||
0ccd713c | 890 | return 0; |
662148c4 DS |
891 | } |
892 | ||
0ccd713c DS |
893 | static int write_midx_large_offsets(struct hashfile *f, |
894 | void *data) | |
662148c4 | 895 | { |
980f525c DS |
896 | struct write_midx_context *ctx = data; |
897 | struct pack_midx_entry *list = ctx->entries; | |
898 | struct pack_midx_entry *end = ctx->entries + ctx->entries_nr; | |
980f525c | 899 | uint32_t nr_large_offset = ctx->num_large_offsets; |
662148c4 DS |
900 | |
901 | while (nr_large_offset) { | |
61b0fcbb JK |
902 | struct pack_midx_entry *obj; |
903 | uint64_t offset; | |
904 | ||
905 | if (list >= end) | |
906 | BUG("too many large-offset objects"); | |
907 | ||
908 | obj = list++; | |
909 | offset = obj->offset; | |
662148c4 DS |
910 | |
911 | if (!(offset >> 31)) | |
912 | continue; | |
913 | ||
0ccd713c | 914 | hashwrite_be64(f, offset); |
662148c4 DS |
915 | |
916 | nr_large_offset--; | |
917 | } | |
918 | ||
0ccd713c | 919 | return 0; |
662148c4 DS |
920 | } |
921 | ||
95e8383b TB |
922 | static int write_midx_revindex(struct hashfile *f, |
923 | void *data) | |
924 | { | |
925 | struct write_midx_context *ctx = data; | |
926 | uint32_t i; | |
927 | ||
928 | for (i = 0; i < ctx->entries_nr; i++) | |
929 | hashwrite_be32(f, ctx->pack_order[i]); | |
930 | ||
931 | return 0; | |
932 | } | |
933 | ||
30077524 JK |
934 | struct midx_pack_order_data { |
935 | uint32_t nr; | |
936 | uint32_t pack; | |
937 | off_t offset; | |
938 | }; | |
38ff7cab | 939 | |
30077524 JK |
940 | static int midx_pack_order_cmp(const void *va, const void *vb) |
941 | { | |
942 | const struct midx_pack_order_data *a = va, *b = vb; | |
943 | if (a->pack < b->pack) | |
38ff7cab | 944 | return -1; |
30077524 | 945 | else if (a->pack > b->pack) |
38ff7cab | 946 | return 1; |
30077524 | 947 | else if (a->offset < b->offset) |
38ff7cab | 948 | return -1; |
30077524 | 949 | else if (a->offset > b->offset) |
38ff7cab | 950 | return 1; |
30077524 JK |
951 | else |
952 | return 0; | |
38ff7cab TB |
953 | } |
954 | ||
955 | static uint32_t *midx_pack_order(struct write_midx_context *ctx) | |
956 | { | |
30077524 | 957 | struct midx_pack_order_data *data; |
38ff7cab TB |
958 | uint32_t *pack_order; |
959 | uint32_t i; | |
960 | ||
2dcff525 TB |
961 | trace2_region_enter("midx", "midx_pack_order", the_repository); |
962 | ||
30077524 JK |
963 | ALLOC_ARRAY(data, ctx->entries_nr); |
964 | for (i = 0; i < ctx->entries_nr; i++) { | |
965 | struct pack_midx_entry *e = &ctx->entries[i]; | |
966 | data[i].nr = i; | |
967 | data[i].pack = ctx->pack_perm[e->pack_int_id]; | |
968 | if (!e->preferred) | |
969 | data[i].pack |= (1U << 31); | |
970 | data[i].offset = e->offset; | |
971 | } | |
972 | ||
973 | QSORT(data, ctx->entries_nr, midx_pack_order_cmp); | |
974 | ||
38ff7cab TB |
975 | ALLOC_ARRAY(pack_order, ctx->entries_nr); |
976 | for (i = 0; i < ctx->entries_nr; i++) | |
30077524 JK |
977 | pack_order[i] = data[i].nr; |
978 | free(data); | |
38ff7cab | 979 | |
2dcff525 TB |
980 | trace2_region_leave("midx", "midx_pack_order", the_repository); |
981 | ||
38ff7cab TB |
982 | return pack_order; |
983 | } | |
984 | ||
985 | static void write_midx_reverse_index(char *midx_name, unsigned char *midx_hash, | |
986 | struct write_midx_context *ctx) | |
987 | { | |
988 | struct strbuf buf = STRBUF_INIT; | |
989 | const char *tmp_file; | |
990 | ||
2dcff525 TB |
991 | trace2_region_enter("midx", "write_midx_reverse_index", the_repository); |
992 | ||
38ff7cab TB |
993 | strbuf_addf(&buf, "%s-%s.rev", midx_name, hash_to_hex(midx_hash)); |
994 | ||
995 | tmp_file = write_rev_file_order(NULL, ctx->pack_order, ctx->entries_nr, | |
996 | midx_hash, WRITE_REV); | |
997 | ||
998 | if (finalize_object_file(tmp_file, buf.buf)) | |
999 | die(_("cannot store reverse index file")); | |
1000 | ||
1001 | strbuf_release(&buf); | |
2dcff525 TB |
1002 | |
1003 | trace2_region_leave("midx", "write_midx_reverse_index", the_repository); | |
38ff7cab TB |
1004 | } |
1005 | ||
426c00e4 | 1006 | static void clear_midx_files_ext(const char *object_dir, const char *ext, |
38ff7cab TB |
1007 | unsigned char *keep_hash); |
1008 | ||
ec1e28ef TB |
1009 | static int midx_checksum_valid(struct multi_pack_index *m) |
1010 | { | |
1011 | return hashfile_checksum_valid(m->data, m->data_len); | |
1012 | } | |
1013 | ||
c528e179 TB |
1014 | static void prepare_midx_packing_data(struct packing_data *pdata, |
1015 | struct write_midx_context *ctx) | |
1016 | { | |
1017 | uint32_t i; | |
1018 | ||
2dcff525 TB |
1019 | trace2_region_enter("midx", "prepare_midx_packing_data", the_repository); |
1020 | ||
c528e179 TB |
1021 | memset(pdata, 0, sizeof(struct packing_data)); |
1022 | prepare_packing_data(the_repository, pdata); | |
1023 | ||
1024 | for (i = 0; i < ctx->entries_nr; i++) { | |
1025 | struct pack_midx_entry *from = &ctx->entries[ctx->pack_order[i]]; | |
1026 | struct object_entry *to = packlist_alloc(pdata, &from->oid); | |
1027 | ||
1028 | oe_set_in_pack(pdata, to, | |
1029 | ctx->info[ctx->pack_perm[from->pack_int_id]].p); | |
1030 | } | |
2dcff525 TB |
1031 | |
1032 | trace2_region_leave("midx", "prepare_midx_packing_data", the_repository); | |
c528e179 TB |
1033 | } |
1034 | ||
1035 | static int add_ref_to_pending(const char *refname, | |
1036 | const struct object_id *oid, | |
1037 | int flag, void *cb_data) | |
1038 | { | |
1039 | struct rev_info *revs = (struct rev_info*)cb_data; | |
1dc4f1ef | 1040 | struct object_id peeled; |
c528e179 TB |
1041 | struct object *object; |
1042 | ||
1043 | if ((flag & REF_ISSYMREF) && (flag & REF_ISBROKEN)) { | |
1044 | warning("symbolic ref is dangling: %s", refname); | |
1045 | return 0; | |
1046 | } | |
1047 | ||
1dc4f1ef TB |
1048 | if (!peel_iterated_oid(oid, &peeled)) |
1049 | oid = &peeled; | |
1050 | ||
c528e179 TB |
1051 | object = parse_object_or_die(oid, refname); |
1052 | if (object->type != OBJ_COMMIT) | |
1053 | return 0; | |
1054 | ||
1055 | add_pending_object(revs, object, ""); | |
1056 | if (bitmap_is_preferred_refname(revs->repo, refname)) | |
1057 | object->flags |= NEEDS_BITMAP; | |
1058 | return 0; | |
1059 | } | |
1060 | ||
1061 | struct bitmap_commit_cb { | |
1062 | struct commit **commits; | |
1063 | size_t commits_nr, commits_alloc; | |
1064 | ||
1065 | struct write_midx_context *ctx; | |
1066 | }; | |
1067 | ||
1068 | static const struct object_id *bitmap_oid_access(size_t index, | |
1069 | const void *_entries) | |
1070 | { | |
1071 | const struct pack_midx_entry *entries = _entries; | |
1072 | return &entries[index].oid; | |
1073 | } | |
1074 | ||
1075 | static void bitmap_show_commit(struct commit *commit, void *_data) | |
1076 | { | |
1077 | struct bitmap_commit_cb *data = _data; | |
1078 | int pos = oid_pos(&commit->object.oid, data->ctx->entries, | |
1079 | data->ctx->entries_nr, | |
1080 | bitmap_oid_access); | |
1081 | if (pos < 0) | |
1082 | return; | |
1083 | ||
1084 | ALLOC_GROW(data->commits, data->commits_nr + 1, data->commits_alloc); | |
1085 | data->commits[data->commits_nr++] = commit; | |
1086 | } | |
1087 | ||
08944d1c TB |
1088 | static int read_refs_snapshot(const char *refs_snapshot, |
1089 | struct rev_info *revs) | |
1090 | { | |
1091 | struct strbuf buf = STRBUF_INIT; | |
1092 | struct object_id oid; | |
1093 | FILE *f = xfopen(refs_snapshot, "r"); | |
1094 | ||
1095 | while (strbuf_getline(&buf, f) != EOF) { | |
1096 | struct object *object; | |
1097 | int preferred = 0; | |
1098 | char *hex = buf.buf; | |
1099 | const char *end = NULL; | |
1100 | ||
1101 | if (buf.len && *buf.buf == '+') { | |
1102 | preferred = 1; | |
1103 | hex = &buf.buf[1]; | |
1104 | } | |
1105 | ||
1106 | if (parse_oid_hex(hex, &oid, &end) < 0) | |
1107 | die(_("could not parse line: %s"), buf.buf); | |
1108 | if (*end) | |
1109 | die(_("malformed line: %s"), buf.buf); | |
1110 | ||
1111 | object = parse_object_or_die(&oid, NULL); | |
1112 | if (preferred) | |
1113 | object->flags |= NEEDS_BITMAP; | |
1114 | ||
1115 | add_pending_object(revs, object, ""); | |
1116 | } | |
1117 | ||
1118 | fclose(f); | |
1119 | strbuf_release(&buf); | |
1120 | return 0; | |
1121 | } | |
1122 | ||
c528e179 | 1123 | static struct commit **find_commits_for_midx_bitmap(uint32_t *indexed_commits_nr_p, |
08944d1c | 1124 | const char *refs_snapshot, |
c528e179 TB |
1125 | struct write_midx_context *ctx) |
1126 | { | |
1127 | struct rev_info revs; | |
1128 | struct bitmap_commit_cb cb = {0}; | |
1129 | ||
2dcff525 TB |
1130 | trace2_region_enter("midx", "find_commits_for_midx_bitmap", |
1131 | the_repository); | |
1132 | ||
c528e179 TB |
1133 | cb.ctx = ctx; |
1134 | ||
1135 | repo_init_revisions(the_repository, &revs, NULL); | |
08944d1c TB |
1136 | if (refs_snapshot) { |
1137 | read_refs_snapshot(refs_snapshot, &revs); | |
1138 | } else { | |
1139 | setup_revisions(0, NULL, &revs, NULL); | |
1140 | for_each_ref(add_ref_to_pending, &revs); | |
1141 | } | |
c528e179 TB |
1142 | |
1143 | /* | |
1144 | * Skipping promisor objects here is intentional, since it only excludes | |
1145 | * them from the list of reachable commits that we want to select from | |
1146 | * when computing the selection of MIDX'd commits to receive bitmaps. | |
1147 | * | |
1148 | * Reachability bitmaps do require that their objects be closed under | |
1149 | * reachability, but fetching any objects missing from promisors at this | |
1150 | * point is too late. But, if one of those objects can be reached from | |
1151 | * an another object that is included in the bitmap, then we will | |
1152 | * complain later that we don't have reachability closure (and fail | |
1153 | * appropriately). | |
1154 | */ | |
1155 | fetch_if_missing = 0; | |
1156 | revs.exclude_promisor_objects = 1; | |
1157 | ||
1158 | if (prepare_revision_walk(&revs)) | |
1159 | die(_("revision walk setup failed")); | |
1160 | ||
1161 | traverse_commit_list(&revs, bitmap_show_commit, NULL, &cb); | |
1162 | if (indexed_commits_nr_p) | |
1163 | *indexed_commits_nr_p = cb.commits_nr; | |
1164 | ||
2108fe4a | 1165 | release_revisions(&revs); |
2dcff525 TB |
1166 | |
1167 | trace2_region_leave("midx", "find_commits_for_midx_bitmap", | |
1168 | the_repository); | |
1169 | ||
c528e179 TB |
1170 | return cb.commits; |
1171 | } | |
1172 | ||
90b2bb71 DS |
1173 | static int write_midx_bitmap(const char *midx_name, |
1174 | const unsigned char *midx_hash, | |
1175 | struct packing_data *pdata, | |
1176 | struct commit **commits, | |
1177 | uint32_t commits_nr, | |
1178 | uint32_t *pack_order, | |
c528e179 TB |
1179 | unsigned flags) |
1180 | { | |
90b2bb71 | 1181 | int ret, i; |
caca3c9f | 1182 | uint16_t options = 0; |
90b2bb71 DS |
1183 | struct pack_idx_entry **index; |
1184 | char *bitmap_name = xstrfmt("%s-%s.bitmap", midx_name, | |
1185 | hash_to_hex(midx_hash)); | |
eb57277b | 1186 | |
2dcff525 TB |
1187 | trace2_region_enter("midx", "write_midx_bitmap", the_repository); |
1188 | ||
caca3c9f TB |
1189 | if (flags & MIDX_WRITE_BITMAP_HASH_CACHE) |
1190 | options |= BITMAP_OPT_HASH_CACHE; | |
1191 | ||
76f14b77 AC |
1192 | if (flags & MIDX_WRITE_BITMAP_LOOKUP_TABLE) |
1193 | options |= BITMAP_OPT_LOOKUP_TABLE; | |
1194 | ||
c528e179 TB |
1195 | /* |
1196 | * Build the MIDX-order index based on pdata.objects (which is already | |
1197 | * in MIDX order; c.f., 'midx_pack_order_cmp()' for the definition of | |
1198 | * this order). | |
1199 | */ | |
90b2bb71 DS |
1200 | ALLOC_ARRAY(index, pdata->nr_objects); |
1201 | for (i = 0; i < pdata->nr_objects; i++) | |
1202 | index[i] = &pdata->objects[i].idx; | |
c528e179 TB |
1203 | |
1204 | bitmap_writer_show_progress(flags & MIDX_PROGRESS); | |
90b2bb71 | 1205 | bitmap_writer_build_type_index(pdata, index, pdata->nr_objects); |
c528e179 TB |
1206 | |
1207 | /* | |
1208 | * bitmap_writer_finish expects objects in lex order, but pack_order | |
1209 | * gives us exactly that. use it directly instead of re-sorting the | |
1210 | * array. | |
1211 | * | |
1212 | * This changes the order of objects in 'index' between | |
1213 | * bitmap_writer_build_type_index and bitmap_writer_finish. | |
1214 | * | |
1215 | * The same re-ordering takes place in the single-pack bitmap code via | |
1216 | * write_idx_file(), which is called by finish_tmp_packfile(), which | |
1217 | * happens between bitmap_writer_build_type_index() and | |
1218 | * bitmap_writer_finish(). | |
1219 | */ | |
90b2bb71 DS |
1220 | for (i = 0; i < pdata->nr_objects; i++) |
1221 | index[pack_order[i]] = &pdata->objects[i].idx; | |
c528e179 TB |
1222 | |
1223 | bitmap_writer_select_commits(commits, commits_nr, -1); | |
90b2bb71 | 1224 | ret = bitmap_writer_build(pdata); |
c528e179 TB |
1225 | if (ret < 0) |
1226 | goto cleanup; | |
1227 | ||
1228 | bitmap_writer_set_checksum(midx_hash); | |
90b2bb71 | 1229 | bitmap_writer_finish(index, pdata->nr_objects, bitmap_name, options); |
c528e179 TB |
1230 | |
1231 | cleanup: | |
1232 | free(index); | |
1233 | free(bitmap_name); | |
2dcff525 TB |
1234 | |
1235 | trace2_region_leave("midx", "write_midx_bitmap", the_repository); | |
1236 | ||
c528e179 TB |
1237 | return ret; |
1238 | } | |
1239 | ||
504131ac TB |
1240 | static struct multi_pack_index *lookup_multi_pack_index(struct repository *r, |
1241 | const char *object_dir) | |
1242 | { | |
eafcc6de | 1243 | struct multi_pack_index *result = NULL; |
504131ac | 1244 | struct multi_pack_index *cur; |
eafcc6de DS |
1245 | char *obj_dir_real = real_pathdup(object_dir, 1); |
1246 | struct strbuf cur_path_real = STRBUF_INIT; | |
504131ac TB |
1247 | |
1248 | /* Ensure the given object_dir is local, or a known alternate. */ | |
eafcc6de | 1249 | find_odb(r, obj_dir_real); |
504131ac TB |
1250 | |
1251 | for (cur = get_multi_pack_index(r); cur; cur = cur->next) { | |
eafcc6de DS |
1252 | strbuf_realpath(&cur_path_real, cur->object_dir, 1); |
1253 | if (!strcmp(obj_dir_real, cur_path_real.buf)) { | |
1254 | result = cur; | |
1255 | goto cleanup; | |
1256 | } | |
504131ac TB |
1257 | } |
1258 | ||
eafcc6de DS |
1259 | cleanup: |
1260 | free(obj_dir_real); | |
1261 | strbuf_release(&cur_path_real); | |
1262 | return result; | |
504131ac TB |
1263 | } |
1264 | ||
f57a7396 | 1265 | static int write_midx_internal(const char *object_dir, |
56d863e9 | 1266 | struct string_list *packs_to_include, |
9218c6a4 TB |
1267 | struct string_list *packs_to_drop, |
1268 | const char *preferred_pack_name, | |
08944d1c | 1269 | const char *refs_snapshot, |
9218c6a4 | 1270 | unsigned flags) |
a3407730 | 1271 | { |
60980aed | 1272 | struct strbuf midx_name = STRBUF_INIT; |
9f191611 | 1273 | unsigned char midx_hash[GIT_MAX_RAWSZ]; |
396f2570 | 1274 | uint32_t i; |
fc59e748 DS |
1275 | struct hashfile *f = NULL; |
1276 | struct lock_file lk; | |
577dc496 | 1277 | struct write_midx_context ctx = { 0 }; |
dba6175c | 1278 | int pack_name_concat_len = 0; |
19575c7c DS |
1279 | int dropped_packs = 0; |
1280 | int result = 0; | |
63a8f0e9 | 1281 | struct chunkfile *cf; |
fc59e748 | 1282 | |
2dcff525 TB |
1283 | trace2_region_enter("midx", "write_midx_internal", the_repository); |
1284 | ||
60980aed TB |
1285 | get_midx_filename(&midx_name, object_dir); |
1286 | if (safe_create_leading_directories(midx_name.buf)) | |
fc59e748 | 1287 | die_errno(_("unable to create leading directories of %s"), |
60980aed | 1288 | midx_name.buf); |
fc59e748 | 1289 | |
56d863e9 TB |
1290 | if (!packs_to_include) { |
1291 | /* | |
1292 | * Only reference an existing MIDX when not filtering which | |
1293 | * packs to include, since all packs and objects are copied | |
1294 | * blindly from an existing MIDX if one is present. | |
1295 | */ | |
504131ac | 1296 | ctx.m = lookup_multi_pack_index(the_repository, object_dir); |
f57a7396 | 1297 | } |
577dc496 | 1298 | |
ec1e28ef TB |
1299 | if (ctx.m && !midx_checksum_valid(ctx.m)) { |
1300 | warning(_("ignoring existing multi-pack-index; checksum mismatch")); | |
1301 | ctx.m = NULL; | |
1302 | } | |
1303 | ||
577dc496 DS |
1304 | ctx.nr = 0; |
1305 | ctx.alloc = ctx.m ? ctx.m->num_packs : 16; | |
1306 | ctx.info = NULL; | |
1307 | ALLOC_ARRAY(ctx.info, ctx.alloc); | |
1308 | ||
1309 | if (ctx.m) { | |
1310 | for (i = 0; i < ctx.m->num_packs; i++) { | |
1311 | ALLOC_GROW(ctx.info, ctx.nr + 1, ctx.alloc); | |
1312 | ||
1313 | ctx.info[ctx.nr].orig_pack_int_id = i; | |
1314 | ctx.info[ctx.nr].pack_name = xstrdup(ctx.m->pack_names[i]); | |
c528e179 | 1315 | ctx.info[ctx.nr].p = ctx.m->packs[i]; |
577dc496 | 1316 | ctx.info[ctx.nr].expired = 0; |
5d3cd09a TB |
1317 | |
1318 | if (flags & MIDX_WRITE_REV_INDEX) { | |
1319 | /* | |
1320 | * If generating a reverse index, need to have | |
1321 | * packed_git's loaded to compare their | |
1322 | * mtimes and object count. | |
1323 | */ | |
1324 | if (prepare_midx_pack(the_repository, ctx.m, i)) { | |
1325 | error(_("could not load pack")); | |
1326 | result = 1; | |
1327 | goto cleanup; | |
1328 | } | |
1329 | ||
1330 | if (open_pack_index(ctx.m->packs[i])) | |
1331 | die(_("could not open index for %s"), | |
1332 | ctx.m->packs[i]->pack_name); | |
1333 | ctx.info[ctx.nr].p = ctx.m->packs[i]; | |
1334 | } | |
1335 | ||
577dc496 | 1336 | ctx.nr++; |
a40498a1 DS |
1337 | } |
1338 | } | |
1339 | ||
577dc496 | 1340 | ctx.pack_paths_checked = 0; |
840cef0c | 1341 | if (flags & MIDX_PROGRESS) |
577dc496 | 1342 | ctx.progress = start_delayed_progress(_("Adding packfiles to multi-pack-index"), 0); |
840cef0c | 1343 | else |
577dc496 | 1344 | ctx.progress = NULL; |
840cef0c | 1345 | |
56d863e9 TB |
1346 | ctx.to_include = packs_to_include; |
1347 | ||
577dc496 DS |
1348 | for_each_file_in_pack_dir(object_dir, add_pack_to_midx, &ctx); |
1349 | stop_progress(&ctx.progress); | |
396f2570 | 1350 | |
56d863e9 TB |
1351 | if ((ctx.m && ctx.nr == ctx.m->num_packs) && |
1352 | !(packs_to_include || packs_to_drop)) { | |
c528e179 TB |
1353 | struct bitmap_index *bitmap_git; |
1354 | int bitmap_exists; | |
1355 | int want_bitmap = flags & MIDX_WRITE_BITMAP; | |
1356 | ||
bfbb60d3 | 1357 | bitmap_git = prepare_midx_bitmap_git(ctx.m); |
c528e179 TB |
1358 | bitmap_exists = bitmap_git && bitmap_is_midx(bitmap_git); |
1359 | free_bitmap_index(bitmap_git); | |
1360 | ||
1361 | if (bitmap_exists || !want_bitmap) { | |
1362 | /* | |
1363 | * The correct MIDX already exists, and so does a | |
1364 | * corresponding bitmap (or one wasn't requested). | |
1365 | */ | |
1366 | if (!want_bitmap) | |
1367 | clear_midx_files_ext(object_dir, ".bitmap", | |
1368 | NULL); | |
1369 | goto cleanup; | |
1370 | } | |
1371 | } | |
a40498a1 | 1372 | |
9218c6a4 | 1373 | if (preferred_pack_name) { |
ceb96a16 PS |
1374 | ctx.preferred_pack_idx = -1; |
1375 | ||
9218c6a4 TB |
1376 | for (i = 0; i < ctx.nr; i++) { |
1377 | if (!cmp_idx_or_pack_name(preferred_pack_name, | |
1378 | ctx.info[i].pack_name)) { | |
1379 | ctx.preferred_pack_idx = i; | |
1380 | break; | |
1381 | } | |
1382 | } | |
177c0d6e | 1383 | |
ceb96a16 | 1384 | if (ctx.preferred_pack_idx == -1) |
177c0d6e TB |
1385 | warning(_("unknown preferred pack: '%s'"), |
1386 | preferred_pack_name); | |
c528e179 TB |
1387 | } else if (ctx.nr && |
1388 | (flags & (MIDX_WRITE_REV_INDEX | MIDX_WRITE_BITMAP))) { | |
177c0d6e TB |
1389 | struct packed_git *oldest = ctx.info[ctx.preferred_pack_idx].p; |
1390 | ctx.preferred_pack_idx = 0; | |
1391 | ||
1392 | if (packs_to_drop && packs_to_drop->nr) | |
1393 | BUG("cannot write a MIDX bitmap during expiration"); | |
1394 | ||
1395 | /* | |
1396 | * set a preferred pack when writing a bitmap to ensure that | |
1397 | * the pack from which the first object is selected in pseudo | |
1398 | * pack-order has all of its objects selected from that pack | |
1399 | * (and not another pack containing a duplicate) | |
1400 | */ | |
1401 | for (i = 1; i < ctx.nr; i++) { | |
1402 | struct packed_git *p = ctx.info[i].p; | |
1403 | ||
1404 | if (!oldest->num_objects || p->mtime < oldest->mtime) { | |
1405 | oldest = p; | |
1406 | ctx.preferred_pack_idx = i; | |
1407 | } | |
1408 | } | |
1409 | ||
1410 | if (!oldest->num_objects) { | |
1411 | /* | |
1412 | * If all packs are empty; unset the preferred index. | |
1413 | * This is acceptable since there will be no duplicate | |
1414 | * objects to resolve, so the preferred value doesn't | |
1415 | * matter. | |
1416 | */ | |
1417 | ctx.preferred_pack_idx = -1; | |
1418 | } | |
1419 | } else { | |
1420 | /* | |
1421 | * otherwise don't mark any pack as preferred to avoid | |
1422 | * interfering with expiration logic below | |
1423 | */ | |
1424 | ctx.preferred_pack_idx = -1; | |
9218c6a4 TB |
1425 | } |
1426 | ||
5d3cd09a TB |
1427 | if (ctx.preferred_pack_idx > -1) { |
1428 | struct packed_git *preferred = ctx.info[ctx.preferred_pack_idx].p; | |
1429 | if (!preferred->num_objects) { | |
1430 | error(_("cannot select preferred pack %s with no objects"), | |
1431 | preferred->pack_name); | |
1432 | result = 1; | |
1433 | goto cleanup; | |
1434 | } | |
1435 | } | |
1436 | ||
9218c6a4 TB |
1437 | ctx.entries = get_sorted_entries(ctx.m, ctx.info, ctx.nr, &ctx.entries_nr, |
1438 | ctx.preferred_pack_idx); | |
a40498a1 | 1439 | |
7a3ada11 | 1440 | ctx.large_offsets_needed = 0; |
31bda9a2 DS |
1441 | for (i = 0; i < ctx.entries_nr; i++) { |
1442 | if (ctx.entries[i].offset > 0x7fffffff) | |
980f525c | 1443 | ctx.num_large_offsets++; |
31bda9a2 | 1444 | if (ctx.entries[i].offset > 0xffffffff) |
7a3ada11 | 1445 | ctx.large_offsets_needed = 1; |
662148c4 | 1446 | } |
fe1ed56f | 1447 | |
577dc496 | 1448 | QSORT(ctx.info, ctx.nr, pack_info_compare); |
d01bf2e6 | 1449 | |
19575c7c DS |
1450 | if (packs_to_drop && packs_to_drop->nr) { |
1451 | int drop_index = 0; | |
1452 | int missing_drops = 0; | |
1453 | ||
577dc496 DS |
1454 | for (i = 0; i < ctx.nr && drop_index < packs_to_drop->nr; i++) { |
1455 | int cmp = strcmp(ctx.info[i].pack_name, | |
19575c7c DS |
1456 | packs_to_drop->items[drop_index].string); |
1457 | ||
1458 | if (!cmp) { | |
1459 | drop_index++; | |
577dc496 | 1460 | ctx.info[i].expired = 1; |
19575c7c DS |
1461 | } else if (cmp > 0) { |
1462 | error(_("did not see pack-file %s to drop"), | |
1463 | packs_to_drop->items[drop_index].string); | |
1464 | drop_index++; | |
1465 | missing_drops++; | |
1466 | i--; | |
1467 | } else { | |
577dc496 | 1468 | ctx.info[i].expired = 0; |
19575c7c DS |
1469 | } |
1470 | } | |
1471 | ||
1472 | if (missing_drops) { | |
1473 | result = 1; | |
1474 | goto cleanup; | |
1475 | } | |
1476 | } | |
1477 | ||
d01bf2e6 DS |
1478 | /* |
1479 | * pack_perm stores a permutation between pack-int-ids from the | |
1480 | * previous multi-pack-index to the new one we are writing: | |
1481 | * | |
1482 | * pack_perm[old_id] = new_id | |
1483 | */ | |
7a3ada11 | 1484 | ALLOC_ARRAY(ctx.pack_perm, ctx.nr); |
577dc496 DS |
1485 | for (i = 0; i < ctx.nr; i++) { |
1486 | if (ctx.info[i].expired) { | |
19575c7c | 1487 | dropped_packs++; |
7a3ada11 | 1488 | ctx.pack_perm[ctx.info[i].orig_pack_int_id] = PACK_EXPIRED; |
19575c7c | 1489 | } else { |
7a3ada11 | 1490 | ctx.pack_perm[ctx.info[i].orig_pack_int_id] = i - dropped_packs; |
19575c7c | 1491 | } |
d01bf2e6 DS |
1492 | } |
1493 | ||
577dc496 DS |
1494 | for (i = 0; i < ctx.nr; i++) { |
1495 | if (!ctx.info[i].expired) | |
1496 | pack_name_concat_len += strlen(ctx.info[i].pack_name) + 1; | |
19575c7c | 1497 | } |
dba6175c | 1498 | |
9218c6a4 TB |
1499 | /* Check that the preferred pack wasn't expired (if given). */ |
1500 | if (preferred_pack_name) { | |
1501 | struct pack_info *preferred = bsearch(preferred_pack_name, | |
1502 | ctx.info, ctx.nr, | |
1503 | sizeof(*ctx.info), | |
1504 | idx_or_pack_name_cmp); | |
177c0d6e | 1505 | if (preferred) { |
9218c6a4 TB |
1506 | uint32_t perm = ctx.pack_perm[preferred->orig_pack_int_id]; |
1507 | if (perm == PACK_EXPIRED) | |
1508 | warning(_("preferred pack '%s' is expired"), | |
1509 | preferred_pack_name); | |
1510 | } | |
1511 | } | |
1512 | ||
dba6175c DS |
1513 | if (pack_name_concat_len % MIDX_CHUNK_ALIGNMENT) |
1514 | pack_name_concat_len += MIDX_CHUNK_ALIGNMENT - | |
1515 | (pack_name_concat_len % MIDX_CHUNK_ALIGNMENT); | |
1516 | ||
60980aed | 1517 | hold_lock_file_for_update(&lk, midx_name.buf, LOCK_DIE_ON_ERROR); |
acd71602 | 1518 | f = hashfd(get_lock_file_fd(&lk), get_lock_file_path(&lk)); |
fc59e748 | 1519 | |
577dc496 | 1520 | if (ctx.nr - dropped_packs == 0) { |
796d61cd DR |
1521 | error(_("no pack files to index.")); |
1522 | result = 1; | |
1523 | goto cleanup; | |
1524 | } | |
1525 | ||
eb57277b TB |
1526 | if (!ctx.entries_nr) { |
1527 | if (flags & MIDX_WRITE_BITMAP) | |
1528 | warning(_("refusing to write multi-pack .bitmap without any objects")); | |
1529 | flags &= ~(MIDX_WRITE_REV_INDEX | MIDX_WRITE_BITMAP); | |
1530 | } | |
1531 | ||
63a8f0e9 | 1532 | cf = init_chunkfile(f); |
32f3c541 | 1533 | |
63a8f0e9 DS |
1534 | add_chunk(cf, MIDX_CHUNKID_PACKNAMES, pack_name_concat_len, |
1535 | write_midx_pack_names); | |
1536 | add_chunk(cf, MIDX_CHUNKID_OIDFANOUT, MIDX_CHUNK_FANOUT_SIZE, | |
1537 | write_midx_oid_fanout); | |
1538 | add_chunk(cf, MIDX_CHUNKID_OIDLOOKUP, | |
2bc764c1 | 1539 | st_mult(ctx.entries_nr, the_hash_algo->rawsz), |
63a8f0e9 DS |
1540 | write_midx_oid_lookup); |
1541 | add_chunk(cf, MIDX_CHUNKID_OBJECTOFFSETS, | |
2bc764c1 | 1542 | st_mult(ctx.entries_nr, MIDX_CHUNK_OFFSET_WIDTH), |
63a8f0e9 | 1543 | write_midx_object_offsets); |
32f3c541 | 1544 | |
63a8f0e9 DS |
1545 | if (ctx.large_offsets_needed) |
1546 | add_chunk(cf, MIDX_CHUNKID_LARGEOFFSETS, | |
2bc764c1 TB |
1547 | st_mult(ctx.num_large_offsets, |
1548 | MIDX_CHUNK_LARGE_OFFSET_WIDTH), | |
63a8f0e9 | 1549 | write_midx_large_offsets); |
32f3c541 | 1550 | |
95e8383b TB |
1551 | if (flags & (MIDX_WRITE_REV_INDEX | MIDX_WRITE_BITMAP)) { |
1552 | ctx.pack_order = midx_pack_order(&ctx); | |
1553 | add_chunk(cf, MIDX_CHUNKID_REVINDEX, | |
2bc764c1 | 1554 | st_mult(ctx.entries_nr, sizeof(uint32_t)), |
95e8383b TB |
1555 | write_midx_revindex); |
1556 | } | |
1557 | ||
63a8f0e9 DS |
1558 | write_midx_header(f, get_num_chunks(cf), ctx.nr - dropped_packs); |
1559 | write_chunkfile(cf, &ctx); | |
fc59e748 | 1560 | |
020406ea NS |
1561 | finalize_hashfile(f, midx_hash, FSYNC_COMPONENT_PACK_METADATA, |
1562 | CSUM_FSYNC | CSUM_HASH_IN_STREAM); | |
63a8f0e9 | 1563 | free_chunkfile(cf); |
38ff7cab | 1564 | |
7f514b7a TB |
1565 | if (flags & MIDX_WRITE_REV_INDEX && |
1566 | git_env_bool("GIT_TEST_MIDX_WRITE_REV", 0)) | |
60980aed | 1567 | write_midx_reverse_index(midx_name.buf, midx_hash, &ctx); |
90b2bb71 | 1568 | |
c528e179 | 1569 | if (flags & MIDX_WRITE_BITMAP) { |
90b2bb71 DS |
1570 | struct packing_data pdata; |
1571 | struct commit **commits; | |
1572 | uint32_t commits_nr; | |
1573 | ||
1574 | if (!ctx.entries_nr) | |
1575 | BUG("cannot write a bitmap without any objects"); | |
1576 | ||
1577 | prepare_midx_packing_data(&pdata, &ctx); | |
1578 | ||
1579 | commits = find_commits_for_midx_bitmap(&commits_nr, refs_snapshot, &ctx); | |
1580 | ||
068fa54c DS |
1581 | /* |
1582 | * The previous steps translated the information from | |
1583 | * 'entries' into information suitable for constructing | |
1584 | * bitmaps. We no longer need that array, so clear it to | |
1585 | * reduce memory pressure. | |
1586 | */ | |
1587 | FREE_AND_NULL(ctx.entries); | |
1588 | ctx.entries_nr = 0; | |
1589 | ||
90b2bb71 DS |
1590 | if (write_midx_bitmap(midx_name.buf, midx_hash, &pdata, |
1591 | commits, commits_nr, ctx.pack_order, | |
51d1b69a | 1592 | flags) < 0) { |
c528e179 TB |
1593 | error(_("could not write multi-pack bitmap")); |
1594 | result = 1; | |
1595 | goto cleanup; | |
1596 | } | |
1597 | } | |
068fa54c DS |
1598 | /* |
1599 | * NOTE: Do not use ctx.entries beyond this point, since it might | |
1600 | * have been freed in the previous if block. | |
1601 | */ | |
c528e179 TB |
1602 | |
1603 | if (ctx.m) | |
1604 | close_object_store(the_repository->objects); | |
38ff7cab | 1605 | |
ae22e841 TB |
1606 | if (commit_lock_file(&lk) < 0) |
1607 | die_errno(_("could not write multi-pack-index")); | |
fc59e748 | 1608 | |
c528e179 | 1609 | clear_midx_files_ext(object_dir, ".bitmap", midx_hash); |
f5909d34 TB |
1610 | clear_midx_files_ext(object_dir, ".rev", midx_hash); |
1611 | ||
a40498a1 | 1612 | cleanup: |
577dc496 DS |
1613 | for (i = 0; i < ctx.nr; i++) { |
1614 | if (ctx.info[i].p) { | |
1615 | close_pack(ctx.info[i].p); | |
1616 | free(ctx.info[i].p); | |
396f2570 | 1617 | } |
577dc496 | 1618 | free(ctx.info[i].pack_name); |
396f2570 DS |
1619 | } |
1620 | ||
577dc496 | 1621 | free(ctx.info); |
31bda9a2 | 1622 | free(ctx.entries); |
7a3ada11 | 1623 | free(ctx.pack_perm); |
38ff7cab | 1624 | free(ctx.pack_order); |
60980aed | 1625 | strbuf_release(&midx_name); |
c528e179 | 1626 | |
2dcff525 TB |
1627 | trace2_region_leave("midx", "write_midx_internal", the_repository); |
1628 | ||
19575c7c DS |
1629 | return result; |
1630 | } | |
1631 | ||
9218c6a4 TB |
1632 | int write_midx_file(const char *object_dir, |
1633 | const char *preferred_pack_name, | |
08944d1c | 1634 | const char *refs_snapshot, |
9218c6a4 | 1635 | unsigned flags) |
19575c7c | 1636 | { |
56d863e9 | 1637 | return write_midx_internal(object_dir, NULL, NULL, preferred_pack_name, |
08944d1c | 1638 | refs_snapshot, flags); |
56d863e9 TB |
1639 | } |
1640 | ||
1641 | int write_midx_file_only(const char *object_dir, | |
1642 | struct string_list *packs_to_include, | |
1643 | const char *preferred_pack_name, | |
08944d1c | 1644 | const char *refs_snapshot, |
56d863e9 TB |
1645 | unsigned flags) |
1646 | { | |
1647 | return write_midx_internal(object_dir, packs_to_include, NULL, | |
08944d1c | 1648 | preferred_pack_name, refs_snapshot, flags); |
a3407730 | 1649 | } |
525e18c0 | 1650 | |
38ff7cab TB |
1651 | struct clear_midx_data { |
1652 | char *keep; | |
1653 | const char *ext; | |
1654 | }; | |
1655 | ||
be252d33 | 1656 | static void clear_midx_file_ext(const char *full_path, size_t full_path_len UNUSED, |
38ff7cab TB |
1657 | const char *file_name, void *_data) |
1658 | { | |
1659 | struct clear_midx_data *data = _data; | |
1660 | ||
1661 | if (!(starts_with(file_name, "multi-pack-index-") && | |
1662 | ends_with(file_name, data->ext))) | |
1663 | return; | |
1664 | if (data->keep && !strcmp(data->keep, file_name)) | |
1665 | return; | |
1666 | ||
1667 | if (unlink(full_path)) | |
1668 | die_errno(_("failed to remove %s"), full_path); | |
1669 | } | |
1670 | ||
426c00e4 | 1671 | static void clear_midx_files_ext(const char *object_dir, const char *ext, |
38ff7cab TB |
1672 | unsigned char *keep_hash) |
1673 | { | |
1674 | struct clear_midx_data data; | |
1675 | memset(&data, 0, sizeof(struct clear_midx_data)); | |
1676 | ||
1677 | if (keep_hash) | |
1678 | data.keep = xstrfmt("multi-pack-index-%s%s", | |
1679 | hash_to_hex(keep_hash), ext); | |
1680 | data.ext = ext; | |
1681 | ||
426c00e4 | 1682 | for_each_file_in_pack_dir(object_dir, |
38ff7cab TB |
1683 | clear_midx_file_ext, |
1684 | &data); | |
1685 | ||
1686 | free(data.keep); | |
a3407730 | 1687 | } |
525e18c0 | 1688 | |
1dcd9f20 | 1689 | void clear_midx_file(struct repository *r) |
525e18c0 | 1690 | { |
60980aed TB |
1691 | struct strbuf midx = STRBUF_INIT; |
1692 | ||
1693 | get_midx_filename(&midx, r->objects->odb->path); | |
1dcd9f20 DS |
1694 | |
1695 | if (r->objects && r->objects->multi_pack_index) { | |
1696 | close_midx(r->objects->multi_pack_index); | |
1697 | r->objects->multi_pack_index = NULL; | |
1698 | } | |
525e18c0 | 1699 | |
60980aed TB |
1700 | if (remove_path(midx.buf)) |
1701 | die(_("failed to clear multi-pack-index at %s"), midx.buf); | |
525e18c0 | 1702 | |
c528e179 | 1703 | clear_midx_files_ext(r->objects->odb->path, ".bitmap", NULL); |
426c00e4 | 1704 | clear_midx_files_ext(r->objects->odb->path, ".rev", NULL); |
38ff7cab | 1705 | |
60980aed | 1706 | strbuf_release(&midx); |
525e18c0 | 1707 | } |
56ee7ff1 DS |
1708 | |
1709 | static int verify_midx_error; | |
1710 | ||
48ca53ca | 1711 | __attribute__((format (printf, 1, 2))) |
d4bf1d88 DS |
1712 | static void midx_report(const char *fmt, ...) |
1713 | { | |
1714 | va_list ap; | |
1715 | verify_midx_error = 1; | |
1716 | va_start(ap, fmt); | |
1717 | vfprintf(stderr, fmt, ap); | |
1718 | fprintf(stderr, "\n"); | |
1719 | va_end(ap); | |
1720 | } | |
1721 | ||
5ae18df9 JH |
1722 | struct pair_pos_vs_id |
1723 | { | |
1724 | uint32_t pos; | |
1725 | uint32_t pack_int_id; | |
1726 | }; | |
1727 | ||
1728 | static int compare_pair_pos_vs_id(const void *_a, const void *_b) | |
1729 | { | |
1730 | struct pair_pos_vs_id *a = (struct pair_pos_vs_id *)_a; | |
1731 | struct pair_pos_vs_id *b = (struct pair_pos_vs_id *)_b; | |
1732 | ||
1733 | return b->pack_int_id - a->pack_int_id; | |
1734 | } | |
1735 | ||
430efb8a JH |
1736 | /* |
1737 | * Limit calls to display_progress() for performance reasons. | |
1738 | * The interval here was arbitrarily chosen. | |
1739 | */ | |
1740 | #define SPARSE_PROGRESS_INTERVAL (1 << 12) | |
1741 | #define midx_display_sparse_progress(progress, n) \ | |
1742 | do { \ | |
1743 | uint64_t _n = (n); \ | |
1744 | if ((_n & (SPARSE_PROGRESS_INTERVAL - 1)) == 0) \ | |
1745 | display_progress(progress, _n); \ | |
1746 | } while (0) | |
1747 | ||
efbc3aee | 1748 | int verify_midx_file(struct repository *r, const char *object_dir, unsigned flags) |
56ee7ff1 | 1749 | { |
5ae18df9 | 1750 | struct pair_pos_vs_id *pairs = NULL; |
d4bf1d88 | 1751 | uint32_t i; |
ad60096d | 1752 | struct progress *progress = NULL; |
56ee7ff1 DS |
1753 | struct multi_pack_index *m = load_multi_pack_index(object_dir, 1); |
1754 | verify_midx_error = 0; | |
1755 | ||
d9607542 DS |
1756 | if (!m) { |
1757 | int result = 0; | |
1758 | struct stat sb; | |
60980aed TB |
1759 | struct strbuf filename = STRBUF_INIT; |
1760 | ||
1761 | get_midx_filename(&filename, object_dir); | |
1762 | ||
1763 | if (!stat(filename.buf, &sb)) { | |
d9607542 DS |
1764 | error(_("multi-pack-index file exists, but failed to parse")); |
1765 | result = 1; | |
1766 | } | |
60980aed | 1767 | strbuf_release(&filename); |
d9607542 DS |
1768 | return result; |
1769 | } | |
56ee7ff1 | 1770 | |
f89ecf79 TB |
1771 | if (!midx_checksum_valid(m)) |
1772 | midx_report(_("incorrect checksum")); | |
1773 | ||
ad60096d | 1774 | if (flags & MIDX_PROGRESS) |
efdd2f0d | 1775 | progress = start_delayed_progress(_("Looking for referenced packfiles"), |
ad60096d | 1776 | m->num_packs); |
d4bf1d88 | 1777 | for (i = 0; i < m->num_packs; i++) { |
64404a24 | 1778 | if (prepare_midx_pack(r, m, i)) |
d4bf1d88 | 1779 | midx_report("failed to load pack in position %d", i); |
430efb8a JH |
1780 | |
1781 | display_progress(progress, i + 1); | |
d4bf1d88 | 1782 | } |
430efb8a | 1783 | stop_progress(&progress); |
d4bf1d88 | 1784 | |
2f23d3f3 DS |
1785 | for (i = 0; i < 255; i++) { |
1786 | uint32_t oid_fanout1 = ntohl(m->chunk_oid_fanout[i]); | |
1787 | uint32_t oid_fanout2 = ntohl(m->chunk_oid_fanout[i + 1]); | |
1788 | ||
1789 | if (oid_fanout1 > oid_fanout2) | |
1790 | midx_report(_("oid fanout out of order: fanout[%d] = %"PRIx32" > %"PRIx32" = fanout[%d]"), | |
1791 | i, oid_fanout1, oid_fanout2, i + 1); | |
1792 | } | |
1793 | ||
796d61cd DR |
1794 | if (m->num_objects == 0) { |
1795 | midx_report(_("the midx contains no oid")); | |
1796 | /* | |
1797 | * Remaining tests assume that we have objects, so we can | |
1798 | * return here. | |
1799 | */ | |
492cb394 | 1800 | goto cleanup; |
796d61cd DR |
1801 | } |
1802 | ||
ad60096d WB |
1803 | if (flags & MIDX_PROGRESS) |
1804 | progress = start_sparse_progress(_("Verifying OID order in multi-pack-index"), | |
1805 | m->num_objects - 1); | |
55c5648d DS |
1806 | for (i = 0; i < m->num_objects - 1; i++) { |
1807 | struct object_id oid1, oid2; | |
1808 | ||
1809 | nth_midxed_object_oid(&oid1, m, i); | |
1810 | nth_midxed_object_oid(&oid2, m, i + 1); | |
1811 | ||
1812 | if (oidcmp(&oid1, &oid2) >= 0) | |
1813 | midx_report(_("oid lookup out of order: oid[%d] = %s >= %s = oid[%d]"), | |
1814 | i, oid_to_hex(&oid1), oid_to_hex(&oid2), i + 1); | |
430efb8a JH |
1815 | |
1816 | midx_display_sparse_progress(progress, i + 1); | |
55c5648d | 1817 | } |
430efb8a | 1818 | stop_progress(&progress); |
55c5648d | 1819 | |
5ae18df9 JH |
1820 | /* |
1821 | * Create an array mapping each object to its packfile id. Sort it | |
1822 | * to group the objects by packfile. Use this permutation to visit | |
1823 | * each of the objects and only require 1 packfile to be open at a | |
1824 | * time. | |
1825 | */ | |
1826 | ALLOC_ARRAY(pairs, m->num_objects); | |
1827 | for (i = 0; i < m->num_objects; i++) { | |
1828 | pairs[i].pos = i; | |
1829 | pairs[i].pack_int_id = nth_midxed_pack_int_id(m, i); | |
1830 | } | |
1831 | ||
ad60096d WB |
1832 | if (flags & MIDX_PROGRESS) |
1833 | progress = start_sparse_progress(_("Sorting objects by packfile"), | |
1834 | m->num_objects); | |
5ae18df9 JH |
1835 | display_progress(progress, 0); /* TODO: Measure QSORT() progress */ |
1836 | QSORT(pairs, m->num_objects, compare_pair_pos_vs_id); | |
1837 | stop_progress(&progress); | |
1838 | ||
ad60096d WB |
1839 | if (flags & MIDX_PROGRESS) |
1840 | progress = start_sparse_progress(_("Verifying object offsets"), m->num_objects); | |
cc6af73c DS |
1841 | for (i = 0; i < m->num_objects; i++) { |
1842 | struct object_id oid; | |
1843 | struct pack_entry e; | |
1844 | off_t m_offset, p_offset; | |
1845 | ||
5ae18df9 JH |
1846 | if (i > 0 && pairs[i-1].pack_int_id != pairs[i].pack_int_id && |
1847 | m->packs[pairs[i-1].pack_int_id]) | |
1848 | { | |
1849 | close_pack_fd(m->packs[pairs[i-1].pack_int_id]); | |
1850 | close_pack_index(m->packs[pairs[i-1].pack_int_id]); | |
1851 | } | |
1852 | ||
1853 | nth_midxed_object_oid(&oid, m, pairs[i].pos); | |
1854 | ||
64404a24 | 1855 | if (!fill_midx_entry(r, &oid, &e, m)) { |
cc6af73c | 1856 | midx_report(_("failed to load pack entry for oid[%d] = %s"), |
5ae18df9 | 1857 | pairs[i].pos, oid_to_hex(&oid)); |
cc6af73c DS |
1858 | continue; |
1859 | } | |
1860 | ||
1861 | if (open_pack_index(e.p)) { | |
1862 | midx_report(_("failed to load pack-index for packfile %s"), | |
1863 | e.p->pack_name); | |
1864 | break; | |
1865 | } | |
1866 | ||
1867 | m_offset = e.offset; | |
1868 | p_offset = find_pack_entry_one(oid.hash, e.p); | |
1869 | ||
1870 | if (m_offset != p_offset) | |
1871 | midx_report(_("incorrect object offset for oid[%d] = %s: %"PRIx64" != %"PRIx64), | |
5ae18df9 | 1872 | pairs[i].pos, oid_to_hex(&oid), m_offset, p_offset); |
144d7033 | 1873 | |
430efb8a | 1874 | midx_display_sparse_progress(progress, i + 1); |
cc6af73c | 1875 | } |
144d7033 | 1876 | stop_progress(&progress); |
cc6af73c | 1877 | |
492cb394 | 1878 | cleanup: |
5ae18df9 | 1879 | free(pairs); |
492cb394 | 1880 | close_midx(m); |
5ae18df9 | 1881 | |
56ee7ff1 DS |
1882 | return verify_midx_error; |
1883 | } | |
cff97116 | 1884 | |
efbc3aee | 1885 | int expire_midx_packs(struct repository *r, const char *object_dir, unsigned flags) |
cff97116 | 1886 | { |
19575c7c DS |
1887 | uint32_t i, *count, result = 0; |
1888 | struct string_list packs_to_drop = STRING_LIST_INIT_DUP; | |
98926e0d | 1889 | struct multi_pack_index *m = lookup_multi_pack_index(r, object_dir); |
8dc18f89 | 1890 | struct progress *progress = NULL; |
19575c7c DS |
1891 | |
1892 | if (!m) | |
1893 | return 0; | |
1894 | ||
ca56dadb | 1895 | CALLOC_ARRAY(count, m->num_packs); |
8dc18f89 WB |
1896 | |
1897 | if (flags & MIDX_PROGRESS) | |
efdd2f0d | 1898 | progress = start_delayed_progress(_("Counting referenced objects"), |
8dc18f89 | 1899 | m->num_objects); |
19575c7c DS |
1900 | for (i = 0; i < m->num_objects; i++) { |
1901 | int pack_int_id = nth_midxed_pack_int_id(m, i); | |
1902 | count[pack_int_id]++; | |
8dc18f89 | 1903 | display_progress(progress, i + 1); |
19575c7c | 1904 | } |
8dc18f89 | 1905 | stop_progress(&progress); |
19575c7c | 1906 | |
8dc18f89 | 1907 | if (flags & MIDX_PROGRESS) |
efdd2f0d | 1908 | progress = start_delayed_progress(_("Finding and deleting unreferenced packfiles"), |
8dc18f89 | 1909 | m->num_packs); |
19575c7c DS |
1910 | for (i = 0; i < m->num_packs; i++) { |
1911 | char *pack_name; | |
8dc18f89 | 1912 | display_progress(progress, i + 1); |
19575c7c DS |
1913 | |
1914 | if (count[i]) | |
1915 | continue; | |
1916 | ||
1917 | if (prepare_midx_pack(r, m, i)) | |
1918 | continue; | |
1919 | ||
757d4579 | 1920 | if (m->packs[i]->pack_keep || m->packs[i]->is_cruft) |
19575c7c DS |
1921 | continue; |
1922 | ||
1923 | pack_name = xstrdup(m->packs[i]->pack_name); | |
1924 | close_pack(m->packs[i]); | |
1925 | ||
1926 | string_list_insert(&packs_to_drop, m->pack_names[i]); | |
1927 | unlink_pack_path(pack_name, 0); | |
1928 | free(pack_name); | |
1929 | } | |
8dc18f89 | 1930 | stop_progress(&progress); |
19575c7c DS |
1931 | |
1932 | free(count); | |
1933 | ||
98926e0d | 1934 | if (packs_to_drop.nr) |
08944d1c | 1935 | result = write_midx_internal(object_dir, NULL, &packs_to_drop, NULL, NULL, flags); |
19575c7c DS |
1936 | |
1937 | string_list_clear(&packs_to_drop, 0); | |
98926e0d | 1938 | |
19575c7c | 1939 | return result; |
cff97116 | 1940 | } |
2af890bb | 1941 | |
ce1e4a10 DS |
1942 | struct repack_info { |
1943 | timestamp_t mtime; | |
1944 | uint32_t referenced_objects; | |
1945 | uint32_t pack_int_id; | |
1946 | }; | |
1947 | ||
1948 | static int compare_by_mtime(const void *a_, const void *b_) | |
2af890bb | 1949 | { |
ce1e4a10 DS |
1950 | const struct repack_info *a, *b; |
1951 | ||
1952 | a = (const struct repack_info *)a_; | |
1953 | b = (const struct repack_info *)b_; | |
1954 | ||
1955 | if (a->mtime < b->mtime) | |
1956 | return -1; | |
1957 | if (a->mtime > b->mtime) | |
1958 | return 1; | |
1959 | return 0; | |
1960 | } | |
1961 | ||
3ce4ca0a DS |
1962 | static int fill_included_packs_all(struct repository *r, |
1963 | struct multi_pack_index *m, | |
ce1e4a10 DS |
1964 | unsigned char *include_pack) |
1965 | { | |
3ce4ca0a DS |
1966 | uint32_t i, count = 0; |
1967 | int pack_kept_objects = 0; | |
1968 | ||
1969 | repo_config_get_bool(r, "repack.packkeptobjects", &pack_kept_objects); | |
1970 | ||
1971 | for (i = 0; i < m->num_packs; i++) { | |
1972 | if (prepare_midx_pack(r, m, i)) | |
1973 | continue; | |
1974 | if (!pack_kept_objects && m->packs[i]->pack_keep) | |
1975 | continue; | |
d9f77214 TB |
1976 | if (m->packs[i]->is_cruft) |
1977 | continue; | |
ce1e4a10 | 1978 | |
ce1e4a10 | 1979 | include_pack[i] = 1; |
3ce4ca0a DS |
1980 | count++; |
1981 | } | |
ce1e4a10 | 1982 | |
3ce4ca0a | 1983 | return count < 2; |
ce1e4a10 DS |
1984 | } |
1985 | ||
1986 | static int fill_included_packs_batch(struct repository *r, | |
1987 | struct multi_pack_index *m, | |
1988 | unsigned char *include_pack, | |
1989 | size_t batch_size) | |
1990 | { | |
1991 | uint32_t i, packs_to_repack; | |
1992 | size_t total_size; | |
cb6c48cb | 1993 | struct repack_info *pack_info; |
3ce4ca0a DS |
1994 | int pack_kept_objects = 0; |
1995 | ||
cb6c48cb TB |
1996 | CALLOC_ARRAY(pack_info, m->num_packs); |
1997 | ||
3ce4ca0a | 1998 | repo_config_get_bool(r, "repack.packkeptobjects", &pack_kept_objects); |
ce1e4a10 DS |
1999 | |
2000 | for (i = 0; i < m->num_packs; i++) { | |
2001 | pack_info[i].pack_int_id = i; | |
2002 | ||
2003 | if (prepare_midx_pack(r, m, i)) | |
2004 | continue; | |
2005 | ||
2006 | pack_info[i].mtime = m->packs[i]->mtime; | |
2007 | } | |
2008 | ||
0a8e5614 | 2009 | for (i = 0; i < m->num_objects; i++) { |
ce1e4a10 DS |
2010 | uint32_t pack_int_id = nth_midxed_pack_int_id(m, i); |
2011 | pack_info[pack_int_id].referenced_objects++; | |
2012 | } | |
2013 | ||
2014 | QSORT(pack_info, m->num_packs, compare_by_mtime); | |
2015 | ||
2016 | total_size = 0; | |
2017 | packs_to_repack = 0; | |
2018 | for (i = 0; total_size < batch_size && i < m->num_packs; i++) { | |
2019 | int pack_int_id = pack_info[i].pack_int_id; | |
2020 | struct packed_git *p = m->packs[pack_int_id]; | |
2021 | size_t expected_size; | |
2022 | ||
2023 | if (!p) | |
2024 | continue; | |
3ce4ca0a DS |
2025 | if (!pack_kept_objects && p->pack_keep) |
2026 | continue; | |
b62ad568 TB |
2027 | if (p->is_cruft) |
2028 | continue; | |
ce1e4a10 DS |
2029 | if (open_pack_index(p) || !p->num_objects) |
2030 | continue; | |
2031 | ||
d67609bd TB |
2032 | expected_size = st_mult(p->pack_size, |
2033 | pack_info[i].referenced_objects); | |
ce1e4a10 DS |
2034 | expected_size /= p->num_objects; |
2035 | ||
2036 | if (expected_size >= batch_size) | |
2037 | continue; | |
2038 | ||
2039 | packs_to_repack++; | |
2040 | total_size += expected_size; | |
2041 | include_pack[pack_int_id] = 1; | |
2042 | } | |
2043 | ||
2044 | free(pack_info); | |
2045 | ||
1eb22c7d | 2046 | if (packs_to_repack < 2) |
ce1e4a10 DS |
2047 | return 1; |
2048 | ||
2af890bb DS |
2049 | return 0; |
2050 | } | |
ce1e4a10 | 2051 | |
efbc3aee | 2052 | int midx_repack(struct repository *r, const char *object_dir, size_t batch_size, unsigned flags) |
ce1e4a10 DS |
2053 | { |
2054 | int result = 0; | |
2055 | uint32_t i; | |
2056 | unsigned char *include_pack; | |
2057 | struct child_process cmd = CHILD_PROCESS_INIT; | |
6af3b00a | 2058 | FILE *cmd_in; |
ce1e4a10 | 2059 | struct strbuf base_name = STRBUF_INIT; |
c0f1f9de | 2060 | struct multi_pack_index *m = lookup_multi_pack_index(r, object_dir); |
ce1e4a10 | 2061 | |
e11d86de SLN |
2062 | /* |
2063 | * When updating the default for these configuration | |
2064 | * variables in builtin/repack.c, these must be adjusted | |
2065 | * to match. | |
2066 | */ | |
2067 | int delta_base_offset = 1; | |
2068 | int use_delta_islands = 0; | |
2069 | ||
ce1e4a10 DS |
2070 | if (!m) |
2071 | return 0; | |
2072 | ||
ca56dadb | 2073 | CALLOC_ARRAY(include_pack, m->num_packs); |
ce1e4a10 DS |
2074 | |
2075 | if (batch_size) { | |
2076 | if (fill_included_packs_batch(r, m, include_pack, batch_size)) | |
2077 | goto cleanup; | |
3ce4ca0a | 2078 | } else if (fill_included_packs_all(r, m, include_pack)) |
ce1e4a10 DS |
2079 | goto cleanup; |
2080 | ||
e11d86de SLN |
2081 | repo_config_get_bool(r, "repack.usedeltabaseoffset", &delta_base_offset); |
2082 | repo_config_get_bool(r, "repack.usedeltaislands", &use_delta_islands); | |
2083 | ||
c972bf4c | 2084 | strvec_push(&cmd.args, "pack-objects"); |
ce1e4a10 DS |
2085 | |
2086 | strbuf_addstr(&base_name, object_dir); | |
2087 | strbuf_addstr(&base_name, "/pack/pack"); | |
c972bf4c | 2088 | strvec_push(&cmd.args, base_name.buf); |
64d80e7d | 2089 | |
e11d86de | 2090 | if (delta_base_offset) |
c972bf4c | 2091 | strvec_push(&cmd.args, "--delta-base-offset"); |
e11d86de | 2092 | if (use_delta_islands) |
c972bf4c | 2093 | strvec_push(&cmd.args, "--delta-islands"); |
e11d86de | 2094 | |
64d80e7d | 2095 | if (flags & MIDX_PROGRESS) |
c972bf4c | 2096 | strvec_push(&cmd.args, "--progress"); |
64d80e7d | 2097 | else |
c972bf4c | 2098 | strvec_push(&cmd.args, "-q"); |
64d80e7d | 2099 | |
ce1e4a10 DS |
2100 | strbuf_release(&base_name); |
2101 | ||
2102 | cmd.git_cmd = 1; | |
2103 | cmd.in = cmd.out = -1; | |
2104 | ||
2105 | if (start_command(&cmd)) { | |
2106 | error(_("could not start pack-objects")); | |
2107 | result = 1; | |
2108 | goto cleanup; | |
2109 | } | |
2110 | ||
6af3b00a RS |
2111 | cmd_in = xfdopen(cmd.in, "w"); |
2112 | ||
ce1e4a10 DS |
2113 | for (i = 0; i < m->num_objects; i++) { |
2114 | struct object_id oid; | |
2115 | uint32_t pack_int_id = nth_midxed_pack_int_id(m, i); | |
2116 | ||
2117 | if (!include_pack[pack_int_id]) | |
2118 | continue; | |
2119 | ||
2120 | nth_midxed_object_oid(&oid, m, i); | |
6af3b00a | 2121 | fprintf(cmd_in, "%s\n", oid_to_hex(&oid)); |
ce1e4a10 | 2122 | } |
6af3b00a | 2123 | fclose(cmd_in); |
ce1e4a10 DS |
2124 | |
2125 | if (finish_command(&cmd)) { | |
2126 | error(_("could not finish pack-objects")); | |
2127 | result = 1; | |
2128 | goto cleanup; | |
2129 | } | |
2130 | ||
08944d1c | 2131 | result = write_midx_internal(object_dir, NULL, NULL, NULL, NULL, flags); |
ce1e4a10 DS |
2132 | |
2133 | cleanup: | |
ce1e4a10 DS |
2134 | free(include_pack); |
2135 | return result; | |
2136 | } |