]>
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" | |
b1e33330 | 24 | #include "pack-revindex.h" |
a3407730 | 25 | |
fc59e748 DS |
26 | #define MIDX_SIGNATURE 0x4d494458 /* "MIDX" */ |
27 | #define MIDX_VERSION 1 | |
4d80560c DS |
28 | #define MIDX_BYTE_FILE_VERSION 4 |
29 | #define MIDX_BYTE_HASH_VERSION 5 | |
30 | #define MIDX_BYTE_NUM_CHUNKS 6 | |
31 | #define MIDX_BYTE_NUM_PACKS 8 | |
fc59e748 | 32 | #define MIDX_HEADER_SIZE 12 |
aaa95dfa | 33 | #define MIDX_MIN_SIZE (MIDX_HEADER_SIZE + the_hash_algo->rawsz) |
fc59e748 | 34 | |
32f3c541 DS |
35 | #define MIDX_CHUNK_ALIGNMENT 4 |
36 | #define MIDX_CHUNKID_PACKNAMES 0x504e414d /* "PNAM" */ | |
5f5ccd95 | 37 | #define MIDX_CHUNKID_BITMAPPEDPACKS 0x42544d50 /* "BTMP" */ |
d7cacf29 | 38 | #define MIDX_CHUNKID_OIDFANOUT 0x4f494446 /* "OIDF" */ |
0d5b3a5e | 39 | #define MIDX_CHUNKID_OIDLOOKUP 0x4f49444c /* "OIDL" */ |
662148c4 DS |
40 | #define MIDX_CHUNKID_OBJECTOFFSETS 0x4f4f4646 /* "OOFF" */ |
41 | #define MIDX_CHUNKID_LARGEOFFSETS 0x4c4f4646 /* "LOFF" */ | |
95e8383b | 42 | #define MIDX_CHUNKID_REVINDEX 0x52494458 /* "RIDX" */ |
d7cacf29 | 43 | #define MIDX_CHUNK_FANOUT_SIZE (sizeof(uint32_t) * 256) |
662148c4 DS |
44 | #define MIDX_CHUNK_OFFSET_WIDTH (2 * sizeof(uint32_t)) |
45 | #define MIDX_CHUNK_LARGE_OFFSET_WIDTH (sizeof(uint64_t)) | |
5f5ccd95 | 46 | #define MIDX_CHUNK_BITMAPPED_PACKS_WIDTH (2 * sizeof(uint32_t)) |
662148c4 | 47 | #define MIDX_LARGE_OFFSET_NEEDED 0x80000000 |
32f3c541 | 48 | |
19575c7c DS |
49 | #define PACK_EXPIRED UINT_MAX |
50 | ||
0f533c72 | 51 | const unsigned char *get_midx_checksum(struct multi_pack_index *m) |
f894081d TB |
52 | { |
53 | return m->data + m->data_len - the_hash_algo->rawsz; | |
54 | } | |
55 | ||
60980aed | 56 | void get_midx_filename(struct strbuf *out, const char *object_dir) |
fc59e748 | 57 | { |
60980aed | 58 | strbuf_addf(out, "%s/pack/multi-pack-index", object_dir); |
fc59e748 DS |
59 | } |
60 | ||
60980aed | 61 | void get_midx_rev_filename(struct strbuf *out, struct multi_pack_index *m) |
f894081d | 62 | { |
60980aed TB |
63 | get_midx_filename(out, m->object_dir); |
64 | strbuf_addf(out, "-%s.rev", hash_to_hex(get_midx_checksum(m))); | |
f894081d TB |
65 | } |
66 | ||
6ab3b8b8 DS |
67 | static int midx_read_oid_fanout(const unsigned char *chunk_start, |
68 | size_t chunk_size, void *data) | |
69 | { | |
9d78fb0e | 70 | int i; |
6ab3b8b8 DS |
71 | struct multi_pack_index *m = data; |
72 | m->chunk_oid_fanout = (uint32_t *)chunk_start; | |
73 | ||
74 | if (chunk_size != 4 * 256) { | |
75 | error(_("multi-pack-index OID fanout is of the wrong size")); | |
76 | return 1; | |
77 | } | |
9d78fb0e JK |
78 | for (i = 0; i < 255; i++) { |
79 | uint32_t oid_fanout1 = ntohl(m->chunk_oid_fanout[i]); | |
80 | uint32_t oid_fanout2 = ntohl(m->chunk_oid_fanout[i+1]); | |
81 | ||
82 | if (oid_fanout1 > oid_fanout2) { | |
83 | error(_("oid fanout out of order: fanout[%d] = %"PRIx32" > %"PRIx32" = fanout[%d]"), | |
84 | i, oid_fanout1, oid_fanout2, i + 1); | |
85 | return 1; | |
86 | } | |
87 | } | |
fc926567 JK |
88 | m->num_objects = ntohl(m->chunk_oid_fanout[255]); |
89 | return 0; | |
90 | } | |
91 | ||
92 | static int midx_read_oid_lookup(const unsigned char *chunk_start, | |
93 | size_t chunk_size, void *data) | |
94 | { | |
95 | struct multi_pack_index *m = data; | |
96 | m->chunk_oid_lookup = chunk_start; | |
97 | ||
98 | if (chunk_size != st_mult(m->hash_len, m->num_objects)) { | |
99 | error(_("multi-pack-index OID lookup chunk is the wrong size")); | |
100 | return 1; | |
101 | } | |
6ab3b8b8 DS |
102 | return 0; |
103 | } | |
104 | ||
0924869b JK |
105 | static int midx_read_object_offsets(const unsigned char *chunk_start, |
106 | size_t chunk_size, void *data) | |
107 | { | |
108 | struct multi_pack_index *m = data; | |
109 | m->chunk_object_offsets = chunk_start; | |
110 | ||
111 | if (chunk_size != st_mult(m->num_objects, MIDX_CHUNK_OFFSET_WIDTH)) { | |
112 | error(_("multi-pack-index object offset chunk is the wrong size")); | |
113 | return 1; | |
114 | } | |
115 | return 0; | |
116 | } | |
117 | ||
2cf489a3 | 118 | struct multi_pack_index *load_multi_pack_index(const char *object_dir, int local) |
4d80560c DS |
119 | { |
120 | struct multi_pack_index *m = NULL; | |
121 | int fd; | |
122 | struct stat st; | |
123 | size_t midx_size; | |
124 | void *midx_map = NULL; | |
125 | uint32_t hash_version; | |
60980aed | 126 | struct strbuf midx_name = STRBUF_INIT; |
32f3c541 | 127 | uint32_t i; |
3227565c | 128 | const char *cur_pack_name; |
6ab3b8b8 | 129 | struct chunkfile *cf = NULL; |
4d80560c | 130 | |
60980aed TB |
131 | get_midx_filename(&midx_name, object_dir); |
132 | ||
133 | fd = git_open(midx_name.buf); | |
4d80560c DS |
134 | |
135 | if (fd < 0) | |
136 | goto cleanup_fail; | |
137 | if (fstat(fd, &st)) { | |
60980aed | 138 | error_errno(_("failed to read %s"), midx_name.buf); |
4d80560c DS |
139 | goto cleanup_fail; |
140 | } | |
141 | ||
142 | midx_size = xsize_t(st.st_size); | |
143 | ||
144 | if (midx_size < MIDX_MIN_SIZE) { | |
60980aed | 145 | error(_("multi-pack-index file %s is too small"), midx_name.buf); |
4d80560c DS |
146 | goto cleanup_fail; |
147 | } | |
148 | ||
60980aed | 149 | strbuf_release(&midx_name); |
4d80560c DS |
150 | |
151 | midx_map = xmmap(NULL, midx_size, PROT_READ, MAP_PRIVATE, fd, 0); | |
6c7ff7cf | 152 | close(fd); |
4d80560c | 153 | |
577314ca | 154 | FLEX_ALLOC_STR(m, object_dir, object_dir); |
4d80560c DS |
155 | m->data = midx_map; |
156 | m->data_len = midx_size; | |
2cf489a3 | 157 | m->local = local; |
4d80560c DS |
158 | |
159 | m->signature = get_be32(m->data); | |
53ad0407 DS |
160 | if (m->signature != MIDX_SIGNATURE) |
161 | die(_("multi-pack-index signature 0x%08x does not match signature 0x%08x"), | |
4d80560c | 162 | m->signature, MIDX_SIGNATURE); |
4d80560c DS |
163 | |
164 | m->version = m->data[MIDX_BYTE_FILE_VERSION]; | |
53ad0407 DS |
165 | if (m->version != MIDX_VERSION) |
166 | die(_("multi-pack-index version %d not recognized"), | |
4d80560c | 167 | m->version); |
4d80560c DS |
168 | |
169 | hash_version = m->data[MIDX_BYTE_HASH_VERSION]; | |
d9fef9d9 | 170 | if (hash_version != oid_version(the_hash_algo)) { |
d9607542 | 171 | error(_("multi-pack-index hash version %u does not match version %u"), |
d9fef9d9 | 172 | hash_version, oid_version(the_hash_algo)); |
d9607542 DS |
173 | goto cleanup_fail; |
174 | } | |
aaa95dfa | 175 | m->hash_len = the_hash_algo->rawsz; |
4d80560c DS |
176 | |
177 | m->num_chunks = m->data[MIDX_BYTE_NUM_CHUNKS]; | |
178 | ||
179 | m->num_packs = get_be32(m->data + MIDX_BYTE_NUM_PACKS); | |
180 | ||
b1e33330 TB |
181 | m->preferred_pack_idx = -1; |
182 | ||
6ab3b8b8 | 183 | cf = init_chunkfile(NULL); |
32f3c541 | 184 | |
6ab3b8b8 | 185 | if (read_table_of_contents(cf, m->data, midx_size, |
c9b9fefc JK |
186 | MIDX_HEADER_SIZE, m->num_chunks, |
187 | MIDX_CHUNK_ALIGNMENT)) | |
6ab3b8b8 | 188 | goto cleanup_fail; |
32f3c541 | 189 | |
72a9a082 | 190 | if (pair_chunk(cf, MIDX_CHUNKID_PACKNAMES, &m->chunk_pack_names, &m->chunk_pack_names_len)) |
e3c96003 JK |
191 | die(_("multi-pack-index required pack-name chunk missing or corrupted")); |
192 | if (read_chunk(cf, MIDX_CHUNKID_OIDFANOUT, midx_read_oid_fanout, m)) | |
193 | die(_("multi-pack-index required OID fanout chunk missing or corrupted")); | |
fc926567 | 194 | if (read_chunk(cf, MIDX_CHUNKID_OIDLOOKUP, midx_read_oid_lookup, m)) |
e3c96003 | 195 | die(_("multi-pack-index required OID lookup chunk missing or corrupted")); |
0924869b | 196 | if (read_chunk(cf, MIDX_CHUNKID_OBJECTOFFSETS, midx_read_object_offsets, m)) |
e3c96003 | 197 | die(_("multi-pack-index required object offsets chunk missing or corrupted")); |
32f3c541 | 198 | |
2abd56e9 JK |
199 | pair_chunk(cf, MIDX_CHUNKID_LARGEOFFSETS, &m->chunk_large_offsets, |
200 | &m->chunk_large_offsets_len); | |
5f5ccd95 TB |
201 | pair_chunk(cf, MIDX_CHUNKID_BITMAPPEDPACKS, |
202 | (const unsigned char **)&m->chunk_bitmapped_packs, | |
203 | &m->chunk_bitmapped_packs_len); | |
6ab3b8b8 | 204 | |
7f514b7a | 205 | if (git_env_bool("GIT_TEST_MIDX_READ_RIDX", 1)) |
c0fe9b2d JK |
206 | pair_chunk(cf, MIDX_CHUNKID_REVINDEX, &m->chunk_revindex, |
207 | &m->chunk_revindex_len); | |
7f514b7a | 208 | |
ca56dadb RS |
209 | CALLOC_ARRAY(m->pack_names, m->num_packs); |
210 | CALLOC_ARRAY(m->packs, m->num_packs); | |
3227565c DS |
211 | |
212 | cur_pack_name = (const char *)m->chunk_pack_names; | |
213 | for (i = 0; i < m->num_packs; i++) { | |
72a9a082 JK |
214 | const char *end; |
215 | size_t avail = m->chunk_pack_names_len - | |
216 | (cur_pack_name - (const char *)m->chunk_pack_names); | |
217 | ||
3227565c DS |
218 | m->pack_names[i] = cur_pack_name; |
219 | ||
72a9a082 JK |
220 | end = memchr(cur_pack_name, '\0', avail); |
221 | if (!end) | |
222 | die(_("multi-pack-index pack-name chunk is too short")); | |
223 | cur_pack_name = end + 1; | |
3227565c | 224 | |
8e72a3c3 DS |
225 | if (i && strcmp(m->pack_names[i], m->pack_names[i - 1]) <= 0) |
226 | die(_("multi-pack-index pack names out of order: '%s' before '%s'"), | |
3227565c DS |
227 | m->pack_names[i - 1], |
228 | m->pack_names[i]); | |
3227565c DS |
229 | } |
230 | ||
d829223a JH |
231 | trace2_data_intmax("midx", the_repository, "load/num_packs", m->num_packs); |
232 | trace2_data_intmax("midx", the_repository, "load/num_objects", m->num_objects); | |
233 | ||
692305ec | 234 | free_chunkfile(cf); |
4d80560c DS |
235 | return m; |
236 | ||
237 | cleanup_fail: | |
238 | free(m); | |
60980aed | 239 | strbuf_release(&midx_name); |
692305ec | 240 | free_chunkfile(cf); |
4d80560c DS |
241 | if (midx_map) |
242 | munmap(midx_map, midx_size); | |
243 | if (0 <= fd) | |
244 | close(fd); | |
245 | return NULL; | |
246 | } | |
247 | ||
1dcd9f20 | 248 | void close_midx(struct multi_pack_index *m) |
a40498a1 DS |
249 | { |
250 | uint32_t i; | |
1dcd9f20 DS |
251 | |
252 | if (!m) | |
253 | return; | |
254 | ||
9bb6c2e5 TB |
255 | close_midx(m->next); |
256 | ||
a40498a1 | 257 | munmap((unsigned char *)m->data, m->data_len); |
a40498a1 DS |
258 | |
259 | for (i = 0; i < m->num_packs; i++) { | |
af96fe33 DS |
260 | if (m->packs[i]) |
261 | m->packs[i]->multi_pack_index = 0; | |
a40498a1 DS |
262 | } |
263 | FREE_AND_NULL(m->packs); | |
264 | FREE_AND_NULL(m->pack_names); | |
9bb6c2e5 | 265 | free(m); |
a40498a1 DS |
266 | } |
267 | ||
64404a24 | 268 | int prepare_midx_pack(struct repository *r, struct multi_pack_index *m, uint32_t pack_int_id) |
3715a633 DS |
269 | { |
270 | struct strbuf pack_name = STRBUF_INIT; | |
af96fe33 | 271 | struct packed_git *p; |
3715a633 DS |
272 | |
273 | if (pack_int_id >= m->num_packs) | |
d355e46a | 274 | die(_("bad pack-int-id: %u (%u total packs)"), |
cc6af73c | 275 | pack_int_id, m->num_packs); |
3715a633 DS |
276 | |
277 | if (m->packs[pack_int_id]) | |
278 | return 0; | |
279 | ||
280 | strbuf_addf(&pack_name, "%s/pack/%s", m->object_dir, | |
281 | m->pack_names[pack_int_id]); | |
282 | ||
af96fe33 | 283 | p = add_packed_git(pack_name.buf, pack_name.len, m->local); |
3715a633 | 284 | strbuf_release(&pack_name); |
af96fe33 DS |
285 | |
286 | if (!p) | |
287 | return 1; | |
288 | ||
289 | p->multi_pack_index = 1; | |
290 | m->packs[pack_int_id] = p; | |
291 | install_packed_git(r, p); | |
292 | list_add_tail(&p->mru, &r->objects->packed_git_mru); | |
293 | ||
294 | return 0; | |
3715a633 DS |
295 | } |
296 | ||
5f5ccd95 TB |
297 | int nth_bitmapped_pack(struct repository *r, struct multi_pack_index *m, |
298 | struct bitmapped_pack *bp, uint32_t pack_int_id) | |
299 | { | |
300 | if (!m->chunk_bitmapped_packs) | |
301 | return error(_("MIDX does not contain the BTMP chunk")); | |
302 | ||
303 | if (prepare_midx_pack(r, m, pack_int_id)) | |
304 | return error(_("could not load bitmapped pack %"PRIu32), pack_int_id); | |
305 | ||
306 | bp->p = m->packs[pack_int_id]; | |
307 | bp->bitmap_pos = get_be32((char *)m->chunk_bitmapped_packs + | |
308 | MIDX_CHUNK_BITMAPPED_PACKS_WIDTH * pack_int_id); | |
309 | bp->bitmap_nr = get_be32((char *)m->chunk_bitmapped_packs + | |
310 | MIDX_CHUNK_BITMAPPED_PACKS_WIDTH * pack_int_id + | |
311 | sizeof(uint32_t)); | |
312 | bp->pack_int_id = pack_int_id; | |
313 | ||
314 | return 0; | |
315 | } | |
316 | ||
3715a633 DS |
317 | int bsearch_midx(const struct object_id *oid, struct multi_pack_index *m, uint32_t *result) |
318 | { | |
319 | return bsearch_hash(oid->hash, m->chunk_oid_fanout, m->chunk_oid_lookup, | |
aaa95dfa | 320 | the_hash_algo->rawsz, result); |
3715a633 DS |
321 | } |
322 | ||
8aac67a1 DS |
323 | struct object_id *nth_midxed_object_oid(struct object_id *oid, |
324 | struct multi_pack_index *m, | |
325 | uint32_t n) | |
326 | { | |
327 | if (n >= m->num_objects) | |
328 | return NULL; | |
329 | ||
c2b24ede | 330 | oidread(oid, m->chunk_oid_lookup + st_mult(m->hash_len, n)); |
8aac67a1 DS |
331 | return oid; |
332 | } | |
333 | ||
62f2c1b5 | 334 | off_t nth_midxed_offset(struct multi_pack_index *m, uint32_t pos) |
3715a633 DS |
335 | { |
336 | const unsigned char *offset_data; | |
337 | uint32_t offset32; | |
338 | ||
329fac3a | 339 | offset_data = m->chunk_object_offsets + (off_t)pos * MIDX_CHUNK_OFFSET_WIDTH; |
3715a633 DS |
340 | offset32 = get_be32(offset_data + sizeof(uint32_t)); |
341 | ||
342 | if (m->chunk_large_offsets && offset32 & MIDX_LARGE_OFFSET_NEEDED) { | |
d8ac9ee1 | 343 | if (sizeof(off_t) < sizeof(uint64_t)) |
3715a633 DS |
344 | die(_("multi-pack-index stores a 64-bit offset, but off_t is too small")); |
345 | ||
346 | offset32 ^= MIDX_LARGE_OFFSET_NEEDED; | |
2abd56e9 JK |
347 | if (offset32 >= m->chunk_large_offsets_len / sizeof(uint64_t)) |
348 | die(_("multi-pack-index large offset out of bounds")); | |
349 | return get_be64(m->chunk_large_offsets + sizeof(uint64_t) * offset32); | |
3715a633 DS |
350 | } |
351 | ||
352 | return offset32; | |
353 | } | |
354 | ||
62f2c1b5 | 355 | uint32_t nth_midxed_pack_int_id(struct multi_pack_index *m, uint32_t pos) |
3715a633 | 356 | { |
329fac3a DS |
357 | return get_be32(m->chunk_object_offsets + |
358 | (off_t)pos * MIDX_CHUNK_OFFSET_WIDTH); | |
3715a633 DS |
359 | } |
360 | ||
a8437f3c | 361 | int fill_midx_entry(struct repository *r, |
893b5635 RS |
362 | const struct object_id *oid, |
363 | struct pack_entry *e, | |
364 | struct multi_pack_index *m) | |
3715a633 | 365 | { |
893b5635 | 366 | uint32_t pos; |
3715a633 DS |
367 | uint32_t pack_int_id; |
368 | struct packed_git *p; | |
369 | ||
893b5635 RS |
370 | if (!bsearch_midx(oid, m, &pos)) |
371 | return 0; | |
372 | ||
3715a633 DS |
373 | if (pos >= m->num_objects) |
374 | return 0; | |
375 | ||
376 | pack_int_id = nth_midxed_pack_int_id(m, pos); | |
377 | ||
64404a24 | 378 | if (prepare_midx_pack(r, m, pack_int_id)) |
506ec2fb | 379 | return 0; |
3715a633 DS |
380 | p = m->packs[pack_int_id]; |
381 | ||
382 | /* | |
383 | * We are about to tell the caller where they can locate the | |
384 | * requested object. We better make sure the packfile is | |
385 | * still here and can be accessed before supplying that | |
386 | * answer, as it may have been deleted since the MIDX was | |
387 | * loaded! | |
388 | */ | |
389 | if (!is_pack_valid(p)) | |
390 | return 0; | |
391 | ||
09ef6617 RS |
392 | if (oidset_size(&p->bad_objects) && |
393 | oidset_contains(&p->bad_objects, oid)) | |
394 | return 0; | |
c39b02ae | 395 | |
3715a633 DS |
396 | e->offset = nth_midxed_offset(m, pos); |
397 | e->p = p; | |
398 | ||
399 | return 1; | |
400 | } | |
401 | ||
013fd7ad JK |
402 | /* Match "foo.idx" against either "foo.pack" _or_ "foo.idx". */ |
403 | static int cmp_idx_or_pack_name(const char *idx_or_pack_name, | |
404 | const char *idx_name) | |
405 | { | |
406 | /* Skip past any initial matching prefix. */ | |
407 | while (*idx_name && *idx_name == *idx_or_pack_name) { | |
408 | idx_name++; | |
409 | idx_or_pack_name++; | |
410 | } | |
411 | ||
412 | /* | |
413 | * If we didn't match completely, we may have matched "pack-1234." and | |
414 | * be left with "idx" and "pack" respectively, which is also OK. We do | |
415 | * not have to check for "idx" and "idx", because that would have been | |
416 | * a complete match (and in that case these strcmps will be false, but | |
417 | * we'll correctly return 0 from the final strcmp() below. | |
418 | * | |
419 | * Technically this matches "fooidx" and "foopack", but we'd never have | |
420 | * such names in the first place. | |
421 | */ | |
422 | if (!strcmp(idx_name, "idx") && !strcmp(idx_or_pack_name, "pack")) | |
423 | return 0; | |
424 | ||
425 | /* | |
426 | * This not only checks for a complete match, but also orders based on | |
427 | * the first non-identical character, which means our ordering will | |
428 | * match a raw strcmp(). That makes it OK to use this to binary search | |
429 | * a naively-sorted list. | |
430 | */ | |
431 | return strcmp(idx_or_pack_name, idx_name); | |
432 | } | |
433 | ||
307d75bb TB |
434 | int midx_locate_pack(struct multi_pack_index *m, const char *idx_or_pack_name, |
435 | uint32_t *pos) | |
a40498a1 DS |
436 | { |
437 | uint32_t first = 0, last = m->num_packs; | |
438 | ||
439 | while (first < last) { | |
440 | uint32_t mid = first + (last - first) / 2; | |
441 | const char *current; | |
442 | int cmp; | |
443 | ||
444 | current = m->pack_names[mid]; | |
013fd7ad | 445 | cmp = cmp_idx_or_pack_name(idx_or_pack_name, current); |
307d75bb TB |
446 | if (!cmp) { |
447 | if (pos) | |
448 | *pos = mid; | |
a40498a1 | 449 | return 1; |
307d75bb | 450 | } |
a40498a1 DS |
451 | if (cmp > 0) { |
452 | first = mid + 1; | |
453 | continue; | |
454 | } | |
455 | last = mid; | |
456 | } | |
457 | ||
458 | return 0; | |
459 | } | |
460 | ||
307d75bb TB |
461 | int midx_contains_pack(struct multi_pack_index *m, const char *idx_or_pack_name) |
462 | { | |
463 | return midx_locate_pack(m, idx_or_pack_name, NULL); | |
464 | } | |
465 | ||
b1e33330 TB |
466 | int midx_preferred_pack(struct multi_pack_index *m, uint32_t *pack_int_id) |
467 | { | |
468 | if (m->preferred_pack_idx == -1) { | |
469 | if (load_midx_revindex(m) < 0) { | |
470 | m->preferred_pack_idx = -2; | |
471 | return -1; | |
472 | } | |
473 | ||
474 | m->preferred_pack_idx = | |
475 | nth_midxed_pack_int_id(m, pack_pos_to_midx(m, 0)); | |
476 | } else if (m->preferred_pack_idx == -2) | |
477 | return -1; /* no revindex */ | |
478 | ||
479 | *pack_int_id = m->preferred_pack_idx; | |
480 | return 0; | |
481 | } | |
482 | ||
2cf489a3 | 483 | int prepare_multi_pack_index_one(struct repository *r, const char *object_dir, int local) |
c4d25228 | 484 | { |
29e2016b | 485 | struct multi_pack_index *m; |
c4d25228 | 486 | struct multi_pack_index *m_search; |
c4d25228 | 487 | |
18e449f8 DS |
488 | prepare_repo_settings(r); |
489 | if (!r->settings.core_multi_pack_index) | |
c4d25228 DS |
490 | return 0; |
491 | ||
29e2016b | 492 | for (m_search = r->objects->multi_pack_index; m_search; m_search = m_search->next) |
c4d25228 DS |
493 | if (!strcmp(object_dir, m_search->object_dir)) |
494 | return 1; | |
495 | ||
29e2016b | 496 | m = load_multi_pack_index(object_dir, local); |
c4d25228 | 497 | |
29e2016b | 498 | if (m) { |
59552fb3 TB |
499 | struct multi_pack_index *mp = r->objects->multi_pack_index; |
500 | if (mp) { | |
501 | m->next = mp->next; | |
502 | mp->next = m; | |
503 | } else | |
504 | r->objects->multi_pack_index = m; | |
c4d25228 DS |
505 | return 1; |
506 | } | |
507 | ||
508 | return 0; | |
509 | } | |
510 | ||
fc59e748 DS |
511 | static size_t write_midx_header(struct hashfile *f, |
512 | unsigned char num_chunks, | |
513 | uint32_t num_packs) | |
514 | { | |
fc59e748 | 515 | hashwrite_be32(f, MIDX_SIGNATURE); |
014f1447 | 516 | hashwrite_u8(f, MIDX_VERSION); |
d9fef9d9 | 517 | hashwrite_u8(f, oid_version(the_hash_algo)); |
014f1447 RS |
518 | hashwrite_u8(f, num_chunks); |
519 | hashwrite_u8(f, 0); /* unused */ | |
fc59e748 DS |
520 | hashwrite_be32(f, num_packs); |
521 | ||
522 | return MIDX_HEADER_SIZE; | |
523 | } | |
524 | ||
5f5ccd95 TB |
525 | #define BITMAP_POS_UNKNOWN (~((uint32_t)0)) |
526 | ||
d01bf2e6 DS |
527 | struct pack_info { |
528 | uint32_t orig_pack_int_id; | |
529 | char *pack_name; | |
530 | struct packed_git *p; | |
5f5ccd95 TB |
531 | |
532 | uint32_t bitmap_pos; | |
533 | uint32_t bitmap_nr; | |
534 | ||
19575c7c | 535 | unsigned expired : 1; |
d01bf2e6 DS |
536 | }; |
537 | ||
fba68184 TB |
538 | static void fill_pack_info(struct pack_info *info, |
539 | struct packed_git *p, const char *pack_name, | |
540 | uint32_t orig_pack_int_id) | |
541 | { | |
542 | memset(info, 0, sizeof(struct pack_info)); | |
543 | ||
544 | info->orig_pack_int_id = orig_pack_int_id; | |
545 | info->pack_name = xstrdup(pack_name); | |
546 | info->p = p; | |
5f5ccd95 | 547 | info->bitmap_pos = BITMAP_POS_UNKNOWN; |
fba68184 TB |
548 | } |
549 | ||
d01bf2e6 DS |
550 | static int pack_info_compare(const void *_a, const void *_b) |
551 | { | |
552 | struct pack_info *a = (struct pack_info *)_a; | |
553 | struct pack_info *b = (struct pack_info *)_b; | |
554 | return strcmp(a->pack_name, b->pack_name); | |
555 | } | |
556 | ||
9218c6a4 TB |
557 | static int idx_or_pack_name_cmp(const void *_va, const void *_vb) |
558 | { | |
559 | const char *pack_name = _va; | |
560 | const struct pack_info *compar = _vb; | |
561 | ||
562 | return cmp_idx_or_pack_name(pack_name, compar->pack_name); | |
563 | } | |
564 | ||
577dc496 | 565 | struct write_midx_context { |
d01bf2e6 | 566 | struct pack_info *info; |
cc381274 TB |
567 | size_t nr; |
568 | size_t alloc; | |
a40498a1 | 569 | struct multi_pack_index *m; |
840cef0c WB |
570 | struct progress *progress; |
571 | unsigned pack_paths_checked; | |
31bda9a2 DS |
572 | |
573 | struct pack_midx_entry *entries; | |
cc381274 | 574 | size_t entries_nr; |
7a3ada11 DS |
575 | |
576 | uint32_t *pack_perm; | |
38ff7cab | 577 | uint32_t *pack_order; |
7a3ada11 | 578 | unsigned large_offsets_needed:1; |
980f525c | 579 | uint32_t num_large_offsets; |
9218c6a4 TB |
580 | |
581 | int preferred_pack_idx; | |
56d863e9 TB |
582 | |
583 | struct string_list *to_include; | |
396f2570 DS |
584 | }; |
585 | ||
586 | static void add_pack_to_midx(const char *full_path, size_t full_path_len, | |
587 | const char *file_name, void *data) | |
588 | { | |
577dc496 | 589 | struct write_midx_context *ctx = data; |
fba68184 | 590 | struct packed_git *p; |
396f2570 DS |
591 | |
592 | if (ends_with(file_name, ".idx")) { | |
577dc496 | 593 | display_progress(ctx->progress, ++ctx->pack_paths_checked); |
56d863e9 TB |
594 | /* |
595 | * Note that at most one of ctx->m and ctx->to_include are set, | |
596 | * so we are testing midx_contains_pack() and | |
597 | * string_list_has_string() independently (guarded by the | |
598 | * appropriate NULL checks). | |
599 | * | |
600 | * We could support passing to_include while reusing an existing | |
601 | * MIDX, but don't currently since the reuse process drags | |
602 | * forward all packs from an existing MIDX (without checking | |
603 | * whether or not they appear in the to_include list). | |
604 | * | |
605 | * If we added support for that, these next two conditional | |
606 | * should be performed independently (likely checking | |
607 | * to_include before the existing MIDX). | |
608 | */ | |
577dc496 | 609 | if (ctx->m && midx_contains_pack(ctx->m, file_name)) |
a40498a1 | 610 | return; |
56d863e9 TB |
611 | else if (ctx->to_include && |
612 | !string_list_has_string(ctx->to_include, file_name)) | |
613 | return; | |
a40498a1 | 614 | |
577dc496 | 615 | ALLOC_GROW(ctx->info, ctx->nr + 1, ctx->alloc); |
396f2570 | 616 | |
fba68184 TB |
617 | p = add_packed_git(full_path, full_path_len, 0); |
618 | if (!p) { | |
396f2570 DS |
619 | warning(_("failed to add packfile '%s'"), |
620 | full_path); | |
621 | return; | |
622 | } | |
623 | ||
fba68184 | 624 | if (open_pack_index(p)) { |
fe1ed56f DS |
625 | warning(_("failed to open pack-index '%s'"), |
626 | full_path); | |
fba68184 TB |
627 | close_pack(p); |
628 | free(p); | |
fe1ed56f DS |
629 | return; |
630 | } | |
631 | ||
fba68184 | 632 | fill_pack_info(&ctx->info[ctx->nr], p, file_name, ctx->nr); |
577dc496 | 633 | ctx->nr++; |
396f2570 DS |
634 | } |
635 | } | |
636 | ||
fe1ed56f DS |
637 | struct pack_midx_entry { |
638 | struct object_id oid; | |
639 | uint32_t pack_int_id; | |
640 | time_t pack_mtime; | |
641 | uint64_t offset; | |
9218c6a4 | 642 | unsigned preferred : 1; |
fe1ed56f DS |
643 | }; |
644 | ||
645 | static int midx_oid_compare(const void *_a, const void *_b) | |
646 | { | |
647 | const struct pack_midx_entry *a = (const struct pack_midx_entry *)_a; | |
648 | const struct pack_midx_entry *b = (const struct pack_midx_entry *)_b; | |
649 | int cmp = oidcmp(&a->oid, &b->oid); | |
650 | ||
651 | if (cmp) | |
652 | return cmp; | |
653 | ||
9218c6a4 TB |
654 | /* Sort objects in a preferred pack first when multiple copies exist. */ |
655 | if (a->preferred > b->preferred) | |
656 | return -1; | |
657 | if (a->preferred < b->preferred) | |
658 | return 1; | |
659 | ||
fe1ed56f DS |
660 | if (a->pack_mtime > b->pack_mtime) |
661 | return -1; | |
662 | else if (a->pack_mtime < b->pack_mtime) | |
663 | return 1; | |
664 | ||
665 | return a->pack_int_id - b->pack_int_id; | |
666 | } | |
667 | ||
a40498a1 | 668 | static int nth_midxed_pack_midx_entry(struct multi_pack_index *m, |
a40498a1 DS |
669 | struct pack_midx_entry *e, |
670 | uint32_t pos) | |
671 | { | |
672 | if (pos >= m->num_objects) | |
673 | return 1; | |
674 | ||
675 | nth_midxed_object_oid(&e->oid, m, pos); | |
d01bf2e6 | 676 | e->pack_int_id = nth_midxed_pack_int_id(m, pos); |
a40498a1 DS |
677 | e->offset = nth_midxed_offset(m, pos); |
678 | ||
679 | /* consider objects in midx to be from "old" packs */ | |
680 | e->pack_mtime = 0; | |
681 | return 0; | |
682 | } | |
683 | ||
fe1ed56f DS |
684 | static void fill_pack_entry(uint32_t pack_int_id, |
685 | struct packed_git *p, | |
686 | uint32_t cur_object, | |
9218c6a4 TB |
687 | struct pack_midx_entry *entry, |
688 | int preferred) | |
fe1ed56f | 689 | { |
0763671b | 690 | if (nth_packed_object_id(&entry->oid, p, cur_object) < 0) |
fe1ed56f DS |
691 | die(_("failed to locate object %d in packfile"), cur_object); |
692 | ||
693 | entry->pack_int_id = pack_int_id; | |
694 | entry->pack_mtime = p->mtime; | |
695 | ||
696 | entry->offset = nth_packed_object_offset(p, cur_object); | |
9218c6a4 | 697 | entry->preferred = !!preferred; |
fe1ed56f DS |
698 | } |
699 | ||
989d9cbd TB |
700 | struct midx_fanout { |
701 | struct pack_midx_entry *entries; | |
e6c71f23 | 702 | size_t nr, alloc; |
989d9cbd TB |
703 | }; |
704 | ||
e6c71f23 | 705 | static void midx_fanout_grow(struct midx_fanout *fanout, size_t nr) |
989d9cbd | 706 | { |
e6c71f23 TB |
707 | if (nr < fanout->nr) |
708 | BUG("negative growth in midx_fanout_grow() (%"PRIuMAX" < %"PRIuMAX")", | |
709 | (uintmax_t)nr, (uintmax_t)fanout->nr); | |
989d9cbd TB |
710 | ALLOC_GROW(fanout->entries, nr, fanout->alloc); |
711 | } | |
712 | ||
713 | static void midx_fanout_sort(struct midx_fanout *fanout) | |
714 | { | |
715 | QSORT(fanout->entries, fanout->nr, midx_oid_compare); | |
716 | } | |
717 | ||
852c5301 TB |
718 | static void midx_fanout_add_midx_fanout(struct midx_fanout *fanout, |
719 | struct multi_pack_index *m, | |
99e4d084 TB |
720 | uint32_t cur_fanout, |
721 | int preferred_pack) | |
852c5301 TB |
722 | { |
723 | uint32_t start = 0, end; | |
724 | uint32_t cur_object; | |
725 | ||
726 | if (cur_fanout) | |
727 | start = ntohl(m->chunk_oid_fanout[cur_fanout - 1]); | |
728 | end = ntohl(m->chunk_oid_fanout[cur_fanout]); | |
729 | ||
730 | for (cur_object = start; cur_object < end; cur_object++) { | |
99e4d084 TB |
731 | if ((preferred_pack > -1) && |
732 | (preferred_pack == nth_midxed_pack_int_id(m, cur_object))) { | |
733 | /* | |
734 | * Objects from preferred packs are added | |
735 | * separately. | |
736 | */ | |
737 | continue; | |
738 | } | |
739 | ||
852c5301 TB |
740 | midx_fanout_grow(fanout, fanout->nr + 1); |
741 | nth_midxed_pack_midx_entry(m, | |
742 | &fanout->entries[fanout->nr], | |
743 | cur_object); | |
cdf517be | 744 | fanout->entries[fanout->nr].preferred = 0; |
852c5301 TB |
745 | fanout->nr++; |
746 | } | |
747 | } | |
748 | ||
1d6f4c64 TB |
749 | static void midx_fanout_add_pack_fanout(struct midx_fanout *fanout, |
750 | struct pack_info *info, | |
751 | uint32_t cur_pack, | |
752 | int preferred, | |
753 | uint32_t cur_fanout) | |
754 | { | |
755 | struct packed_git *pack = info[cur_pack].p; | |
756 | uint32_t start = 0, end; | |
757 | uint32_t cur_object; | |
758 | ||
759 | if (cur_fanout) | |
760 | start = get_pack_fanout(pack, cur_fanout - 1); | |
761 | end = get_pack_fanout(pack, cur_fanout); | |
762 | ||
763 | for (cur_object = start; cur_object < end; cur_object++) { | |
764 | midx_fanout_grow(fanout, fanout->nr + 1); | |
765 | fill_pack_entry(cur_pack, | |
766 | info[cur_pack].p, | |
767 | cur_object, | |
768 | &fanout->entries[fanout->nr], | |
769 | preferred); | |
770 | fanout->nr++; | |
771 | } | |
772 | } | |
773 | ||
fe1ed56f DS |
774 | /* |
775 | * It is possible to artificially get into a state where there are many | |
776 | * duplicate copies of objects. That can create high memory pressure if | |
777 | * we are to create a list of all objects before de-duplication. To reduce | |
778 | * this memory pressure without a significant performance drop, automatically | |
779 | * group objects by the first byte of their object id. Use the IDX fanout | |
780 | * tables to group the data, copy to a local array, then sort. | |
781 | * | |
782 | * Copy only the de-duplicated entries (selected by most-recent modified time | |
783 | * of a packfile containing the object). | |
784 | */ | |
a40498a1 | 785 | static struct pack_midx_entry *get_sorted_entries(struct multi_pack_index *m, |
d01bf2e6 | 786 | struct pack_info *info, |
fe1ed56f | 787 | uint32_t nr_packs, |
cc381274 | 788 | size_t *nr_objects, |
9218c6a4 | 789 | int preferred_pack) |
fe1ed56f DS |
790 | { |
791 | uint32_t cur_fanout, cur_pack, cur_object; | |
cc381274 | 792 | size_t alloc_objects, total_objects = 0; |
989d9cbd | 793 | struct midx_fanout fanout = { 0 }; |
fe1ed56f | 794 | struct pack_midx_entry *deduplicated_entries = NULL; |
a40498a1 | 795 | uint32_t start_pack = m ? m->num_packs : 0; |
fe1ed56f | 796 | |
a40498a1 | 797 | for (cur_pack = start_pack; cur_pack < nr_packs; cur_pack++) |
cc381274 TB |
798 | total_objects = st_add(total_objects, |
799 | info[cur_pack].p->num_objects); | |
fe1ed56f DS |
800 | |
801 | /* | |
802 | * As we de-duplicate by fanout value, we expect the fanout | |
803 | * slices to be evenly distributed, with some noise. Hence, | |
804 | * allocate slightly more than one 256th. | |
805 | */ | |
989d9cbd | 806 | alloc_objects = fanout.alloc = total_objects > 3200 ? total_objects / 200 : 16; |
fe1ed56f | 807 | |
989d9cbd | 808 | ALLOC_ARRAY(fanout.entries, fanout.alloc); |
fe1ed56f DS |
809 | ALLOC_ARRAY(deduplicated_entries, alloc_objects); |
810 | *nr_objects = 0; | |
811 | ||
812 | for (cur_fanout = 0; cur_fanout < 256; cur_fanout++) { | |
989d9cbd | 813 | fanout.nr = 0; |
fe1ed56f | 814 | |
852c5301 | 815 | if (m) |
99e4d084 TB |
816 | midx_fanout_add_midx_fanout(&fanout, m, cur_fanout, |
817 | preferred_pack); | |
a40498a1 DS |
818 | |
819 | for (cur_pack = start_pack; cur_pack < nr_packs; cur_pack++) { | |
9218c6a4 | 820 | int preferred = cur_pack == preferred_pack; |
1d6f4c64 TB |
821 | midx_fanout_add_pack_fanout(&fanout, |
822 | info, cur_pack, | |
823 | preferred, cur_fanout); | |
fe1ed56f DS |
824 | } |
825 | ||
cdf517be TB |
826 | if (-1 < preferred_pack && preferred_pack < start_pack) |
827 | midx_fanout_add_pack_fanout(&fanout, info, | |
828 | preferred_pack, 1, | |
829 | cur_fanout); | |
830 | ||
989d9cbd | 831 | midx_fanout_sort(&fanout); |
fe1ed56f DS |
832 | |
833 | /* | |
834 | * The batch is now sorted by OID and then mtime (descending). | |
835 | * Take only the first duplicate. | |
836 | */ | |
989d9cbd TB |
837 | for (cur_object = 0; cur_object < fanout.nr; cur_object++) { |
838 | if (cur_object && oideq(&fanout.entries[cur_object - 1].oid, | |
839 | &fanout.entries[cur_object].oid)) | |
fe1ed56f DS |
840 | continue; |
841 | ||
cc381274 TB |
842 | ALLOC_GROW(deduplicated_entries, st_add(*nr_objects, 1), |
843 | alloc_objects); | |
fe1ed56f | 844 | memcpy(&deduplicated_entries[*nr_objects], |
989d9cbd | 845 | &fanout.entries[cur_object], |
fe1ed56f DS |
846 | sizeof(struct pack_midx_entry)); |
847 | (*nr_objects)++; | |
848 | } | |
849 | } | |
850 | ||
989d9cbd | 851 | free(fanout.entries); |
fe1ed56f DS |
852 | return deduplicated_entries; |
853 | } | |
854 | ||
0ccd713c | 855 | static int write_midx_pack_names(struct hashfile *f, void *data) |
32f3c541 | 856 | { |
b4d94142 | 857 | struct write_midx_context *ctx = data; |
32f3c541 DS |
858 | uint32_t i; |
859 | unsigned char padding[MIDX_CHUNK_ALIGNMENT]; | |
860 | size_t written = 0; | |
861 | ||
b4d94142 | 862 | for (i = 0; i < ctx->nr; i++) { |
19575c7c DS |
863 | size_t writelen; |
864 | ||
b4d94142 | 865 | if (ctx->info[i].expired) |
19575c7c | 866 | continue; |
32f3c541 | 867 | |
b4d94142 | 868 | if (i && strcmp(ctx->info[i].pack_name, ctx->info[i - 1].pack_name) <= 0) |
32f3c541 | 869 | BUG("incorrect pack-file order: %s before %s", |
b4d94142 DS |
870 | ctx->info[i - 1].pack_name, |
871 | ctx->info[i].pack_name); | |
32f3c541 | 872 | |
b4d94142 DS |
873 | writelen = strlen(ctx->info[i].pack_name) + 1; |
874 | hashwrite(f, ctx->info[i].pack_name, writelen); | |
32f3c541 DS |
875 | written += writelen; |
876 | } | |
877 | ||
878 | /* add padding to be aligned */ | |
879 | i = MIDX_CHUNK_ALIGNMENT - (written % MIDX_CHUNK_ALIGNMENT); | |
880 | if (i < MIDX_CHUNK_ALIGNMENT) { | |
881 | memset(padding, 0, sizeof(padding)); | |
882 | hashwrite(f, padding, i); | |
32f3c541 DS |
883 | } |
884 | ||
0ccd713c | 885 | return 0; |
32f3c541 DS |
886 | } |
887 | ||
5f5ccd95 TB |
888 | static int write_midx_bitmapped_packs(struct hashfile *f, void *data) |
889 | { | |
890 | struct write_midx_context *ctx = data; | |
891 | size_t i; | |
892 | ||
893 | for (i = 0; i < ctx->nr; i++) { | |
894 | struct pack_info *pack = &ctx->info[i]; | |
895 | if (pack->expired) | |
896 | continue; | |
897 | ||
898 | if (pack->bitmap_pos == BITMAP_POS_UNKNOWN && pack->bitmap_nr) | |
899 | BUG("pack '%s' has no bitmap position, but has %d bitmapped object(s)", | |
900 | pack->pack_name, pack->bitmap_nr); | |
901 | ||
902 | hashwrite_be32(f, pack->bitmap_pos); | |
903 | hashwrite_be32(f, pack->bitmap_nr); | |
904 | } | |
905 | return 0; | |
906 | } | |
907 | ||
0ccd713c DS |
908 | static int write_midx_oid_fanout(struct hashfile *f, |
909 | void *data) | |
d7cacf29 | 910 | { |
31bda9a2 DS |
911 | struct write_midx_context *ctx = data; |
912 | struct pack_midx_entry *list = ctx->entries; | |
913 | struct pack_midx_entry *last = ctx->entries + ctx->entries_nr; | |
d7cacf29 DS |
914 | uint32_t count = 0; |
915 | uint32_t i; | |
916 | ||
917 | /* | |
918 | * Write the first-level table (the list is sorted, | |
919 | * but we use a 256-entry lookup to be able to avoid | |
920 | * having to do eight extra binary search iterations). | |
921 | */ | |
922 | for (i = 0; i < 256; i++) { | |
923 | struct pack_midx_entry *next = list; | |
924 | ||
925 | while (next < last && next->oid.hash[0] == i) { | |
926 | count++; | |
927 | next++; | |
928 | } | |
929 | ||
930 | hashwrite_be32(f, count); | |
931 | list = next; | |
932 | } | |
933 | ||
0ccd713c | 934 | return 0; |
d7cacf29 DS |
935 | } |
936 | ||
0ccd713c DS |
937 | static int write_midx_oid_lookup(struct hashfile *f, |
938 | void *data) | |
0d5b3a5e | 939 | { |
31bda9a2 DS |
940 | struct write_midx_context *ctx = data; |
941 | unsigned char hash_len = the_hash_algo->rawsz; | |
942 | struct pack_midx_entry *list = ctx->entries; | |
0d5b3a5e | 943 | uint32_t i; |
0d5b3a5e | 944 | |
31bda9a2 | 945 | for (i = 0; i < ctx->entries_nr; i++) { |
0d5b3a5e DS |
946 | struct pack_midx_entry *obj = list++; |
947 | ||
31bda9a2 | 948 | if (i < ctx->entries_nr - 1) { |
0d5b3a5e DS |
949 | struct pack_midx_entry *next = list; |
950 | if (oidcmp(&obj->oid, &next->oid) >= 0) | |
951 | BUG("OIDs not in order: %s >= %s", | |
952 | oid_to_hex(&obj->oid), | |
953 | oid_to_hex(&next->oid)); | |
954 | } | |
955 | ||
956 | hashwrite(f, obj->oid.hash, (int)hash_len); | |
0d5b3a5e DS |
957 | } |
958 | ||
0ccd713c | 959 | return 0; |
0d5b3a5e DS |
960 | } |
961 | ||
0ccd713c DS |
962 | static int write_midx_object_offsets(struct hashfile *f, |
963 | void *data) | |
662148c4 | 964 | { |
7a3ada11 DS |
965 | struct write_midx_context *ctx = data; |
966 | struct pack_midx_entry *list = ctx->entries; | |
662148c4 | 967 | uint32_t i, nr_large_offset = 0; |
662148c4 | 968 | |
7a3ada11 | 969 | for (i = 0; i < ctx->entries_nr; i++) { |
662148c4 DS |
970 | struct pack_midx_entry *obj = list++; |
971 | ||
7a3ada11 | 972 | if (ctx->pack_perm[obj->pack_int_id] == PACK_EXPIRED) |
19575c7c DS |
973 | BUG("object %s is in an expired pack with int-id %d", |
974 | oid_to_hex(&obj->oid), | |
975 | obj->pack_int_id); | |
976 | ||
7a3ada11 | 977 | hashwrite_be32(f, ctx->pack_perm[obj->pack_int_id]); |
662148c4 | 978 | |
7a3ada11 | 979 | if (ctx->large_offsets_needed && obj->offset >> 31) |
662148c4 | 980 | hashwrite_be32(f, MIDX_LARGE_OFFSET_NEEDED | nr_large_offset++); |
7a3ada11 | 981 | else if (!ctx->large_offsets_needed && obj->offset >> 32) |
662148c4 DS |
982 | BUG("object %s requires a large offset (%"PRIx64") but the MIDX is not writing large offsets!", |
983 | oid_to_hex(&obj->oid), | |
984 | obj->offset); | |
985 | else | |
986 | hashwrite_be32(f, (uint32_t)obj->offset); | |
662148c4 DS |
987 | } |
988 | ||
0ccd713c | 989 | return 0; |
662148c4 DS |
990 | } |
991 | ||
0ccd713c DS |
992 | static int write_midx_large_offsets(struct hashfile *f, |
993 | void *data) | |
662148c4 | 994 | { |
980f525c DS |
995 | struct write_midx_context *ctx = data; |
996 | struct pack_midx_entry *list = ctx->entries; | |
997 | struct pack_midx_entry *end = ctx->entries + ctx->entries_nr; | |
980f525c | 998 | uint32_t nr_large_offset = ctx->num_large_offsets; |
662148c4 DS |
999 | |
1000 | while (nr_large_offset) { | |
61b0fcbb JK |
1001 | struct pack_midx_entry *obj; |
1002 | uint64_t offset; | |
1003 | ||
1004 | if (list >= end) | |
1005 | BUG("too many large-offset objects"); | |
1006 | ||
1007 | obj = list++; | |
1008 | offset = obj->offset; | |
662148c4 DS |
1009 | |
1010 | if (!(offset >> 31)) | |
1011 | continue; | |
1012 | ||
0ccd713c | 1013 | hashwrite_be64(f, offset); |
662148c4 DS |
1014 | |
1015 | nr_large_offset--; | |
1016 | } | |
1017 | ||
0ccd713c | 1018 | return 0; |
662148c4 DS |
1019 | } |
1020 | ||
95e8383b TB |
1021 | static int write_midx_revindex(struct hashfile *f, |
1022 | void *data) | |
1023 | { | |
1024 | struct write_midx_context *ctx = data; | |
1025 | uint32_t i; | |
1026 | ||
1027 | for (i = 0; i < ctx->entries_nr; i++) | |
1028 | hashwrite_be32(f, ctx->pack_order[i]); | |
1029 | ||
1030 | return 0; | |
1031 | } | |
1032 | ||
30077524 JK |
1033 | struct midx_pack_order_data { |
1034 | uint32_t nr; | |
1035 | uint32_t pack; | |
1036 | off_t offset; | |
1037 | }; | |
38ff7cab | 1038 | |
30077524 JK |
1039 | static int midx_pack_order_cmp(const void *va, const void *vb) |
1040 | { | |
1041 | const struct midx_pack_order_data *a = va, *b = vb; | |
1042 | if (a->pack < b->pack) | |
38ff7cab | 1043 | return -1; |
30077524 | 1044 | else if (a->pack > b->pack) |
38ff7cab | 1045 | return 1; |
30077524 | 1046 | else if (a->offset < b->offset) |
38ff7cab | 1047 | return -1; |
30077524 | 1048 | else if (a->offset > b->offset) |
38ff7cab | 1049 | return 1; |
30077524 JK |
1050 | else |
1051 | return 0; | |
38ff7cab TB |
1052 | } |
1053 | ||
1054 | static uint32_t *midx_pack_order(struct write_midx_context *ctx) | |
1055 | { | |
30077524 | 1056 | struct midx_pack_order_data *data; |
38ff7cab TB |
1057 | uint32_t *pack_order; |
1058 | uint32_t i; | |
1059 | ||
2dcff525 TB |
1060 | trace2_region_enter("midx", "midx_pack_order", the_repository); |
1061 | ||
30077524 JK |
1062 | ALLOC_ARRAY(data, ctx->entries_nr); |
1063 | for (i = 0; i < ctx->entries_nr; i++) { | |
1064 | struct pack_midx_entry *e = &ctx->entries[i]; | |
1065 | data[i].nr = i; | |
1066 | data[i].pack = ctx->pack_perm[e->pack_int_id]; | |
1067 | if (!e->preferred) | |
1068 | data[i].pack |= (1U << 31); | |
1069 | data[i].offset = e->offset; | |
1070 | } | |
1071 | ||
1072 | QSORT(data, ctx->entries_nr, midx_pack_order_cmp); | |
1073 | ||
38ff7cab | 1074 | ALLOC_ARRAY(pack_order, ctx->entries_nr); |
5f5ccd95 TB |
1075 | for (i = 0; i < ctx->entries_nr; i++) { |
1076 | struct pack_midx_entry *e = &ctx->entries[data[i].nr]; | |
1077 | struct pack_info *pack = &ctx->info[ctx->pack_perm[e->pack_int_id]]; | |
1078 | if (pack->bitmap_pos == BITMAP_POS_UNKNOWN) | |
1079 | pack->bitmap_pos = i; | |
1080 | pack->bitmap_nr++; | |
30077524 | 1081 | pack_order[i] = data[i].nr; |
5f5ccd95 TB |
1082 | } |
1083 | for (i = 0; i < ctx->nr; i++) { | |
1084 | struct pack_info *pack = &ctx->info[ctx->pack_perm[i]]; | |
1085 | if (pack->bitmap_pos == BITMAP_POS_UNKNOWN) | |
1086 | pack->bitmap_pos = 0; | |
1087 | } | |
30077524 | 1088 | free(data); |
38ff7cab | 1089 | |
2dcff525 TB |
1090 | trace2_region_leave("midx", "midx_pack_order", the_repository); |
1091 | ||
38ff7cab TB |
1092 | return pack_order; |
1093 | } | |
1094 | ||
1095 | static void write_midx_reverse_index(char *midx_name, unsigned char *midx_hash, | |
1096 | struct write_midx_context *ctx) | |
1097 | { | |
1098 | struct strbuf buf = STRBUF_INIT; | |
1099 | const char *tmp_file; | |
1100 | ||
2dcff525 TB |
1101 | trace2_region_enter("midx", "write_midx_reverse_index", the_repository); |
1102 | ||
38ff7cab TB |
1103 | strbuf_addf(&buf, "%s-%s.rev", midx_name, hash_to_hex(midx_hash)); |
1104 | ||
1105 | tmp_file = write_rev_file_order(NULL, ctx->pack_order, ctx->entries_nr, | |
1106 | midx_hash, WRITE_REV); | |
1107 | ||
1108 | if (finalize_object_file(tmp_file, buf.buf)) | |
1109 | die(_("cannot store reverse index file")); | |
1110 | ||
1111 | strbuf_release(&buf); | |
2dcff525 TB |
1112 | |
1113 | trace2_region_leave("midx", "write_midx_reverse_index", the_repository); | |
38ff7cab TB |
1114 | } |
1115 | ||
426c00e4 | 1116 | static void clear_midx_files_ext(const char *object_dir, const char *ext, |
38ff7cab TB |
1117 | unsigned char *keep_hash); |
1118 | ||
ec1e28ef TB |
1119 | static int midx_checksum_valid(struct multi_pack_index *m) |
1120 | { | |
1121 | return hashfile_checksum_valid(m->data, m->data_len); | |
1122 | } | |
1123 | ||
c528e179 TB |
1124 | static void prepare_midx_packing_data(struct packing_data *pdata, |
1125 | struct write_midx_context *ctx) | |
1126 | { | |
1127 | uint32_t i; | |
1128 | ||
2dcff525 TB |
1129 | trace2_region_enter("midx", "prepare_midx_packing_data", the_repository); |
1130 | ||
c528e179 TB |
1131 | memset(pdata, 0, sizeof(struct packing_data)); |
1132 | prepare_packing_data(the_repository, pdata); | |
1133 | ||
1134 | for (i = 0; i < ctx->entries_nr; i++) { | |
1135 | struct pack_midx_entry *from = &ctx->entries[ctx->pack_order[i]]; | |
1136 | struct object_entry *to = packlist_alloc(pdata, &from->oid); | |
1137 | ||
1138 | oe_set_in_pack(pdata, to, | |
1139 | ctx->info[ctx->pack_perm[from->pack_int_id]].p); | |
1140 | } | |
2dcff525 TB |
1141 | |
1142 | trace2_region_leave("midx", "prepare_midx_packing_data", the_repository); | |
c528e179 TB |
1143 | } |
1144 | ||
1145 | static int add_ref_to_pending(const char *refname, | |
1146 | const struct object_id *oid, | |
1147 | int flag, void *cb_data) | |
1148 | { | |
1149 | struct rev_info *revs = (struct rev_info*)cb_data; | |
1dc4f1ef | 1150 | struct object_id peeled; |
c528e179 TB |
1151 | struct object *object; |
1152 | ||
1153 | if ((flag & REF_ISSYMREF) && (flag & REF_ISBROKEN)) { | |
1154 | warning("symbolic ref is dangling: %s", refname); | |
1155 | return 0; | |
1156 | } | |
1157 | ||
1dc4f1ef TB |
1158 | if (!peel_iterated_oid(oid, &peeled)) |
1159 | oid = &peeled; | |
1160 | ||
c528e179 TB |
1161 | object = parse_object_or_die(oid, refname); |
1162 | if (object->type != OBJ_COMMIT) | |
1163 | return 0; | |
1164 | ||
1165 | add_pending_object(revs, object, ""); | |
1166 | if (bitmap_is_preferred_refname(revs->repo, refname)) | |
1167 | object->flags |= NEEDS_BITMAP; | |
1168 | return 0; | |
1169 | } | |
1170 | ||
1171 | struct bitmap_commit_cb { | |
1172 | struct commit **commits; | |
1173 | size_t commits_nr, commits_alloc; | |
1174 | ||
1175 | struct write_midx_context *ctx; | |
1176 | }; | |
1177 | ||
1178 | static const struct object_id *bitmap_oid_access(size_t index, | |
1179 | const void *_entries) | |
1180 | { | |
1181 | const struct pack_midx_entry *entries = _entries; | |
1182 | return &entries[index].oid; | |
1183 | } | |
1184 | ||
1185 | static void bitmap_show_commit(struct commit *commit, void *_data) | |
1186 | { | |
1187 | struct bitmap_commit_cb *data = _data; | |
1188 | int pos = oid_pos(&commit->object.oid, data->ctx->entries, | |
1189 | data->ctx->entries_nr, | |
1190 | bitmap_oid_access); | |
1191 | if (pos < 0) | |
1192 | return; | |
1193 | ||
1194 | ALLOC_GROW(data->commits, data->commits_nr + 1, data->commits_alloc); | |
1195 | data->commits[data->commits_nr++] = commit; | |
1196 | } | |
1197 | ||
08944d1c TB |
1198 | static int read_refs_snapshot(const char *refs_snapshot, |
1199 | struct rev_info *revs) | |
1200 | { | |
1201 | struct strbuf buf = STRBUF_INIT; | |
1202 | struct object_id oid; | |
1203 | FILE *f = xfopen(refs_snapshot, "r"); | |
1204 | ||
1205 | while (strbuf_getline(&buf, f) != EOF) { | |
1206 | struct object *object; | |
1207 | int preferred = 0; | |
1208 | char *hex = buf.buf; | |
1209 | const char *end = NULL; | |
1210 | ||
1211 | if (buf.len && *buf.buf == '+') { | |
1212 | preferred = 1; | |
1213 | hex = &buf.buf[1]; | |
1214 | } | |
1215 | ||
1216 | if (parse_oid_hex(hex, &oid, &end) < 0) | |
1217 | die(_("could not parse line: %s"), buf.buf); | |
1218 | if (*end) | |
1219 | die(_("malformed line: %s"), buf.buf); | |
1220 | ||
1221 | object = parse_object_or_die(&oid, NULL); | |
1222 | if (preferred) | |
1223 | object->flags |= NEEDS_BITMAP; | |
1224 | ||
1225 | add_pending_object(revs, object, ""); | |
1226 | } | |
1227 | ||
1228 | fclose(f); | |
1229 | strbuf_release(&buf); | |
1230 | return 0; | |
1231 | } | |
1232 | ||
c528e179 | 1233 | static struct commit **find_commits_for_midx_bitmap(uint32_t *indexed_commits_nr_p, |
08944d1c | 1234 | const char *refs_snapshot, |
c528e179 TB |
1235 | struct write_midx_context *ctx) |
1236 | { | |
1237 | struct rev_info revs; | |
1238 | struct bitmap_commit_cb cb = {0}; | |
1239 | ||
2dcff525 TB |
1240 | trace2_region_enter("midx", "find_commits_for_midx_bitmap", |
1241 | the_repository); | |
1242 | ||
c528e179 TB |
1243 | cb.ctx = ctx; |
1244 | ||
1245 | repo_init_revisions(the_repository, &revs, NULL); | |
08944d1c TB |
1246 | if (refs_snapshot) { |
1247 | read_refs_snapshot(refs_snapshot, &revs); | |
1248 | } else { | |
1249 | setup_revisions(0, NULL, &revs, NULL); | |
1250 | for_each_ref(add_ref_to_pending, &revs); | |
1251 | } | |
c528e179 TB |
1252 | |
1253 | /* | |
1254 | * Skipping promisor objects here is intentional, since it only excludes | |
1255 | * them from the list of reachable commits that we want to select from | |
1256 | * when computing the selection of MIDX'd commits to receive bitmaps. | |
1257 | * | |
1258 | * Reachability bitmaps do require that their objects be closed under | |
1259 | * reachability, but fetching any objects missing from promisors at this | |
1260 | * point is too late. But, if one of those objects can be reached from | |
1261 | * an another object that is included in the bitmap, then we will | |
1262 | * complain later that we don't have reachability closure (and fail | |
1263 | * appropriately). | |
1264 | */ | |
1265 | fetch_if_missing = 0; | |
1266 | revs.exclude_promisor_objects = 1; | |
1267 | ||
1268 | if (prepare_revision_walk(&revs)) | |
1269 | die(_("revision walk setup failed")); | |
1270 | ||
1271 | traverse_commit_list(&revs, bitmap_show_commit, NULL, &cb); | |
1272 | if (indexed_commits_nr_p) | |
1273 | *indexed_commits_nr_p = cb.commits_nr; | |
1274 | ||
2108fe4a | 1275 | release_revisions(&revs); |
2dcff525 TB |
1276 | |
1277 | trace2_region_leave("midx", "find_commits_for_midx_bitmap", | |
1278 | the_repository); | |
1279 | ||
c528e179 TB |
1280 | return cb.commits; |
1281 | } | |
1282 | ||
90b2bb71 DS |
1283 | static int write_midx_bitmap(const char *midx_name, |
1284 | const unsigned char *midx_hash, | |
1285 | struct packing_data *pdata, | |
1286 | struct commit **commits, | |
1287 | uint32_t commits_nr, | |
1288 | uint32_t *pack_order, | |
c528e179 TB |
1289 | unsigned flags) |
1290 | { | |
90b2bb71 | 1291 | int ret, i; |
caca3c9f | 1292 | uint16_t options = 0; |
90b2bb71 DS |
1293 | struct pack_idx_entry **index; |
1294 | char *bitmap_name = xstrfmt("%s-%s.bitmap", midx_name, | |
1295 | hash_to_hex(midx_hash)); | |
eb57277b | 1296 | |
2dcff525 TB |
1297 | trace2_region_enter("midx", "write_midx_bitmap", the_repository); |
1298 | ||
caca3c9f TB |
1299 | if (flags & MIDX_WRITE_BITMAP_HASH_CACHE) |
1300 | options |= BITMAP_OPT_HASH_CACHE; | |
1301 | ||
76f14b77 AC |
1302 | if (flags & MIDX_WRITE_BITMAP_LOOKUP_TABLE) |
1303 | options |= BITMAP_OPT_LOOKUP_TABLE; | |
1304 | ||
c528e179 TB |
1305 | /* |
1306 | * Build the MIDX-order index based on pdata.objects (which is already | |
1307 | * in MIDX order; c.f., 'midx_pack_order_cmp()' for the definition of | |
1308 | * this order). | |
1309 | */ | |
90b2bb71 DS |
1310 | ALLOC_ARRAY(index, pdata->nr_objects); |
1311 | for (i = 0; i < pdata->nr_objects; i++) | |
1312 | index[i] = &pdata->objects[i].idx; | |
c528e179 TB |
1313 | |
1314 | bitmap_writer_show_progress(flags & MIDX_PROGRESS); | |
90b2bb71 | 1315 | bitmap_writer_build_type_index(pdata, index, pdata->nr_objects); |
c528e179 TB |
1316 | |
1317 | /* | |
1318 | * bitmap_writer_finish expects objects in lex order, but pack_order | |
1319 | * gives us exactly that. use it directly instead of re-sorting the | |
1320 | * array. | |
1321 | * | |
1322 | * This changes the order of objects in 'index' between | |
1323 | * bitmap_writer_build_type_index and bitmap_writer_finish. | |
1324 | * | |
1325 | * The same re-ordering takes place in the single-pack bitmap code via | |
1326 | * write_idx_file(), which is called by finish_tmp_packfile(), which | |
1327 | * happens between bitmap_writer_build_type_index() and | |
1328 | * bitmap_writer_finish(). | |
1329 | */ | |
90b2bb71 DS |
1330 | for (i = 0; i < pdata->nr_objects; i++) |
1331 | index[pack_order[i]] = &pdata->objects[i].idx; | |
c528e179 TB |
1332 | |
1333 | bitmap_writer_select_commits(commits, commits_nr, -1); | |
90b2bb71 | 1334 | ret = bitmap_writer_build(pdata); |
c528e179 TB |
1335 | if (ret < 0) |
1336 | goto cleanup; | |
1337 | ||
1338 | bitmap_writer_set_checksum(midx_hash); | |
90b2bb71 | 1339 | bitmap_writer_finish(index, pdata->nr_objects, bitmap_name, options); |
c528e179 TB |
1340 | |
1341 | cleanup: | |
1342 | free(index); | |
1343 | free(bitmap_name); | |
2dcff525 TB |
1344 | |
1345 | trace2_region_leave("midx", "write_midx_bitmap", the_repository); | |
1346 | ||
c528e179 TB |
1347 | return ret; |
1348 | } | |
1349 | ||
504131ac TB |
1350 | static struct multi_pack_index *lookup_multi_pack_index(struct repository *r, |
1351 | const char *object_dir) | |
1352 | { | |
eafcc6de | 1353 | struct multi_pack_index *result = NULL; |
504131ac | 1354 | struct multi_pack_index *cur; |
eafcc6de DS |
1355 | char *obj_dir_real = real_pathdup(object_dir, 1); |
1356 | struct strbuf cur_path_real = STRBUF_INIT; | |
504131ac TB |
1357 | |
1358 | /* Ensure the given object_dir is local, or a known alternate. */ | |
eafcc6de | 1359 | find_odb(r, obj_dir_real); |
504131ac TB |
1360 | |
1361 | for (cur = get_multi_pack_index(r); cur; cur = cur->next) { | |
eafcc6de DS |
1362 | strbuf_realpath(&cur_path_real, cur->object_dir, 1); |
1363 | if (!strcmp(obj_dir_real, cur_path_real.buf)) { | |
1364 | result = cur; | |
1365 | goto cleanup; | |
1366 | } | |
504131ac TB |
1367 | } |
1368 | ||
eafcc6de DS |
1369 | cleanup: |
1370 | free(obj_dir_real); | |
1371 | strbuf_release(&cur_path_real); | |
1372 | return result; | |
504131ac TB |
1373 | } |
1374 | ||
f57a7396 | 1375 | static int write_midx_internal(const char *object_dir, |
56d863e9 | 1376 | struct string_list *packs_to_include, |
9218c6a4 TB |
1377 | struct string_list *packs_to_drop, |
1378 | const char *preferred_pack_name, | |
08944d1c | 1379 | const char *refs_snapshot, |
9218c6a4 | 1380 | unsigned flags) |
a3407730 | 1381 | { |
60980aed | 1382 | struct strbuf midx_name = STRBUF_INIT; |
9f191611 | 1383 | unsigned char midx_hash[GIT_MAX_RAWSZ]; |
396f2570 | 1384 | uint32_t i; |
fc59e748 DS |
1385 | struct hashfile *f = NULL; |
1386 | struct lock_file lk; | |
577dc496 | 1387 | struct write_midx_context ctx = { 0 }; |
5f5ccd95 | 1388 | int bitmapped_packs_concat_len = 0; |
dba6175c | 1389 | int pack_name_concat_len = 0; |
19575c7c DS |
1390 | int dropped_packs = 0; |
1391 | int result = 0; | |
63a8f0e9 | 1392 | struct chunkfile *cf; |
fc59e748 | 1393 | |
2dcff525 TB |
1394 | trace2_region_enter("midx", "write_midx_internal", the_repository); |
1395 | ||
60980aed TB |
1396 | get_midx_filename(&midx_name, object_dir); |
1397 | if (safe_create_leading_directories(midx_name.buf)) | |
fc59e748 | 1398 | die_errno(_("unable to create leading directories of %s"), |
60980aed | 1399 | midx_name.buf); |
fc59e748 | 1400 | |
56d863e9 TB |
1401 | if (!packs_to_include) { |
1402 | /* | |
1403 | * Only reference an existing MIDX when not filtering which | |
1404 | * packs to include, since all packs and objects are copied | |
1405 | * blindly from an existing MIDX if one is present. | |
1406 | */ | |
504131ac | 1407 | ctx.m = lookup_multi_pack_index(the_repository, object_dir); |
f57a7396 | 1408 | } |
577dc496 | 1409 | |
ec1e28ef TB |
1410 | if (ctx.m && !midx_checksum_valid(ctx.m)) { |
1411 | warning(_("ignoring existing multi-pack-index; checksum mismatch")); | |
1412 | ctx.m = NULL; | |
1413 | } | |
1414 | ||
577dc496 DS |
1415 | ctx.nr = 0; |
1416 | ctx.alloc = ctx.m ? ctx.m->num_packs : 16; | |
1417 | ctx.info = NULL; | |
1418 | ALLOC_ARRAY(ctx.info, ctx.alloc); | |
1419 | ||
1420 | if (ctx.m) { | |
1421 | for (i = 0; i < ctx.m->num_packs; i++) { | |
1422 | ALLOC_GROW(ctx.info, ctx.nr + 1, ctx.alloc); | |
1423 | ||
5d3cd09a TB |
1424 | if (flags & MIDX_WRITE_REV_INDEX) { |
1425 | /* | |
1426 | * If generating a reverse index, need to have | |
1427 | * packed_git's loaded to compare their | |
1428 | * mtimes and object count. | |
1429 | */ | |
1430 | if (prepare_midx_pack(the_repository, ctx.m, i)) { | |
1431 | error(_("could not load pack")); | |
1432 | result = 1; | |
1433 | goto cleanup; | |
1434 | } | |
1435 | ||
1436 | if (open_pack_index(ctx.m->packs[i])) | |
1437 | die(_("could not open index for %s"), | |
1438 | ctx.m->packs[i]->pack_name); | |
5d3cd09a TB |
1439 | } |
1440 | ||
fba68184 TB |
1441 | fill_pack_info(&ctx.info[ctx.nr++], ctx.m->packs[i], |
1442 | ctx.m->pack_names[i], i); | |
a40498a1 DS |
1443 | } |
1444 | } | |
1445 | ||
577dc496 | 1446 | ctx.pack_paths_checked = 0; |
840cef0c | 1447 | if (flags & MIDX_PROGRESS) |
577dc496 | 1448 | ctx.progress = start_delayed_progress(_("Adding packfiles to multi-pack-index"), 0); |
840cef0c | 1449 | else |
577dc496 | 1450 | ctx.progress = NULL; |
840cef0c | 1451 | |
56d863e9 TB |
1452 | ctx.to_include = packs_to_include; |
1453 | ||
577dc496 DS |
1454 | for_each_file_in_pack_dir(object_dir, add_pack_to_midx, &ctx); |
1455 | stop_progress(&ctx.progress); | |
396f2570 | 1456 | |
56d863e9 TB |
1457 | if ((ctx.m && ctx.nr == ctx.m->num_packs) && |
1458 | !(packs_to_include || packs_to_drop)) { | |
c528e179 TB |
1459 | struct bitmap_index *bitmap_git; |
1460 | int bitmap_exists; | |
1461 | int want_bitmap = flags & MIDX_WRITE_BITMAP; | |
1462 | ||
bfbb60d3 | 1463 | bitmap_git = prepare_midx_bitmap_git(ctx.m); |
c528e179 TB |
1464 | bitmap_exists = bitmap_git && bitmap_is_midx(bitmap_git); |
1465 | free_bitmap_index(bitmap_git); | |
1466 | ||
1467 | if (bitmap_exists || !want_bitmap) { | |
1468 | /* | |
1469 | * The correct MIDX already exists, and so does a | |
1470 | * corresponding bitmap (or one wasn't requested). | |
1471 | */ | |
1472 | if (!want_bitmap) | |
1473 | clear_midx_files_ext(object_dir, ".bitmap", | |
1474 | NULL); | |
1475 | goto cleanup; | |
1476 | } | |
1477 | } | |
a40498a1 | 1478 | |
9218c6a4 | 1479 | if (preferred_pack_name) { |
ceb96a16 PS |
1480 | ctx.preferred_pack_idx = -1; |
1481 | ||
9218c6a4 TB |
1482 | for (i = 0; i < ctx.nr; i++) { |
1483 | if (!cmp_idx_or_pack_name(preferred_pack_name, | |
1484 | ctx.info[i].pack_name)) { | |
1485 | ctx.preferred_pack_idx = i; | |
1486 | break; | |
1487 | } | |
1488 | } | |
177c0d6e | 1489 | |
ceb96a16 | 1490 | if (ctx.preferred_pack_idx == -1) |
177c0d6e TB |
1491 | warning(_("unknown preferred pack: '%s'"), |
1492 | preferred_pack_name); | |
c528e179 TB |
1493 | } else if (ctx.nr && |
1494 | (flags & (MIDX_WRITE_REV_INDEX | MIDX_WRITE_BITMAP))) { | |
177c0d6e TB |
1495 | struct packed_git *oldest = ctx.info[ctx.preferred_pack_idx].p; |
1496 | ctx.preferred_pack_idx = 0; | |
1497 | ||
1498 | if (packs_to_drop && packs_to_drop->nr) | |
1499 | BUG("cannot write a MIDX bitmap during expiration"); | |
1500 | ||
1501 | /* | |
1502 | * set a preferred pack when writing a bitmap to ensure that | |
1503 | * the pack from which the first object is selected in pseudo | |
1504 | * pack-order has all of its objects selected from that pack | |
1505 | * (and not another pack containing a duplicate) | |
1506 | */ | |
1507 | for (i = 1; i < ctx.nr; i++) { | |
1508 | struct packed_git *p = ctx.info[i].p; | |
1509 | ||
1510 | if (!oldest->num_objects || p->mtime < oldest->mtime) { | |
1511 | oldest = p; | |
1512 | ctx.preferred_pack_idx = i; | |
1513 | } | |
1514 | } | |
1515 | ||
1516 | if (!oldest->num_objects) { | |
1517 | /* | |
1518 | * If all packs are empty; unset the preferred index. | |
1519 | * This is acceptable since there will be no duplicate | |
1520 | * objects to resolve, so the preferred value doesn't | |
1521 | * matter. | |
1522 | */ | |
1523 | ctx.preferred_pack_idx = -1; | |
1524 | } | |
1525 | } else { | |
1526 | /* | |
1527 | * otherwise don't mark any pack as preferred to avoid | |
1528 | * interfering with expiration logic below | |
1529 | */ | |
1530 | ctx.preferred_pack_idx = -1; | |
9218c6a4 TB |
1531 | } |
1532 | ||
5d3cd09a TB |
1533 | if (ctx.preferred_pack_idx > -1) { |
1534 | struct packed_git *preferred = ctx.info[ctx.preferred_pack_idx].p; | |
1535 | if (!preferred->num_objects) { | |
1536 | error(_("cannot select preferred pack %s with no objects"), | |
1537 | preferred->pack_name); | |
1538 | result = 1; | |
1539 | goto cleanup; | |
1540 | } | |
1541 | } | |
1542 | ||
9218c6a4 TB |
1543 | ctx.entries = get_sorted_entries(ctx.m, ctx.info, ctx.nr, &ctx.entries_nr, |
1544 | ctx.preferred_pack_idx); | |
a40498a1 | 1545 | |
7a3ada11 | 1546 | ctx.large_offsets_needed = 0; |
31bda9a2 DS |
1547 | for (i = 0; i < ctx.entries_nr; i++) { |
1548 | if (ctx.entries[i].offset > 0x7fffffff) | |
980f525c | 1549 | ctx.num_large_offsets++; |
31bda9a2 | 1550 | if (ctx.entries[i].offset > 0xffffffff) |
7a3ada11 | 1551 | ctx.large_offsets_needed = 1; |
662148c4 | 1552 | } |
fe1ed56f | 1553 | |
577dc496 | 1554 | QSORT(ctx.info, ctx.nr, pack_info_compare); |
d01bf2e6 | 1555 | |
19575c7c DS |
1556 | if (packs_to_drop && packs_to_drop->nr) { |
1557 | int drop_index = 0; | |
1558 | int missing_drops = 0; | |
1559 | ||
577dc496 DS |
1560 | for (i = 0; i < ctx.nr && drop_index < packs_to_drop->nr; i++) { |
1561 | int cmp = strcmp(ctx.info[i].pack_name, | |
19575c7c DS |
1562 | packs_to_drop->items[drop_index].string); |
1563 | ||
1564 | if (!cmp) { | |
1565 | drop_index++; | |
577dc496 | 1566 | ctx.info[i].expired = 1; |
19575c7c DS |
1567 | } else if (cmp > 0) { |
1568 | error(_("did not see pack-file %s to drop"), | |
1569 | packs_to_drop->items[drop_index].string); | |
1570 | drop_index++; | |
1571 | missing_drops++; | |
1572 | i--; | |
1573 | } else { | |
577dc496 | 1574 | ctx.info[i].expired = 0; |
19575c7c DS |
1575 | } |
1576 | } | |
1577 | ||
1578 | if (missing_drops) { | |
1579 | result = 1; | |
1580 | goto cleanup; | |
1581 | } | |
1582 | } | |
1583 | ||
d01bf2e6 DS |
1584 | /* |
1585 | * pack_perm stores a permutation between pack-int-ids from the | |
1586 | * previous multi-pack-index to the new one we are writing: | |
1587 | * | |
1588 | * pack_perm[old_id] = new_id | |
1589 | */ | |
7a3ada11 | 1590 | ALLOC_ARRAY(ctx.pack_perm, ctx.nr); |
577dc496 DS |
1591 | for (i = 0; i < ctx.nr; i++) { |
1592 | if (ctx.info[i].expired) { | |
19575c7c | 1593 | dropped_packs++; |
7a3ada11 | 1594 | ctx.pack_perm[ctx.info[i].orig_pack_int_id] = PACK_EXPIRED; |
19575c7c | 1595 | } else { |
7a3ada11 | 1596 | ctx.pack_perm[ctx.info[i].orig_pack_int_id] = i - dropped_packs; |
19575c7c | 1597 | } |
d01bf2e6 DS |
1598 | } |
1599 | ||
577dc496 | 1600 | for (i = 0; i < ctx.nr; i++) { |
5f5ccd95 TB |
1601 | if (ctx.info[i].expired) |
1602 | continue; | |
1603 | pack_name_concat_len += strlen(ctx.info[i].pack_name) + 1; | |
1604 | bitmapped_packs_concat_len += 2 * sizeof(uint32_t); | |
19575c7c | 1605 | } |
dba6175c | 1606 | |
9218c6a4 TB |
1607 | /* Check that the preferred pack wasn't expired (if given). */ |
1608 | if (preferred_pack_name) { | |
1609 | struct pack_info *preferred = bsearch(preferred_pack_name, | |
1610 | ctx.info, ctx.nr, | |
1611 | sizeof(*ctx.info), | |
1612 | idx_or_pack_name_cmp); | |
177c0d6e | 1613 | if (preferred) { |
9218c6a4 TB |
1614 | uint32_t perm = ctx.pack_perm[preferred->orig_pack_int_id]; |
1615 | if (perm == PACK_EXPIRED) | |
1616 | warning(_("preferred pack '%s' is expired"), | |
1617 | preferred_pack_name); | |
1618 | } | |
1619 | } | |
1620 | ||
dba6175c DS |
1621 | if (pack_name_concat_len % MIDX_CHUNK_ALIGNMENT) |
1622 | pack_name_concat_len += MIDX_CHUNK_ALIGNMENT - | |
1623 | (pack_name_concat_len % MIDX_CHUNK_ALIGNMENT); | |
1624 | ||
60980aed | 1625 | hold_lock_file_for_update(&lk, midx_name.buf, LOCK_DIE_ON_ERROR); |
acd71602 | 1626 | f = hashfd(get_lock_file_fd(&lk), get_lock_file_path(&lk)); |
fc59e748 | 1627 | |
577dc496 | 1628 | if (ctx.nr - dropped_packs == 0) { |
796d61cd DR |
1629 | error(_("no pack files to index.")); |
1630 | result = 1; | |
1631 | goto cleanup; | |
1632 | } | |
1633 | ||
eb57277b TB |
1634 | if (!ctx.entries_nr) { |
1635 | if (flags & MIDX_WRITE_BITMAP) | |
1636 | warning(_("refusing to write multi-pack .bitmap without any objects")); | |
1637 | flags &= ~(MIDX_WRITE_REV_INDEX | MIDX_WRITE_BITMAP); | |
1638 | } | |
1639 | ||
63a8f0e9 | 1640 | cf = init_chunkfile(f); |
32f3c541 | 1641 | |
63a8f0e9 DS |
1642 | add_chunk(cf, MIDX_CHUNKID_PACKNAMES, pack_name_concat_len, |
1643 | write_midx_pack_names); | |
1644 | add_chunk(cf, MIDX_CHUNKID_OIDFANOUT, MIDX_CHUNK_FANOUT_SIZE, | |
1645 | write_midx_oid_fanout); | |
1646 | add_chunk(cf, MIDX_CHUNKID_OIDLOOKUP, | |
2bc764c1 | 1647 | st_mult(ctx.entries_nr, the_hash_algo->rawsz), |
63a8f0e9 DS |
1648 | write_midx_oid_lookup); |
1649 | add_chunk(cf, MIDX_CHUNKID_OBJECTOFFSETS, | |
2bc764c1 | 1650 | st_mult(ctx.entries_nr, MIDX_CHUNK_OFFSET_WIDTH), |
63a8f0e9 | 1651 | write_midx_object_offsets); |
32f3c541 | 1652 | |
63a8f0e9 DS |
1653 | if (ctx.large_offsets_needed) |
1654 | add_chunk(cf, MIDX_CHUNKID_LARGEOFFSETS, | |
2bc764c1 TB |
1655 | st_mult(ctx.num_large_offsets, |
1656 | MIDX_CHUNK_LARGE_OFFSET_WIDTH), | |
63a8f0e9 | 1657 | write_midx_large_offsets); |
32f3c541 | 1658 | |
95e8383b TB |
1659 | if (flags & (MIDX_WRITE_REV_INDEX | MIDX_WRITE_BITMAP)) { |
1660 | ctx.pack_order = midx_pack_order(&ctx); | |
1661 | add_chunk(cf, MIDX_CHUNKID_REVINDEX, | |
2bc764c1 | 1662 | st_mult(ctx.entries_nr, sizeof(uint32_t)), |
95e8383b | 1663 | write_midx_revindex); |
5f5ccd95 TB |
1664 | add_chunk(cf, MIDX_CHUNKID_BITMAPPEDPACKS, |
1665 | bitmapped_packs_concat_len, | |
1666 | write_midx_bitmapped_packs); | |
95e8383b TB |
1667 | } |
1668 | ||
63a8f0e9 DS |
1669 | write_midx_header(f, get_num_chunks(cf), ctx.nr - dropped_packs); |
1670 | write_chunkfile(cf, &ctx); | |
fc59e748 | 1671 | |
020406ea NS |
1672 | finalize_hashfile(f, midx_hash, FSYNC_COMPONENT_PACK_METADATA, |
1673 | CSUM_FSYNC | CSUM_HASH_IN_STREAM); | |
63a8f0e9 | 1674 | free_chunkfile(cf); |
38ff7cab | 1675 | |
7f514b7a TB |
1676 | if (flags & MIDX_WRITE_REV_INDEX && |
1677 | git_env_bool("GIT_TEST_MIDX_WRITE_REV", 0)) | |
60980aed | 1678 | write_midx_reverse_index(midx_name.buf, midx_hash, &ctx); |
90b2bb71 | 1679 | |
c528e179 | 1680 | if (flags & MIDX_WRITE_BITMAP) { |
90b2bb71 DS |
1681 | struct packing_data pdata; |
1682 | struct commit **commits; | |
1683 | uint32_t commits_nr; | |
1684 | ||
1685 | if (!ctx.entries_nr) | |
1686 | BUG("cannot write a bitmap without any objects"); | |
1687 | ||
1688 | prepare_midx_packing_data(&pdata, &ctx); | |
1689 | ||
1690 | commits = find_commits_for_midx_bitmap(&commits_nr, refs_snapshot, &ctx); | |
1691 | ||
068fa54c DS |
1692 | /* |
1693 | * The previous steps translated the information from | |
1694 | * 'entries' into information suitable for constructing | |
1695 | * bitmaps. We no longer need that array, so clear it to | |
1696 | * reduce memory pressure. | |
1697 | */ | |
1698 | FREE_AND_NULL(ctx.entries); | |
1699 | ctx.entries_nr = 0; | |
1700 | ||
90b2bb71 DS |
1701 | if (write_midx_bitmap(midx_name.buf, midx_hash, &pdata, |
1702 | commits, commits_nr, ctx.pack_order, | |
51d1b69a | 1703 | flags) < 0) { |
c528e179 TB |
1704 | error(_("could not write multi-pack bitmap")); |
1705 | result = 1; | |
66f0c710 TB |
1706 | clear_packing_data(&pdata); |
1707 | free(commits); | |
c528e179 TB |
1708 | goto cleanup; |
1709 | } | |
66f0c710 TB |
1710 | |
1711 | clear_packing_data(&pdata); | |
1712 | free(commits); | |
c528e179 | 1713 | } |
068fa54c DS |
1714 | /* |
1715 | * NOTE: Do not use ctx.entries beyond this point, since it might | |
1716 | * have been freed in the previous if block. | |
1717 | */ | |
c528e179 TB |
1718 | |
1719 | if (ctx.m) | |
1720 | close_object_store(the_repository->objects); | |
38ff7cab | 1721 | |
ae22e841 TB |
1722 | if (commit_lock_file(&lk) < 0) |
1723 | die_errno(_("could not write multi-pack-index")); | |
fc59e748 | 1724 | |
c528e179 | 1725 | clear_midx_files_ext(object_dir, ".bitmap", midx_hash); |
f5909d34 TB |
1726 | clear_midx_files_ext(object_dir, ".rev", midx_hash); |
1727 | ||
a40498a1 | 1728 | cleanup: |
577dc496 DS |
1729 | for (i = 0; i < ctx.nr; i++) { |
1730 | if (ctx.info[i].p) { | |
1731 | close_pack(ctx.info[i].p); | |
1732 | free(ctx.info[i].p); | |
396f2570 | 1733 | } |
577dc496 | 1734 | free(ctx.info[i].pack_name); |
396f2570 DS |
1735 | } |
1736 | ||
577dc496 | 1737 | free(ctx.info); |
31bda9a2 | 1738 | free(ctx.entries); |
7a3ada11 | 1739 | free(ctx.pack_perm); |
38ff7cab | 1740 | free(ctx.pack_order); |
60980aed | 1741 | strbuf_release(&midx_name); |
c528e179 | 1742 | |
2dcff525 TB |
1743 | trace2_region_leave("midx", "write_midx_internal", the_repository); |
1744 | ||
19575c7c DS |
1745 | return result; |
1746 | } | |
1747 | ||
9218c6a4 TB |
1748 | int write_midx_file(const char *object_dir, |
1749 | const char *preferred_pack_name, | |
08944d1c | 1750 | const char *refs_snapshot, |
9218c6a4 | 1751 | unsigned flags) |
19575c7c | 1752 | { |
56d863e9 | 1753 | return write_midx_internal(object_dir, NULL, NULL, preferred_pack_name, |
08944d1c | 1754 | refs_snapshot, flags); |
56d863e9 TB |
1755 | } |
1756 | ||
1757 | int write_midx_file_only(const char *object_dir, | |
1758 | struct string_list *packs_to_include, | |
1759 | const char *preferred_pack_name, | |
08944d1c | 1760 | const char *refs_snapshot, |
56d863e9 TB |
1761 | unsigned flags) |
1762 | { | |
1763 | return write_midx_internal(object_dir, packs_to_include, NULL, | |
08944d1c | 1764 | preferred_pack_name, refs_snapshot, flags); |
a3407730 | 1765 | } |
525e18c0 | 1766 | |
38ff7cab TB |
1767 | struct clear_midx_data { |
1768 | char *keep; | |
1769 | const char *ext; | |
1770 | }; | |
1771 | ||
be252d33 | 1772 | static void clear_midx_file_ext(const char *full_path, size_t full_path_len UNUSED, |
38ff7cab TB |
1773 | const char *file_name, void *_data) |
1774 | { | |
1775 | struct clear_midx_data *data = _data; | |
1776 | ||
1777 | if (!(starts_with(file_name, "multi-pack-index-") && | |
1778 | ends_with(file_name, data->ext))) | |
1779 | return; | |
1780 | if (data->keep && !strcmp(data->keep, file_name)) | |
1781 | return; | |
1782 | ||
1783 | if (unlink(full_path)) | |
1784 | die_errno(_("failed to remove %s"), full_path); | |
1785 | } | |
1786 | ||
426c00e4 | 1787 | static void clear_midx_files_ext(const char *object_dir, const char *ext, |
38ff7cab TB |
1788 | unsigned char *keep_hash) |
1789 | { | |
1790 | struct clear_midx_data data; | |
1791 | memset(&data, 0, sizeof(struct clear_midx_data)); | |
1792 | ||
1793 | if (keep_hash) | |
1794 | data.keep = xstrfmt("multi-pack-index-%s%s", | |
1795 | hash_to_hex(keep_hash), ext); | |
1796 | data.ext = ext; | |
1797 | ||
426c00e4 | 1798 | for_each_file_in_pack_dir(object_dir, |
38ff7cab TB |
1799 | clear_midx_file_ext, |
1800 | &data); | |
1801 | ||
1802 | free(data.keep); | |
a3407730 | 1803 | } |
525e18c0 | 1804 | |
1dcd9f20 | 1805 | void clear_midx_file(struct repository *r) |
525e18c0 | 1806 | { |
60980aed TB |
1807 | struct strbuf midx = STRBUF_INIT; |
1808 | ||
1809 | get_midx_filename(&midx, r->objects->odb->path); | |
1dcd9f20 DS |
1810 | |
1811 | if (r->objects && r->objects->multi_pack_index) { | |
1812 | close_midx(r->objects->multi_pack_index); | |
1813 | r->objects->multi_pack_index = NULL; | |
1814 | } | |
525e18c0 | 1815 | |
60980aed TB |
1816 | if (remove_path(midx.buf)) |
1817 | die(_("failed to clear multi-pack-index at %s"), midx.buf); | |
525e18c0 | 1818 | |
c528e179 | 1819 | clear_midx_files_ext(r->objects->odb->path, ".bitmap", NULL); |
426c00e4 | 1820 | clear_midx_files_ext(r->objects->odb->path, ".rev", NULL); |
38ff7cab | 1821 | |
60980aed | 1822 | strbuf_release(&midx); |
525e18c0 | 1823 | } |
56ee7ff1 DS |
1824 | |
1825 | static int verify_midx_error; | |
1826 | ||
48ca53ca | 1827 | __attribute__((format (printf, 1, 2))) |
d4bf1d88 DS |
1828 | static void midx_report(const char *fmt, ...) |
1829 | { | |
1830 | va_list ap; | |
1831 | verify_midx_error = 1; | |
1832 | va_start(ap, fmt); | |
1833 | vfprintf(stderr, fmt, ap); | |
1834 | fprintf(stderr, "\n"); | |
1835 | va_end(ap); | |
1836 | } | |
1837 | ||
5ae18df9 JH |
1838 | struct pair_pos_vs_id |
1839 | { | |
1840 | uint32_t pos; | |
1841 | uint32_t pack_int_id; | |
1842 | }; | |
1843 | ||
1844 | static int compare_pair_pos_vs_id(const void *_a, const void *_b) | |
1845 | { | |
1846 | struct pair_pos_vs_id *a = (struct pair_pos_vs_id *)_a; | |
1847 | struct pair_pos_vs_id *b = (struct pair_pos_vs_id *)_b; | |
1848 | ||
1849 | return b->pack_int_id - a->pack_int_id; | |
1850 | } | |
1851 | ||
430efb8a JH |
1852 | /* |
1853 | * Limit calls to display_progress() for performance reasons. | |
1854 | * The interval here was arbitrarily chosen. | |
1855 | */ | |
1856 | #define SPARSE_PROGRESS_INTERVAL (1 << 12) | |
1857 | #define midx_display_sparse_progress(progress, n) \ | |
1858 | do { \ | |
1859 | uint64_t _n = (n); \ | |
1860 | if ((_n & (SPARSE_PROGRESS_INTERVAL - 1)) == 0) \ | |
1861 | display_progress(progress, _n); \ | |
1862 | } while (0) | |
1863 | ||
efbc3aee | 1864 | int verify_midx_file(struct repository *r, const char *object_dir, unsigned flags) |
56ee7ff1 | 1865 | { |
5ae18df9 | 1866 | struct pair_pos_vs_id *pairs = NULL; |
d4bf1d88 | 1867 | uint32_t i; |
ad60096d | 1868 | struct progress *progress = NULL; |
56ee7ff1 DS |
1869 | struct multi_pack_index *m = load_multi_pack_index(object_dir, 1); |
1870 | verify_midx_error = 0; | |
1871 | ||
d9607542 DS |
1872 | if (!m) { |
1873 | int result = 0; | |
1874 | struct stat sb; | |
60980aed TB |
1875 | struct strbuf filename = STRBUF_INIT; |
1876 | ||
1877 | get_midx_filename(&filename, object_dir); | |
1878 | ||
1879 | if (!stat(filename.buf, &sb)) { | |
d9607542 DS |
1880 | error(_("multi-pack-index file exists, but failed to parse")); |
1881 | result = 1; | |
1882 | } | |
60980aed | 1883 | strbuf_release(&filename); |
d9607542 DS |
1884 | return result; |
1885 | } | |
56ee7ff1 | 1886 | |
f89ecf79 TB |
1887 | if (!midx_checksum_valid(m)) |
1888 | midx_report(_("incorrect checksum")); | |
1889 | ||
ad60096d | 1890 | if (flags & MIDX_PROGRESS) |
efdd2f0d | 1891 | progress = start_delayed_progress(_("Looking for referenced packfiles"), |
ad60096d | 1892 | m->num_packs); |
d4bf1d88 | 1893 | for (i = 0; i < m->num_packs; i++) { |
64404a24 | 1894 | if (prepare_midx_pack(r, m, i)) |
d4bf1d88 | 1895 | midx_report("failed to load pack in position %d", i); |
430efb8a JH |
1896 | |
1897 | display_progress(progress, i + 1); | |
d4bf1d88 | 1898 | } |
430efb8a | 1899 | stop_progress(&progress); |
d4bf1d88 | 1900 | |
796d61cd DR |
1901 | if (m->num_objects == 0) { |
1902 | midx_report(_("the midx contains no oid")); | |
1903 | /* | |
1904 | * Remaining tests assume that we have objects, so we can | |
1905 | * return here. | |
1906 | */ | |
492cb394 | 1907 | goto cleanup; |
796d61cd DR |
1908 | } |
1909 | ||
ad60096d WB |
1910 | if (flags & MIDX_PROGRESS) |
1911 | progress = start_sparse_progress(_("Verifying OID order in multi-pack-index"), | |
1912 | m->num_objects - 1); | |
55c5648d DS |
1913 | for (i = 0; i < m->num_objects - 1; i++) { |
1914 | struct object_id oid1, oid2; | |
1915 | ||
1916 | nth_midxed_object_oid(&oid1, m, i); | |
1917 | nth_midxed_object_oid(&oid2, m, i + 1); | |
1918 | ||
1919 | if (oidcmp(&oid1, &oid2) >= 0) | |
1920 | midx_report(_("oid lookup out of order: oid[%d] = %s >= %s = oid[%d]"), | |
1921 | i, oid_to_hex(&oid1), oid_to_hex(&oid2), i + 1); | |
430efb8a JH |
1922 | |
1923 | midx_display_sparse_progress(progress, i + 1); | |
55c5648d | 1924 | } |
430efb8a | 1925 | stop_progress(&progress); |
55c5648d | 1926 | |
5ae18df9 JH |
1927 | /* |
1928 | * Create an array mapping each object to its packfile id. Sort it | |
1929 | * to group the objects by packfile. Use this permutation to visit | |
1930 | * each of the objects and only require 1 packfile to be open at a | |
1931 | * time. | |
1932 | */ | |
1933 | ALLOC_ARRAY(pairs, m->num_objects); | |
1934 | for (i = 0; i < m->num_objects; i++) { | |
1935 | pairs[i].pos = i; | |
1936 | pairs[i].pack_int_id = nth_midxed_pack_int_id(m, i); | |
1937 | } | |
1938 | ||
ad60096d WB |
1939 | if (flags & MIDX_PROGRESS) |
1940 | progress = start_sparse_progress(_("Sorting objects by packfile"), | |
1941 | m->num_objects); | |
5ae18df9 JH |
1942 | display_progress(progress, 0); /* TODO: Measure QSORT() progress */ |
1943 | QSORT(pairs, m->num_objects, compare_pair_pos_vs_id); | |
1944 | stop_progress(&progress); | |
1945 | ||
ad60096d WB |
1946 | if (flags & MIDX_PROGRESS) |
1947 | progress = start_sparse_progress(_("Verifying object offsets"), m->num_objects); | |
cc6af73c DS |
1948 | for (i = 0; i < m->num_objects; i++) { |
1949 | struct object_id oid; | |
1950 | struct pack_entry e; | |
1951 | off_t m_offset, p_offset; | |
1952 | ||
5ae18df9 JH |
1953 | if (i > 0 && pairs[i-1].pack_int_id != pairs[i].pack_int_id && |
1954 | m->packs[pairs[i-1].pack_int_id]) | |
1955 | { | |
1956 | close_pack_fd(m->packs[pairs[i-1].pack_int_id]); | |
1957 | close_pack_index(m->packs[pairs[i-1].pack_int_id]); | |
1958 | } | |
1959 | ||
1960 | nth_midxed_object_oid(&oid, m, pairs[i].pos); | |
1961 | ||
64404a24 | 1962 | if (!fill_midx_entry(r, &oid, &e, m)) { |
cc6af73c | 1963 | midx_report(_("failed to load pack entry for oid[%d] = %s"), |
5ae18df9 | 1964 | pairs[i].pos, oid_to_hex(&oid)); |
cc6af73c DS |
1965 | continue; |
1966 | } | |
1967 | ||
1968 | if (open_pack_index(e.p)) { | |
1969 | midx_report(_("failed to load pack-index for packfile %s"), | |
1970 | e.p->pack_name); | |
1971 | break; | |
1972 | } | |
1973 | ||
1974 | m_offset = e.offset; | |
1975 | p_offset = find_pack_entry_one(oid.hash, e.p); | |
1976 | ||
1977 | if (m_offset != p_offset) | |
1978 | midx_report(_("incorrect object offset for oid[%d] = %s: %"PRIx64" != %"PRIx64), | |
5ae18df9 | 1979 | pairs[i].pos, oid_to_hex(&oid), m_offset, p_offset); |
144d7033 | 1980 | |
430efb8a | 1981 | midx_display_sparse_progress(progress, i + 1); |
cc6af73c | 1982 | } |
144d7033 | 1983 | stop_progress(&progress); |
cc6af73c | 1984 | |
492cb394 | 1985 | cleanup: |
5ae18df9 | 1986 | free(pairs); |
492cb394 | 1987 | close_midx(m); |
5ae18df9 | 1988 | |
56ee7ff1 DS |
1989 | return verify_midx_error; |
1990 | } | |
cff97116 | 1991 | |
efbc3aee | 1992 | int expire_midx_packs(struct repository *r, const char *object_dir, unsigned flags) |
cff97116 | 1993 | { |
19575c7c DS |
1994 | uint32_t i, *count, result = 0; |
1995 | struct string_list packs_to_drop = STRING_LIST_INIT_DUP; | |
98926e0d | 1996 | struct multi_pack_index *m = lookup_multi_pack_index(r, object_dir); |
8dc18f89 | 1997 | struct progress *progress = NULL; |
19575c7c DS |
1998 | |
1999 | if (!m) | |
2000 | return 0; | |
2001 | ||
ca56dadb | 2002 | CALLOC_ARRAY(count, m->num_packs); |
8dc18f89 WB |
2003 | |
2004 | if (flags & MIDX_PROGRESS) | |
efdd2f0d | 2005 | progress = start_delayed_progress(_("Counting referenced objects"), |
8dc18f89 | 2006 | m->num_objects); |
19575c7c DS |
2007 | for (i = 0; i < m->num_objects; i++) { |
2008 | int pack_int_id = nth_midxed_pack_int_id(m, i); | |
2009 | count[pack_int_id]++; | |
8dc18f89 | 2010 | display_progress(progress, i + 1); |
19575c7c | 2011 | } |
8dc18f89 | 2012 | stop_progress(&progress); |
19575c7c | 2013 | |
8dc18f89 | 2014 | if (flags & MIDX_PROGRESS) |
efdd2f0d | 2015 | progress = start_delayed_progress(_("Finding and deleting unreferenced packfiles"), |
8dc18f89 | 2016 | m->num_packs); |
19575c7c DS |
2017 | for (i = 0; i < m->num_packs; i++) { |
2018 | char *pack_name; | |
8dc18f89 | 2019 | display_progress(progress, i + 1); |
19575c7c DS |
2020 | |
2021 | if (count[i]) | |
2022 | continue; | |
2023 | ||
2024 | if (prepare_midx_pack(r, m, i)) | |
2025 | continue; | |
2026 | ||
757d4579 | 2027 | if (m->packs[i]->pack_keep || m->packs[i]->is_cruft) |
19575c7c DS |
2028 | continue; |
2029 | ||
2030 | pack_name = xstrdup(m->packs[i]->pack_name); | |
2031 | close_pack(m->packs[i]); | |
2032 | ||
2033 | string_list_insert(&packs_to_drop, m->pack_names[i]); | |
2034 | unlink_pack_path(pack_name, 0); | |
2035 | free(pack_name); | |
2036 | } | |
8dc18f89 | 2037 | stop_progress(&progress); |
19575c7c DS |
2038 | |
2039 | free(count); | |
2040 | ||
98926e0d | 2041 | if (packs_to_drop.nr) |
08944d1c | 2042 | result = write_midx_internal(object_dir, NULL, &packs_to_drop, NULL, NULL, flags); |
19575c7c DS |
2043 | |
2044 | string_list_clear(&packs_to_drop, 0); | |
98926e0d | 2045 | |
19575c7c | 2046 | return result; |
cff97116 | 2047 | } |
2af890bb | 2048 | |
ce1e4a10 DS |
2049 | struct repack_info { |
2050 | timestamp_t mtime; | |
2051 | uint32_t referenced_objects; | |
2052 | uint32_t pack_int_id; | |
2053 | }; | |
2054 | ||
2055 | static int compare_by_mtime(const void *a_, const void *b_) | |
2af890bb | 2056 | { |
ce1e4a10 DS |
2057 | const struct repack_info *a, *b; |
2058 | ||
2059 | a = (const struct repack_info *)a_; | |
2060 | b = (const struct repack_info *)b_; | |
2061 | ||
2062 | if (a->mtime < b->mtime) | |
2063 | return -1; | |
2064 | if (a->mtime > b->mtime) | |
2065 | return 1; | |
2066 | return 0; | |
2067 | } | |
2068 | ||
3ce4ca0a DS |
2069 | static int fill_included_packs_all(struct repository *r, |
2070 | struct multi_pack_index *m, | |
ce1e4a10 DS |
2071 | unsigned char *include_pack) |
2072 | { | |
3ce4ca0a DS |
2073 | uint32_t i, count = 0; |
2074 | int pack_kept_objects = 0; | |
2075 | ||
2076 | repo_config_get_bool(r, "repack.packkeptobjects", &pack_kept_objects); | |
2077 | ||
2078 | for (i = 0; i < m->num_packs; i++) { | |
2079 | if (prepare_midx_pack(r, m, i)) | |
2080 | continue; | |
2081 | if (!pack_kept_objects && m->packs[i]->pack_keep) | |
2082 | continue; | |
d9f77214 TB |
2083 | if (m->packs[i]->is_cruft) |
2084 | continue; | |
ce1e4a10 | 2085 | |
ce1e4a10 | 2086 | include_pack[i] = 1; |
3ce4ca0a DS |
2087 | count++; |
2088 | } | |
ce1e4a10 | 2089 | |
3ce4ca0a | 2090 | return count < 2; |
ce1e4a10 DS |
2091 | } |
2092 | ||
2093 | static int fill_included_packs_batch(struct repository *r, | |
2094 | struct multi_pack_index *m, | |
2095 | unsigned char *include_pack, | |
2096 | size_t batch_size) | |
2097 | { | |
2098 | uint32_t i, packs_to_repack; | |
2099 | size_t total_size; | |
cb6c48cb | 2100 | struct repack_info *pack_info; |
3ce4ca0a DS |
2101 | int pack_kept_objects = 0; |
2102 | ||
cb6c48cb TB |
2103 | CALLOC_ARRAY(pack_info, m->num_packs); |
2104 | ||
3ce4ca0a | 2105 | repo_config_get_bool(r, "repack.packkeptobjects", &pack_kept_objects); |
ce1e4a10 DS |
2106 | |
2107 | for (i = 0; i < m->num_packs; i++) { | |
2108 | pack_info[i].pack_int_id = i; | |
2109 | ||
2110 | if (prepare_midx_pack(r, m, i)) | |
2111 | continue; | |
2112 | ||
2113 | pack_info[i].mtime = m->packs[i]->mtime; | |
2114 | } | |
2115 | ||
0a8e5614 | 2116 | for (i = 0; i < m->num_objects; i++) { |
ce1e4a10 DS |
2117 | uint32_t pack_int_id = nth_midxed_pack_int_id(m, i); |
2118 | pack_info[pack_int_id].referenced_objects++; | |
2119 | } | |
2120 | ||
2121 | QSORT(pack_info, m->num_packs, compare_by_mtime); | |
2122 | ||
2123 | total_size = 0; | |
2124 | packs_to_repack = 0; | |
2125 | for (i = 0; total_size < batch_size && i < m->num_packs; i++) { | |
2126 | int pack_int_id = pack_info[i].pack_int_id; | |
2127 | struct packed_git *p = m->packs[pack_int_id]; | |
2128 | size_t expected_size; | |
2129 | ||
2130 | if (!p) | |
2131 | continue; | |
3ce4ca0a DS |
2132 | if (!pack_kept_objects && p->pack_keep) |
2133 | continue; | |
b62ad568 TB |
2134 | if (p->is_cruft) |
2135 | continue; | |
ce1e4a10 DS |
2136 | if (open_pack_index(p) || !p->num_objects) |
2137 | continue; | |
2138 | ||
d67609bd TB |
2139 | expected_size = st_mult(p->pack_size, |
2140 | pack_info[i].referenced_objects); | |
ce1e4a10 DS |
2141 | expected_size /= p->num_objects; |
2142 | ||
2143 | if (expected_size >= batch_size) | |
2144 | continue; | |
2145 | ||
2146 | packs_to_repack++; | |
2147 | total_size += expected_size; | |
2148 | include_pack[pack_int_id] = 1; | |
2149 | } | |
2150 | ||
2151 | free(pack_info); | |
2152 | ||
1eb22c7d | 2153 | if (packs_to_repack < 2) |
ce1e4a10 DS |
2154 | return 1; |
2155 | ||
2af890bb DS |
2156 | return 0; |
2157 | } | |
ce1e4a10 | 2158 | |
efbc3aee | 2159 | int midx_repack(struct repository *r, const char *object_dir, size_t batch_size, unsigned flags) |
ce1e4a10 DS |
2160 | { |
2161 | int result = 0; | |
2162 | uint32_t i; | |
2163 | unsigned char *include_pack; | |
2164 | struct child_process cmd = CHILD_PROCESS_INIT; | |
6af3b00a | 2165 | FILE *cmd_in; |
ce1e4a10 | 2166 | struct strbuf base_name = STRBUF_INIT; |
c0f1f9de | 2167 | struct multi_pack_index *m = lookup_multi_pack_index(r, object_dir); |
ce1e4a10 | 2168 | |
e11d86de SLN |
2169 | /* |
2170 | * When updating the default for these configuration | |
2171 | * variables in builtin/repack.c, these must be adjusted | |
2172 | * to match. | |
2173 | */ | |
2174 | int delta_base_offset = 1; | |
2175 | int use_delta_islands = 0; | |
2176 | ||
ce1e4a10 DS |
2177 | if (!m) |
2178 | return 0; | |
2179 | ||
ca56dadb | 2180 | CALLOC_ARRAY(include_pack, m->num_packs); |
ce1e4a10 DS |
2181 | |
2182 | if (batch_size) { | |
2183 | if (fill_included_packs_batch(r, m, include_pack, batch_size)) | |
2184 | goto cleanup; | |
3ce4ca0a | 2185 | } else if (fill_included_packs_all(r, m, include_pack)) |
ce1e4a10 DS |
2186 | goto cleanup; |
2187 | ||
e11d86de SLN |
2188 | repo_config_get_bool(r, "repack.usedeltabaseoffset", &delta_base_offset); |
2189 | repo_config_get_bool(r, "repack.usedeltaislands", &use_delta_islands); | |
2190 | ||
c972bf4c | 2191 | strvec_push(&cmd.args, "pack-objects"); |
ce1e4a10 DS |
2192 | |
2193 | strbuf_addstr(&base_name, object_dir); | |
2194 | strbuf_addstr(&base_name, "/pack/pack"); | |
c972bf4c | 2195 | strvec_push(&cmd.args, base_name.buf); |
64d80e7d | 2196 | |
e11d86de | 2197 | if (delta_base_offset) |
c972bf4c | 2198 | strvec_push(&cmd.args, "--delta-base-offset"); |
e11d86de | 2199 | if (use_delta_islands) |
c972bf4c | 2200 | strvec_push(&cmd.args, "--delta-islands"); |
e11d86de | 2201 | |
64d80e7d | 2202 | if (flags & MIDX_PROGRESS) |
c972bf4c | 2203 | strvec_push(&cmd.args, "--progress"); |
64d80e7d | 2204 | else |
c972bf4c | 2205 | strvec_push(&cmd.args, "-q"); |
64d80e7d | 2206 | |
ce1e4a10 DS |
2207 | strbuf_release(&base_name); |
2208 | ||
2209 | cmd.git_cmd = 1; | |
2210 | cmd.in = cmd.out = -1; | |
2211 | ||
2212 | if (start_command(&cmd)) { | |
2213 | error(_("could not start pack-objects")); | |
2214 | result = 1; | |
2215 | goto cleanup; | |
2216 | } | |
2217 | ||
6af3b00a RS |
2218 | cmd_in = xfdopen(cmd.in, "w"); |
2219 | ||
ce1e4a10 DS |
2220 | for (i = 0; i < m->num_objects; i++) { |
2221 | struct object_id oid; | |
2222 | uint32_t pack_int_id = nth_midxed_pack_int_id(m, i); | |
2223 | ||
2224 | if (!include_pack[pack_int_id]) | |
2225 | continue; | |
2226 | ||
2227 | nth_midxed_object_oid(&oid, m, i); | |
6af3b00a | 2228 | fprintf(cmd_in, "%s\n", oid_to_hex(&oid)); |
ce1e4a10 | 2229 | } |
6af3b00a | 2230 | fclose(cmd_in); |
ce1e4a10 DS |
2231 | |
2232 | if (finish_command(&cmd)) { | |
2233 | error(_("could not finish pack-objects")); | |
2234 | result = 1; | |
2235 | goto cleanup; | |
2236 | } | |
2237 | ||
08944d1c | 2238 | result = write_midx_internal(object_dir, NULL, NULL, NULL, NULL, flags); |
ce1e4a10 DS |
2239 | |
2240 | cleanup: | |
ce1e4a10 DS |
2241 | free(include_pack); |
2242 | return result; | |
2243 | } |