]>
Commit | Line | Data |
---|---|---|
0fcfd160 LT |
1 | /* |
2 | * GIT - The information manager from hell | |
3 | * | |
4 | * Copyright (C) Linus Torvalds, 2005 | |
5 | * | |
6 | * This handles basic git sha1 object files - packing, unpacking, | |
7 | * creation etc. | |
8 | */ | |
0fcfd160 | 9 | #include "cache.h" |
6eac50d8 | 10 | #include "string-list.h" |
697cc8ef | 11 | #include "lockfile.h" |
1f688557 | 12 | #include "delta.h" |
a733cb60 | 13 | #include "pack.h" |
8e440259 PE |
14 | #include "blob.h" |
15 | #include "commit.h" | |
4dd1fbc7 | 16 | #include "run-command.h" |
8e440259 PE |
17 | #include "tag.h" |
18 | #include "tree.h" | |
c879daa2 | 19 | #include "tree-walk.h" |
f35a6d3b | 20 | #include "refs.h" |
70f5d5d3 | 21 | #include "pack-revindex.h" |
628522ec | 22 | #include "sha1-lookup.h" |
568508e7 | 23 | #include "bulk-checkin.h" |
090ea126 | 24 | #include "streaming.h" |
543c5caa | 25 | #include "dir.h" |
0fcfd160 | 26 | |
144bde78 LT |
27 | #ifndef O_NOATIME |
28 | #if defined(__linux__) && (defined(__i386__) || defined(__PPC__)) | |
29 | #define O_NOATIME 01000000 | |
30 | #else | |
31 | #define O_NOATIME 0 | |
32 | #endif | |
33 | #endif | |
34 | ||
28bd70d8 JN |
35 | #define SZ_FMT PRIuMAX |
36 | static inline uintmax_t sz_fmt(size_t s) { return s; } | |
e05db0fd | 37 | |
96f1e58f | 38 | const unsigned char null_sha1[20]; |
88cd621d | 39 | |
c597ba80 NTND |
40 | /* |
41 | * This is meant to hold a *small* number of objects that you would | |
42 | * want read_sha1_file() to be able to return, but yet you do not want | |
43 | * to write them into the object store (e.g. a browse-only | |
44 | * application). | |
45 | */ | |
46 | static struct cached_object { | |
47 | unsigned char sha1[20]; | |
48 | enum object_type type; | |
49 | void *buf; | |
50 | unsigned long size; | |
51 | } *cached_objects; | |
52 | static int cached_object_nr, cached_object_alloc; | |
53 | ||
54 | static struct cached_object empty_tree = { | |
dab0d410 | 55 | EMPTY_TREE_SHA1_BIN_LITERAL, |
c597ba80 NTND |
56 | OBJ_TREE, |
57 | "", | |
58 | 0 | |
59 | }; | |
60 | ||
1b1005d1 MH |
61 | /* |
62 | * A pointer to the last packed_git in which an object was found. | |
63 | * When an object is sought, we look in this packfile first, because | |
64 | * objects that are looked up at similar times are often in the same | |
65 | * packfile as one another. | |
66 | */ | |
c01f51cc NTND |
67 | static struct packed_git *last_found_pack; |
68 | ||
c597ba80 NTND |
69 | static struct cached_object *find_cached_object(const unsigned char *sha1) |
70 | { | |
71 | int i; | |
72 | struct cached_object *co = cached_objects; | |
73 | ||
74 | for (i = 0; i < cached_object_nr; i++, co++) { | |
75 | if (!hashcmp(co->sha1, sha1)) | |
76 | return co; | |
77 | } | |
78 | if (!hashcmp(sha1, empty_tree.sha1)) | |
79 | return &empty_tree; | |
80 | return NULL; | |
81 | } | |
82 | ||
90a6464b JH |
83 | int mkdir_in_gitdir(const char *path) |
84 | { | |
85 | if (mkdir(path, 0777)) { | |
86 | int saved_errno = errno; | |
87 | struct stat st; | |
88 | struct strbuf sb = STRBUF_INIT; | |
89 | ||
90 | if (errno != EEXIST) | |
91 | return -1; | |
92 | /* | |
93 | * Are we looking at a path in a symlinked worktree | |
94 | * whose original repository does not yet have it? | |
95 | * e.g. .git/rr-cache pointing at its original | |
96 | * repository in which the user hasn't performed any | |
97 | * conflict resolution yet? | |
98 | */ | |
99 | if (lstat(path, &st) || !S_ISLNK(st.st_mode) || | |
100 | strbuf_readlink(&sb, path, st.st_size) || | |
101 | !is_absolute_path(sb.buf) || | |
102 | mkdir(sb.buf, 0777)) { | |
103 | strbuf_release(&sb); | |
104 | errno = saved_errno; | |
105 | return -1; | |
106 | } | |
107 | strbuf_release(&sb); | |
108 | } | |
109 | return adjust_shared_perm(path); | |
110 | } | |
111 | ||
0be0521b | 112 | enum scld_error safe_create_leading_directories(char *path) |
b2cb9425 | 113 | { |
26c8ae2a | 114 | char *next_component = path + offset_1st_component(path); |
0be0521b | 115 | enum scld_error ret = SCLD_OK; |
67d42212 | 116 | |
0be0521b | 117 | while (ret == SCLD_OK && next_component) { |
f0502332 | 118 | struct stat st; |
0f527403 | 119 | char *slash = next_component, slash_character; |
f0502332 | 120 | |
0f527403 MH |
121 | while (*slash && !is_dir_sep(*slash)) |
122 | slash++; | |
123 | ||
124 | if (!*slash) | |
b2cb9425 | 125 | break; |
bf10cf70 | 126 | |
26c8ae2a | 127 | next_component = slash + 1; |
0f527403 | 128 | while (is_dir_sep(*next_component)) |
bf10cf70 | 129 | next_component++; |
26c8ae2a | 130 | if (!*next_component) |
5f0bdf50 | 131 | break; |
831651fd | 132 | |
0f527403 | 133 | slash_character = *slash; |
831651fd | 134 | *slash = '\0'; |
67d42212 JR |
135 | if (!stat(path, &st)) { |
136 | /* path exists */ | |
9e6f885d | 137 | if (!S_ISDIR(st.st_mode)) |
0be0521b | 138 | ret = SCLD_EXISTS; |
53a39721 | 139 | } else if (mkdir(path, 0777)) { |
928734d9 | 140 | if (errno == EEXIST && |
9e6f885d | 141 | !stat(path, &st) && S_ISDIR(st.st_mode)) |
928734d9 | 142 | ; /* somebody created it since we checked */ |
18d37e86 MH |
143 | else if (errno == ENOENT) |
144 | /* | |
145 | * Either mkdir() failed because | |
146 | * somebody just pruned the containing | |
147 | * directory, or stat() failed because | |
148 | * the file that was in our way was | |
149 | * just removed. Either way, inform | |
150 | * the caller that it might be worth | |
151 | * trying again: | |
152 | */ | |
153 | ret = SCLD_VANISHED; | |
9e6f885d | 154 | else |
0be0521b | 155 | ret = SCLD_FAILED; |
53a39721 | 156 | } else if (adjust_shared_perm(path)) { |
0be0521b | 157 | ret = SCLD_PERMS; |
457f06d6 | 158 | } |
0f527403 | 159 | *slash = slash_character; |
b2cb9425 | 160 | } |
9e6f885d | 161 | return ret; |
b2cb9425 | 162 | } |
723c31fe | 163 | |
0be0521b | 164 | enum scld_error safe_create_leading_directories_const(const char *path) |
8e21d63b JK |
165 | { |
166 | /* path points to cache entries, so xstrdup before messing with it */ | |
167 | char *buf = xstrdup(path); | |
0be0521b | 168 | enum scld_error result = safe_create_leading_directories(buf); |
8e21d63b JK |
169 | free(buf); |
170 | return result; | |
171 | } | |
172 | ||
ace1534d JH |
173 | static void fill_sha1_path(char *pathbuf, const unsigned char *sha1) |
174 | { | |
175 | int i; | |
176 | for (i = 0; i < 20; i++) { | |
177 | static char hex[] = "0123456789abcdef"; | |
178 | unsigned int val = sha1[i]; | |
179 | char *pos = pathbuf + i*2 + (i > 0); | |
180 | *pos++ = hex[val >> 4]; | |
181 | *pos = hex[val & 0xf]; | |
182 | } | |
183 | } | |
184 | ||
30d6c6ea | 185 | const char *sha1_file_name(const unsigned char *sha1) |
0fcfd160 | 186 | { |
560fb6a1 JK |
187 | static char buf[PATH_MAX]; |
188 | const char *objdir; | |
189 | int len; | |
0fcfd160 | 190 | |
560fb6a1 JK |
191 | objdir = get_object_directory(); |
192 | len = strlen(objdir); | |
193 | ||
194 | /* '/' + sha1(2) + '/' + sha1(38) + '\0' */ | |
195 | if (len + 43 > PATH_MAX) | |
196 | die("insanely long object directory %s", objdir); | |
197 | memcpy(buf, objdir, len); | |
198 | buf[len] = '/'; | |
199 | buf[len+3] = '/'; | |
200 | buf[len+42] = '\0'; | |
201 | fill_sha1_path(buf + len + 1, sha1); | |
202 | return buf; | |
0fcfd160 LT |
203 | } |
204 | ||
d40d535b MH |
205 | /* |
206 | * Return the name of the pack or index file with the specified sha1 | |
207 | * in its filename. *base and *name are scratch space that must be | |
208 | * provided by the caller. which should be "pack" or "idx". | |
209 | */ | |
633f43e1 | 210 | static char *sha1_get_pack_name(const unsigned char *sha1, |
ac5190cc JK |
211 | struct strbuf *buf, |
212 | const char *which) | |
bf592c50 | 213 | { |
ac5190cc JK |
214 | strbuf_reset(buf); |
215 | strbuf_addf(buf, "%s/pack/pack-%s.%s", get_object_directory(), | |
216 | sha1_to_hex(sha1), which); | |
217 | return buf->buf; | |
bf592c50 DB |
218 | } |
219 | ||
633f43e1 | 220 | char *sha1_pack_name(const unsigned char *sha1) |
bf592c50 | 221 | { |
ac5190cc JK |
222 | static struct strbuf buf = STRBUF_INIT; |
223 | return sha1_get_pack_name(sha1, &buf, "pack"); | |
633f43e1 | 224 | } |
bf592c50 | 225 | |
633f43e1 HO |
226 | char *sha1_pack_index_name(const unsigned char *sha1) |
227 | { | |
ac5190cc JK |
228 | static struct strbuf buf = STRBUF_INIT; |
229 | return sha1_get_pack_name(sha1, &buf, "idx"); | |
bf592c50 DB |
230 | } |
231 | ||
d5a63b99 JH |
232 | struct alternate_object_database *alt_odb_list; |
233 | static struct alternate_object_database **alt_odb_tail; | |
ace1534d | 234 | |
ddd5d056 JH |
235 | /* |
236 | * Prepare alternate object database registry. | |
d5a63b99 JH |
237 | * |
238 | * The variable alt_odb_list points at the list of struct | |
239 | * alternate_object_database. The elements on this list come from | |
240 | * non-empty elements from colon separated ALTERNATE_DB_ENVIRONMENT | |
241 | * environment variable, and $GIT_OBJECT_DIRECTORY/info/alternates, | |
1494e038 JH |
242 | * whose contents is similar to that environment variable but can be |
243 | * LF separated. Its base points at a statically allocated buffer that | |
d5a63b99 JH |
244 | * contains "/the/directory/corresponding/to/.git/objects/...", while |
245 | * its name points just after the slash at the end of ".git/objects/" | |
246 | * in the example above, and has enough space to hold 40-byte hex | |
247 | * SHA1, an extra slash for the first level indirection, and the | |
248 | * terminating NUL. | |
ddd5d056 | 249 | */ |
539e7506 EK |
250 | static int link_alt_odb_entry(const char *entry, const char *relative_base, |
251 | int depth, const char *normalized_objdir) | |
ace1534d | 252 | { |
c2f493a4 MW |
253 | struct alternate_object_database *ent; |
254 | struct alternate_object_database *alt; | |
5bdf0a84 HW |
255 | int pfxlen, entlen; |
256 | struct strbuf pathbuf = STRBUF_INIT; | |
d5a63b99 | 257 | |
85dadc38 | 258 | if (!is_absolute_path(entry) && relative_base) { |
5bdf0a84 HW |
259 | strbuf_addstr(&pathbuf, real_path(relative_base)); |
260 | strbuf_addch(&pathbuf, '/'); | |
c2f493a4 | 261 | } |
6eac50d8 | 262 | strbuf_addstr(&pathbuf, entry); |
c2f493a4 | 263 | |
5bdf0a84 HW |
264 | normalize_path_copy(pathbuf.buf, pathbuf.buf); |
265 | ||
266 | pfxlen = strlen(pathbuf.buf); | |
267 | ||
268 | /* | |
269 | * The trailing slash after the directory name is given by | |
270 | * this function at the end. Remove duplicates. | |
271 | */ | |
272 | while (pfxlen && pathbuf.buf[pfxlen-1] == '/') | |
273 | pfxlen -= 1; | |
274 | ||
275 | entlen = pfxlen + 43; /* '/' + 2 hex + '/' + 38 hex + NUL */ | |
276 | ent = xmalloc(sizeof(*ent) + entlen); | |
277 | memcpy(ent->base, pathbuf.buf, pfxlen); | |
278 | strbuf_release(&pathbuf); | |
c2f493a4 MW |
279 | |
280 | ent->name = ent->base + pfxlen + 1; | |
281 | ent->base[pfxlen + 3] = '/'; | |
282 | ent->base[pfxlen] = ent->base[entlen-1] = 0; | |
283 | ||
284 | /* Detect cases where alternate disappeared */ | |
90b4a71c | 285 | if (!is_directory(ent->base)) { |
c2f493a4 MW |
286 | error("object directory %s does not exist; " |
287 | "check .git/objects/info/alternates.", | |
288 | ent->base); | |
289 | free(ent); | |
290 | return -1; | |
291 | } | |
292 | ||
293 | /* Prevent the common mistake of listing the same | |
294 | * thing twice, or object directory itself. | |
295 | */ | |
296 | for (alt = alt_odb_list; alt; alt = alt->next) { | |
80b47854 RS |
297 | if (pfxlen == alt->name - alt->base - 1 && |
298 | !memcmp(ent->base, alt->base, pfxlen)) { | |
c2f493a4 MW |
299 | free(ent); |
300 | return -1; | |
301 | } | |
302 | } | |
539e7506 | 303 | if (!strcmp_icase(ent->base, normalized_objdir)) { |
c2f493a4 MW |
304 | free(ent); |
305 | return -1; | |
306 | } | |
307 | ||
308 | /* add the alternate entry */ | |
309 | *alt_odb_tail = ent; | |
310 | alt_odb_tail = &(ent->next); | |
311 | ent->next = NULL; | |
312 | ||
313 | /* recursively add alternates */ | |
314 | read_info_alternates(ent->base, depth + 1); | |
315 | ||
316 | ent->base[pfxlen] = '/'; | |
317 | ||
318 | return 0; | |
319 | } | |
320 | ||
c5950164 | 321 | static void link_alt_odb_entries(const char *alt, int len, int sep, |
c2f493a4 MW |
322 | const char *relative_base, int depth) |
323 | { | |
6eac50d8 MH |
324 | struct string_list entries = STRING_LIST_INIT_NODUP; |
325 | char *alt_copy; | |
326 | int i; | |
539e7506 | 327 | struct strbuf objdirbuf = STRBUF_INIT; |
c2f493a4 MW |
328 | |
329 | if (depth > 5) { | |
330 | error("%s: ignoring alternate object stores, nesting too deep.", | |
331 | relative_base); | |
332 | return; | |
333 | } | |
334 | ||
f655651e | 335 | strbuf_add_absolute_path(&objdirbuf, get_object_directory()); |
539e7506 EK |
336 | normalize_path_copy(objdirbuf.buf, objdirbuf.buf); |
337 | ||
c5950164 | 338 | alt_copy = xmemdupz(alt, len); |
6eac50d8 MH |
339 | string_list_split_in_place(&entries, alt_copy, sep, -1); |
340 | for (i = 0; i < entries.nr; i++) { | |
341 | const char *entry = entries.items[i].string; | |
342 | if (entry[0] == '\0' || entry[0] == '#') | |
9577e7e3 | 343 | continue; |
6eac50d8 MH |
344 | if (!is_absolute_path(entry) && depth) { |
345 | error("%s: ignoring relative alternate object store %s", | |
346 | relative_base, entry); | |
347 | } else { | |
539e7506 | 348 | link_alt_odb_entry(entry, relative_base, depth, objdirbuf.buf); |
9577e7e3 | 349 | } |
9577e7e3 | 350 | } |
6eac50d8 MH |
351 | string_list_clear(&entries, 0); |
352 | free(alt_copy); | |
539e7506 | 353 | strbuf_release(&objdirbuf); |
d5a63b99 JH |
354 | } |
355 | ||
5e73633d | 356 | void read_info_alternates(const char * relative_base, int depth) |
d5a63b99 | 357 | { |
9577e7e3 | 358 | char *map; |
dc49cd76 | 359 | size_t mapsz; |
d5a63b99 | 360 | struct stat st; |
5015f01c | 361 | char *path; |
c2f493a4 | 362 | int fd; |
d5a63b99 | 363 | |
5015f01c | 364 | path = xstrfmt("%s/info/alternates", relative_base); |
f4e51683 | 365 | fd = git_open_noatime(path); |
5015f01c | 366 | free(path); |
d5a63b99 JH |
367 | if (fd < 0) |
368 | return; | |
369 | if (fstat(fd, &st) || (st.st_size == 0)) { | |
370 | close(fd); | |
9a217f2a | 371 | return; |
ace1534d | 372 | } |
dc49cd76 SP |
373 | mapsz = xsize_t(st.st_size); |
374 | map = xmmap(NULL, mapsz, PROT_READ, MAP_PRIVATE, fd, 0); | |
d5a63b99 | 375 | close(fd); |
d5a63b99 | 376 | |
c5950164 | 377 | link_alt_odb_entries(map, mapsz, '\n', relative_base, depth); |
c2f493a4 | 378 | |
dc49cd76 | 379 | munmap(map, mapsz); |
ace1534d JH |
380 | } |
381 | ||
bef70b22 DB |
382 | void add_to_alternates_file(const char *reference) |
383 | { | |
384 | struct lock_file *lock = xcalloc(1, sizeof(struct lock_file)); | |
77b9b1d1 JK |
385 | char *alts = git_pathdup("objects/info/alternates"); |
386 | FILE *in, *out; | |
387 | ||
388 | hold_lock_file_for_update(lock, alts, LOCK_DIE_ON_ERROR); | |
389 | out = fdopen_lock_file(lock, "w"); | |
390 | if (!out) | |
391 | die_errno("unable to fdopen alternates lockfile"); | |
392 | ||
393 | in = fopen(alts, "r"); | |
394 | if (in) { | |
395 | struct strbuf line = STRBUF_INIT; | |
396 | int found = 0; | |
397 | ||
398 | while (strbuf_getline(&line, in, '\n') != EOF) { | |
399 | if (!strcmp(reference, line.buf)) { | |
400 | found = 1; | |
401 | break; | |
402 | } | |
403 | fprintf_or_die(out, "%s\n", line.buf); | |
404 | } | |
405 | ||
406 | strbuf_release(&line); | |
407 | fclose(in); | |
408 | ||
409 | if (found) { | |
410 | rollback_lock_file(lock); | |
411 | lock = NULL; | |
412 | } | |
413 | } | |
414 | else if (errno != ENOENT) | |
415 | die_errno("unable to read alternates file"); | |
416 | ||
417 | if (lock) { | |
418 | fprintf_or_die(out, "%s\n", reference); | |
419 | if (commit_lock_file(lock)) | |
420 | die_errno("unable to move new alternates file into place"); | |
421 | if (alt_odb_tail) | |
422 | link_alt_odb_entries(reference, strlen(reference), '\n', NULL, 0); | |
423 | } | |
424 | free(alts); | |
bef70b22 DB |
425 | } |
426 | ||
fe1b2268 | 427 | int foreach_alt_odb(alt_odb_fn fn, void *cb) |
d79796bc JH |
428 | { |
429 | struct alternate_object_database *ent; | |
fe1b2268 | 430 | int r = 0; |
d79796bc JH |
431 | |
432 | prepare_alt_odb(); | |
fe1b2268 JK |
433 | for (ent = alt_odb_list; ent; ent = ent->next) { |
434 | r = fn(ent, cb); | |
435 | if (r) | |
436 | break; | |
437 | } | |
438 | return r; | |
d79796bc JH |
439 | } |
440 | ||
c2f493a4 MW |
441 | void prepare_alt_odb(void) |
442 | { | |
554fe20d | 443 | const char *alt; |
c2f493a4 | 444 | |
7dc24aa5 SP |
445 | if (alt_odb_tail) |
446 | return; | |
447 | ||
c2f493a4 MW |
448 | alt = getenv(ALTERNATE_DB_ENVIRONMENT); |
449 | if (!alt) alt = ""; | |
450 | ||
c2f493a4 | 451 | alt_odb_tail = &alt_odb_list; |
c5950164 | 452 | link_alt_odb_entries(alt, strlen(alt), PATH_SEP, NULL, 0); |
c2f493a4 MW |
453 | |
454 | read_info_alternates(get_object_directory(), 0); | |
455 | } | |
456 | ||
3096b2ec | 457 | /* Returns 1 if we have successfully freshened the file, 0 otherwise. */ |
33d4221c | 458 | static int freshen_file(const char *fn) |
ace1534d | 459 | { |
33d4221c JK |
460 | struct utimbuf t; |
461 | t.actime = t.modtime = time(NULL); | |
462 | return !utime(fn, &t); | |
0f4dc14a | 463 | } |
ace1534d | 464 | |
3096b2ec JK |
465 | /* |
466 | * All of the check_and_freshen functions return 1 if the file exists and was | |
467 | * freshened (if freshening was requested), 0 otherwise. If they return | |
468 | * 0, you should not assume that it is safe to skip a write of the object (it | |
469 | * either does not exist on disk, or has a stale mtime and may be subject to | |
470 | * pruning). | |
471 | */ | |
33d4221c JK |
472 | static int check_and_freshen_file(const char *fn, int freshen) |
473 | { | |
474 | if (access(fn, F_OK)) | |
475 | return 0; | |
3096b2ec | 476 | if (freshen && !freshen_file(fn)) |
33d4221c JK |
477 | return 0; |
478 | return 1; | |
479 | } | |
480 | ||
481 | static int check_and_freshen_local(const unsigned char *sha1, int freshen) | |
482 | { | |
483 | return check_and_freshen_file(sha1_file_name(sha1), freshen); | |
484 | } | |
485 | ||
486 | static int check_and_freshen_nonlocal(const unsigned char *sha1, int freshen) | |
0f4dc14a BC |
487 | { |
488 | struct alternate_object_database *alt; | |
9a217f2a | 489 | prepare_alt_odb(); |
d5a63b99 | 490 | for (alt = alt_odb_list; alt; alt = alt->next) { |
0f4dc14a | 491 | fill_sha1_path(alt->name, sha1); |
33d4221c | 492 | if (check_and_freshen_file(alt->base, freshen)) |
c529d75a | 493 | return 1; |
ace1534d | 494 | } |
c529d75a | 495 | return 0; |
ace1534d JH |
496 | } |
497 | ||
33d4221c JK |
498 | static int check_and_freshen(const unsigned char *sha1, int freshen) |
499 | { | |
500 | return check_and_freshen_local(sha1, freshen) || | |
501 | check_and_freshen_nonlocal(sha1, freshen); | |
502 | } | |
503 | ||
504 | int has_loose_object_nonlocal(const unsigned char *sha1) | |
505 | { | |
506 | return check_and_freshen_nonlocal(sha1, 0); | |
507 | } | |
508 | ||
0f4dc14a BC |
509 | static int has_loose_object(const unsigned char *sha1) |
510 | { | |
33d4221c | 511 | return check_and_freshen(sha1, 0); |
0f4dc14a BC |
512 | } |
513 | ||
60bb8b14 | 514 | static unsigned int pack_used_ctr; |
a53128b6 SP |
515 | static unsigned int pack_mmap_calls; |
516 | static unsigned int peak_pack_open_windows; | |
517 | static unsigned int pack_open_windows; | |
c7934306 SP |
518 | static unsigned int pack_open_fds; |
519 | static unsigned int pack_max_fds; | |
a53128b6 | 520 | static size_t peak_pack_mapped; |
60bb8b14 | 521 | static size_t pack_mapped; |
9a217f2a | 522 | struct packed_git *packed_git; |
1f688557 | 523 | |
b79d18c9 | 524 | void pack_report(void) |
a53128b6 SP |
525 | { |
526 | fprintf(stderr, | |
e05db0fd PR |
527 | "pack_report: getpagesize() = %10" SZ_FMT "\n" |
528 | "pack_report: core.packedGitWindowSize = %10" SZ_FMT "\n" | |
529 | "pack_report: core.packedGitLimit = %10" SZ_FMT "\n", | |
9e42d6a1 SP |
530 | sz_fmt(getpagesize()), |
531 | sz_fmt(packed_git_window_size), | |
532 | sz_fmt(packed_git_limit)); | |
a53128b6 SP |
533 | fprintf(stderr, |
534 | "pack_report: pack_used_ctr = %10u\n" | |
535 | "pack_report: pack_mmap_calls = %10u\n" | |
536 | "pack_report: pack_open_windows = %10u / %10u\n" | |
e05db0fd PR |
537 | "pack_report: pack_mapped = " |
538 | "%10" SZ_FMT " / %10" SZ_FMT "\n", | |
a53128b6 SP |
539 | pack_used_ctr, |
540 | pack_mmap_calls, | |
541 | pack_open_windows, peak_pack_open_windows, | |
9e42d6a1 | 542 | sz_fmt(pack_mapped), sz_fmt(peak_pack_mapped)); |
a53128b6 SP |
543 | } |
544 | ||
d40d535b MH |
545 | /* |
546 | * Open and mmap the index file at path, perform a couple of | |
547 | * consistency checks, then record its information to p. Return 0 on | |
548 | * success. | |
549 | */ | |
550 | static int check_packed_git_idx(const char *path, struct packed_git *p) | |
1f688557 JH |
551 | { |
552 | void *idx_map; | |
42873078 | 553 | struct pack_idx_header *hdr; |
dc49cd76 | 554 | size_t idx_size; |
74e34e1f | 555 | uint32_t version, nr, i, *index; |
f4e51683 | 556 | int fd = git_open_noatime(path); |
1f688557 | 557 | struct stat st; |
42873078 | 558 | |
1f688557 JH |
559 | if (fd < 0) |
560 | return -1; | |
561 | if (fstat(fd, &st)) { | |
562 | close(fd); | |
563 | return -1; | |
564 | } | |
dc49cd76 | 565 | idx_size = xsize_t(st.st_size); |
2d88451b SP |
566 | if (idx_size < 4 * 256 + 20 + 20) { |
567 | close(fd); | |
568 | return error("index file %s is too small", path); | |
569 | } | |
c4712e45 | 570 | idx_map = xmmap(NULL, idx_size, PROT_READ, MAP_PRIVATE, fd, 0); |
1f688557 | 571 | close(fd); |
1f688557 | 572 | |
42873078 NP |
573 | hdr = idx_map; |
574 | if (hdr->idx_signature == htonl(PACK_IDX_SIGNATURE)) { | |
74e34e1f NP |
575 | version = ntohl(hdr->idx_version); |
576 | if (version < 2 || version > 2) { | |
577 | munmap(idx_map, idx_size); | |
6e1c2344 | 578 | return error("index file %s is version %"PRIu32 |
74e34e1f NP |
579 | " and is not supported by this binary" |
580 | " (try upgrading GIT to a newer version)", | |
581 | path, version); | |
582 | } | |
583 | } else | |
584 | version = 1; | |
df1b059d | 585 | |
1f688557 | 586 | nr = 0; |
42873078 | 587 | index = idx_map; |
74e34e1f NP |
588 | if (version > 1) |
589 | index += 2; /* skip index header */ | |
1f688557 | 590 | for (i = 0; i < 256; i++) { |
326bf396 | 591 | uint32_t n = ntohl(index[i]); |
2d88451b SP |
592 | if (n < nr) { |
593 | munmap(idx_map, idx_size); | |
df1b059d | 594 | return error("non-monotonic index %s", path); |
2d88451b | 595 | } |
1f688557 JH |
596 | nr = n; |
597 | } | |
598 | ||
74e34e1f NP |
599 | if (version == 1) { |
600 | /* | |
601 | * Total size: | |
602 | * - 256 index entries 4 bytes each | |
603 | * - 24-byte entries * nr (20-byte sha1 + 4-byte offset) | |
604 | * - 20-byte SHA1 of the packfile | |
605 | * - 20-byte SHA1 file checksum | |
606 | */ | |
607 | if (idx_size != 4*256 + nr * 24 + 20 + 20) { | |
608 | munmap(idx_map, idx_size); | |
eef427a0 | 609 | return error("wrong index v1 file size in %s", path); |
74e34e1f NP |
610 | } |
611 | } else if (version == 2) { | |
612 | /* | |
613 | * Minimum size: | |
614 | * - 8 bytes of header | |
615 | * - 256 index entries 4 bytes each | |
616 | * - 20-byte sha1 entry * nr | |
617 | * - 4-byte crc entry * nr | |
618 | * - 4-byte offset entry * nr | |
619 | * - 20-byte SHA1 of the packfile | |
620 | * - 20-byte SHA1 file checksum | |
621 | * And after the 4-byte offset table might be a | |
622 | * variable sized table containing 8-byte entries | |
623 | * for offsets larger than 2^31. | |
624 | */ | |
625 | unsigned long min_size = 8 + 4*256 + nr*(20 + 4 + 4) + 20 + 20; | |
1164f1e4 LT |
626 | unsigned long max_size = min_size; |
627 | if (nr) | |
628 | max_size += (nr - 1)*8; | |
629 | if (idx_size < min_size || idx_size > max_size) { | |
74e34e1f | 630 | munmap(idx_map, idx_size); |
eef427a0 | 631 | return error("wrong index v2 file size in %s", path); |
74e34e1f | 632 | } |
7109c889 JH |
633 | if (idx_size != min_size && |
634 | /* | |
635 | * make sure we can deal with large pack offsets. | |
636 | * 31-bit signed offset won't be enough, neither | |
637 | * 32-bit unsigned one will be. | |
638 | */ | |
639 | (sizeof(off_t) <= 4)) { | |
640 | munmap(idx_map, idx_size); | |
641 | return error("pack too large for current definition of off_t in %s", path); | |
74e34e1f | 642 | } |
2d88451b | 643 | } |
1f688557 | 644 | |
74e34e1f | 645 | p->index_version = version; |
42873078 NP |
646 | p->index_data = idx_map; |
647 | p->index_size = idx_size; | |
57059091 | 648 | p->num_objects = nr; |
1f688557 JH |
649 | return 0; |
650 | } | |
651 | ||
bc8e478a | 652 | int open_pack_index(struct packed_git *p) |
d079837e SP |
653 | { |
654 | char *idx_name; | |
9ae97018 | 655 | size_t len; |
d079837e SP |
656 | int ret; |
657 | ||
658 | if (p->index_data) | |
659 | return 0; | |
660 | ||
9ae97018 JK |
661 | if (!strip_suffix(p->pack_name, ".pack", &len)) |
662 | die("BUG: pack_name does not end in .pack"); | |
663 | idx_name = xstrfmt("%.*s.idx", (int)len, p->pack_name); | |
d079837e SP |
664 | ret = check_packed_git_idx(idx_name, p); |
665 | free(idx_name); | |
666 | return ret; | |
667 | } | |
668 | ||
11daf39b SP |
669 | static void scan_windows(struct packed_git *p, |
670 | struct packed_git **lru_p, | |
671 | struct pack_window **lru_w, | |
672 | struct pack_window **lru_l) | |
1f688557 | 673 | { |
11daf39b SP |
674 | struct pack_window *w, *w_l; |
675 | ||
676 | for (w_l = NULL, w = p->windows; w; w = w->next) { | |
677 | if (!w->inuse_cnt) { | |
678 | if (!*lru_w || w->last_used < (*lru_w)->last_used) { | |
679 | *lru_p = p; | |
680 | *lru_w = w; | |
681 | *lru_l = w_l; | |
54044bf8 | 682 | } |
54044bf8 | 683 | } |
11daf39b | 684 | w_l = w; |
f9253394 | 685 | } |
11daf39b SP |
686 | } |
687 | ||
7c3ecb32 | 688 | static int unuse_one_window(struct packed_git *current) |
11daf39b SP |
689 | { |
690 | struct packed_git *p, *lru_p = NULL; | |
691 | struct pack_window *lru_w = NULL, *lru_l = NULL; | |
692 | ||
693 | if (current) | |
694 | scan_windows(current, &lru_p, &lru_w, &lru_l); | |
695 | for (p = packed_git; p; p = p->next) | |
696 | scan_windows(p, &lru_p, &lru_w, &lru_l); | |
54044bf8 SP |
697 | if (lru_p) { |
698 | munmap(lru_w->base, lru_w->len); | |
699 | pack_mapped -= lru_w->len; | |
700 | if (lru_l) | |
701 | lru_l->next = lru_w->next; | |
7c3ecb32 | 702 | else |
54044bf8 | 703 | lru_p->windows = lru_w->next; |
54044bf8 | 704 | free(lru_w); |
a53128b6 | 705 | pack_open_windows--; |
54044bf8 SP |
706 | return 1; |
707 | } | |
708 | return 0; | |
f9253394 JH |
709 | } |
710 | ||
7c3ecb32 | 711 | void release_pack_memory(size_t need) |
97bfeb34 SP |
712 | { |
713 | size_t cur = pack_mapped; | |
7c3ecb32 | 714 | while (need >= (cur - pack_mapped) && unuse_one_window(NULL)) |
97bfeb34 SP |
715 | ; /* nothing */ |
716 | } | |
717 | ||
02710228 SP |
718 | static void mmap_limit_check(size_t length) |
719 | { | |
720 | static size_t limit = 0; | |
721 | if (!limit) { | |
722 | limit = git_env_ulong("GIT_MMAP_LIMIT", 0); | |
723 | if (!limit) | |
724 | limit = SIZE_MAX; | |
725 | } | |
726 | if (length > limit) | |
727 | die("attempting to mmap %"PRIuMAX" over limit %"PRIuMAX, | |
728 | (uintmax_t)length, (uintmax_t)limit); | |
729 | } | |
730 | ||
1570856b JK |
731 | void *xmmap_gently(void *start, size_t length, |
732 | int prot, int flags, int fd, off_t offset) | |
58ecbd5e | 733 | { |
02710228 SP |
734 | void *ret; |
735 | ||
736 | mmap_limit_check(length); | |
737 | ret = mmap(start, length, prot, flags, fd, offset); | |
58ecbd5e JN |
738 | if (ret == MAP_FAILED) { |
739 | if (!length) | |
740 | return NULL; | |
7c3ecb32 | 741 | release_pack_memory(length); |
58ecbd5e | 742 | ret = mmap(start, length, prot, flags, fd, offset); |
58ecbd5e JN |
743 | } |
744 | return ret; | |
745 | } | |
746 | ||
1570856b JK |
747 | void *xmmap(void *start, size_t length, |
748 | int prot, int flags, int fd, off_t offset) | |
749 | { | |
750 | void *ret = xmmap_gently(start, length, prot, flags, fd, offset); | |
751 | if (ret == MAP_FAILED) | |
9ca0aaf6 | 752 | die_errno("mmap failed"); |
1570856b JK |
753 | return ret; |
754 | } | |
755 | ||
c9ced051 SP |
756 | void close_pack_windows(struct packed_git *p) |
757 | { | |
758 | while (p->windows) { | |
759 | struct pack_window *w = p->windows; | |
760 | ||
761 | if (w->inuse_cnt) | |
762 | die("pack '%s' still has open windows to it", | |
763 | p->pack_name); | |
764 | munmap(w->base, w->len); | |
765 | pack_mapped -= w->len; | |
766 | pack_open_windows--; | |
767 | p->windows = w->next; | |
768 | free(w); | |
769 | } | |
770 | } | |
771 | ||
71fe5d7f JS |
772 | static int close_pack_fd(struct packed_git *p) |
773 | { | |
774 | if (p->pack_fd < 0) | |
775 | return 0; | |
776 | ||
777 | close(p->pack_fd); | |
778 | pack_open_fds--; | |
779 | p->pack_fd = -1; | |
780 | ||
781 | return 1; | |
782 | } | |
783 | ||
38849a81 JS |
784 | static void close_pack(struct packed_git *p) |
785 | { | |
786 | close_pack_windows(p); | |
787 | close_pack_fd(p); | |
788 | close_pack_index(p); | |
789 | } | |
790 | ||
791 | void close_all_packs(void) | |
792 | { | |
793 | struct packed_git *p; | |
794 | ||
795 | for (p = packed_git; p; p = p->next) | |
796 | if (p->do_not_close) | |
797 | die("BUG! Want to close pack marked 'do-not-close'"); | |
798 | else | |
799 | close_pack(p); | |
800 | } | |
801 | ||
802 | ||
88d0db55 BC |
803 | /* |
804 | * The LRU pack is the one with the oldest MRU window, preferring packs | |
805 | * with no used windows, or the oldest mtime if it has no windows allocated. | |
806 | */ | |
807 | static void find_lru_pack(struct packed_git *p, struct packed_git **lru_p, struct pack_window **mru_w, int *accept_windows_inuse) | |
808 | { | |
809 | struct pack_window *w, *this_mru_w; | |
810 | int has_windows_inuse = 0; | |
811 | ||
812 | /* | |
813 | * Reject this pack if it has windows and the previously selected | |
814 | * one does not. If this pack does not have windows, reject | |
815 | * it if the pack file is newer than the previously selected one. | |
816 | */ | |
817 | if (*lru_p && !*mru_w && (p->windows || p->mtime > (*lru_p)->mtime)) | |
818 | return; | |
819 | ||
820 | for (w = this_mru_w = p->windows; w; w = w->next) { | |
821 | /* | |
822 | * Reject this pack if any of its windows are in use, | |
823 | * but the previously selected pack did not have any | |
824 | * inuse windows. Otherwise, record that this pack | |
825 | * has windows in use. | |
826 | */ | |
827 | if (w->inuse_cnt) { | |
828 | if (*accept_windows_inuse) | |
829 | has_windows_inuse = 1; | |
830 | else | |
831 | return; | |
832 | } | |
833 | ||
834 | if (w->last_used > this_mru_w->last_used) | |
835 | this_mru_w = w; | |
836 | ||
837 | /* | |
838 | * Reject this pack if it has windows that have been | |
839 | * used more recently than the previously selected pack. | |
840 | * If the previously selected pack had windows inuse and | |
841 | * we have not encountered a window in this pack that is | |
842 | * inuse, skip this check since we prefer a pack with no | |
843 | * inuse windows to one that has inuse windows. | |
844 | */ | |
845 | if (*mru_w && *accept_windows_inuse == has_windows_inuse && | |
846 | this_mru_w->last_used > (*mru_w)->last_used) | |
847 | return; | |
848 | } | |
849 | ||
850 | /* | |
851 | * Select this pack. | |
852 | */ | |
853 | *mru_w = this_mru_w; | |
854 | *lru_p = p; | |
855 | *accept_windows_inuse = has_windows_inuse; | |
856 | } | |
857 | ||
858 | static int close_one_pack(void) | |
859 | { | |
860 | struct packed_git *p, *lru_p = NULL; | |
861 | struct pack_window *mru_w = NULL; | |
862 | int accept_windows_inuse = 1; | |
863 | ||
864 | for (p = packed_git; p; p = p->next) { | |
865 | if (p->pack_fd == -1) | |
866 | continue; | |
867 | find_lru_pack(p, &lru_p, &mru_w, &accept_windows_inuse); | |
868 | } | |
869 | ||
71fe5d7f JS |
870 | if (lru_p) |
871 | return close_pack_fd(lru_p); | |
88d0db55 BC |
872 | |
873 | return 0; | |
874 | } | |
875 | ||
03e79c88 | 876 | void unuse_pack(struct pack_window **w_cursor) |
f9253394 | 877 | { |
03e79c88 SP |
878 | struct pack_window *w = *w_cursor; |
879 | if (w) { | |
880 | w->inuse_cnt--; | |
881 | *w_cursor = NULL; | |
882 | } | |
1f688557 JH |
883 | } |
884 | ||
fa5fc15d SP |
885 | void close_pack_index(struct packed_git *p) |
886 | { | |
887 | if (p->index_data) { | |
888 | munmap((void *)p->index_data, p->index_size); | |
889 | p->index_data = NULL; | |
890 | } | |
891 | } | |
892 | ||
c74faea1 NP |
893 | /* |
894 | * This is used by git-repack in case a newly created pack happens to | |
895 | * contain the same set of objects as an existing one. In that case | |
896 | * the resulting file might be different even if its name would be the | |
897 | * same. It is best to close any reference to the old pack before it is | |
01689909 | 898 | * replaced on disk. Of course no index pointers or windows for given pack |
c74faea1 NP |
899 | * must subsist at this point. If ever objects from this pack are requested |
900 | * again, the new version of the pack will be reinitialized through | |
901 | * reprepare_packed_git(). | |
902 | */ | |
903 | void free_pack_by_name(const char *pack_name) | |
904 | { | |
905 | struct packed_git *p, **pp = &packed_git; | |
906 | ||
907 | while (*pp) { | |
908 | p = *pp; | |
909 | if (strcmp(pack_name, p->pack_name) == 0) { | |
fa3a0c94 | 910 | clear_delta_base_cache(); |
38849a81 | 911 | close_pack(p); |
c74faea1 NP |
912 | free(p->bad_object_sha1); |
913 | *pp = p->next; | |
c01f51cc NTND |
914 | if (last_found_pack == p) |
915 | last_found_pack = NULL; | |
c74faea1 NP |
916 | free(p); |
917 | return; | |
918 | } | |
919 | pp = &p->next; | |
920 | } | |
921 | } | |
922 | ||
a0788266 JS |
923 | static unsigned int get_max_fd_limit(void) |
924 | { | |
925 | #ifdef RLIMIT_NOFILE | |
491a8dec JH |
926 | { |
927 | struct rlimit lim; | |
a0788266 | 928 | |
491a8dec JH |
929 | if (!getrlimit(RLIMIT_NOFILE, &lim)) |
930 | return lim.rlim_cur; | |
931 | } | |
932 | #endif | |
933 | ||
934 | #ifdef _SC_OPEN_MAX | |
935 | { | |
936 | long open_max = sysconf(_SC_OPEN_MAX); | |
937 | if (0 < open_max) | |
938 | return open_max; | |
939 | /* | |
940 | * Otherwise, we got -1 for one of the two | |
941 | * reasons: | |
942 | * | |
943 | * (1) sysconf() did not understand _SC_OPEN_MAX | |
944 | * and signaled an error with -1; or | |
945 | * (2) sysconf() said there is no limit. | |
946 | * | |
947 | * We _could_ clear errno before calling sysconf() to | |
948 | * tell these two cases apart and return a huge number | |
949 | * in the latter case to let the caller cap it to a | |
950 | * value that is not so selfish, but letting the | |
951 | * fallback OPEN_MAX codepath take care of these cases | |
952 | * is a lot simpler. | |
953 | */ | |
954 | } | |
955 | #endif | |
a0788266 | 956 | |
491a8dec | 957 | #ifdef OPEN_MAX |
a0788266 JS |
958 | return OPEN_MAX; |
959 | #else | |
960 | return 1; /* see the caller ;-) */ | |
961 | #endif | |
962 | } | |
963 | ||
3cf8b462 SP |
964 | /* |
965 | * Do not call this directly as this leaks p->pack_fd on error return; | |
966 | * call open_packed_git() instead. | |
967 | */ | |
968 | static int open_packed_git_1(struct packed_git *p) | |
1f688557 | 969 | { |
9bc879c1 SP |
970 | struct stat st; |
971 | struct pack_header hdr; | |
972 | unsigned char sha1[20]; | |
973 | unsigned char *idx_sha1; | |
2c039da8 | 974 | long fd_flag; |
9bc879c1 | 975 | |
d079837e SP |
976 | if (!p->index_data && open_pack_index(p)) |
977 | return error("packfile %s index unavailable", p->pack_name); | |
978 | ||
c7934306 | 979 | if (!pack_max_fds) { |
a0788266 | 980 | unsigned int max_fds = get_max_fd_limit(); |
c7934306 SP |
981 | |
982 | /* Save 3 for stdin/stdout/stderr, 22 for work */ | |
983 | if (25 < max_fds) | |
984 | pack_max_fds = max_fds - 25; | |
985 | else | |
986 | pack_max_fds = 1; | |
987 | } | |
988 | ||
88d0db55 | 989 | while (pack_max_fds <= pack_open_fds && close_one_pack()) |
c7934306 SP |
990 | ; /* nothing */ |
991 | ||
f4e51683 | 992 | p->pack_fd = git_open_noatime(p->pack_name); |
9bc879c1 | 993 | if (p->pack_fd < 0 || fstat(p->pack_fd, &st)) |
072db278 | 994 | return -1; |
c7934306 | 995 | pack_open_fds++; |
9bc879c1 SP |
996 | |
997 | /* If we created the struct before we had the pack we lack size. */ | |
bf592c50 | 998 | if (!p->pack_size) { |
bf592c50 | 999 | if (!S_ISREG(st.st_mode)) |
072db278 | 1000 | return error("packfile %s not a regular file", p->pack_name); |
bf592c50 | 1001 | p->pack_size = st.st_size; |
9bc879c1 | 1002 | } else if (p->pack_size != st.st_size) |
072db278 | 1003 | return error("packfile %s size changed", p->pack_name); |
9bc879c1 | 1004 | |
2c039da8 JH |
1005 | /* We leave these file descriptors open with sliding mmap; |
1006 | * there is no point keeping them open across exec(), though. | |
1007 | */ | |
1008 | fd_flag = fcntl(p->pack_fd, F_GETFD, 0); | |
1009 | if (fd_flag < 0) | |
072db278 | 1010 | return error("cannot determine file descriptor flags"); |
2c039da8 JH |
1011 | fd_flag |= FD_CLOEXEC; |
1012 | if (fcntl(p->pack_fd, F_SETFD, fd_flag) == -1) | |
072db278 | 1013 | return error("cannot set FD_CLOEXEC"); |
2c039da8 | 1014 | |
9bc879c1 | 1015 | /* Verify we recognize this pack file format. */ |
e6e2bd62 | 1016 | if (read_in_full(p->pack_fd, &hdr, sizeof(hdr)) != sizeof(hdr)) |
072db278 | 1017 | return error("file %s is far too short to be a packfile", p->pack_name); |
9bc879c1 | 1018 | if (hdr.hdr_signature != htonl(PACK_SIGNATURE)) |
072db278 | 1019 | return error("file %s is not a GIT packfile", p->pack_name); |
9bc879c1 | 1020 | if (!pack_version_ok(hdr.hdr_version)) |
6e1c2344 RJ |
1021 | return error("packfile %s is version %"PRIu32" and not" |
1022 | " supported (try upgrading GIT to a newer version)", | |
9bc879c1 SP |
1023 | p->pack_name, ntohl(hdr.hdr_version)); |
1024 | ||
1025 | /* Verify the pack matches its index. */ | |
57059091 | 1026 | if (p->num_objects != ntohl(hdr.hdr_entries)) |
6e1c2344 RJ |
1027 | return error("packfile %s claims to have %"PRIu32" objects" |
1028 | " while index indicates %"PRIu32" objects", | |
57059091 NP |
1029 | p->pack_name, ntohl(hdr.hdr_entries), |
1030 | p->num_objects); | |
9bc879c1 | 1031 | if (lseek(p->pack_fd, p->pack_size - sizeof(sha1), SEEK_SET) == -1) |
072db278 | 1032 | return error("end of packfile %s is unavailable", p->pack_name); |
e6e2bd62 | 1033 | if (read_in_full(p->pack_fd, sha1, sizeof(sha1)) != sizeof(sha1)) |
072db278 | 1034 | return error("packfile %s signature is unavailable", p->pack_name); |
42873078 | 1035 | idx_sha1 = ((unsigned char *)p->index_data) + p->index_size - 40; |
9bc879c1 | 1036 | if (hashcmp(sha1, idx_sha1)) |
072db278 SP |
1037 | return error("packfile %s does not match index", p->pack_name); |
1038 | return 0; | |
9bc879c1 SP |
1039 | } |
1040 | ||
3cf8b462 SP |
1041 | static int open_packed_git(struct packed_git *p) |
1042 | { | |
1043 | if (!open_packed_git_1(p)) | |
1044 | return 0; | |
71fe5d7f | 1045 | close_pack_fd(p); |
3cf8b462 SP |
1046 | return -1; |
1047 | } | |
1048 | ||
c4001d92 | 1049 | static int in_window(struct pack_window *win, off_t offset) |
60bb8b14 SP |
1050 | { |
1051 | /* We must promise at least 20 bytes (one hash) after the | |
1052 | * offset is available from this window, otherwise the offset | |
1053 | * is not actually in this window and a different window (which | |
1054 | * has that one hash excess) must be used. This is to support | |
1055 | * the object header and delta base parsing routines below. | |
1056 | */ | |
1057 | off_t win_off = win->offset; | |
1058 | return win_off <= offset | |
1059 | && (offset + 20) <= (win_off + win->len); | |
1060 | } | |
1061 | ||
4b25d091 | 1062 | unsigned char *use_pack(struct packed_git *p, |
03e79c88 | 1063 | struct pack_window **w_cursor, |
c4001d92 | 1064 | off_t offset, |
ef49a7a0 | 1065 | unsigned long *left) |
9bc879c1 | 1066 | { |
60bb8b14 | 1067 | struct pack_window *win = *w_cursor; |
03e79c88 | 1068 | |
a9d98a14 | 1069 | /* Since packfiles end in a hash of their content and it's |
60bb8b14 SP |
1070 | * pointless to ask for an offset into the middle of that |
1071 | * hash, and the in_window function above wouldn't match | |
1072 | * don't allow an offset too close to the end of the file. | |
1073 | */ | |
d131b7af SP |
1074 | if (!p->pack_size && p->pack_fd == -1 && open_packed_git(p)) |
1075 | die("packfile %s cannot be accessed", p->pack_name); | |
60bb8b14 SP |
1076 | if (offset > (p->pack_size - 20)) |
1077 | die("offset beyond end of packfile (truncated pack?)"); | |
1078 | ||
1079 | if (!win || !in_window(win, offset)) { | |
1080 | if (win) | |
1081 | win->inuse_cnt--; | |
1082 | for (win = p->windows; win; win = win->next) { | |
1083 | if (in_window(win, offset)) | |
1084 | break; | |
1085 | } | |
1086 | if (!win) { | |
78a28df9 | 1087 | size_t window_align = packed_git_window_size / 2; |
dc49cd76 | 1088 | off_t len; |
d131b7af SP |
1089 | |
1090 | if (p->pack_fd == -1 && open_packed_git(p)) | |
1091 | die("packfile %s cannot be accessed", p->pack_name); | |
1092 | ||
60bb8b14 | 1093 | win = xcalloc(1, sizeof(*win)); |
78a28df9 | 1094 | win->offset = (offset / window_align) * window_align; |
dc49cd76 SP |
1095 | len = p->pack_size - win->offset; |
1096 | if (len > packed_git_window_size) | |
1097 | len = packed_git_window_size; | |
1098 | win->len = (size_t)len; | |
60bb8b14 | 1099 | pack_mapped += win->len; |
11daf39b | 1100 | while (packed_git_limit < pack_mapped |
7c3ecb32 | 1101 | && unuse_one_window(p)) |
60bb8b14 | 1102 | ; /* nothing */ |
c4712e45 | 1103 | win->base = xmmap(NULL, win->len, |
60bb8b14 SP |
1104 | PROT_READ, MAP_PRIVATE, |
1105 | p->pack_fd, win->offset); | |
1106 | if (win->base == MAP_FAILED) | |
73b4e4be SP |
1107 | die("packfile %s cannot be mapped: %s", |
1108 | p->pack_name, | |
1109 | strerror(errno)); | |
d131b7af | 1110 | if (!win->offset && win->len == p->pack_size |
71fe5d7f JS |
1111 | && !p->do_not_close) |
1112 | close_pack_fd(p); | |
a53128b6 SP |
1113 | pack_mmap_calls++; |
1114 | pack_open_windows++; | |
1115 | if (pack_mapped > peak_pack_mapped) | |
1116 | peak_pack_mapped = pack_mapped; | |
1117 | if (pack_open_windows > peak_pack_open_windows) | |
1118 | peak_pack_open_windows = pack_open_windows; | |
60bb8b14 SP |
1119 | win->next = p->windows; |
1120 | p->windows = win; | |
1121 | } | |
1f688557 | 1122 | } |
03e79c88 SP |
1123 | if (win != *w_cursor) { |
1124 | win->last_used = pack_used_ctr++; | |
1125 | win->inuse_cnt++; | |
1126 | *w_cursor = win; | |
1127 | } | |
60bb8b14 | 1128 | offset -= win->offset; |
03e79c88 | 1129 | if (left) |
dc49cd76 | 1130 | *left = win->len - xsize_t(offset); |
03e79c88 | 1131 | return win->base + offset; |
1f688557 JH |
1132 | } |
1133 | ||
27d69a46 NP |
1134 | static struct packed_git *alloc_packed_git(int extra) |
1135 | { | |
1136 | struct packed_git *p = xmalloc(sizeof(*p) + extra); | |
1137 | memset(p, 0, sizeof(*p)); | |
1138 | p->pack_fd = -1; | |
1139 | return p; | |
1140 | } | |
1141 | ||
e0500293 JN |
1142 | static void try_to_free_pack_memory(size_t size) |
1143 | { | |
7c3ecb32 | 1144 | release_pack_memory(size); |
e0500293 JN |
1145 | } |
1146 | ||
48bcc1c3 | 1147 | struct packed_git *add_packed_git(const char *path, size_t path_len, int local) |
1f688557 | 1148 | { |
e0500293 | 1149 | static int have_set_try_to_free_routine; |
1f688557 | 1150 | struct stat st; |
48bcc1c3 JK |
1151 | size_t alloc; |
1152 | struct packed_git *p; | |
1f688557 | 1153 | |
e0500293 JN |
1154 | if (!have_set_try_to_free_routine) { |
1155 | have_set_try_to_free_routine = 1; | |
1156 | set_try_to_free_routine(try_to_free_pack_memory); | |
1157 | } | |
1158 | ||
42873078 NP |
1159 | /* |
1160 | * Make sure a corresponding .pack file exists and that | |
1161 | * the index looks sane. | |
1162 | */ | |
48bcc1c3 | 1163 | if (!strip_suffix_mem(path, &path_len, ".idx")) |
1f688557 | 1164 | return NULL; |
48bcc1c3 JK |
1165 | |
1166 | /* | |
1167 | * ".pack" is long enough to hold any suffix we're adding (and | |
1168 | * the use xsnprintf double-checks that) | |
1169 | */ | |
1170 | alloc = path_len + strlen(".pack") + 1; | |
1171 | p = alloc_packed_git(alloc); | |
42873078 | 1172 | memcpy(p->pack_name, path, path_len); |
8d25931d | 1173 | |
48bcc1c3 | 1174 | xsnprintf(p->pack_name + path_len, alloc - path_len, ".keep"); |
8d25931d BC |
1175 | if (!access(p->pack_name, F_OK)) |
1176 | p->pack_keep = 1; | |
1177 | ||
48bcc1c3 | 1178 | xsnprintf(p->pack_name + path_len, alloc - path_len, ".pack"); |
d079837e | 1179 | if (stat(p->pack_name, &st) || !S_ISREG(st.st_mode)) { |
42873078 | 1180 | free(p); |
1f688557 JH |
1181 | return NULL; |
1182 | } | |
42873078 | 1183 | |
1f688557 JH |
1184 | /* ok, it looks sane as far as we can check without |
1185 | * actually mapping the pack file. | |
1186 | */ | |
1f688557 | 1187 | p->pack_size = st.st_size; |
9d835df2 | 1188 | p->pack_local = local; |
b867092f | 1189 | p->mtime = st.st_mtime; |
42873078 NP |
1190 | if (path_len < 40 || get_sha1_hex(path + path_len - 40, p->sha1)) |
1191 | hashclr(p->sha1); | |
1f688557 JH |
1192 | return p; |
1193 | } | |
1194 | ||
7b64469a | 1195 | struct packed_git *parse_pack_index(unsigned char *sha1, const char *idx_path) |
c508df5e | 1196 | { |
42873078 | 1197 | const char *path = sha1_pack_name(sha1); |
c7ab0ba3 JK |
1198 | int alloc = strlen(path) + 1; |
1199 | struct packed_git *p = alloc_packed_git(alloc); | |
bf592c50 | 1200 | |
c7ab0ba3 | 1201 | memcpy(p->pack_name, path, alloc); /* includes NUL */ |
27d69a46 | 1202 | hashcpy(p->sha1, sha1); |
42873078 NP |
1203 | if (check_packed_git_idx(idx_path, p)) { |
1204 | free(p); | |
bf592c50 | 1205 | return NULL; |
42873078 | 1206 | } |
bf592c50 | 1207 | |
bf592c50 DB |
1208 | return p; |
1209 | } | |
1210 | ||
1211 | void install_packed_git(struct packed_git *pack) | |
1212 | { | |
c7934306 SP |
1213 | if (pack->pack_fd != -1) |
1214 | pack_open_fds++; | |
1215 | ||
bf592c50 DB |
1216 | pack->next = packed_git; |
1217 | packed_git = pack; | |
1218 | } | |
1219 | ||
543c5caa NTND |
1220 | void (*report_garbage)(const char *desc, const char *path); |
1221 | ||
1222 | static void report_helper(const struct string_list *list, | |
1223 | int seen_bits, int first, int last) | |
1224 | { | |
1225 | const char *msg; | |
1226 | switch (seen_bits) { | |
1227 | case 0: | |
235e8d59 | 1228 | msg = "no corresponding .idx or .pack"; |
543c5caa NTND |
1229 | break; |
1230 | case 1: | |
1231 | msg = "no corresponding .idx"; | |
1232 | break; | |
1233 | case 2: | |
1234 | msg = "no corresponding .pack"; | |
1235 | break; | |
1236 | default: | |
1237 | return; | |
1238 | } | |
1239 | for (; first < last; first++) | |
1240 | report_garbage(msg, list->items[first].string); | |
1241 | } | |
1242 | ||
1243 | static void report_pack_garbage(struct string_list *list) | |
1244 | { | |
1245 | int i, baselen = -1, first = 0, seen_bits = 0; | |
1246 | ||
1247 | if (!report_garbage) | |
1248 | return; | |
1249 | ||
3383e199 | 1250 | string_list_sort(list); |
543c5caa NTND |
1251 | |
1252 | for (i = 0; i < list->nr; i++) { | |
1253 | const char *path = list->items[i].string; | |
1254 | if (baselen != -1 && | |
1255 | strncmp(path, list->items[first].string, baselen)) { | |
1256 | report_helper(list, seen_bits, first, i); | |
1257 | baselen = -1; | |
1258 | seen_bits = 0; | |
1259 | } | |
1260 | if (baselen == -1) { | |
1261 | const char *dot = strrchr(path, '.'); | |
1262 | if (!dot) { | |
1263 | report_garbage("garbage found", path); | |
1264 | continue; | |
1265 | } | |
1266 | baselen = dot - path + 1; | |
1267 | first = i; | |
1268 | } | |
1269 | if (!strcmp(path + baselen, "pack")) | |
1270 | seen_bits |= 1; | |
1271 | else if (!strcmp(path + baselen, "idx")) | |
1272 | seen_bits |= 2; | |
1273 | } | |
1274 | report_helper(list, seen_bits, first, list->nr); | |
1275 | } | |
1276 | ||
9d835df2 | 1277 | static void prepare_packed_git_one(char *objdir, int local) |
1f688557 | 1278 | { |
880fb8de RS |
1279 | struct strbuf path = STRBUF_INIT; |
1280 | size_t dirnamelen; | |
1f688557 JH |
1281 | DIR *dir; |
1282 | struct dirent *de; | |
543c5caa | 1283 | struct string_list garbage = STRING_LIST_INIT_DUP; |
1f688557 | 1284 | |
880fb8de RS |
1285 | strbuf_addstr(&path, objdir); |
1286 | strbuf_addstr(&path, "/pack"); | |
1287 | dir = opendir(path.buf); | |
b5b16990 | 1288 | if (!dir) { |
26125f6b | 1289 | if (errno != ENOENT) |
bd2afde8 | 1290 | error("unable to open object pack directory: %s: %s", |
880fb8de RS |
1291 | path.buf, strerror(errno)); |
1292 | strbuf_release(&path); | |
1f688557 | 1293 | return; |
b5b16990 | 1294 | } |
880fb8de RS |
1295 | strbuf_addch(&path, '/'); |
1296 | dirnamelen = path.len; | |
1f688557 | 1297 | while ((de = readdir(dir)) != NULL) { |
1f688557 | 1298 | struct packed_git *p; |
47bf4b0f | 1299 | size_t base_len; |
543c5caa NTND |
1300 | |
1301 | if (is_dot_or_dotdot(de->d_name)) | |
9cb18f56 JM |
1302 | continue; |
1303 | ||
880fb8de RS |
1304 | strbuf_setlen(&path, dirnamelen); |
1305 | strbuf_addstr(&path, de->d_name); | |
d90906a9 | 1306 | |
47bf4b0f JK |
1307 | base_len = path.len; |
1308 | if (strip_suffix_mem(path.buf, &base_len, ".idx")) { | |
d90906a9 NTND |
1309 | /* Don't reopen a pack we already have. */ |
1310 | for (p = packed_git; p; p = p->next) { | |
47bf4b0f JK |
1311 | size_t len; |
1312 | if (strip_suffix(p->pack_name, ".pack", &len) && | |
1313 | len == base_len && | |
1314 | !memcmp(p->pack_name, path.buf, len)) | |
d90906a9 NTND |
1315 | break; |
1316 | } | |
1317 | if (p == NULL && | |
1318 | /* | |
1319 | * See if it really is a valid .idx file with | |
1320 | * corresponding .pack file that we can map. | |
1321 | */ | |
880fb8de | 1322 | (p = add_packed_git(path.buf, path.len, local)) != NULL) |
d90906a9 | 1323 | install_packed_git(p); |
86f7780c | 1324 | } |
543c5caa NTND |
1325 | |
1326 | if (!report_garbage) | |
1327 | continue; | |
1328 | ||
2975c770 JK |
1329 | if (ends_with(de->d_name, ".idx") || |
1330 | ends_with(de->d_name, ".pack") || | |
1331 | ends_with(de->d_name, ".bitmap") || | |
1332 | ends_with(de->d_name, ".keep")) | |
880fb8de | 1333 | string_list_append(&garbage, path.buf); |
543c5caa | 1334 | else |
880fb8de | 1335 | report_garbage("garbage found", path.buf); |
1f688557 | 1336 | } |
5b35bcd5 | 1337 | closedir(dir); |
543c5caa NTND |
1338 | report_pack_garbage(&garbage); |
1339 | string_list_clear(&garbage, 0); | |
880fb8de | 1340 | strbuf_release(&path); |
1f688557 JH |
1341 | } |
1342 | ||
b867092f JH |
1343 | static int sort_pack(const void *a_, const void *b_) |
1344 | { | |
1345 | struct packed_git *a = *((struct packed_git **)a_); | |
1346 | struct packed_git *b = *((struct packed_git **)b_); | |
1347 | int st; | |
1348 | ||
1349 | /* | |
1350 | * Local packs tend to contain objects specific to our | |
1351 | * variant of the project than remote ones. In addition, | |
1352 | * remote ones could be on a network mounted filesystem. | |
1353 | * Favor local ones for these reasons. | |
1354 | */ | |
1355 | st = a->pack_local - b->pack_local; | |
1356 | if (st) | |
1357 | return -st; | |
1358 | ||
1359 | /* | |
1360 | * Younger packs tend to contain more recent objects, | |
1361 | * and more recent objects tend to get accessed more | |
1362 | * often. | |
1363 | */ | |
1364 | if (a->mtime < b->mtime) | |
1365 | return 1; | |
1366 | else if (a->mtime == b->mtime) | |
1367 | return 0; | |
1368 | return -1; | |
1369 | } | |
1370 | ||
1371 | static void rearrange_packed_git(void) | |
1372 | { | |
1373 | struct packed_git **ary, *p; | |
1374 | int i, n; | |
1375 | ||
1376 | for (n = 0, p = packed_git; p; p = p->next) | |
1377 | n++; | |
1378 | if (n < 2) | |
1379 | return; | |
1380 | ||
1381 | /* prepare an array of packed_git for easier sorting */ | |
1382 | ary = xcalloc(n, sizeof(struct packed_git *)); | |
1383 | for (n = 0, p = packed_git; p; p = p->next) | |
1384 | ary[n++] = p; | |
1385 | ||
1386 | qsort(ary, n, sizeof(struct packed_git *), sort_pack); | |
1387 | ||
1388 | /* link them back again */ | |
1389 | for (i = 0; i < n - 1; i++) | |
1390 | ary[i]->next = ary[i + 1]; | |
1391 | ary[n - 1]->next = NULL; | |
1392 | packed_git = ary[0]; | |
1393 | ||
1394 | free(ary); | |
1395 | } | |
1396 | ||
637cdd9d | 1397 | static int prepare_packed_git_run_once = 0; |
9a217f2a | 1398 | void prepare_packed_git(void) |
1f688557 | 1399 | { |
d5a63b99 | 1400 | struct alternate_object_database *alt; |
1f688557 | 1401 | |
637cdd9d | 1402 | if (prepare_packed_git_run_once) |
1f688557 | 1403 | return; |
9d835df2 | 1404 | prepare_packed_git_one(get_object_directory(), 1); |
9a217f2a | 1405 | prepare_alt_odb(); |
d5a63b99 | 1406 | for (alt = alt_odb_list; alt; alt = alt->next) { |
1494e038 | 1407 | alt->name[-1] = 0; |
9d835df2 | 1408 | prepare_packed_git_one(alt->base, 0); |
1494e038 | 1409 | alt->name[-1] = '/'; |
1f688557 | 1410 | } |
b867092f | 1411 | rearrange_packed_git(); |
637cdd9d JK |
1412 | prepare_packed_git_run_once = 1; |
1413 | } | |
1414 | ||
fc04c412 | 1415 | void reprepare_packed_git(void) |
637cdd9d JK |
1416 | { |
1417 | prepare_packed_git_run_once = 0; | |
1418 | prepare_packed_git(); | |
1f688557 JH |
1419 | } |
1420 | ||
8eca0b47 NP |
1421 | static void mark_bad_packed_object(struct packed_git *p, |
1422 | const unsigned char *sha1) | |
1423 | { | |
1424 | unsigned i; | |
1425 | for (i = 0; i < p->num_bad_objects; i++) | |
1426 | if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i)) | |
1427 | return; | |
1428 | p->bad_object_sha1 = xrealloc(p->bad_object_sha1, 20 * (p->num_bad_objects + 1)); | |
1429 | hashcpy(p->bad_object_sha1 + 20 * p->num_bad_objects, sha1); | |
1430 | p->num_bad_objects++; | |
1431 | } | |
1432 | ||
b6c4cecc | 1433 | static const struct packed_git *has_packed_and_bad(const unsigned char *sha1) |
ac939109 NP |
1434 | { |
1435 | struct packed_git *p; | |
1436 | unsigned i; | |
1437 | ||
1438 | for (p = packed_git; p; p = p->next) | |
1439 | for (i = 0; i < p->num_bad_objects; i++) | |
1440 | if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i)) | |
b6c4cecc JH |
1441 | return p; |
1442 | return NULL; | |
ac939109 NP |
1443 | } |
1444 | ||
090ea126 NTND |
1445 | /* |
1446 | * With an in-core object data in "map", rehash it to make sure the | |
1447 | * object name actually matches "sha1" to detect object corruption. | |
1448 | * With "map" == NULL, try reading the object named with "sha1" using | |
1449 | * the streaming interface and rehash it to do the same. | |
1450 | */ | |
1451 | int check_sha1_signature(const unsigned char *sha1, void *map, | |
1452 | unsigned long size, const char *type) | |
0fcfd160 LT |
1453 | { |
1454 | unsigned char real_sha1[20]; | |
090ea126 NTND |
1455 | enum object_type obj_type; |
1456 | struct git_istream *st; | |
1457 | git_SHA_CTX c; | |
1458 | char hdr[32]; | |
1459 | int hdrlen; | |
1460 | ||
1461 | if (map) { | |
1462 | hash_sha1_file(map, size, type, real_sha1); | |
1463 | return hashcmp(sha1, real_sha1) ? -1 : 0; | |
1464 | } | |
1465 | ||
1466 | st = open_istream(sha1, &obj_type, &size, NULL); | |
1467 | if (!st) | |
1468 | return -1; | |
1469 | ||
1470 | /* Generate the header */ | |
ef1286d3 | 1471 | hdrlen = xsnprintf(hdr, sizeof(hdr), "%s %lu", typename(obj_type), size) + 1; |
090ea126 NTND |
1472 | |
1473 | /* Sha1.. */ | |
1474 | git_SHA1_Init(&c); | |
1475 | git_SHA1_Update(&c, hdr, hdrlen); | |
1476 | for (;;) { | |
1477 | char buf[1024 * 16]; | |
1478 | ssize_t readlen = read_istream(st, buf, sizeof(buf)); | |
1479 | ||
f54fac53 JK |
1480 | if (readlen < 0) { |
1481 | close_istream(st); | |
1482 | return -1; | |
1483 | } | |
090ea126 NTND |
1484 | if (!readlen) |
1485 | break; | |
1486 | git_SHA1_Update(&c, buf, readlen); | |
1487 | } | |
1488 | git_SHA1_Final(real_sha1, &c); | |
1489 | close_istream(st); | |
a89fccd2 | 1490 | return hashcmp(sha1, real_sha1) ? -1 : 0; |
0fcfd160 LT |
1491 | } |
1492 | ||
ec73f580 | 1493 | int git_open_noatime(const char *name) |
44d1c19e LT |
1494 | { |
1495 | static int sha1_file_open_flag = O_NOATIME; | |
44d1c19e | 1496 | |
f2e872aa | 1497 | for (;;) { |
dff6f280 CB |
1498 | int fd; |
1499 | ||
1500 | errno = 0; | |
1501 | fd = open(name, O_RDONLY | sha1_file_open_flag); | |
44d1c19e | 1502 | if (fd >= 0) |
f2e872aa SP |
1503 | return fd; |
1504 | ||
f2e872aa SP |
1505 | /* Might the failure be due to O_NOATIME? */ |
1506 | if (errno != ENOENT && sha1_file_open_flag) { | |
44d1c19e | 1507 | sha1_file_open_flag = 0; |
f2e872aa SP |
1508 | continue; |
1509 | } | |
1510 | ||
1511 | return -1; | |
44d1c19e | 1512 | } |
44d1c19e LT |
1513 | } |
1514 | ||
052fe5ea JK |
1515 | static int stat_sha1_file(const unsigned char *sha1, struct stat *st) |
1516 | { | |
052fe5ea JK |
1517 | struct alternate_object_database *alt; |
1518 | ||
30d6c6ea | 1519 | if (!lstat(sha1_file_name(sha1), st)) |
052fe5ea JK |
1520 | return 0; |
1521 | ||
1522 | prepare_alt_odb(); | |
1523 | errno = ENOENT; | |
1524 | for (alt = alt_odb_list; alt; alt = alt->next) { | |
30d6c6ea | 1525 | fill_sha1_path(alt->name, sha1); |
052fe5ea JK |
1526 | if (!lstat(alt->base, st)) |
1527 | return 0; | |
1528 | } | |
1529 | ||
1530 | return -1; | |
1531 | } | |
1532 | ||
44d1c19e LT |
1533 | static int open_sha1_file(const unsigned char *sha1) |
1534 | { | |
1535 | int fd; | |
44d1c19e | 1536 | struct alternate_object_database *alt; |
d6c8a05b | 1537 | int most_interesting_errno; |
44d1c19e | 1538 | |
30d6c6ea | 1539 | fd = git_open_noatime(sha1_file_name(sha1)); |
44d1c19e LT |
1540 | if (fd >= 0) |
1541 | return fd; | |
d6c8a05b | 1542 | most_interesting_errno = errno; |
44d1c19e LT |
1543 | |
1544 | prepare_alt_odb(); | |
44d1c19e | 1545 | for (alt = alt_odb_list; alt; alt = alt->next) { |
30d6c6ea | 1546 | fill_sha1_path(alt->name, sha1); |
f4e51683 | 1547 | fd = git_open_noatime(alt->base); |
44d1c19e LT |
1548 | if (fd >= 0) |
1549 | return fd; | |
d6c8a05b JK |
1550 | if (most_interesting_errno == ENOENT) |
1551 | most_interesting_errno = errno; | |
44d1c19e | 1552 | } |
d6c8a05b | 1553 | errno = most_interesting_errno; |
44d1c19e LT |
1554 | return -1; |
1555 | } | |
1556 | ||
f0270efd | 1557 | void *map_sha1_file(const unsigned char *sha1, unsigned long *size) |
0fcfd160 | 1558 | { |
0fcfd160 | 1559 | void *map; |
144bde78 | 1560 | int fd; |
ace1534d | 1561 | |
44d1c19e LT |
1562 | fd = open_sha1_file(sha1); |
1563 | map = NULL; | |
1564 | if (fd >= 0) { | |
1565 | struct stat st; | |
0fcfd160 | 1566 | |
44d1c19e LT |
1567 | if (!fstat(fd, &st)) { |
1568 | *size = xsize_t(st.st_size); | |
33e42de0 MM |
1569 | if (!*size) { |
1570 | /* mmap() is forbidden on empty files */ | |
1571 | error("object file %s is empty", sha1_file_name(sha1)); | |
1572 | return NULL; | |
1573 | } | |
44d1c19e | 1574 | map = xmmap(NULL, *size, PROT_READ, MAP_PRIVATE, fd, 0); |
144bde78 | 1575 | } |
44d1c19e | 1576 | close(fd); |
0fcfd160 | 1577 | } |
0fcfd160 LT |
1578 | return map; |
1579 | } | |
1580 | ||
09ded04b NP |
1581 | unsigned long unpack_object_header_buffer(const unsigned char *buf, |
1582 | unsigned long len, enum object_type *type, unsigned long *sizep) | |
c4483576 | 1583 | { |
ad1ed5ee | 1584 | unsigned shift; |
48fb7deb | 1585 | unsigned long size, c; |
ad1ed5ee JH |
1586 | unsigned long used = 0; |
1587 | ||
1588 | c = buf[used++]; | |
1589 | *type = (c >> 4) & 7; | |
1590 | size = c & 15; | |
1591 | shift = 4; | |
1592 | while (c & 0x80) { | |
f630cfda | 1593 | if (len <= used || bitsizeof(long) <= shift) { |
09ded04b | 1594 | error("bad object header"); |
ea4f9685 JH |
1595 | size = used = 0; |
1596 | break; | |
09ded04b | 1597 | } |
ad1ed5ee JH |
1598 | c = buf[used++]; |
1599 | size += (c & 0x7f) << shift; | |
1600 | shift += 7; | |
1601 | } | |
1602 | *sizep = size; | |
1603 | return used; | |
1604 | } | |
1605 | ||
eb4f4076 | 1606 | int unpack_sha1_header(git_zstream *stream, unsigned char *map, unsigned long mapsize, void *buffer, unsigned long bufsiz) |
ad1ed5ee | 1607 | { |
c4483576 LT |
1608 | /* Get the data stream */ |
1609 | memset(stream, 0, sizeof(*stream)); | |
1610 | stream->next_in = map; | |
1611 | stream->avail_in = mapsize; | |
1612 | stream->next_out = buffer; | |
93821bd9 LT |
1613 | stream->avail_out = bufsiz; |
1614 | ||
39c68542 | 1615 | git_inflate_init(stream); |
cc5c54e7 | 1616 | return git_inflate(stream, 0); |
c4483576 LT |
1617 | } |
1618 | ||
46f03448 KN |
1619 | static int unpack_sha1_header_to_strbuf(git_zstream *stream, unsigned char *map, |
1620 | unsigned long mapsize, void *buffer, | |
1621 | unsigned long bufsiz, struct strbuf *header) | |
1622 | { | |
1623 | int status; | |
1624 | ||
1625 | status = unpack_sha1_header(stream, map, mapsize, buffer, bufsiz); | |
1626 | ||
1627 | /* | |
1628 | * Check if entire header is unpacked in the first iteration. | |
1629 | */ | |
1630 | if (memchr(buffer, '\0', stream->next_out - (unsigned char *)buffer)) | |
1631 | return 0; | |
1632 | ||
1633 | /* | |
1634 | * buffer[0..bufsiz] was not large enough. Copy the partial | |
1635 | * result out to header, and then append the result of further | |
1636 | * reading the stream. | |
1637 | */ | |
1638 | strbuf_add(header, buffer, stream->next_out - (unsigned char *)buffer); | |
1639 | stream->next_out = buffer; | |
1640 | stream->avail_out = bufsiz; | |
1641 | ||
1642 | do { | |
1643 | status = git_inflate(stream, 0); | |
1644 | strbuf_add(header, buffer, stream->next_out - (unsigned char *)buffer); | |
1645 | if (memchr(buffer, '\0', stream->next_out - (unsigned char *)buffer)) | |
1646 | return 0; | |
1647 | stream->next_out = buffer; | |
1648 | stream->avail_out = bufsiz; | |
1649 | } while (status != Z_STREAM_END); | |
1650 | return -1; | |
1651 | } | |
1652 | ||
ef49a7a0 | 1653 | static void *unpack_sha1_rest(git_zstream *stream, void *buffer, unsigned long size, const unsigned char *sha1) |
5180cacc LT |
1654 | { |
1655 | int bytes = strlen(buffer) + 1; | |
3aee68aa | 1656 | unsigned char *buf = xmallocz(size); |
93821bd9 | 1657 | unsigned long n; |
7efbff75 | 1658 | int status = Z_OK; |
5180cacc | 1659 | |
93821bd9 LT |
1660 | n = stream->total_out - bytes; |
1661 | if (n > size) | |
1662 | n = size; | |
1663 | memcpy(buf, (char *) buffer + bytes, n); | |
1664 | bytes = n; | |
456cdf6e LT |
1665 | if (bytes <= size) { |
1666 | /* | |
1667 | * The above condition must be (bytes <= size), not | |
1668 | * (bytes < size). In other words, even though we | |
ccf5ace0 | 1669 | * expect no more output and set avail_out to zero, |
456cdf6e LT |
1670 | * the input zlib stream may have bytes that express |
1671 | * "this concludes the stream", and we *do* want to | |
1672 | * eat that input. | |
1673 | * | |
1674 | * Otherwise we would not be able to test that we | |
1675 | * consumed all the input to reach the expected size; | |
1676 | * we also want to check that zlib tells us that all | |
1677 | * went well with status == Z_STREAM_END at the end. | |
1678 | */ | |
5180cacc LT |
1679 | stream->next_out = buf + bytes; |
1680 | stream->avail_out = size - bytes; | |
7efbff75 | 1681 | while (status == Z_OK) |
39c68542 | 1682 | status = git_inflate(stream, Z_FINISH); |
5180cacc | 1683 | } |
456cdf6e | 1684 | if (status == Z_STREAM_END && !stream->avail_in) { |
39c68542 | 1685 | git_inflate_end(stream); |
7efbff75 JH |
1686 | return buf; |
1687 | } | |
1688 | ||
1689 | if (status < 0) | |
1690 | error("corrupt loose object '%s'", sha1_to_hex(sha1)); | |
1691 | else if (stream->avail_in) | |
1692 | error("garbage at end of loose object '%s'", | |
1693 | sha1_to_hex(sha1)); | |
1694 | free(buf); | |
1695 | return NULL; | |
5180cacc LT |
1696 | } |
1697 | ||
1698 | /* | |
1699 | * We used to just use "sscanf()", but that's actually way | |
1700 | * too permissive for what we want to check. So do an anal | |
1701 | * object header parse by hand. | |
1702 | */ | |
46f03448 KN |
1703 | static int parse_sha1_header_extended(const char *hdr, struct object_info *oi, |
1704 | unsigned int flags) | |
5180cacc | 1705 | { |
46f03448 | 1706 | const char *type_buf = hdr; |
5180cacc | 1707 | unsigned long size; |
46f03448 | 1708 | int type, type_len = 0; |
5180cacc LT |
1709 | |
1710 | /* | |
46f03448 | 1711 | * The type can be of any size but is followed by |
21666f1a | 1712 | * a space. |
5180cacc | 1713 | */ |
5180cacc LT |
1714 | for (;;) { |
1715 | char c = *hdr++; | |
1716 | if (c == ' ') | |
1717 | break; | |
46f03448 | 1718 | type_len++; |
5180cacc | 1719 | } |
46f03448 KN |
1720 | |
1721 | type = type_from_string_gently(type_buf, type_len, 1); | |
1722 | if (oi->typename) | |
1723 | strbuf_add(oi->typename, type_buf, type_len); | |
1724 | /* | |
1725 | * Set type to 0 if its an unknown object and | |
1726 | * we're obtaining the type using '--allow-unkown-type' | |
1727 | * option. | |
1728 | */ | |
1729 | if ((flags & LOOKUP_UNKNOWN_OBJECT) && (type < 0)) | |
1730 | type = 0; | |
1731 | else if (type < 0) | |
1732 | die("invalid object type"); | |
1733 | if (oi->typep) | |
1734 | *oi->typep = type; | |
5180cacc LT |
1735 | |
1736 | /* | |
1737 | * The length must follow immediately, and be in canonical | |
1738 | * decimal format (ie "010" is not valid). | |
1739 | */ | |
1740 | size = *hdr++ - '0'; | |
1741 | if (size > 9) | |
1742 | return -1; | |
1743 | if (size) { | |
1744 | for (;;) { | |
1745 | unsigned long c = *hdr - '0'; | |
1746 | if (c > 9) | |
1747 | break; | |
1748 | hdr++; | |
1749 | size = size * 10 + c; | |
1750 | } | |
1751 | } | |
46f03448 KN |
1752 | |
1753 | if (oi->sizep) | |
1754 | *oi->sizep = size; | |
5180cacc LT |
1755 | |
1756 | /* | |
1757 | * The length must be followed by a zero byte | |
1758 | */ | |
46f03448 KN |
1759 | return *hdr ? -1 : type; |
1760 | } | |
1761 | ||
1762 | int parse_sha1_header(const char *hdr, unsigned long *sizep) | |
1763 | { | |
1764 | struct object_info oi; | |
1765 | ||
1766 | oi.sizep = sizep; | |
1767 | oi.typename = NULL; | |
1768 | oi.typep = NULL; | |
1769 | return parse_sha1_header_extended(hdr, &oi, LOOKUP_REPLACE_OBJECT); | |
5180cacc LT |
1770 | } |
1771 | ||
7efbff75 | 1772 | static void *unpack_sha1_file(void *map, unsigned long mapsize, enum object_type *type, unsigned long *size, const unsigned char *sha1) |
0fcfd160 | 1773 | { |
5180cacc | 1774 | int ret; |
ef49a7a0 | 1775 | git_zstream stream; |
5180cacc | 1776 | char hdr[8192]; |
0fcfd160 | 1777 | |
5180cacc | 1778 | ret = unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)); |
21666f1a | 1779 | if (ret < Z_OK || (*type = parse_sha1_header(hdr, size)) < 0) |
0fcfd160 LT |
1780 | return NULL; |
1781 | ||
7efbff75 | 1782 | return unpack_sha1_rest(&stream, hdr, *size, sha1); |
0fcfd160 LT |
1783 | } |
1784 | ||
54dab52a NP |
1785 | unsigned long get_size_from_delta(struct packed_git *p, |
1786 | struct pack_window **w_curs, | |
1787 | off_t curpos) | |
1788 | { | |
1789 | const unsigned char *data; | |
1790 | unsigned char delta_head[20], *in; | |
ef49a7a0 | 1791 | git_zstream stream; |
54dab52a NP |
1792 | int st; |
1793 | ||
1794 | memset(&stream, 0, sizeof(stream)); | |
1795 | stream.next_out = delta_head; | |
1796 | stream.avail_out = sizeof(delta_head); | |
1797 | ||
39c68542 | 1798 | git_inflate_init(&stream); |
54dab52a NP |
1799 | do { |
1800 | in = use_pack(p, w_curs, curpos, &stream.avail_in); | |
1801 | stream.next_in = in; | |
39c68542 | 1802 | st = git_inflate(&stream, Z_FINISH); |
54dab52a NP |
1803 | curpos += stream.next_in - in; |
1804 | } while ((st == Z_OK || st == Z_BUF_ERROR) && | |
1805 | stream.total_out < sizeof(delta_head)); | |
39c68542 | 1806 | git_inflate_end(&stream); |
3d77d877 NP |
1807 | if ((st != Z_STREAM_END) && stream.total_out != sizeof(delta_head)) { |
1808 | error("delta data unpack-initial failed"); | |
1809 | return 0; | |
1810 | } | |
54dab52a NP |
1811 | |
1812 | /* Examine the initial part of the delta to figure out | |
1813 | * the result size. | |
1814 | */ | |
1815 | data = delta_head; | |
1816 | ||
1817 | /* ignore base size */ | |
1818 | get_delta_hdr_size(&data, delta_head+sizeof(delta_head)); | |
1819 | ||
1820 | /* Read the result size */ | |
1821 | return get_delta_hdr_size(&data, delta_head+sizeof(delta_head)); | |
1822 | } | |
1823 | ||
c4001d92 | 1824 | static off_t get_delta_base(struct packed_git *p, |
03e79c88 | 1825 | struct pack_window **w_curs, |
c4001d92 | 1826 | off_t *curpos, |
21666f1a | 1827 | enum object_type type, |
c4001d92 | 1828 | off_t delta_obj_offset) |
eb32d236 | 1829 | { |
2b87c45b | 1830 | unsigned char *base_info = use_pack(p, w_curs, *curpos, NULL); |
c4001d92 | 1831 | off_t base_offset; |
eb32d236 | 1832 | |
8d8a4ea5 SP |
1833 | /* use_pack() assured us we have [base_info, base_info + 20) |
1834 | * as a range that we can look at without walking off the | |
1835 | * end of the mapped window. Its actually the hash size | |
1836 | * that is assured. An OFS_DELTA longer than the hash size | |
1837 | * is stupid, as then a REF_DELTA would be smaller to store. | |
1838 | */ | |
21666f1a | 1839 | if (type == OBJ_OFS_DELTA) { |
eb32d236 NP |
1840 | unsigned used = 0; |
1841 | unsigned char c = base_info[used++]; | |
1842 | base_offset = c & 127; | |
1843 | while (c & 128) { | |
1844 | base_offset += 1; | |
8723f216 | 1845 | if (!base_offset || MSB(base_offset, 7)) |
8eca0b47 | 1846 | return 0; /* overflow */ |
eb32d236 NP |
1847 | c = base_info[used++]; |
1848 | base_offset = (base_offset << 7) + (c & 127); | |
1849 | } | |
1850 | base_offset = delta_obj_offset - base_offset; | |
d8f32556 | 1851 | if (base_offset <= 0 || base_offset >= delta_obj_offset) |
8eca0b47 | 1852 | return 0; /* out of bound */ |
2b87c45b | 1853 | *curpos += used; |
21666f1a | 1854 | } else if (type == OBJ_REF_DELTA) { |
eb32d236 NP |
1855 | /* The base entry _must_ be in the same pack */ |
1856 | base_offset = find_pack_entry_one(base_info, p); | |
2b87c45b | 1857 | *curpos += 20; |
eb32d236 NP |
1858 | } else |
1859 | die("I am totally screwed"); | |
2b87c45b | 1860 | return base_offset; |
eb32d236 NP |
1861 | } |
1862 | ||
5d642e75 JK |
1863 | /* |
1864 | * Like get_delta_base above, but we return the sha1 instead of the pack | |
1865 | * offset. This means it is cheaper for REF deltas (we do not have to do | |
1866 | * the final object lookup), but more expensive for OFS deltas (we | |
1867 | * have to load the revidx to convert the offset back into a sha1). | |
1868 | */ | |
1869 | static const unsigned char *get_delta_base_sha1(struct packed_git *p, | |
1870 | struct pack_window **w_curs, | |
1871 | off_t curpos, | |
1872 | enum object_type type, | |
1873 | off_t delta_obj_offset) | |
1874 | { | |
1875 | if (type == OBJ_REF_DELTA) { | |
1876 | unsigned char *base = use_pack(p, w_curs, curpos, NULL); | |
1877 | return base; | |
1878 | } else if (type == OBJ_OFS_DELTA) { | |
1879 | struct revindex_entry *revidx; | |
1880 | off_t base_offset = get_delta_base(p, w_curs, &curpos, | |
1881 | type, delta_obj_offset); | |
1882 | ||
1883 | if (!base_offset) | |
1884 | return NULL; | |
1885 | ||
1886 | revidx = find_pack_revindex(p, base_offset); | |
1887 | if (!revidx) | |
1888 | return NULL; | |
1889 | ||
1890 | return nth_packed_object_sha1(p, revidx->nr); | |
1891 | } else | |
1892 | return NULL; | |
1893 | } | |
1894 | ||
f8c8abc5 JH |
1895 | int unpack_object_header(struct packed_git *p, |
1896 | struct pack_window **w_curs, | |
1897 | off_t *curpos, | |
1898 | unsigned long *sizep) | |
a733cb60 | 1899 | { |
03e79c88 | 1900 | unsigned char *base; |
ef49a7a0 | 1901 | unsigned long left; |
ad1ed5ee | 1902 | unsigned long used; |
2b87c45b | 1903 | enum object_type type; |
a733cb60 | 1904 | |
8d8a4ea5 | 1905 | /* use_pack() assures us we have [base, base + 20) available |
0353a0c4 | 1906 | * as a range that we can look at. (Its actually the hash |
3dff5379 | 1907 | * size that is assured.) With our object header encoding |
8d8a4ea5 SP |
1908 | * the maximum deflated object size is 2^137, which is just |
1909 | * insane, so we know won't exceed what we have been given. | |
1910 | */ | |
2b87c45b | 1911 | base = use_pack(p, w_curs, *curpos, &left); |
09ded04b NP |
1912 | used = unpack_object_header_buffer(base, left, &type, sizep); |
1913 | if (!used) { | |
1914 | type = OBJ_BAD; | |
1915 | } else | |
1916 | *curpos += used; | |
ad1ed5ee | 1917 | |
2b87c45b | 1918 | return type; |
a733cb60 LT |
1919 | } |
1920 | ||
790d96c0 TR |
1921 | static int retry_bad_packed_offset(struct packed_git *p, off_t obj_offset) |
1922 | { | |
1923 | int type; | |
1924 | struct revindex_entry *revidx; | |
1925 | const unsigned char *sha1; | |
1926 | revidx = find_pack_revindex(p, obj_offset); | |
1927 | if (!revidx) | |
1928 | return OBJ_BAD; | |
1929 | sha1 = nth_packed_object_sha1(p, revidx->nr); | |
1930 | mark_bad_packed_object(p, sha1); | |
1931 | type = sha1_object_info(sha1, NULL); | |
1932 | if (type <= OBJ_NONE) | |
1933 | return OBJ_BAD; | |
1934 | return type; | |
1935 | } | |
1936 | ||
790d96c0 TR |
1937 | #define POI_STACK_PREALLOC 64 |
1938 | ||
90191d37 JK |
1939 | static enum object_type packed_to_object_type(struct packed_git *p, |
1940 | off_t obj_offset, | |
1941 | enum object_type type, | |
1942 | struct pack_window **w_curs, | |
1943 | off_t curpos) | |
1f688557 | 1944 | { |
790d96c0 TR |
1945 | off_t small_poi_stack[POI_STACK_PREALLOC]; |
1946 | off_t *poi_stack = small_poi_stack; | |
1947 | int poi_stack_nr = 0, poi_stack_alloc = POI_STACK_PREALLOC; | |
5db47c2b | 1948 | |
790d96c0 TR |
1949 | while (type == OBJ_OFS_DELTA || type == OBJ_REF_DELTA) { |
1950 | off_t base_offset; | |
90191d37 | 1951 | unsigned long size; |
790d96c0 TR |
1952 | /* Push the object we're going to leave behind */ |
1953 | if (poi_stack_nr >= poi_stack_alloc && poi_stack == small_poi_stack) { | |
1954 | poi_stack_alloc = alloc_nr(poi_stack_nr); | |
1955 | poi_stack = xmalloc(sizeof(off_t)*poi_stack_alloc); | |
1956 | memcpy(poi_stack, small_poi_stack, sizeof(off_t)*poi_stack_nr); | |
1957 | } else { | |
1958 | ALLOC_GROW(poi_stack, poi_stack_nr+1, poi_stack_alloc); | |
1959 | } | |
1960 | poi_stack[poi_stack_nr++] = obj_offset; | |
1961 | /* If parsing the base offset fails, just unwind */ | |
90191d37 | 1962 | base_offset = get_delta_base(p, w_curs, &curpos, type, obj_offset); |
790d96c0 TR |
1963 | if (!base_offset) |
1964 | goto unwind; | |
1965 | curpos = obj_offset = base_offset; | |
90191d37 | 1966 | type = unpack_object_header(p, w_curs, &curpos, &size); |
790d96c0 TR |
1967 | if (type <= OBJ_NONE) { |
1968 | /* If getting the base itself fails, we first | |
1969 | * retry the base, otherwise unwind */ | |
1970 | type = retry_bad_packed_offset(p, base_offset); | |
1971 | if (type > OBJ_NONE) | |
1972 | goto out; | |
1973 | goto unwind; | |
1974 | } | |
1975 | } | |
1976 | ||
21666f1a | 1977 | switch (type) { |
790d96c0 | 1978 | case OBJ_BAD: |
a733cb60 | 1979 | case OBJ_COMMIT: |
a733cb60 | 1980 | case OBJ_TREE: |
a733cb60 | 1981 | case OBJ_BLOB: |
a733cb60 | 1982 | case OBJ_TAG: |
a69d0943 | 1983 | break; |
1f688557 | 1984 | default: |
3d77d877 NP |
1985 | error("unknown object type %i at offset %"PRIuMAX" in %s", |
1986 | type, (uintmax_t)obj_offset, p->pack_name); | |
1987 | type = OBJ_BAD; | |
1f688557 | 1988 | } |
790d96c0 TR |
1989 | |
1990 | out: | |
1991 | if (poi_stack != small_poi_stack) | |
1992 | free(poi_stack); | |
21666f1a | 1993 | return type; |
790d96c0 TR |
1994 | |
1995 | unwind: | |
1996 | while (poi_stack_nr) { | |
1997 | obj_offset = poi_stack[--poi_stack_nr]; | |
1998 | type = retry_bad_packed_offset(p, obj_offset); | |
1999 | if (type > OBJ_NONE) | |
2000 | goto out; | |
2001 | } | |
2002 | type = OBJ_BAD; | |
2003 | goto out; | |
1f688557 JH |
2004 | } |
2005 | ||
90191d37 | 2006 | static int packed_object_info(struct packed_git *p, off_t obj_offset, |
23c339c0 | 2007 | struct object_info *oi) |
90191d37 JK |
2008 | { |
2009 | struct pack_window *w_curs = NULL; | |
2010 | unsigned long size; | |
2011 | off_t curpos = obj_offset; | |
2012 | enum object_type type; | |
2013 | ||
412916ee JK |
2014 | /* |
2015 | * We always get the representation type, but only convert it to | |
2016 | * a "real" type later if the caller is interested. | |
2017 | */ | |
90191d37 JK |
2018 | type = unpack_object_header(p, &w_curs, &curpos, &size); |
2019 | ||
23c339c0 | 2020 | if (oi->sizep) { |
90191d37 JK |
2021 | if (type == OBJ_OFS_DELTA || type == OBJ_REF_DELTA) { |
2022 | off_t tmp_pos = curpos; | |
2023 | off_t base_offset = get_delta_base(p, &w_curs, &tmp_pos, | |
2024 | type, obj_offset); | |
2025 | if (!base_offset) { | |
2026 | type = OBJ_BAD; | |
2027 | goto out; | |
2028 | } | |
23c339c0 JK |
2029 | *oi->sizep = get_size_from_delta(p, &w_curs, tmp_pos); |
2030 | if (*oi->sizep == 0) { | |
90191d37 JK |
2031 | type = OBJ_BAD; |
2032 | goto out; | |
2033 | } | |
2034 | } else { | |
23c339c0 | 2035 | *oi->sizep = size; |
90191d37 JK |
2036 | } |
2037 | } | |
2038 | ||
23c339c0 | 2039 | if (oi->disk_sizep) { |
90191d37 | 2040 | struct revindex_entry *revidx = find_pack_revindex(p, obj_offset); |
23c339c0 | 2041 | *oi->disk_sizep = revidx[1].offset - obj_offset; |
90191d37 JK |
2042 | } |
2043 | ||
23c339c0 JK |
2044 | if (oi->typep) { |
2045 | *oi->typep = packed_to_object_type(p, obj_offset, type, &w_curs, curpos); | |
2046 | if (*oi->typep < 0) { | |
412916ee JK |
2047 | type = OBJ_BAD; |
2048 | goto out; | |
2049 | } | |
2050 | } | |
90191d37 | 2051 | |
5d642e75 JK |
2052 | if (oi->delta_base_sha1) { |
2053 | if (type == OBJ_OFS_DELTA || type == OBJ_REF_DELTA) { | |
2054 | const unsigned char *base; | |
2055 | ||
2056 | base = get_delta_base_sha1(p, &w_curs, curpos, | |
2057 | type, obj_offset); | |
2058 | if (!base) { | |
2059 | type = OBJ_BAD; | |
2060 | goto out; | |
2061 | } | |
2062 | ||
2063 | hashcpy(oi->delta_base_sha1, base); | |
2064 | } else | |
2065 | hashclr(oi->delta_base_sha1); | |
2066 | } | |
2067 | ||
90191d37 JK |
2068 | out: |
2069 | unuse_pack(&w_curs); | |
2070 | return type; | |
2071 | } | |
2072 | ||
eb950c19 | 2073 | static void *unpack_compressed_entry(struct packed_git *p, |
03e79c88 | 2074 | struct pack_window **w_curs, |
c4001d92 | 2075 | off_t curpos, |
eb950c19 | 2076 | unsigned long size) |
de530aaa SP |
2077 | { |
2078 | int st; | |
ef49a7a0 | 2079 | git_zstream stream; |
079afb18 | 2080 | unsigned char *buffer, *in; |
de530aaa | 2081 | |
735efde8 NTND |
2082 | buffer = xmallocz_gently(size); |
2083 | if (!buffer) | |
2084 | return NULL; | |
de530aaa | 2085 | memset(&stream, 0, sizeof(stream)); |
de530aaa | 2086 | stream.next_out = buffer; |
39eea7bd | 2087 | stream.avail_out = size + 1; |
de530aaa | 2088 | |
39c68542 | 2089 | git_inflate_init(&stream); |
079afb18 | 2090 | do { |
2b87c45b | 2091 | in = use_pack(p, w_curs, curpos, &stream.avail_in); |
079afb18 | 2092 | stream.next_in = in; |
39c68542 | 2093 | st = git_inflate(&stream, Z_FINISH); |
39eea7bd JH |
2094 | if (!stream.avail_out) |
2095 | break; /* the payload is larger than it should be */ | |
2b87c45b | 2096 | curpos += stream.next_in - in; |
079afb18 | 2097 | } while (st == Z_OK || st == Z_BUF_ERROR); |
39c68542 | 2098 | git_inflate_end(&stream); |
de530aaa SP |
2099 | if ((st != Z_STREAM_END) || stream.total_out != size) { |
2100 | free(buffer); | |
2101 | return NULL; | |
2102 | } | |
2103 | ||
2104 | return buffer; | |
2105 | } | |
2106 | ||
e5e01619 LT |
2107 | #define MAX_DELTA_CACHE (256) |
2108 | ||
18bdec11 | 2109 | static size_t delta_base_cached; |
5e08ecbf NP |
2110 | |
2111 | static struct delta_base_cache_lru_list { | |
2112 | struct delta_base_cache_lru_list *prev; | |
2113 | struct delta_base_cache_lru_list *next; | |
2114 | } delta_base_cache_lru = { &delta_base_cache_lru, &delta_base_cache_lru }; | |
2115 | ||
e5e01619 | 2116 | static struct delta_base_cache_entry { |
5e08ecbf NP |
2117 | struct delta_base_cache_lru_list lru; |
2118 | void *data; | |
e5e01619 LT |
2119 | struct packed_git *p; |
2120 | off_t base_offset; | |
2121 | unsigned long size; | |
e5e01619 LT |
2122 | enum object_type type; |
2123 | } delta_base_cache[MAX_DELTA_CACHE]; | |
2124 | ||
2125 | static unsigned long pack_entry_hash(struct packed_git *p, off_t base_offset) | |
2126 | { | |
2127 | unsigned long hash; | |
2128 | ||
56a1a3ab | 2129 | hash = (unsigned long)(intptr_t)p + (unsigned long)base_offset; |
e5e01619 | 2130 | hash += (hash >> 8) + (hash >> 16); |
3358004a | 2131 | return hash % MAX_DELTA_CACHE; |
e5e01619 LT |
2132 | } |
2133 | ||
84dd81c1 TR |
2134 | static struct delta_base_cache_entry * |
2135 | get_delta_base_cache_entry(struct packed_git *p, off_t base_offset) | |
5266d369 JH |
2136 | { |
2137 | unsigned long hash = pack_entry_hash(p, base_offset); | |
84dd81c1 TR |
2138 | return delta_base_cache + hash; |
2139 | } | |
2140 | ||
2141 | static int eq_delta_base_cache_entry(struct delta_base_cache_entry *ent, | |
2142 | struct packed_git *p, off_t base_offset) | |
2143 | { | |
5266d369 JH |
2144 | return (ent->data && ent->p == p && ent->base_offset == base_offset); |
2145 | } | |
2146 | ||
84dd81c1 TR |
2147 | static int in_delta_base_cache(struct packed_git *p, off_t base_offset) |
2148 | { | |
2149 | struct delta_base_cache_entry *ent; | |
2150 | ent = get_delta_base_cache_entry(p, base_offset); | |
2151 | return eq_delta_base_cache_entry(ent, p, base_offset); | |
2152 | } | |
2153 | ||
2154 | static void clear_delta_base_cache_entry(struct delta_base_cache_entry *ent) | |
2155 | { | |
2156 | ent->data = NULL; | |
2157 | ent->lru.next->prev = ent->lru.prev; | |
2158 | ent->lru.prev->next = ent->lru.next; | |
2159 | delta_base_cached -= ent->size; | |
2160 | } | |
2161 | ||
62f255ad | 2162 | static void *cache_or_unpack_entry(struct packed_git *p, off_t base_offset, |
a0cba108 | 2163 | unsigned long *base_size, enum object_type *type, int keep_cache) |
62f255ad | 2164 | { |
84dd81c1 | 2165 | struct delta_base_cache_entry *ent; |
e5e01619 | 2166 | void *ret; |
e5e01619 | 2167 | |
84dd81c1 TR |
2168 | ent = get_delta_base_cache_entry(p, base_offset); |
2169 | ||
2170 | if (!eq_delta_base_cache_entry(ent, p, base_offset)) | |
749bc58c | 2171 | return unpack_entry(p, base_offset, type, base_size); |
e5e01619 | 2172 | |
84dd81c1 TR |
2173 | ret = ent->data; |
2174 | ||
2175 | if (!keep_cache) | |
2176 | clear_delta_base_cache_entry(ent); | |
2177 | else | |
182af834 | 2178 | ret = xmemdupz(ent->data, ent->size); |
e5e01619 LT |
2179 | *type = ent->type; |
2180 | *base_size = ent->size; | |
2181 | return ret; | |
62f255ad LT |
2182 | } |
2183 | ||
18bdec11 SP |
2184 | static inline void release_delta_base_cache(struct delta_base_cache_entry *ent) |
2185 | { | |
2186 | if (ent->data) { | |
2187 | free(ent->data); | |
2188 | ent->data = NULL; | |
5e08ecbf NP |
2189 | ent->lru.next->prev = ent->lru.prev; |
2190 | ent->lru.prev->next = ent->lru.next; | |
18bdec11 SP |
2191 | delta_base_cached -= ent->size; |
2192 | } | |
2193 | } | |
2194 | ||
3d20c636 SP |
2195 | void clear_delta_base_cache(void) |
2196 | { | |
2197 | unsigned long p; | |
2198 | for (p = 0; p < MAX_DELTA_CACHE; p++) | |
2199 | release_delta_base_cache(&delta_base_cache[p]); | |
2200 | } | |
2201 | ||
62f255ad LT |
2202 | static void add_delta_base_cache(struct packed_git *p, off_t base_offset, |
2203 | void *base, unsigned long base_size, enum object_type type) | |
2204 | { | |
5e08ecbf | 2205 | unsigned long hash = pack_entry_hash(p, base_offset); |
e5e01619 | 2206 | struct delta_base_cache_entry *ent = delta_base_cache + hash; |
5e08ecbf | 2207 | struct delta_base_cache_lru_list *lru; |
e5e01619 | 2208 | |
18bdec11 SP |
2209 | release_delta_base_cache(ent); |
2210 | delta_base_cached += base_size; | |
5e08ecbf NP |
2211 | |
2212 | for (lru = delta_base_cache_lru.next; | |
2213 | delta_base_cached > delta_base_cache_limit | |
2214 | && lru != &delta_base_cache_lru; | |
2215 | lru = lru->next) { | |
2216 | struct delta_base_cache_entry *f = (void *)lru; | |
18bdec11 SP |
2217 | if (f->type == OBJ_BLOB) |
2218 | release_delta_base_cache(f); | |
2219 | } | |
5e08ecbf NP |
2220 | for (lru = delta_base_cache_lru.next; |
2221 | delta_base_cached > delta_base_cache_limit | |
2222 | && lru != &delta_base_cache_lru; | |
2223 | lru = lru->next) { | |
2224 | struct delta_base_cache_entry *f = (void *)lru; | |
2225 | release_delta_base_cache(f); | |
2226 | } | |
18bdec11 | 2227 | |
e5e01619 LT |
2228 | ent->p = p; |
2229 | ent->base_offset = base_offset; | |
2230 | ent->type = type; | |
2231 | ent->data = base; | |
2232 | ent->size = base_size; | |
5e08ecbf NP |
2233 | ent->lru.next = &delta_base_cache_lru; |
2234 | ent->lru.prev = delta_base_cache_lru.prev; | |
2235 | delta_base_cache_lru.prev->next = &ent->lru; | |
2236 | delta_base_cache_lru.prev = &ent->lru; | |
62f255ad LT |
2237 | } |
2238 | ||
c2c5b270 CC |
2239 | static void *read_object(const unsigned char *sha1, enum object_type *type, |
2240 | unsigned long *size); | |
2241 | ||
5f44324d JH |
2242 | static void write_pack_access_log(struct packed_git *p, off_t obj_offset) |
2243 | { | |
67dc598e KB |
2244 | static struct trace_key pack_access = TRACE_KEY_INIT(PACK_ACCESS); |
2245 | trace_printf_key(&pack_access, "%s %"PRIuMAX"\n", | |
2246 | p->pack_name, (uintmax_t)obj_offset); | |
5f44324d JH |
2247 | } |
2248 | ||
0e8189e2 NP |
2249 | int do_check_packed_object_crc; |
2250 | ||
abe601bb TR |
2251 | #define UNPACK_ENTRY_STACK_PREALLOC 64 |
2252 | struct unpack_entry_stack_ent { | |
2253 | off_t obj_offset; | |
2254 | off_t curpos; | |
2255 | unsigned long size; | |
2256 | }; | |
2257 | ||
c4001d92 | 2258 | void *unpack_entry(struct packed_git *p, off_t obj_offset, |
abe601bb | 2259 | enum object_type *final_type, unsigned long *final_size) |
1f688557 | 2260 | { |
03e79c88 | 2261 | struct pack_window *w_curs = NULL; |
c4001d92 | 2262 | off_t curpos = obj_offset; |
abe601bb TR |
2263 | void *data = NULL; |
2264 | unsigned long size; | |
2265 | enum object_type type; | |
2266 | struct unpack_entry_stack_ent small_delta_stack[UNPACK_ENTRY_STACK_PREALLOC]; | |
2267 | struct unpack_entry_stack_ent *delta_stack = small_delta_stack; | |
2268 | int delta_stack_nr = 0, delta_stack_alloc = UNPACK_ENTRY_STACK_PREALLOC; | |
2269 | int base_from_cache = 0; | |
1f688557 | 2270 | |
67dc598e | 2271 | write_pack_access_log(p, obj_offset); |
5f44324d | 2272 | |
abe601bb TR |
2273 | /* PHASE 1: drill down to the innermost base object */ |
2274 | for (;;) { | |
2275 | off_t base_offset; | |
2276 | int i; | |
2277 | struct delta_base_cache_entry *ent; | |
2278 | ||
77965f8b NTND |
2279 | ent = get_delta_base_cache_entry(p, curpos); |
2280 | if (eq_delta_base_cache_entry(ent, p, curpos)) { | |
2281 | type = ent->type; | |
2282 | data = ent->data; | |
2283 | size = ent->size; | |
2284 | clear_delta_base_cache_entry(ent); | |
2285 | base_from_cache = 1; | |
2286 | break; | |
2287 | } | |
2288 | ||
abe601bb TR |
2289 | if (do_check_packed_object_crc && p->index_version > 1) { |
2290 | struct revindex_entry *revidx = find_pack_revindex(p, obj_offset); | |
2291 | unsigned long len = revidx[1].offset - obj_offset; | |
2292 | if (check_pack_crc(p, &w_curs, obj_offset, len, revidx->nr)) { | |
2293 | const unsigned char *sha1 = | |
2294 | nth_packed_object_sha1(p, revidx->nr); | |
2295 | error("bad packed object CRC for %s", | |
2296 | sha1_to_hex(sha1)); | |
2297 | mark_bad_packed_object(p, sha1); | |
2298 | unuse_pack(&w_curs); | |
2299 | return NULL; | |
2300 | } | |
2301 | } | |
2302 | ||
abe601bb TR |
2303 | type = unpack_object_header(p, &w_curs, &curpos, &size); |
2304 | if (type != OBJ_OFS_DELTA && type != OBJ_REF_DELTA) | |
2305 | break; | |
2306 | ||
2307 | base_offset = get_delta_base(p, &w_curs, &curpos, type, obj_offset); | |
2308 | if (!base_offset) { | |
2309 | error("failed to validate delta base reference " | |
2310 | "at offset %"PRIuMAX" from %s", | |
2311 | (uintmax_t)curpos, p->pack_name); | |
2312 | /* bail to phase 2, in hopes of recovery */ | |
2313 | data = NULL; | |
2314 | break; | |
2315 | } | |
2316 | ||
2317 | /* push object, proceed to base */ | |
2318 | if (delta_stack_nr >= delta_stack_alloc | |
2319 | && delta_stack == small_delta_stack) { | |
2320 | delta_stack_alloc = alloc_nr(delta_stack_nr); | |
2321 | delta_stack = xmalloc(sizeof(*delta_stack)*delta_stack_alloc); | |
2322 | memcpy(delta_stack, small_delta_stack, | |
2323 | sizeof(*delta_stack)*delta_stack_nr); | |
2324 | } else { | |
2325 | ALLOC_GROW(delta_stack, delta_stack_nr+1, delta_stack_alloc); | |
2326 | } | |
2327 | i = delta_stack_nr++; | |
2328 | delta_stack[i].obj_offset = obj_offset; | |
2329 | delta_stack[i].curpos = curpos; | |
2330 | delta_stack[i].size = size; | |
2331 | ||
2332 | curpos = obj_offset = base_offset; | |
0e8189e2 NP |
2333 | } |
2334 | ||
abe601bb TR |
2335 | /* PHASE 2: handle the base */ |
2336 | switch (type) { | |
eb32d236 NP |
2337 | case OBJ_OFS_DELTA: |
2338 | case OBJ_REF_DELTA: | |
abe601bb TR |
2339 | if (data) |
2340 | die("BUG in unpack_entry: left loop at a valid delta"); | |
4d703a1a | 2341 | break; |
a733cb60 | 2342 | case OBJ_COMMIT: |
a733cb60 | 2343 | case OBJ_TREE: |
a733cb60 | 2344 | case OBJ_BLOB: |
a733cb60 | 2345 | case OBJ_TAG: |
abe601bb TR |
2346 | if (!base_from_cache) |
2347 | data = unpack_compressed_entry(p, &w_curs, curpos, size); | |
4d703a1a | 2348 | break; |
1f688557 | 2349 | default: |
8eca0b47 NP |
2350 | data = NULL; |
2351 | error("unknown object type %i at offset %"PRIuMAX" in %s", | |
abe601bb TR |
2352 | type, (uintmax_t)obj_offset, p->pack_name); |
2353 | } | |
2354 | ||
2355 | /* PHASE 3: apply deltas in order */ | |
2356 | ||
2357 | /* invariants: | |
2358 | * 'data' holds the base data, or NULL if there was corruption | |
2359 | */ | |
2360 | while (delta_stack_nr) { | |
2361 | void *delta_data; | |
2362 | void *base = data; | |
2363 | unsigned long delta_size, base_size = size; | |
2364 | int i; | |
2365 | ||
2366 | data = NULL; | |
2367 | ||
2368 | if (base) | |
2369 | add_delta_base_cache(p, obj_offset, base, base_size, type); | |
2370 | ||
2371 | if (!base) { | |
2372 | /* | |
2373 | * We're probably in deep shit, but let's try to fetch | |
2374 | * the required base anyway from another pack or loose. | |
2375 | * This is costly but should happen only in the presence | |
2376 | * of a corrupted pack, and is better than failing outright. | |
2377 | */ | |
2378 | struct revindex_entry *revidx; | |
2379 | const unsigned char *base_sha1; | |
2380 | revidx = find_pack_revindex(p, obj_offset); | |
2381 | if (revidx) { | |
2382 | base_sha1 = nth_packed_object_sha1(p, revidx->nr); | |
2383 | error("failed to read delta base object %s" | |
2384 | " at offset %"PRIuMAX" from %s", | |
2385 | sha1_to_hex(base_sha1), (uintmax_t)obj_offset, | |
2386 | p->pack_name); | |
2387 | mark_bad_packed_object(p, base_sha1); | |
2388 | base = read_object(base_sha1, &type, &base_size); | |
2389 | } | |
2390 | } | |
2391 | ||
2392 | i = --delta_stack_nr; | |
2393 | obj_offset = delta_stack[i].obj_offset; | |
2394 | curpos = delta_stack[i].curpos; | |
2395 | delta_size = delta_stack[i].size; | |
2396 | ||
2397 | if (!base) | |
2398 | continue; | |
2399 | ||
2400 | delta_data = unpack_compressed_entry(p, &w_curs, curpos, delta_size); | |
2401 | ||
2402 | if (!delta_data) { | |
2403 | error("failed to unpack compressed delta " | |
2404 | "at offset %"PRIuMAX" from %s", | |
2405 | (uintmax_t)curpos, p->pack_name); | |
abe601bb TR |
2406 | data = NULL; |
2407 | continue; | |
2408 | } | |
2409 | ||
2410 | data = patch_delta(base, base_size, | |
2411 | delta_data, delta_size, | |
2412 | &size); | |
1ee886c1 JK |
2413 | |
2414 | /* | |
2415 | * We could not apply the delta; warn the user, but keep going. | |
2416 | * Our failure will be noticed either in the next iteration of | |
2417 | * the loop, or if this is the final delta, in the caller when | |
2418 | * we return NULL. Those code paths will take care of making | |
2419 | * a more explicit warning and retrying with another copy of | |
2420 | * the object. | |
2421 | */ | |
abe601bb | 2422 | if (!data) |
1ee886c1 | 2423 | error("failed to apply delta"); |
abe601bb | 2424 | |
4b8f772c | 2425 | free(delta_data); |
1f688557 | 2426 | } |
abe601bb TR |
2427 | |
2428 | *final_type = type; | |
2429 | *final_size = size; | |
2430 | ||
03e79c88 | 2431 | unuse_pack(&w_curs); |
019d1e65 NTND |
2432 | |
2433 | if (delta_stack != small_delta_stack) | |
2434 | free(delta_stack); | |
2435 | ||
21666f1a | 2436 | return data; |
1f688557 JH |
2437 | } |
2438 | ||
d079837e | 2439 | const unsigned char *nth_packed_object_sha1(struct packed_git *p, |
d72308e0 | 2440 | uint32_t n) |
9a217f2a | 2441 | { |
42873078 | 2442 | const unsigned char *index = p->index_data; |
d079837e SP |
2443 | if (!index) { |
2444 | if (open_pack_index(p)) | |
2445 | return NULL; | |
2446 | index = p->index_data; | |
2447 | } | |
57059091 | 2448 | if (n >= p->num_objects) |
d72308e0 | 2449 | return NULL; |
74e34e1f NP |
2450 | index += 4 * 256; |
2451 | if (p->index_version == 1) { | |
2452 | return index + 24 * n + 4; | |
2453 | } else { | |
2454 | index += 8; | |
2455 | return index + 20 * n; | |
2456 | } | |
2457 | } | |
2458 | ||
99093238 | 2459 | off_t nth_packed_object_offset(const struct packed_git *p, uint32_t n) |
74e34e1f NP |
2460 | { |
2461 | const unsigned char *index = p->index_data; | |
2462 | index += 4 * 256; | |
2463 | if (p->index_version == 1) { | |
2464 | return ntohl(*((uint32_t *)(index + 24 * n))); | |
2465 | } else { | |
2466 | uint32_t off; | |
2467 | index += 8 + p->num_objects * (20 + 4); | |
2468 | off = ntohl(*((uint32_t *)(index + 4 * n))); | |
2469 | if (!(off & 0x80000000)) | |
2470 | return off; | |
2471 | index += p->num_objects * 4 + (off & 0x7fffffff) * 8; | |
2472 | return (((uint64_t)ntohl(*((uint32_t *)(index + 0)))) << 32) | | |
2473 | ntohl(*((uint32_t *)(index + 4))); | |
2474 | } | |
9a217f2a JH |
2475 | } |
2476 | ||
c4001d92 | 2477 | off_t find_pack_entry_one(const unsigned char *sha1, |
43057304 | 2478 | struct packed_git *p) |
1f688557 | 2479 | { |
42873078 | 2480 | const uint32_t *level1_ofs = p->index_data; |
42873078 | 2481 | const unsigned char *index = p->index_data; |
628522ec JH |
2482 | unsigned hi, lo, stride; |
2483 | static int use_lookup = -1; | |
2484 | static int debug_lookup = -1; | |
2485 | ||
2486 | if (debug_lookup < 0) | |
2487 | debug_lookup = !!getenv("GIT_DEBUG_LOOKUP"); | |
42873078 | 2488 | |
d079837e SP |
2489 | if (!index) { |
2490 | if (open_pack_index(p)) | |
2491 | return 0; | |
2492 | level1_ofs = p->index_data; | |
2493 | index = p->index_data; | |
2494 | } | |
74e34e1f NP |
2495 | if (p->index_version > 1) { |
2496 | level1_ofs += 2; | |
2497 | index += 8; | |
2498 | } | |
42873078 | 2499 | index += 4 * 256; |
74e34e1f NP |
2500 | hi = ntohl(level1_ofs[*sha1]); |
2501 | lo = ((*sha1 == 0x0) ? 0 : ntohl(level1_ofs[*sha1 - 1])); | |
628522ec JH |
2502 | if (p->index_version > 1) { |
2503 | stride = 20; | |
2504 | } else { | |
2505 | stride = 24; | |
2506 | index += 4; | |
2507 | } | |
2508 | ||
2509 | if (debug_lookup) | |
6e1c2344 | 2510 | printf("%02x%02x%02x... lo %u hi %u nr %"PRIu32"\n", |
628522ec JH |
2511 | sha1[0], sha1[1], sha1[2], lo, hi, p->num_objects); |
2512 | ||
2513 | if (use_lookup < 0) | |
2514 | use_lookup = !!getenv("GIT_USE_LOOKUP"); | |
2515 | if (use_lookup) { | |
2516 | int pos = sha1_entry_pos(index, stride, 0, | |
2517 | lo, hi, p->num_objects, sha1); | |
2518 | if (pos < 0) | |
2519 | return 0; | |
2520 | return nth_packed_object_offset(p, pos); | |
2521 | } | |
1f688557 JH |
2522 | |
2523 | do { | |
74e34e1f | 2524 | unsigned mi = (lo + hi) / 2; |
628522ec JH |
2525 | int cmp = hashcmp(index + mi * stride, sha1); |
2526 | ||
2527 | if (debug_lookup) | |
2528 | printf("lo %u hi %u rg %u mi %u\n", | |
2529 | lo, hi, hi - lo, mi); | |
43057304 | 2530 | if (!cmp) |
74e34e1f | 2531 | return nth_packed_object_offset(p, mi); |
1f688557 JH |
2532 | if (cmp > 0) |
2533 | hi = mi; | |
2534 | else | |
2535 | lo = mi+1; | |
2536 | } while (lo < hi); | |
2537 | return 0; | |
2538 | } | |
2539 | ||
4c080182 | 2540 | int is_pack_valid(struct packed_git *p) |
d131b7af SP |
2541 | { |
2542 | /* An already open pack is known to be valid. */ | |
2543 | if (p->pack_fd != -1) | |
2544 | return 1; | |
2545 | ||
2546 | /* If the pack has one window completely covering the | |
2547 | * file size, the pack is known to be valid even if | |
2548 | * the descriptor is not currently open. | |
2549 | */ | |
2550 | if (p->windows) { | |
2551 | struct pack_window *w = p->windows; | |
2552 | ||
2553 | if (!w->offset && w->len == p->pack_size) | |
2554 | return 1; | |
2555 | } | |
2556 | ||
2557 | /* Force the pack to open to prove its valid. */ | |
2558 | return !open_packed_git(p); | |
2559 | } | |
2560 | ||
95099731 NTND |
2561 | static int fill_pack_entry(const unsigned char *sha1, |
2562 | struct pack_entry *e, | |
2563 | struct packed_git *p) | |
2564 | { | |
2565 | off_t offset; | |
2566 | ||
2567 | if (p->num_bad_objects) { | |
2568 | unsigned i; | |
2569 | for (i = 0; i < p->num_bad_objects; i++) | |
2570 | if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i)) | |
2571 | return 0; | |
2572 | } | |
2573 | ||
2574 | offset = find_pack_entry_one(sha1, p); | |
2575 | if (!offset) | |
2576 | return 0; | |
2577 | ||
2578 | /* | |
2579 | * We are about to tell the caller where they can locate the | |
2580 | * requested object. We better make sure the packfile is | |
2581 | * still here and can be accessed before supplying that | |
2582 | * answer, as it may have been deleted since the index was | |
2583 | * loaded! | |
2584 | */ | |
319b678a | 2585 | if (!is_pack_valid(p)) |
95099731 | 2586 | return 0; |
95099731 NTND |
2587 | e->offset = offset; |
2588 | e->p = p; | |
2589 | hashcpy(e->sha1, sha1); | |
2590 | return 1; | |
2591 | } | |
2592 | ||
d40d535b MH |
2593 | /* |
2594 | * Iff a pack file contains the object named by sha1, return true and | |
2595 | * store its location to e. | |
2596 | */ | |
4d6acb70 | 2597 | static int find_pack_entry(const unsigned char *sha1, struct pack_entry *e) |
1f688557 JH |
2598 | { |
2599 | struct packed_git *p; | |
43057304 | 2600 | |
1f688557 | 2601 | prepare_packed_git(); |
f7c22cc6 NP |
2602 | if (!packed_git) |
2603 | return 0; | |
1f688557 | 2604 | |
c01f51cc NTND |
2605 | if (last_found_pack && fill_pack_entry(sha1, e, last_found_pack)) |
2606 | return 1; | |
8eca0b47 | 2607 | |
c01f51cc | 2608 | for (p = packed_git; p; p = p->next) { |
1b1005d1 MH |
2609 | if (p == last_found_pack) |
2610 | continue; /* we already checked this one */ | |
f7c22cc6 | 2611 | |
1b1005d1 MH |
2612 | if (fill_pack_entry(sha1, e, p)) { |
2613 | last_found_pack = p; | |
2614 | return 1; | |
2615 | } | |
c01f51cc | 2616 | } |
1f688557 JH |
2617 | return 0; |
2618 | } | |
2619 | ||
a6080a0a | 2620 | struct packed_git *find_sha1_pack(const unsigned char *sha1, |
bf592c50 DB |
2621 | struct packed_git *packs) |
2622 | { | |
2623 | struct packed_git *p; | |
bf592c50 DB |
2624 | |
2625 | for (p = packs; p; p = p->next) { | |
43057304 | 2626 | if (find_pack_entry_one(sha1, p)) |
bf592c50 DB |
2627 | return p; |
2628 | } | |
2629 | return NULL; | |
21666f1a | 2630 | |
bf592c50 DB |
2631 | } |
2632 | ||
052fe5ea | 2633 | static int sha1_loose_object_info(const unsigned char *sha1, |
46f03448 KN |
2634 | struct object_info *oi, |
2635 | int flags) | |
65c2e0c3 | 2636 | { |
46f03448 KN |
2637 | int status = 0; |
2638 | unsigned long mapsize; | |
65c2e0c3 | 2639 | void *map; |
ef49a7a0 | 2640 | git_zstream stream; |
d65a16f6 | 2641 | char hdr[32]; |
46f03448 | 2642 | struct strbuf hdrbuf = STRBUF_INIT; |
65c2e0c3 | 2643 | |
5d642e75 JK |
2644 | if (oi->delta_base_sha1) |
2645 | hashclr(oi->delta_base_sha1); | |
2646 | ||
052fe5ea JK |
2647 | /* |
2648 | * If we don't care about type or size, then we don't | |
4ef8d1dd JH |
2649 | * need to look inside the object at all. Note that we |
2650 | * do not optimize out the stat call, even if the | |
2651 | * caller doesn't care about the disk-size, since our | |
2652 | * return value implicitly indicates whether the | |
2653 | * object even exists. | |
052fe5ea | 2654 | */ |
46f03448 | 2655 | if (!oi->typep && !oi->typename && !oi->sizep) { |
4ef8d1dd JH |
2656 | struct stat st; |
2657 | if (stat_sha1_file(sha1, &st) < 0) | |
2658 | return -1; | |
2659 | if (oi->disk_sizep) | |
23c339c0 | 2660 | *oi->disk_sizep = st.st_size; |
052fe5ea JK |
2661 | return 0; |
2662 | } | |
2663 | ||
bb6b8e4f | 2664 | map = map_sha1_file(sha1, &mapsize); |
f0df4ed5 | 2665 | if (!map) |
dbea72a8 | 2666 | return -1; |
23c339c0 JK |
2667 | if (oi->disk_sizep) |
2668 | *oi->disk_sizep = mapsize; | |
46f03448 KN |
2669 | if ((flags & LOOKUP_UNKNOWN_OBJECT)) { |
2670 | if (unpack_sha1_header_to_strbuf(&stream, map, mapsize, hdr, sizeof(hdr), &hdrbuf) < 0) | |
2671 | status = error("unable to unpack %s header with --allow-unknown-type", | |
2672 | sha1_to_hex(sha1)); | |
2673 | } else if (unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)) < 0) | |
36e4d74a JH |
2674 | status = error("unable to unpack %s header", |
2675 | sha1_to_hex(sha1)); | |
46f03448 KN |
2676 | if (status < 0) |
2677 | ; /* Do nothing */ | |
2678 | else if (hdrbuf.len) { | |
2679 | if ((status = parse_sha1_header_extended(hdrbuf.buf, oi, flags)) < 0) | |
2680 | status = error("unable to parse %s header with --allow-unknown-type", | |
2681 | sha1_to_hex(sha1)); | |
2682 | } else if ((status = parse_sha1_header_extended(hdr, oi, flags)) < 0) | |
36e4d74a | 2683 | status = error("unable to parse %s header", sha1_to_hex(sha1)); |
39c68542 | 2684 | git_inflate_end(&stream); |
65c2e0c3 | 2685 | munmap(map, mapsize); |
46f03448 | 2686 | if (status && oi->typep) |
23c339c0 | 2687 | *oi->typep = status; |
46f03448 | 2688 | strbuf_release(&hdrbuf); |
052fe5ea | 2689 | return 0; |
65c2e0c3 JH |
2690 | } |
2691 | ||
de7b5d62 | 2692 | int sha1_object_info_extended(const unsigned char *sha1, struct object_info *oi, unsigned flags) |
f0df4ed5 | 2693 | { |
c4d9986f | 2694 | struct cached_object *co; |
f0df4ed5 | 2695 | struct pack_entry e; |
5b086407 | 2696 | int rtype; |
46f03448 | 2697 | enum object_type real_type; |
1f7117ef | 2698 | const unsigned char *real = lookup_replace_object_extended(sha1, flags); |
f0df4ed5 | 2699 | |
1f7117ef | 2700 | co = find_cached_object(real); |
c4d9986f | 2701 | if (co) { |
5b086407 JK |
2702 | if (oi->typep) |
2703 | *(oi->typep) = co->type; | |
9a490590 JH |
2704 | if (oi->sizep) |
2705 | *(oi->sizep) = co->size; | |
161f00e7 JK |
2706 | if (oi->disk_sizep) |
2707 | *(oi->disk_sizep) = 0; | |
5d642e75 JK |
2708 | if (oi->delta_base_sha1) |
2709 | hashclr(oi->delta_base_sha1); | |
46f03448 KN |
2710 | if (oi->typename) |
2711 | strbuf_addstr(oi->typename, typename(co->type)); | |
9a490590 | 2712 | oi->whence = OI_CACHED; |
5b086407 | 2713 | return 0; |
c4d9986f NTND |
2714 | } |
2715 | ||
1f7117ef | 2716 | if (!find_pack_entry(real, &e)) { |
ddd63e64 | 2717 | /* Most likely it's a loose object. */ |
46f03448 | 2718 | if (!sha1_loose_object_info(real, oi, flags)) { |
9a490590 | 2719 | oi->whence = OI_LOOSE; |
5b086407 | 2720 | return 0; |
9a490590 | 2721 | } |
ddd63e64 SG |
2722 | |
2723 | /* Not a loose object; someone else may have just packed it. */ | |
f0df4ed5 | 2724 | reprepare_packed_git(); |
1f7117ef | 2725 | if (!find_pack_entry(real, &e)) |
052fe5ea | 2726 | return -1; |
f0df4ed5 | 2727 | } |
3d77d877 | 2728 | |
46f03448 KN |
2729 | /* |
2730 | * packed_object_info() does not follow the delta chain to | |
2731 | * find out the real type, unless it is given oi->typep. | |
2732 | */ | |
2733 | if (oi->typename && !oi->typep) | |
2734 | oi->typep = &real_type; | |
2735 | ||
23c339c0 | 2736 | rtype = packed_object_info(e.p, e.offset, oi); |
412916ee | 2737 | if (rtype < 0) { |
1f7117ef | 2738 | mark_bad_packed_object(e.p, real); |
46f03448 KN |
2739 | if (oi->typep == &real_type) |
2740 | oi->typep = NULL; | |
1f7117ef | 2741 | return sha1_object_info_extended(real, oi, 0); |
5266d369 JH |
2742 | } else if (in_delta_base_cache(e.p, e.offset)) { |
2743 | oi->whence = OI_DBCACHED; | |
9a490590 JH |
2744 | } else { |
2745 | oi->whence = OI_PACKED; | |
2746 | oi->u.packed.offset = e.offset; | |
2747 | oi->u.packed.pack = e.p; | |
2748 | oi->u.packed.is_delta = (rtype == OBJ_REF_DELTA || | |
2749 | rtype == OBJ_OFS_DELTA); | |
3d77d877 | 2750 | } |
46f03448 KN |
2751 | if (oi->typename) |
2752 | strbuf_addstr(oi->typename, typename(*oi->typep)); | |
2753 | if (oi->typep == &real_type) | |
2754 | oi->typep = NULL; | |
3d77d877 | 2755 | |
5b086407 | 2756 | return 0; |
f0df4ed5 JS |
2757 | } |
2758 | ||
3fc0dca9 | 2759 | /* returns enum object_type or negative */ |
9a490590 JH |
2760 | int sha1_object_info(const unsigned char *sha1, unsigned long *sizep) |
2761 | { | |
5b086407 | 2762 | enum object_type type; |
d099b717 | 2763 | struct object_info oi = {NULL}; |
9a490590 | 2764 | |
5b086407 | 2765 | oi.typep = &type; |
9a490590 | 2766 | oi.sizep = sizep; |
de7b5d62 | 2767 | if (sha1_object_info_extended(sha1, &oi, LOOKUP_REPLACE_OBJECT) < 0) |
5b086407 JK |
2768 | return -1; |
2769 | return type; | |
9a490590 JH |
2770 | } |
2771 | ||
21666f1a NP |
2772 | static void *read_packed_sha1(const unsigned char *sha1, |
2773 | enum object_type *type, unsigned long *size) | |
1f688557 JH |
2774 | { |
2775 | struct pack_entry e; | |
8eca0b47 | 2776 | void *data; |
1f688557 | 2777 | |
cd673c1f | 2778 | if (!find_pack_entry(sha1, &e)) |
1f688557 | 2779 | return NULL; |
8eca0b47 NP |
2780 | data = cache_or_unpack_entry(e.p, e.offset, size, type, 1); |
2781 | if (!data) { | |
2782 | /* | |
2783 | * We're probably in deep shit, but let's try to fetch | |
2784 | * the required object anyway from another pack or loose. | |
2785 | * This should happen only in the presence of a corrupted | |
2786 | * pack, and is better than failing outright. | |
2787 | */ | |
2788 | error("failed to read object %s at offset %"PRIuMAX" from %s", | |
2789 | sha1_to_hex(sha1), (uintmax_t)e.offset, e.p->pack_name); | |
2790 | mark_bad_packed_object(e.p, sha1); | |
ac939109 | 2791 | data = read_object(sha1, type, size); |
8eca0b47 NP |
2792 | } |
2793 | return data; | |
1f688557 JH |
2794 | } |
2795 | ||
21666f1a NP |
2796 | int pretend_sha1_file(void *buf, unsigned long len, enum object_type type, |
2797 | unsigned char *sha1) | |
d66b37bb JH |
2798 | { |
2799 | struct cached_object *co; | |
2800 | ||
21666f1a | 2801 | hash_sha1_file(buf, len, typename(type), sha1); |
d66b37bb JH |
2802 | if (has_sha1_file(sha1) || find_cached_object(sha1)) |
2803 | return 0; | |
c7353967 | 2804 | ALLOC_GROW(cached_objects, cached_object_nr + 1, cached_object_alloc); |
d66b37bb JH |
2805 | co = &cached_objects[cached_object_nr++]; |
2806 | co->size = len; | |
21666f1a | 2807 | co->type = type; |
efa13f7b JH |
2808 | co->buf = xmalloc(len); |
2809 | memcpy(co->buf, buf, len); | |
d66b37bb JH |
2810 | hashcpy(co->sha1, sha1); |
2811 | return 0; | |
2812 | } | |
2813 | ||
c2c5b270 CC |
2814 | static void *read_object(const unsigned char *sha1, enum object_type *type, |
2815 | unsigned long *size) | |
0fcfd160 LT |
2816 | { |
2817 | unsigned long mapsize; | |
2818 | void *map, *buf; | |
d66b37bb JH |
2819 | struct cached_object *co; |
2820 | ||
2821 | co = find_cached_object(sha1); | |
2822 | if (co) { | |
21666f1a | 2823 | *type = co->type; |
d66b37bb | 2824 | *size = co->size; |
182af834 | 2825 | return xmemdupz(co->buf, co->size); |
d66b37bb | 2826 | } |
0fcfd160 | 2827 | |
8276c007 PE |
2828 | buf = read_packed_sha1(sha1, type, size); |
2829 | if (buf) | |
2830 | return buf; | |
bb6b8e4f | 2831 | map = map_sha1_file(sha1, &mapsize); |
0fcfd160 | 2832 | if (map) { |
7efbff75 | 2833 | buf = unpack_sha1_file(map, mapsize, type, size, sha1); |
0fcfd160 LT |
2834 | munmap(map, mapsize); |
2835 | return buf; | |
2836 | } | |
637cdd9d | 2837 | reprepare_packed_git(); |
8276c007 | 2838 | return read_packed_sha1(sha1, type, size); |
0fcfd160 LT |
2839 | } |
2840 | ||
b6c4cecc JH |
2841 | /* |
2842 | * This function dies on corrupt objects; the callers who want to | |
2843 | * deal with them should arrange to call read_object() and give error | |
2844 | * messages themselves. | |
2845 | */ | |
5bf29b95 JH |
2846 | void *read_sha1_file_extended(const unsigned char *sha1, |
2847 | enum object_type *type, | |
2848 | unsigned long *size, | |
2849 | unsigned flag) | |
ac939109 | 2850 | { |
3ba7a065 | 2851 | void *data; |
b6c4cecc | 2852 | const struct packed_git *p; |
bf93eea0 | 2853 | const unsigned char *repl = lookup_replace_object_extended(sha1, flag); |
b6c4cecc | 2854 | |
3ba7a065 JH |
2855 | errno = 0; |
2856 | data = read_object(repl, type, size); | |
4bbf5a26 | 2857 | if (data) |
b6c4cecc | 2858 | return data; |
68095570 | 2859 | |
25f3af3f | 2860 | if (errno && errno != ENOENT) |
3ba7a065 JH |
2861 | die_errno("failed to read object %s", sha1_to_hex(sha1)); |
2862 | ||
68095570 | 2863 | /* die if we replaced an object with one that does not exist */ |
b6c4cecc | 2864 | if (repl != sha1) |
68095570 CC |
2865 | die("replacement %s not found for %s", |
2866 | sha1_to_hex(repl), sha1_to_hex(sha1)); | |
2867 | ||
b6c4cecc | 2868 | if (has_loose_object(repl)) { |
30d6c6ea MH |
2869 | const char *path = sha1_file_name(sha1); |
2870 | ||
b6c4cecc JH |
2871 | die("loose object %s (stored in %s) is corrupt", |
2872 | sha1_to_hex(repl), path); | |
e8b15e61 | 2873 | } |
68095570 | 2874 | |
b6c4cecc JH |
2875 | if ((p = has_packed_and_bad(repl)) != NULL) |
2876 | die("packed object %s (stored in %s) is corrupt", | |
2877 | sha1_to_hex(repl), p->pack_name); | |
f5552aee | 2878 | |
b6c4cecc | 2879 | return NULL; |
ac939109 NP |
2880 | } |
2881 | ||
40469ee9 | 2882 | void *read_object_with_reference(const unsigned char *sha1, |
21666f1a | 2883 | const char *required_type_name, |
40469ee9 JH |
2884 | unsigned long *size, |
2885 | unsigned char *actual_sha1_return) | |
f4913f91 | 2886 | { |
21666f1a | 2887 | enum object_type type, required_type; |
f4913f91 JH |
2888 | void *buffer; |
2889 | unsigned long isize; | |
40469ee9 | 2890 | unsigned char actual_sha1[20]; |
f4913f91 | 2891 | |
21666f1a | 2892 | required_type = type_from_string(required_type_name); |
e702496e | 2893 | hashcpy(actual_sha1, sha1); |
40469ee9 JH |
2894 | while (1) { |
2895 | int ref_length = -1; | |
2896 | const char *ref_type = NULL; | |
f4913f91 | 2897 | |
21666f1a | 2898 | buffer = read_sha1_file(actual_sha1, &type, &isize); |
40469ee9 JH |
2899 | if (!buffer) |
2900 | return NULL; | |
21666f1a | 2901 | if (type == required_type) { |
40469ee9 JH |
2902 | *size = isize; |
2903 | if (actual_sha1_return) | |
e702496e | 2904 | hashcpy(actual_sha1_return, actual_sha1); |
40469ee9 JH |
2905 | return buffer; |
2906 | } | |
2907 | /* Handle references */ | |
21666f1a | 2908 | else if (type == OBJ_COMMIT) |
40469ee9 | 2909 | ref_type = "tree "; |
21666f1a | 2910 | else if (type == OBJ_TAG) |
40469ee9 JH |
2911 | ref_type = "object "; |
2912 | else { | |
2913 | free(buffer); | |
2914 | return NULL; | |
2915 | } | |
2916 | ref_length = strlen(ref_type); | |
f4913f91 | 2917 | |
50974ec9 MK |
2918 | if (ref_length + 40 > isize || |
2919 | memcmp(buffer, ref_type, ref_length) || | |
1d7f171c | 2920 | get_sha1_hex((char *) buffer + ref_length, actual_sha1)) { |
40469ee9 JH |
2921 | free(buffer); |
2922 | return NULL; | |
2923 | } | |
1cf58e72 | 2924 | free(buffer); |
40469ee9 JH |
2925 | /* Now we have the ID of the referred-to object in |
2926 | * actual_sha1. Check again. */ | |
f4913f91 | 2927 | } |
f4913f91 JH |
2928 | } |
2929 | ||
ce9fbf16 | 2930 | static void write_sha1_file_prepare(const void *buf, unsigned long len, |
972a9155 | 2931 | const char *type, unsigned char *sha1, |
d65a16f6 | 2932 | char *hdr, int *hdrlen) |
d410c0f5 | 2933 | { |
9126f009 | 2934 | git_SHA_CTX c; |
d410c0f5 JH |
2935 | |
2936 | /* Generate the header */ | |
ef1286d3 | 2937 | *hdrlen = xsnprintf(hdr, *hdrlen, "%s %lu", type, len)+1; |
d410c0f5 JH |
2938 | |
2939 | /* Sha1.. */ | |
9126f009 NP |
2940 | git_SHA1_Init(&c); |
2941 | git_SHA1_Update(&c, hdr, *hdrlen); | |
2942 | git_SHA1_Update(&c, buf, len); | |
2943 | git_SHA1_Final(sha1, &c); | |
d410c0f5 JH |
2944 | } |
2945 | ||
230f1322 | 2946 | /* |
5a688fe4 | 2947 | * Move the just written object into its final resting place. |
230f1322 | 2948 | */ |
cb5add58 | 2949 | int finalize_object_file(const char *tmpfile, const char *filename) |
230f1322 | 2950 | { |
e32c0a9c | 2951 | int ret = 0; |
5a688fe4 | 2952 | |
348df166 | 2953 | if (object_creation_mode == OBJECT_CREATION_USES_RENAMES) |
be66a6c4 JS |
2954 | goto try_rename; |
2955 | else if (link(tmpfile, filename)) | |
e32c0a9c | 2956 | ret = errno; |
7ebb6fca LT |
2957 | |
2958 | /* | |
2959 | * Coda hack - coda doesn't like cross-directory links, | |
2960 | * so we fall back to a rename, which will mean that it | |
2961 | * won't be able to check collisions, but that's not a | |
2962 | * big deal. | |
2963 | * | |
2964 | * The same holds for FAT formatted media. | |
2965 | * | |
3be1f18e | 2966 | * When this succeeds, we just return. We have nothing |
7ebb6fca LT |
2967 | * left to unlink. |
2968 | */ | |
2969 | if (ret && ret != EEXIST) { | |
be66a6c4 | 2970 | try_rename: |
7ebb6fca | 2971 | if (!rename(tmpfile, filename)) |
3be1f18e | 2972 | goto out; |
9e48b389 | 2973 | ret = errno; |
230f1322 | 2974 | } |
691f1a28 | 2975 | unlink_or_warn(tmpfile); |
230f1322 LT |
2976 | if (ret) { |
2977 | if (ret != EEXIST) { | |
82247e9b | 2978 | return error("unable to write sha1 filename %s: %s", filename, strerror(ret)); |
230f1322 LT |
2979 | } |
2980 | /* FIXME!!! Collision check here ? */ | |
2981 | } | |
2982 | ||
3be1f18e | 2983 | out: |
5256b006 | 2984 | if (adjust_shared_perm(filename)) |
5a688fe4 | 2985 | return error("unable to set permission to '%s'", filename); |
230f1322 LT |
2986 | return 0; |
2987 | } | |
2988 | ||
4d548150 LT |
2989 | static int write_buffer(int fd, const void *buf, size_t len) |
2990 | { | |
d34cf19b | 2991 | if (write_in_full(fd, buf, len) < 0) |
93822c22 | 2992 | return error("file write error (%s)", strerror(errno)); |
4d548150 LT |
2993 | return 0; |
2994 | } | |
2995 | ||
ce9fbf16 | 2996 | int hash_sha1_file(const void *buf, unsigned long len, const char *type, |
abdc3fc8 RS |
2997 | unsigned char *sha1) |
2998 | { | |
d65a16f6 | 2999 | char hdr[32]; |
ef1286d3 | 3000 | int hdrlen = sizeof(hdr); |
abdc3fc8 RS |
3001 | write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen); |
3002 | return 0; | |
3003 | } | |
3004 | ||
e9039dd3 LT |
3005 | /* Finalize a file on disk, and close it. */ |
3006 | static void close_sha1_file(int fd) | |
3007 | { | |
aafe9fba LT |
3008 | if (fsync_object_files) |
3009 | fsync_or_die(fd, "sha1 file"); | |
e9039dd3 | 3010 | if (close(fd) != 0) |
d824cbba | 3011 | die_errno("error when closing sha1 file"); |
e9039dd3 LT |
3012 | } |
3013 | ||
5723fe7e LT |
3014 | /* Size of directory component, including the ending '/' */ |
3015 | static inline int directory_size(const char *filename) | |
3016 | { | |
3017 | const char *s = strrchr(filename, '/'); | |
3018 | if (!s) | |
3019 | return 0; | |
3020 | return s - filename + 1; | |
3021 | } | |
3022 | ||
3023 | /* | |
3024 | * This creates a temporary file in the same directory as the final | |
3025 | * 'filename' | |
3026 | * | |
3027 | * We want to avoid cross-directory filename renames, because those | |
3028 | * can have problems on various filesystems (FAT, NFS, Coda). | |
3029 | */ | |
d4b3d11a | 3030 | static int create_tmpfile(struct strbuf *tmp, const char *filename) |
5723fe7e LT |
3031 | { |
3032 | int fd, dirlen = directory_size(filename); | |
3033 | ||
d4b3d11a JK |
3034 | strbuf_reset(tmp); |
3035 | strbuf_add(tmp, filename, dirlen); | |
3036 | strbuf_addstr(tmp, "tmp_obj_XXXXXX"); | |
3037 | fd = git_mkstemp_mode(tmp->buf, 0444); | |
cbacbf4e | 3038 | if (fd < 0 && dirlen && errno == ENOENT) { |
d4b3d11a JK |
3039 | /* |
3040 | * Make sure the directory exists; note that the contents | |
3041 | * of the buffer are undefined after mkstemp returns an | |
3042 | * error, so we have to rewrite the whole buffer from | |
3043 | * scratch. | |
3044 | */ | |
3045 | strbuf_reset(tmp); | |
3046 | strbuf_add(tmp, filename, dirlen - 1); | |
3047 | if (mkdir(tmp->buf, 0777) && errno != EEXIST) | |
b2476a60 | 3048 | return -1; |
d4b3d11a | 3049 | if (adjust_shared_perm(tmp->buf)) |
5723fe7e LT |
3050 | return -1; |
3051 | ||
3052 | /* Try again */ | |
d4b3d11a JK |
3053 | strbuf_addstr(tmp, "/tmp_obj_XXXXXX"); |
3054 | fd = git_mkstemp_mode(tmp->buf, 0444); | |
5723fe7e LT |
3055 | } |
3056 | return fd; | |
3057 | } | |
3058 | ||
bbac7311 | 3059 | static int write_loose_object(const unsigned char *sha1, char *hdr, int hdrlen, |
c00e657d | 3060 | const void *buf, unsigned long len, time_t mtime) |
0fcfd160 | 3061 | { |
915308b1 | 3062 | int fd, ret; |
9892beba | 3063 | unsigned char compressed[4096]; |
ef49a7a0 | 3064 | git_zstream stream; |
748af44c NP |
3065 | git_SHA_CTX c; |
3066 | unsigned char parano_sha1[20]; | |
d4b3d11a | 3067 | static struct strbuf tmp_file = STRBUF_INIT; |
30d6c6ea | 3068 | const char *filename = sha1_file_name(sha1); |
a44c9a5e | 3069 | |
d4b3d11a | 3070 | fd = create_tmpfile(&tmp_file, filename); |
aac17941 | 3071 | if (fd < 0) { |
35243577 | 3072 | if (errno == EACCES) |
82247e9b | 3073 | return error("insufficient permission for adding an object to repository database %s", get_object_directory()); |
916d081b | 3074 | else |
5eaeda70 | 3075 | return error("unable to create temporary file: %s", strerror(errno)); |
aac17941 LT |
3076 | } |
3077 | ||
0fcfd160 | 3078 | /* Set it up */ |
55bb5c91 | 3079 | git_deflate_init(&stream, zlib_compression_level); |
0fcfd160 | 3080 | stream.next_out = compressed; |
9892beba | 3081 | stream.avail_out = sizeof(compressed); |
748af44c | 3082 | git_SHA1_Init(&c); |
a44c9a5e LT |
3083 | |
3084 | /* First header.. */ | |
d65a16f6 | 3085 | stream.next_in = (unsigned char *)hdr; |
a44c9a5e | 3086 | stream.avail_in = hdrlen; |
55bb5c91 JH |
3087 | while (git_deflate(&stream, 0) == Z_OK) |
3088 | ; /* nothing */ | |
748af44c | 3089 | git_SHA1_Update(&c, hdr, hdrlen); |
a44c9a5e LT |
3090 | |
3091 | /* Then the data itself.. */ | |
c00e657d | 3092 | stream.next_in = (void *)buf; |
a44c9a5e | 3093 | stream.avail_in = len; |
9892beba | 3094 | do { |
748af44c | 3095 | unsigned char *in0 = stream.next_in; |
55bb5c91 | 3096 | ret = git_deflate(&stream, Z_FINISH); |
748af44c | 3097 | git_SHA1_Update(&c, in0, stream.next_in - in0); |
9892beba NP |
3098 | if (write_buffer(fd, compressed, stream.next_out - compressed) < 0) |
3099 | die("unable to write sha1 file"); | |
3100 | stream.next_out = compressed; | |
3101 | stream.avail_out = sizeof(compressed); | |
3102 | } while (ret == Z_OK); | |
3103 | ||
ac54c277 LT |
3104 | if (ret != Z_STREAM_END) |
3105 | die("unable to deflate new object %s (%d)", sha1_to_hex(sha1), ret); | |
55bb5c91 | 3106 | ret = git_deflate_end_gently(&stream); |
ac54c277 LT |
3107 | if (ret != Z_OK) |
3108 | die("deflateEnd on object %s failed (%d)", sha1_to_hex(sha1), ret); | |
748af44c NP |
3109 | git_SHA1_Final(parano_sha1, &c); |
3110 | if (hashcmp(sha1, parano_sha1) != 0) | |
3111 | die("confused by unstable object source data for %s", sha1_to_hex(sha1)); | |
ac54c277 | 3112 | |
e9039dd3 | 3113 | close_sha1_file(fd); |
0fcfd160 | 3114 | |
bbac7311 NP |
3115 | if (mtime) { |
3116 | struct utimbuf utb; | |
3117 | utb.actime = mtime; | |
3118 | utb.modtime = mtime; | |
d4b3d11a | 3119 | if (utime(tmp_file.buf, &utb) < 0) |
bbac7311 | 3120 | warning("failed utime() on %s: %s", |
d4b3d11a | 3121 | tmp_file.buf, strerror(errno)); |
bbac7311 NP |
3122 | } |
3123 | ||
d4b3d11a | 3124 | return finalize_object_file(tmp_file.buf, filename); |
0fcfd160 | 3125 | } |
8237b185 | 3126 | |
33d4221c JK |
3127 | static int freshen_loose_object(const unsigned char *sha1) |
3128 | { | |
3129 | return check_and_freshen(sha1, 1); | |
3130 | } | |
3131 | ||
3132 | static int freshen_packed_object(const unsigned char *sha1) | |
3133 | { | |
3134 | struct pack_entry e; | |
ee1c6c34 JK |
3135 | if (!find_pack_entry(sha1, &e)) |
3136 | return 0; | |
3137 | if (e.p->freshened) | |
3138 | return 1; | |
3139 | if (!freshen_file(e.p->pack_name)) | |
3140 | return 0; | |
3141 | e.p->freshened = 1; | |
3142 | return 1; | |
33d4221c JK |
3143 | } |
3144 | ||
1427a7ff | 3145 | int write_sha1_file(const void *buf, unsigned long len, const char *type, unsigned char *sha1) |
bbac7311 | 3146 | { |
bbac7311 | 3147 | char hdr[32]; |
ef1286d3 | 3148 | int hdrlen = sizeof(hdr); |
bbac7311 NP |
3149 | |
3150 | /* Normally if we have it in the pack then we do not bother writing | |
3151 | * it out into .git/objects/??/?{38} file. | |
3152 | */ | |
3153 | write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen); | |
b5f52f37 | 3154 | if (freshen_packed_object(sha1) || freshen_loose_object(sha1)) |
bbac7311 NP |
3155 | return 0; |
3156 | return write_loose_object(sha1, hdr, hdrlen, buf, len, 0); | |
3157 | } | |
3158 | ||
0c3db67c ES |
3159 | int hash_sha1_file_literally(const void *buf, unsigned long len, const char *type, |
3160 | unsigned char *sha1, unsigned flags) | |
3161 | { | |
3162 | char *header; | |
3163 | int hdrlen, status = 0; | |
3164 | ||
3165 | /* type string, SP, %lu of the length plus NUL must fit this */ | |
ef1286d3 JK |
3166 | hdrlen = strlen(type) + 32; |
3167 | header = xmalloc(hdrlen); | |
0c3db67c ES |
3168 | write_sha1_file_prepare(buf, len, type, sha1, header, &hdrlen); |
3169 | ||
3170 | if (!(flags & HASH_WRITE_OBJECT)) | |
3171 | goto cleanup; | |
051086b9 | 3172 | if (freshen_packed_object(sha1) || freshen_loose_object(sha1)) |
0c3db67c ES |
3173 | goto cleanup; |
3174 | status = write_loose_object(sha1, header, hdrlen, buf, len, 0); | |
3175 | ||
3176 | cleanup: | |
3177 | free(header); | |
3178 | return status; | |
3179 | } | |
3180 | ||
bbac7311 NP |
3181 | int force_object_loose(const unsigned char *sha1, time_t mtime) |
3182 | { | |
bbac7311 NP |
3183 | void *buf; |
3184 | unsigned long len; | |
3185 | enum object_type type; | |
3186 | char hdr[32]; | |
3187 | int hdrlen; | |
1fb23e65 | 3188 | int ret; |
bbac7311 | 3189 | |
c529d75a | 3190 | if (has_loose_object(sha1)) |
bbac7311 NP |
3191 | return 0; |
3192 | buf = read_packed_sha1(sha1, &type, &len); | |
3193 | if (!buf) | |
3194 | return error("cannot read sha1_file for %s", sha1_to_hex(sha1)); | |
ef1286d3 | 3195 | hdrlen = xsnprintf(hdr, sizeof(hdr), "%s %lu", typename(type), len) + 1; |
1fb23e65 BS |
3196 | ret = write_loose_object(sha1, hdr, hdrlen, buf, len, mtime); |
3197 | free(buf); | |
3198 | ||
3199 | return ret; | |
bbac7311 NP |
3200 | } |
3201 | ||
bf592c50 DB |
3202 | int has_pack_index(const unsigned char *sha1) |
3203 | { | |
3204 | struct stat st; | |
3205 | if (stat(sha1_pack_index_name(sha1), &st)) | |
3206 | return 0; | |
3207 | return 1; | |
bf592c50 DB |
3208 | } |
3209 | ||
cd673c1f JH |
3210 | int has_sha1_pack(const unsigned char *sha1) |
3211 | { | |
3212 | struct pack_entry e; | |
3213 | return find_pack_entry(sha1, &e); | |
3214 | } | |
3215 | ||
0eeb077b | 3216 | int has_sha1_file_with_flags(const unsigned char *sha1, int flags) |
8237b185 | 3217 | { |
1f688557 JH |
3218 | struct pack_entry e; |
3219 | ||
cd673c1f | 3220 | if (find_pack_entry(sha1, &e)) |
1f688557 | 3221 | return 1; |
45e8a748 JK |
3222 | if (has_loose_object(sha1)) |
3223 | return 1; | |
0eeb077b JK |
3224 | if (flags & HAS_SHA1_QUICK) |
3225 | return 0; | |
45e8a748 JK |
3226 | reprepare_packed_git(); |
3227 | return find_pack_entry(sha1, &e); | |
8237b185 | 3228 | } |
74400e71 | 3229 | |
c879daa2 NTND |
3230 | static void check_tree(const void *buf, size_t size) |
3231 | { | |
3232 | struct tree_desc desc; | |
3233 | struct name_entry entry; | |
3234 | ||
3235 | init_tree_desc(&desc, buf, size); | |
3236 | while (tree_entry(&desc, &entry)) | |
3237 | /* do nothing | |
3238 | * tree_entry() will die() on malformed entries */ | |
3239 | ; | |
3240 | } | |
3241 | ||
3242 | static void check_commit(const void *buf, size_t size) | |
3243 | { | |
3244 | struct commit c; | |
3245 | memset(&c, 0, sizeof(c)); | |
3246 | if (parse_commit_buffer(&c, buf, size)) | |
3247 | die("corrupt commit"); | |
3248 | } | |
3249 | ||
3250 | static void check_tag(const void *buf, size_t size) | |
3251 | { | |
3252 | struct tag t; | |
3253 | memset(&t, 0, sizeof(t)); | |
3254 | if (parse_tag_buffer(&t, buf, size)) | |
3255 | die("corrupt tag"); | |
3256 | } | |
3257 | ||
43df4f86 | 3258 | static int index_mem(unsigned char *sha1, void *buf, size_t size, |
c4ce46fc JH |
3259 | enum object_type type, |
3260 | const char *path, unsigned flags) | |
e7332f96 | 3261 | { |
6c510bee | 3262 | int ret, re_allocated = 0; |
c4ce46fc | 3263 | int write_object = flags & HASH_WRITE_OBJECT; |
74400e71 | 3264 | |
7672db20 | 3265 | if (!type) |
edaec3fb | 3266 | type = OBJ_BLOB; |
6c510bee LT |
3267 | |
3268 | /* | |
3269 | * Convert blobs to git internal format | |
3270 | */ | |
43df4f86 | 3271 | if ((type == OBJ_BLOB) && path) { |
f285a2d7 | 3272 | struct strbuf nbuf = STRBUF_INIT; |
21e5ad50 | 3273 | if (convert_to_git(path, buf, size, &nbuf, |
5e12e78e | 3274 | write_object ? safe_crlf : SAFE_CRLF_FALSE)) { |
b315c5c0 | 3275 | buf = strbuf_detach(&nbuf, &size); |
6c510bee LT |
3276 | re_allocated = 1; |
3277 | } | |
3278 | } | |
c4ce46fc | 3279 | if (flags & HASH_FORMAT_CHECK) { |
c879daa2 NTND |
3280 | if (type == OBJ_TREE) |
3281 | check_tree(buf, size); | |
3282 | if (type == OBJ_COMMIT) | |
3283 | check_commit(buf, size); | |
3284 | if (type == OBJ_TAG) | |
3285 | check_tag(buf, size); | |
3286 | } | |
6c510bee | 3287 | |
7672db20 | 3288 | if (write_object) |
edaec3fb | 3289 | ret = write_sha1_file(buf, size, typename(type), sha1); |
abdc3fc8 | 3290 | else |
edaec3fb | 3291 | ret = hash_sha1_file(buf, size, typename(type), sha1); |
43df4f86 | 3292 | if (re_allocated) |
6c510bee | 3293 | free(buf); |
43df4f86 DP |
3294 | return ret; |
3295 | } | |
3296 | ||
9035d75a SP |
3297 | static int index_stream_convert_blob(unsigned char *sha1, int fd, |
3298 | const char *path, unsigned flags) | |
3299 | { | |
3300 | int ret; | |
3301 | const int write_object = flags & HASH_WRITE_OBJECT; | |
3302 | struct strbuf sbuf = STRBUF_INIT; | |
3303 | ||
3304 | assert(path); | |
3305 | assert(would_convert_to_git_filter_fd(path)); | |
3306 | ||
3307 | convert_to_git_filter_fd(path, fd, &sbuf, | |
3308 | write_object ? safe_crlf : SAFE_CRLF_FALSE); | |
3309 | ||
3310 | if (write_object) | |
3311 | ret = write_sha1_file(sbuf.buf, sbuf.len, typename(OBJ_BLOB), | |
3312 | sha1); | |
3313 | else | |
3314 | ret = hash_sha1_file(sbuf.buf, sbuf.len, typename(OBJ_BLOB), | |
3315 | sha1); | |
3316 | strbuf_release(&sbuf); | |
3317 | return ret; | |
3318 | } | |
3319 | ||
7b41e1e1 JH |
3320 | static int index_pipe(unsigned char *sha1, int fd, enum object_type type, |
3321 | const char *path, unsigned flags) | |
3322 | { | |
3323 | struct strbuf sbuf = STRBUF_INIT; | |
3324 | int ret; | |
3325 | ||
3326 | if (strbuf_read(&sbuf, fd, 4096) >= 0) | |
3327 | ret = index_mem(sha1, sbuf.buf, sbuf.len, type, path, flags); | |
3328 | else | |
3329 | ret = -1; | |
3330 | strbuf_release(&sbuf); | |
3331 | return ret; | |
3332 | } | |
3333 | ||
ea68b0ce DP |
3334 | #define SMALL_FILE_SIZE (32*1024) |
3335 | ||
7b41e1e1 JH |
3336 | static int index_core(unsigned char *sha1, int fd, size_t size, |
3337 | enum object_type type, const char *path, | |
3338 | unsigned flags) | |
43df4f86 DP |
3339 | { |
3340 | int ret; | |
43df4f86 | 3341 | |
7b41e1e1 | 3342 | if (!size) { |
f6a1e1e2 | 3343 | ret = index_mem(sha1, "", size, type, path, flags); |
ea68b0ce DP |
3344 | } else if (size <= SMALL_FILE_SIZE) { |
3345 | char *buf = xmalloc(size); | |
3346 | if (size == read_in_full(fd, buf, size)) | |
c4ce46fc | 3347 | ret = index_mem(sha1, buf, size, type, path, flags); |
ea68b0ce DP |
3348 | else |
3349 | ret = error("short read %s", strerror(errno)); | |
3350 | free(buf); | |
08bda208 | 3351 | } else { |
43df4f86 | 3352 | void *buf = xmmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0); |
c4ce46fc | 3353 | ret = index_mem(sha1, buf, size, type, path, flags); |
aac17941 | 3354 | munmap(buf, size); |
08bda208 | 3355 | } |
7b41e1e1 JH |
3356 | return ret; |
3357 | } | |
3358 | ||
4dd1fbc7 | 3359 | /* |
568508e7 JH |
3360 | * This creates one packfile per large blob unless bulk-checkin |
3361 | * machinery is "plugged". | |
4dd1fbc7 JH |
3362 | * |
3363 | * This also bypasses the usual "convert-to-git" dance, and that is on | |
3364 | * purpose. We could write a streaming version of the converting | |
3365 | * functions and insert that before feeding the data to fast-import | |
4f22b101 JK |
3366 | * (or equivalent in-core API described above). However, that is |
3367 | * somewhat complicated, as we do not know the size of the filter | |
3368 | * result, which we need to know beforehand when writing a git object. | |
3369 | * Since the primary motivation for trying to stream from the working | |
3370 | * tree file and to avoid mmaping it in core is to deal with large | |
3371 | * binary blobs, they generally do not want to get any conversion, and | |
3372 | * callers should avoid this code path when filters are requested. | |
4dd1fbc7 JH |
3373 | */ |
3374 | static int index_stream(unsigned char *sha1, int fd, size_t size, | |
3375 | enum object_type type, const char *path, | |
3376 | unsigned flags) | |
3377 | { | |
568508e7 | 3378 | return index_bulk_checkin(sha1, fd, size, type, path, flags); |
4dd1fbc7 JH |
3379 | } |
3380 | ||
7b41e1e1 JH |
3381 | int index_fd(unsigned char *sha1, int fd, struct stat *st, |
3382 | enum object_type type, const char *path, unsigned flags) | |
3383 | { | |
3384 | int ret; | |
7b41e1e1 | 3385 | |
9079ab7c SP |
3386 | /* |
3387 | * Call xsize_t() only when needed to avoid potentially unnecessary | |
3388 | * die() for large files. | |
3389 | */ | |
9035d75a SP |
3390 | if (type == OBJ_BLOB && path && would_convert_to_git_filter_fd(path)) |
3391 | ret = index_stream_convert_blob(sha1, fd, path, flags); | |
3392 | else if (!S_ISREG(st->st_mode)) | |
7b41e1e1 | 3393 | ret = index_pipe(sha1, fd, type, path, flags); |
9079ab7c | 3394 | else if (st->st_size <= big_file_threshold || type != OBJ_BLOB || |
7ce7c760 | 3395 | (path && would_convert_to_git(path))) |
9079ab7c SP |
3396 | ret = index_core(sha1, fd, xsize_t(st->st_size), type, path, |
3397 | flags); | |
4dd1fbc7 | 3398 | else |
9079ab7c SP |
3399 | ret = index_stream(sha1, fd, xsize_t(st->st_size), type, path, |
3400 | flags); | |
43df4f86 | 3401 | close(fd); |
aac17941 | 3402 | return ret; |
74400e71 | 3403 | } |
ec1fcc16 | 3404 | |
c4ce46fc | 3405 | int index_path(unsigned char *sha1, const char *path, struct stat *st, unsigned flags) |
ec1fcc16 JH |
3406 | { |
3407 | int fd; | |
b760d3aa | 3408 | struct strbuf sb = STRBUF_INIT; |
ec1fcc16 JH |
3409 | |
3410 | switch (st->st_mode & S_IFMT) { | |
3411 | case S_IFREG: | |
3412 | fd = open(path, O_RDONLY); | |
3413 | if (fd < 0) | |
3414 | return error("open(\"%s\"): %s", path, | |
3415 | strerror(errno)); | |
c4ce46fc | 3416 | if (index_fd(sha1, fd, st, OBJ_BLOB, path, flags) < 0) |
ec1fcc16 JH |
3417 | return error("%s: failed to insert into database", |
3418 | path); | |
3419 | break; | |
3420 | case S_IFLNK: | |
b760d3aa | 3421 | if (strbuf_readlink(&sb, path, st->st_size)) { |
ec1fcc16 | 3422 | char *errstr = strerror(errno); |
ec1fcc16 JH |
3423 | return error("readlink(\"%s\"): %s", path, |
3424 | errstr); | |
3425 | } | |
c4ce46fc | 3426 | if (!(flags & HASH_WRITE_OBJECT)) |
b760d3aa LT |
3427 | hash_sha1_file(sb.buf, sb.len, blob_type, sha1); |
3428 | else if (write_sha1_file(sb.buf, sb.len, blob_type, sha1)) | |
ec1fcc16 JH |
3429 | return error("%s: failed to insert into database", |
3430 | path); | |
b760d3aa | 3431 | strbuf_release(&sb); |
ec1fcc16 | 3432 | break; |
f35a6d3b LT |
3433 | case S_IFDIR: |
3434 | return resolve_gitlink_ref(path, "HEAD", sha1); | |
ec1fcc16 JH |
3435 | default: |
3436 | return error("%s: unsupported file type", path); | |
3437 | } | |
3438 | return 0; | |
3439 | } | |
a69e5429 JH |
3440 | |
3441 | int read_pack_header(int fd, struct pack_header *header) | |
3442 | { | |
c697ad14 HO |
3443 | if (read_in_full(fd, header, sizeof(*header)) < sizeof(*header)) |
3444 | /* "eof before pack header was fully read" */ | |
3445 | return PH_ERROR_EOF; | |
3446 | ||
a69e5429 JH |
3447 | if (header->hdr_signature != htonl(PACK_SIGNATURE)) |
3448 | /* "protocol error (pack signature mismatch detected)" */ | |
3449 | return PH_ERROR_PACK_SIGNATURE; | |
3450 | if (!pack_version_ok(header->hdr_version)) | |
3451 | /* "protocol error (pack version unsupported)" */ | |
3452 | return PH_ERROR_PROTOCOL; | |
3453 | return 0; | |
3454 | } | |
40d52ff7 JK |
3455 | |
3456 | void assert_sha1_type(const unsigned char *sha1, enum object_type expect) | |
3457 | { | |
3458 | enum object_type type = sha1_object_info(sha1, NULL); | |
3459 | if (type < 0) | |
3460 | die("%s is not a valid object", sha1_to_hex(sha1)); | |
3461 | if (type != expect) | |
3462 | die("%s is not a valid '%s' object", sha1_to_hex(sha1), | |
3463 | typename(expect)); | |
3464 | } | |
27e1e22d JK |
3465 | |
3466 | static int for_each_file_in_obj_subdir(int subdir_nr, | |
3467 | struct strbuf *path, | |
3468 | each_loose_object_fn obj_cb, | |
3469 | each_loose_cruft_fn cruft_cb, | |
3470 | each_loose_subdir_fn subdir_cb, | |
3471 | void *data) | |
3472 | { | |
3473 | size_t baselen = path->len; | |
3474 | DIR *dir = opendir(path->buf); | |
3475 | struct dirent *de; | |
3476 | int r = 0; | |
3477 | ||
3478 | if (!dir) { | |
3479 | if (errno == ENOENT) | |
3480 | return 0; | |
3481 | return error("unable to open %s: %s", path->buf, strerror(errno)); | |
3482 | } | |
3483 | ||
3484 | while ((de = readdir(dir))) { | |
3485 | if (is_dot_or_dotdot(de->d_name)) | |
3486 | continue; | |
3487 | ||
3488 | strbuf_setlen(path, baselen); | |
3489 | strbuf_addf(path, "/%s", de->d_name); | |
3490 | ||
3491 | if (strlen(de->d_name) == 38) { | |
3492 | char hex[41]; | |
3493 | unsigned char sha1[20]; | |
3494 | ||
3495 | snprintf(hex, sizeof(hex), "%02x%s", | |
3496 | subdir_nr, de->d_name); | |
3497 | if (!get_sha1_hex(hex, sha1)) { | |
3498 | if (obj_cb) { | |
3499 | r = obj_cb(sha1, path->buf, data); | |
3500 | if (r) | |
3501 | break; | |
3502 | } | |
3503 | continue; | |
3504 | } | |
3505 | } | |
3506 | ||
3507 | if (cruft_cb) { | |
3508 | r = cruft_cb(de->d_name, path->buf, data); | |
3509 | if (r) | |
3510 | break; | |
3511 | } | |
3512 | } | |
3513 | strbuf_setlen(path, baselen); | |
3514 | ||
3515 | if (!r && subdir_cb) | |
3516 | r = subdir_cb(subdir_nr, path->buf, data); | |
3517 | ||
3518 | closedir(dir); | |
3519 | return r; | |
3520 | } | |
3521 | ||
e6f875e0 | 3522 | int for_each_loose_file_in_objdir_buf(struct strbuf *path, |
27e1e22d JK |
3523 | each_loose_object_fn obj_cb, |
3524 | each_loose_cruft_fn cruft_cb, | |
3525 | each_loose_subdir_fn subdir_cb, | |
3526 | void *data) | |
3527 | { | |
e6f875e0 | 3528 | size_t baselen = path->len; |
27e1e22d JK |
3529 | int r = 0; |
3530 | int i; | |
3531 | ||
27e1e22d | 3532 | for (i = 0; i < 256; i++) { |
e6f875e0 JK |
3533 | strbuf_addf(path, "/%02x", i); |
3534 | r = for_each_file_in_obj_subdir(i, path, obj_cb, cruft_cb, | |
27e1e22d | 3535 | subdir_cb, data); |
e6f875e0 | 3536 | strbuf_setlen(path, baselen); |
27e1e22d JK |
3537 | if (r) |
3538 | break; | |
3539 | } | |
3540 | ||
e6f875e0 JK |
3541 | return r; |
3542 | } | |
3543 | ||
3544 | int for_each_loose_file_in_objdir(const char *path, | |
3545 | each_loose_object_fn obj_cb, | |
3546 | each_loose_cruft_fn cruft_cb, | |
3547 | each_loose_subdir_fn subdir_cb, | |
3548 | void *data) | |
3549 | { | |
3550 | struct strbuf buf = STRBUF_INIT; | |
3551 | int r; | |
3552 | ||
3553 | strbuf_addstr(&buf, path); | |
3554 | r = for_each_loose_file_in_objdir_buf(&buf, obj_cb, cruft_cb, | |
3555 | subdir_cb, data); | |
27e1e22d | 3556 | strbuf_release(&buf); |
e6f875e0 | 3557 | |
27e1e22d JK |
3558 | return r; |
3559 | } | |
660c889e JK |
3560 | |
3561 | struct loose_alt_odb_data { | |
3562 | each_loose_object_fn *cb; | |
3563 | void *data; | |
3564 | }; | |
3565 | ||
3566 | static int loose_from_alt_odb(struct alternate_object_database *alt, | |
3567 | void *vdata) | |
3568 | { | |
3569 | struct loose_alt_odb_data *data = vdata; | |
b0a42642 JM |
3570 | struct strbuf buf = STRBUF_INIT; |
3571 | int r; | |
3572 | ||
3573 | /* copy base not including trailing '/' */ | |
3574 | strbuf_add(&buf, alt->base, alt->name - alt->base - 1); | |
3575 | r = for_each_loose_file_in_objdir_buf(&buf, | |
3576 | data->cb, NULL, NULL, | |
3577 | data->data); | |
3578 | strbuf_release(&buf); | |
3579 | return r; | |
660c889e JK |
3580 | } |
3581 | ||
1385bb7b | 3582 | int for_each_loose_object(each_loose_object_fn cb, void *data, unsigned flags) |
660c889e JK |
3583 | { |
3584 | struct loose_alt_odb_data alt; | |
3585 | int r; | |
3586 | ||
3587 | r = for_each_loose_file_in_objdir(get_object_directory(), | |
3588 | cb, NULL, NULL, data); | |
3589 | if (r) | |
3590 | return r; | |
3591 | ||
1385bb7b JK |
3592 | if (flags & FOR_EACH_OBJECT_LOCAL_ONLY) |
3593 | return 0; | |
3594 | ||
660c889e JK |
3595 | alt.cb = cb; |
3596 | alt.data = data; | |
3597 | return foreach_alt_odb(loose_from_alt_odb, &alt); | |
3598 | } | |
3599 | ||
3600 | static int for_each_object_in_pack(struct packed_git *p, each_packed_object_fn cb, void *data) | |
3601 | { | |
3602 | uint32_t i; | |
3603 | int r = 0; | |
3604 | ||
3605 | for (i = 0; i < p->num_objects; i++) { | |
3606 | const unsigned char *sha1 = nth_packed_object_sha1(p, i); | |
3607 | ||
3608 | if (!sha1) | |
3609 | return error("unable to get sha1 of object %u in %s", | |
3610 | i, p->pack_name); | |
3611 | ||
3612 | r = cb(sha1, p, i, data); | |
3613 | if (r) | |
3614 | break; | |
3615 | } | |
3616 | return r; | |
3617 | } | |
3618 | ||
1385bb7b | 3619 | int for_each_packed_object(each_packed_object_fn cb, void *data, unsigned flags) |
660c889e JK |
3620 | { |
3621 | struct packed_git *p; | |
3622 | int r = 0; | |
f813e9ea | 3623 | int pack_errors = 0; |
660c889e JK |
3624 | |
3625 | prepare_packed_git(); | |
3626 | for (p = packed_git; p; p = p->next) { | |
1385bb7b JK |
3627 | if ((flags & FOR_EACH_OBJECT_LOCAL_ONLY) && !p->pack_local) |
3628 | continue; | |
f813e9ea JK |
3629 | if (open_pack_index(p)) { |
3630 | pack_errors = 1; | |
3631 | continue; | |
3632 | } | |
660c889e JK |
3633 | r = for_each_object_in_pack(p, cb, data); |
3634 | if (r) | |
3635 | break; | |
3636 | } | |
f813e9ea | 3637 | return r ? r : pack_errors; |
660c889e | 3638 | } |