]>
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" |
1f688557 | 11 | #include "delta.h" |
a733cb60 | 12 | #include "pack.h" |
8e440259 PE |
13 | #include "blob.h" |
14 | #include "commit.h" | |
4dd1fbc7 | 15 | #include "run-command.h" |
8e440259 PE |
16 | #include "tag.h" |
17 | #include "tree.h" | |
c879daa2 | 18 | #include "tree-walk.h" |
f35a6d3b | 19 | #include "refs.h" |
70f5d5d3 | 20 | #include "pack-revindex.h" |
628522ec | 21 | #include "sha1-lookup.h" |
568508e7 | 22 | #include "bulk-checkin.h" |
090ea126 | 23 | #include "streaming.h" |
543c5caa | 24 | #include "dir.h" |
0fcfd160 | 25 | |
144bde78 LT |
26 | #ifndef O_NOATIME |
27 | #if defined(__linux__) && (defined(__i386__) || defined(__PPC__)) | |
28 | #define O_NOATIME 01000000 | |
29 | #else | |
30 | #define O_NOATIME 0 | |
31 | #endif | |
32 | #endif | |
33 | ||
28bd70d8 JN |
34 | #define SZ_FMT PRIuMAX |
35 | static inline uintmax_t sz_fmt(size_t s) { return s; } | |
e05db0fd | 36 | |
96f1e58f | 37 | const unsigned char null_sha1[20]; |
88cd621d | 38 | |
b12ca963 NTND |
39 | static const char *no_log_pack_access = "no_log_pack_access"; |
40 | static const char *log_pack_access; | |
41 | ||
c597ba80 NTND |
42 | /* |
43 | * This is meant to hold a *small* number of objects that you would | |
44 | * want read_sha1_file() to be able to return, but yet you do not want | |
45 | * to write them into the object store (e.g. a browse-only | |
46 | * application). | |
47 | */ | |
48 | static struct cached_object { | |
49 | unsigned char sha1[20]; | |
50 | enum object_type type; | |
51 | void *buf; | |
52 | unsigned long size; | |
53 | } *cached_objects; | |
54 | static int cached_object_nr, cached_object_alloc; | |
55 | ||
56 | static struct cached_object empty_tree = { | |
dab0d410 | 57 | EMPTY_TREE_SHA1_BIN_LITERAL, |
c597ba80 NTND |
58 | OBJ_TREE, |
59 | "", | |
60 | 0 | |
61 | }; | |
62 | ||
c01f51cc NTND |
63 | static struct packed_git *last_found_pack; |
64 | ||
c597ba80 NTND |
65 | static struct cached_object *find_cached_object(const unsigned char *sha1) |
66 | { | |
67 | int i; | |
68 | struct cached_object *co = cached_objects; | |
69 | ||
70 | for (i = 0; i < cached_object_nr; i++, co++) { | |
71 | if (!hashcmp(co->sha1, sha1)) | |
72 | return co; | |
73 | } | |
74 | if (!hashcmp(sha1, empty_tree.sha1)) | |
75 | return &empty_tree; | |
76 | return NULL; | |
77 | } | |
78 | ||
90a6464b JH |
79 | int mkdir_in_gitdir(const char *path) |
80 | { | |
81 | if (mkdir(path, 0777)) { | |
82 | int saved_errno = errno; | |
83 | struct stat st; | |
84 | struct strbuf sb = STRBUF_INIT; | |
85 | ||
86 | if (errno != EEXIST) | |
87 | return -1; | |
88 | /* | |
89 | * Are we looking at a path in a symlinked worktree | |
90 | * whose original repository does not yet have it? | |
91 | * e.g. .git/rr-cache pointing at its original | |
92 | * repository in which the user hasn't performed any | |
93 | * conflict resolution yet? | |
94 | */ | |
95 | if (lstat(path, &st) || !S_ISLNK(st.st_mode) || | |
96 | strbuf_readlink(&sb, path, st.st_size) || | |
97 | !is_absolute_path(sb.buf) || | |
98 | mkdir(sb.buf, 0777)) { | |
99 | strbuf_release(&sb); | |
100 | errno = saved_errno; | |
101 | return -1; | |
102 | } | |
103 | strbuf_release(&sb); | |
104 | } | |
105 | return adjust_shared_perm(path); | |
106 | } | |
107 | ||
b2cb9425 JH |
108 | int safe_create_leading_directories(char *path) |
109 | { | |
8385abfd | 110 | char *pos = path + offset_1st_component(path); |
67d42212 JR |
111 | struct stat st; |
112 | ||
b2cb9425 JH |
113 | while (pos) { |
114 | pos = strchr(pos, '/'); | |
115 | if (!pos) | |
116 | break; | |
5f0bdf50 JH |
117 | while (*++pos == '/') |
118 | ; | |
119 | if (!*pos) | |
120 | break; | |
121 | *--pos = '\0'; | |
67d42212 JR |
122 | if (!stat(path, &st)) { |
123 | /* path exists */ | |
124 | if (!S_ISDIR(st.st_mode)) { | |
b2cb9425 | 125 | *pos = '/'; |
67d42212 | 126 | return -3; |
b2cb9425 | 127 | } |
457f06d6 | 128 | } |
67d42212 | 129 | else if (mkdir(path, 0777)) { |
928734d9 SW |
130 | if (errno == EEXIST && |
131 | !stat(path, &st) && S_ISDIR(st.st_mode)) { | |
132 | ; /* somebody created it since we checked */ | |
133 | } else { | |
134 | *pos = '/'; | |
135 | return -1; | |
136 | } | |
67d42212 | 137 | } |
457f06d6 JS |
138 | else if (adjust_shared_perm(path)) { |
139 | *pos = '/'; | |
140 | return -2; | |
141 | } | |
b2cb9425 JH |
142 | *pos++ = '/'; |
143 | } | |
144 | return 0; | |
145 | } | |
723c31fe | 146 | |
8e21d63b JK |
147 | int safe_create_leading_directories_const(const char *path) |
148 | { | |
149 | /* path points to cache entries, so xstrdup before messing with it */ | |
150 | char *buf = xstrdup(path); | |
151 | int result = safe_create_leading_directories(buf); | |
152 | free(buf); | |
153 | return result; | |
154 | } | |
155 | ||
ace1534d JH |
156 | static void fill_sha1_path(char *pathbuf, const unsigned char *sha1) |
157 | { | |
158 | int i; | |
159 | for (i = 0; i < 20; i++) { | |
160 | static char hex[] = "0123456789abcdef"; | |
161 | unsigned int val = sha1[i]; | |
162 | char *pos = pathbuf + i*2 + (i > 0); | |
163 | *pos++ = hex[val >> 4]; | |
164 | *pos = hex[val & 0xf]; | |
165 | } | |
166 | } | |
167 | ||
0fcfd160 LT |
168 | /* |
169 | * NOTE! This returns a statically allocated buffer, so you have to be | |
790296fd | 170 | * careful about using it. Do an "xstrdup()" if you need to save the |
0fcfd160 | 171 | * filename. |
ace1534d JH |
172 | * |
173 | * Also note that this returns the location for creating. Reading | |
174 | * SHA1 file can happen from any alternate directory listed in the | |
d19938ab | 175 | * DB_ENVIRONMENT environment variable if it is not found in |
ace1534d | 176 | * the primary object database. |
0fcfd160 LT |
177 | */ |
178 | char *sha1_file_name(const unsigned char *sha1) | |
179 | { | |
560fb6a1 JK |
180 | static char buf[PATH_MAX]; |
181 | const char *objdir; | |
182 | int len; | |
0fcfd160 | 183 | |
560fb6a1 JK |
184 | objdir = get_object_directory(); |
185 | len = strlen(objdir); | |
186 | ||
187 | /* '/' + sha1(2) + '/' + sha1(38) + '\0' */ | |
188 | if (len + 43 > PATH_MAX) | |
189 | die("insanely long object directory %s", objdir); | |
190 | memcpy(buf, objdir, len); | |
191 | buf[len] = '/'; | |
192 | buf[len+3] = '/'; | |
193 | buf[len+42] = '\0'; | |
194 | fill_sha1_path(buf + len + 1, sha1); | |
195 | return buf; | |
0fcfd160 LT |
196 | } |
197 | ||
633f43e1 | 198 | static char *sha1_get_pack_name(const unsigned char *sha1, |
6eec46bd | 199 | char **name, char **base, const char *which) |
bf592c50 DB |
200 | { |
201 | static const char hex[] = "0123456789abcdef"; | |
633f43e1 | 202 | char *buf; |
bf592c50 DB |
203 | int i; |
204 | ||
633f43e1 | 205 | if (!*base) { |
bf592c50 DB |
206 | const char *sha1_file_directory = get_object_directory(); |
207 | int len = strlen(sha1_file_directory); | |
633f43e1 | 208 | *base = xmalloc(len + 60); |
6eec46bd JH |
209 | sprintf(*base, "%s/pack/pack-1234567890123456789012345678901234567890.%s", |
210 | sha1_file_directory, which); | |
633f43e1 | 211 | *name = *base + len + 11; |
bf592c50 DB |
212 | } |
213 | ||
633f43e1 | 214 | buf = *name; |
bf592c50 DB |
215 | |
216 | for (i = 0; i < 20; i++) { | |
217 | unsigned int val = *sha1++; | |
218 | *buf++ = hex[val >> 4]; | |
219 | *buf++ = hex[val & 0xf]; | |
220 | } | |
a6080a0a | 221 | |
633f43e1 | 222 | return *base; |
bf592c50 DB |
223 | } |
224 | ||
633f43e1 | 225 | char *sha1_pack_name(const unsigned char *sha1) |
bf592c50 | 226 | { |
633f43e1 | 227 | static char *name, *base; |
bf592c50 | 228 | |
6eec46bd | 229 | return sha1_get_pack_name(sha1, &name, &base, "pack"); |
633f43e1 | 230 | } |
bf592c50 | 231 | |
633f43e1 HO |
232 | char *sha1_pack_index_name(const unsigned char *sha1) |
233 | { | |
234 | static char *name, *base; | |
a6080a0a | 235 | |
6eec46bd | 236 | return sha1_get_pack_name(sha1, &name, &base, "idx"); |
bf592c50 DB |
237 | } |
238 | ||
d5a63b99 JH |
239 | struct alternate_object_database *alt_odb_list; |
240 | static struct alternate_object_database **alt_odb_tail; | |
ace1534d | 241 | |
f4e51683 | 242 | static int git_open_noatime(const char *name); |
c2f493a4 | 243 | |
ddd5d056 JH |
244 | /* |
245 | * Prepare alternate object database registry. | |
d5a63b99 JH |
246 | * |
247 | * The variable alt_odb_list points at the list of struct | |
248 | * alternate_object_database. The elements on this list come from | |
249 | * non-empty elements from colon separated ALTERNATE_DB_ENVIRONMENT | |
250 | * environment variable, and $GIT_OBJECT_DIRECTORY/info/alternates, | |
1494e038 JH |
251 | * whose contents is similar to that environment variable but can be |
252 | * LF separated. Its base points at a statically allocated buffer that | |
d5a63b99 JH |
253 | * contains "/the/directory/corresponding/to/.git/objects/...", while |
254 | * its name points just after the slash at the end of ".git/objects/" | |
255 | * in the example above, and has enough space to hold 40-byte hex | |
256 | * SHA1, an extra slash for the first level indirection, and the | |
257 | * terminating NUL. | |
ddd5d056 | 258 | */ |
6eac50d8 | 259 | static int link_alt_odb_entry(const char *entry, const char *relative_base, int depth) |
ace1534d | 260 | { |
1494e038 | 261 | const char *objdir = get_object_directory(); |
c2f493a4 MW |
262 | struct alternate_object_database *ent; |
263 | struct alternate_object_database *alt; | |
5bdf0a84 HW |
264 | int pfxlen, entlen; |
265 | struct strbuf pathbuf = STRBUF_INIT; | |
d5a63b99 | 266 | |
85dadc38 | 267 | if (!is_absolute_path(entry) && relative_base) { |
5bdf0a84 HW |
268 | strbuf_addstr(&pathbuf, real_path(relative_base)); |
269 | strbuf_addch(&pathbuf, '/'); | |
c2f493a4 | 270 | } |
6eac50d8 | 271 | strbuf_addstr(&pathbuf, entry); |
c2f493a4 | 272 | |
5bdf0a84 HW |
273 | normalize_path_copy(pathbuf.buf, pathbuf.buf); |
274 | ||
275 | pfxlen = strlen(pathbuf.buf); | |
276 | ||
277 | /* | |
278 | * The trailing slash after the directory name is given by | |
279 | * this function at the end. Remove duplicates. | |
280 | */ | |
281 | while (pfxlen && pathbuf.buf[pfxlen-1] == '/') | |
282 | pfxlen -= 1; | |
283 | ||
284 | entlen = pfxlen + 43; /* '/' + 2 hex + '/' + 38 hex + NUL */ | |
285 | ent = xmalloc(sizeof(*ent) + entlen); | |
286 | memcpy(ent->base, pathbuf.buf, pfxlen); | |
287 | strbuf_release(&pathbuf); | |
c2f493a4 MW |
288 | |
289 | ent->name = ent->base + pfxlen + 1; | |
290 | ent->base[pfxlen + 3] = '/'; | |
291 | ent->base[pfxlen] = ent->base[entlen-1] = 0; | |
292 | ||
293 | /* Detect cases where alternate disappeared */ | |
90b4a71c | 294 | if (!is_directory(ent->base)) { |
c2f493a4 MW |
295 | error("object directory %s does not exist; " |
296 | "check .git/objects/info/alternates.", | |
297 | ent->base); | |
298 | free(ent); | |
299 | return -1; | |
300 | } | |
301 | ||
302 | /* Prevent the common mistake of listing the same | |
303 | * thing twice, or object directory itself. | |
304 | */ | |
305 | for (alt = alt_odb_list; alt; alt = alt->next) { | |
306 | if (!memcmp(ent->base, alt->base, pfxlen)) { | |
307 | free(ent); | |
308 | return -1; | |
309 | } | |
310 | } | |
cb2912c3 | 311 | if (!strcmp(ent->base, objdir)) { |
c2f493a4 MW |
312 | free(ent); |
313 | return -1; | |
314 | } | |
315 | ||
316 | /* add the alternate entry */ | |
317 | *alt_odb_tail = ent; | |
318 | alt_odb_tail = &(ent->next); | |
319 | ent->next = NULL; | |
320 | ||
321 | /* recursively add alternates */ | |
322 | read_info_alternates(ent->base, depth + 1); | |
323 | ||
324 | ent->base[pfxlen] = '/'; | |
325 | ||
326 | return 0; | |
327 | } | |
328 | ||
c5950164 | 329 | static void link_alt_odb_entries(const char *alt, int len, int sep, |
c2f493a4 MW |
330 | const char *relative_base, int depth) |
331 | { | |
6eac50d8 MH |
332 | struct string_list entries = STRING_LIST_INIT_NODUP; |
333 | char *alt_copy; | |
334 | int i; | |
c2f493a4 MW |
335 | |
336 | if (depth > 5) { | |
337 | error("%s: ignoring alternate object stores, nesting too deep.", | |
338 | relative_base); | |
339 | return; | |
340 | } | |
341 | ||
c5950164 | 342 | alt_copy = xmemdupz(alt, len); |
6eac50d8 MH |
343 | string_list_split_in_place(&entries, alt_copy, sep, -1); |
344 | for (i = 0; i < entries.nr; i++) { | |
345 | const char *entry = entries.items[i].string; | |
346 | if (entry[0] == '\0' || entry[0] == '#') | |
9577e7e3 | 347 | continue; |
6eac50d8 MH |
348 | if (!is_absolute_path(entry) && depth) { |
349 | error("%s: ignoring relative alternate object store %s", | |
350 | relative_base, entry); | |
351 | } else { | |
352 | link_alt_odb_entry(entry, relative_base, depth); | |
9577e7e3 | 353 | } |
9577e7e3 | 354 | } |
6eac50d8 MH |
355 | string_list_clear(&entries, 0); |
356 | free(alt_copy); | |
d5a63b99 JH |
357 | } |
358 | ||
5e73633d | 359 | void read_info_alternates(const char * relative_base, int depth) |
d5a63b99 | 360 | { |
9577e7e3 | 361 | char *map; |
dc49cd76 | 362 | size_t mapsz; |
d5a63b99 | 363 | struct stat st; |
9cb18f56 JM |
364 | const char alt_file_name[] = "info/alternates"; |
365 | /* Given that relative_base is no longer than PATH_MAX, | |
366 | ensure that "path" has enough space to append "/", the | |
367 | file name, "info/alternates", and a trailing NUL. */ | |
368 | char path[PATH_MAX + 1 + sizeof alt_file_name]; | |
c2f493a4 | 369 | int fd; |
d5a63b99 | 370 | |
9cb18f56 | 371 | sprintf(path, "%s/%s", relative_base, alt_file_name); |
f4e51683 | 372 | fd = git_open_noatime(path); |
d5a63b99 JH |
373 | if (fd < 0) |
374 | return; | |
375 | if (fstat(fd, &st) || (st.st_size == 0)) { | |
376 | close(fd); | |
9a217f2a | 377 | return; |
ace1534d | 378 | } |
dc49cd76 SP |
379 | mapsz = xsize_t(st.st_size); |
380 | map = xmmap(NULL, mapsz, PROT_READ, MAP_PRIVATE, fd, 0); | |
d5a63b99 | 381 | close(fd); |
d5a63b99 | 382 | |
c5950164 | 383 | link_alt_odb_entries(map, mapsz, '\n', relative_base, depth); |
c2f493a4 | 384 | |
dc49cd76 | 385 | munmap(map, mapsz); |
ace1534d JH |
386 | } |
387 | ||
bef70b22 DB |
388 | void add_to_alternates_file(const char *reference) |
389 | { | |
390 | struct lock_file *lock = xcalloc(1, sizeof(struct lock_file)); | |
acd3b9ec | 391 | int fd = hold_lock_file_for_append(lock, git_path("objects/info/alternates"), LOCK_DIE_ON_ERROR); |
e6baf4a1 | 392 | char *alt = mkpath("%s\n", reference); |
bef70b22 DB |
393 | write_or_die(fd, alt, strlen(alt)); |
394 | if (commit_lock_file(lock)) | |
395 | die("could not close alternates file"); | |
396 | if (alt_odb_tail) | |
c5950164 | 397 | link_alt_odb_entries(alt, strlen(alt), '\n', NULL, 0); |
bef70b22 DB |
398 | } |
399 | ||
d79796bc JH |
400 | void foreach_alt_odb(alt_odb_fn fn, void *cb) |
401 | { | |
402 | struct alternate_object_database *ent; | |
403 | ||
404 | prepare_alt_odb(); | |
405 | for (ent = alt_odb_list; ent; ent = ent->next) | |
406 | if (fn(ent, cb)) | |
407 | return; | |
408 | } | |
409 | ||
c2f493a4 MW |
410 | void prepare_alt_odb(void) |
411 | { | |
554fe20d | 412 | const char *alt; |
c2f493a4 | 413 | |
7dc24aa5 SP |
414 | if (alt_odb_tail) |
415 | return; | |
416 | ||
c2f493a4 MW |
417 | alt = getenv(ALTERNATE_DB_ENVIRONMENT); |
418 | if (!alt) alt = ""; | |
419 | ||
c2f493a4 | 420 | alt_odb_tail = &alt_odb_list; |
c5950164 | 421 | link_alt_odb_entries(alt, strlen(alt), PATH_SEP, NULL, 0); |
c2f493a4 MW |
422 | |
423 | read_info_alternates(get_object_directory(), 0); | |
424 | } | |
425 | ||
0f4dc14a | 426 | static int has_loose_object_local(const unsigned char *sha1) |
ace1534d | 427 | { |
ace1534d | 428 | char *name = sha1_file_name(sha1); |
0f4dc14a BC |
429 | return !access(name, F_OK); |
430 | } | |
ace1534d | 431 | |
0f4dc14a BC |
432 | int has_loose_object_nonlocal(const unsigned char *sha1) |
433 | { | |
434 | struct alternate_object_database *alt; | |
9a217f2a | 435 | prepare_alt_odb(); |
d5a63b99 | 436 | for (alt = alt_odb_list; alt; alt = alt->next) { |
0f4dc14a | 437 | fill_sha1_path(alt->name, sha1); |
c529d75a LT |
438 | if (!access(alt->base, F_OK)) |
439 | return 1; | |
ace1534d | 440 | } |
c529d75a | 441 | return 0; |
ace1534d JH |
442 | } |
443 | ||
0f4dc14a BC |
444 | static int has_loose_object(const unsigned char *sha1) |
445 | { | |
446 | return has_loose_object_local(sha1) || | |
447 | has_loose_object_nonlocal(sha1); | |
448 | } | |
449 | ||
60bb8b14 | 450 | static unsigned int pack_used_ctr; |
a53128b6 SP |
451 | static unsigned int pack_mmap_calls; |
452 | static unsigned int peak_pack_open_windows; | |
453 | static unsigned int pack_open_windows; | |
c7934306 SP |
454 | static unsigned int pack_open_fds; |
455 | static unsigned int pack_max_fds; | |
a53128b6 | 456 | static size_t peak_pack_mapped; |
60bb8b14 | 457 | static size_t pack_mapped; |
9a217f2a | 458 | struct packed_git *packed_git; |
1f688557 | 459 | |
b79d18c9 | 460 | void pack_report(void) |
a53128b6 SP |
461 | { |
462 | fprintf(stderr, | |
e05db0fd PR |
463 | "pack_report: getpagesize() = %10" SZ_FMT "\n" |
464 | "pack_report: core.packedGitWindowSize = %10" SZ_FMT "\n" | |
465 | "pack_report: core.packedGitLimit = %10" SZ_FMT "\n", | |
9e42d6a1 SP |
466 | sz_fmt(getpagesize()), |
467 | sz_fmt(packed_git_window_size), | |
468 | sz_fmt(packed_git_limit)); | |
a53128b6 SP |
469 | fprintf(stderr, |
470 | "pack_report: pack_used_ctr = %10u\n" | |
471 | "pack_report: pack_mmap_calls = %10u\n" | |
472 | "pack_report: pack_open_windows = %10u / %10u\n" | |
e05db0fd PR |
473 | "pack_report: pack_mapped = " |
474 | "%10" SZ_FMT " / %10" SZ_FMT "\n", | |
a53128b6 SP |
475 | pack_used_ctr, |
476 | pack_mmap_calls, | |
477 | pack_open_windows, peak_pack_open_windows, | |
9e42d6a1 | 478 | sz_fmt(pack_mapped), sz_fmt(peak_pack_mapped)); |
a53128b6 SP |
479 | } |
480 | ||
42873078 | 481 | static int check_packed_git_idx(const char *path, struct packed_git *p) |
1f688557 JH |
482 | { |
483 | void *idx_map; | |
42873078 | 484 | struct pack_idx_header *hdr; |
dc49cd76 | 485 | size_t idx_size; |
74e34e1f | 486 | uint32_t version, nr, i, *index; |
f4e51683 | 487 | int fd = git_open_noatime(path); |
1f688557 | 488 | struct stat st; |
42873078 | 489 | |
1f688557 JH |
490 | if (fd < 0) |
491 | return -1; | |
492 | if (fstat(fd, &st)) { | |
493 | close(fd); | |
494 | return -1; | |
495 | } | |
dc49cd76 | 496 | idx_size = xsize_t(st.st_size); |
2d88451b SP |
497 | if (idx_size < 4 * 256 + 20 + 20) { |
498 | close(fd); | |
499 | return error("index file %s is too small", path); | |
500 | } | |
c4712e45 | 501 | idx_map = xmmap(NULL, idx_size, PROT_READ, MAP_PRIVATE, fd, 0); |
1f688557 | 502 | close(fd); |
1f688557 | 503 | |
42873078 NP |
504 | hdr = idx_map; |
505 | if (hdr->idx_signature == htonl(PACK_IDX_SIGNATURE)) { | |
74e34e1f NP |
506 | version = ntohl(hdr->idx_version); |
507 | if (version < 2 || version > 2) { | |
508 | munmap(idx_map, idx_size); | |
6e1c2344 | 509 | return error("index file %s is version %"PRIu32 |
74e34e1f NP |
510 | " and is not supported by this binary" |
511 | " (try upgrading GIT to a newer version)", | |
512 | path, version); | |
513 | } | |
514 | } else | |
515 | version = 1; | |
df1b059d | 516 | |
1f688557 | 517 | nr = 0; |
42873078 | 518 | index = idx_map; |
74e34e1f NP |
519 | if (version > 1) |
520 | index += 2; /* skip index header */ | |
1f688557 | 521 | for (i = 0; i < 256; i++) { |
326bf396 | 522 | uint32_t n = ntohl(index[i]); |
2d88451b SP |
523 | if (n < nr) { |
524 | munmap(idx_map, idx_size); | |
df1b059d | 525 | return error("non-monotonic index %s", path); |
2d88451b | 526 | } |
1f688557 JH |
527 | nr = n; |
528 | } | |
529 | ||
74e34e1f NP |
530 | if (version == 1) { |
531 | /* | |
532 | * Total size: | |
533 | * - 256 index entries 4 bytes each | |
534 | * - 24-byte entries * nr (20-byte sha1 + 4-byte offset) | |
535 | * - 20-byte SHA1 of the packfile | |
536 | * - 20-byte SHA1 file checksum | |
537 | */ | |
538 | if (idx_size != 4*256 + nr * 24 + 20 + 20) { | |
539 | munmap(idx_map, idx_size); | |
eef427a0 | 540 | return error("wrong index v1 file size in %s", path); |
74e34e1f NP |
541 | } |
542 | } else if (version == 2) { | |
543 | /* | |
544 | * Minimum size: | |
545 | * - 8 bytes of header | |
546 | * - 256 index entries 4 bytes each | |
547 | * - 20-byte sha1 entry * nr | |
548 | * - 4-byte crc entry * nr | |
549 | * - 4-byte offset entry * nr | |
550 | * - 20-byte SHA1 of the packfile | |
551 | * - 20-byte SHA1 file checksum | |
552 | * And after the 4-byte offset table might be a | |
553 | * variable sized table containing 8-byte entries | |
554 | * for offsets larger than 2^31. | |
555 | */ | |
556 | unsigned long min_size = 8 + 4*256 + nr*(20 + 4 + 4) + 20 + 20; | |
1164f1e4 LT |
557 | unsigned long max_size = min_size; |
558 | if (nr) | |
559 | max_size += (nr - 1)*8; | |
560 | if (idx_size < min_size || idx_size > max_size) { | |
74e34e1f | 561 | munmap(idx_map, idx_size); |
eef427a0 | 562 | return error("wrong index v2 file size in %s", path); |
74e34e1f | 563 | } |
7109c889 JH |
564 | if (idx_size != min_size && |
565 | /* | |
566 | * make sure we can deal with large pack offsets. | |
567 | * 31-bit signed offset won't be enough, neither | |
568 | * 32-bit unsigned one will be. | |
569 | */ | |
570 | (sizeof(off_t) <= 4)) { | |
571 | munmap(idx_map, idx_size); | |
572 | return error("pack too large for current definition of off_t in %s", path); | |
74e34e1f | 573 | } |
2d88451b | 574 | } |
1f688557 | 575 | |
74e34e1f | 576 | p->index_version = version; |
42873078 NP |
577 | p->index_data = idx_map; |
578 | p->index_size = idx_size; | |
57059091 | 579 | p->num_objects = nr; |
1f688557 JH |
580 | return 0; |
581 | } | |
582 | ||
bc8e478a | 583 | int open_pack_index(struct packed_git *p) |
d079837e SP |
584 | { |
585 | char *idx_name; | |
586 | int ret; | |
587 | ||
588 | if (p->index_data) | |
589 | return 0; | |
590 | ||
591 | idx_name = xstrdup(p->pack_name); | |
592 | strcpy(idx_name + strlen(idx_name) - strlen(".pack"), ".idx"); | |
593 | ret = check_packed_git_idx(idx_name, p); | |
594 | free(idx_name); | |
595 | return ret; | |
596 | } | |
597 | ||
11daf39b SP |
598 | static void scan_windows(struct packed_git *p, |
599 | struct packed_git **lru_p, | |
600 | struct pack_window **lru_w, | |
601 | struct pack_window **lru_l) | |
1f688557 | 602 | { |
11daf39b SP |
603 | struct pack_window *w, *w_l; |
604 | ||
605 | for (w_l = NULL, w = p->windows; w; w = w->next) { | |
606 | if (!w->inuse_cnt) { | |
607 | if (!*lru_w || w->last_used < (*lru_w)->last_used) { | |
608 | *lru_p = p; | |
609 | *lru_w = w; | |
610 | *lru_l = w_l; | |
54044bf8 | 611 | } |
54044bf8 | 612 | } |
11daf39b | 613 | w_l = w; |
f9253394 | 614 | } |
11daf39b SP |
615 | } |
616 | ||
7c3ecb32 | 617 | static int unuse_one_window(struct packed_git *current) |
11daf39b SP |
618 | { |
619 | struct packed_git *p, *lru_p = NULL; | |
620 | struct pack_window *lru_w = NULL, *lru_l = NULL; | |
621 | ||
622 | if (current) | |
623 | scan_windows(current, &lru_p, &lru_w, &lru_l); | |
624 | for (p = packed_git; p; p = p->next) | |
625 | scan_windows(p, &lru_p, &lru_w, &lru_l); | |
54044bf8 SP |
626 | if (lru_p) { |
627 | munmap(lru_w->base, lru_w->len); | |
628 | pack_mapped -= lru_w->len; | |
629 | if (lru_l) | |
630 | lru_l->next = lru_w->next; | |
7c3ecb32 | 631 | else |
54044bf8 | 632 | lru_p->windows = lru_w->next; |
54044bf8 | 633 | free(lru_w); |
a53128b6 | 634 | pack_open_windows--; |
54044bf8 SP |
635 | return 1; |
636 | } | |
637 | return 0; | |
f9253394 JH |
638 | } |
639 | ||
7c3ecb32 | 640 | void release_pack_memory(size_t need) |
97bfeb34 SP |
641 | { |
642 | size_t cur = pack_mapped; | |
7c3ecb32 | 643 | while (need >= (cur - pack_mapped) && unuse_one_window(NULL)) |
97bfeb34 SP |
644 | ; /* nothing */ |
645 | } | |
646 | ||
58ecbd5e JN |
647 | void *xmmap(void *start, size_t length, |
648 | int prot, int flags, int fd, off_t offset) | |
649 | { | |
650 | void *ret = mmap(start, length, prot, flags, fd, offset); | |
651 | if (ret == MAP_FAILED) { | |
652 | if (!length) | |
653 | return NULL; | |
7c3ecb32 | 654 | release_pack_memory(length); |
58ecbd5e JN |
655 | ret = mmap(start, length, prot, flags, fd, offset); |
656 | if (ret == MAP_FAILED) | |
657 | die_errno("Out of memory? mmap failed"); | |
658 | } | |
659 | return ret; | |
660 | } | |
661 | ||
c9ced051 SP |
662 | void close_pack_windows(struct packed_git *p) |
663 | { | |
664 | while (p->windows) { | |
665 | struct pack_window *w = p->windows; | |
666 | ||
667 | if (w->inuse_cnt) | |
668 | die("pack '%s' still has open windows to it", | |
669 | p->pack_name); | |
670 | munmap(w->base, w->len); | |
671 | pack_mapped -= w->len; | |
672 | pack_open_windows--; | |
673 | p->windows = w->next; | |
674 | free(w); | |
675 | } | |
676 | } | |
677 | ||
88d0db55 BC |
678 | /* |
679 | * The LRU pack is the one with the oldest MRU window, preferring packs | |
680 | * with no used windows, or the oldest mtime if it has no windows allocated. | |
681 | */ | |
682 | static void find_lru_pack(struct packed_git *p, struct packed_git **lru_p, struct pack_window **mru_w, int *accept_windows_inuse) | |
683 | { | |
684 | struct pack_window *w, *this_mru_w; | |
685 | int has_windows_inuse = 0; | |
686 | ||
687 | /* | |
688 | * Reject this pack if it has windows and the previously selected | |
689 | * one does not. If this pack does not have windows, reject | |
690 | * it if the pack file is newer than the previously selected one. | |
691 | */ | |
692 | if (*lru_p && !*mru_w && (p->windows || p->mtime > (*lru_p)->mtime)) | |
693 | return; | |
694 | ||
695 | for (w = this_mru_w = p->windows; w; w = w->next) { | |
696 | /* | |
697 | * Reject this pack if any of its windows are in use, | |
698 | * but the previously selected pack did not have any | |
699 | * inuse windows. Otherwise, record that this pack | |
700 | * has windows in use. | |
701 | */ | |
702 | if (w->inuse_cnt) { | |
703 | if (*accept_windows_inuse) | |
704 | has_windows_inuse = 1; | |
705 | else | |
706 | return; | |
707 | } | |
708 | ||
709 | if (w->last_used > this_mru_w->last_used) | |
710 | this_mru_w = w; | |
711 | ||
712 | /* | |
713 | * Reject this pack if it has windows that have been | |
714 | * used more recently than the previously selected pack. | |
715 | * If the previously selected pack had windows inuse and | |
716 | * we have not encountered a window in this pack that is | |
717 | * inuse, skip this check since we prefer a pack with no | |
718 | * inuse windows to one that has inuse windows. | |
719 | */ | |
720 | if (*mru_w && *accept_windows_inuse == has_windows_inuse && | |
721 | this_mru_w->last_used > (*mru_w)->last_used) | |
722 | return; | |
723 | } | |
724 | ||
725 | /* | |
726 | * Select this pack. | |
727 | */ | |
728 | *mru_w = this_mru_w; | |
729 | *lru_p = p; | |
730 | *accept_windows_inuse = has_windows_inuse; | |
731 | } | |
732 | ||
733 | static int close_one_pack(void) | |
734 | { | |
735 | struct packed_git *p, *lru_p = NULL; | |
736 | struct pack_window *mru_w = NULL; | |
737 | int accept_windows_inuse = 1; | |
738 | ||
739 | for (p = packed_git; p; p = p->next) { | |
740 | if (p->pack_fd == -1) | |
741 | continue; | |
742 | find_lru_pack(p, &lru_p, &mru_w, &accept_windows_inuse); | |
743 | } | |
744 | ||
745 | if (lru_p) { | |
746 | close(lru_p->pack_fd); | |
747 | pack_open_fds--; | |
748 | lru_p->pack_fd = -1; | |
749 | return 1; | |
750 | } | |
751 | ||
752 | return 0; | |
753 | } | |
754 | ||
03e79c88 | 755 | void unuse_pack(struct pack_window **w_cursor) |
f9253394 | 756 | { |
03e79c88 SP |
757 | struct pack_window *w = *w_cursor; |
758 | if (w) { | |
759 | w->inuse_cnt--; | |
760 | *w_cursor = NULL; | |
761 | } | |
1f688557 JH |
762 | } |
763 | ||
fa5fc15d SP |
764 | void close_pack_index(struct packed_git *p) |
765 | { | |
766 | if (p->index_data) { | |
767 | munmap((void *)p->index_data, p->index_size); | |
768 | p->index_data = NULL; | |
769 | } | |
770 | } | |
771 | ||
c74faea1 NP |
772 | /* |
773 | * This is used by git-repack in case a newly created pack happens to | |
774 | * contain the same set of objects as an existing one. In that case | |
775 | * the resulting file might be different even if its name would be the | |
776 | * same. It is best to close any reference to the old pack before it is | |
777 | * replaced on disk. Of course no index pointers nor windows for given pack | |
778 | * must subsist at this point. If ever objects from this pack are requested | |
779 | * again, the new version of the pack will be reinitialized through | |
780 | * reprepare_packed_git(). | |
781 | */ | |
782 | void free_pack_by_name(const char *pack_name) | |
783 | { | |
784 | struct packed_git *p, **pp = &packed_git; | |
785 | ||
786 | while (*pp) { | |
787 | p = *pp; | |
788 | if (strcmp(pack_name, p->pack_name) == 0) { | |
fa3a0c94 | 789 | clear_delta_base_cache(); |
c74faea1 | 790 | close_pack_windows(p); |
c7934306 | 791 | if (p->pack_fd != -1) { |
c74faea1 | 792 | close(p->pack_fd); |
c7934306 SP |
793 | pack_open_fds--; |
794 | } | |
fa5fc15d | 795 | close_pack_index(p); |
c74faea1 NP |
796 | free(p->bad_object_sha1); |
797 | *pp = p->next; | |
c01f51cc NTND |
798 | if (last_found_pack == p) |
799 | last_found_pack = NULL; | |
c74faea1 NP |
800 | free(p); |
801 | return; | |
802 | } | |
803 | pp = &p->next; | |
804 | } | |
805 | } | |
806 | ||
a0788266 JS |
807 | static unsigned int get_max_fd_limit(void) |
808 | { | |
809 | #ifdef RLIMIT_NOFILE | |
491a8dec JH |
810 | { |
811 | struct rlimit lim; | |
a0788266 | 812 | |
491a8dec JH |
813 | if (!getrlimit(RLIMIT_NOFILE, &lim)) |
814 | return lim.rlim_cur; | |
815 | } | |
816 | #endif | |
817 | ||
818 | #ifdef _SC_OPEN_MAX | |
819 | { | |
820 | long open_max = sysconf(_SC_OPEN_MAX); | |
821 | if (0 < open_max) | |
822 | return open_max; | |
823 | /* | |
824 | * Otherwise, we got -1 for one of the two | |
825 | * reasons: | |
826 | * | |
827 | * (1) sysconf() did not understand _SC_OPEN_MAX | |
828 | * and signaled an error with -1; or | |
829 | * (2) sysconf() said there is no limit. | |
830 | * | |
831 | * We _could_ clear errno before calling sysconf() to | |
832 | * tell these two cases apart and return a huge number | |
833 | * in the latter case to let the caller cap it to a | |
834 | * value that is not so selfish, but letting the | |
835 | * fallback OPEN_MAX codepath take care of these cases | |
836 | * is a lot simpler. | |
837 | */ | |
838 | } | |
839 | #endif | |
a0788266 | 840 | |
491a8dec | 841 | #ifdef OPEN_MAX |
a0788266 JS |
842 | return OPEN_MAX; |
843 | #else | |
844 | return 1; /* see the caller ;-) */ | |
845 | #endif | |
846 | } | |
847 | ||
3cf8b462 SP |
848 | /* |
849 | * Do not call this directly as this leaks p->pack_fd on error return; | |
850 | * call open_packed_git() instead. | |
851 | */ | |
852 | static int open_packed_git_1(struct packed_git *p) | |
1f688557 | 853 | { |
9bc879c1 SP |
854 | struct stat st; |
855 | struct pack_header hdr; | |
856 | unsigned char sha1[20]; | |
857 | unsigned char *idx_sha1; | |
2c039da8 | 858 | long fd_flag; |
9bc879c1 | 859 | |
d079837e SP |
860 | if (!p->index_data && open_pack_index(p)) |
861 | return error("packfile %s index unavailable", p->pack_name); | |
862 | ||
c7934306 | 863 | if (!pack_max_fds) { |
a0788266 | 864 | unsigned int max_fds = get_max_fd_limit(); |
c7934306 SP |
865 | |
866 | /* Save 3 for stdin/stdout/stderr, 22 for work */ | |
867 | if (25 < max_fds) | |
868 | pack_max_fds = max_fds - 25; | |
869 | else | |
870 | pack_max_fds = 1; | |
871 | } | |
872 | ||
88d0db55 | 873 | while (pack_max_fds <= pack_open_fds && close_one_pack()) |
c7934306 SP |
874 | ; /* nothing */ |
875 | ||
f4e51683 | 876 | p->pack_fd = git_open_noatime(p->pack_name); |
9bc879c1 | 877 | if (p->pack_fd < 0 || fstat(p->pack_fd, &st)) |
072db278 | 878 | return -1; |
c7934306 | 879 | pack_open_fds++; |
9bc879c1 SP |
880 | |
881 | /* If we created the struct before we had the pack we lack size. */ | |
bf592c50 | 882 | if (!p->pack_size) { |
bf592c50 | 883 | if (!S_ISREG(st.st_mode)) |
072db278 | 884 | return error("packfile %s not a regular file", p->pack_name); |
bf592c50 | 885 | p->pack_size = st.st_size; |
9bc879c1 | 886 | } else if (p->pack_size != st.st_size) |
072db278 | 887 | return error("packfile %s size changed", p->pack_name); |
9bc879c1 | 888 | |
2c039da8 JH |
889 | /* We leave these file descriptors open with sliding mmap; |
890 | * there is no point keeping them open across exec(), though. | |
891 | */ | |
892 | fd_flag = fcntl(p->pack_fd, F_GETFD, 0); | |
893 | if (fd_flag < 0) | |
072db278 | 894 | return error("cannot determine file descriptor flags"); |
2c039da8 JH |
895 | fd_flag |= FD_CLOEXEC; |
896 | if (fcntl(p->pack_fd, F_SETFD, fd_flag) == -1) | |
072db278 | 897 | return error("cannot set FD_CLOEXEC"); |
2c039da8 | 898 | |
9bc879c1 | 899 | /* Verify we recognize this pack file format. */ |
e6e2bd62 | 900 | if (read_in_full(p->pack_fd, &hdr, sizeof(hdr)) != sizeof(hdr)) |
072db278 | 901 | return error("file %s is far too short to be a packfile", p->pack_name); |
9bc879c1 | 902 | if (hdr.hdr_signature != htonl(PACK_SIGNATURE)) |
072db278 | 903 | return error("file %s is not a GIT packfile", p->pack_name); |
9bc879c1 | 904 | if (!pack_version_ok(hdr.hdr_version)) |
6e1c2344 RJ |
905 | return error("packfile %s is version %"PRIu32" and not" |
906 | " supported (try upgrading GIT to a newer version)", | |
9bc879c1 SP |
907 | p->pack_name, ntohl(hdr.hdr_version)); |
908 | ||
909 | /* Verify the pack matches its index. */ | |
57059091 | 910 | if (p->num_objects != ntohl(hdr.hdr_entries)) |
6e1c2344 RJ |
911 | return error("packfile %s claims to have %"PRIu32" objects" |
912 | " while index indicates %"PRIu32" objects", | |
57059091 NP |
913 | p->pack_name, ntohl(hdr.hdr_entries), |
914 | p->num_objects); | |
9bc879c1 | 915 | if (lseek(p->pack_fd, p->pack_size - sizeof(sha1), SEEK_SET) == -1) |
072db278 | 916 | return error("end of packfile %s is unavailable", p->pack_name); |
e6e2bd62 | 917 | if (read_in_full(p->pack_fd, sha1, sizeof(sha1)) != sizeof(sha1)) |
072db278 | 918 | return error("packfile %s signature is unavailable", p->pack_name); |
42873078 | 919 | idx_sha1 = ((unsigned char *)p->index_data) + p->index_size - 40; |
9bc879c1 | 920 | if (hashcmp(sha1, idx_sha1)) |
072db278 SP |
921 | return error("packfile %s does not match index", p->pack_name); |
922 | return 0; | |
9bc879c1 SP |
923 | } |
924 | ||
3cf8b462 SP |
925 | static int open_packed_git(struct packed_git *p) |
926 | { | |
927 | if (!open_packed_git_1(p)) | |
928 | return 0; | |
929 | if (p->pack_fd != -1) { | |
930 | close(p->pack_fd); | |
c7934306 | 931 | pack_open_fds--; |
3cf8b462 SP |
932 | p->pack_fd = -1; |
933 | } | |
934 | return -1; | |
935 | } | |
936 | ||
c4001d92 | 937 | static int in_window(struct pack_window *win, off_t offset) |
60bb8b14 SP |
938 | { |
939 | /* We must promise at least 20 bytes (one hash) after the | |
940 | * offset is available from this window, otherwise the offset | |
941 | * is not actually in this window and a different window (which | |
942 | * has that one hash excess) must be used. This is to support | |
943 | * the object header and delta base parsing routines below. | |
944 | */ | |
945 | off_t win_off = win->offset; | |
946 | return win_off <= offset | |
947 | && (offset + 20) <= (win_off + win->len); | |
948 | } | |
949 | ||
4b25d091 | 950 | unsigned char *use_pack(struct packed_git *p, |
03e79c88 | 951 | struct pack_window **w_cursor, |
c4001d92 | 952 | off_t offset, |
ef49a7a0 | 953 | unsigned long *left) |
9bc879c1 | 954 | { |
60bb8b14 | 955 | struct pack_window *win = *w_cursor; |
03e79c88 | 956 | |
a9d98a14 | 957 | /* Since packfiles end in a hash of their content and it's |
60bb8b14 SP |
958 | * pointless to ask for an offset into the middle of that |
959 | * hash, and the in_window function above wouldn't match | |
960 | * don't allow an offset too close to the end of the file. | |
961 | */ | |
d131b7af SP |
962 | if (!p->pack_size && p->pack_fd == -1 && open_packed_git(p)) |
963 | die("packfile %s cannot be accessed", p->pack_name); | |
60bb8b14 SP |
964 | if (offset > (p->pack_size - 20)) |
965 | die("offset beyond end of packfile (truncated pack?)"); | |
966 | ||
967 | if (!win || !in_window(win, offset)) { | |
968 | if (win) | |
969 | win->inuse_cnt--; | |
970 | for (win = p->windows; win; win = win->next) { | |
971 | if (in_window(win, offset)) | |
972 | break; | |
973 | } | |
974 | if (!win) { | |
78a28df9 | 975 | size_t window_align = packed_git_window_size / 2; |
dc49cd76 | 976 | off_t len; |
d131b7af SP |
977 | |
978 | if (p->pack_fd == -1 && open_packed_git(p)) | |
979 | die("packfile %s cannot be accessed", p->pack_name); | |
980 | ||
60bb8b14 | 981 | win = xcalloc(1, sizeof(*win)); |
78a28df9 | 982 | win->offset = (offset / window_align) * window_align; |
dc49cd76 SP |
983 | len = p->pack_size - win->offset; |
984 | if (len > packed_git_window_size) | |
985 | len = packed_git_window_size; | |
986 | win->len = (size_t)len; | |
60bb8b14 | 987 | pack_mapped += win->len; |
11daf39b | 988 | while (packed_git_limit < pack_mapped |
7c3ecb32 | 989 | && unuse_one_window(p)) |
60bb8b14 | 990 | ; /* nothing */ |
c4712e45 | 991 | win->base = xmmap(NULL, win->len, |
60bb8b14 SP |
992 | PROT_READ, MAP_PRIVATE, |
993 | p->pack_fd, win->offset); | |
994 | if (win->base == MAP_FAILED) | |
73b4e4be SP |
995 | die("packfile %s cannot be mapped: %s", |
996 | p->pack_name, | |
997 | strerror(errno)); | |
d131b7af SP |
998 | if (!win->offset && win->len == p->pack_size |
999 | && !p->do_not_close) { | |
1000 | close(p->pack_fd); | |
1001 | pack_open_fds--; | |
1002 | p->pack_fd = -1; | |
1003 | } | |
a53128b6 SP |
1004 | pack_mmap_calls++; |
1005 | pack_open_windows++; | |
1006 | if (pack_mapped > peak_pack_mapped) | |
1007 | peak_pack_mapped = pack_mapped; | |
1008 | if (pack_open_windows > peak_pack_open_windows) | |
1009 | peak_pack_open_windows = pack_open_windows; | |
60bb8b14 SP |
1010 | win->next = p->windows; |
1011 | p->windows = win; | |
1012 | } | |
1f688557 | 1013 | } |
03e79c88 SP |
1014 | if (win != *w_cursor) { |
1015 | win->last_used = pack_used_ctr++; | |
1016 | win->inuse_cnt++; | |
1017 | *w_cursor = win; | |
1018 | } | |
60bb8b14 | 1019 | offset -= win->offset; |
03e79c88 | 1020 | if (left) |
dc49cd76 | 1021 | *left = win->len - xsize_t(offset); |
03e79c88 | 1022 | return win->base + offset; |
1f688557 JH |
1023 | } |
1024 | ||
27d69a46 NP |
1025 | static struct packed_git *alloc_packed_git(int extra) |
1026 | { | |
1027 | struct packed_git *p = xmalloc(sizeof(*p) + extra); | |
1028 | memset(p, 0, sizeof(*p)); | |
1029 | p->pack_fd = -1; | |
1030 | return p; | |
1031 | } | |
1032 | ||
e0500293 JN |
1033 | static void try_to_free_pack_memory(size_t size) |
1034 | { | |
7c3ecb32 | 1035 | release_pack_memory(size); |
e0500293 JN |
1036 | } |
1037 | ||
42873078 | 1038 | struct packed_git *add_packed_git(const char *path, int path_len, int local) |
1f688557 | 1039 | { |
e0500293 | 1040 | static int have_set_try_to_free_routine; |
1f688557 | 1041 | struct stat st; |
27d69a46 | 1042 | struct packed_git *p = alloc_packed_git(path_len + 2); |
1f688557 | 1043 | |
e0500293 JN |
1044 | if (!have_set_try_to_free_routine) { |
1045 | have_set_try_to_free_routine = 1; | |
1046 | set_try_to_free_routine(try_to_free_pack_memory); | |
1047 | } | |
1048 | ||
42873078 NP |
1049 | /* |
1050 | * Make sure a corresponding .pack file exists and that | |
1051 | * the index looks sane. | |
1052 | */ | |
1053 | path_len -= strlen(".idx"); | |
27d69a46 NP |
1054 | if (path_len < 1) { |
1055 | free(p); | |
1f688557 | 1056 | return NULL; |
27d69a46 | 1057 | } |
42873078 | 1058 | memcpy(p->pack_name, path, path_len); |
8d25931d BC |
1059 | |
1060 | strcpy(p->pack_name + path_len, ".keep"); | |
1061 | if (!access(p->pack_name, F_OK)) | |
1062 | p->pack_keep = 1; | |
1063 | ||
42873078 | 1064 | strcpy(p->pack_name + path_len, ".pack"); |
d079837e | 1065 | if (stat(p->pack_name, &st) || !S_ISREG(st.st_mode)) { |
42873078 | 1066 | free(p); |
1f688557 JH |
1067 | return NULL; |
1068 | } | |
42873078 | 1069 | |
1f688557 JH |
1070 | /* ok, it looks sane as far as we can check without |
1071 | * actually mapping the pack file. | |
1072 | */ | |
1f688557 | 1073 | p->pack_size = st.st_size; |
9d835df2 | 1074 | p->pack_local = local; |
b867092f | 1075 | p->mtime = st.st_mtime; |
42873078 NP |
1076 | if (path_len < 40 || get_sha1_hex(path + path_len - 40, p->sha1)) |
1077 | hashclr(p->sha1); | |
1f688557 JH |
1078 | return p; |
1079 | } | |
1080 | ||
7b64469a | 1081 | struct packed_git *parse_pack_index(unsigned char *sha1, const char *idx_path) |
c508df5e | 1082 | { |
42873078 | 1083 | const char *path = sha1_pack_name(sha1); |
27d69a46 | 1084 | struct packed_git *p = alloc_packed_git(strlen(path) + 1); |
bf592c50 | 1085 | |
27d69a46 NP |
1086 | strcpy(p->pack_name, path); |
1087 | hashcpy(p->sha1, sha1); | |
42873078 NP |
1088 | if (check_packed_git_idx(idx_path, p)) { |
1089 | free(p); | |
bf592c50 | 1090 | return NULL; |
42873078 | 1091 | } |
bf592c50 | 1092 | |
bf592c50 DB |
1093 | return p; |
1094 | } | |
1095 | ||
1096 | void install_packed_git(struct packed_git *pack) | |
1097 | { | |
c7934306 SP |
1098 | if (pack->pack_fd != -1) |
1099 | pack_open_fds++; | |
1100 | ||
bf592c50 DB |
1101 | pack->next = packed_git; |
1102 | packed_git = pack; | |
1103 | } | |
1104 | ||
543c5caa NTND |
1105 | void (*report_garbage)(const char *desc, const char *path); |
1106 | ||
1107 | static void report_helper(const struct string_list *list, | |
1108 | int seen_bits, int first, int last) | |
1109 | { | |
1110 | const char *msg; | |
1111 | switch (seen_bits) { | |
1112 | case 0: | |
1113 | msg = "no corresponding .idx nor .pack"; | |
1114 | break; | |
1115 | case 1: | |
1116 | msg = "no corresponding .idx"; | |
1117 | break; | |
1118 | case 2: | |
1119 | msg = "no corresponding .pack"; | |
1120 | break; | |
1121 | default: | |
1122 | return; | |
1123 | } | |
1124 | for (; first < last; first++) | |
1125 | report_garbage(msg, list->items[first].string); | |
1126 | } | |
1127 | ||
1128 | static void report_pack_garbage(struct string_list *list) | |
1129 | { | |
1130 | int i, baselen = -1, first = 0, seen_bits = 0; | |
1131 | ||
1132 | if (!report_garbage) | |
1133 | return; | |
1134 | ||
1135 | sort_string_list(list); | |
1136 | ||
1137 | for (i = 0; i < list->nr; i++) { | |
1138 | const char *path = list->items[i].string; | |
1139 | if (baselen != -1 && | |
1140 | strncmp(path, list->items[first].string, baselen)) { | |
1141 | report_helper(list, seen_bits, first, i); | |
1142 | baselen = -1; | |
1143 | seen_bits = 0; | |
1144 | } | |
1145 | if (baselen == -1) { | |
1146 | const char *dot = strrchr(path, '.'); | |
1147 | if (!dot) { | |
1148 | report_garbage("garbage found", path); | |
1149 | continue; | |
1150 | } | |
1151 | baselen = dot - path + 1; | |
1152 | first = i; | |
1153 | } | |
1154 | if (!strcmp(path + baselen, "pack")) | |
1155 | seen_bits |= 1; | |
1156 | else if (!strcmp(path + baselen, "idx")) | |
1157 | seen_bits |= 2; | |
1158 | } | |
1159 | report_helper(list, seen_bits, first, list->nr); | |
1160 | } | |
1161 | ||
9d835df2 | 1162 | static void prepare_packed_git_one(char *objdir, int local) |
1f688557 | 1163 | { |
9cb18f56 JM |
1164 | /* Ensure that this buffer is large enough so that we can |
1165 | append "/pack/" without clobbering the stack even if | |
1166 | strlen(objdir) were PATH_MAX. */ | |
1167 | char path[PATH_MAX + 1 + 4 + 1 + 1]; | |
1f688557 JH |
1168 | int len; |
1169 | DIR *dir; | |
1170 | struct dirent *de; | |
543c5caa | 1171 | struct string_list garbage = STRING_LIST_INIT_DUP; |
1f688557 JH |
1172 | |
1173 | sprintf(path, "%s/pack", objdir); | |
1174 | len = strlen(path); | |
1175 | dir = opendir(path); | |
b5b16990 | 1176 | if (!dir) { |
26125f6b | 1177 | if (errno != ENOENT) |
bd2afde8 | 1178 | error("unable to open object pack directory: %s: %s", |
26125f6b | 1179 | path, strerror(errno)); |
1f688557 | 1180 | return; |
b5b16990 | 1181 | } |
1f688557 JH |
1182 | path[len++] = '/'; |
1183 | while ((de = readdir(dir)) != NULL) { | |
1184 | int namelen = strlen(de->d_name); | |
1185 | struct packed_git *p; | |
1186 | ||
543c5caa NTND |
1187 | if (len + namelen + 1 > sizeof(path)) { |
1188 | if (report_garbage) { | |
1189 | struct strbuf sb = STRBUF_INIT; | |
1190 | strbuf_addf(&sb, "%.*s/%s", len - 1, path, de->d_name); | |
1191 | report_garbage("path too long", sb.buf); | |
1192 | strbuf_release(&sb); | |
1193 | } | |
1194 | continue; | |
1195 | } | |
1196 | ||
1197 | if (is_dot_or_dotdot(de->d_name)) | |
9cb18f56 JM |
1198 | continue; |
1199 | ||
1f688557 | 1200 | strcpy(path + len, de->d_name); |
d90906a9 NTND |
1201 | |
1202 | if (has_extension(de->d_name, ".idx")) { | |
1203 | /* Don't reopen a pack we already have. */ | |
1204 | for (p = packed_git; p; p = p->next) { | |
1205 | if (!memcmp(path, p->pack_name, len + namelen - 4)) | |
1206 | break; | |
1207 | } | |
1208 | if (p == NULL && | |
1209 | /* | |
1210 | * See if it really is a valid .idx file with | |
1211 | * corresponding .pack file that we can map. | |
1212 | */ | |
1213 | (p = add_packed_git(path, len + namelen, local)) != NULL) | |
1214 | install_packed_git(p); | |
86f7780c | 1215 | } |
543c5caa NTND |
1216 | |
1217 | if (!report_garbage) | |
1218 | continue; | |
1219 | ||
1220 | if (has_extension(de->d_name, ".idx") || | |
1221 | has_extension(de->d_name, ".pack") || | |
1222 | has_extension(de->d_name, ".keep")) | |
1223 | string_list_append(&garbage, path); | |
1224 | else | |
1225 | report_garbage("garbage found", path); | |
1f688557 | 1226 | } |
5b35bcd5 | 1227 | closedir(dir); |
543c5caa NTND |
1228 | report_pack_garbage(&garbage); |
1229 | string_list_clear(&garbage, 0); | |
1f688557 JH |
1230 | } |
1231 | ||
b867092f JH |
1232 | static int sort_pack(const void *a_, const void *b_) |
1233 | { | |
1234 | struct packed_git *a = *((struct packed_git **)a_); | |
1235 | struct packed_git *b = *((struct packed_git **)b_); | |
1236 | int st; | |
1237 | ||
1238 | /* | |
1239 | * Local packs tend to contain objects specific to our | |
1240 | * variant of the project than remote ones. In addition, | |
1241 | * remote ones could be on a network mounted filesystem. | |
1242 | * Favor local ones for these reasons. | |
1243 | */ | |
1244 | st = a->pack_local - b->pack_local; | |
1245 | if (st) | |
1246 | return -st; | |
1247 | ||
1248 | /* | |
1249 | * Younger packs tend to contain more recent objects, | |
1250 | * and more recent objects tend to get accessed more | |
1251 | * often. | |
1252 | */ | |
1253 | if (a->mtime < b->mtime) | |
1254 | return 1; | |
1255 | else if (a->mtime == b->mtime) | |
1256 | return 0; | |
1257 | return -1; | |
1258 | } | |
1259 | ||
1260 | static void rearrange_packed_git(void) | |
1261 | { | |
1262 | struct packed_git **ary, *p; | |
1263 | int i, n; | |
1264 | ||
1265 | for (n = 0, p = packed_git; p; p = p->next) | |
1266 | n++; | |
1267 | if (n < 2) | |
1268 | return; | |
1269 | ||
1270 | /* prepare an array of packed_git for easier sorting */ | |
1271 | ary = xcalloc(n, sizeof(struct packed_git *)); | |
1272 | for (n = 0, p = packed_git; p; p = p->next) | |
1273 | ary[n++] = p; | |
1274 | ||
1275 | qsort(ary, n, sizeof(struct packed_git *), sort_pack); | |
1276 | ||
1277 | /* link them back again */ | |
1278 | for (i = 0; i < n - 1; i++) | |
1279 | ary[i]->next = ary[i + 1]; | |
1280 | ary[n - 1]->next = NULL; | |
1281 | packed_git = ary[0]; | |
1282 | ||
1283 | free(ary); | |
1284 | } | |
1285 | ||
637cdd9d | 1286 | static int prepare_packed_git_run_once = 0; |
9a217f2a | 1287 | void prepare_packed_git(void) |
1f688557 | 1288 | { |
d5a63b99 | 1289 | struct alternate_object_database *alt; |
1f688557 | 1290 | |
637cdd9d | 1291 | if (prepare_packed_git_run_once) |
1f688557 | 1292 | return; |
9d835df2 | 1293 | prepare_packed_git_one(get_object_directory(), 1); |
9a217f2a | 1294 | prepare_alt_odb(); |
d5a63b99 | 1295 | for (alt = alt_odb_list; alt; alt = alt->next) { |
1494e038 | 1296 | alt->name[-1] = 0; |
9d835df2 | 1297 | prepare_packed_git_one(alt->base, 0); |
1494e038 | 1298 | alt->name[-1] = '/'; |
1f688557 | 1299 | } |
b867092f | 1300 | rearrange_packed_git(); |
637cdd9d JK |
1301 | prepare_packed_git_run_once = 1; |
1302 | } | |
1303 | ||
fc04c412 | 1304 | void reprepare_packed_git(void) |
637cdd9d | 1305 | { |
4b480c67 | 1306 | discard_revindex(); |
637cdd9d JK |
1307 | prepare_packed_git_run_once = 0; |
1308 | prepare_packed_git(); | |
1f688557 JH |
1309 | } |
1310 | ||
8eca0b47 NP |
1311 | static void mark_bad_packed_object(struct packed_git *p, |
1312 | const unsigned char *sha1) | |
1313 | { | |
1314 | unsigned i; | |
1315 | for (i = 0; i < p->num_bad_objects; i++) | |
1316 | if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i)) | |
1317 | return; | |
1318 | p->bad_object_sha1 = xrealloc(p->bad_object_sha1, 20 * (p->num_bad_objects + 1)); | |
1319 | hashcpy(p->bad_object_sha1 + 20 * p->num_bad_objects, sha1); | |
1320 | p->num_bad_objects++; | |
1321 | } | |
1322 | ||
b6c4cecc | 1323 | static const struct packed_git *has_packed_and_bad(const unsigned char *sha1) |
ac939109 NP |
1324 | { |
1325 | struct packed_git *p; | |
1326 | unsigned i; | |
1327 | ||
1328 | for (p = packed_git; p; p = p->next) | |
1329 | for (i = 0; i < p->num_bad_objects; i++) | |
1330 | if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i)) | |
b6c4cecc JH |
1331 | return p; |
1332 | return NULL; | |
ac939109 NP |
1333 | } |
1334 | ||
090ea126 NTND |
1335 | /* |
1336 | * With an in-core object data in "map", rehash it to make sure the | |
1337 | * object name actually matches "sha1" to detect object corruption. | |
1338 | * With "map" == NULL, try reading the object named with "sha1" using | |
1339 | * the streaming interface and rehash it to do the same. | |
1340 | */ | |
1341 | int check_sha1_signature(const unsigned char *sha1, void *map, | |
1342 | unsigned long size, const char *type) | |
0fcfd160 LT |
1343 | { |
1344 | unsigned char real_sha1[20]; | |
090ea126 NTND |
1345 | enum object_type obj_type; |
1346 | struct git_istream *st; | |
1347 | git_SHA_CTX c; | |
1348 | char hdr[32]; | |
1349 | int hdrlen; | |
1350 | ||
1351 | if (map) { | |
1352 | hash_sha1_file(map, size, type, real_sha1); | |
1353 | return hashcmp(sha1, real_sha1) ? -1 : 0; | |
1354 | } | |
1355 | ||
1356 | st = open_istream(sha1, &obj_type, &size, NULL); | |
1357 | if (!st) | |
1358 | return -1; | |
1359 | ||
1360 | /* Generate the header */ | |
1361 | hdrlen = sprintf(hdr, "%s %lu", typename(obj_type), size) + 1; | |
1362 | ||
1363 | /* Sha1.. */ | |
1364 | git_SHA1_Init(&c); | |
1365 | git_SHA1_Update(&c, hdr, hdrlen); | |
1366 | for (;;) { | |
1367 | char buf[1024 * 16]; | |
1368 | ssize_t readlen = read_istream(st, buf, sizeof(buf)); | |
1369 | ||
f54fac53 JK |
1370 | if (readlen < 0) { |
1371 | close_istream(st); | |
1372 | return -1; | |
1373 | } | |
090ea126 NTND |
1374 | if (!readlen) |
1375 | break; | |
1376 | git_SHA1_Update(&c, buf, readlen); | |
1377 | } | |
1378 | git_SHA1_Final(real_sha1, &c); | |
1379 | close_istream(st); | |
a89fccd2 | 1380 | return hashcmp(sha1, real_sha1) ? -1 : 0; |
0fcfd160 LT |
1381 | } |
1382 | ||
f4e51683 | 1383 | static int git_open_noatime(const char *name) |
44d1c19e LT |
1384 | { |
1385 | static int sha1_file_open_flag = O_NOATIME; | |
44d1c19e | 1386 | |
f2e872aa SP |
1387 | for (;;) { |
1388 | int fd = open(name, O_RDONLY | sha1_file_open_flag); | |
44d1c19e | 1389 | if (fd >= 0) |
f2e872aa SP |
1390 | return fd; |
1391 | ||
f2e872aa SP |
1392 | /* Might the failure be due to O_NOATIME? */ |
1393 | if (errno != ENOENT && sha1_file_open_flag) { | |
44d1c19e | 1394 | sha1_file_open_flag = 0; |
f2e872aa SP |
1395 | continue; |
1396 | } | |
1397 | ||
1398 | return -1; | |
44d1c19e | 1399 | } |
44d1c19e LT |
1400 | } |
1401 | ||
052fe5ea JK |
1402 | static int stat_sha1_file(const unsigned char *sha1, struct stat *st) |
1403 | { | |
1404 | char *name = sha1_file_name(sha1); | |
1405 | struct alternate_object_database *alt; | |
1406 | ||
1407 | if (!lstat(name, st)) | |
1408 | return 0; | |
1409 | ||
1410 | prepare_alt_odb(); | |
1411 | errno = ENOENT; | |
1412 | for (alt = alt_odb_list; alt; alt = alt->next) { | |
1413 | name = alt->name; | |
1414 | fill_sha1_path(name, sha1); | |
1415 | if (!lstat(alt->base, st)) | |
1416 | return 0; | |
1417 | } | |
1418 | ||
1419 | return -1; | |
1420 | } | |
1421 | ||
44d1c19e LT |
1422 | static int open_sha1_file(const unsigned char *sha1) |
1423 | { | |
1424 | int fd; | |
1425 | char *name = sha1_file_name(sha1); | |
1426 | struct alternate_object_database *alt; | |
1427 | ||
f4e51683 | 1428 | fd = git_open_noatime(name); |
44d1c19e LT |
1429 | if (fd >= 0) |
1430 | return fd; | |
1431 | ||
1432 | prepare_alt_odb(); | |
1433 | errno = ENOENT; | |
1434 | for (alt = alt_odb_list; alt; alt = alt->next) { | |
1435 | name = alt->name; | |
1436 | fill_sha1_path(name, sha1); | |
f4e51683 | 1437 | fd = git_open_noatime(alt->base); |
44d1c19e LT |
1438 | if (fd >= 0) |
1439 | return fd; | |
1440 | } | |
1441 | return -1; | |
1442 | } | |
1443 | ||
f0270efd | 1444 | void *map_sha1_file(const unsigned char *sha1, unsigned long *size) |
0fcfd160 | 1445 | { |
0fcfd160 | 1446 | void *map; |
144bde78 | 1447 | int fd; |
ace1534d | 1448 | |
44d1c19e LT |
1449 | fd = open_sha1_file(sha1); |
1450 | map = NULL; | |
1451 | if (fd >= 0) { | |
1452 | struct stat st; | |
0fcfd160 | 1453 | |
44d1c19e LT |
1454 | if (!fstat(fd, &st)) { |
1455 | *size = xsize_t(st.st_size); | |
33e42de0 MM |
1456 | if (!*size) { |
1457 | /* mmap() is forbidden on empty files */ | |
1458 | error("object file %s is empty", sha1_file_name(sha1)); | |
1459 | return NULL; | |
1460 | } | |
44d1c19e | 1461 | map = xmmap(NULL, *size, PROT_READ, MAP_PRIVATE, fd, 0); |
144bde78 | 1462 | } |
44d1c19e | 1463 | close(fd); |
0fcfd160 | 1464 | } |
0fcfd160 LT |
1465 | return map; |
1466 | } | |
1467 | ||
cc5c54e7 JH |
1468 | /* |
1469 | * There used to be a second loose object header format which | |
1470 | * was meant to mimic the in-pack format, allowing for direct | |
1471 | * copy of the object data. This format turned up not to be | |
1472 | * really worth it and we no longer write loose objects in that | |
1473 | * format. | |
1474 | */ | |
1475 | static int experimental_loose_object(unsigned char *map) | |
bb6b8e4f JH |
1476 | { |
1477 | unsigned int word; | |
1478 | ||
1479 | /* | |
7f684a2a RT |
1480 | * We must determine if the buffer contains the standard |
1481 | * zlib-deflated stream or the experimental format based | |
1482 | * on the in-pack object format. Compare the header byte | |
1483 | * for each format: | |
1484 | * | |
1485 | * RFC1950 zlib w/ deflate : 0www1000 : 0 <= www <= 7 | |
1486 | * Experimental pack-based : Stttssss : ttt = 1,2,3,4 | |
1487 | * | |
1488 | * If bit 7 is clear and bits 0-3 equal 8, the buffer MUST be | |
1489 | * in standard loose-object format, UNLESS it is a Git-pack | |
1490 | * format object *exactly* 8 bytes in size when inflated. | |
1491 | * | |
1492 | * However, RFC1950 also specifies that the 1st 16-bit word | |
1493 | * must be divisible by 31 - this checksum tells us our buffer | |
1494 | * is in the standard format, giving a false positive only if | |
1495 | * the 1st word of the Git-pack format object happens to be | |
1496 | * divisible by 31, ie: | |
1497 | * ((byte0 * 256) + byte1) % 31 = 0 | |
1498 | * => 0ttt10000www1000 % 31 = 0 | |
1499 | * | |
1500 | * As it happens, this case can only arise for www=3 & ttt=1 | |
1501 | * - ie, a Commit object, which would have to be 8 bytes in | |
1502 | * size. As no Commit can be that small, we find that the | |
1503 | * combination of these two criteria (bitmask & checksum) | |
1504 | * can always correctly determine the buffer format. | |
bb6b8e4f JH |
1505 | */ |
1506 | word = (map[0] << 8) + map[1]; | |
7f684a2a | 1507 | if ((map[0] & 0x8F) == 0x08 && !(word % 31)) |
bb6b8e4f | 1508 | return 0; |
cc5c54e7 JH |
1509 | else |
1510 | return 1; | |
bb6b8e4f JH |
1511 | } |
1512 | ||
09ded04b NP |
1513 | unsigned long unpack_object_header_buffer(const unsigned char *buf, |
1514 | unsigned long len, enum object_type *type, unsigned long *sizep) | |
c4483576 | 1515 | { |
ad1ed5ee | 1516 | unsigned shift; |
48fb7deb | 1517 | unsigned long size, c; |
ad1ed5ee JH |
1518 | unsigned long used = 0; |
1519 | ||
1520 | c = buf[used++]; | |
1521 | *type = (c >> 4) & 7; | |
1522 | size = c & 15; | |
1523 | shift = 4; | |
1524 | while (c & 0x80) { | |
f630cfda | 1525 | if (len <= used || bitsizeof(long) <= shift) { |
09ded04b | 1526 | error("bad object header"); |
ea4f9685 JH |
1527 | size = used = 0; |
1528 | break; | |
09ded04b | 1529 | } |
ad1ed5ee JH |
1530 | c = buf[used++]; |
1531 | size += (c & 0x7f) << shift; | |
1532 | shift += 7; | |
1533 | } | |
1534 | *sizep = size; | |
1535 | return used; | |
1536 | } | |
1537 | ||
eb4f4076 | 1538 | int unpack_sha1_header(git_zstream *stream, unsigned char *map, unsigned long mapsize, void *buffer, unsigned long bufsiz) |
ad1ed5ee JH |
1539 | { |
1540 | unsigned long size, used; | |
1541 | static const char valid_loose_object_type[8] = { | |
1542 | 0, /* OBJ_EXT */ | |
1543 | 1, 1, 1, 1, /* "commit", "tree", "blob", "tag" */ | |
1544 | 0, /* "delta" and others are invalid in a loose object */ | |
93821bd9 | 1545 | }; |
ad1ed5ee | 1546 | enum object_type type; |
93821bd9 | 1547 | |
c4483576 LT |
1548 | /* Get the data stream */ |
1549 | memset(stream, 0, sizeof(*stream)); | |
1550 | stream->next_in = map; | |
1551 | stream->avail_in = mapsize; | |
1552 | stream->next_out = buffer; | |
93821bd9 LT |
1553 | stream->avail_out = bufsiz; |
1554 | ||
cc5c54e7 JH |
1555 | if (experimental_loose_object(map)) { |
1556 | /* | |
1557 | * The old experimental format we no longer produce; | |
1558 | * we can still read it. | |
1559 | */ | |
1560 | used = unpack_object_header_buffer(map, mapsize, &type, &size); | |
1561 | if (!used || !valid_loose_object_type[type]) | |
1562 | return -1; | |
1563 | map += used; | |
1564 | mapsize -= used; | |
726f852b | 1565 | |
cc5c54e7 JH |
1566 | /* Set up the stream for the rest.. */ |
1567 | stream->next_in = map; | |
1568 | stream->avail_in = mapsize; | |
1569 | git_inflate_init(stream); | |
93821bd9 | 1570 | |
cc5c54e7 JH |
1571 | /* And generate the fake traditional header */ |
1572 | stream->total_out = 1 + snprintf(buffer, bufsiz, "%s %lu", | |
1573 | typename(type), size); | |
1574 | return 0; | |
1575 | } | |
39c68542 | 1576 | git_inflate_init(stream); |
cc5c54e7 | 1577 | return git_inflate(stream, 0); |
c4483576 LT |
1578 | } |
1579 | ||
ef49a7a0 | 1580 | static void *unpack_sha1_rest(git_zstream *stream, void *buffer, unsigned long size, const unsigned char *sha1) |
5180cacc LT |
1581 | { |
1582 | int bytes = strlen(buffer) + 1; | |
3aee68aa | 1583 | unsigned char *buf = xmallocz(size); |
93821bd9 | 1584 | unsigned long n; |
7efbff75 | 1585 | int status = Z_OK; |
5180cacc | 1586 | |
93821bd9 LT |
1587 | n = stream->total_out - bytes; |
1588 | if (n > size) | |
1589 | n = size; | |
1590 | memcpy(buf, (char *) buffer + bytes, n); | |
1591 | bytes = n; | |
456cdf6e LT |
1592 | if (bytes <= size) { |
1593 | /* | |
1594 | * The above condition must be (bytes <= size), not | |
1595 | * (bytes < size). In other words, even though we | |
ccf5ace0 | 1596 | * expect no more output and set avail_out to zero, |
456cdf6e LT |
1597 | * the input zlib stream may have bytes that express |
1598 | * "this concludes the stream", and we *do* want to | |
1599 | * eat that input. | |
1600 | * | |
1601 | * Otherwise we would not be able to test that we | |
1602 | * consumed all the input to reach the expected size; | |
1603 | * we also want to check that zlib tells us that all | |
1604 | * went well with status == Z_STREAM_END at the end. | |
1605 | */ | |
5180cacc LT |
1606 | stream->next_out = buf + bytes; |
1607 | stream->avail_out = size - bytes; | |
7efbff75 | 1608 | while (status == Z_OK) |
39c68542 | 1609 | status = git_inflate(stream, Z_FINISH); |
5180cacc | 1610 | } |
456cdf6e | 1611 | if (status == Z_STREAM_END && !stream->avail_in) { |
39c68542 | 1612 | git_inflate_end(stream); |
7efbff75 JH |
1613 | return buf; |
1614 | } | |
1615 | ||
1616 | if (status < 0) | |
1617 | error("corrupt loose object '%s'", sha1_to_hex(sha1)); | |
1618 | else if (stream->avail_in) | |
1619 | error("garbage at end of loose object '%s'", | |
1620 | sha1_to_hex(sha1)); | |
1621 | free(buf); | |
1622 | return NULL; | |
5180cacc LT |
1623 | } |
1624 | ||
1625 | /* | |
1626 | * We used to just use "sscanf()", but that's actually way | |
1627 | * too permissive for what we want to check. So do an anal | |
1628 | * object header parse by hand. | |
1629 | */ | |
f0270efd | 1630 | int parse_sha1_header(const char *hdr, unsigned long *sizep) |
5180cacc | 1631 | { |
21666f1a | 1632 | char type[10]; |
5180cacc LT |
1633 | int i; |
1634 | unsigned long size; | |
1635 | ||
1636 | /* | |
a6080a0a | 1637 | * The type can be at most ten bytes (including the |
5180cacc | 1638 | * terminating '\0' that we add), and is followed by |
21666f1a | 1639 | * a space. |
5180cacc | 1640 | */ |
21666f1a | 1641 | i = 0; |
5180cacc LT |
1642 | for (;;) { |
1643 | char c = *hdr++; | |
1644 | if (c == ' ') | |
1645 | break; | |
21666f1a NP |
1646 | type[i++] = c; |
1647 | if (i >= sizeof(type)) | |
5180cacc | 1648 | return -1; |
5180cacc | 1649 | } |
21666f1a | 1650 | type[i] = 0; |
5180cacc LT |
1651 | |
1652 | /* | |
1653 | * The length must follow immediately, and be in canonical | |
1654 | * decimal format (ie "010" is not valid). | |
1655 | */ | |
1656 | size = *hdr++ - '0'; | |
1657 | if (size > 9) | |
1658 | return -1; | |
1659 | if (size) { | |
1660 | for (;;) { | |
1661 | unsigned long c = *hdr - '0'; | |
1662 | if (c > 9) | |
1663 | break; | |
1664 | hdr++; | |
1665 | size = size * 10 + c; | |
1666 | } | |
1667 | } | |
1668 | *sizep = size; | |
1669 | ||
1670 | /* | |
1671 | * The length must be followed by a zero byte | |
1672 | */ | |
21666f1a | 1673 | return *hdr ? -1 : type_from_string(type); |
5180cacc LT |
1674 | } |
1675 | ||
7efbff75 | 1676 | static void *unpack_sha1_file(void *map, unsigned long mapsize, enum object_type *type, unsigned long *size, const unsigned char *sha1) |
0fcfd160 | 1677 | { |
5180cacc | 1678 | int ret; |
ef49a7a0 | 1679 | git_zstream stream; |
5180cacc | 1680 | char hdr[8192]; |
0fcfd160 | 1681 | |
5180cacc | 1682 | ret = unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)); |
21666f1a | 1683 | if (ret < Z_OK || (*type = parse_sha1_header(hdr, size)) < 0) |
0fcfd160 LT |
1684 | return NULL; |
1685 | ||
7efbff75 | 1686 | return unpack_sha1_rest(&stream, hdr, *size, sha1); |
0fcfd160 LT |
1687 | } |
1688 | ||
54dab52a NP |
1689 | unsigned long get_size_from_delta(struct packed_git *p, |
1690 | struct pack_window **w_curs, | |
1691 | off_t curpos) | |
1692 | { | |
1693 | const unsigned char *data; | |
1694 | unsigned char delta_head[20], *in; | |
ef49a7a0 | 1695 | git_zstream stream; |
54dab52a NP |
1696 | int st; |
1697 | ||
1698 | memset(&stream, 0, sizeof(stream)); | |
1699 | stream.next_out = delta_head; | |
1700 | stream.avail_out = sizeof(delta_head); | |
1701 | ||
39c68542 | 1702 | git_inflate_init(&stream); |
54dab52a NP |
1703 | do { |
1704 | in = use_pack(p, w_curs, curpos, &stream.avail_in); | |
1705 | stream.next_in = in; | |
39c68542 | 1706 | st = git_inflate(&stream, Z_FINISH); |
54dab52a NP |
1707 | curpos += stream.next_in - in; |
1708 | } while ((st == Z_OK || st == Z_BUF_ERROR) && | |
1709 | stream.total_out < sizeof(delta_head)); | |
39c68542 | 1710 | git_inflate_end(&stream); |
3d77d877 NP |
1711 | if ((st != Z_STREAM_END) && stream.total_out != sizeof(delta_head)) { |
1712 | error("delta data unpack-initial failed"); | |
1713 | return 0; | |
1714 | } | |
54dab52a NP |
1715 | |
1716 | /* Examine the initial part of the delta to figure out | |
1717 | * the result size. | |
1718 | */ | |
1719 | data = delta_head; | |
1720 | ||
1721 | /* ignore base size */ | |
1722 | get_delta_hdr_size(&data, delta_head+sizeof(delta_head)); | |
1723 | ||
1724 | /* Read the result size */ | |
1725 | return get_delta_hdr_size(&data, delta_head+sizeof(delta_head)); | |
1726 | } | |
1727 | ||
c4001d92 | 1728 | static off_t get_delta_base(struct packed_git *p, |
03e79c88 | 1729 | struct pack_window **w_curs, |
c4001d92 | 1730 | off_t *curpos, |
21666f1a | 1731 | enum object_type type, |
c4001d92 | 1732 | off_t delta_obj_offset) |
eb32d236 | 1733 | { |
2b87c45b | 1734 | unsigned char *base_info = use_pack(p, w_curs, *curpos, NULL); |
c4001d92 | 1735 | off_t base_offset; |
eb32d236 | 1736 | |
8d8a4ea5 SP |
1737 | /* use_pack() assured us we have [base_info, base_info + 20) |
1738 | * as a range that we can look at without walking off the | |
1739 | * end of the mapped window. Its actually the hash size | |
1740 | * that is assured. An OFS_DELTA longer than the hash size | |
1741 | * is stupid, as then a REF_DELTA would be smaller to store. | |
1742 | */ | |
21666f1a | 1743 | if (type == OBJ_OFS_DELTA) { |
eb32d236 NP |
1744 | unsigned used = 0; |
1745 | unsigned char c = base_info[used++]; | |
1746 | base_offset = c & 127; | |
1747 | while (c & 128) { | |
1748 | base_offset += 1; | |
8723f216 | 1749 | if (!base_offset || MSB(base_offset, 7)) |
8eca0b47 | 1750 | return 0; /* overflow */ |
eb32d236 NP |
1751 | c = base_info[used++]; |
1752 | base_offset = (base_offset << 7) + (c & 127); | |
1753 | } | |
1754 | base_offset = delta_obj_offset - base_offset; | |
d8f32556 | 1755 | if (base_offset <= 0 || base_offset >= delta_obj_offset) |
8eca0b47 | 1756 | return 0; /* out of bound */ |
2b87c45b | 1757 | *curpos += used; |
21666f1a | 1758 | } else if (type == OBJ_REF_DELTA) { |
eb32d236 NP |
1759 | /* The base entry _must_ be in the same pack */ |
1760 | base_offset = find_pack_entry_one(base_info, p); | |
2b87c45b | 1761 | *curpos += 20; |
eb32d236 NP |
1762 | } else |
1763 | die("I am totally screwed"); | |
2b87c45b | 1764 | return base_offset; |
eb32d236 NP |
1765 | } |
1766 | ||
f8c8abc5 JH |
1767 | int unpack_object_header(struct packed_git *p, |
1768 | struct pack_window **w_curs, | |
1769 | off_t *curpos, | |
1770 | unsigned long *sizep) | |
a733cb60 | 1771 | { |
03e79c88 | 1772 | unsigned char *base; |
ef49a7a0 | 1773 | unsigned long left; |
ad1ed5ee | 1774 | unsigned long used; |
2b87c45b | 1775 | enum object_type type; |
a733cb60 | 1776 | |
8d8a4ea5 | 1777 | /* use_pack() assures us we have [base, base + 20) available |
0353a0c4 | 1778 | * as a range that we can look at. (Its actually the hash |
3dff5379 | 1779 | * size that is assured.) With our object header encoding |
8d8a4ea5 SP |
1780 | * the maximum deflated object size is 2^137, which is just |
1781 | * insane, so we know won't exceed what we have been given. | |
1782 | */ | |
2b87c45b | 1783 | base = use_pack(p, w_curs, *curpos, &left); |
09ded04b NP |
1784 | used = unpack_object_header_buffer(base, left, &type, sizep); |
1785 | if (!used) { | |
1786 | type = OBJ_BAD; | |
1787 | } else | |
1788 | *curpos += used; | |
ad1ed5ee | 1789 | |
2b87c45b | 1790 | return type; |
a733cb60 LT |
1791 | } |
1792 | ||
790d96c0 TR |
1793 | static int retry_bad_packed_offset(struct packed_git *p, off_t obj_offset) |
1794 | { | |
1795 | int type; | |
1796 | struct revindex_entry *revidx; | |
1797 | const unsigned char *sha1; | |
1798 | revidx = find_pack_revindex(p, obj_offset); | |
1799 | if (!revidx) | |
1800 | return OBJ_BAD; | |
1801 | sha1 = nth_packed_object_sha1(p, revidx->nr); | |
1802 | mark_bad_packed_object(p, sha1); | |
1803 | type = sha1_object_info(sha1, NULL); | |
1804 | if (type <= OBJ_NONE) | |
1805 | return OBJ_BAD; | |
1806 | return type; | |
1807 | } | |
1808 | ||
790d96c0 TR |
1809 | #define POI_STACK_PREALLOC 64 |
1810 | ||
90191d37 JK |
1811 | static enum object_type packed_to_object_type(struct packed_git *p, |
1812 | off_t obj_offset, | |
1813 | enum object_type type, | |
1814 | struct pack_window **w_curs, | |
1815 | off_t curpos) | |
1f688557 | 1816 | { |
790d96c0 TR |
1817 | off_t small_poi_stack[POI_STACK_PREALLOC]; |
1818 | off_t *poi_stack = small_poi_stack; | |
1819 | int poi_stack_nr = 0, poi_stack_alloc = POI_STACK_PREALLOC; | |
5db47c2b | 1820 | |
790d96c0 TR |
1821 | while (type == OBJ_OFS_DELTA || type == OBJ_REF_DELTA) { |
1822 | off_t base_offset; | |
90191d37 | 1823 | unsigned long size; |
790d96c0 TR |
1824 | /* Push the object we're going to leave behind */ |
1825 | if (poi_stack_nr >= poi_stack_alloc && poi_stack == small_poi_stack) { | |
1826 | poi_stack_alloc = alloc_nr(poi_stack_nr); | |
1827 | poi_stack = xmalloc(sizeof(off_t)*poi_stack_alloc); | |
1828 | memcpy(poi_stack, small_poi_stack, sizeof(off_t)*poi_stack_nr); | |
1829 | } else { | |
1830 | ALLOC_GROW(poi_stack, poi_stack_nr+1, poi_stack_alloc); | |
1831 | } | |
1832 | poi_stack[poi_stack_nr++] = obj_offset; | |
1833 | /* If parsing the base offset fails, just unwind */ | |
90191d37 | 1834 | base_offset = get_delta_base(p, w_curs, &curpos, type, obj_offset); |
790d96c0 TR |
1835 | if (!base_offset) |
1836 | goto unwind; | |
1837 | curpos = obj_offset = base_offset; | |
90191d37 | 1838 | type = unpack_object_header(p, w_curs, &curpos, &size); |
790d96c0 TR |
1839 | if (type <= OBJ_NONE) { |
1840 | /* If getting the base itself fails, we first | |
1841 | * retry the base, otherwise unwind */ | |
1842 | type = retry_bad_packed_offset(p, base_offset); | |
1843 | if (type > OBJ_NONE) | |
1844 | goto out; | |
1845 | goto unwind; | |
1846 | } | |
1847 | } | |
1848 | ||
21666f1a | 1849 | switch (type) { |
790d96c0 | 1850 | case OBJ_BAD: |
a733cb60 | 1851 | case OBJ_COMMIT: |
a733cb60 | 1852 | case OBJ_TREE: |
a733cb60 | 1853 | case OBJ_BLOB: |
a733cb60 | 1854 | case OBJ_TAG: |
a69d0943 | 1855 | break; |
1f688557 | 1856 | default: |
3d77d877 NP |
1857 | error("unknown object type %i at offset %"PRIuMAX" in %s", |
1858 | type, (uintmax_t)obj_offset, p->pack_name); | |
1859 | type = OBJ_BAD; | |
1f688557 | 1860 | } |
790d96c0 TR |
1861 | |
1862 | out: | |
1863 | if (poi_stack != small_poi_stack) | |
1864 | free(poi_stack); | |
21666f1a | 1865 | return type; |
790d96c0 TR |
1866 | |
1867 | unwind: | |
1868 | while (poi_stack_nr) { | |
1869 | obj_offset = poi_stack[--poi_stack_nr]; | |
1870 | type = retry_bad_packed_offset(p, obj_offset); | |
1871 | if (type > OBJ_NONE) | |
1872 | goto out; | |
1873 | } | |
1874 | type = OBJ_BAD; | |
1875 | goto out; | |
1f688557 JH |
1876 | } |
1877 | ||
90191d37 | 1878 | static int packed_object_info(struct packed_git *p, off_t obj_offset, |
23c339c0 | 1879 | struct object_info *oi) |
90191d37 JK |
1880 | { |
1881 | struct pack_window *w_curs = NULL; | |
1882 | unsigned long size; | |
1883 | off_t curpos = obj_offset; | |
1884 | enum object_type type; | |
1885 | ||
412916ee JK |
1886 | /* |
1887 | * We always get the representation type, but only convert it to | |
1888 | * a "real" type later if the caller is interested. | |
1889 | */ | |
90191d37 JK |
1890 | type = unpack_object_header(p, &w_curs, &curpos, &size); |
1891 | ||
23c339c0 | 1892 | if (oi->sizep) { |
90191d37 JK |
1893 | if (type == OBJ_OFS_DELTA || type == OBJ_REF_DELTA) { |
1894 | off_t tmp_pos = curpos; | |
1895 | off_t base_offset = get_delta_base(p, &w_curs, &tmp_pos, | |
1896 | type, obj_offset); | |
1897 | if (!base_offset) { | |
1898 | type = OBJ_BAD; | |
1899 | goto out; | |
1900 | } | |
23c339c0 JK |
1901 | *oi->sizep = get_size_from_delta(p, &w_curs, tmp_pos); |
1902 | if (*oi->sizep == 0) { | |
90191d37 JK |
1903 | type = OBJ_BAD; |
1904 | goto out; | |
1905 | } | |
1906 | } else { | |
23c339c0 | 1907 | *oi->sizep = size; |
90191d37 JK |
1908 | } |
1909 | } | |
1910 | ||
23c339c0 | 1911 | if (oi->disk_sizep) { |
90191d37 | 1912 | struct revindex_entry *revidx = find_pack_revindex(p, obj_offset); |
23c339c0 | 1913 | *oi->disk_sizep = revidx[1].offset - obj_offset; |
90191d37 JK |
1914 | } |
1915 | ||
23c339c0 JK |
1916 | if (oi->typep) { |
1917 | *oi->typep = packed_to_object_type(p, obj_offset, type, &w_curs, curpos); | |
1918 | if (*oi->typep < 0) { | |
412916ee JK |
1919 | type = OBJ_BAD; |
1920 | goto out; | |
1921 | } | |
1922 | } | |
90191d37 JK |
1923 | |
1924 | out: | |
1925 | unuse_pack(&w_curs); | |
1926 | return type; | |
1927 | } | |
1928 | ||
eb950c19 | 1929 | static void *unpack_compressed_entry(struct packed_git *p, |
03e79c88 | 1930 | struct pack_window **w_curs, |
c4001d92 | 1931 | off_t curpos, |
eb950c19 | 1932 | unsigned long size) |
de530aaa SP |
1933 | { |
1934 | int st; | |
ef49a7a0 | 1935 | git_zstream stream; |
079afb18 | 1936 | unsigned char *buffer, *in; |
de530aaa | 1937 | |
4ab07e4d | 1938 | buffer = xmallocz(size); |
de530aaa | 1939 | memset(&stream, 0, sizeof(stream)); |
de530aaa | 1940 | stream.next_out = buffer; |
39eea7bd | 1941 | stream.avail_out = size + 1; |
de530aaa | 1942 | |
39c68542 | 1943 | git_inflate_init(&stream); |
079afb18 | 1944 | do { |
2b87c45b | 1945 | in = use_pack(p, w_curs, curpos, &stream.avail_in); |
079afb18 | 1946 | stream.next_in = in; |
39c68542 | 1947 | st = git_inflate(&stream, Z_FINISH); |
39eea7bd JH |
1948 | if (!stream.avail_out) |
1949 | break; /* the payload is larger than it should be */ | |
2b87c45b | 1950 | curpos += stream.next_in - in; |
079afb18 | 1951 | } while (st == Z_OK || st == Z_BUF_ERROR); |
39c68542 | 1952 | git_inflate_end(&stream); |
de530aaa SP |
1953 | if ((st != Z_STREAM_END) || stream.total_out != size) { |
1954 | free(buffer); | |
1955 | return NULL; | |
1956 | } | |
1957 | ||
1958 | return buffer; | |
1959 | } | |
1960 | ||
e5e01619 LT |
1961 | #define MAX_DELTA_CACHE (256) |
1962 | ||
18bdec11 | 1963 | static size_t delta_base_cached; |
5e08ecbf NP |
1964 | |
1965 | static struct delta_base_cache_lru_list { | |
1966 | struct delta_base_cache_lru_list *prev; | |
1967 | struct delta_base_cache_lru_list *next; | |
1968 | } delta_base_cache_lru = { &delta_base_cache_lru, &delta_base_cache_lru }; | |
1969 | ||
e5e01619 | 1970 | static struct delta_base_cache_entry { |
5e08ecbf NP |
1971 | struct delta_base_cache_lru_list lru; |
1972 | void *data; | |
e5e01619 LT |
1973 | struct packed_git *p; |
1974 | off_t base_offset; | |
1975 | unsigned long size; | |
e5e01619 LT |
1976 | enum object_type type; |
1977 | } delta_base_cache[MAX_DELTA_CACHE]; | |
1978 | ||
1979 | static unsigned long pack_entry_hash(struct packed_git *p, off_t base_offset) | |
1980 | { | |
1981 | unsigned long hash; | |
1982 | ||
1983 | hash = (unsigned long)p + (unsigned long)base_offset; | |
1984 | hash += (hash >> 8) + (hash >> 16); | |
3358004a | 1985 | return hash % MAX_DELTA_CACHE; |
e5e01619 LT |
1986 | } |
1987 | ||
84dd81c1 TR |
1988 | static struct delta_base_cache_entry * |
1989 | get_delta_base_cache_entry(struct packed_git *p, off_t base_offset) | |
5266d369 JH |
1990 | { |
1991 | unsigned long hash = pack_entry_hash(p, base_offset); | |
84dd81c1 TR |
1992 | return delta_base_cache + hash; |
1993 | } | |
1994 | ||
1995 | static int eq_delta_base_cache_entry(struct delta_base_cache_entry *ent, | |
1996 | struct packed_git *p, off_t base_offset) | |
1997 | { | |
5266d369 JH |
1998 | return (ent->data && ent->p == p && ent->base_offset == base_offset); |
1999 | } | |
2000 | ||
84dd81c1 TR |
2001 | static int in_delta_base_cache(struct packed_git *p, off_t base_offset) |
2002 | { | |
2003 | struct delta_base_cache_entry *ent; | |
2004 | ent = get_delta_base_cache_entry(p, base_offset); | |
2005 | return eq_delta_base_cache_entry(ent, p, base_offset); | |
2006 | } | |
2007 | ||
2008 | static void clear_delta_base_cache_entry(struct delta_base_cache_entry *ent) | |
2009 | { | |
2010 | ent->data = NULL; | |
2011 | ent->lru.next->prev = ent->lru.prev; | |
2012 | ent->lru.prev->next = ent->lru.next; | |
2013 | delta_base_cached -= ent->size; | |
2014 | } | |
2015 | ||
62f255ad | 2016 | static void *cache_or_unpack_entry(struct packed_git *p, off_t base_offset, |
a0cba108 | 2017 | unsigned long *base_size, enum object_type *type, int keep_cache) |
62f255ad | 2018 | { |
84dd81c1 | 2019 | struct delta_base_cache_entry *ent; |
e5e01619 | 2020 | void *ret; |
e5e01619 | 2021 | |
84dd81c1 TR |
2022 | ent = get_delta_base_cache_entry(p, base_offset); |
2023 | ||
2024 | if (!eq_delta_base_cache_entry(ent, p, base_offset)) | |
749bc58c | 2025 | return unpack_entry(p, base_offset, type, base_size); |
e5e01619 | 2026 | |
84dd81c1 TR |
2027 | ret = ent->data; |
2028 | ||
2029 | if (!keep_cache) | |
2030 | clear_delta_base_cache_entry(ent); | |
2031 | else | |
182af834 | 2032 | ret = xmemdupz(ent->data, ent->size); |
e5e01619 LT |
2033 | *type = ent->type; |
2034 | *base_size = ent->size; | |
2035 | return ret; | |
62f255ad LT |
2036 | } |
2037 | ||
18bdec11 SP |
2038 | static inline void release_delta_base_cache(struct delta_base_cache_entry *ent) |
2039 | { | |
2040 | if (ent->data) { | |
2041 | free(ent->data); | |
2042 | ent->data = NULL; | |
5e08ecbf NP |
2043 | ent->lru.next->prev = ent->lru.prev; |
2044 | ent->lru.prev->next = ent->lru.next; | |
18bdec11 SP |
2045 | delta_base_cached -= ent->size; |
2046 | } | |
2047 | } | |
2048 | ||
3d20c636 SP |
2049 | void clear_delta_base_cache(void) |
2050 | { | |
2051 | unsigned long p; | |
2052 | for (p = 0; p < MAX_DELTA_CACHE; p++) | |
2053 | release_delta_base_cache(&delta_base_cache[p]); | |
2054 | } | |
2055 | ||
62f255ad LT |
2056 | static void add_delta_base_cache(struct packed_git *p, off_t base_offset, |
2057 | void *base, unsigned long base_size, enum object_type type) | |
2058 | { | |
5e08ecbf | 2059 | unsigned long hash = pack_entry_hash(p, base_offset); |
e5e01619 | 2060 | struct delta_base_cache_entry *ent = delta_base_cache + hash; |
5e08ecbf | 2061 | struct delta_base_cache_lru_list *lru; |
e5e01619 | 2062 | |
18bdec11 SP |
2063 | release_delta_base_cache(ent); |
2064 | delta_base_cached += base_size; | |
5e08ecbf NP |
2065 | |
2066 | for (lru = delta_base_cache_lru.next; | |
2067 | delta_base_cached > delta_base_cache_limit | |
2068 | && lru != &delta_base_cache_lru; | |
2069 | lru = lru->next) { | |
2070 | struct delta_base_cache_entry *f = (void *)lru; | |
18bdec11 SP |
2071 | if (f->type == OBJ_BLOB) |
2072 | release_delta_base_cache(f); | |
2073 | } | |
5e08ecbf NP |
2074 | for (lru = delta_base_cache_lru.next; |
2075 | delta_base_cached > delta_base_cache_limit | |
2076 | && lru != &delta_base_cache_lru; | |
2077 | lru = lru->next) { | |
2078 | struct delta_base_cache_entry *f = (void *)lru; | |
2079 | release_delta_base_cache(f); | |
2080 | } | |
18bdec11 | 2081 | |
e5e01619 LT |
2082 | ent->p = p; |
2083 | ent->base_offset = base_offset; | |
2084 | ent->type = type; | |
2085 | ent->data = base; | |
2086 | ent->size = base_size; | |
5e08ecbf NP |
2087 | ent->lru.next = &delta_base_cache_lru; |
2088 | ent->lru.prev = delta_base_cache_lru.prev; | |
2089 | delta_base_cache_lru.prev->next = &ent->lru; | |
2090 | delta_base_cache_lru.prev = &ent->lru; | |
62f255ad LT |
2091 | } |
2092 | ||
c2c5b270 CC |
2093 | static void *read_object(const unsigned char *sha1, enum object_type *type, |
2094 | unsigned long *size); | |
2095 | ||
5f44324d JH |
2096 | static void write_pack_access_log(struct packed_git *p, off_t obj_offset) |
2097 | { | |
2098 | static FILE *log_file; | |
2099 | ||
b12ca963 NTND |
2100 | if (!log_pack_access) |
2101 | log_pack_access = getenv("GIT_TRACE_PACK_ACCESS"); | |
2102 | if (!log_pack_access) | |
2103 | log_pack_access = no_log_pack_access; | |
2104 | if (log_pack_access == no_log_pack_access) | |
2105 | return; | |
2106 | ||
5f44324d JH |
2107 | if (!log_file) { |
2108 | log_file = fopen(log_pack_access, "w"); | |
2109 | if (!log_file) { | |
2110 | error("cannot open pack access log '%s' for writing: %s", | |
2111 | log_pack_access, strerror(errno)); | |
b12ca963 | 2112 | log_pack_access = no_log_pack_access; |
5f44324d JH |
2113 | return; |
2114 | } | |
2115 | } | |
2116 | fprintf(log_file, "%s %"PRIuMAX"\n", | |
2117 | p->pack_name, (uintmax_t)obj_offset); | |
2118 | fflush(log_file); | |
2119 | } | |
2120 | ||
0e8189e2 NP |
2121 | int do_check_packed_object_crc; |
2122 | ||
abe601bb TR |
2123 | #define UNPACK_ENTRY_STACK_PREALLOC 64 |
2124 | struct unpack_entry_stack_ent { | |
2125 | off_t obj_offset; | |
2126 | off_t curpos; | |
2127 | unsigned long size; | |
2128 | }; | |
2129 | ||
c4001d92 | 2130 | void *unpack_entry(struct packed_git *p, off_t obj_offset, |
abe601bb | 2131 | enum object_type *final_type, unsigned long *final_size) |
1f688557 | 2132 | { |
03e79c88 | 2133 | struct pack_window *w_curs = NULL; |
c4001d92 | 2134 | off_t curpos = obj_offset; |
abe601bb TR |
2135 | void *data = NULL; |
2136 | unsigned long size; | |
2137 | enum object_type type; | |
2138 | struct unpack_entry_stack_ent small_delta_stack[UNPACK_ENTRY_STACK_PREALLOC]; | |
2139 | struct unpack_entry_stack_ent *delta_stack = small_delta_stack; | |
2140 | int delta_stack_nr = 0, delta_stack_alloc = UNPACK_ENTRY_STACK_PREALLOC; | |
2141 | int base_from_cache = 0; | |
1f688557 | 2142 | |
b12ca963 | 2143 | if (log_pack_access != no_log_pack_access) |
5f44324d JH |
2144 | write_pack_access_log(p, obj_offset); |
2145 | ||
abe601bb TR |
2146 | /* PHASE 1: drill down to the innermost base object */ |
2147 | for (;;) { | |
2148 | off_t base_offset; | |
2149 | int i; | |
2150 | struct delta_base_cache_entry *ent; | |
2151 | ||
77965f8b NTND |
2152 | ent = get_delta_base_cache_entry(p, curpos); |
2153 | if (eq_delta_base_cache_entry(ent, p, curpos)) { | |
2154 | type = ent->type; | |
2155 | data = ent->data; | |
2156 | size = ent->size; | |
2157 | clear_delta_base_cache_entry(ent); | |
2158 | base_from_cache = 1; | |
2159 | break; | |
2160 | } | |
2161 | ||
abe601bb TR |
2162 | if (do_check_packed_object_crc && p->index_version > 1) { |
2163 | struct revindex_entry *revidx = find_pack_revindex(p, obj_offset); | |
2164 | unsigned long len = revidx[1].offset - obj_offset; | |
2165 | if (check_pack_crc(p, &w_curs, obj_offset, len, revidx->nr)) { | |
2166 | const unsigned char *sha1 = | |
2167 | nth_packed_object_sha1(p, revidx->nr); | |
2168 | error("bad packed object CRC for %s", | |
2169 | sha1_to_hex(sha1)); | |
2170 | mark_bad_packed_object(p, sha1); | |
2171 | unuse_pack(&w_curs); | |
2172 | return NULL; | |
2173 | } | |
2174 | } | |
2175 | ||
abe601bb TR |
2176 | type = unpack_object_header(p, &w_curs, &curpos, &size); |
2177 | if (type != OBJ_OFS_DELTA && type != OBJ_REF_DELTA) | |
2178 | break; | |
2179 | ||
2180 | base_offset = get_delta_base(p, &w_curs, &curpos, type, obj_offset); | |
2181 | if (!base_offset) { | |
2182 | error("failed to validate delta base reference " | |
2183 | "at offset %"PRIuMAX" from %s", | |
2184 | (uintmax_t)curpos, p->pack_name); | |
2185 | /* bail to phase 2, in hopes of recovery */ | |
2186 | data = NULL; | |
2187 | break; | |
2188 | } | |
2189 | ||
2190 | /* push object, proceed to base */ | |
2191 | if (delta_stack_nr >= delta_stack_alloc | |
2192 | && delta_stack == small_delta_stack) { | |
2193 | delta_stack_alloc = alloc_nr(delta_stack_nr); | |
2194 | delta_stack = xmalloc(sizeof(*delta_stack)*delta_stack_alloc); | |
2195 | memcpy(delta_stack, small_delta_stack, | |
2196 | sizeof(*delta_stack)*delta_stack_nr); | |
2197 | } else { | |
2198 | ALLOC_GROW(delta_stack, delta_stack_nr+1, delta_stack_alloc); | |
2199 | } | |
2200 | i = delta_stack_nr++; | |
2201 | delta_stack[i].obj_offset = obj_offset; | |
2202 | delta_stack[i].curpos = curpos; | |
2203 | delta_stack[i].size = size; | |
2204 | ||
2205 | curpos = obj_offset = base_offset; | |
0e8189e2 NP |
2206 | } |
2207 | ||
abe601bb TR |
2208 | /* PHASE 2: handle the base */ |
2209 | switch (type) { | |
eb32d236 NP |
2210 | case OBJ_OFS_DELTA: |
2211 | case OBJ_REF_DELTA: | |
abe601bb TR |
2212 | if (data) |
2213 | die("BUG in unpack_entry: left loop at a valid delta"); | |
4d703a1a | 2214 | break; |
a733cb60 | 2215 | case OBJ_COMMIT: |
a733cb60 | 2216 | case OBJ_TREE: |
a733cb60 | 2217 | case OBJ_BLOB: |
a733cb60 | 2218 | case OBJ_TAG: |
abe601bb TR |
2219 | if (!base_from_cache) |
2220 | data = unpack_compressed_entry(p, &w_curs, curpos, size); | |
4d703a1a | 2221 | break; |
1f688557 | 2222 | default: |
8eca0b47 NP |
2223 | data = NULL; |
2224 | error("unknown object type %i at offset %"PRIuMAX" in %s", | |
abe601bb TR |
2225 | type, (uintmax_t)obj_offset, p->pack_name); |
2226 | } | |
2227 | ||
2228 | /* PHASE 3: apply deltas in order */ | |
2229 | ||
2230 | /* invariants: | |
2231 | * 'data' holds the base data, or NULL if there was corruption | |
2232 | */ | |
2233 | while (delta_stack_nr) { | |
2234 | void *delta_data; | |
2235 | void *base = data; | |
2236 | unsigned long delta_size, base_size = size; | |
2237 | int i; | |
2238 | ||
2239 | data = NULL; | |
2240 | ||
2241 | if (base) | |
2242 | add_delta_base_cache(p, obj_offset, base, base_size, type); | |
2243 | ||
2244 | if (!base) { | |
2245 | /* | |
2246 | * We're probably in deep shit, but let's try to fetch | |
2247 | * the required base anyway from another pack or loose. | |
2248 | * This is costly but should happen only in the presence | |
2249 | * of a corrupted pack, and is better than failing outright. | |
2250 | */ | |
2251 | struct revindex_entry *revidx; | |
2252 | const unsigned char *base_sha1; | |
2253 | revidx = find_pack_revindex(p, obj_offset); | |
2254 | if (revidx) { | |
2255 | base_sha1 = nth_packed_object_sha1(p, revidx->nr); | |
2256 | error("failed to read delta base object %s" | |
2257 | " at offset %"PRIuMAX" from %s", | |
2258 | sha1_to_hex(base_sha1), (uintmax_t)obj_offset, | |
2259 | p->pack_name); | |
2260 | mark_bad_packed_object(p, base_sha1); | |
2261 | base = read_object(base_sha1, &type, &base_size); | |
2262 | } | |
2263 | } | |
2264 | ||
2265 | i = --delta_stack_nr; | |
2266 | obj_offset = delta_stack[i].obj_offset; | |
2267 | curpos = delta_stack[i].curpos; | |
2268 | delta_size = delta_stack[i].size; | |
2269 | ||
2270 | if (!base) | |
2271 | continue; | |
2272 | ||
2273 | delta_data = unpack_compressed_entry(p, &w_curs, curpos, delta_size); | |
2274 | ||
2275 | if (!delta_data) { | |
2276 | error("failed to unpack compressed delta " | |
2277 | "at offset %"PRIuMAX" from %s", | |
2278 | (uintmax_t)curpos, p->pack_name); | |
abe601bb TR |
2279 | data = NULL; |
2280 | continue; | |
2281 | } | |
2282 | ||
2283 | data = patch_delta(base, base_size, | |
2284 | delta_data, delta_size, | |
2285 | &size); | |
1ee886c1 JK |
2286 | |
2287 | /* | |
2288 | * We could not apply the delta; warn the user, but keep going. | |
2289 | * Our failure will be noticed either in the next iteration of | |
2290 | * the loop, or if this is the final delta, in the caller when | |
2291 | * we return NULL. Those code paths will take care of making | |
2292 | * a more explicit warning and retrying with another copy of | |
2293 | * the object. | |
2294 | */ | |
abe601bb | 2295 | if (!data) |
1ee886c1 | 2296 | error("failed to apply delta"); |
abe601bb | 2297 | |
4b8f772c | 2298 | free(delta_data); |
1f688557 | 2299 | } |
abe601bb TR |
2300 | |
2301 | *final_type = type; | |
2302 | *final_size = size; | |
2303 | ||
03e79c88 | 2304 | unuse_pack(&w_curs); |
21666f1a | 2305 | return data; |
1f688557 JH |
2306 | } |
2307 | ||
d079837e | 2308 | const unsigned char *nth_packed_object_sha1(struct packed_git *p, |
d72308e0 | 2309 | uint32_t n) |
9a217f2a | 2310 | { |
42873078 | 2311 | const unsigned char *index = p->index_data; |
d079837e SP |
2312 | if (!index) { |
2313 | if (open_pack_index(p)) | |
2314 | return NULL; | |
2315 | index = p->index_data; | |
2316 | } | |
57059091 | 2317 | if (n >= p->num_objects) |
d72308e0 | 2318 | return NULL; |
74e34e1f NP |
2319 | index += 4 * 256; |
2320 | if (p->index_version == 1) { | |
2321 | return index + 24 * n + 4; | |
2322 | } else { | |
2323 | index += 8; | |
2324 | return index + 20 * n; | |
2325 | } | |
2326 | } | |
2327 | ||
99093238 | 2328 | off_t nth_packed_object_offset(const struct packed_git *p, uint32_t n) |
74e34e1f NP |
2329 | { |
2330 | const unsigned char *index = p->index_data; | |
2331 | index += 4 * 256; | |
2332 | if (p->index_version == 1) { | |
2333 | return ntohl(*((uint32_t *)(index + 24 * n))); | |
2334 | } else { | |
2335 | uint32_t off; | |
2336 | index += 8 + p->num_objects * (20 + 4); | |
2337 | off = ntohl(*((uint32_t *)(index + 4 * n))); | |
2338 | if (!(off & 0x80000000)) | |
2339 | return off; | |
2340 | index += p->num_objects * 4 + (off & 0x7fffffff) * 8; | |
2341 | return (((uint64_t)ntohl(*((uint32_t *)(index + 0)))) << 32) | | |
2342 | ntohl(*((uint32_t *)(index + 4))); | |
2343 | } | |
9a217f2a JH |
2344 | } |
2345 | ||
c4001d92 | 2346 | off_t find_pack_entry_one(const unsigned char *sha1, |
43057304 | 2347 | struct packed_git *p) |
1f688557 | 2348 | { |
42873078 | 2349 | const uint32_t *level1_ofs = p->index_data; |
42873078 | 2350 | const unsigned char *index = p->index_data; |
628522ec JH |
2351 | unsigned hi, lo, stride; |
2352 | static int use_lookup = -1; | |
2353 | static int debug_lookup = -1; | |
2354 | ||
2355 | if (debug_lookup < 0) | |
2356 | debug_lookup = !!getenv("GIT_DEBUG_LOOKUP"); | |
42873078 | 2357 | |
d079837e SP |
2358 | if (!index) { |
2359 | if (open_pack_index(p)) | |
2360 | return 0; | |
2361 | level1_ofs = p->index_data; | |
2362 | index = p->index_data; | |
2363 | } | |
74e34e1f NP |
2364 | if (p->index_version > 1) { |
2365 | level1_ofs += 2; | |
2366 | index += 8; | |
2367 | } | |
42873078 | 2368 | index += 4 * 256; |
74e34e1f NP |
2369 | hi = ntohl(level1_ofs[*sha1]); |
2370 | lo = ((*sha1 == 0x0) ? 0 : ntohl(level1_ofs[*sha1 - 1])); | |
628522ec JH |
2371 | if (p->index_version > 1) { |
2372 | stride = 20; | |
2373 | } else { | |
2374 | stride = 24; | |
2375 | index += 4; | |
2376 | } | |
2377 | ||
2378 | if (debug_lookup) | |
6e1c2344 | 2379 | printf("%02x%02x%02x... lo %u hi %u nr %"PRIu32"\n", |
628522ec JH |
2380 | sha1[0], sha1[1], sha1[2], lo, hi, p->num_objects); |
2381 | ||
2382 | if (use_lookup < 0) | |
2383 | use_lookup = !!getenv("GIT_USE_LOOKUP"); | |
2384 | if (use_lookup) { | |
2385 | int pos = sha1_entry_pos(index, stride, 0, | |
2386 | lo, hi, p->num_objects, sha1); | |
2387 | if (pos < 0) | |
2388 | return 0; | |
2389 | return nth_packed_object_offset(p, pos); | |
2390 | } | |
1f688557 JH |
2391 | |
2392 | do { | |
74e34e1f | 2393 | unsigned mi = (lo + hi) / 2; |
628522ec JH |
2394 | int cmp = hashcmp(index + mi * stride, sha1); |
2395 | ||
2396 | if (debug_lookup) | |
2397 | printf("lo %u hi %u rg %u mi %u\n", | |
2398 | lo, hi, hi - lo, mi); | |
43057304 | 2399 | if (!cmp) |
74e34e1f | 2400 | return nth_packed_object_offset(p, mi); |
1f688557 JH |
2401 | if (cmp > 0) |
2402 | hi = mi; | |
2403 | else | |
2404 | lo = mi+1; | |
2405 | } while (lo < hi); | |
2406 | return 0; | |
2407 | } | |
2408 | ||
4c080182 | 2409 | int is_pack_valid(struct packed_git *p) |
d131b7af SP |
2410 | { |
2411 | /* An already open pack is known to be valid. */ | |
2412 | if (p->pack_fd != -1) | |
2413 | return 1; | |
2414 | ||
2415 | /* If the pack has one window completely covering the | |
2416 | * file size, the pack is known to be valid even if | |
2417 | * the descriptor is not currently open. | |
2418 | */ | |
2419 | if (p->windows) { | |
2420 | struct pack_window *w = p->windows; | |
2421 | ||
2422 | if (!w->offset && w->len == p->pack_size) | |
2423 | return 1; | |
2424 | } | |
2425 | ||
2426 | /* Force the pack to open to prove its valid. */ | |
2427 | return !open_packed_git(p); | |
2428 | } | |
2429 | ||
95099731 NTND |
2430 | static int fill_pack_entry(const unsigned char *sha1, |
2431 | struct pack_entry *e, | |
2432 | struct packed_git *p) | |
2433 | { | |
2434 | off_t offset; | |
2435 | ||
2436 | if (p->num_bad_objects) { | |
2437 | unsigned i; | |
2438 | for (i = 0; i < p->num_bad_objects; i++) | |
2439 | if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i)) | |
2440 | return 0; | |
2441 | } | |
2442 | ||
2443 | offset = find_pack_entry_one(sha1, p); | |
2444 | if (!offset) | |
2445 | return 0; | |
2446 | ||
2447 | /* | |
2448 | * We are about to tell the caller where they can locate the | |
2449 | * requested object. We better make sure the packfile is | |
2450 | * still here and can be accessed before supplying that | |
2451 | * answer, as it may have been deleted since the index was | |
2452 | * loaded! | |
2453 | */ | |
2454 | if (!is_pack_valid(p)) { | |
2455 | warning("packfile %s cannot be accessed", p->pack_name); | |
2456 | return 0; | |
2457 | } | |
2458 | e->offset = offset; | |
2459 | e->p = p; | |
2460 | hashcpy(e->sha1, sha1); | |
2461 | return 1; | |
2462 | } | |
2463 | ||
4d6acb70 | 2464 | static int find_pack_entry(const unsigned char *sha1, struct pack_entry *e) |
1f688557 JH |
2465 | { |
2466 | struct packed_git *p; | |
43057304 | 2467 | |
1f688557 | 2468 | prepare_packed_git(); |
f7c22cc6 NP |
2469 | if (!packed_git) |
2470 | return 0; | |
1f688557 | 2471 | |
c01f51cc NTND |
2472 | if (last_found_pack && fill_pack_entry(sha1, e, last_found_pack)) |
2473 | return 1; | |
8eca0b47 | 2474 | |
c01f51cc NTND |
2475 | for (p = packed_git; p; p = p->next) { |
2476 | if (p == last_found_pack || !fill_pack_entry(sha1, e, p)) | |
2477 | continue; | |
f7c22cc6 | 2478 | |
c01f51cc NTND |
2479 | last_found_pack = p; |
2480 | return 1; | |
2481 | } | |
1f688557 JH |
2482 | return 0; |
2483 | } | |
2484 | ||
a6080a0a | 2485 | struct packed_git *find_sha1_pack(const unsigned char *sha1, |
bf592c50 DB |
2486 | struct packed_git *packs) |
2487 | { | |
2488 | struct packed_git *p; | |
bf592c50 DB |
2489 | |
2490 | for (p = packs; p; p = p->next) { | |
43057304 | 2491 | if (find_pack_entry_one(sha1, p)) |
bf592c50 DB |
2492 | return p; |
2493 | } | |
2494 | return NULL; | |
21666f1a | 2495 | |
bf592c50 DB |
2496 | } |
2497 | ||
052fe5ea | 2498 | static int sha1_loose_object_info(const unsigned char *sha1, |
23c339c0 | 2499 | struct object_info *oi) |
65c2e0c3 | 2500 | { |
36e4d74a | 2501 | int status; |
65c2e0c3 JH |
2502 | unsigned long mapsize, size; |
2503 | void *map; | |
ef49a7a0 | 2504 | git_zstream stream; |
d65a16f6 | 2505 | char hdr[32]; |
65c2e0c3 | 2506 | |
052fe5ea JK |
2507 | /* |
2508 | * If we don't care about type or size, then we don't | |
4ef8d1dd JH |
2509 | * need to look inside the object at all. Note that we |
2510 | * do not optimize out the stat call, even if the | |
2511 | * caller doesn't care about the disk-size, since our | |
2512 | * return value implicitly indicates whether the | |
2513 | * object even exists. | |
052fe5ea | 2514 | */ |
23c339c0 | 2515 | if (!oi->typep && !oi->sizep) { |
4ef8d1dd JH |
2516 | struct stat st; |
2517 | if (stat_sha1_file(sha1, &st) < 0) | |
2518 | return -1; | |
2519 | if (oi->disk_sizep) | |
23c339c0 | 2520 | *oi->disk_sizep = st.st_size; |
052fe5ea JK |
2521 | return 0; |
2522 | } | |
2523 | ||
bb6b8e4f | 2524 | map = map_sha1_file(sha1, &mapsize); |
f0df4ed5 | 2525 | if (!map) |
dbea72a8 | 2526 | return -1; |
23c339c0 JK |
2527 | if (oi->disk_sizep) |
2528 | *oi->disk_sizep = mapsize; | |
36e4d74a JH |
2529 | if (unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)) < 0) |
2530 | status = error("unable to unpack %s header", | |
2531 | sha1_to_hex(sha1)); | |
21666f1a | 2532 | else if ((status = parse_sha1_header(hdr, &size)) < 0) |
36e4d74a | 2533 | status = error("unable to parse %s header", sha1_to_hex(sha1)); |
23c339c0 JK |
2534 | else if (oi->sizep) |
2535 | *oi->sizep = size; | |
39c68542 | 2536 | git_inflate_end(&stream); |
65c2e0c3 | 2537 | munmap(map, mapsize); |
23c339c0 JK |
2538 | if (oi->typep) |
2539 | *oi->typep = status; | |
052fe5ea | 2540 | return 0; |
65c2e0c3 JH |
2541 | } |
2542 | ||
9a490590 | 2543 | int sha1_object_info_extended(const unsigned char *sha1, struct object_info *oi) |
f0df4ed5 | 2544 | { |
c4d9986f | 2545 | struct cached_object *co; |
f0df4ed5 | 2546 | struct pack_entry e; |
5b086407 | 2547 | int rtype; |
f0df4ed5 | 2548 | |
c4d9986f NTND |
2549 | co = find_cached_object(sha1); |
2550 | if (co) { | |
5b086407 JK |
2551 | if (oi->typep) |
2552 | *(oi->typep) = co->type; | |
9a490590 JH |
2553 | if (oi->sizep) |
2554 | *(oi->sizep) = co->size; | |
161f00e7 JK |
2555 | if (oi->disk_sizep) |
2556 | *(oi->disk_sizep) = 0; | |
9a490590 | 2557 | oi->whence = OI_CACHED; |
5b086407 | 2558 | return 0; |
c4d9986f NTND |
2559 | } |
2560 | ||
cd673c1f | 2561 | if (!find_pack_entry(sha1, &e)) { |
ddd63e64 | 2562 | /* Most likely it's a loose object. */ |
23c339c0 | 2563 | if (!sha1_loose_object_info(sha1, oi)) { |
9a490590 | 2564 | oi->whence = OI_LOOSE; |
5b086407 | 2565 | return 0; |
9a490590 | 2566 | } |
ddd63e64 SG |
2567 | |
2568 | /* Not a loose object; someone else may have just packed it. */ | |
f0df4ed5 | 2569 | reprepare_packed_git(); |
cd673c1f | 2570 | if (!find_pack_entry(sha1, &e)) |
052fe5ea | 2571 | return -1; |
f0df4ed5 | 2572 | } |
3d77d877 | 2573 | |
23c339c0 | 2574 | rtype = packed_object_info(e.p, e.offset, oi); |
412916ee | 2575 | if (rtype < 0) { |
3d77d877 | 2576 | mark_bad_packed_object(e.p, sha1); |
412916ee | 2577 | return sha1_object_info_extended(sha1, oi); |
5266d369 JH |
2578 | } else if (in_delta_base_cache(e.p, e.offset)) { |
2579 | oi->whence = OI_DBCACHED; | |
9a490590 JH |
2580 | } else { |
2581 | oi->whence = OI_PACKED; | |
2582 | oi->u.packed.offset = e.offset; | |
2583 | oi->u.packed.pack = e.p; | |
2584 | oi->u.packed.is_delta = (rtype == OBJ_REF_DELTA || | |
2585 | rtype == OBJ_OFS_DELTA); | |
3d77d877 NP |
2586 | } |
2587 | ||
5b086407 | 2588 | return 0; |
f0df4ed5 JS |
2589 | } |
2590 | ||
3fc0dca9 | 2591 | /* returns enum object_type or negative */ |
9a490590 JH |
2592 | int sha1_object_info(const unsigned char *sha1, unsigned long *sizep) |
2593 | { | |
5b086407 | 2594 | enum object_type type; |
d099b717 | 2595 | struct object_info oi = {NULL}; |
9a490590 | 2596 | |
5b086407 | 2597 | oi.typep = &type; |
9a490590 | 2598 | oi.sizep = sizep; |
5b086407 JK |
2599 | if (sha1_object_info_extended(sha1, &oi) < 0) |
2600 | return -1; | |
2601 | return type; | |
9a490590 JH |
2602 | } |
2603 | ||
21666f1a NP |
2604 | static void *read_packed_sha1(const unsigned char *sha1, |
2605 | enum object_type *type, unsigned long *size) | |
1f688557 JH |
2606 | { |
2607 | struct pack_entry e; | |
8eca0b47 | 2608 | void *data; |
1f688557 | 2609 | |
cd673c1f | 2610 | if (!find_pack_entry(sha1, &e)) |
1f688557 | 2611 | return NULL; |
8eca0b47 NP |
2612 | data = cache_or_unpack_entry(e.p, e.offset, size, type, 1); |
2613 | if (!data) { | |
2614 | /* | |
2615 | * We're probably in deep shit, but let's try to fetch | |
2616 | * the required object anyway from another pack or loose. | |
2617 | * This should happen only in the presence of a corrupted | |
2618 | * pack, and is better than failing outright. | |
2619 | */ | |
2620 | error("failed to read object %s at offset %"PRIuMAX" from %s", | |
2621 | sha1_to_hex(sha1), (uintmax_t)e.offset, e.p->pack_name); | |
2622 | mark_bad_packed_object(e.p, sha1); | |
ac939109 | 2623 | data = read_object(sha1, type, size); |
8eca0b47 NP |
2624 | } |
2625 | return data; | |
1f688557 JH |
2626 | } |
2627 | ||
21666f1a NP |
2628 | int pretend_sha1_file(void *buf, unsigned long len, enum object_type type, |
2629 | unsigned char *sha1) | |
d66b37bb JH |
2630 | { |
2631 | struct cached_object *co; | |
2632 | ||
21666f1a | 2633 | hash_sha1_file(buf, len, typename(type), sha1); |
d66b37bb JH |
2634 | if (has_sha1_file(sha1) || find_cached_object(sha1)) |
2635 | return 0; | |
2636 | if (cached_object_alloc <= cached_object_nr) { | |
2637 | cached_object_alloc = alloc_nr(cached_object_alloc); | |
2638 | cached_objects = xrealloc(cached_objects, | |
2639 | sizeof(*cached_objects) * | |
2640 | cached_object_alloc); | |
2641 | } | |
2642 | co = &cached_objects[cached_object_nr++]; | |
2643 | co->size = len; | |
21666f1a | 2644 | co->type = type; |
efa13f7b JH |
2645 | co->buf = xmalloc(len); |
2646 | memcpy(co->buf, buf, len); | |
d66b37bb JH |
2647 | hashcpy(co->sha1, sha1); |
2648 | return 0; | |
2649 | } | |
2650 | ||
c2c5b270 CC |
2651 | static void *read_object(const unsigned char *sha1, enum object_type *type, |
2652 | unsigned long *size) | |
0fcfd160 LT |
2653 | { |
2654 | unsigned long mapsize; | |
2655 | void *map, *buf; | |
d66b37bb JH |
2656 | struct cached_object *co; |
2657 | ||
2658 | co = find_cached_object(sha1); | |
2659 | if (co) { | |
21666f1a | 2660 | *type = co->type; |
d66b37bb | 2661 | *size = co->size; |
182af834 | 2662 | return xmemdupz(co->buf, co->size); |
d66b37bb | 2663 | } |
0fcfd160 | 2664 | |
8276c007 PE |
2665 | buf = read_packed_sha1(sha1, type, size); |
2666 | if (buf) | |
2667 | return buf; | |
bb6b8e4f | 2668 | map = map_sha1_file(sha1, &mapsize); |
0fcfd160 | 2669 | if (map) { |
7efbff75 | 2670 | buf = unpack_sha1_file(map, mapsize, type, size, sha1); |
0fcfd160 LT |
2671 | munmap(map, mapsize); |
2672 | return buf; | |
2673 | } | |
637cdd9d | 2674 | reprepare_packed_git(); |
8276c007 | 2675 | return read_packed_sha1(sha1, type, size); |
0fcfd160 LT |
2676 | } |
2677 | ||
b6c4cecc JH |
2678 | /* |
2679 | * This function dies on corrupt objects; the callers who want to | |
2680 | * deal with them should arrange to call read_object() and give error | |
2681 | * messages themselves. | |
2682 | */ | |
5bf29b95 JH |
2683 | void *read_sha1_file_extended(const unsigned char *sha1, |
2684 | enum object_type *type, | |
2685 | unsigned long *size, | |
2686 | unsigned flag) | |
ac939109 | 2687 | { |
3ba7a065 | 2688 | void *data; |
e8b15e61 | 2689 | char *path; |
b6c4cecc | 2690 | const struct packed_git *p; |
5bf29b95 JH |
2691 | const unsigned char *repl = (flag & READ_SHA1_FILE_REPLACE) |
2692 | ? lookup_replace_object(sha1) : sha1; | |
b6c4cecc | 2693 | |
3ba7a065 JH |
2694 | errno = 0; |
2695 | data = read_object(repl, type, size); | |
4bbf5a26 | 2696 | if (data) |
b6c4cecc | 2697 | return data; |
68095570 | 2698 | |
25f3af3f | 2699 | if (errno && errno != ENOENT) |
3ba7a065 JH |
2700 | die_errno("failed to read object %s", sha1_to_hex(sha1)); |
2701 | ||
68095570 | 2702 | /* die if we replaced an object with one that does not exist */ |
b6c4cecc | 2703 | if (repl != sha1) |
68095570 CC |
2704 | die("replacement %s not found for %s", |
2705 | sha1_to_hex(repl), sha1_to_hex(sha1)); | |
2706 | ||
b6c4cecc JH |
2707 | if (has_loose_object(repl)) { |
2708 | path = sha1_file_name(sha1); | |
2709 | die("loose object %s (stored in %s) is corrupt", | |
2710 | sha1_to_hex(repl), path); | |
e8b15e61 | 2711 | } |
68095570 | 2712 | |
b6c4cecc JH |
2713 | if ((p = has_packed_and_bad(repl)) != NULL) |
2714 | die("packed object %s (stored in %s) is corrupt", | |
2715 | sha1_to_hex(repl), p->pack_name); | |
f5552aee | 2716 | |
b6c4cecc | 2717 | return NULL; |
ac939109 NP |
2718 | } |
2719 | ||
40469ee9 | 2720 | void *read_object_with_reference(const unsigned char *sha1, |
21666f1a | 2721 | const char *required_type_name, |
40469ee9 JH |
2722 | unsigned long *size, |
2723 | unsigned char *actual_sha1_return) | |
f4913f91 | 2724 | { |
21666f1a | 2725 | enum object_type type, required_type; |
f4913f91 JH |
2726 | void *buffer; |
2727 | unsigned long isize; | |
40469ee9 | 2728 | unsigned char actual_sha1[20]; |
f4913f91 | 2729 | |
21666f1a | 2730 | required_type = type_from_string(required_type_name); |
e702496e | 2731 | hashcpy(actual_sha1, sha1); |
40469ee9 JH |
2732 | while (1) { |
2733 | int ref_length = -1; | |
2734 | const char *ref_type = NULL; | |
f4913f91 | 2735 | |
21666f1a | 2736 | buffer = read_sha1_file(actual_sha1, &type, &isize); |
40469ee9 JH |
2737 | if (!buffer) |
2738 | return NULL; | |
21666f1a | 2739 | if (type == required_type) { |
40469ee9 JH |
2740 | *size = isize; |
2741 | if (actual_sha1_return) | |
e702496e | 2742 | hashcpy(actual_sha1_return, actual_sha1); |
40469ee9 JH |
2743 | return buffer; |
2744 | } | |
2745 | /* Handle references */ | |
21666f1a | 2746 | else if (type == OBJ_COMMIT) |
40469ee9 | 2747 | ref_type = "tree "; |
21666f1a | 2748 | else if (type == OBJ_TAG) |
40469ee9 JH |
2749 | ref_type = "object "; |
2750 | else { | |
2751 | free(buffer); | |
2752 | return NULL; | |
2753 | } | |
2754 | ref_length = strlen(ref_type); | |
f4913f91 | 2755 | |
50974ec9 MK |
2756 | if (ref_length + 40 > isize || |
2757 | memcmp(buffer, ref_type, ref_length) || | |
1d7f171c | 2758 | get_sha1_hex((char *) buffer + ref_length, actual_sha1)) { |
40469ee9 JH |
2759 | free(buffer); |
2760 | return NULL; | |
2761 | } | |
1cf58e72 | 2762 | free(buffer); |
40469ee9 JH |
2763 | /* Now we have the ID of the referred-to object in |
2764 | * actual_sha1. Check again. */ | |
f4913f91 | 2765 | } |
f4913f91 JH |
2766 | } |
2767 | ||
ce9fbf16 | 2768 | static void write_sha1_file_prepare(const void *buf, unsigned long len, |
972a9155 | 2769 | const char *type, unsigned char *sha1, |
d65a16f6 | 2770 | char *hdr, int *hdrlen) |
d410c0f5 | 2771 | { |
9126f009 | 2772 | git_SHA_CTX c; |
d410c0f5 JH |
2773 | |
2774 | /* Generate the header */ | |
d65a16f6 | 2775 | *hdrlen = sprintf(hdr, "%s %lu", type, len)+1; |
d410c0f5 JH |
2776 | |
2777 | /* Sha1.. */ | |
9126f009 NP |
2778 | git_SHA1_Init(&c); |
2779 | git_SHA1_Update(&c, hdr, *hdrlen); | |
2780 | git_SHA1_Update(&c, buf, len); | |
2781 | git_SHA1_Final(sha1, &c); | |
d410c0f5 JH |
2782 | } |
2783 | ||
230f1322 | 2784 | /* |
5a688fe4 JH |
2785 | * Move the just written object into its final resting place. |
2786 | * NEEDSWORK: this should be renamed to finalize_temp_file() as | |
2787 | * "moving" is only a part of what it does, when no patch between | |
2788 | * master to pu changes the call sites of this function. | |
230f1322 | 2789 | */ |
839837b9 | 2790 | int move_temp_to_file(const char *tmpfile, const char *filename) |
230f1322 | 2791 | { |
e32c0a9c | 2792 | int ret = 0; |
5a688fe4 | 2793 | |
348df166 | 2794 | if (object_creation_mode == OBJECT_CREATION_USES_RENAMES) |
be66a6c4 JS |
2795 | goto try_rename; |
2796 | else if (link(tmpfile, filename)) | |
e32c0a9c | 2797 | ret = errno; |
7ebb6fca LT |
2798 | |
2799 | /* | |
2800 | * Coda hack - coda doesn't like cross-directory links, | |
2801 | * so we fall back to a rename, which will mean that it | |
2802 | * won't be able to check collisions, but that's not a | |
2803 | * big deal. | |
2804 | * | |
2805 | * The same holds for FAT formatted media. | |
2806 | * | |
3be1f18e | 2807 | * When this succeeds, we just return. We have nothing |
7ebb6fca LT |
2808 | * left to unlink. |
2809 | */ | |
2810 | if (ret && ret != EEXIST) { | |
be66a6c4 | 2811 | try_rename: |
7ebb6fca | 2812 | if (!rename(tmpfile, filename)) |
3be1f18e | 2813 | goto out; |
9e48b389 | 2814 | ret = errno; |
230f1322 | 2815 | } |
691f1a28 | 2816 | unlink_or_warn(tmpfile); |
230f1322 LT |
2817 | if (ret) { |
2818 | if (ret != EEXIST) { | |
82247e9b | 2819 | return error("unable to write sha1 filename %s: %s", filename, strerror(ret)); |
230f1322 LT |
2820 | } |
2821 | /* FIXME!!! Collision check here ? */ | |
2822 | } | |
2823 | ||
3be1f18e | 2824 | out: |
5256b006 | 2825 | if (adjust_shared_perm(filename)) |
5a688fe4 | 2826 | return error("unable to set permission to '%s'", filename); |
230f1322 LT |
2827 | return 0; |
2828 | } | |
2829 | ||
4d548150 LT |
2830 | static int write_buffer(int fd, const void *buf, size_t len) |
2831 | { | |
d34cf19b | 2832 | if (write_in_full(fd, buf, len) < 0) |
93822c22 | 2833 | return error("file write error (%s)", strerror(errno)); |
4d548150 LT |
2834 | return 0; |
2835 | } | |
2836 | ||
ce9fbf16 | 2837 | int hash_sha1_file(const void *buf, unsigned long len, const char *type, |
abdc3fc8 RS |
2838 | unsigned char *sha1) |
2839 | { | |
d65a16f6 | 2840 | char hdr[32]; |
abdc3fc8 RS |
2841 | int hdrlen; |
2842 | write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen); | |
2843 | return 0; | |
2844 | } | |
2845 | ||
e9039dd3 LT |
2846 | /* Finalize a file on disk, and close it. */ |
2847 | static void close_sha1_file(int fd) | |
2848 | { | |
aafe9fba LT |
2849 | if (fsync_object_files) |
2850 | fsync_or_die(fd, "sha1 file"); | |
e9039dd3 | 2851 | if (close(fd) != 0) |
d824cbba | 2852 | die_errno("error when closing sha1 file"); |
e9039dd3 LT |
2853 | } |
2854 | ||
5723fe7e LT |
2855 | /* Size of directory component, including the ending '/' */ |
2856 | static inline int directory_size(const char *filename) | |
2857 | { | |
2858 | const char *s = strrchr(filename, '/'); | |
2859 | if (!s) | |
2860 | return 0; | |
2861 | return s - filename + 1; | |
2862 | } | |
2863 | ||
2864 | /* | |
2865 | * This creates a temporary file in the same directory as the final | |
2866 | * 'filename' | |
2867 | * | |
2868 | * We want to avoid cross-directory filename renames, because those | |
2869 | * can have problems on various filesystems (FAT, NFS, Coda). | |
2870 | */ | |
2871 | static int create_tmpfile(char *buffer, size_t bufsiz, const char *filename) | |
2872 | { | |
2873 | int fd, dirlen = directory_size(filename); | |
2874 | ||
2875 | if (dirlen + 20 > bufsiz) { | |
2876 | errno = ENAMETOOLONG; | |
2877 | return -1; | |
2878 | } | |
2879 | memcpy(buffer, filename, dirlen); | |
2880 | strcpy(buffer + dirlen, "tmp_obj_XXXXXX"); | |
5256b006 | 2881 | fd = git_mkstemp_mode(buffer, 0444); |
cbacbf4e | 2882 | if (fd < 0 && dirlen && errno == ENOENT) { |
5723fe7e | 2883 | /* Make sure the directory exists */ |
6ff6af62 | 2884 | memcpy(buffer, filename, dirlen); |
5723fe7e | 2885 | buffer[dirlen-1] = 0; |
b2476a60 JH |
2886 | if (mkdir(buffer, 0777) && errno != EEXIST) |
2887 | return -1; | |
2888 | if (adjust_shared_perm(buffer)) | |
5723fe7e LT |
2889 | return -1; |
2890 | ||
2891 | /* Try again */ | |
2892 | strcpy(buffer + dirlen - 1, "/tmp_obj_XXXXXX"); | |
5256b006 | 2893 | fd = git_mkstemp_mode(buffer, 0444); |
5723fe7e LT |
2894 | } |
2895 | return fd; | |
2896 | } | |
2897 | ||
bbac7311 | 2898 | static int write_loose_object(const unsigned char *sha1, char *hdr, int hdrlen, |
c00e657d | 2899 | const void *buf, unsigned long len, time_t mtime) |
0fcfd160 | 2900 | { |
915308b1 | 2901 | int fd, ret; |
9892beba | 2902 | unsigned char compressed[4096]; |
ef49a7a0 | 2903 | git_zstream stream; |
748af44c NP |
2904 | git_SHA_CTX c; |
2905 | unsigned char parano_sha1[20]; | |
706bc531 | 2906 | char *filename; |
ab1900a3 | 2907 | static char tmp_file[PATH_MAX]; |
a44c9a5e | 2908 | |
972a9155 | 2909 | filename = sha1_file_name(sha1); |
ab1900a3 | 2910 | fd = create_tmpfile(tmp_file, sizeof(tmp_file), filename); |
aac17941 | 2911 | if (fd < 0) { |
35243577 | 2912 | if (errno == EACCES) |
82247e9b | 2913 | return error("insufficient permission for adding an object to repository database %s", get_object_directory()); |
916d081b | 2914 | else |
5eaeda70 | 2915 | return error("unable to create temporary file: %s", strerror(errno)); |
aac17941 LT |
2916 | } |
2917 | ||
0fcfd160 LT |
2918 | /* Set it up */ |
2919 | memset(&stream, 0, sizeof(stream)); | |
55bb5c91 | 2920 | git_deflate_init(&stream, zlib_compression_level); |
0fcfd160 | 2921 | stream.next_out = compressed; |
9892beba | 2922 | stream.avail_out = sizeof(compressed); |
748af44c | 2923 | git_SHA1_Init(&c); |
a44c9a5e LT |
2924 | |
2925 | /* First header.. */ | |
d65a16f6 | 2926 | stream.next_in = (unsigned char *)hdr; |
a44c9a5e | 2927 | stream.avail_in = hdrlen; |
55bb5c91 JH |
2928 | while (git_deflate(&stream, 0) == Z_OK) |
2929 | ; /* nothing */ | |
748af44c | 2930 | git_SHA1_Update(&c, hdr, hdrlen); |
a44c9a5e LT |
2931 | |
2932 | /* Then the data itself.. */ | |
c00e657d | 2933 | stream.next_in = (void *)buf; |
a44c9a5e | 2934 | stream.avail_in = len; |
9892beba | 2935 | do { |
748af44c | 2936 | unsigned char *in0 = stream.next_in; |
55bb5c91 | 2937 | ret = git_deflate(&stream, Z_FINISH); |
748af44c | 2938 | git_SHA1_Update(&c, in0, stream.next_in - in0); |
9892beba NP |
2939 | if (write_buffer(fd, compressed, stream.next_out - compressed) < 0) |
2940 | die("unable to write sha1 file"); | |
2941 | stream.next_out = compressed; | |
2942 | stream.avail_out = sizeof(compressed); | |
2943 | } while (ret == Z_OK); | |
2944 | ||
ac54c277 LT |
2945 | if (ret != Z_STREAM_END) |
2946 | die("unable to deflate new object %s (%d)", sha1_to_hex(sha1), ret); | |
55bb5c91 | 2947 | ret = git_deflate_end_gently(&stream); |
ac54c277 LT |
2948 | if (ret != Z_OK) |
2949 | die("deflateEnd on object %s failed (%d)", sha1_to_hex(sha1), ret); | |
748af44c NP |
2950 | git_SHA1_Final(parano_sha1, &c); |
2951 | if (hashcmp(sha1, parano_sha1) != 0) | |
2952 | die("confused by unstable object source data for %s", sha1_to_hex(sha1)); | |
ac54c277 | 2953 | |
e9039dd3 | 2954 | close_sha1_file(fd); |
0fcfd160 | 2955 | |
bbac7311 NP |
2956 | if (mtime) { |
2957 | struct utimbuf utb; | |
2958 | utb.actime = mtime; | |
2959 | utb.modtime = mtime; | |
ab1900a3 | 2960 | if (utime(tmp_file, &utb) < 0) |
bbac7311 | 2961 | warning("failed utime() on %s: %s", |
ab1900a3 | 2962 | tmp_file, strerror(errno)); |
bbac7311 NP |
2963 | } |
2964 | ||
ab1900a3 | 2965 | return move_temp_to_file(tmp_file, filename); |
0fcfd160 | 2966 | } |
8237b185 | 2967 | |
c00e657d | 2968 | int write_sha1_file(const void *buf, unsigned long len, const char *type, unsigned char *returnsha1) |
bbac7311 NP |
2969 | { |
2970 | unsigned char sha1[20]; | |
2971 | char hdr[32]; | |
2972 | int hdrlen; | |
2973 | ||
2974 | /* Normally if we have it in the pack then we do not bother writing | |
2975 | * it out into .git/objects/??/?{38} file. | |
2976 | */ | |
2977 | write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen); | |
2978 | if (returnsha1) | |
2979 | hashcpy(returnsha1, sha1); | |
2980 | if (has_sha1_file(sha1)) | |
2981 | return 0; | |
2982 | return write_loose_object(sha1, hdr, hdrlen, buf, len, 0); | |
2983 | } | |
2984 | ||
2985 | int force_object_loose(const unsigned char *sha1, time_t mtime) | |
2986 | { | |
bbac7311 NP |
2987 | void *buf; |
2988 | unsigned long len; | |
2989 | enum object_type type; | |
2990 | char hdr[32]; | |
2991 | int hdrlen; | |
1fb23e65 | 2992 | int ret; |
bbac7311 | 2993 | |
c529d75a | 2994 | if (has_loose_object(sha1)) |
bbac7311 NP |
2995 | return 0; |
2996 | buf = read_packed_sha1(sha1, &type, &len); | |
2997 | if (!buf) | |
2998 | return error("cannot read sha1_file for %s", sha1_to_hex(sha1)); | |
2999 | hdrlen = sprintf(hdr, "%s %lu", typename(type), len) + 1; | |
1fb23e65 BS |
3000 | ret = write_loose_object(sha1, hdr, hdrlen, buf, len, mtime); |
3001 | free(buf); | |
3002 | ||
3003 | return ret; | |
bbac7311 NP |
3004 | } |
3005 | ||
bf592c50 DB |
3006 | int has_pack_index(const unsigned char *sha1) |
3007 | { | |
3008 | struct stat st; | |
3009 | if (stat(sha1_pack_index_name(sha1), &st)) | |
3010 | return 0; | |
3011 | return 1; | |
bf592c50 DB |
3012 | } |
3013 | ||
cd673c1f JH |
3014 | int has_sha1_pack(const unsigned char *sha1) |
3015 | { | |
3016 | struct pack_entry e; | |
3017 | return find_pack_entry(sha1, &e); | |
3018 | } | |
3019 | ||
8237b185 DB |
3020 | int has_sha1_file(const unsigned char *sha1) |
3021 | { | |
1f688557 JH |
3022 | struct pack_entry e; |
3023 | ||
cd673c1f | 3024 | if (find_pack_entry(sha1, &e)) |
1f688557 | 3025 | return 1; |
45e8a748 JK |
3026 | if (has_loose_object(sha1)) |
3027 | return 1; | |
3028 | reprepare_packed_git(); | |
3029 | return find_pack_entry(sha1, &e); | |
8237b185 | 3030 | } |
74400e71 | 3031 | |
c879daa2 NTND |
3032 | static void check_tree(const void *buf, size_t size) |
3033 | { | |
3034 | struct tree_desc desc; | |
3035 | struct name_entry entry; | |
3036 | ||
3037 | init_tree_desc(&desc, buf, size); | |
3038 | while (tree_entry(&desc, &entry)) | |
3039 | /* do nothing | |
3040 | * tree_entry() will die() on malformed entries */ | |
3041 | ; | |
3042 | } | |
3043 | ||
3044 | static void check_commit(const void *buf, size_t size) | |
3045 | { | |
3046 | struct commit c; | |
3047 | memset(&c, 0, sizeof(c)); | |
3048 | if (parse_commit_buffer(&c, buf, size)) | |
3049 | die("corrupt commit"); | |
3050 | } | |
3051 | ||
3052 | static void check_tag(const void *buf, size_t size) | |
3053 | { | |
3054 | struct tag t; | |
3055 | memset(&t, 0, sizeof(t)); | |
3056 | if (parse_tag_buffer(&t, buf, size)) | |
3057 | die("corrupt tag"); | |
3058 | } | |
3059 | ||
43df4f86 | 3060 | static int index_mem(unsigned char *sha1, void *buf, size_t size, |
c4ce46fc JH |
3061 | enum object_type type, |
3062 | const char *path, unsigned flags) | |
e7332f96 | 3063 | { |
6c510bee | 3064 | int ret, re_allocated = 0; |
c4ce46fc | 3065 | int write_object = flags & HASH_WRITE_OBJECT; |
74400e71 | 3066 | |
7672db20 | 3067 | if (!type) |
edaec3fb | 3068 | type = OBJ_BLOB; |
6c510bee LT |
3069 | |
3070 | /* | |
3071 | * Convert blobs to git internal format | |
3072 | */ | |
43df4f86 | 3073 | if ((type == OBJ_BLOB) && path) { |
f285a2d7 | 3074 | struct strbuf nbuf = STRBUF_INIT; |
21e5ad50 | 3075 | if (convert_to_git(path, buf, size, &nbuf, |
5e12e78e | 3076 | write_object ? safe_crlf : SAFE_CRLF_FALSE)) { |
b315c5c0 | 3077 | buf = strbuf_detach(&nbuf, &size); |
6c510bee LT |
3078 | re_allocated = 1; |
3079 | } | |
3080 | } | |
c4ce46fc | 3081 | if (flags & HASH_FORMAT_CHECK) { |
c879daa2 NTND |
3082 | if (type == OBJ_TREE) |
3083 | check_tree(buf, size); | |
3084 | if (type == OBJ_COMMIT) | |
3085 | check_commit(buf, size); | |
3086 | if (type == OBJ_TAG) | |
3087 | check_tag(buf, size); | |
3088 | } | |
6c510bee | 3089 | |
7672db20 | 3090 | if (write_object) |
edaec3fb | 3091 | ret = write_sha1_file(buf, size, typename(type), sha1); |
abdc3fc8 | 3092 | else |
edaec3fb | 3093 | ret = hash_sha1_file(buf, size, typename(type), sha1); |
43df4f86 | 3094 | if (re_allocated) |
6c510bee | 3095 | free(buf); |
43df4f86 DP |
3096 | return ret; |
3097 | } | |
3098 | ||
7b41e1e1 JH |
3099 | static int index_pipe(unsigned char *sha1, int fd, enum object_type type, |
3100 | const char *path, unsigned flags) | |
3101 | { | |
3102 | struct strbuf sbuf = STRBUF_INIT; | |
3103 | int ret; | |
3104 | ||
3105 | if (strbuf_read(&sbuf, fd, 4096) >= 0) | |
3106 | ret = index_mem(sha1, sbuf.buf, sbuf.len, type, path, flags); | |
3107 | else | |
3108 | ret = -1; | |
3109 | strbuf_release(&sbuf); | |
3110 | return ret; | |
3111 | } | |
3112 | ||
ea68b0ce DP |
3113 | #define SMALL_FILE_SIZE (32*1024) |
3114 | ||
7b41e1e1 JH |
3115 | static int index_core(unsigned char *sha1, int fd, size_t size, |
3116 | enum object_type type, const char *path, | |
3117 | unsigned flags) | |
43df4f86 DP |
3118 | { |
3119 | int ret; | |
43df4f86 | 3120 | |
7b41e1e1 | 3121 | if (!size) { |
c4ce46fc | 3122 | ret = index_mem(sha1, NULL, size, type, path, flags); |
ea68b0ce DP |
3123 | } else if (size <= SMALL_FILE_SIZE) { |
3124 | char *buf = xmalloc(size); | |
3125 | if (size == read_in_full(fd, buf, size)) | |
c4ce46fc | 3126 | ret = index_mem(sha1, buf, size, type, path, flags); |
ea68b0ce DP |
3127 | else |
3128 | ret = error("short read %s", strerror(errno)); | |
3129 | free(buf); | |
08bda208 | 3130 | } else { |
43df4f86 | 3131 | void *buf = xmmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0); |
c4ce46fc | 3132 | ret = index_mem(sha1, buf, size, type, path, flags); |
aac17941 | 3133 | munmap(buf, size); |
08bda208 | 3134 | } |
7b41e1e1 JH |
3135 | return ret; |
3136 | } | |
3137 | ||
4dd1fbc7 | 3138 | /* |
568508e7 JH |
3139 | * This creates one packfile per large blob unless bulk-checkin |
3140 | * machinery is "plugged". | |
4dd1fbc7 JH |
3141 | * |
3142 | * This also bypasses the usual "convert-to-git" dance, and that is on | |
3143 | * purpose. We could write a streaming version of the converting | |
3144 | * functions and insert that before feeding the data to fast-import | |
4f22b101 JK |
3145 | * (or equivalent in-core API described above). However, that is |
3146 | * somewhat complicated, as we do not know the size of the filter | |
3147 | * result, which we need to know beforehand when writing a git object. | |
3148 | * Since the primary motivation for trying to stream from the working | |
3149 | * tree file and to avoid mmaping it in core is to deal with large | |
3150 | * binary blobs, they generally do not want to get any conversion, and | |
3151 | * callers should avoid this code path when filters are requested. | |
4dd1fbc7 JH |
3152 | */ |
3153 | static int index_stream(unsigned char *sha1, int fd, size_t size, | |
3154 | enum object_type type, const char *path, | |
3155 | unsigned flags) | |
3156 | { | |
568508e7 | 3157 | return index_bulk_checkin(sha1, fd, size, type, path, flags); |
4dd1fbc7 JH |
3158 | } |
3159 | ||
7b41e1e1 JH |
3160 | int index_fd(unsigned char *sha1, int fd, struct stat *st, |
3161 | enum object_type type, const char *path, unsigned flags) | |
3162 | { | |
3163 | int ret; | |
3164 | size_t size = xsize_t(st->st_size); | |
3165 | ||
3166 | if (!S_ISREG(st->st_mode)) | |
3167 | ret = index_pipe(sha1, fd, type, path, flags); | |
4f22b101 JK |
3168 | else if (size <= big_file_threshold || type != OBJ_BLOB || |
3169 | (path && would_convert_to_git(path, NULL, 0, 0))) | |
7b41e1e1 | 3170 | ret = index_core(sha1, fd, size, type, path, flags); |
4dd1fbc7 JH |
3171 | else |
3172 | ret = index_stream(sha1, fd, size, type, path, flags); | |
43df4f86 | 3173 | close(fd); |
aac17941 | 3174 | return ret; |
74400e71 | 3175 | } |
ec1fcc16 | 3176 | |
c4ce46fc | 3177 | int index_path(unsigned char *sha1, const char *path, struct stat *st, unsigned flags) |
ec1fcc16 JH |
3178 | { |
3179 | int fd; | |
b760d3aa | 3180 | struct strbuf sb = STRBUF_INIT; |
ec1fcc16 JH |
3181 | |
3182 | switch (st->st_mode & S_IFMT) { | |
3183 | case S_IFREG: | |
3184 | fd = open(path, O_RDONLY); | |
3185 | if (fd < 0) | |
3186 | return error("open(\"%s\"): %s", path, | |
3187 | strerror(errno)); | |
c4ce46fc | 3188 | if (index_fd(sha1, fd, st, OBJ_BLOB, path, flags) < 0) |
ec1fcc16 JH |
3189 | return error("%s: failed to insert into database", |
3190 | path); | |
3191 | break; | |
3192 | case S_IFLNK: | |
b760d3aa | 3193 | if (strbuf_readlink(&sb, path, st->st_size)) { |
ec1fcc16 | 3194 | char *errstr = strerror(errno); |
ec1fcc16 JH |
3195 | return error("readlink(\"%s\"): %s", path, |
3196 | errstr); | |
3197 | } | |
c4ce46fc | 3198 | if (!(flags & HASH_WRITE_OBJECT)) |
b760d3aa LT |
3199 | hash_sha1_file(sb.buf, sb.len, blob_type, sha1); |
3200 | else if (write_sha1_file(sb.buf, sb.len, blob_type, sha1)) | |
ec1fcc16 JH |
3201 | return error("%s: failed to insert into database", |
3202 | path); | |
b760d3aa | 3203 | strbuf_release(&sb); |
ec1fcc16 | 3204 | break; |
f35a6d3b LT |
3205 | case S_IFDIR: |
3206 | return resolve_gitlink_ref(path, "HEAD", sha1); | |
ec1fcc16 JH |
3207 | default: |
3208 | return error("%s: unsupported file type", path); | |
3209 | } | |
3210 | return 0; | |
3211 | } | |
a69e5429 JH |
3212 | |
3213 | int read_pack_header(int fd, struct pack_header *header) | |
3214 | { | |
c697ad14 HO |
3215 | if (read_in_full(fd, header, sizeof(*header)) < sizeof(*header)) |
3216 | /* "eof before pack header was fully read" */ | |
3217 | return PH_ERROR_EOF; | |
3218 | ||
a69e5429 JH |
3219 | if (header->hdr_signature != htonl(PACK_SIGNATURE)) |
3220 | /* "protocol error (pack signature mismatch detected)" */ | |
3221 | return PH_ERROR_PACK_SIGNATURE; | |
3222 | if (!pack_version_ok(header->hdr_version)) | |
3223 | /* "protocol error (pack version unsupported)" */ | |
3224 | return PH_ERROR_PROTOCOL; | |
3225 | return 0; | |
3226 | } | |
40d52ff7 JK |
3227 | |
3228 | void assert_sha1_type(const unsigned char *sha1, enum object_type expect) | |
3229 | { | |
3230 | enum object_type type = sha1_object_info(sha1, NULL); | |
3231 | if (type < 0) | |
3232 | die("%s is not a valid object", sha1_to_hex(sha1)); | |
3233 | if (type != expect) | |
3234 | die("%s is not a valid '%s' object", sha1_to_hex(sha1), | |
3235 | typename(expect)); | |
3236 | } |