]>
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" |
1f688557 | 10 | #include "delta.h" |
a733cb60 | 11 | #include "pack.h" |
8e440259 PE |
12 | #include "blob.h" |
13 | #include "commit.h" | |
4dd1fbc7 | 14 | #include "run-command.h" |
8e440259 PE |
15 | #include "tag.h" |
16 | #include "tree.h" | |
c879daa2 | 17 | #include "tree-walk.h" |
f35a6d3b | 18 | #include "refs.h" |
70f5d5d3 | 19 | #include "pack-revindex.h" |
628522ec | 20 | #include "sha1-lookup.h" |
568508e7 | 21 | #include "bulk-checkin.h" |
090ea126 | 22 | #include "streaming.h" |
0fcfd160 | 23 | |
144bde78 LT |
24 | #ifndef O_NOATIME |
25 | #if defined(__linux__) && (defined(__i386__) || defined(__PPC__)) | |
26 | #define O_NOATIME 01000000 | |
27 | #else | |
28 | #define O_NOATIME 0 | |
29 | #endif | |
30 | #endif | |
31 | ||
28bd70d8 JN |
32 | #define SZ_FMT PRIuMAX |
33 | static inline uintmax_t sz_fmt(size_t s) { return s; } | |
e05db0fd | 34 | |
96f1e58f | 35 | const unsigned char null_sha1[20]; |
88cd621d | 36 | |
c597ba80 NTND |
37 | /* |
38 | * This is meant to hold a *small* number of objects that you would | |
39 | * want read_sha1_file() to be able to return, but yet you do not want | |
40 | * to write them into the object store (e.g. a browse-only | |
41 | * application). | |
42 | */ | |
43 | static struct cached_object { | |
44 | unsigned char sha1[20]; | |
45 | enum object_type type; | |
46 | void *buf; | |
47 | unsigned long size; | |
48 | } *cached_objects; | |
49 | static int cached_object_nr, cached_object_alloc; | |
50 | ||
51 | static struct cached_object empty_tree = { | |
dab0d410 | 52 | EMPTY_TREE_SHA1_BIN_LITERAL, |
c597ba80 NTND |
53 | OBJ_TREE, |
54 | "", | |
55 | 0 | |
56 | }; | |
57 | ||
c01f51cc NTND |
58 | static struct packed_git *last_found_pack; |
59 | ||
c597ba80 NTND |
60 | static struct cached_object *find_cached_object(const unsigned char *sha1) |
61 | { | |
62 | int i; | |
63 | struct cached_object *co = cached_objects; | |
64 | ||
65 | for (i = 0; i < cached_object_nr; i++, co++) { | |
66 | if (!hashcmp(co->sha1, sha1)) | |
67 | return co; | |
68 | } | |
69 | if (!hashcmp(sha1, empty_tree.sha1)) | |
70 | return &empty_tree; | |
71 | return NULL; | |
72 | } | |
73 | ||
90a6464b JH |
74 | int mkdir_in_gitdir(const char *path) |
75 | { | |
76 | if (mkdir(path, 0777)) { | |
77 | int saved_errno = errno; | |
78 | struct stat st; | |
79 | struct strbuf sb = STRBUF_INIT; | |
80 | ||
81 | if (errno != EEXIST) | |
82 | return -1; | |
83 | /* | |
84 | * Are we looking at a path in a symlinked worktree | |
85 | * whose original repository does not yet have it? | |
86 | * e.g. .git/rr-cache pointing at its original | |
87 | * repository in which the user hasn't performed any | |
88 | * conflict resolution yet? | |
89 | */ | |
90 | if (lstat(path, &st) || !S_ISLNK(st.st_mode) || | |
91 | strbuf_readlink(&sb, path, st.st_size) || | |
92 | !is_absolute_path(sb.buf) || | |
93 | mkdir(sb.buf, 0777)) { | |
94 | strbuf_release(&sb); | |
95 | errno = saved_errno; | |
96 | return -1; | |
97 | } | |
98 | strbuf_release(&sb); | |
99 | } | |
100 | return adjust_shared_perm(path); | |
101 | } | |
102 | ||
b2cb9425 JH |
103 | int safe_create_leading_directories(char *path) |
104 | { | |
8385abfd | 105 | char *pos = path + offset_1st_component(path); |
67d42212 JR |
106 | struct stat st; |
107 | ||
b2cb9425 JH |
108 | while (pos) { |
109 | pos = strchr(pos, '/'); | |
110 | if (!pos) | |
111 | break; | |
5f0bdf50 JH |
112 | while (*++pos == '/') |
113 | ; | |
114 | if (!*pos) | |
115 | break; | |
116 | *--pos = '\0'; | |
67d42212 JR |
117 | if (!stat(path, &st)) { |
118 | /* path exists */ | |
119 | if (!S_ISDIR(st.st_mode)) { | |
b2cb9425 | 120 | *pos = '/'; |
67d42212 | 121 | return -3; |
b2cb9425 | 122 | } |
457f06d6 | 123 | } |
67d42212 JR |
124 | else if (mkdir(path, 0777)) { |
125 | *pos = '/'; | |
126 | return -1; | |
127 | } | |
457f06d6 JS |
128 | else if (adjust_shared_perm(path)) { |
129 | *pos = '/'; | |
130 | return -2; | |
131 | } | |
b2cb9425 JH |
132 | *pos++ = '/'; |
133 | } | |
134 | return 0; | |
135 | } | |
723c31fe | 136 | |
8e21d63b JK |
137 | int safe_create_leading_directories_const(const char *path) |
138 | { | |
139 | /* path points to cache entries, so xstrdup before messing with it */ | |
140 | char *buf = xstrdup(path); | |
141 | int result = safe_create_leading_directories(buf); | |
142 | free(buf); | |
143 | return result; | |
144 | } | |
145 | ||
ace1534d JH |
146 | static void fill_sha1_path(char *pathbuf, const unsigned char *sha1) |
147 | { | |
148 | int i; | |
149 | for (i = 0; i < 20; i++) { | |
150 | static char hex[] = "0123456789abcdef"; | |
151 | unsigned int val = sha1[i]; | |
152 | char *pos = pathbuf + i*2 + (i > 0); | |
153 | *pos++ = hex[val >> 4]; | |
154 | *pos = hex[val & 0xf]; | |
155 | } | |
156 | } | |
157 | ||
0fcfd160 LT |
158 | /* |
159 | * NOTE! This returns a statically allocated buffer, so you have to be | |
790296fd | 160 | * careful about using it. Do an "xstrdup()" if you need to save the |
0fcfd160 | 161 | * filename. |
ace1534d JH |
162 | * |
163 | * Also note that this returns the location for creating. Reading | |
164 | * SHA1 file can happen from any alternate directory listed in the | |
d19938ab | 165 | * DB_ENVIRONMENT environment variable if it is not found in |
ace1534d | 166 | * the primary object database. |
0fcfd160 LT |
167 | */ |
168 | char *sha1_file_name(const unsigned char *sha1) | |
169 | { | |
560fb6a1 JK |
170 | static char buf[PATH_MAX]; |
171 | const char *objdir; | |
172 | int len; | |
0fcfd160 | 173 | |
560fb6a1 JK |
174 | objdir = get_object_directory(); |
175 | len = strlen(objdir); | |
176 | ||
177 | /* '/' + sha1(2) + '/' + sha1(38) + '\0' */ | |
178 | if (len + 43 > PATH_MAX) | |
179 | die("insanely long object directory %s", objdir); | |
180 | memcpy(buf, objdir, len); | |
181 | buf[len] = '/'; | |
182 | buf[len+3] = '/'; | |
183 | buf[len+42] = '\0'; | |
184 | fill_sha1_path(buf + len + 1, sha1); | |
185 | return buf; | |
0fcfd160 LT |
186 | } |
187 | ||
633f43e1 | 188 | static char *sha1_get_pack_name(const unsigned char *sha1, |
6eec46bd | 189 | char **name, char **base, const char *which) |
bf592c50 DB |
190 | { |
191 | static const char hex[] = "0123456789abcdef"; | |
633f43e1 | 192 | char *buf; |
bf592c50 DB |
193 | int i; |
194 | ||
633f43e1 | 195 | if (!*base) { |
bf592c50 DB |
196 | const char *sha1_file_directory = get_object_directory(); |
197 | int len = strlen(sha1_file_directory); | |
633f43e1 | 198 | *base = xmalloc(len + 60); |
6eec46bd JH |
199 | sprintf(*base, "%s/pack/pack-1234567890123456789012345678901234567890.%s", |
200 | sha1_file_directory, which); | |
633f43e1 | 201 | *name = *base + len + 11; |
bf592c50 DB |
202 | } |
203 | ||
633f43e1 | 204 | buf = *name; |
bf592c50 DB |
205 | |
206 | for (i = 0; i < 20; i++) { | |
207 | unsigned int val = *sha1++; | |
208 | *buf++ = hex[val >> 4]; | |
209 | *buf++ = hex[val & 0xf]; | |
210 | } | |
a6080a0a | 211 | |
633f43e1 | 212 | return *base; |
bf592c50 DB |
213 | } |
214 | ||
633f43e1 | 215 | char *sha1_pack_name(const unsigned char *sha1) |
bf592c50 | 216 | { |
633f43e1 | 217 | static char *name, *base; |
bf592c50 | 218 | |
6eec46bd | 219 | return sha1_get_pack_name(sha1, &name, &base, "pack"); |
633f43e1 | 220 | } |
bf592c50 | 221 | |
633f43e1 HO |
222 | char *sha1_pack_index_name(const unsigned char *sha1) |
223 | { | |
224 | static char *name, *base; | |
a6080a0a | 225 | |
6eec46bd | 226 | return sha1_get_pack_name(sha1, &name, &base, "idx"); |
bf592c50 DB |
227 | } |
228 | ||
d5a63b99 JH |
229 | struct alternate_object_database *alt_odb_list; |
230 | static struct alternate_object_database **alt_odb_tail; | |
ace1534d | 231 | |
f4e51683 | 232 | static int git_open_noatime(const char *name); |
c2f493a4 | 233 | |
ddd5d056 JH |
234 | /* |
235 | * Prepare alternate object database registry. | |
d5a63b99 JH |
236 | * |
237 | * The variable alt_odb_list points at the list of struct | |
238 | * alternate_object_database. The elements on this list come from | |
239 | * non-empty elements from colon separated ALTERNATE_DB_ENVIRONMENT | |
240 | * environment variable, and $GIT_OBJECT_DIRECTORY/info/alternates, | |
1494e038 JH |
241 | * whose contents is similar to that environment variable but can be |
242 | * LF separated. Its base points at a statically allocated buffer that | |
d5a63b99 JH |
243 | * contains "/the/directory/corresponding/to/.git/objects/...", while |
244 | * its name points just after the slash at the end of ".git/objects/" | |
245 | * in the example above, and has enough space to hold 40-byte hex | |
246 | * SHA1, an extra slash for the first level indirection, and the | |
247 | * terminating NUL. | |
ddd5d056 | 248 | */ |
c2f493a4 | 249 | static int link_alt_odb_entry(const char * entry, int len, const char * relative_base, int depth) |
ace1534d | 250 | { |
1494e038 | 251 | const char *objdir = get_object_directory(); |
c2f493a4 MW |
252 | struct alternate_object_database *ent; |
253 | struct alternate_object_database *alt; | |
5bdf0a84 HW |
254 | int pfxlen, entlen; |
255 | struct strbuf pathbuf = STRBUF_INIT; | |
d5a63b99 | 256 | |
85dadc38 | 257 | if (!is_absolute_path(entry) && relative_base) { |
5bdf0a84 HW |
258 | strbuf_addstr(&pathbuf, real_path(relative_base)); |
259 | strbuf_addch(&pathbuf, '/'); | |
c2f493a4 | 260 | } |
5bdf0a84 | 261 | strbuf_add(&pathbuf, entry, len); |
c2f493a4 | 262 | |
5bdf0a84 HW |
263 | normalize_path_copy(pathbuf.buf, pathbuf.buf); |
264 | ||
265 | pfxlen = strlen(pathbuf.buf); | |
266 | ||
267 | /* | |
268 | * The trailing slash after the directory name is given by | |
269 | * this function at the end. Remove duplicates. | |
270 | */ | |
271 | while (pfxlen && pathbuf.buf[pfxlen-1] == '/') | |
272 | pfxlen -= 1; | |
273 | ||
274 | entlen = pfxlen + 43; /* '/' + 2 hex + '/' + 38 hex + NUL */ | |
275 | ent = xmalloc(sizeof(*ent) + entlen); | |
276 | memcpy(ent->base, pathbuf.buf, pfxlen); | |
277 | strbuf_release(&pathbuf); | |
c2f493a4 MW |
278 | |
279 | ent->name = ent->base + pfxlen + 1; | |
280 | ent->base[pfxlen + 3] = '/'; | |
281 | ent->base[pfxlen] = ent->base[entlen-1] = 0; | |
282 | ||
283 | /* Detect cases where alternate disappeared */ | |
90b4a71c | 284 | if (!is_directory(ent->base)) { |
c2f493a4 MW |
285 | error("object directory %s does not exist; " |
286 | "check .git/objects/info/alternates.", | |
287 | ent->base); | |
288 | free(ent); | |
289 | return -1; | |
290 | } | |
291 | ||
292 | /* Prevent the common mistake of listing the same | |
293 | * thing twice, or object directory itself. | |
294 | */ | |
295 | for (alt = alt_odb_list; alt; alt = alt->next) { | |
296 | if (!memcmp(ent->base, alt->base, pfxlen)) { | |
297 | free(ent); | |
298 | return -1; | |
299 | } | |
300 | } | |
cb2912c3 | 301 | if (!strcmp(ent->base, objdir)) { |
c2f493a4 MW |
302 | free(ent); |
303 | return -1; | |
304 | } | |
305 | ||
306 | /* add the alternate entry */ | |
307 | *alt_odb_tail = ent; | |
308 | alt_odb_tail = &(ent->next); | |
309 | ent->next = NULL; | |
310 | ||
311 | /* recursively add alternates */ | |
312 | read_info_alternates(ent->base, depth + 1); | |
313 | ||
314 | ent->base[pfxlen] = '/'; | |
315 | ||
316 | return 0; | |
317 | } | |
318 | ||
319 | static void link_alt_odb_entries(const char *alt, const char *ep, int sep, | |
320 | const char *relative_base, int depth) | |
321 | { | |
322 | const char *cp, *last; | |
323 | ||
324 | if (depth > 5) { | |
325 | error("%s: ignoring alternate object stores, nesting too deep.", | |
326 | relative_base); | |
327 | return; | |
328 | } | |
329 | ||
d5a63b99 | 330 | last = alt; |
9577e7e3 JH |
331 | while (last < ep) { |
332 | cp = last; | |
333 | if (cp < ep && *cp == '#') { | |
334 | while (cp < ep && *cp != sep) | |
335 | cp++; | |
336 | last = cp + 1; | |
337 | continue; | |
338 | } | |
c2f493a4 MW |
339 | while (cp < ep && *cp != sep) |
340 | cp++; | |
d5a63b99 | 341 | if (last != cp) { |
85dadc38 | 342 | if (!is_absolute_path(last) && depth) { |
c2f493a4 MW |
343 | error("%s: ignoring relative alternate object store %s", |
344 | relative_base, last); | |
345 | } else { | |
346 | link_alt_odb_entry(last, cp - last, | |
347 | relative_base, depth); | |
1494e038 | 348 | } |
d5a63b99 | 349 | } |
9577e7e3 | 350 | while (cp < ep && *cp == sep) |
d5a63b99 JH |
351 | cp++; |
352 | last = cp; | |
9577e7e3 | 353 | } |
d5a63b99 JH |
354 | } |
355 | ||
5e73633d | 356 | void read_info_alternates(const char * relative_base, int depth) |
d5a63b99 | 357 | { |
9577e7e3 | 358 | char *map; |
dc49cd76 | 359 | size_t mapsz; |
d5a63b99 | 360 | struct stat st; |
9cb18f56 JM |
361 | const char alt_file_name[] = "info/alternates"; |
362 | /* Given that relative_base is no longer than PATH_MAX, | |
363 | ensure that "path" has enough space to append "/", the | |
364 | file name, "info/alternates", and a trailing NUL. */ | |
365 | char path[PATH_MAX + 1 + sizeof alt_file_name]; | |
c2f493a4 | 366 | int fd; |
d5a63b99 | 367 | |
9cb18f56 | 368 | sprintf(path, "%s/%s", relative_base, alt_file_name); |
f4e51683 | 369 | fd = git_open_noatime(path); |
d5a63b99 JH |
370 | if (fd < 0) |
371 | return; | |
372 | if (fstat(fd, &st) || (st.st_size == 0)) { | |
373 | close(fd); | |
9a217f2a | 374 | return; |
ace1534d | 375 | } |
dc49cd76 SP |
376 | mapsz = xsize_t(st.st_size); |
377 | map = xmmap(NULL, mapsz, PROT_READ, MAP_PRIVATE, fd, 0); | |
d5a63b99 | 378 | close(fd); |
d5a63b99 | 379 | |
dc49cd76 | 380 | link_alt_odb_entries(map, map + mapsz, '\n', relative_base, depth); |
c2f493a4 | 381 | |
dc49cd76 | 382 | munmap(map, mapsz); |
ace1534d JH |
383 | } |
384 | ||
bef70b22 DB |
385 | void add_to_alternates_file(const char *reference) |
386 | { | |
387 | struct lock_file *lock = xcalloc(1, sizeof(struct lock_file)); | |
acd3b9ec | 388 | int fd = hold_lock_file_for_append(lock, git_path("objects/info/alternates"), LOCK_DIE_ON_ERROR); |
e6baf4a1 | 389 | char *alt = mkpath("%s\n", reference); |
bef70b22 DB |
390 | write_or_die(fd, alt, strlen(alt)); |
391 | if (commit_lock_file(lock)) | |
392 | die("could not close alternates file"); | |
393 | if (alt_odb_tail) | |
394 | link_alt_odb_entries(alt, alt + strlen(alt), '\n', NULL, 0); | |
395 | } | |
396 | ||
d79796bc JH |
397 | void foreach_alt_odb(alt_odb_fn fn, void *cb) |
398 | { | |
399 | struct alternate_object_database *ent; | |
400 | ||
401 | prepare_alt_odb(); | |
402 | for (ent = alt_odb_list; ent; ent = ent->next) | |
403 | if (fn(ent, cb)) | |
404 | return; | |
405 | } | |
406 | ||
c2f493a4 MW |
407 | void prepare_alt_odb(void) |
408 | { | |
554fe20d | 409 | const char *alt; |
c2f493a4 | 410 | |
7dc24aa5 SP |
411 | if (alt_odb_tail) |
412 | return; | |
413 | ||
c2f493a4 MW |
414 | alt = getenv(ALTERNATE_DB_ENVIRONMENT); |
415 | if (!alt) alt = ""; | |
416 | ||
c2f493a4 | 417 | alt_odb_tail = &alt_odb_list; |
80ba074f | 418 | link_alt_odb_entries(alt, alt + strlen(alt), PATH_SEP, NULL, 0); |
c2f493a4 MW |
419 | |
420 | read_info_alternates(get_object_directory(), 0); | |
421 | } | |
422 | ||
0f4dc14a | 423 | static int has_loose_object_local(const unsigned char *sha1) |
ace1534d | 424 | { |
ace1534d | 425 | char *name = sha1_file_name(sha1); |
0f4dc14a BC |
426 | return !access(name, F_OK); |
427 | } | |
ace1534d | 428 | |
0f4dc14a BC |
429 | int has_loose_object_nonlocal(const unsigned char *sha1) |
430 | { | |
431 | struct alternate_object_database *alt; | |
9a217f2a | 432 | prepare_alt_odb(); |
d5a63b99 | 433 | for (alt = alt_odb_list; alt; alt = alt->next) { |
0f4dc14a | 434 | fill_sha1_path(alt->name, sha1); |
c529d75a LT |
435 | if (!access(alt->base, F_OK)) |
436 | return 1; | |
ace1534d | 437 | } |
c529d75a | 438 | return 0; |
ace1534d JH |
439 | } |
440 | ||
0f4dc14a BC |
441 | static int has_loose_object(const unsigned char *sha1) |
442 | { | |
443 | return has_loose_object_local(sha1) || | |
444 | has_loose_object_nonlocal(sha1); | |
445 | } | |
446 | ||
60bb8b14 | 447 | static unsigned int pack_used_ctr; |
a53128b6 SP |
448 | static unsigned int pack_mmap_calls; |
449 | static unsigned int peak_pack_open_windows; | |
450 | static unsigned int pack_open_windows; | |
c7934306 SP |
451 | static unsigned int pack_open_fds; |
452 | static unsigned int pack_max_fds; | |
a53128b6 | 453 | static size_t peak_pack_mapped; |
60bb8b14 | 454 | static size_t pack_mapped; |
9a217f2a | 455 | struct packed_git *packed_git; |
1f688557 | 456 | |
b79d18c9 | 457 | void pack_report(void) |
a53128b6 SP |
458 | { |
459 | fprintf(stderr, | |
e05db0fd PR |
460 | "pack_report: getpagesize() = %10" SZ_FMT "\n" |
461 | "pack_report: core.packedGitWindowSize = %10" SZ_FMT "\n" | |
462 | "pack_report: core.packedGitLimit = %10" SZ_FMT "\n", | |
9e42d6a1 SP |
463 | sz_fmt(getpagesize()), |
464 | sz_fmt(packed_git_window_size), | |
465 | sz_fmt(packed_git_limit)); | |
a53128b6 SP |
466 | fprintf(stderr, |
467 | "pack_report: pack_used_ctr = %10u\n" | |
468 | "pack_report: pack_mmap_calls = %10u\n" | |
469 | "pack_report: pack_open_windows = %10u / %10u\n" | |
e05db0fd PR |
470 | "pack_report: pack_mapped = " |
471 | "%10" SZ_FMT " / %10" SZ_FMT "\n", | |
a53128b6 SP |
472 | pack_used_ctr, |
473 | pack_mmap_calls, | |
474 | pack_open_windows, peak_pack_open_windows, | |
9e42d6a1 | 475 | sz_fmt(pack_mapped), sz_fmt(peak_pack_mapped)); |
a53128b6 SP |
476 | } |
477 | ||
42873078 | 478 | static int check_packed_git_idx(const char *path, struct packed_git *p) |
1f688557 JH |
479 | { |
480 | void *idx_map; | |
42873078 | 481 | struct pack_idx_header *hdr; |
dc49cd76 | 482 | size_t idx_size; |
74e34e1f | 483 | uint32_t version, nr, i, *index; |
f4e51683 | 484 | int fd = git_open_noatime(path); |
1f688557 | 485 | struct stat st; |
42873078 | 486 | |
1f688557 JH |
487 | if (fd < 0) |
488 | return -1; | |
489 | if (fstat(fd, &st)) { | |
490 | close(fd); | |
491 | return -1; | |
492 | } | |
dc49cd76 | 493 | idx_size = xsize_t(st.st_size); |
2d88451b SP |
494 | if (idx_size < 4 * 256 + 20 + 20) { |
495 | close(fd); | |
496 | return error("index file %s is too small", path); | |
497 | } | |
c4712e45 | 498 | idx_map = xmmap(NULL, idx_size, PROT_READ, MAP_PRIVATE, fd, 0); |
1f688557 | 499 | close(fd); |
1f688557 | 500 | |
42873078 NP |
501 | hdr = idx_map; |
502 | if (hdr->idx_signature == htonl(PACK_IDX_SIGNATURE)) { | |
74e34e1f NP |
503 | version = ntohl(hdr->idx_version); |
504 | if (version < 2 || version > 2) { | |
505 | munmap(idx_map, idx_size); | |
6e1c2344 | 506 | return error("index file %s is version %"PRIu32 |
74e34e1f NP |
507 | " and is not supported by this binary" |
508 | " (try upgrading GIT to a newer version)", | |
509 | path, version); | |
510 | } | |
511 | } else | |
512 | version = 1; | |
df1b059d | 513 | |
1f688557 | 514 | nr = 0; |
42873078 | 515 | index = idx_map; |
74e34e1f NP |
516 | if (version > 1) |
517 | index += 2; /* skip index header */ | |
1f688557 | 518 | for (i = 0; i < 256; i++) { |
326bf396 | 519 | uint32_t n = ntohl(index[i]); |
2d88451b SP |
520 | if (n < nr) { |
521 | munmap(idx_map, idx_size); | |
df1b059d | 522 | return error("non-monotonic index %s", path); |
2d88451b | 523 | } |
1f688557 JH |
524 | nr = n; |
525 | } | |
526 | ||
74e34e1f NP |
527 | if (version == 1) { |
528 | /* | |
529 | * Total size: | |
530 | * - 256 index entries 4 bytes each | |
531 | * - 24-byte entries * nr (20-byte sha1 + 4-byte offset) | |
532 | * - 20-byte SHA1 of the packfile | |
533 | * - 20-byte SHA1 file checksum | |
534 | */ | |
535 | if (idx_size != 4*256 + nr * 24 + 20 + 20) { | |
536 | munmap(idx_map, idx_size); | |
eef427a0 | 537 | return error("wrong index v1 file size in %s", path); |
74e34e1f NP |
538 | } |
539 | } else if (version == 2) { | |
540 | /* | |
541 | * Minimum size: | |
542 | * - 8 bytes of header | |
543 | * - 256 index entries 4 bytes each | |
544 | * - 20-byte sha1 entry * nr | |
545 | * - 4-byte crc entry * nr | |
546 | * - 4-byte offset entry * nr | |
547 | * - 20-byte SHA1 of the packfile | |
548 | * - 20-byte SHA1 file checksum | |
549 | * And after the 4-byte offset table might be a | |
550 | * variable sized table containing 8-byte entries | |
551 | * for offsets larger than 2^31. | |
552 | */ | |
553 | unsigned long min_size = 8 + 4*256 + nr*(20 + 4 + 4) + 20 + 20; | |
1164f1e4 LT |
554 | unsigned long max_size = min_size; |
555 | if (nr) | |
556 | max_size += (nr - 1)*8; | |
557 | if (idx_size < min_size || idx_size > max_size) { | |
74e34e1f | 558 | munmap(idx_map, idx_size); |
eef427a0 | 559 | return error("wrong index v2 file size in %s", path); |
74e34e1f | 560 | } |
7109c889 JH |
561 | if (idx_size != min_size && |
562 | /* | |
563 | * make sure we can deal with large pack offsets. | |
564 | * 31-bit signed offset won't be enough, neither | |
565 | * 32-bit unsigned one will be. | |
566 | */ | |
567 | (sizeof(off_t) <= 4)) { | |
568 | munmap(idx_map, idx_size); | |
569 | return error("pack too large for current definition of off_t in %s", path); | |
74e34e1f | 570 | } |
2d88451b | 571 | } |
1f688557 | 572 | |
74e34e1f | 573 | p->index_version = version; |
42873078 NP |
574 | p->index_data = idx_map; |
575 | p->index_size = idx_size; | |
57059091 | 576 | p->num_objects = nr; |
1f688557 JH |
577 | return 0; |
578 | } | |
579 | ||
bc8e478a | 580 | int open_pack_index(struct packed_git *p) |
d079837e SP |
581 | { |
582 | char *idx_name; | |
583 | int ret; | |
584 | ||
585 | if (p->index_data) | |
586 | return 0; | |
587 | ||
588 | idx_name = xstrdup(p->pack_name); | |
589 | strcpy(idx_name + strlen(idx_name) - strlen(".pack"), ".idx"); | |
590 | ret = check_packed_git_idx(idx_name, p); | |
591 | free(idx_name); | |
592 | return ret; | |
593 | } | |
594 | ||
11daf39b SP |
595 | static void scan_windows(struct packed_git *p, |
596 | struct packed_git **lru_p, | |
597 | struct pack_window **lru_w, | |
598 | struct pack_window **lru_l) | |
1f688557 | 599 | { |
11daf39b SP |
600 | struct pack_window *w, *w_l; |
601 | ||
602 | for (w_l = NULL, w = p->windows; w; w = w->next) { | |
603 | if (!w->inuse_cnt) { | |
604 | if (!*lru_w || w->last_used < (*lru_w)->last_used) { | |
605 | *lru_p = p; | |
606 | *lru_w = w; | |
607 | *lru_l = w_l; | |
54044bf8 | 608 | } |
54044bf8 | 609 | } |
11daf39b | 610 | w_l = w; |
f9253394 | 611 | } |
11daf39b SP |
612 | } |
613 | ||
d1efefa4 | 614 | static int unuse_one_window(struct packed_git *current, int keep_fd) |
11daf39b SP |
615 | { |
616 | struct packed_git *p, *lru_p = NULL; | |
617 | struct pack_window *lru_w = NULL, *lru_l = NULL; | |
618 | ||
619 | if (current) | |
620 | scan_windows(current, &lru_p, &lru_w, &lru_l); | |
621 | for (p = packed_git; p; p = p->next) | |
622 | scan_windows(p, &lru_p, &lru_w, &lru_l); | |
54044bf8 SP |
623 | if (lru_p) { |
624 | munmap(lru_w->base, lru_w->len); | |
625 | pack_mapped -= lru_w->len; | |
626 | if (lru_l) | |
627 | lru_l->next = lru_w->next; | |
628 | else { | |
629 | lru_p->windows = lru_w->next; | |
d131b7af SP |
630 | if (!lru_p->windows && lru_p->pack_fd != -1 |
631 | && lru_p->pack_fd != keep_fd) { | |
54044bf8 | 632 | close(lru_p->pack_fd); |
c7934306 | 633 | pack_open_fds--; |
54044bf8 SP |
634 | lru_p->pack_fd = -1; |
635 | } | |
636 | } | |
637 | free(lru_w); | |
a53128b6 | 638 | pack_open_windows--; |
54044bf8 SP |
639 | return 1; |
640 | } | |
641 | return 0; | |
f9253394 JH |
642 | } |
643 | ||
d1efefa4 | 644 | void release_pack_memory(size_t need, int fd) |
97bfeb34 SP |
645 | { |
646 | size_t cur = pack_mapped; | |
d1efefa4 | 647 | while (need >= (cur - pack_mapped) && unuse_one_window(NULL, fd)) |
97bfeb34 SP |
648 | ; /* nothing */ |
649 | } | |
650 | ||
58ecbd5e JN |
651 | void *xmmap(void *start, size_t length, |
652 | int prot, int flags, int fd, off_t offset) | |
653 | { | |
654 | void *ret = mmap(start, length, prot, flags, fd, offset); | |
655 | if (ret == MAP_FAILED) { | |
656 | if (!length) | |
657 | return NULL; | |
658 | release_pack_memory(length, fd); | |
659 | ret = mmap(start, length, prot, flags, fd, offset); | |
660 | if (ret == MAP_FAILED) | |
661 | die_errno("Out of memory? mmap failed"); | |
662 | } | |
663 | return ret; | |
664 | } | |
665 | ||
c9ced051 SP |
666 | void close_pack_windows(struct packed_git *p) |
667 | { | |
668 | while (p->windows) { | |
669 | struct pack_window *w = p->windows; | |
670 | ||
671 | if (w->inuse_cnt) | |
672 | die("pack '%s' still has open windows to it", | |
673 | p->pack_name); | |
674 | munmap(w->base, w->len); | |
675 | pack_mapped -= w->len; | |
676 | pack_open_windows--; | |
677 | p->windows = w->next; | |
678 | free(w); | |
679 | } | |
680 | } | |
681 | ||
03e79c88 | 682 | void unuse_pack(struct pack_window **w_cursor) |
f9253394 | 683 | { |
03e79c88 SP |
684 | struct pack_window *w = *w_cursor; |
685 | if (w) { | |
686 | w->inuse_cnt--; | |
687 | *w_cursor = NULL; | |
688 | } | |
1f688557 JH |
689 | } |
690 | ||
fa5fc15d SP |
691 | void close_pack_index(struct packed_git *p) |
692 | { | |
693 | if (p->index_data) { | |
694 | munmap((void *)p->index_data, p->index_size); | |
695 | p->index_data = NULL; | |
696 | } | |
697 | } | |
698 | ||
c74faea1 NP |
699 | /* |
700 | * This is used by git-repack in case a newly created pack happens to | |
701 | * contain the same set of objects as an existing one. In that case | |
702 | * the resulting file might be different even if its name would be the | |
703 | * same. It is best to close any reference to the old pack before it is | |
704 | * replaced on disk. Of course no index pointers nor windows for given pack | |
705 | * must subsist at this point. If ever objects from this pack are requested | |
706 | * again, the new version of the pack will be reinitialized through | |
707 | * reprepare_packed_git(). | |
708 | */ | |
709 | void free_pack_by_name(const char *pack_name) | |
710 | { | |
711 | struct packed_git *p, **pp = &packed_git; | |
712 | ||
713 | while (*pp) { | |
714 | p = *pp; | |
715 | if (strcmp(pack_name, p->pack_name) == 0) { | |
fa3a0c94 | 716 | clear_delta_base_cache(); |
c74faea1 | 717 | close_pack_windows(p); |
c7934306 | 718 | if (p->pack_fd != -1) { |
c74faea1 | 719 | close(p->pack_fd); |
c7934306 SP |
720 | pack_open_fds--; |
721 | } | |
fa5fc15d | 722 | close_pack_index(p); |
c74faea1 NP |
723 | free(p->bad_object_sha1); |
724 | *pp = p->next; | |
c01f51cc NTND |
725 | if (last_found_pack == p) |
726 | last_found_pack = NULL; | |
c74faea1 NP |
727 | free(p); |
728 | return; | |
729 | } | |
730 | pp = &p->next; | |
731 | } | |
732 | } | |
733 | ||
a0788266 JS |
734 | static unsigned int get_max_fd_limit(void) |
735 | { | |
736 | #ifdef RLIMIT_NOFILE | |
737 | struct rlimit lim; | |
738 | ||
739 | if (getrlimit(RLIMIT_NOFILE, &lim)) | |
740 | die_errno("cannot get RLIMIT_NOFILE"); | |
741 | ||
742 | return lim.rlim_cur; | |
743 | #elif defined(_SC_OPEN_MAX) | |
744 | return sysconf(_SC_OPEN_MAX); | |
745 | #elif defined(OPEN_MAX) | |
746 | return OPEN_MAX; | |
747 | #else | |
748 | return 1; /* see the caller ;-) */ | |
749 | #endif | |
750 | } | |
751 | ||
3cf8b462 SP |
752 | /* |
753 | * Do not call this directly as this leaks p->pack_fd on error return; | |
754 | * call open_packed_git() instead. | |
755 | */ | |
756 | static int open_packed_git_1(struct packed_git *p) | |
1f688557 | 757 | { |
9bc879c1 SP |
758 | struct stat st; |
759 | struct pack_header hdr; | |
760 | unsigned char sha1[20]; | |
761 | unsigned char *idx_sha1; | |
2c039da8 | 762 | long fd_flag; |
9bc879c1 | 763 | |
d079837e SP |
764 | if (!p->index_data && open_pack_index(p)) |
765 | return error("packfile %s index unavailable", p->pack_name); | |
766 | ||
c7934306 | 767 | if (!pack_max_fds) { |
a0788266 | 768 | unsigned int max_fds = get_max_fd_limit(); |
c7934306 SP |
769 | |
770 | /* Save 3 for stdin/stdout/stderr, 22 for work */ | |
771 | if (25 < max_fds) | |
772 | pack_max_fds = max_fds - 25; | |
773 | else | |
774 | pack_max_fds = 1; | |
775 | } | |
776 | ||
777 | while (pack_max_fds <= pack_open_fds && unuse_one_window(NULL, -1)) | |
778 | ; /* nothing */ | |
779 | ||
f4e51683 | 780 | p->pack_fd = git_open_noatime(p->pack_name); |
9bc879c1 | 781 | if (p->pack_fd < 0 || fstat(p->pack_fd, &st)) |
072db278 | 782 | return -1; |
c7934306 | 783 | pack_open_fds++; |
9bc879c1 SP |
784 | |
785 | /* If we created the struct before we had the pack we lack size. */ | |
bf592c50 | 786 | if (!p->pack_size) { |
bf592c50 | 787 | if (!S_ISREG(st.st_mode)) |
072db278 | 788 | return error("packfile %s not a regular file", p->pack_name); |
bf592c50 | 789 | p->pack_size = st.st_size; |
9bc879c1 | 790 | } else if (p->pack_size != st.st_size) |
072db278 | 791 | return error("packfile %s size changed", p->pack_name); |
9bc879c1 | 792 | |
2c039da8 JH |
793 | /* We leave these file descriptors open with sliding mmap; |
794 | * there is no point keeping them open across exec(), though. | |
795 | */ | |
796 | fd_flag = fcntl(p->pack_fd, F_GETFD, 0); | |
797 | if (fd_flag < 0) | |
072db278 | 798 | return error("cannot determine file descriptor flags"); |
2c039da8 JH |
799 | fd_flag |= FD_CLOEXEC; |
800 | if (fcntl(p->pack_fd, F_SETFD, fd_flag) == -1) | |
072db278 | 801 | return error("cannot set FD_CLOEXEC"); |
2c039da8 | 802 | |
9bc879c1 | 803 | /* Verify we recognize this pack file format. */ |
e6e2bd62 | 804 | if (read_in_full(p->pack_fd, &hdr, sizeof(hdr)) != sizeof(hdr)) |
072db278 | 805 | return error("file %s is far too short to be a packfile", p->pack_name); |
9bc879c1 | 806 | if (hdr.hdr_signature != htonl(PACK_SIGNATURE)) |
072db278 | 807 | return error("file %s is not a GIT packfile", p->pack_name); |
9bc879c1 | 808 | if (!pack_version_ok(hdr.hdr_version)) |
6e1c2344 RJ |
809 | return error("packfile %s is version %"PRIu32" and not" |
810 | " supported (try upgrading GIT to a newer version)", | |
9bc879c1 SP |
811 | p->pack_name, ntohl(hdr.hdr_version)); |
812 | ||
813 | /* Verify the pack matches its index. */ | |
57059091 | 814 | if (p->num_objects != ntohl(hdr.hdr_entries)) |
6e1c2344 RJ |
815 | return error("packfile %s claims to have %"PRIu32" objects" |
816 | " while index indicates %"PRIu32" objects", | |
57059091 NP |
817 | p->pack_name, ntohl(hdr.hdr_entries), |
818 | p->num_objects); | |
9bc879c1 | 819 | if (lseek(p->pack_fd, p->pack_size - sizeof(sha1), SEEK_SET) == -1) |
072db278 | 820 | return error("end of packfile %s is unavailable", p->pack_name); |
e6e2bd62 | 821 | if (read_in_full(p->pack_fd, sha1, sizeof(sha1)) != sizeof(sha1)) |
072db278 | 822 | return error("packfile %s signature is unavailable", p->pack_name); |
42873078 | 823 | idx_sha1 = ((unsigned char *)p->index_data) + p->index_size - 40; |
9bc879c1 | 824 | if (hashcmp(sha1, idx_sha1)) |
072db278 SP |
825 | return error("packfile %s does not match index", p->pack_name); |
826 | return 0; | |
9bc879c1 SP |
827 | } |
828 | ||
3cf8b462 SP |
829 | static int open_packed_git(struct packed_git *p) |
830 | { | |
831 | if (!open_packed_git_1(p)) | |
832 | return 0; | |
833 | if (p->pack_fd != -1) { | |
834 | close(p->pack_fd); | |
c7934306 | 835 | pack_open_fds--; |
3cf8b462 SP |
836 | p->pack_fd = -1; |
837 | } | |
838 | return -1; | |
839 | } | |
840 | ||
c4001d92 | 841 | static int in_window(struct pack_window *win, off_t offset) |
60bb8b14 SP |
842 | { |
843 | /* We must promise at least 20 bytes (one hash) after the | |
844 | * offset is available from this window, otherwise the offset | |
845 | * is not actually in this window and a different window (which | |
846 | * has that one hash excess) must be used. This is to support | |
847 | * the object header and delta base parsing routines below. | |
848 | */ | |
849 | off_t win_off = win->offset; | |
850 | return win_off <= offset | |
851 | && (offset + 20) <= (win_off + win->len); | |
852 | } | |
853 | ||
4b25d091 | 854 | unsigned char *use_pack(struct packed_git *p, |
03e79c88 | 855 | struct pack_window **w_cursor, |
c4001d92 | 856 | off_t offset, |
ef49a7a0 | 857 | unsigned long *left) |
9bc879c1 | 858 | { |
60bb8b14 | 859 | struct pack_window *win = *w_cursor; |
03e79c88 | 860 | |
a9d98a14 | 861 | /* Since packfiles end in a hash of their content and it's |
60bb8b14 SP |
862 | * pointless to ask for an offset into the middle of that |
863 | * hash, and the in_window function above wouldn't match | |
864 | * don't allow an offset too close to the end of the file. | |
865 | */ | |
d131b7af SP |
866 | if (!p->pack_size && p->pack_fd == -1 && open_packed_git(p)) |
867 | die("packfile %s cannot be accessed", p->pack_name); | |
60bb8b14 SP |
868 | if (offset > (p->pack_size - 20)) |
869 | die("offset beyond end of packfile (truncated pack?)"); | |
870 | ||
871 | if (!win || !in_window(win, offset)) { | |
872 | if (win) | |
873 | win->inuse_cnt--; | |
874 | for (win = p->windows; win; win = win->next) { | |
875 | if (in_window(win, offset)) | |
876 | break; | |
877 | } | |
878 | if (!win) { | |
78a28df9 | 879 | size_t window_align = packed_git_window_size / 2; |
dc49cd76 | 880 | off_t len; |
d131b7af SP |
881 | |
882 | if (p->pack_fd == -1 && open_packed_git(p)) | |
883 | die("packfile %s cannot be accessed", p->pack_name); | |
884 | ||
60bb8b14 | 885 | win = xcalloc(1, sizeof(*win)); |
78a28df9 | 886 | win->offset = (offset / window_align) * window_align; |
dc49cd76 SP |
887 | len = p->pack_size - win->offset; |
888 | if (len > packed_git_window_size) | |
889 | len = packed_git_window_size; | |
890 | win->len = (size_t)len; | |
60bb8b14 | 891 | pack_mapped += win->len; |
11daf39b | 892 | while (packed_git_limit < pack_mapped |
d1efefa4 | 893 | && unuse_one_window(p, p->pack_fd)) |
60bb8b14 | 894 | ; /* nothing */ |
c4712e45 | 895 | win->base = xmmap(NULL, win->len, |
60bb8b14 SP |
896 | PROT_READ, MAP_PRIVATE, |
897 | p->pack_fd, win->offset); | |
898 | if (win->base == MAP_FAILED) | |
73b4e4be SP |
899 | die("packfile %s cannot be mapped: %s", |
900 | p->pack_name, | |
901 | strerror(errno)); | |
d131b7af SP |
902 | if (!win->offset && win->len == p->pack_size |
903 | && !p->do_not_close) { | |
904 | close(p->pack_fd); | |
905 | pack_open_fds--; | |
906 | p->pack_fd = -1; | |
907 | } | |
a53128b6 SP |
908 | pack_mmap_calls++; |
909 | pack_open_windows++; | |
910 | if (pack_mapped > peak_pack_mapped) | |
911 | peak_pack_mapped = pack_mapped; | |
912 | if (pack_open_windows > peak_pack_open_windows) | |
913 | peak_pack_open_windows = pack_open_windows; | |
60bb8b14 SP |
914 | win->next = p->windows; |
915 | p->windows = win; | |
916 | } | |
1f688557 | 917 | } |
03e79c88 SP |
918 | if (win != *w_cursor) { |
919 | win->last_used = pack_used_ctr++; | |
920 | win->inuse_cnt++; | |
921 | *w_cursor = win; | |
922 | } | |
60bb8b14 | 923 | offset -= win->offset; |
03e79c88 | 924 | if (left) |
dc49cd76 | 925 | *left = win->len - xsize_t(offset); |
03e79c88 | 926 | return win->base + offset; |
1f688557 JH |
927 | } |
928 | ||
27d69a46 NP |
929 | static struct packed_git *alloc_packed_git(int extra) |
930 | { | |
931 | struct packed_git *p = xmalloc(sizeof(*p) + extra); | |
932 | memset(p, 0, sizeof(*p)); | |
933 | p->pack_fd = -1; | |
934 | return p; | |
935 | } | |
936 | ||
e0500293 JN |
937 | static void try_to_free_pack_memory(size_t size) |
938 | { | |
939 | release_pack_memory(size, -1); | |
940 | } | |
941 | ||
42873078 | 942 | struct packed_git *add_packed_git(const char *path, int path_len, int local) |
1f688557 | 943 | { |
e0500293 | 944 | static int have_set_try_to_free_routine; |
1f688557 | 945 | struct stat st; |
27d69a46 | 946 | struct packed_git *p = alloc_packed_git(path_len + 2); |
1f688557 | 947 | |
e0500293 JN |
948 | if (!have_set_try_to_free_routine) { |
949 | have_set_try_to_free_routine = 1; | |
950 | set_try_to_free_routine(try_to_free_pack_memory); | |
951 | } | |
952 | ||
42873078 NP |
953 | /* |
954 | * Make sure a corresponding .pack file exists and that | |
955 | * the index looks sane. | |
956 | */ | |
957 | path_len -= strlen(".idx"); | |
27d69a46 NP |
958 | if (path_len < 1) { |
959 | free(p); | |
1f688557 | 960 | return NULL; |
27d69a46 | 961 | } |
42873078 | 962 | memcpy(p->pack_name, path, path_len); |
8d25931d BC |
963 | |
964 | strcpy(p->pack_name + path_len, ".keep"); | |
965 | if (!access(p->pack_name, F_OK)) | |
966 | p->pack_keep = 1; | |
967 | ||
42873078 | 968 | strcpy(p->pack_name + path_len, ".pack"); |
d079837e | 969 | if (stat(p->pack_name, &st) || !S_ISREG(st.st_mode)) { |
42873078 | 970 | free(p); |
1f688557 JH |
971 | return NULL; |
972 | } | |
42873078 | 973 | |
1f688557 JH |
974 | /* ok, it looks sane as far as we can check without |
975 | * actually mapping the pack file. | |
976 | */ | |
1f688557 | 977 | p->pack_size = st.st_size; |
9d835df2 | 978 | p->pack_local = local; |
b867092f | 979 | p->mtime = st.st_mtime; |
42873078 NP |
980 | if (path_len < 40 || get_sha1_hex(path + path_len - 40, p->sha1)) |
981 | hashclr(p->sha1); | |
1f688557 JH |
982 | return p; |
983 | } | |
984 | ||
7b64469a | 985 | struct packed_git *parse_pack_index(unsigned char *sha1, const char *idx_path) |
c508df5e | 986 | { |
42873078 | 987 | const char *path = sha1_pack_name(sha1); |
27d69a46 | 988 | struct packed_git *p = alloc_packed_git(strlen(path) + 1); |
bf592c50 | 989 | |
27d69a46 NP |
990 | strcpy(p->pack_name, path); |
991 | hashcpy(p->sha1, sha1); | |
42873078 NP |
992 | if (check_packed_git_idx(idx_path, p)) { |
993 | free(p); | |
bf592c50 | 994 | return NULL; |
42873078 | 995 | } |
bf592c50 | 996 | |
bf592c50 DB |
997 | return p; |
998 | } | |
999 | ||
1000 | void install_packed_git(struct packed_git *pack) | |
1001 | { | |
c7934306 SP |
1002 | if (pack->pack_fd != -1) |
1003 | pack_open_fds++; | |
1004 | ||
bf592c50 DB |
1005 | pack->next = packed_git; |
1006 | packed_git = pack; | |
1007 | } | |
1008 | ||
9d835df2 | 1009 | static void prepare_packed_git_one(char *objdir, int local) |
1f688557 | 1010 | { |
9cb18f56 JM |
1011 | /* Ensure that this buffer is large enough so that we can |
1012 | append "/pack/" without clobbering the stack even if | |
1013 | strlen(objdir) were PATH_MAX. */ | |
1014 | char path[PATH_MAX + 1 + 4 + 1 + 1]; | |
1f688557 JH |
1015 | int len; |
1016 | DIR *dir; | |
1017 | struct dirent *de; | |
1018 | ||
1019 | sprintf(path, "%s/pack", objdir); | |
1020 | len = strlen(path); | |
1021 | dir = opendir(path); | |
b5b16990 | 1022 | if (!dir) { |
26125f6b | 1023 | if (errno != ENOENT) |
bd2afde8 | 1024 | error("unable to open object pack directory: %s: %s", |
26125f6b | 1025 | path, strerror(errno)); |
1f688557 | 1026 | return; |
b5b16990 | 1027 | } |
1f688557 JH |
1028 | path[len++] = '/'; |
1029 | while ((de = readdir(dir)) != NULL) { | |
1030 | int namelen = strlen(de->d_name); | |
1031 | struct packed_git *p; | |
1032 | ||
5bb1cda5 | 1033 | if (!has_extension(de->d_name, ".idx")) |
1f688557 JH |
1034 | continue; |
1035 | ||
9cb18f56 JM |
1036 | if (len + namelen + 1 > sizeof(path)) |
1037 | continue; | |
1038 | ||
54a15a8d | 1039 | /* Don't reopen a pack we already have. */ |
1f688557 | 1040 | strcpy(path + len, de->d_name); |
86f7780c JK |
1041 | for (p = packed_git; p; p = p->next) { |
1042 | if (!memcmp(path, p->pack_name, len + namelen - 4)) | |
1043 | break; | |
1044 | } | |
1045 | if (p) | |
1046 | continue; | |
54a15a8d SP |
1047 | /* See if it really is a valid .idx file with corresponding |
1048 | * .pack file that we can map. | |
1049 | */ | |
9d835df2 | 1050 | p = add_packed_git(path, len + namelen, local); |
1f688557 JH |
1051 | if (!p) |
1052 | continue; | |
625e9421 | 1053 | install_packed_git(p); |
1f688557 | 1054 | } |
5b35bcd5 | 1055 | closedir(dir); |
1f688557 JH |
1056 | } |
1057 | ||
b867092f JH |
1058 | static int sort_pack(const void *a_, const void *b_) |
1059 | { | |
1060 | struct packed_git *a = *((struct packed_git **)a_); | |
1061 | struct packed_git *b = *((struct packed_git **)b_); | |
1062 | int st; | |
1063 | ||
1064 | /* | |
1065 | * Local packs tend to contain objects specific to our | |
1066 | * variant of the project than remote ones. In addition, | |
1067 | * remote ones could be on a network mounted filesystem. | |
1068 | * Favor local ones for these reasons. | |
1069 | */ | |
1070 | st = a->pack_local - b->pack_local; | |
1071 | if (st) | |
1072 | return -st; | |
1073 | ||
1074 | /* | |
1075 | * Younger packs tend to contain more recent objects, | |
1076 | * and more recent objects tend to get accessed more | |
1077 | * often. | |
1078 | */ | |
1079 | if (a->mtime < b->mtime) | |
1080 | return 1; | |
1081 | else if (a->mtime == b->mtime) | |
1082 | return 0; | |
1083 | return -1; | |
1084 | } | |
1085 | ||
1086 | static void rearrange_packed_git(void) | |
1087 | { | |
1088 | struct packed_git **ary, *p; | |
1089 | int i, n; | |
1090 | ||
1091 | for (n = 0, p = packed_git; p; p = p->next) | |
1092 | n++; | |
1093 | if (n < 2) | |
1094 | return; | |
1095 | ||
1096 | /* prepare an array of packed_git for easier sorting */ | |
1097 | ary = xcalloc(n, sizeof(struct packed_git *)); | |
1098 | for (n = 0, p = packed_git; p; p = p->next) | |
1099 | ary[n++] = p; | |
1100 | ||
1101 | qsort(ary, n, sizeof(struct packed_git *), sort_pack); | |
1102 | ||
1103 | /* link them back again */ | |
1104 | for (i = 0; i < n - 1; i++) | |
1105 | ary[i]->next = ary[i + 1]; | |
1106 | ary[n - 1]->next = NULL; | |
1107 | packed_git = ary[0]; | |
1108 | ||
1109 | free(ary); | |
1110 | } | |
1111 | ||
637cdd9d | 1112 | static int prepare_packed_git_run_once = 0; |
9a217f2a | 1113 | void prepare_packed_git(void) |
1f688557 | 1114 | { |
d5a63b99 | 1115 | struct alternate_object_database *alt; |
1f688557 | 1116 | |
637cdd9d | 1117 | if (prepare_packed_git_run_once) |
1f688557 | 1118 | return; |
9d835df2 | 1119 | prepare_packed_git_one(get_object_directory(), 1); |
9a217f2a | 1120 | prepare_alt_odb(); |
d5a63b99 | 1121 | for (alt = alt_odb_list; alt; alt = alt->next) { |
1494e038 | 1122 | alt->name[-1] = 0; |
9d835df2 | 1123 | prepare_packed_git_one(alt->base, 0); |
1494e038 | 1124 | alt->name[-1] = '/'; |
1f688557 | 1125 | } |
b867092f | 1126 | rearrange_packed_git(); |
637cdd9d JK |
1127 | prepare_packed_git_run_once = 1; |
1128 | } | |
1129 | ||
fc04c412 | 1130 | void reprepare_packed_git(void) |
637cdd9d | 1131 | { |
4b480c67 | 1132 | discard_revindex(); |
637cdd9d JK |
1133 | prepare_packed_git_run_once = 0; |
1134 | prepare_packed_git(); | |
1f688557 JH |
1135 | } |
1136 | ||
8eca0b47 NP |
1137 | static void mark_bad_packed_object(struct packed_git *p, |
1138 | const unsigned char *sha1) | |
1139 | { | |
1140 | unsigned i; | |
1141 | for (i = 0; i < p->num_bad_objects; i++) | |
1142 | if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i)) | |
1143 | return; | |
1144 | p->bad_object_sha1 = xrealloc(p->bad_object_sha1, 20 * (p->num_bad_objects + 1)); | |
1145 | hashcpy(p->bad_object_sha1 + 20 * p->num_bad_objects, sha1); | |
1146 | p->num_bad_objects++; | |
1147 | } | |
1148 | ||
b6c4cecc | 1149 | static const struct packed_git *has_packed_and_bad(const unsigned char *sha1) |
ac939109 NP |
1150 | { |
1151 | struct packed_git *p; | |
1152 | unsigned i; | |
1153 | ||
1154 | for (p = packed_git; p; p = p->next) | |
1155 | for (i = 0; i < p->num_bad_objects; i++) | |
1156 | if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i)) | |
b6c4cecc JH |
1157 | return p; |
1158 | return NULL; | |
ac939109 NP |
1159 | } |
1160 | ||
090ea126 NTND |
1161 | /* |
1162 | * With an in-core object data in "map", rehash it to make sure the | |
1163 | * object name actually matches "sha1" to detect object corruption. | |
1164 | * With "map" == NULL, try reading the object named with "sha1" using | |
1165 | * the streaming interface and rehash it to do the same. | |
1166 | */ | |
1167 | int check_sha1_signature(const unsigned char *sha1, void *map, | |
1168 | unsigned long size, const char *type) | |
0fcfd160 LT |
1169 | { |
1170 | unsigned char real_sha1[20]; | |
090ea126 NTND |
1171 | enum object_type obj_type; |
1172 | struct git_istream *st; | |
1173 | git_SHA_CTX c; | |
1174 | char hdr[32]; | |
1175 | int hdrlen; | |
1176 | ||
1177 | if (map) { | |
1178 | hash_sha1_file(map, size, type, real_sha1); | |
1179 | return hashcmp(sha1, real_sha1) ? -1 : 0; | |
1180 | } | |
1181 | ||
1182 | st = open_istream(sha1, &obj_type, &size, NULL); | |
1183 | if (!st) | |
1184 | return -1; | |
1185 | ||
1186 | /* Generate the header */ | |
1187 | hdrlen = sprintf(hdr, "%s %lu", typename(obj_type), size) + 1; | |
1188 | ||
1189 | /* Sha1.. */ | |
1190 | git_SHA1_Init(&c); | |
1191 | git_SHA1_Update(&c, hdr, hdrlen); | |
1192 | for (;;) { | |
1193 | char buf[1024 * 16]; | |
1194 | ssize_t readlen = read_istream(st, buf, sizeof(buf)); | |
1195 | ||
1196 | if (!readlen) | |
1197 | break; | |
1198 | git_SHA1_Update(&c, buf, readlen); | |
1199 | } | |
1200 | git_SHA1_Final(real_sha1, &c); | |
1201 | close_istream(st); | |
a89fccd2 | 1202 | return hashcmp(sha1, real_sha1) ? -1 : 0; |
0fcfd160 LT |
1203 | } |
1204 | ||
f4e51683 | 1205 | static int git_open_noatime(const char *name) |
44d1c19e LT |
1206 | { |
1207 | static int sha1_file_open_flag = O_NOATIME; | |
44d1c19e | 1208 | |
f2e872aa SP |
1209 | for (;;) { |
1210 | int fd = open(name, O_RDONLY | sha1_file_open_flag); | |
44d1c19e | 1211 | if (fd >= 0) |
f2e872aa SP |
1212 | return fd; |
1213 | ||
f2e872aa SP |
1214 | /* Might the failure be due to O_NOATIME? */ |
1215 | if (errno != ENOENT && sha1_file_open_flag) { | |
44d1c19e | 1216 | sha1_file_open_flag = 0; |
f2e872aa SP |
1217 | continue; |
1218 | } | |
1219 | ||
1220 | return -1; | |
44d1c19e | 1221 | } |
44d1c19e LT |
1222 | } |
1223 | ||
1224 | static int open_sha1_file(const unsigned char *sha1) | |
1225 | { | |
1226 | int fd; | |
1227 | char *name = sha1_file_name(sha1); | |
1228 | struct alternate_object_database *alt; | |
1229 | ||
f4e51683 | 1230 | fd = git_open_noatime(name); |
44d1c19e LT |
1231 | if (fd >= 0) |
1232 | return fd; | |
1233 | ||
1234 | prepare_alt_odb(); | |
1235 | errno = ENOENT; | |
1236 | for (alt = alt_odb_list; alt; alt = alt->next) { | |
1237 | name = alt->name; | |
1238 | fill_sha1_path(name, sha1); | |
f4e51683 | 1239 | fd = git_open_noatime(alt->base); |
44d1c19e LT |
1240 | if (fd >= 0) |
1241 | return fd; | |
1242 | } | |
1243 | return -1; | |
1244 | } | |
1245 | ||
f0270efd | 1246 | void *map_sha1_file(const unsigned char *sha1, unsigned long *size) |
0fcfd160 | 1247 | { |
0fcfd160 | 1248 | void *map; |
144bde78 | 1249 | int fd; |
ace1534d | 1250 | |
44d1c19e LT |
1251 | fd = open_sha1_file(sha1); |
1252 | map = NULL; | |
1253 | if (fd >= 0) { | |
1254 | struct stat st; | |
0fcfd160 | 1255 | |
44d1c19e LT |
1256 | if (!fstat(fd, &st)) { |
1257 | *size = xsize_t(st.st_size); | |
33e42de0 MM |
1258 | if (!*size) { |
1259 | /* mmap() is forbidden on empty files */ | |
1260 | error("object file %s is empty", sha1_file_name(sha1)); | |
1261 | return NULL; | |
1262 | } | |
44d1c19e | 1263 | map = xmmap(NULL, *size, PROT_READ, MAP_PRIVATE, fd, 0); |
144bde78 | 1264 | } |
44d1c19e | 1265 | close(fd); |
0fcfd160 | 1266 | } |
0fcfd160 LT |
1267 | return map; |
1268 | } | |
1269 | ||
cc5c54e7 JH |
1270 | /* |
1271 | * There used to be a second loose object header format which | |
1272 | * was meant to mimic the in-pack format, allowing for direct | |
1273 | * copy of the object data. This format turned up not to be | |
1274 | * really worth it and we no longer write loose objects in that | |
1275 | * format. | |
1276 | */ | |
1277 | static int experimental_loose_object(unsigned char *map) | |
bb6b8e4f JH |
1278 | { |
1279 | unsigned int word; | |
1280 | ||
1281 | /* | |
7f684a2a RT |
1282 | * We must determine if the buffer contains the standard |
1283 | * zlib-deflated stream or the experimental format based | |
1284 | * on the in-pack object format. Compare the header byte | |
1285 | * for each format: | |
1286 | * | |
1287 | * RFC1950 zlib w/ deflate : 0www1000 : 0 <= www <= 7 | |
1288 | * Experimental pack-based : Stttssss : ttt = 1,2,3,4 | |
1289 | * | |
1290 | * If bit 7 is clear and bits 0-3 equal 8, the buffer MUST be | |
1291 | * in standard loose-object format, UNLESS it is a Git-pack | |
1292 | * format object *exactly* 8 bytes in size when inflated. | |
1293 | * | |
1294 | * However, RFC1950 also specifies that the 1st 16-bit word | |
1295 | * must be divisible by 31 - this checksum tells us our buffer | |
1296 | * is in the standard format, giving a false positive only if | |
1297 | * the 1st word of the Git-pack format object happens to be | |
1298 | * divisible by 31, ie: | |
1299 | * ((byte0 * 256) + byte1) % 31 = 0 | |
1300 | * => 0ttt10000www1000 % 31 = 0 | |
1301 | * | |
1302 | * As it happens, this case can only arise for www=3 & ttt=1 | |
1303 | * - ie, a Commit object, which would have to be 8 bytes in | |
1304 | * size. As no Commit can be that small, we find that the | |
1305 | * combination of these two criteria (bitmask & checksum) | |
1306 | * can always correctly determine the buffer format. | |
bb6b8e4f JH |
1307 | */ |
1308 | word = (map[0] << 8) + map[1]; | |
7f684a2a | 1309 | if ((map[0] & 0x8F) == 0x08 && !(word % 31)) |
bb6b8e4f | 1310 | return 0; |
cc5c54e7 JH |
1311 | else |
1312 | return 1; | |
bb6b8e4f JH |
1313 | } |
1314 | ||
09ded04b NP |
1315 | unsigned long unpack_object_header_buffer(const unsigned char *buf, |
1316 | unsigned long len, enum object_type *type, unsigned long *sizep) | |
c4483576 | 1317 | { |
ad1ed5ee | 1318 | unsigned shift; |
48fb7deb | 1319 | unsigned long size, c; |
ad1ed5ee JH |
1320 | unsigned long used = 0; |
1321 | ||
1322 | c = buf[used++]; | |
1323 | *type = (c >> 4) & 7; | |
1324 | size = c & 15; | |
1325 | shift = 4; | |
1326 | while (c & 0x80) { | |
f630cfda | 1327 | if (len <= used || bitsizeof(long) <= shift) { |
09ded04b | 1328 | error("bad object header"); |
ea4f9685 JH |
1329 | size = used = 0; |
1330 | break; | |
09ded04b | 1331 | } |
ad1ed5ee JH |
1332 | c = buf[used++]; |
1333 | size += (c & 0x7f) << shift; | |
1334 | shift += 7; | |
1335 | } | |
1336 | *sizep = size; | |
1337 | return used; | |
1338 | } | |
1339 | ||
eb4f4076 | 1340 | int unpack_sha1_header(git_zstream *stream, unsigned char *map, unsigned long mapsize, void *buffer, unsigned long bufsiz) |
ad1ed5ee JH |
1341 | { |
1342 | unsigned long size, used; | |
1343 | static const char valid_loose_object_type[8] = { | |
1344 | 0, /* OBJ_EXT */ | |
1345 | 1, 1, 1, 1, /* "commit", "tree", "blob", "tag" */ | |
1346 | 0, /* "delta" and others are invalid in a loose object */ | |
93821bd9 | 1347 | }; |
ad1ed5ee | 1348 | enum object_type type; |
93821bd9 | 1349 | |
c4483576 LT |
1350 | /* Get the data stream */ |
1351 | memset(stream, 0, sizeof(*stream)); | |
1352 | stream->next_in = map; | |
1353 | stream->avail_in = mapsize; | |
1354 | stream->next_out = buffer; | |
93821bd9 LT |
1355 | stream->avail_out = bufsiz; |
1356 | ||
cc5c54e7 JH |
1357 | if (experimental_loose_object(map)) { |
1358 | /* | |
1359 | * The old experimental format we no longer produce; | |
1360 | * we can still read it. | |
1361 | */ | |
1362 | used = unpack_object_header_buffer(map, mapsize, &type, &size); | |
1363 | if (!used || !valid_loose_object_type[type]) | |
1364 | return -1; | |
1365 | map += used; | |
1366 | mapsize -= used; | |
726f852b | 1367 | |
cc5c54e7 JH |
1368 | /* Set up the stream for the rest.. */ |
1369 | stream->next_in = map; | |
1370 | stream->avail_in = mapsize; | |
1371 | git_inflate_init(stream); | |
93821bd9 | 1372 | |
cc5c54e7 JH |
1373 | /* And generate the fake traditional header */ |
1374 | stream->total_out = 1 + snprintf(buffer, bufsiz, "%s %lu", | |
1375 | typename(type), size); | |
1376 | return 0; | |
1377 | } | |
39c68542 | 1378 | git_inflate_init(stream); |
cc5c54e7 | 1379 | return git_inflate(stream, 0); |
c4483576 LT |
1380 | } |
1381 | ||
ef49a7a0 | 1382 | static void *unpack_sha1_rest(git_zstream *stream, void *buffer, unsigned long size, const unsigned char *sha1) |
5180cacc LT |
1383 | { |
1384 | int bytes = strlen(buffer) + 1; | |
3aee68aa | 1385 | unsigned char *buf = xmallocz(size); |
93821bd9 | 1386 | unsigned long n; |
7efbff75 | 1387 | int status = Z_OK; |
5180cacc | 1388 | |
93821bd9 LT |
1389 | n = stream->total_out - bytes; |
1390 | if (n > size) | |
1391 | n = size; | |
1392 | memcpy(buf, (char *) buffer + bytes, n); | |
1393 | bytes = n; | |
456cdf6e LT |
1394 | if (bytes <= size) { |
1395 | /* | |
1396 | * The above condition must be (bytes <= size), not | |
1397 | * (bytes < size). In other words, even though we | |
ccf5ace0 | 1398 | * expect no more output and set avail_out to zero, |
456cdf6e LT |
1399 | * the input zlib stream may have bytes that express |
1400 | * "this concludes the stream", and we *do* want to | |
1401 | * eat that input. | |
1402 | * | |
1403 | * Otherwise we would not be able to test that we | |
1404 | * consumed all the input to reach the expected size; | |
1405 | * we also want to check that zlib tells us that all | |
1406 | * went well with status == Z_STREAM_END at the end. | |
1407 | */ | |
5180cacc LT |
1408 | stream->next_out = buf + bytes; |
1409 | stream->avail_out = size - bytes; | |
7efbff75 | 1410 | while (status == Z_OK) |
39c68542 | 1411 | status = git_inflate(stream, Z_FINISH); |
5180cacc | 1412 | } |
456cdf6e | 1413 | if (status == Z_STREAM_END && !stream->avail_in) { |
39c68542 | 1414 | git_inflate_end(stream); |
7efbff75 JH |
1415 | return buf; |
1416 | } | |
1417 | ||
1418 | if (status < 0) | |
1419 | error("corrupt loose object '%s'", sha1_to_hex(sha1)); | |
1420 | else if (stream->avail_in) | |
1421 | error("garbage at end of loose object '%s'", | |
1422 | sha1_to_hex(sha1)); | |
1423 | free(buf); | |
1424 | return NULL; | |
5180cacc LT |
1425 | } |
1426 | ||
1427 | /* | |
1428 | * We used to just use "sscanf()", but that's actually way | |
1429 | * too permissive for what we want to check. So do an anal | |
1430 | * object header parse by hand. | |
1431 | */ | |
f0270efd | 1432 | int parse_sha1_header(const char *hdr, unsigned long *sizep) |
5180cacc | 1433 | { |
21666f1a | 1434 | char type[10]; |
5180cacc LT |
1435 | int i; |
1436 | unsigned long size; | |
1437 | ||
1438 | /* | |
a6080a0a | 1439 | * The type can be at most ten bytes (including the |
5180cacc | 1440 | * terminating '\0' that we add), and is followed by |
21666f1a | 1441 | * a space. |
5180cacc | 1442 | */ |
21666f1a | 1443 | i = 0; |
5180cacc LT |
1444 | for (;;) { |
1445 | char c = *hdr++; | |
1446 | if (c == ' ') | |
1447 | break; | |
21666f1a NP |
1448 | type[i++] = c; |
1449 | if (i >= sizeof(type)) | |
5180cacc | 1450 | return -1; |
5180cacc | 1451 | } |
21666f1a | 1452 | type[i] = 0; |
5180cacc LT |
1453 | |
1454 | /* | |
1455 | * The length must follow immediately, and be in canonical | |
1456 | * decimal format (ie "010" is not valid). | |
1457 | */ | |
1458 | size = *hdr++ - '0'; | |
1459 | if (size > 9) | |
1460 | return -1; | |
1461 | if (size) { | |
1462 | for (;;) { | |
1463 | unsigned long c = *hdr - '0'; | |
1464 | if (c > 9) | |
1465 | break; | |
1466 | hdr++; | |
1467 | size = size * 10 + c; | |
1468 | } | |
1469 | } | |
1470 | *sizep = size; | |
1471 | ||
1472 | /* | |
1473 | * The length must be followed by a zero byte | |
1474 | */ | |
21666f1a | 1475 | return *hdr ? -1 : type_from_string(type); |
5180cacc LT |
1476 | } |
1477 | ||
7efbff75 | 1478 | static void *unpack_sha1_file(void *map, unsigned long mapsize, enum object_type *type, unsigned long *size, const unsigned char *sha1) |
0fcfd160 | 1479 | { |
5180cacc | 1480 | int ret; |
ef49a7a0 | 1481 | git_zstream stream; |
5180cacc | 1482 | char hdr[8192]; |
0fcfd160 | 1483 | |
5180cacc | 1484 | ret = unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)); |
21666f1a | 1485 | if (ret < Z_OK || (*type = parse_sha1_header(hdr, size)) < 0) |
0fcfd160 LT |
1486 | return NULL; |
1487 | ||
7efbff75 | 1488 | return unpack_sha1_rest(&stream, hdr, *size, sha1); |
0fcfd160 LT |
1489 | } |
1490 | ||
54dab52a NP |
1491 | unsigned long get_size_from_delta(struct packed_git *p, |
1492 | struct pack_window **w_curs, | |
1493 | off_t curpos) | |
1494 | { | |
1495 | const unsigned char *data; | |
1496 | unsigned char delta_head[20], *in; | |
ef49a7a0 | 1497 | git_zstream stream; |
54dab52a NP |
1498 | int st; |
1499 | ||
1500 | memset(&stream, 0, sizeof(stream)); | |
1501 | stream.next_out = delta_head; | |
1502 | stream.avail_out = sizeof(delta_head); | |
1503 | ||
39c68542 | 1504 | git_inflate_init(&stream); |
54dab52a NP |
1505 | do { |
1506 | in = use_pack(p, w_curs, curpos, &stream.avail_in); | |
1507 | stream.next_in = in; | |
39c68542 | 1508 | st = git_inflate(&stream, Z_FINISH); |
54dab52a NP |
1509 | curpos += stream.next_in - in; |
1510 | } while ((st == Z_OK || st == Z_BUF_ERROR) && | |
1511 | stream.total_out < sizeof(delta_head)); | |
39c68542 | 1512 | git_inflate_end(&stream); |
3d77d877 NP |
1513 | if ((st != Z_STREAM_END) && stream.total_out != sizeof(delta_head)) { |
1514 | error("delta data unpack-initial failed"); | |
1515 | return 0; | |
1516 | } | |
54dab52a NP |
1517 | |
1518 | /* Examine the initial part of the delta to figure out | |
1519 | * the result size. | |
1520 | */ | |
1521 | data = delta_head; | |
1522 | ||
1523 | /* ignore base size */ | |
1524 | get_delta_hdr_size(&data, delta_head+sizeof(delta_head)); | |
1525 | ||
1526 | /* Read the result size */ | |
1527 | return get_delta_hdr_size(&data, delta_head+sizeof(delta_head)); | |
1528 | } | |
1529 | ||
c4001d92 | 1530 | static off_t get_delta_base(struct packed_git *p, |
03e79c88 | 1531 | struct pack_window **w_curs, |
c4001d92 | 1532 | off_t *curpos, |
21666f1a | 1533 | enum object_type type, |
c4001d92 | 1534 | off_t delta_obj_offset) |
eb32d236 | 1535 | { |
2b87c45b | 1536 | unsigned char *base_info = use_pack(p, w_curs, *curpos, NULL); |
c4001d92 | 1537 | off_t base_offset; |
eb32d236 | 1538 | |
8d8a4ea5 SP |
1539 | /* use_pack() assured us we have [base_info, base_info + 20) |
1540 | * as a range that we can look at without walking off the | |
1541 | * end of the mapped window. Its actually the hash size | |
1542 | * that is assured. An OFS_DELTA longer than the hash size | |
1543 | * is stupid, as then a REF_DELTA would be smaller to store. | |
1544 | */ | |
21666f1a | 1545 | if (type == OBJ_OFS_DELTA) { |
eb32d236 NP |
1546 | unsigned used = 0; |
1547 | unsigned char c = base_info[used++]; | |
1548 | base_offset = c & 127; | |
1549 | while (c & 128) { | |
1550 | base_offset += 1; | |
8723f216 | 1551 | if (!base_offset || MSB(base_offset, 7)) |
8eca0b47 | 1552 | return 0; /* overflow */ |
eb32d236 NP |
1553 | c = base_info[used++]; |
1554 | base_offset = (base_offset << 7) + (c & 127); | |
1555 | } | |
1556 | base_offset = delta_obj_offset - base_offset; | |
d8f32556 | 1557 | if (base_offset <= 0 || base_offset >= delta_obj_offset) |
8eca0b47 | 1558 | return 0; /* out of bound */ |
2b87c45b | 1559 | *curpos += used; |
21666f1a | 1560 | } else if (type == OBJ_REF_DELTA) { |
eb32d236 NP |
1561 | /* The base entry _must_ be in the same pack */ |
1562 | base_offset = find_pack_entry_one(base_info, p); | |
2b87c45b | 1563 | *curpos += 20; |
eb32d236 NP |
1564 | } else |
1565 | die("I am totally screwed"); | |
2b87c45b | 1566 | return base_offset; |
eb32d236 NP |
1567 | } |
1568 | ||
f3bf9224 | 1569 | /* forward declaration for a mutually recursive function */ |
c4001d92 | 1570 | static int packed_object_info(struct packed_git *p, off_t offset, |
9a490590 | 1571 | unsigned long *sizep, int *rtype); |
f3bf9224 | 1572 | |
43057304 | 1573 | static int packed_delta_info(struct packed_git *p, |
03e79c88 | 1574 | struct pack_window **w_curs, |
c4001d92 | 1575 | off_t curpos, |
21666f1a | 1576 | enum object_type type, |
c4001d92 | 1577 | off_t obj_offset, |
43057304 | 1578 | unsigned long *sizep) |
5db47c2b | 1579 | { |
c4001d92 | 1580 | off_t base_offset; |
f3bf9224 | 1581 | |
21666f1a | 1582 | base_offset = get_delta_base(p, w_curs, &curpos, type, obj_offset); |
3d77d877 NP |
1583 | if (!base_offset) |
1584 | return OBJ_BAD; | |
9a490590 | 1585 | type = packed_object_info(p, base_offset, NULL, NULL); |
3d77d877 | 1586 | if (type <= OBJ_NONE) { |
08698b1e NP |
1587 | struct revindex_entry *revidx; |
1588 | const unsigned char *base_sha1; | |
1589 | revidx = find_pack_revindex(p, base_offset); | |
1590 | if (!revidx) | |
1591 | return OBJ_BAD; | |
1592 | base_sha1 = nth_packed_object_sha1(p, revidx->nr); | |
3d77d877 NP |
1593 | mark_bad_packed_object(p, base_sha1); |
1594 | type = sha1_object_info(base_sha1, NULL); | |
1595 | if (type <= OBJ_NONE) | |
1596 | return OBJ_BAD; | |
1597 | } | |
f3bf9224 | 1598 | |
c62266f3 JH |
1599 | /* We choose to only get the type of the base object and |
1600 | * ignore potentially corrupt pack file that expects the delta | |
1601 | * based on a base with a wrong size. This saves tons of | |
1602 | * inflate() calls. | |
1603 | */ | |
3d77d877 | 1604 | if (sizep) { |
54dab52a | 1605 | *sizep = get_size_from_delta(p, w_curs, curpos); |
3d77d877 NP |
1606 | if (*sizep == 0) |
1607 | type = OBJ_BAD; | |
1608 | } | |
21666f1a NP |
1609 | |
1610 | return type; | |
5db47c2b JH |
1611 | } |
1612 | ||
f8c8abc5 JH |
1613 | int unpack_object_header(struct packed_git *p, |
1614 | struct pack_window **w_curs, | |
1615 | off_t *curpos, | |
1616 | unsigned long *sizep) | |
a733cb60 | 1617 | { |
03e79c88 | 1618 | unsigned char *base; |
ef49a7a0 | 1619 | unsigned long left; |
ad1ed5ee | 1620 | unsigned long used; |
2b87c45b | 1621 | enum object_type type; |
a733cb60 | 1622 | |
8d8a4ea5 | 1623 | /* use_pack() assures us we have [base, base + 20) available |
0353a0c4 | 1624 | * as a range that we can look at. (Its actually the hash |
3dff5379 | 1625 | * size that is assured.) With our object header encoding |
8d8a4ea5 SP |
1626 | * the maximum deflated object size is 2^137, which is just |
1627 | * insane, so we know won't exceed what we have been given. | |
1628 | */ | |
2b87c45b | 1629 | base = use_pack(p, w_curs, *curpos, &left); |
09ded04b NP |
1630 | used = unpack_object_header_buffer(base, left, &type, sizep); |
1631 | if (!used) { | |
1632 | type = OBJ_BAD; | |
1633 | } else | |
1634 | *curpos += used; | |
ad1ed5ee | 1635 | |
2b87c45b | 1636 | return type; |
a733cb60 LT |
1637 | } |
1638 | ||
c4001d92 | 1639 | static int packed_object_info(struct packed_git *p, off_t obj_offset, |
9a490590 | 1640 | unsigned long *sizep, int *rtype) |
1f688557 | 1641 | { |
03e79c88 | 1642 | struct pack_window *w_curs = NULL; |
c4001d92 SP |
1643 | unsigned long size; |
1644 | off_t curpos = obj_offset; | |
21666f1a | 1645 | enum object_type type; |
5db47c2b | 1646 | |
21666f1a | 1647 | type = unpack_object_header(p, &w_curs, &curpos, &size); |
9a490590 JH |
1648 | if (rtype) |
1649 | *rtype = type; /* representation type */ | |
5db47c2b | 1650 | |
21666f1a | 1651 | switch (type) { |
eb32d236 NP |
1652 | case OBJ_OFS_DELTA: |
1653 | case OBJ_REF_DELTA: | |
21666f1a NP |
1654 | type = packed_delta_info(p, &w_curs, curpos, |
1655 | type, obj_offset, sizep); | |
2b87c45b | 1656 | break; |
a733cb60 | 1657 | case OBJ_COMMIT: |
a733cb60 | 1658 | case OBJ_TREE: |
a733cb60 | 1659 | case OBJ_BLOB: |
a733cb60 | 1660 | case OBJ_TAG: |
2b87c45b NP |
1661 | if (sizep) |
1662 | *sizep = size; | |
a69d0943 | 1663 | break; |
1f688557 | 1664 | default: |
3d77d877 NP |
1665 | error("unknown object type %i at offset %"PRIuMAX" in %s", |
1666 | type, (uintmax_t)obj_offset, p->pack_name); | |
1667 | type = OBJ_BAD; | |
1f688557 | 1668 | } |
2b87c45b | 1669 | unuse_pack(&w_curs); |
21666f1a | 1670 | return type; |
1f688557 JH |
1671 | } |
1672 | ||
eb950c19 | 1673 | static void *unpack_compressed_entry(struct packed_git *p, |
03e79c88 | 1674 | struct pack_window **w_curs, |
c4001d92 | 1675 | off_t curpos, |
eb950c19 | 1676 | unsigned long size) |
de530aaa SP |
1677 | { |
1678 | int st; | |
ef49a7a0 | 1679 | git_zstream stream; |
079afb18 | 1680 | unsigned char *buffer, *in; |
de530aaa | 1681 | |
4ab07e4d | 1682 | buffer = xmallocz(size); |
de530aaa | 1683 | memset(&stream, 0, sizeof(stream)); |
de530aaa | 1684 | stream.next_out = buffer; |
39eea7bd | 1685 | stream.avail_out = size + 1; |
de530aaa | 1686 | |
39c68542 | 1687 | git_inflate_init(&stream); |
079afb18 | 1688 | do { |
2b87c45b | 1689 | in = use_pack(p, w_curs, curpos, &stream.avail_in); |
079afb18 | 1690 | stream.next_in = in; |
39c68542 | 1691 | st = git_inflate(&stream, Z_FINISH); |
39eea7bd JH |
1692 | if (!stream.avail_out) |
1693 | break; /* the payload is larger than it should be */ | |
2b87c45b | 1694 | curpos += stream.next_in - in; |
079afb18 | 1695 | } while (st == Z_OK || st == Z_BUF_ERROR); |
39c68542 | 1696 | git_inflate_end(&stream); |
de530aaa SP |
1697 | if ((st != Z_STREAM_END) || stream.total_out != size) { |
1698 | free(buffer); | |
1699 | return NULL; | |
1700 | } | |
1701 | ||
1702 | return buffer; | |
1703 | } | |
1704 | ||
e5e01619 LT |
1705 | #define MAX_DELTA_CACHE (256) |
1706 | ||
18bdec11 | 1707 | static size_t delta_base_cached; |
5e08ecbf NP |
1708 | |
1709 | static struct delta_base_cache_lru_list { | |
1710 | struct delta_base_cache_lru_list *prev; | |
1711 | struct delta_base_cache_lru_list *next; | |
1712 | } delta_base_cache_lru = { &delta_base_cache_lru, &delta_base_cache_lru }; | |
1713 | ||
e5e01619 | 1714 | static struct delta_base_cache_entry { |
5e08ecbf NP |
1715 | struct delta_base_cache_lru_list lru; |
1716 | void *data; | |
e5e01619 LT |
1717 | struct packed_git *p; |
1718 | off_t base_offset; | |
1719 | unsigned long size; | |
e5e01619 LT |
1720 | enum object_type type; |
1721 | } delta_base_cache[MAX_DELTA_CACHE]; | |
1722 | ||
1723 | static unsigned long pack_entry_hash(struct packed_git *p, off_t base_offset) | |
1724 | { | |
1725 | unsigned long hash; | |
1726 | ||
1727 | hash = (unsigned long)p + (unsigned long)base_offset; | |
1728 | hash += (hash >> 8) + (hash >> 16); | |
3358004a | 1729 | return hash % MAX_DELTA_CACHE; |
e5e01619 LT |
1730 | } |
1731 | ||
5266d369 JH |
1732 | static int in_delta_base_cache(struct packed_git *p, off_t base_offset) |
1733 | { | |
1734 | unsigned long hash = pack_entry_hash(p, base_offset); | |
1735 | struct delta_base_cache_entry *ent = delta_base_cache + hash; | |
1736 | return (ent->data && ent->p == p && ent->base_offset == base_offset); | |
1737 | } | |
1738 | ||
62f255ad | 1739 | static void *cache_or_unpack_entry(struct packed_git *p, off_t base_offset, |
a0cba108 | 1740 | unsigned long *base_size, enum object_type *type, int keep_cache) |
62f255ad | 1741 | { |
e5e01619 LT |
1742 | void *ret; |
1743 | unsigned long hash = pack_entry_hash(p, base_offset); | |
1744 | struct delta_base_cache_entry *ent = delta_base_cache + hash; | |
1745 | ||
1746 | ret = ent->data; | |
749bc58c MV |
1747 | if (!ret || ent->p != p || ent->base_offset != base_offset) |
1748 | return unpack_entry(p, base_offset, type, base_size); | |
e5e01619 | 1749 | |
18bdec11 | 1750 | if (!keep_cache) { |
a0cba108 | 1751 | ent->data = NULL; |
5e08ecbf NP |
1752 | ent->lru.next->prev = ent->lru.prev; |
1753 | ent->lru.prev->next = ent->lru.next; | |
18bdec11 | 1754 | delta_base_cached -= ent->size; |
182af834 PH |
1755 | } else { |
1756 | ret = xmemdupz(ent->data, ent->size); | |
a0cba108 | 1757 | } |
e5e01619 LT |
1758 | *type = ent->type; |
1759 | *base_size = ent->size; | |
1760 | return ret; | |
62f255ad LT |
1761 | } |
1762 | ||
18bdec11 SP |
1763 | static inline void release_delta_base_cache(struct delta_base_cache_entry *ent) |
1764 | { | |
1765 | if (ent->data) { | |
1766 | free(ent->data); | |
1767 | ent->data = NULL; | |
5e08ecbf NP |
1768 | ent->lru.next->prev = ent->lru.prev; |
1769 | ent->lru.prev->next = ent->lru.next; | |
18bdec11 SP |
1770 | delta_base_cached -= ent->size; |
1771 | } | |
1772 | } | |
1773 | ||
3d20c636 SP |
1774 | void clear_delta_base_cache(void) |
1775 | { | |
1776 | unsigned long p; | |
1777 | for (p = 0; p < MAX_DELTA_CACHE; p++) | |
1778 | release_delta_base_cache(&delta_base_cache[p]); | |
1779 | } | |
1780 | ||
62f255ad LT |
1781 | static void add_delta_base_cache(struct packed_git *p, off_t base_offset, |
1782 | void *base, unsigned long base_size, enum object_type type) | |
1783 | { | |
5e08ecbf | 1784 | unsigned long hash = pack_entry_hash(p, base_offset); |
e5e01619 | 1785 | struct delta_base_cache_entry *ent = delta_base_cache + hash; |
5e08ecbf | 1786 | struct delta_base_cache_lru_list *lru; |
e5e01619 | 1787 | |
18bdec11 SP |
1788 | release_delta_base_cache(ent); |
1789 | delta_base_cached += base_size; | |
5e08ecbf NP |
1790 | |
1791 | for (lru = delta_base_cache_lru.next; | |
1792 | delta_base_cached > delta_base_cache_limit | |
1793 | && lru != &delta_base_cache_lru; | |
1794 | lru = lru->next) { | |
1795 | struct delta_base_cache_entry *f = (void *)lru; | |
18bdec11 SP |
1796 | if (f->type == OBJ_BLOB) |
1797 | release_delta_base_cache(f); | |
1798 | } | |
5e08ecbf NP |
1799 | for (lru = delta_base_cache_lru.next; |
1800 | delta_base_cached > delta_base_cache_limit | |
1801 | && lru != &delta_base_cache_lru; | |
1802 | lru = lru->next) { | |
1803 | struct delta_base_cache_entry *f = (void *)lru; | |
1804 | release_delta_base_cache(f); | |
1805 | } | |
18bdec11 | 1806 | |
e5e01619 LT |
1807 | ent->p = p; |
1808 | ent->base_offset = base_offset; | |
1809 | ent->type = type; | |
1810 | ent->data = base; | |
1811 | ent->size = base_size; | |
5e08ecbf NP |
1812 | ent->lru.next = &delta_base_cache_lru; |
1813 | ent->lru.prev = delta_base_cache_lru.prev; | |
1814 | delta_base_cache_lru.prev->next = &ent->lru; | |
1815 | delta_base_cache_lru.prev = &ent->lru; | |
62f255ad LT |
1816 | } |
1817 | ||
c2c5b270 CC |
1818 | static void *read_object(const unsigned char *sha1, enum object_type *type, |
1819 | unsigned long *size); | |
1820 | ||
eb950c19 | 1821 | static void *unpack_delta_entry(struct packed_git *p, |
03e79c88 | 1822 | struct pack_window **w_curs, |
c4001d92 | 1823 | off_t curpos, |
1f688557 | 1824 | unsigned long delta_size, |
c4001d92 | 1825 | off_t obj_offset, |
21666f1a | 1826 | enum object_type *type, |
eb950c19 | 1827 | unsigned long *sizep) |
1f688557 | 1828 | { |
7c3e8be3 | 1829 | void *delta_data, *result, *base; |
c4001d92 SP |
1830 | unsigned long base_size; |
1831 | off_t base_offset; | |
43057304 | 1832 | |
21666f1a | 1833 | base_offset = get_delta_base(p, w_curs, &curpos, *type, obj_offset); |
8eca0b47 NP |
1834 | if (!base_offset) { |
1835 | error("failed to validate delta base reference " | |
1836 | "at offset %"PRIuMAX" from %s", | |
1837 | (uintmax_t)curpos, p->pack_name); | |
1838 | return NULL; | |
1839 | } | |
eac12e2d | 1840 | unuse_pack(w_curs); |
a0cba108 | 1841 | base = cache_or_unpack_entry(p, base_offset, &base_size, type, 0); |
8eca0b47 NP |
1842 | if (!base) { |
1843 | /* | |
1844 | * We're probably in deep shit, but let's try to fetch | |
1845 | * the required base anyway from another pack or loose. | |
1846 | * This is costly but should happen only in the presence | |
1847 | * of a corrupted pack, and is better than failing outright. | |
1848 | */ | |
08698b1e NP |
1849 | struct revindex_entry *revidx; |
1850 | const unsigned char *base_sha1; | |
1851 | revidx = find_pack_revindex(p, base_offset); | |
1852 | if (!revidx) | |
1853 | return NULL; | |
1854 | base_sha1 = nth_packed_object_sha1(p, revidx->nr); | |
8eca0b47 NP |
1855 | error("failed to read delta base object %s" |
1856 | " at offset %"PRIuMAX" from %s", | |
1857 | sha1_to_hex(base_sha1), (uintmax_t)base_offset, | |
1858 | p->pack_name); | |
1859 | mark_bad_packed_object(p, base_sha1); | |
ac939109 | 1860 | base = read_object(base_sha1, type, &base_size); |
8eca0b47 NP |
1861 | if (!base) |
1862 | return NULL; | |
1863 | } | |
67686d95 | 1864 | |
2b87c45b | 1865 | delta_data = unpack_compressed_entry(p, w_curs, curpos, delta_size); |
8eca0b47 NP |
1866 | if (!delta_data) { |
1867 | error("failed to unpack compressed delta " | |
1868 | "at offset %"PRIuMAX" from %s", | |
1869 | (uintmax_t)curpos, p->pack_name); | |
1870 | free(base); | |
1871 | return NULL; | |
1872 | } | |
1f688557 JH |
1873 | result = patch_delta(base, base_size, |
1874 | delta_data, delta_size, | |
21666f1a | 1875 | sizep); |
1f688557 JH |
1876 | if (!result) |
1877 | die("failed to apply delta"); | |
1878 | free(delta_data); | |
62f255ad | 1879 | add_delta_base_cache(p, base_offset, base, base_size, *type); |
1f688557 JH |
1880 | return result; |
1881 | } | |
1882 | ||
5f44324d JH |
1883 | static void write_pack_access_log(struct packed_git *p, off_t obj_offset) |
1884 | { | |
1885 | static FILE *log_file; | |
1886 | ||
1887 | if (!log_file) { | |
1888 | log_file = fopen(log_pack_access, "w"); | |
1889 | if (!log_file) { | |
1890 | error("cannot open pack access log '%s' for writing: %s", | |
1891 | log_pack_access, strerror(errno)); | |
1892 | log_pack_access = NULL; | |
1893 | return; | |
1894 | } | |
1895 | } | |
1896 | fprintf(log_file, "%s %"PRIuMAX"\n", | |
1897 | p->pack_name, (uintmax_t)obj_offset); | |
1898 | fflush(log_file); | |
1899 | } | |
1900 | ||
0e8189e2 NP |
1901 | int do_check_packed_object_crc; |
1902 | ||
c4001d92 | 1903 | void *unpack_entry(struct packed_git *p, off_t obj_offset, |
21666f1a | 1904 | enum object_type *type, unsigned long *sizep) |
1f688557 | 1905 | { |
03e79c88 | 1906 | struct pack_window *w_curs = NULL; |
c4001d92 | 1907 | off_t curpos = obj_offset; |
21666f1a | 1908 | void *data; |
1f688557 | 1909 | |
5f44324d JH |
1910 | if (log_pack_access) |
1911 | write_pack_access_log(p, obj_offset); | |
1912 | ||
0e8189e2 NP |
1913 | if (do_check_packed_object_crc && p->index_version > 1) { |
1914 | struct revindex_entry *revidx = find_pack_revindex(p, obj_offset); | |
1915 | unsigned long len = revidx[1].offset - obj_offset; | |
1916 | if (check_pack_crc(p, &w_curs, obj_offset, len, revidx->nr)) { | |
1917 | const unsigned char *sha1 = | |
1918 | nth_packed_object_sha1(p, revidx->nr); | |
1919 | error("bad packed object CRC for %s", | |
1920 | sha1_to_hex(sha1)); | |
1921 | mark_bad_packed_object(p, sha1); | |
f755bb99 | 1922 | unuse_pack(&w_curs); |
0e8189e2 NP |
1923 | return NULL; |
1924 | } | |
1925 | } | |
1926 | ||
21666f1a NP |
1927 | *type = unpack_object_header(p, &w_curs, &curpos, sizep); |
1928 | switch (*type) { | |
eb32d236 NP |
1929 | case OBJ_OFS_DELTA: |
1930 | case OBJ_REF_DELTA: | |
21666f1a NP |
1931 | data = unpack_delta_entry(p, &w_curs, curpos, *sizep, |
1932 | obj_offset, type, sizep); | |
4d703a1a | 1933 | break; |
a733cb60 | 1934 | case OBJ_COMMIT: |
a733cb60 | 1935 | case OBJ_TREE: |
a733cb60 | 1936 | case OBJ_BLOB: |
a733cb60 | 1937 | case OBJ_TAG: |
21666f1a | 1938 | data = unpack_compressed_entry(p, &w_curs, curpos, *sizep); |
4d703a1a | 1939 | break; |
1f688557 | 1940 | default: |
8eca0b47 NP |
1941 | data = NULL; |
1942 | error("unknown object type %i at offset %"PRIuMAX" in %s", | |
1943 | *type, (uintmax_t)obj_offset, p->pack_name); | |
1f688557 | 1944 | } |
03e79c88 | 1945 | unuse_pack(&w_curs); |
21666f1a | 1946 | return data; |
1f688557 JH |
1947 | } |
1948 | ||
d079837e | 1949 | const unsigned char *nth_packed_object_sha1(struct packed_git *p, |
d72308e0 | 1950 | uint32_t n) |
9a217f2a | 1951 | { |
42873078 | 1952 | const unsigned char *index = p->index_data; |
d079837e SP |
1953 | if (!index) { |
1954 | if (open_pack_index(p)) | |
1955 | return NULL; | |
1956 | index = p->index_data; | |
1957 | } | |
57059091 | 1958 | if (n >= p->num_objects) |
d72308e0 | 1959 | return NULL; |
74e34e1f NP |
1960 | index += 4 * 256; |
1961 | if (p->index_version == 1) { | |
1962 | return index + 24 * n + 4; | |
1963 | } else { | |
1964 | index += 8; | |
1965 | return index + 20 * n; | |
1966 | } | |
1967 | } | |
1968 | ||
99093238 | 1969 | off_t nth_packed_object_offset(const struct packed_git *p, uint32_t n) |
74e34e1f NP |
1970 | { |
1971 | const unsigned char *index = p->index_data; | |
1972 | index += 4 * 256; | |
1973 | if (p->index_version == 1) { | |
1974 | return ntohl(*((uint32_t *)(index + 24 * n))); | |
1975 | } else { | |
1976 | uint32_t off; | |
1977 | index += 8 + p->num_objects * (20 + 4); | |
1978 | off = ntohl(*((uint32_t *)(index + 4 * n))); | |
1979 | if (!(off & 0x80000000)) | |
1980 | return off; | |
1981 | index += p->num_objects * 4 + (off & 0x7fffffff) * 8; | |
1982 | return (((uint64_t)ntohl(*((uint32_t *)(index + 0)))) << 32) | | |
1983 | ntohl(*((uint32_t *)(index + 4))); | |
1984 | } | |
9a217f2a JH |
1985 | } |
1986 | ||
c4001d92 | 1987 | off_t find_pack_entry_one(const unsigned char *sha1, |
43057304 | 1988 | struct packed_git *p) |
1f688557 | 1989 | { |
42873078 | 1990 | const uint32_t *level1_ofs = p->index_data; |
42873078 | 1991 | const unsigned char *index = p->index_data; |
628522ec JH |
1992 | unsigned hi, lo, stride; |
1993 | static int use_lookup = -1; | |
1994 | static int debug_lookup = -1; | |
1995 | ||
1996 | if (debug_lookup < 0) | |
1997 | debug_lookup = !!getenv("GIT_DEBUG_LOOKUP"); | |
42873078 | 1998 | |
d079837e SP |
1999 | if (!index) { |
2000 | if (open_pack_index(p)) | |
2001 | return 0; | |
2002 | level1_ofs = p->index_data; | |
2003 | index = p->index_data; | |
2004 | } | |
74e34e1f NP |
2005 | if (p->index_version > 1) { |
2006 | level1_ofs += 2; | |
2007 | index += 8; | |
2008 | } | |
42873078 | 2009 | index += 4 * 256; |
74e34e1f NP |
2010 | hi = ntohl(level1_ofs[*sha1]); |
2011 | lo = ((*sha1 == 0x0) ? 0 : ntohl(level1_ofs[*sha1 - 1])); | |
628522ec JH |
2012 | if (p->index_version > 1) { |
2013 | stride = 20; | |
2014 | } else { | |
2015 | stride = 24; | |
2016 | index += 4; | |
2017 | } | |
2018 | ||
2019 | if (debug_lookup) | |
6e1c2344 | 2020 | printf("%02x%02x%02x... lo %u hi %u nr %"PRIu32"\n", |
628522ec JH |
2021 | sha1[0], sha1[1], sha1[2], lo, hi, p->num_objects); |
2022 | ||
2023 | if (use_lookup < 0) | |
2024 | use_lookup = !!getenv("GIT_USE_LOOKUP"); | |
2025 | if (use_lookup) { | |
2026 | int pos = sha1_entry_pos(index, stride, 0, | |
2027 | lo, hi, p->num_objects, sha1); | |
2028 | if (pos < 0) | |
2029 | return 0; | |
2030 | return nth_packed_object_offset(p, pos); | |
2031 | } | |
1f688557 JH |
2032 | |
2033 | do { | |
74e34e1f | 2034 | unsigned mi = (lo + hi) / 2; |
628522ec JH |
2035 | int cmp = hashcmp(index + mi * stride, sha1); |
2036 | ||
2037 | if (debug_lookup) | |
2038 | printf("lo %u hi %u rg %u mi %u\n", | |
2039 | lo, hi, hi - lo, mi); | |
43057304 | 2040 | if (!cmp) |
74e34e1f | 2041 | return nth_packed_object_offset(p, mi); |
1f688557 JH |
2042 | if (cmp > 0) |
2043 | hi = mi; | |
2044 | else | |
2045 | lo = mi+1; | |
2046 | } while (lo < hi); | |
2047 | return 0; | |
2048 | } | |
2049 | ||
4c080182 | 2050 | int is_pack_valid(struct packed_git *p) |
d131b7af SP |
2051 | { |
2052 | /* An already open pack is known to be valid. */ | |
2053 | if (p->pack_fd != -1) | |
2054 | return 1; | |
2055 | ||
2056 | /* If the pack has one window completely covering the | |
2057 | * file size, the pack is known to be valid even if | |
2058 | * the descriptor is not currently open. | |
2059 | */ | |
2060 | if (p->windows) { | |
2061 | struct pack_window *w = p->windows; | |
2062 | ||
2063 | if (!w->offset && w->len == p->pack_size) | |
2064 | return 1; | |
2065 | } | |
2066 | ||
2067 | /* Force the pack to open to prove its valid. */ | |
2068 | return !open_packed_git(p); | |
2069 | } | |
2070 | ||
95099731 NTND |
2071 | static int fill_pack_entry(const unsigned char *sha1, |
2072 | struct pack_entry *e, | |
2073 | struct packed_git *p) | |
2074 | { | |
2075 | off_t offset; | |
2076 | ||
2077 | if (p->num_bad_objects) { | |
2078 | unsigned i; | |
2079 | for (i = 0; i < p->num_bad_objects; i++) | |
2080 | if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i)) | |
2081 | return 0; | |
2082 | } | |
2083 | ||
2084 | offset = find_pack_entry_one(sha1, p); | |
2085 | if (!offset) | |
2086 | return 0; | |
2087 | ||
2088 | /* | |
2089 | * We are about to tell the caller where they can locate the | |
2090 | * requested object. We better make sure the packfile is | |
2091 | * still here and can be accessed before supplying that | |
2092 | * answer, as it may have been deleted since the index was | |
2093 | * loaded! | |
2094 | */ | |
2095 | if (!is_pack_valid(p)) { | |
2096 | warning("packfile %s cannot be accessed", p->pack_name); | |
2097 | return 0; | |
2098 | } | |
2099 | e->offset = offset; | |
2100 | e->p = p; | |
2101 | hashcpy(e->sha1, sha1); | |
2102 | return 1; | |
2103 | } | |
2104 | ||
4d6acb70 | 2105 | static int find_pack_entry(const unsigned char *sha1, struct pack_entry *e) |
1f688557 JH |
2106 | { |
2107 | struct packed_git *p; | |
43057304 | 2108 | |
1f688557 | 2109 | prepare_packed_git(); |
f7c22cc6 NP |
2110 | if (!packed_git) |
2111 | return 0; | |
1f688557 | 2112 | |
c01f51cc NTND |
2113 | if (last_found_pack && fill_pack_entry(sha1, e, last_found_pack)) |
2114 | return 1; | |
8eca0b47 | 2115 | |
c01f51cc NTND |
2116 | for (p = packed_git; p; p = p->next) { |
2117 | if (p == last_found_pack || !fill_pack_entry(sha1, e, p)) | |
2118 | continue; | |
f7c22cc6 | 2119 | |
c01f51cc NTND |
2120 | last_found_pack = p; |
2121 | return 1; | |
2122 | } | |
1f688557 JH |
2123 | return 0; |
2124 | } | |
2125 | ||
a6080a0a | 2126 | struct packed_git *find_sha1_pack(const unsigned char *sha1, |
bf592c50 DB |
2127 | struct packed_git *packs) |
2128 | { | |
2129 | struct packed_git *p; | |
bf592c50 DB |
2130 | |
2131 | for (p = packs; p; p = p->next) { | |
43057304 | 2132 | if (find_pack_entry_one(sha1, p)) |
bf592c50 DB |
2133 | return p; |
2134 | } | |
2135 | return NULL; | |
21666f1a | 2136 | |
bf592c50 DB |
2137 | } |
2138 | ||
21666f1a | 2139 | static int sha1_loose_object_info(const unsigned char *sha1, unsigned long *sizep) |
65c2e0c3 | 2140 | { |
36e4d74a | 2141 | int status; |
65c2e0c3 JH |
2142 | unsigned long mapsize, size; |
2143 | void *map; | |
ef49a7a0 | 2144 | git_zstream stream; |
d65a16f6 | 2145 | char hdr[32]; |
65c2e0c3 | 2146 | |
bb6b8e4f | 2147 | map = map_sha1_file(sha1, &mapsize); |
f0df4ed5 JS |
2148 | if (!map) |
2149 | return error("unable to find %s", sha1_to_hex(sha1)); | |
36e4d74a JH |
2150 | if (unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)) < 0) |
2151 | status = error("unable to unpack %s header", | |
2152 | sha1_to_hex(sha1)); | |
21666f1a | 2153 | else if ((status = parse_sha1_header(hdr, &size)) < 0) |
36e4d74a | 2154 | status = error("unable to parse %s header", sha1_to_hex(sha1)); |
21666f1a NP |
2155 | else if (sizep) |
2156 | *sizep = size; | |
39c68542 | 2157 | git_inflate_end(&stream); |
65c2e0c3 JH |
2158 | munmap(map, mapsize); |
2159 | return status; | |
2160 | } | |
2161 | ||
9a490590 JH |
2162 | /* returns enum object_type or negative */ |
2163 | int sha1_object_info_extended(const unsigned char *sha1, struct object_info *oi) | |
f0df4ed5 | 2164 | { |
c4d9986f | 2165 | struct cached_object *co; |
f0df4ed5 | 2166 | struct pack_entry e; |
9a490590 | 2167 | int status, rtype; |
f0df4ed5 | 2168 | |
c4d9986f NTND |
2169 | co = find_cached_object(sha1); |
2170 | if (co) { | |
9a490590 JH |
2171 | if (oi->sizep) |
2172 | *(oi->sizep) = co->size; | |
2173 | oi->whence = OI_CACHED; | |
c4d9986f NTND |
2174 | return co->type; |
2175 | } | |
2176 | ||
cd673c1f | 2177 | if (!find_pack_entry(sha1, &e)) { |
ddd63e64 | 2178 | /* Most likely it's a loose object. */ |
9a490590 JH |
2179 | status = sha1_loose_object_info(sha1, oi->sizep); |
2180 | if (status >= 0) { | |
2181 | oi->whence = OI_LOOSE; | |
ddd63e64 | 2182 | return status; |
9a490590 | 2183 | } |
ddd63e64 SG |
2184 | |
2185 | /* Not a loose object; someone else may have just packed it. */ | |
f0df4ed5 | 2186 | reprepare_packed_git(); |
cd673c1f | 2187 | if (!find_pack_entry(sha1, &e)) |
ddd63e64 | 2188 | return status; |
f0df4ed5 | 2189 | } |
3d77d877 | 2190 | |
9a490590 | 2191 | status = packed_object_info(e.p, e.offset, oi->sizep, &rtype); |
3d77d877 NP |
2192 | if (status < 0) { |
2193 | mark_bad_packed_object(e.p, sha1); | |
9a490590 | 2194 | status = sha1_object_info_extended(sha1, oi); |
5266d369 JH |
2195 | } else if (in_delta_base_cache(e.p, e.offset)) { |
2196 | oi->whence = OI_DBCACHED; | |
9a490590 JH |
2197 | } else { |
2198 | oi->whence = OI_PACKED; | |
2199 | oi->u.packed.offset = e.offset; | |
2200 | oi->u.packed.pack = e.p; | |
2201 | oi->u.packed.is_delta = (rtype == OBJ_REF_DELTA || | |
2202 | rtype == OBJ_OFS_DELTA); | |
3d77d877 NP |
2203 | } |
2204 | ||
2205 | return status; | |
f0df4ed5 JS |
2206 | } |
2207 | ||
9a490590 JH |
2208 | int sha1_object_info(const unsigned char *sha1, unsigned long *sizep) |
2209 | { | |
2210 | struct object_info oi; | |
2211 | ||
2212 | oi.sizep = sizep; | |
2213 | return sha1_object_info_extended(sha1, &oi); | |
2214 | } | |
2215 | ||
21666f1a NP |
2216 | static void *read_packed_sha1(const unsigned char *sha1, |
2217 | enum object_type *type, unsigned long *size) | |
1f688557 JH |
2218 | { |
2219 | struct pack_entry e; | |
8eca0b47 | 2220 | void *data; |
1f688557 | 2221 | |
cd673c1f | 2222 | if (!find_pack_entry(sha1, &e)) |
1f688557 | 2223 | return NULL; |
8eca0b47 NP |
2224 | data = cache_or_unpack_entry(e.p, e.offset, size, type, 1); |
2225 | if (!data) { | |
2226 | /* | |
2227 | * We're probably in deep shit, but let's try to fetch | |
2228 | * the required object anyway from another pack or loose. | |
2229 | * This should happen only in the presence of a corrupted | |
2230 | * pack, and is better than failing outright. | |
2231 | */ | |
2232 | error("failed to read object %s at offset %"PRIuMAX" from %s", | |
2233 | sha1_to_hex(sha1), (uintmax_t)e.offset, e.p->pack_name); | |
2234 | mark_bad_packed_object(e.p, sha1); | |
ac939109 | 2235 | data = read_object(sha1, type, size); |
8eca0b47 NP |
2236 | } |
2237 | return data; | |
1f688557 JH |
2238 | } |
2239 | ||
21666f1a NP |
2240 | int pretend_sha1_file(void *buf, unsigned long len, enum object_type type, |
2241 | unsigned char *sha1) | |
d66b37bb JH |
2242 | { |
2243 | struct cached_object *co; | |
2244 | ||
21666f1a | 2245 | hash_sha1_file(buf, len, typename(type), sha1); |
d66b37bb JH |
2246 | if (has_sha1_file(sha1) || find_cached_object(sha1)) |
2247 | return 0; | |
2248 | if (cached_object_alloc <= cached_object_nr) { | |
2249 | cached_object_alloc = alloc_nr(cached_object_alloc); | |
2250 | cached_objects = xrealloc(cached_objects, | |
2251 | sizeof(*cached_objects) * | |
2252 | cached_object_alloc); | |
2253 | } | |
2254 | co = &cached_objects[cached_object_nr++]; | |
2255 | co->size = len; | |
21666f1a | 2256 | co->type = type; |
efa13f7b JH |
2257 | co->buf = xmalloc(len); |
2258 | memcpy(co->buf, buf, len); | |
d66b37bb JH |
2259 | hashcpy(co->sha1, sha1); |
2260 | return 0; | |
2261 | } | |
2262 | ||
c2c5b270 CC |
2263 | static void *read_object(const unsigned char *sha1, enum object_type *type, |
2264 | unsigned long *size) | |
0fcfd160 LT |
2265 | { |
2266 | unsigned long mapsize; | |
2267 | void *map, *buf; | |
d66b37bb JH |
2268 | struct cached_object *co; |
2269 | ||
2270 | co = find_cached_object(sha1); | |
2271 | if (co) { | |
21666f1a | 2272 | *type = co->type; |
d66b37bb | 2273 | *size = co->size; |
182af834 | 2274 | return xmemdupz(co->buf, co->size); |
d66b37bb | 2275 | } |
0fcfd160 | 2276 | |
8276c007 PE |
2277 | buf = read_packed_sha1(sha1, type, size); |
2278 | if (buf) | |
2279 | return buf; | |
bb6b8e4f | 2280 | map = map_sha1_file(sha1, &mapsize); |
0fcfd160 | 2281 | if (map) { |
7efbff75 | 2282 | buf = unpack_sha1_file(map, mapsize, type, size, sha1); |
0fcfd160 LT |
2283 | munmap(map, mapsize); |
2284 | return buf; | |
2285 | } | |
637cdd9d | 2286 | reprepare_packed_git(); |
8276c007 | 2287 | return read_packed_sha1(sha1, type, size); |
0fcfd160 LT |
2288 | } |
2289 | ||
b6c4cecc JH |
2290 | /* |
2291 | * This function dies on corrupt objects; the callers who want to | |
2292 | * deal with them should arrange to call read_object() and give error | |
2293 | * messages themselves. | |
2294 | */ | |
5bf29b95 JH |
2295 | void *read_sha1_file_extended(const unsigned char *sha1, |
2296 | enum object_type *type, | |
2297 | unsigned long *size, | |
2298 | unsigned flag) | |
ac939109 | 2299 | { |
3ba7a065 | 2300 | void *data; |
e8b15e61 | 2301 | char *path; |
b6c4cecc | 2302 | const struct packed_git *p; |
5bf29b95 JH |
2303 | const unsigned char *repl = (flag & READ_SHA1_FILE_REPLACE) |
2304 | ? lookup_replace_object(sha1) : sha1; | |
b6c4cecc | 2305 | |
3ba7a065 JH |
2306 | errno = 0; |
2307 | data = read_object(repl, type, size); | |
4bbf5a26 | 2308 | if (data) |
b6c4cecc | 2309 | return data; |
68095570 | 2310 | |
25f3af3f | 2311 | if (errno && errno != ENOENT) |
3ba7a065 JH |
2312 | die_errno("failed to read object %s", sha1_to_hex(sha1)); |
2313 | ||
68095570 | 2314 | /* die if we replaced an object with one that does not exist */ |
b6c4cecc | 2315 | if (repl != sha1) |
68095570 CC |
2316 | die("replacement %s not found for %s", |
2317 | sha1_to_hex(repl), sha1_to_hex(sha1)); | |
2318 | ||
b6c4cecc JH |
2319 | if (has_loose_object(repl)) { |
2320 | path = sha1_file_name(sha1); | |
2321 | die("loose object %s (stored in %s) is corrupt", | |
2322 | sha1_to_hex(repl), path); | |
e8b15e61 | 2323 | } |
68095570 | 2324 | |
b6c4cecc JH |
2325 | if ((p = has_packed_and_bad(repl)) != NULL) |
2326 | die("packed object %s (stored in %s) is corrupt", | |
2327 | sha1_to_hex(repl), p->pack_name); | |
f5552aee | 2328 | |
b6c4cecc | 2329 | return NULL; |
ac939109 NP |
2330 | } |
2331 | ||
40469ee9 | 2332 | void *read_object_with_reference(const unsigned char *sha1, |
21666f1a | 2333 | const char *required_type_name, |
40469ee9 JH |
2334 | unsigned long *size, |
2335 | unsigned char *actual_sha1_return) | |
f4913f91 | 2336 | { |
21666f1a | 2337 | enum object_type type, required_type; |
f4913f91 JH |
2338 | void *buffer; |
2339 | unsigned long isize; | |
40469ee9 | 2340 | unsigned char actual_sha1[20]; |
f4913f91 | 2341 | |
21666f1a | 2342 | required_type = type_from_string(required_type_name); |
e702496e | 2343 | hashcpy(actual_sha1, sha1); |
40469ee9 JH |
2344 | while (1) { |
2345 | int ref_length = -1; | |
2346 | const char *ref_type = NULL; | |
f4913f91 | 2347 | |
21666f1a | 2348 | buffer = read_sha1_file(actual_sha1, &type, &isize); |
40469ee9 JH |
2349 | if (!buffer) |
2350 | return NULL; | |
21666f1a | 2351 | if (type == required_type) { |
40469ee9 JH |
2352 | *size = isize; |
2353 | if (actual_sha1_return) | |
e702496e | 2354 | hashcpy(actual_sha1_return, actual_sha1); |
40469ee9 JH |
2355 | return buffer; |
2356 | } | |
2357 | /* Handle references */ | |
21666f1a | 2358 | else if (type == OBJ_COMMIT) |
40469ee9 | 2359 | ref_type = "tree "; |
21666f1a | 2360 | else if (type == OBJ_TAG) |
40469ee9 JH |
2361 | ref_type = "object "; |
2362 | else { | |
2363 | free(buffer); | |
2364 | return NULL; | |
2365 | } | |
2366 | ref_length = strlen(ref_type); | |
f4913f91 | 2367 | |
50974ec9 MK |
2368 | if (ref_length + 40 > isize || |
2369 | memcmp(buffer, ref_type, ref_length) || | |
1d7f171c | 2370 | get_sha1_hex((char *) buffer + ref_length, actual_sha1)) { |
40469ee9 JH |
2371 | free(buffer); |
2372 | return NULL; | |
2373 | } | |
1cf58e72 | 2374 | free(buffer); |
40469ee9 JH |
2375 | /* Now we have the ID of the referred-to object in |
2376 | * actual_sha1. Check again. */ | |
f4913f91 | 2377 | } |
f4913f91 JH |
2378 | } |
2379 | ||
ce9fbf16 | 2380 | static void write_sha1_file_prepare(const void *buf, unsigned long len, |
972a9155 | 2381 | const char *type, unsigned char *sha1, |
d65a16f6 | 2382 | char *hdr, int *hdrlen) |
d410c0f5 | 2383 | { |
9126f009 | 2384 | git_SHA_CTX c; |
d410c0f5 JH |
2385 | |
2386 | /* Generate the header */ | |
d65a16f6 | 2387 | *hdrlen = sprintf(hdr, "%s %lu", type, len)+1; |
d410c0f5 JH |
2388 | |
2389 | /* Sha1.. */ | |
9126f009 NP |
2390 | git_SHA1_Init(&c); |
2391 | git_SHA1_Update(&c, hdr, *hdrlen); | |
2392 | git_SHA1_Update(&c, buf, len); | |
2393 | git_SHA1_Final(sha1, &c); | |
d410c0f5 JH |
2394 | } |
2395 | ||
230f1322 | 2396 | /* |
5a688fe4 JH |
2397 | * Move the just written object into its final resting place. |
2398 | * NEEDSWORK: this should be renamed to finalize_temp_file() as | |
2399 | * "moving" is only a part of what it does, when no patch between | |
2400 | * master to pu changes the call sites of this function. | |
230f1322 | 2401 | */ |
839837b9 | 2402 | int move_temp_to_file(const char *tmpfile, const char *filename) |
230f1322 | 2403 | { |
e32c0a9c | 2404 | int ret = 0; |
5a688fe4 | 2405 | |
348df166 | 2406 | if (object_creation_mode == OBJECT_CREATION_USES_RENAMES) |
be66a6c4 JS |
2407 | goto try_rename; |
2408 | else if (link(tmpfile, filename)) | |
e32c0a9c | 2409 | ret = errno; |
7ebb6fca LT |
2410 | |
2411 | /* | |
2412 | * Coda hack - coda doesn't like cross-directory links, | |
2413 | * so we fall back to a rename, which will mean that it | |
2414 | * won't be able to check collisions, but that's not a | |
2415 | * big deal. | |
2416 | * | |
2417 | * The same holds for FAT formatted media. | |
2418 | * | |
3be1f18e | 2419 | * When this succeeds, we just return. We have nothing |
7ebb6fca LT |
2420 | * left to unlink. |
2421 | */ | |
2422 | if (ret && ret != EEXIST) { | |
be66a6c4 | 2423 | try_rename: |
7ebb6fca | 2424 | if (!rename(tmpfile, filename)) |
3be1f18e | 2425 | goto out; |
9e48b389 | 2426 | ret = errno; |
230f1322 | 2427 | } |
691f1a28 | 2428 | unlink_or_warn(tmpfile); |
230f1322 LT |
2429 | if (ret) { |
2430 | if (ret != EEXIST) { | |
82247e9b | 2431 | return error("unable to write sha1 filename %s: %s", filename, strerror(ret)); |
230f1322 LT |
2432 | } |
2433 | /* FIXME!!! Collision check here ? */ | |
2434 | } | |
2435 | ||
3be1f18e | 2436 | out: |
5256b006 | 2437 | if (adjust_shared_perm(filename)) |
5a688fe4 | 2438 | return error("unable to set permission to '%s'", filename); |
230f1322 LT |
2439 | return 0; |
2440 | } | |
2441 | ||
4d548150 LT |
2442 | static int write_buffer(int fd, const void *buf, size_t len) |
2443 | { | |
d34cf19b | 2444 | if (write_in_full(fd, buf, len) < 0) |
93822c22 | 2445 | return error("file write error (%s)", strerror(errno)); |
4d548150 LT |
2446 | return 0; |
2447 | } | |
2448 | ||
ce9fbf16 | 2449 | int hash_sha1_file(const void *buf, unsigned long len, const char *type, |
abdc3fc8 RS |
2450 | unsigned char *sha1) |
2451 | { | |
d65a16f6 | 2452 | char hdr[32]; |
abdc3fc8 RS |
2453 | int hdrlen; |
2454 | write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen); | |
2455 | return 0; | |
2456 | } | |
2457 | ||
e9039dd3 LT |
2458 | /* Finalize a file on disk, and close it. */ |
2459 | static void close_sha1_file(int fd) | |
2460 | { | |
aafe9fba LT |
2461 | if (fsync_object_files) |
2462 | fsync_or_die(fd, "sha1 file"); | |
e9039dd3 | 2463 | if (close(fd) != 0) |
d824cbba | 2464 | die_errno("error when closing sha1 file"); |
e9039dd3 LT |
2465 | } |
2466 | ||
5723fe7e LT |
2467 | /* Size of directory component, including the ending '/' */ |
2468 | static inline int directory_size(const char *filename) | |
2469 | { | |
2470 | const char *s = strrchr(filename, '/'); | |
2471 | if (!s) | |
2472 | return 0; | |
2473 | return s - filename + 1; | |
2474 | } | |
2475 | ||
2476 | /* | |
2477 | * This creates a temporary file in the same directory as the final | |
2478 | * 'filename' | |
2479 | * | |
2480 | * We want to avoid cross-directory filename renames, because those | |
2481 | * can have problems on various filesystems (FAT, NFS, Coda). | |
2482 | */ | |
2483 | static int create_tmpfile(char *buffer, size_t bufsiz, const char *filename) | |
2484 | { | |
2485 | int fd, dirlen = directory_size(filename); | |
2486 | ||
2487 | if (dirlen + 20 > bufsiz) { | |
2488 | errno = ENAMETOOLONG; | |
2489 | return -1; | |
2490 | } | |
2491 | memcpy(buffer, filename, dirlen); | |
2492 | strcpy(buffer + dirlen, "tmp_obj_XXXXXX"); | |
5256b006 | 2493 | fd = git_mkstemp_mode(buffer, 0444); |
cbacbf4e | 2494 | if (fd < 0 && dirlen && errno == ENOENT) { |
5723fe7e | 2495 | /* Make sure the directory exists */ |
6ff6af62 | 2496 | memcpy(buffer, filename, dirlen); |
5723fe7e | 2497 | buffer[dirlen-1] = 0; |
3bfaf018 | 2498 | if (mkdir(buffer, 0777) || adjust_shared_perm(buffer)) |
5723fe7e LT |
2499 | return -1; |
2500 | ||
2501 | /* Try again */ | |
2502 | strcpy(buffer + dirlen - 1, "/tmp_obj_XXXXXX"); | |
5256b006 | 2503 | fd = git_mkstemp_mode(buffer, 0444); |
5723fe7e LT |
2504 | } |
2505 | return fd; | |
2506 | } | |
2507 | ||
bbac7311 | 2508 | static int write_loose_object(const unsigned char *sha1, char *hdr, int hdrlen, |
c00e657d | 2509 | const void *buf, unsigned long len, time_t mtime) |
0fcfd160 | 2510 | { |
915308b1 | 2511 | int fd, ret; |
9892beba | 2512 | unsigned char compressed[4096]; |
ef49a7a0 | 2513 | git_zstream stream; |
748af44c NP |
2514 | git_SHA_CTX c; |
2515 | unsigned char parano_sha1[20]; | |
706bc531 | 2516 | char *filename; |
ab1900a3 | 2517 | static char tmp_file[PATH_MAX]; |
a44c9a5e | 2518 | |
972a9155 | 2519 | filename = sha1_file_name(sha1); |
ab1900a3 | 2520 | fd = create_tmpfile(tmp_file, sizeof(tmp_file), filename); |
aac17941 | 2521 | if (fd < 0) { |
35243577 | 2522 | if (errno == EACCES) |
82247e9b | 2523 | return error("insufficient permission for adding an object to repository database %s", get_object_directory()); |
916d081b | 2524 | else |
5eaeda70 | 2525 | return error("unable to create temporary file: %s", strerror(errno)); |
aac17941 LT |
2526 | } |
2527 | ||
0fcfd160 LT |
2528 | /* Set it up */ |
2529 | memset(&stream, 0, sizeof(stream)); | |
55bb5c91 | 2530 | git_deflate_init(&stream, zlib_compression_level); |
0fcfd160 | 2531 | stream.next_out = compressed; |
9892beba | 2532 | stream.avail_out = sizeof(compressed); |
748af44c | 2533 | git_SHA1_Init(&c); |
a44c9a5e LT |
2534 | |
2535 | /* First header.. */ | |
d65a16f6 | 2536 | stream.next_in = (unsigned char *)hdr; |
a44c9a5e | 2537 | stream.avail_in = hdrlen; |
55bb5c91 JH |
2538 | while (git_deflate(&stream, 0) == Z_OK) |
2539 | ; /* nothing */ | |
748af44c | 2540 | git_SHA1_Update(&c, hdr, hdrlen); |
a44c9a5e LT |
2541 | |
2542 | /* Then the data itself.. */ | |
c00e657d | 2543 | stream.next_in = (void *)buf; |
a44c9a5e | 2544 | stream.avail_in = len; |
9892beba | 2545 | do { |
748af44c | 2546 | unsigned char *in0 = stream.next_in; |
55bb5c91 | 2547 | ret = git_deflate(&stream, Z_FINISH); |
748af44c | 2548 | git_SHA1_Update(&c, in0, stream.next_in - in0); |
9892beba NP |
2549 | if (write_buffer(fd, compressed, stream.next_out - compressed) < 0) |
2550 | die("unable to write sha1 file"); | |
2551 | stream.next_out = compressed; | |
2552 | stream.avail_out = sizeof(compressed); | |
2553 | } while (ret == Z_OK); | |
2554 | ||
ac54c277 LT |
2555 | if (ret != Z_STREAM_END) |
2556 | die("unable to deflate new object %s (%d)", sha1_to_hex(sha1), ret); | |
55bb5c91 | 2557 | ret = git_deflate_end_gently(&stream); |
ac54c277 LT |
2558 | if (ret != Z_OK) |
2559 | die("deflateEnd on object %s failed (%d)", sha1_to_hex(sha1), ret); | |
748af44c NP |
2560 | git_SHA1_Final(parano_sha1, &c); |
2561 | if (hashcmp(sha1, parano_sha1) != 0) | |
2562 | die("confused by unstable object source data for %s", sha1_to_hex(sha1)); | |
ac54c277 | 2563 | |
e9039dd3 | 2564 | close_sha1_file(fd); |
0fcfd160 | 2565 | |
bbac7311 NP |
2566 | if (mtime) { |
2567 | struct utimbuf utb; | |
2568 | utb.actime = mtime; | |
2569 | utb.modtime = mtime; | |
ab1900a3 | 2570 | if (utime(tmp_file, &utb) < 0) |
bbac7311 | 2571 | warning("failed utime() on %s: %s", |
ab1900a3 | 2572 | tmp_file, strerror(errno)); |
bbac7311 NP |
2573 | } |
2574 | ||
ab1900a3 | 2575 | return move_temp_to_file(tmp_file, filename); |
0fcfd160 | 2576 | } |
8237b185 | 2577 | |
c00e657d | 2578 | int write_sha1_file(const void *buf, unsigned long len, const char *type, unsigned char *returnsha1) |
bbac7311 NP |
2579 | { |
2580 | unsigned char sha1[20]; | |
2581 | char hdr[32]; | |
2582 | int hdrlen; | |
2583 | ||
2584 | /* Normally if we have it in the pack then we do not bother writing | |
2585 | * it out into .git/objects/??/?{38} file. | |
2586 | */ | |
2587 | write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen); | |
2588 | if (returnsha1) | |
2589 | hashcpy(returnsha1, sha1); | |
2590 | if (has_sha1_file(sha1)) | |
2591 | return 0; | |
2592 | return write_loose_object(sha1, hdr, hdrlen, buf, len, 0); | |
2593 | } | |
2594 | ||
2595 | int force_object_loose(const unsigned char *sha1, time_t mtime) | |
2596 | { | |
bbac7311 NP |
2597 | void *buf; |
2598 | unsigned long len; | |
2599 | enum object_type type; | |
2600 | char hdr[32]; | |
2601 | int hdrlen; | |
1fb23e65 | 2602 | int ret; |
bbac7311 | 2603 | |
c529d75a | 2604 | if (has_loose_object(sha1)) |
bbac7311 NP |
2605 | return 0; |
2606 | buf = read_packed_sha1(sha1, &type, &len); | |
2607 | if (!buf) | |
2608 | return error("cannot read sha1_file for %s", sha1_to_hex(sha1)); | |
2609 | hdrlen = sprintf(hdr, "%s %lu", typename(type), len) + 1; | |
1fb23e65 BS |
2610 | ret = write_loose_object(sha1, hdr, hdrlen, buf, len, mtime); |
2611 | free(buf); | |
2612 | ||
2613 | return ret; | |
bbac7311 NP |
2614 | } |
2615 | ||
bf592c50 DB |
2616 | int has_pack_index(const unsigned char *sha1) |
2617 | { | |
2618 | struct stat st; | |
2619 | if (stat(sha1_pack_index_name(sha1), &st)) | |
2620 | return 0; | |
2621 | return 1; | |
bf592c50 DB |
2622 | } |
2623 | ||
cd673c1f JH |
2624 | int has_sha1_pack(const unsigned char *sha1) |
2625 | { | |
2626 | struct pack_entry e; | |
2627 | return find_pack_entry(sha1, &e); | |
2628 | } | |
2629 | ||
8237b185 DB |
2630 | int has_sha1_file(const unsigned char *sha1) |
2631 | { | |
1f688557 JH |
2632 | struct pack_entry e; |
2633 | ||
cd673c1f | 2634 | if (find_pack_entry(sha1, &e)) |
1f688557 | 2635 | return 1; |
c529d75a | 2636 | return has_loose_object(sha1); |
8237b185 | 2637 | } |
74400e71 | 2638 | |
c879daa2 NTND |
2639 | static void check_tree(const void *buf, size_t size) |
2640 | { | |
2641 | struct tree_desc desc; | |
2642 | struct name_entry entry; | |
2643 | ||
2644 | init_tree_desc(&desc, buf, size); | |
2645 | while (tree_entry(&desc, &entry)) | |
2646 | /* do nothing | |
2647 | * tree_entry() will die() on malformed entries */ | |
2648 | ; | |
2649 | } | |
2650 | ||
2651 | static void check_commit(const void *buf, size_t size) | |
2652 | { | |
2653 | struct commit c; | |
2654 | memset(&c, 0, sizeof(c)); | |
2655 | if (parse_commit_buffer(&c, buf, size)) | |
2656 | die("corrupt commit"); | |
2657 | } | |
2658 | ||
2659 | static void check_tag(const void *buf, size_t size) | |
2660 | { | |
2661 | struct tag t; | |
2662 | memset(&t, 0, sizeof(t)); | |
2663 | if (parse_tag_buffer(&t, buf, size)) | |
2664 | die("corrupt tag"); | |
2665 | } | |
2666 | ||
43df4f86 | 2667 | static int index_mem(unsigned char *sha1, void *buf, size_t size, |
c4ce46fc JH |
2668 | enum object_type type, |
2669 | const char *path, unsigned flags) | |
e7332f96 | 2670 | { |
6c510bee | 2671 | int ret, re_allocated = 0; |
c4ce46fc | 2672 | int write_object = flags & HASH_WRITE_OBJECT; |
74400e71 | 2673 | |
7672db20 | 2674 | if (!type) |
edaec3fb | 2675 | type = OBJ_BLOB; |
6c510bee LT |
2676 | |
2677 | /* | |
2678 | * Convert blobs to git internal format | |
2679 | */ | |
43df4f86 | 2680 | if ((type == OBJ_BLOB) && path) { |
f285a2d7 | 2681 | struct strbuf nbuf = STRBUF_INIT; |
21e5ad50 | 2682 | if (convert_to_git(path, buf, size, &nbuf, |
5e12e78e | 2683 | write_object ? safe_crlf : SAFE_CRLF_FALSE)) { |
b315c5c0 | 2684 | buf = strbuf_detach(&nbuf, &size); |
6c510bee LT |
2685 | re_allocated = 1; |
2686 | } | |
2687 | } | |
c4ce46fc | 2688 | if (flags & HASH_FORMAT_CHECK) { |
c879daa2 NTND |
2689 | if (type == OBJ_TREE) |
2690 | check_tree(buf, size); | |
2691 | if (type == OBJ_COMMIT) | |
2692 | check_commit(buf, size); | |
2693 | if (type == OBJ_TAG) | |
2694 | check_tag(buf, size); | |
2695 | } | |
6c510bee | 2696 | |
7672db20 | 2697 | if (write_object) |
edaec3fb | 2698 | ret = write_sha1_file(buf, size, typename(type), sha1); |
abdc3fc8 | 2699 | else |
edaec3fb | 2700 | ret = hash_sha1_file(buf, size, typename(type), sha1); |
43df4f86 | 2701 | if (re_allocated) |
6c510bee | 2702 | free(buf); |
43df4f86 DP |
2703 | return ret; |
2704 | } | |
2705 | ||
7b41e1e1 JH |
2706 | static int index_pipe(unsigned char *sha1, int fd, enum object_type type, |
2707 | const char *path, unsigned flags) | |
2708 | { | |
2709 | struct strbuf sbuf = STRBUF_INIT; | |
2710 | int ret; | |
2711 | ||
2712 | if (strbuf_read(&sbuf, fd, 4096) >= 0) | |
2713 | ret = index_mem(sha1, sbuf.buf, sbuf.len, type, path, flags); | |
2714 | else | |
2715 | ret = -1; | |
2716 | strbuf_release(&sbuf); | |
2717 | return ret; | |
2718 | } | |
2719 | ||
ea68b0ce DP |
2720 | #define SMALL_FILE_SIZE (32*1024) |
2721 | ||
7b41e1e1 JH |
2722 | static int index_core(unsigned char *sha1, int fd, size_t size, |
2723 | enum object_type type, const char *path, | |
2724 | unsigned flags) | |
43df4f86 DP |
2725 | { |
2726 | int ret; | |
43df4f86 | 2727 | |
7b41e1e1 | 2728 | if (!size) { |
c4ce46fc | 2729 | ret = index_mem(sha1, NULL, size, type, path, flags); |
ea68b0ce DP |
2730 | } else if (size <= SMALL_FILE_SIZE) { |
2731 | char *buf = xmalloc(size); | |
2732 | if (size == read_in_full(fd, buf, size)) | |
c4ce46fc | 2733 | ret = index_mem(sha1, buf, size, type, path, flags); |
ea68b0ce DP |
2734 | else |
2735 | ret = error("short read %s", strerror(errno)); | |
2736 | free(buf); | |
08bda208 | 2737 | } else { |
43df4f86 | 2738 | void *buf = xmmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0); |
c4ce46fc | 2739 | ret = index_mem(sha1, buf, size, type, path, flags); |
aac17941 | 2740 | munmap(buf, size); |
08bda208 | 2741 | } |
7b41e1e1 JH |
2742 | return ret; |
2743 | } | |
2744 | ||
4dd1fbc7 | 2745 | /* |
568508e7 JH |
2746 | * This creates one packfile per large blob unless bulk-checkin |
2747 | * machinery is "plugged". | |
4dd1fbc7 JH |
2748 | * |
2749 | * This also bypasses the usual "convert-to-git" dance, and that is on | |
2750 | * purpose. We could write a streaming version of the converting | |
2751 | * functions and insert that before feeding the data to fast-import | |
4f22b101 JK |
2752 | * (or equivalent in-core API described above). However, that is |
2753 | * somewhat complicated, as we do not know the size of the filter | |
2754 | * result, which we need to know beforehand when writing a git object. | |
2755 | * Since the primary motivation for trying to stream from the working | |
2756 | * tree file and to avoid mmaping it in core is to deal with large | |
2757 | * binary blobs, they generally do not want to get any conversion, and | |
2758 | * callers should avoid this code path when filters are requested. | |
4dd1fbc7 JH |
2759 | */ |
2760 | static int index_stream(unsigned char *sha1, int fd, size_t size, | |
2761 | enum object_type type, const char *path, | |
2762 | unsigned flags) | |
2763 | { | |
568508e7 | 2764 | return index_bulk_checkin(sha1, fd, size, type, path, flags); |
4dd1fbc7 JH |
2765 | } |
2766 | ||
7b41e1e1 JH |
2767 | int index_fd(unsigned char *sha1, int fd, struct stat *st, |
2768 | enum object_type type, const char *path, unsigned flags) | |
2769 | { | |
2770 | int ret; | |
2771 | size_t size = xsize_t(st->st_size); | |
2772 | ||
2773 | if (!S_ISREG(st->st_mode)) | |
2774 | ret = index_pipe(sha1, fd, type, path, flags); | |
4f22b101 JK |
2775 | else if (size <= big_file_threshold || type != OBJ_BLOB || |
2776 | (path && would_convert_to_git(path, NULL, 0, 0))) | |
7b41e1e1 | 2777 | ret = index_core(sha1, fd, size, type, path, flags); |
4dd1fbc7 JH |
2778 | else |
2779 | ret = index_stream(sha1, fd, size, type, path, flags); | |
43df4f86 | 2780 | close(fd); |
aac17941 | 2781 | return ret; |
74400e71 | 2782 | } |
ec1fcc16 | 2783 | |
c4ce46fc | 2784 | int index_path(unsigned char *sha1, const char *path, struct stat *st, unsigned flags) |
ec1fcc16 JH |
2785 | { |
2786 | int fd; | |
b760d3aa | 2787 | struct strbuf sb = STRBUF_INIT; |
ec1fcc16 JH |
2788 | |
2789 | switch (st->st_mode & S_IFMT) { | |
2790 | case S_IFREG: | |
2791 | fd = open(path, O_RDONLY); | |
2792 | if (fd < 0) | |
2793 | return error("open(\"%s\"): %s", path, | |
2794 | strerror(errno)); | |
c4ce46fc | 2795 | if (index_fd(sha1, fd, st, OBJ_BLOB, path, flags) < 0) |
ec1fcc16 JH |
2796 | return error("%s: failed to insert into database", |
2797 | path); | |
2798 | break; | |
2799 | case S_IFLNK: | |
b760d3aa | 2800 | if (strbuf_readlink(&sb, path, st->st_size)) { |
ec1fcc16 | 2801 | char *errstr = strerror(errno); |
ec1fcc16 JH |
2802 | return error("readlink(\"%s\"): %s", path, |
2803 | errstr); | |
2804 | } | |
c4ce46fc | 2805 | if (!(flags & HASH_WRITE_OBJECT)) |
b760d3aa LT |
2806 | hash_sha1_file(sb.buf, sb.len, blob_type, sha1); |
2807 | else if (write_sha1_file(sb.buf, sb.len, blob_type, sha1)) | |
ec1fcc16 JH |
2808 | return error("%s: failed to insert into database", |
2809 | path); | |
b760d3aa | 2810 | strbuf_release(&sb); |
ec1fcc16 | 2811 | break; |
f35a6d3b LT |
2812 | case S_IFDIR: |
2813 | return resolve_gitlink_ref(path, "HEAD", sha1); | |
ec1fcc16 JH |
2814 | default: |
2815 | return error("%s: unsupported file type", path); | |
2816 | } | |
2817 | return 0; | |
2818 | } | |
a69e5429 JH |
2819 | |
2820 | int read_pack_header(int fd, struct pack_header *header) | |
2821 | { | |
c697ad14 HO |
2822 | if (read_in_full(fd, header, sizeof(*header)) < sizeof(*header)) |
2823 | /* "eof before pack header was fully read" */ | |
2824 | return PH_ERROR_EOF; | |
2825 | ||
a69e5429 JH |
2826 | if (header->hdr_signature != htonl(PACK_SIGNATURE)) |
2827 | /* "protocol error (pack signature mismatch detected)" */ | |
2828 | return PH_ERROR_PACK_SIGNATURE; | |
2829 | if (!pack_version_ok(header->hdr_version)) | |
2830 | /* "protocol error (pack version unsupported)" */ | |
2831 | return PH_ERROR_PROTOCOL; | |
2832 | return 0; | |
2833 | } | |
40d52ff7 JK |
2834 | |
2835 | void assert_sha1_type(const unsigned char *sha1, enum object_type expect) | |
2836 | { | |
2837 | enum object_type type = sha1_object_info(sha1, NULL); | |
2838 | if (type < 0) | |
2839 | die("%s is not a valid object", sha1_to_hex(sha1)); | |
2840 | if (type != expect) | |
2841 | die("%s is not a valid '%s' object", sha1_to_hex(sha1), | |
2842 | typename(expect)); | |
2843 | } |