]>
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" | |
14 | #include "tag.h" | |
15 | #include "tree.h" | |
f35a6d3b | 16 | #include "refs.h" |
70f5d5d3 | 17 | #include "pack-revindex.h" |
628522ec | 18 | #include "sha1-lookup.h" |
0fcfd160 | 19 | |
144bde78 LT |
20 | #ifndef O_NOATIME |
21 | #if defined(__linux__) && (defined(__i386__) || defined(__PPC__)) | |
22 | #define O_NOATIME 01000000 | |
23 | #else | |
24 | #define O_NOATIME 0 | |
25 | #endif | |
26 | #endif | |
27 | ||
e05db0fd PR |
28 | #ifdef NO_C99_FORMAT |
29 | #define SZ_FMT "lu" | |
9e42d6a1 | 30 | static unsigned long sz_fmt(size_t s) { return (unsigned long)s; } |
e05db0fd PR |
31 | #else |
32 | #define SZ_FMT "zu" | |
9e42d6a1 | 33 | static size_t sz_fmt(size_t s) { return s; } |
e05db0fd PR |
34 | #endif |
35 | ||
96f1e58f | 36 | const unsigned char null_sha1[20]; |
88cd621d | 37 | |
192a6be2 | 38 | const signed char hexval_table[256] = { |
e49521b5 JH |
39 | -1, -1, -1, -1, -1, -1, -1, -1, /* 00-07 */ |
40 | -1, -1, -1, -1, -1, -1, -1, -1, /* 08-0f */ | |
41 | -1, -1, -1, -1, -1, -1, -1, -1, /* 10-17 */ | |
42 | -1, -1, -1, -1, -1, -1, -1, -1, /* 18-1f */ | |
43 | -1, -1, -1, -1, -1, -1, -1, -1, /* 20-27 */ | |
44 | -1, -1, -1, -1, -1, -1, -1, -1, /* 28-2f */ | |
45 | 0, 1, 2, 3, 4, 5, 6, 7, /* 30-37 */ | |
46 | 8, 9, -1, -1, -1, -1, -1, -1, /* 38-3f */ | |
47 | -1, 10, 11, 12, 13, 14, 15, -1, /* 40-47 */ | |
48 | -1, -1, -1, -1, -1, -1, -1, -1, /* 48-4f */ | |
49 | -1, -1, -1, -1, -1, -1, -1, -1, /* 50-57 */ | |
50 | -1, -1, -1, -1, -1, -1, -1, -1, /* 58-5f */ | |
51 | -1, 10, 11, 12, 13, 14, 15, -1, /* 60-67 */ | |
52 | -1, -1, -1, -1, -1, -1, -1, -1, /* 68-67 */ | |
53 | -1, -1, -1, -1, -1, -1, -1, -1, /* 70-77 */ | |
54 | -1, -1, -1, -1, -1, -1, -1, -1, /* 78-7f */ | |
55 | -1, -1, -1, -1, -1, -1, -1, -1, /* 80-87 */ | |
56 | -1, -1, -1, -1, -1, -1, -1, -1, /* 88-8f */ | |
57 | -1, -1, -1, -1, -1, -1, -1, -1, /* 90-97 */ | |
58 | -1, -1, -1, -1, -1, -1, -1, -1, /* 98-9f */ | |
59 | -1, -1, -1, -1, -1, -1, -1, -1, /* a0-a7 */ | |
60 | -1, -1, -1, -1, -1, -1, -1, -1, /* a8-af */ | |
61 | -1, -1, -1, -1, -1, -1, -1, -1, /* b0-b7 */ | |
62 | -1, -1, -1, -1, -1, -1, -1, -1, /* b8-bf */ | |
63 | -1, -1, -1, -1, -1, -1, -1, -1, /* c0-c7 */ | |
64 | -1, -1, -1, -1, -1, -1, -1, -1, /* c8-cf */ | |
65 | -1, -1, -1, -1, -1, -1, -1, -1, /* d0-d7 */ | |
66 | -1, -1, -1, -1, -1, -1, -1, -1, /* d8-df */ | |
67 | -1, -1, -1, -1, -1, -1, -1, -1, /* e0-e7 */ | |
68 | -1, -1, -1, -1, -1, -1, -1, -1, /* e8-ef */ | |
69 | -1, -1, -1, -1, -1, -1, -1, -1, /* f0-f7 */ | |
70 | -1, -1, -1, -1, -1, -1, -1, -1, /* f8-ff */ | |
71 | }; | |
0fcfd160 LT |
72 | |
73 | int get_sha1_hex(const char *hex, unsigned char *sha1) | |
74 | { | |
75 | int i; | |
76 | for (i = 0; i < 20; i++) { | |
77 | unsigned int val = (hexval(hex[0]) << 4) | hexval(hex[1]); | |
78 | if (val & ~0xff) | |
79 | return -1; | |
80 | *sha1++ = val; | |
81 | hex += 2; | |
82 | } | |
83 | return 0; | |
84 | } | |
85 | ||
8385abfd JS |
86 | static inline int offset_1st_component(const char *path) |
87 | { | |
88 | if (has_dos_drive_prefix(path)) | |
89 | return 2 + (path[2] == '/'); | |
90 | return *path == '/'; | |
91 | } | |
92 | ||
b2cb9425 JH |
93 | int safe_create_leading_directories(char *path) |
94 | { | |
8385abfd | 95 | char *pos = path + offset_1st_component(path); |
67d42212 JR |
96 | struct stat st; |
97 | ||
b2cb9425 JH |
98 | while (pos) { |
99 | pos = strchr(pos, '/'); | |
100 | if (!pos) | |
101 | break; | |
5f0bdf50 JH |
102 | while (*++pos == '/') |
103 | ; | |
104 | if (!*pos) | |
105 | break; | |
106 | *--pos = '\0'; | |
67d42212 JR |
107 | if (!stat(path, &st)) { |
108 | /* path exists */ | |
109 | if (!S_ISDIR(st.st_mode)) { | |
b2cb9425 | 110 | *pos = '/'; |
67d42212 | 111 | return -3; |
b2cb9425 | 112 | } |
457f06d6 | 113 | } |
67d42212 JR |
114 | else if (mkdir(path, 0777)) { |
115 | *pos = '/'; | |
116 | return -1; | |
117 | } | |
457f06d6 JS |
118 | else if (adjust_shared_perm(path)) { |
119 | *pos = '/'; | |
120 | return -2; | |
121 | } | |
b2cb9425 JH |
122 | *pos++ = '/'; |
123 | } | |
124 | return 0; | |
125 | } | |
723c31fe | 126 | |
8e21d63b JK |
127 | int safe_create_leading_directories_const(const char *path) |
128 | { | |
129 | /* path points to cache entries, so xstrdup before messing with it */ | |
130 | char *buf = xstrdup(path); | |
131 | int result = safe_create_leading_directories(buf); | |
132 | free(buf); | |
133 | return result; | |
134 | } | |
135 | ||
3bfaf018 | 136 | char *sha1_to_hex(const unsigned char *sha1) |
0fcfd160 | 137 | { |
dcb3450f LT |
138 | static int bufno; |
139 | static char hexbuffer[4][50]; | |
0fcfd160 | 140 | static const char hex[] = "0123456789abcdef"; |
dcb3450f | 141 | char *buffer = hexbuffer[3 & ++bufno], *buf = buffer; |
0fcfd160 LT |
142 | int i; |
143 | ||
144 | for (i = 0; i < 20; i++) { | |
145 | unsigned int val = *sha1++; | |
146 | *buf++ = hex[val >> 4]; | |
147 | *buf++ = hex[val & 0xf]; | |
148 | } | |
1e80e044 JS |
149 | *buf = '\0'; |
150 | ||
0fcfd160 LT |
151 | return buffer; |
152 | } | |
153 | ||
ace1534d JH |
154 | static void fill_sha1_path(char *pathbuf, const unsigned char *sha1) |
155 | { | |
156 | int i; | |
157 | for (i = 0; i < 20; i++) { | |
158 | static char hex[] = "0123456789abcdef"; | |
159 | unsigned int val = sha1[i]; | |
160 | char *pos = pathbuf + i*2 + (i > 0); | |
161 | *pos++ = hex[val >> 4]; | |
162 | *pos = hex[val & 0xf]; | |
163 | } | |
164 | } | |
165 | ||
0fcfd160 LT |
166 | /* |
167 | * NOTE! This returns a statically allocated buffer, so you have to be | |
790296fd | 168 | * careful about using it. Do an "xstrdup()" if you need to save the |
0fcfd160 | 169 | * filename. |
ace1534d JH |
170 | * |
171 | * Also note that this returns the location for creating. Reading | |
172 | * SHA1 file can happen from any alternate directory listed in the | |
d19938ab | 173 | * DB_ENVIRONMENT environment variable if it is not found in |
ace1534d | 174 | * the primary object database. |
0fcfd160 LT |
175 | */ |
176 | char *sha1_file_name(const unsigned char *sha1) | |
177 | { | |
0fcfd160 LT |
178 | static char *name, *base; |
179 | ||
180 | if (!base) { | |
d19938ab | 181 | const char *sha1_file_directory = get_object_directory(); |
0fcfd160 | 182 | int len = strlen(sha1_file_directory); |
812666c8 | 183 | base = xmalloc(len + 60); |
0fcfd160 LT |
184 | memcpy(base, sha1_file_directory, len); |
185 | memset(base+len, 0, 60); | |
186 | base[len] = '/'; | |
187 | base[len+3] = '/'; | |
188 | name = base + len + 1; | |
189 | } | |
ace1534d | 190 | fill_sha1_path(name, sha1); |
0fcfd160 LT |
191 | return base; |
192 | } | |
193 | ||
633f43e1 | 194 | static char *sha1_get_pack_name(const unsigned char *sha1, |
6eec46bd | 195 | char **name, char **base, const char *which) |
bf592c50 DB |
196 | { |
197 | static const char hex[] = "0123456789abcdef"; | |
633f43e1 | 198 | char *buf; |
bf592c50 DB |
199 | int i; |
200 | ||
633f43e1 | 201 | if (!*base) { |
bf592c50 DB |
202 | const char *sha1_file_directory = get_object_directory(); |
203 | int len = strlen(sha1_file_directory); | |
633f43e1 | 204 | *base = xmalloc(len + 60); |
6eec46bd JH |
205 | sprintf(*base, "%s/pack/pack-1234567890123456789012345678901234567890.%s", |
206 | sha1_file_directory, which); | |
633f43e1 | 207 | *name = *base + len + 11; |
bf592c50 DB |
208 | } |
209 | ||
633f43e1 | 210 | buf = *name; |
bf592c50 DB |
211 | |
212 | for (i = 0; i < 20; i++) { | |
213 | unsigned int val = *sha1++; | |
214 | *buf++ = hex[val >> 4]; | |
215 | *buf++ = hex[val & 0xf]; | |
216 | } | |
a6080a0a | 217 | |
633f43e1 | 218 | return *base; |
bf592c50 DB |
219 | } |
220 | ||
633f43e1 | 221 | char *sha1_pack_name(const unsigned char *sha1) |
bf592c50 | 222 | { |
633f43e1 | 223 | static char *name, *base; |
bf592c50 | 224 | |
6eec46bd | 225 | return sha1_get_pack_name(sha1, &name, &base, "pack"); |
633f43e1 | 226 | } |
bf592c50 | 227 | |
633f43e1 HO |
228 | char *sha1_pack_index_name(const unsigned char *sha1) |
229 | { | |
230 | static char *name, *base; | |
a6080a0a | 231 | |
6eec46bd | 232 | return sha1_get_pack_name(sha1, &name, &base, "idx"); |
bf592c50 DB |
233 | } |
234 | ||
d5a63b99 JH |
235 | struct alternate_object_database *alt_odb_list; |
236 | static struct alternate_object_database **alt_odb_tail; | |
ace1534d | 237 | |
c2f493a4 MW |
238 | static void read_info_alternates(const char * alternates, int depth); |
239 | ||
ddd5d056 JH |
240 | /* |
241 | * Prepare alternate object database registry. | |
d5a63b99 JH |
242 | * |
243 | * The variable alt_odb_list points at the list of struct | |
244 | * alternate_object_database. The elements on this list come from | |
245 | * non-empty elements from colon separated ALTERNATE_DB_ENVIRONMENT | |
246 | * environment variable, and $GIT_OBJECT_DIRECTORY/info/alternates, | |
1494e038 JH |
247 | * whose contents is similar to that environment variable but can be |
248 | * LF separated. Its base points at a statically allocated buffer that | |
d5a63b99 JH |
249 | * contains "/the/directory/corresponding/to/.git/objects/...", while |
250 | * its name points just after the slash at the end of ".git/objects/" | |
251 | * in the example above, and has enough space to hold 40-byte hex | |
252 | * SHA1, an extra slash for the first level indirection, and the | |
253 | * terminating NUL. | |
ddd5d056 | 254 | */ |
c2f493a4 | 255 | static int link_alt_odb_entry(const char * entry, int len, const char * relative_base, int depth) |
ace1534d | 256 | { |
1494e038 | 257 | const char *objdir = get_object_directory(); |
c2f493a4 MW |
258 | struct alternate_object_database *ent; |
259 | struct alternate_object_database *alt; | |
260 | /* 43 = 40-byte + 2 '/' + terminating NUL */ | |
261 | int pfxlen = len; | |
262 | int entlen = pfxlen + 43; | |
ccfd3e99 | 263 | int base_len = -1; |
d5a63b99 | 264 | |
85dadc38 | 265 | if (!is_absolute_path(entry) && relative_base) { |
c2f493a4 MW |
266 | /* Relative alt-odb */ |
267 | if (base_len < 0) | |
268 | base_len = strlen(relative_base) + 1; | |
269 | entlen += base_len; | |
270 | pfxlen += base_len; | |
271 | } | |
272 | ent = xmalloc(sizeof(*ent) + entlen); | |
273 | ||
85dadc38 | 274 | if (!is_absolute_path(entry) && relative_base) { |
c2f493a4 MW |
275 | memcpy(ent->base, relative_base, base_len - 1); |
276 | ent->base[base_len - 1] = '/'; | |
277 | memcpy(ent->base + base_len, entry, len); | |
278 | } | |
279 | else | |
280 | memcpy(ent->base, entry, pfxlen); | |
281 | ||
282 | ent->name = ent->base + pfxlen + 1; | |
283 | ent->base[pfxlen + 3] = '/'; | |
284 | ent->base[pfxlen] = ent->base[entlen-1] = 0; | |
285 | ||
286 | /* Detect cases where alternate disappeared */ | |
90b4a71c | 287 | if (!is_directory(ent->base)) { |
c2f493a4 MW |
288 | error("object directory %s does not exist; " |
289 | "check .git/objects/info/alternates.", | |
290 | ent->base); | |
291 | free(ent); | |
292 | return -1; | |
293 | } | |
294 | ||
295 | /* Prevent the common mistake of listing the same | |
296 | * thing twice, or object directory itself. | |
297 | */ | |
298 | for (alt = alt_odb_list; alt; alt = alt->next) { | |
299 | if (!memcmp(ent->base, alt->base, pfxlen)) { | |
300 | free(ent); | |
301 | return -1; | |
302 | } | |
303 | } | |
304 | if (!memcmp(ent->base, objdir, pfxlen)) { | |
305 | free(ent); | |
306 | return -1; | |
307 | } | |
308 | ||
309 | /* add the alternate entry */ | |
310 | *alt_odb_tail = ent; | |
311 | alt_odb_tail = &(ent->next); | |
312 | ent->next = NULL; | |
313 | ||
314 | /* recursively add alternates */ | |
315 | read_info_alternates(ent->base, depth + 1); | |
316 | ||
317 | ent->base[pfxlen] = '/'; | |
318 | ||
319 | return 0; | |
320 | } | |
321 | ||
322 | static void link_alt_odb_entries(const char *alt, const char *ep, int sep, | |
323 | const char *relative_base, int depth) | |
324 | { | |
325 | const char *cp, *last; | |
326 | ||
327 | if (depth > 5) { | |
328 | error("%s: ignoring alternate object stores, nesting too deep.", | |
329 | relative_base); | |
330 | return; | |
331 | } | |
332 | ||
d5a63b99 | 333 | last = alt; |
9577e7e3 JH |
334 | while (last < ep) { |
335 | cp = last; | |
336 | if (cp < ep && *cp == '#') { | |
337 | while (cp < ep && *cp != sep) | |
338 | cp++; | |
339 | last = cp + 1; | |
340 | continue; | |
341 | } | |
c2f493a4 MW |
342 | while (cp < ep && *cp != sep) |
343 | cp++; | |
d5a63b99 | 344 | if (last != cp) { |
85dadc38 | 345 | if (!is_absolute_path(last) && depth) { |
c2f493a4 MW |
346 | error("%s: ignoring relative alternate object store %s", |
347 | relative_base, last); | |
348 | } else { | |
349 | link_alt_odb_entry(last, cp - last, | |
350 | relative_base, depth); | |
1494e038 | 351 | } |
d5a63b99 | 352 | } |
9577e7e3 | 353 | while (cp < ep && *cp == sep) |
d5a63b99 JH |
354 | cp++; |
355 | last = cp; | |
9577e7e3 | 356 | } |
d5a63b99 JH |
357 | } |
358 | ||
c2f493a4 | 359 | static void read_info_alternates(const char * relative_base, int depth) |
d5a63b99 | 360 | { |
9577e7e3 | 361 | char *map; |
dc49cd76 | 362 | size_t mapsz; |
d5a63b99 | 363 | struct stat st; |
9cb18f56 JM |
364 | const char alt_file_name[] = "info/alternates"; |
365 | /* Given that relative_base is no longer than PATH_MAX, | |
366 | ensure that "path" has enough space to append "/", the | |
367 | file name, "info/alternates", and a trailing NUL. */ | |
368 | char path[PATH_MAX + 1 + sizeof alt_file_name]; | |
c2f493a4 | 369 | int fd; |
d5a63b99 | 370 | |
9cb18f56 | 371 | sprintf(path, "%s/%s", relative_base, alt_file_name); |
d5a63b99 JH |
372 | fd = open(path, O_RDONLY); |
373 | if (fd < 0) | |
374 | return; | |
375 | if (fstat(fd, &st) || (st.st_size == 0)) { | |
376 | close(fd); | |
9a217f2a | 377 | return; |
ace1534d | 378 | } |
dc49cd76 SP |
379 | mapsz = xsize_t(st.st_size); |
380 | map = xmmap(NULL, mapsz, PROT_READ, MAP_PRIVATE, fd, 0); | |
d5a63b99 | 381 | close(fd); |
d5a63b99 | 382 | |
dc49cd76 | 383 | link_alt_odb_entries(map, map + mapsz, '\n', relative_base, depth); |
c2f493a4 | 384 | |
dc49cd76 | 385 | munmap(map, mapsz); |
ace1534d JH |
386 | } |
387 | ||
bef70b22 DB |
388 | void add_to_alternates_file(const char *reference) |
389 | { | |
390 | struct lock_file *lock = xcalloc(1, sizeof(struct lock_file)); | |
acd3b9ec | 391 | int fd = hold_lock_file_for_append(lock, git_path("objects/info/alternates"), LOCK_DIE_ON_ERROR); |
bef70b22 DB |
392 | char *alt = mkpath("%s/objects\n", reference); |
393 | write_or_die(fd, alt, strlen(alt)); | |
394 | if (commit_lock_file(lock)) | |
395 | die("could not close alternates file"); | |
396 | if (alt_odb_tail) | |
397 | link_alt_odb_entries(alt, alt + strlen(alt), '\n', NULL, 0); | |
398 | } | |
399 | ||
d79796bc JH |
400 | void foreach_alt_odb(alt_odb_fn fn, void *cb) |
401 | { | |
402 | struct alternate_object_database *ent; | |
403 | ||
404 | prepare_alt_odb(); | |
405 | for (ent = alt_odb_list; ent; ent = ent->next) | |
406 | if (fn(ent, cb)) | |
407 | return; | |
408 | } | |
409 | ||
c2f493a4 MW |
410 | void prepare_alt_odb(void) |
411 | { | |
554fe20d | 412 | const char *alt; |
c2f493a4 | 413 | |
7dc24aa5 SP |
414 | if (alt_odb_tail) |
415 | return; | |
416 | ||
c2f493a4 MW |
417 | alt = getenv(ALTERNATE_DB_ENVIRONMENT); |
418 | if (!alt) alt = ""; | |
419 | ||
c2f493a4 | 420 | alt_odb_tail = &alt_odb_list; |
80ba074f | 421 | link_alt_odb_entries(alt, alt + strlen(alt), PATH_SEP, NULL, 0); |
c2f493a4 MW |
422 | |
423 | read_info_alternates(get_object_directory(), 0); | |
424 | } | |
425 | ||
0f4dc14a | 426 | static int has_loose_object_local(const unsigned char *sha1) |
ace1534d | 427 | { |
ace1534d | 428 | char *name = sha1_file_name(sha1); |
0f4dc14a BC |
429 | return !access(name, F_OK); |
430 | } | |
ace1534d | 431 | |
0f4dc14a BC |
432 | int has_loose_object_nonlocal(const unsigned char *sha1) |
433 | { | |
434 | struct alternate_object_database *alt; | |
9a217f2a | 435 | prepare_alt_odb(); |
d5a63b99 | 436 | for (alt = alt_odb_list; alt; alt = alt->next) { |
0f4dc14a | 437 | fill_sha1_path(alt->name, sha1); |
c529d75a LT |
438 | if (!access(alt->base, F_OK)) |
439 | return 1; | |
ace1534d | 440 | } |
c529d75a | 441 | return 0; |
ace1534d JH |
442 | } |
443 | ||
0f4dc14a BC |
444 | static int has_loose_object(const unsigned char *sha1) |
445 | { | |
446 | return has_loose_object_local(sha1) || | |
447 | has_loose_object_nonlocal(sha1); | |
448 | } | |
449 | ||
60bb8b14 | 450 | static unsigned int pack_used_ctr; |
a53128b6 SP |
451 | static unsigned int pack_mmap_calls; |
452 | static unsigned int peak_pack_open_windows; | |
453 | static unsigned int pack_open_windows; | |
454 | static size_t peak_pack_mapped; | |
60bb8b14 | 455 | static size_t pack_mapped; |
9a217f2a | 456 | struct packed_git *packed_git; |
1f688557 | 457 | |
b79d18c9 | 458 | void pack_report(void) |
a53128b6 SP |
459 | { |
460 | fprintf(stderr, | |
e05db0fd PR |
461 | "pack_report: getpagesize() = %10" SZ_FMT "\n" |
462 | "pack_report: core.packedGitWindowSize = %10" SZ_FMT "\n" | |
463 | "pack_report: core.packedGitLimit = %10" SZ_FMT "\n", | |
9e42d6a1 SP |
464 | sz_fmt(getpagesize()), |
465 | sz_fmt(packed_git_window_size), | |
466 | sz_fmt(packed_git_limit)); | |
a53128b6 SP |
467 | fprintf(stderr, |
468 | "pack_report: pack_used_ctr = %10u\n" | |
469 | "pack_report: pack_mmap_calls = %10u\n" | |
470 | "pack_report: pack_open_windows = %10u / %10u\n" | |
e05db0fd PR |
471 | "pack_report: pack_mapped = " |
472 | "%10" SZ_FMT " / %10" SZ_FMT "\n", | |
a53128b6 SP |
473 | pack_used_ctr, |
474 | pack_mmap_calls, | |
475 | pack_open_windows, peak_pack_open_windows, | |
9e42d6a1 | 476 | sz_fmt(pack_mapped), sz_fmt(peak_pack_mapped)); |
a53128b6 SP |
477 | } |
478 | ||
42873078 | 479 | static int check_packed_git_idx(const char *path, struct packed_git *p) |
1f688557 JH |
480 | { |
481 | void *idx_map; | |
42873078 | 482 | struct pack_idx_header *hdr; |
dc49cd76 | 483 | size_t idx_size; |
74e34e1f | 484 | uint32_t version, nr, i, *index; |
e93ec6f9 | 485 | int fd = open(path, O_RDONLY); |
1f688557 | 486 | struct stat st; |
42873078 | 487 | |
1f688557 JH |
488 | if (fd < 0) |
489 | return -1; | |
490 | if (fstat(fd, &st)) { | |
491 | close(fd); | |
492 | return -1; | |
493 | } | |
dc49cd76 | 494 | idx_size = xsize_t(st.st_size); |
2d88451b SP |
495 | if (idx_size < 4 * 256 + 20 + 20) { |
496 | close(fd); | |
497 | return error("index file %s is too small", path); | |
498 | } | |
c4712e45 | 499 | idx_map = xmmap(NULL, idx_size, PROT_READ, MAP_PRIVATE, fd, 0); |
1f688557 | 500 | close(fd); |
1f688557 | 501 | |
42873078 NP |
502 | hdr = idx_map; |
503 | if (hdr->idx_signature == htonl(PACK_IDX_SIGNATURE)) { | |
74e34e1f NP |
504 | version = ntohl(hdr->idx_version); |
505 | if (version < 2 || version > 2) { | |
506 | munmap(idx_map, idx_size); | |
6e1c2344 | 507 | return error("index file %s is version %"PRIu32 |
74e34e1f NP |
508 | " and is not supported by this binary" |
509 | " (try upgrading GIT to a newer version)", | |
510 | path, version); | |
511 | } | |
512 | } else | |
513 | version = 1; | |
df1b059d | 514 | |
1f688557 | 515 | nr = 0; |
42873078 | 516 | index = idx_map; |
74e34e1f NP |
517 | if (version > 1) |
518 | index += 2; /* skip index header */ | |
1f688557 | 519 | for (i = 0; i < 256; i++) { |
326bf396 | 520 | uint32_t n = ntohl(index[i]); |
2d88451b SP |
521 | if (n < nr) { |
522 | munmap(idx_map, idx_size); | |
df1b059d | 523 | return error("non-monotonic index %s", path); |
2d88451b | 524 | } |
1f688557 JH |
525 | nr = n; |
526 | } | |
527 | ||
74e34e1f NP |
528 | if (version == 1) { |
529 | /* | |
530 | * Total size: | |
531 | * - 256 index entries 4 bytes each | |
532 | * - 24-byte entries * nr (20-byte sha1 + 4-byte offset) | |
533 | * - 20-byte SHA1 of the packfile | |
534 | * - 20-byte SHA1 file checksum | |
535 | */ | |
536 | if (idx_size != 4*256 + nr * 24 + 20 + 20) { | |
537 | munmap(idx_map, idx_size); | |
eef427a0 | 538 | return error("wrong index v1 file size in %s", path); |
74e34e1f NP |
539 | } |
540 | } else if (version == 2) { | |
541 | /* | |
542 | * Minimum size: | |
543 | * - 8 bytes of header | |
544 | * - 256 index entries 4 bytes each | |
545 | * - 20-byte sha1 entry * nr | |
546 | * - 4-byte crc entry * nr | |
547 | * - 4-byte offset entry * nr | |
548 | * - 20-byte SHA1 of the packfile | |
549 | * - 20-byte SHA1 file checksum | |
550 | * And after the 4-byte offset table might be a | |
551 | * variable sized table containing 8-byte entries | |
552 | * for offsets larger than 2^31. | |
553 | */ | |
554 | unsigned long min_size = 8 + 4*256 + nr*(20 + 4 + 4) + 20 + 20; | |
1164f1e4 LT |
555 | unsigned long max_size = min_size; |
556 | if (nr) | |
557 | max_size += (nr - 1)*8; | |
558 | if (idx_size < min_size || idx_size > max_size) { | |
74e34e1f | 559 | munmap(idx_map, idx_size); |
eef427a0 | 560 | return error("wrong index v2 file size in %s", path); |
74e34e1f | 561 | } |
7109c889 JH |
562 | if (idx_size != min_size && |
563 | /* | |
564 | * make sure we can deal with large pack offsets. | |
565 | * 31-bit signed offset won't be enough, neither | |
566 | * 32-bit unsigned one will be. | |
567 | */ | |
568 | (sizeof(off_t) <= 4)) { | |
569 | munmap(idx_map, idx_size); | |
570 | return error("pack too large for current definition of off_t in %s", path); | |
74e34e1f | 571 | } |
2d88451b | 572 | } |
1f688557 | 573 | |
74e34e1f | 574 | p->index_version = version; |
42873078 NP |
575 | p->index_data = idx_map; |
576 | p->index_size = idx_size; | |
57059091 | 577 | p->num_objects = nr; |
1f688557 JH |
578 | return 0; |
579 | } | |
580 | ||
bc8e478a | 581 | int open_pack_index(struct packed_git *p) |
d079837e SP |
582 | { |
583 | char *idx_name; | |
584 | int ret; | |
585 | ||
586 | if (p->index_data) | |
587 | return 0; | |
588 | ||
589 | idx_name = xstrdup(p->pack_name); | |
590 | strcpy(idx_name + strlen(idx_name) - strlen(".pack"), ".idx"); | |
591 | ret = check_packed_git_idx(idx_name, p); | |
592 | free(idx_name); | |
593 | return ret; | |
594 | } | |
595 | ||
11daf39b SP |
596 | static void scan_windows(struct packed_git *p, |
597 | struct packed_git **lru_p, | |
598 | struct pack_window **lru_w, | |
599 | struct pack_window **lru_l) | |
1f688557 | 600 | { |
11daf39b SP |
601 | struct pack_window *w, *w_l; |
602 | ||
603 | for (w_l = NULL, w = p->windows; w; w = w->next) { | |
604 | if (!w->inuse_cnt) { | |
605 | if (!*lru_w || w->last_used < (*lru_w)->last_used) { | |
606 | *lru_p = p; | |
607 | *lru_w = w; | |
608 | *lru_l = w_l; | |
54044bf8 | 609 | } |
54044bf8 | 610 | } |
11daf39b | 611 | w_l = w; |
f9253394 | 612 | } |
11daf39b SP |
613 | } |
614 | ||
d1efefa4 | 615 | static int unuse_one_window(struct packed_git *current, int keep_fd) |
11daf39b SP |
616 | { |
617 | struct packed_git *p, *lru_p = NULL; | |
618 | struct pack_window *lru_w = NULL, *lru_l = NULL; | |
619 | ||
620 | if (current) | |
621 | scan_windows(current, &lru_p, &lru_w, &lru_l); | |
622 | for (p = packed_git; p; p = p->next) | |
623 | scan_windows(p, &lru_p, &lru_w, &lru_l); | |
54044bf8 SP |
624 | if (lru_p) { |
625 | munmap(lru_w->base, lru_w->len); | |
626 | pack_mapped -= lru_w->len; | |
627 | if (lru_l) | |
628 | lru_l->next = lru_w->next; | |
629 | else { | |
630 | lru_p->windows = lru_w->next; | |
d1efefa4 | 631 | if (!lru_p->windows && lru_p->pack_fd != keep_fd) { |
54044bf8 SP |
632 | close(lru_p->pack_fd); |
633 | lru_p->pack_fd = -1; | |
634 | } | |
635 | } | |
636 | free(lru_w); | |
a53128b6 | 637 | pack_open_windows--; |
54044bf8 SP |
638 | return 1; |
639 | } | |
640 | return 0; | |
f9253394 JH |
641 | } |
642 | ||
d1efefa4 | 643 | void release_pack_memory(size_t need, int fd) |
97bfeb34 SP |
644 | { |
645 | size_t cur = pack_mapped; | |
d1efefa4 | 646 | while (need >= (cur - pack_mapped) && unuse_one_window(NULL, fd)) |
97bfeb34 SP |
647 | ; /* nothing */ |
648 | } | |
649 | ||
c9ced051 SP |
650 | void close_pack_windows(struct packed_git *p) |
651 | { | |
652 | while (p->windows) { | |
653 | struct pack_window *w = p->windows; | |
654 | ||
655 | if (w->inuse_cnt) | |
656 | die("pack '%s' still has open windows to it", | |
657 | p->pack_name); | |
658 | munmap(w->base, w->len); | |
659 | pack_mapped -= w->len; | |
660 | pack_open_windows--; | |
661 | p->windows = w->next; | |
662 | free(w); | |
663 | } | |
664 | } | |
665 | ||
03e79c88 | 666 | void unuse_pack(struct pack_window **w_cursor) |
f9253394 | 667 | { |
03e79c88 SP |
668 | struct pack_window *w = *w_cursor; |
669 | if (w) { | |
670 | w->inuse_cnt--; | |
671 | *w_cursor = NULL; | |
672 | } | |
1f688557 JH |
673 | } |
674 | ||
c74faea1 NP |
675 | /* |
676 | * This is used by git-repack in case a newly created pack happens to | |
677 | * contain the same set of objects as an existing one. In that case | |
678 | * the resulting file might be different even if its name would be the | |
679 | * same. It is best to close any reference to the old pack before it is | |
680 | * replaced on disk. Of course no index pointers nor windows for given pack | |
681 | * must subsist at this point. If ever objects from this pack are requested | |
682 | * again, the new version of the pack will be reinitialized through | |
683 | * reprepare_packed_git(). | |
684 | */ | |
685 | void free_pack_by_name(const char *pack_name) | |
686 | { | |
687 | struct packed_git *p, **pp = &packed_git; | |
688 | ||
689 | while (*pp) { | |
690 | p = *pp; | |
691 | if (strcmp(pack_name, p->pack_name) == 0) { | |
692 | close_pack_windows(p); | |
693 | if (p->pack_fd != -1) | |
694 | close(p->pack_fd); | |
695 | if (p->index_data) | |
696 | munmap((void *)p->index_data, p->index_size); | |
697 | free(p->bad_object_sha1); | |
698 | *pp = p->next; | |
699 | free(p); | |
700 | return; | |
701 | } | |
702 | pp = &p->next; | |
703 | } | |
704 | } | |
705 | ||
3cf8b462 SP |
706 | /* |
707 | * Do not call this directly as this leaks p->pack_fd on error return; | |
708 | * call open_packed_git() instead. | |
709 | */ | |
710 | static int open_packed_git_1(struct packed_git *p) | |
1f688557 | 711 | { |
9bc879c1 SP |
712 | struct stat st; |
713 | struct pack_header hdr; | |
714 | unsigned char sha1[20]; | |
715 | unsigned char *idx_sha1; | |
2c039da8 | 716 | long fd_flag; |
9bc879c1 | 717 | |
d079837e SP |
718 | if (!p->index_data && open_pack_index(p)) |
719 | return error("packfile %s index unavailable", p->pack_name); | |
720 | ||
9bc879c1 SP |
721 | p->pack_fd = open(p->pack_name, O_RDONLY); |
722 | if (p->pack_fd < 0 || fstat(p->pack_fd, &st)) | |
072db278 | 723 | return -1; |
9bc879c1 SP |
724 | |
725 | /* If we created the struct before we had the pack we lack size. */ | |
bf592c50 | 726 | if (!p->pack_size) { |
bf592c50 | 727 | if (!S_ISREG(st.st_mode)) |
072db278 | 728 | return error("packfile %s not a regular file", p->pack_name); |
bf592c50 | 729 | p->pack_size = st.st_size; |
9bc879c1 | 730 | } else if (p->pack_size != st.st_size) |
072db278 | 731 | return error("packfile %s size changed", p->pack_name); |
9bc879c1 | 732 | |
2c039da8 JH |
733 | /* We leave these file descriptors open with sliding mmap; |
734 | * there is no point keeping them open across exec(), though. | |
735 | */ | |
736 | fd_flag = fcntl(p->pack_fd, F_GETFD, 0); | |
737 | if (fd_flag < 0) | |
072db278 | 738 | return error("cannot determine file descriptor flags"); |
2c039da8 JH |
739 | fd_flag |= FD_CLOEXEC; |
740 | if (fcntl(p->pack_fd, F_SETFD, fd_flag) == -1) | |
072db278 | 741 | return error("cannot set FD_CLOEXEC"); |
2c039da8 | 742 | |
9bc879c1 | 743 | /* Verify we recognize this pack file format. */ |
e6e2bd62 | 744 | if (read_in_full(p->pack_fd, &hdr, sizeof(hdr)) != sizeof(hdr)) |
072db278 | 745 | return error("file %s is far too short to be a packfile", p->pack_name); |
9bc879c1 | 746 | if (hdr.hdr_signature != htonl(PACK_SIGNATURE)) |
072db278 | 747 | return error("file %s is not a GIT packfile", p->pack_name); |
9bc879c1 | 748 | if (!pack_version_ok(hdr.hdr_version)) |
6e1c2344 RJ |
749 | return error("packfile %s is version %"PRIu32" and not" |
750 | " supported (try upgrading GIT to a newer version)", | |
9bc879c1 SP |
751 | p->pack_name, ntohl(hdr.hdr_version)); |
752 | ||
753 | /* Verify the pack matches its index. */ | |
57059091 | 754 | if (p->num_objects != ntohl(hdr.hdr_entries)) |
6e1c2344 RJ |
755 | return error("packfile %s claims to have %"PRIu32" objects" |
756 | " while index indicates %"PRIu32" objects", | |
57059091 NP |
757 | p->pack_name, ntohl(hdr.hdr_entries), |
758 | p->num_objects); | |
9bc879c1 | 759 | if (lseek(p->pack_fd, p->pack_size - sizeof(sha1), SEEK_SET) == -1) |
072db278 | 760 | return error("end of packfile %s is unavailable", p->pack_name); |
e6e2bd62 | 761 | if (read_in_full(p->pack_fd, sha1, sizeof(sha1)) != sizeof(sha1)) |
072db278 | 762 | return error("packfile %s signature is unavailable", p->pack_name); |
42873078 | 763 | idx_sha1 = ((unsigned char *)p->index_data) + p->index_size - 40; |
9bc879c1 | 764 | if (hashcmp(sha1, idx_sha1)) |
072db278 SP |
765 | return error("packfile %s does not match index", p->pack_name); |
766 | return 0; | |
9bc879c1 SP |
767 | } |
768 | ||
3cf8b462 SP |
769 | static int open_packed_git(struct packed_git *p) |
770 | { | |
771 | if (!open_packed_git_1(p)) | |
772 | return 0; | |
773 | if (p->pack_fd != -1) { | |
774 | close(p->pack_fd); | |
775 | p->pack_fd = -1; | |
776 | } | |
777 | return -1; | |
778 | } | |
779 | ||
c4001d92 | 780 | static int in_window(struct pack_window *win, off_t offset) |
60bb8b14 SP |
781 | { |
782 | /* We must promise at least 20 bytes (one hash) after the | |
783 | * offset is available from this window, otherwise the offset | |
784 | * is not actually in this window and a different window (which | |
785 | * has that one hash excess) must be used. This is to support | |
786 | * the object header and delta base parsing routines below. | |
787 | */ | |
788 | off_t win_off = win->offset; | |
789 | return win_off <= offset | |
790 | && (offset + 20) <= (win_off + win->len); | |
791 | } | |
792 | ||
03e79c88 SP |
793 | unsigned char* use_pack(struct packed_git *p, |
794 | struct pack_window **w_cursor, | |
c4001d92 | 795 | off_t offset, |
03e79c88 | 796 | unsigned int *left) |
9bc879c1 | 797 | { |
60bb8b14 | 798 | struct pack_window *win = *w_cursor; |
03e79c88 | 799 | |
072db278 SP |
800 | if (p->pack_fd == -1 && open_packed_git(p)) |
801 | die("packfile %s cannot be accessed", p->pack_name); | |
60bb8b14 SP |
802 | |
803 | /* Since packfiles end in a hash of their content and its | |
804 | * pointless to ask for an offset into the middle of that | |
805 | * hash, and the in_window function above wouldn't match | |
806 | * don't allow an offset too close to the end of the file. | |
807 | */ | |
808 | if (offset > (p->pack_size - 20)) | |
809 | die("offset beyond end of packfile (truncated pack?)"); | |
810 | ||
811 | if (!win || !in_window(win, offset)) { | |
812 | if (win) | |
813 | win->inuse_cnt--; | |
814 | for (win = p->windows; win; win = win->next) { | |
815 | if (in_window(win, offset)) | |
816 | break; | |
817 | } | |
818 | if (!win) { | |
78a28df9 | 819 | size_t window_align = packed_git_window_size / 2; |
dc49cd76 | 820 | off_t len; |
60bb8b14 | 821 | win = xcalloc(1, sizeof(*win)); |
78a28df9 | 822 | win->offset = (offset / window_align) * window_align; |
dc49cd76 SP |
823 | len = p->pack_size - win->offset; |
824 | if (len > packed_git_window_size) | |
825 | len = packed_git_window_size; | |
826 | win->len = (size_t)len; | |
60bb8b14 | 827 | pack_mapped += win->len; |
11daf39b | 828 | while (packed_git_limit < pack_mapped |
d1efefa4 | 829 | && unuse_one_window(p, p->pack_fd)) |
60bb8b14 | 830 | ; /* nothing */ |
c4712e45 | 831 | win->base = xmmap(NULL, win->len, |
60bb8b14 SP |
832 | PROT_READ, MAP_PRIVATE, |
833 | p->pack_fd, win->offset); | |
834 | if (win->base == MAP_FAILED) | |
73b4e4be SP |
835 | die("packfile %s cannot be mapped: %s", |
836 | p->pack_name, | |
837 | strerror(errno)); | |
a53128b6 SP |
838 | pack_mmap_calls++; |
839 | pack_open_windows++; | |
840 | if (pack_mapped > peak_pack_mapped) | |
841 | peak_pack_mapped = pack_mapped; | |
842 | if (pack_open_windows > peak_pack_open_windows) | |
843 | peak_pack_open_windows = pack_open_windows; | |
60bb8b14 SP |
844 | win->next = p->windows; |
845 | p->windows = win; | |
846 | } | |
1f688557 | 847 | } |
03e79c88 SP |
848 | if (win != *w_cursor) { |
849 | win->last_used = pack_used_ctr++; | |
850 | win->inuse_cnt++; | |
851 | *w_cursor = win; | |
852 | } | |
60bb8b14 | 853 | offset -= win->offset; |
03e79c88 | 854 | if (left) |
dc49cd76 | 855 | *left = win->len - xsize_t(offset); |
03e79c88 | 856 | return win->base + offset; |
1f688557 JH |
857 | } |
858 | ||
27d69a46 NP |
859 | static struct packed_git *alloc_packed_git(int extra) |
860 | { | |
861 | struct packed_git *p = xmalloc(sizeof(*p) + extra); | |
862 | memset(p, 0, sizeof(*p)); | |
863 | p->pack_fd = -1; | |
864 | return p; | |
865 | } | |
866 | ||
42873078 | 867 | struct packed_git *add_packed_git(const char *path, int path_len, int local) |
1f688557 JH |
868 | { |
869 | struct stat st; | |
27d69a46 | 870 | struct packed_git *p = alloc_packed_git(path_len + 2); |
1f688557 | 871 | |
42873078 NP |
872 | /* |
873 | * Make sure a corresponding .pack file exists and that | |
874 | * the index looks sane. | |
875 | */ | |
876 | path_len -= strlen(".idx"); | |
27d69a46 NP |
877 | if (path_len < 1) { |
878 | free(p); | |
1f688557 | 879 | return NULL; |
27d69a46 | 880 | } |
42873078 | 881 | memcpy(p->pack_name, path, path_len); |
8d25931d BC |
882 | |
883 | strcpy(p->pack_name + path_len, ".keep"); | |
884 | if (!access(p->pack_name, F_OK)) | |
885 | p->pack_keep = 1; | |
886 | ||
42873078 | 887 | strcpy(p->pack_name + path_len, ".pack"); |
d079837e | 888 | if (stat(p->pack_name, &st) || !S_ISREG(st.st_mode)) { |
42873078 | 889 | free(p); |
1f688557 JH |
890 | return NULL; |
891 | } | |
42873078 | 892 | |
1f688557 JH |
893 | /* ok, it looks sane as far as we can check without |
894 | * actually mapping the pack file. | |
895 | */ | |
1f688557 | 896 | p->pack_size = st.st_size; |
9d835df2 | 897 | p->pack_local = local; |
b867092f | 898 | p->mtime = st.st_mtime; |
42873078 NP |
899 | if (path_len < 40 || get_sha1_hex(path + path_len - 40, p->sha1)) |
900 | hashclr(p->sha1); | |
1f688557 JH |
901 | return p; |
902 | } | |
903 | ||
bf592c50 | 904 | struct packed_git *parse_pack_index(unsigned char *sha1) |
c508df5e | 905 | { |
79c6dca4 | 906 | const char *idx_path = sha1_pack_index_name(sha1); |
42873078 | 907 | const char *path = sha1_pack_name(sha1); |
27d69a46 | 908 | struct packed_git *p = alloc_packed_git(strlen(path) + 1); |
bf592c50 | 909 | |
27d69a46 NP |
910 | strcpy(p->pack_name, path); |
911 | hashcpy(p->sha1, sha1); | |
42873078 NP |
912 | if (check_packed_git_idx(idx_path, p)) { |
913 | free(p); | |
bf592c50 | 914 | return NULL; |
42873078 | 915 | } |
bf592c50 | 916 | |
bf592c50 DB |
917 | return p; |
918 | } | |
919 | ||
920 | void install_packed_git(struct packed_git *pack) | |
921 | { | |
922 | pack->next = packed_git; | |
923 | packed_git = pack; | |
924 | } | |
925 | ||
9d835df2 | 926 | static void prepare_packed_git_one(char *objdir, int local) |
1f688557 | 927 | { |
9cb18f56 JM |
928 | /* Ensure that this buffer is large enough so that we can |
929 | append "/pack/" without clobbering the stack even if | |
930 | strlen(objdir) were PATH_MAX. */ | |
931 | char path[PATH_MAX + 1 + 4 + 1 + 1]; | |
1f688557 JH |
932 | int len; |
933 | DIR *dir; | |
934 | struct dirent *de; | |
935 | ||
936 | sprintf(path, "%s/pack", objdir); | |
937 | len = strlen(path); | |
938 | dir = opendir(path); | |
b5b16990 | 939 | if (!dir) { |
26125f6b | 940 | if (errno != ENOENT) |
bd2afde8 | 941 | error("unable to open object pack directory: %s: %s", |
26125f6b | 942 | path, strerror(errno)); |
1f688557 | 943 | return; |
b5b16990 | 944 | } |
1f688557 JH |
945 | path[len++] = '/'; |
946 | while ((de = readdir(dir)) != NULL) { | |
947 | int namelen = strlen(de->d_name); | |
948 | struct packed_git *p; | |
949 | ||
5bb1cda5 | 950 | if (!has_extension(de->d_name, ".idx")) |
1f688557 JH |
951 | continue; |
952 | ||
9cb18f56 JM |
953 | if (len + namelen + 1 > sizeof(path)) |
954 | continue; | |
955 | ||
54a15a8d | 956 | /* Don't reopen a pack we already have. */ |
1f688557 | 957 | strcpy(path + len, de->d_name); |
86f7780c JK |
958 | for (p = packed_git; p; p = p->next) { |
959 | if (!memcmp(path, p->pack_name, len + namelen - 4)) | |
960 | break; | |
961 | } | |
962 | if (p) | |
963 | continue; | |
54a15a8d SP |
964 | /* See if it really is a valid .idx file with corresponding |
965 | * .pack file that we can map. | |
966 | */ | |
9d835df2 | 967 | p = add_packed_git(path, len + namelen, local); |
1f688557 JH |
968 | if (!p) |
969 | continue; | |
625e9421 | 970 | install_packed_git(p); |
1f688557 | 971 | } |
5b35bcd5 | 972 | closedir(dir); |
1f688557 JH |
973 | } |
974 | ||
b867092f JH |
975 | static int sort_pack(const void *a_, const void *b_) |
976 | { | |
977 | struct packed_git *a = *((struct packed_git **)a_); | |
978 | struct packed_git *b = *((struct packed_git **)b_); | |
979 | int st; | |
980 | ||
981 | /* | |
982 | * Local packs tend to contain objects specific to our | |
983 | * variant of the project than remote ones. In addition, | |
984 | * remote ones could be on a network mounted filesystem. | |
985 | * Favor local ones for these reasons. | |
986 | */ | |
987 | st = a->pack_local - b->pack_local; | |
988 | if (st) | |
989 | return -st; | |
990 | ||
991 | /* | |
992 | * Younger packs tend to contain more recent objects, | |
993 | * and more recent objects tend to get accessed more | |
994 | * often. | |
995 | */ | |
996 | if (a->mtime < b->mtime) | |
997 | return 1; | |
998 | else if (a->mtime == b->mtime) | |
999 | return 0; | |
1000 | return -1; | |
1001 | } | |
1002 | ||
1003 | static void rearrange_packed_git(void) | |
1004 | { | |
1005 | struct packed_git **ary, *p; | |
1006 | int i, n; | |
1007 | ||
1008 | for (n = 0, p = packed_git; p; p = p->next) | |
1009 | n++; | |
1010 | if (n < 2) | |
1011 | return; | |
1012 | ||
1013 | /* prepare an array of packed_git for easier sorting */ | |
1014 | ary = xcalloc(n, sizeof(struct packed_git *)); | |
1015 | for (n = 0, p = packed_git; p; p = p->next) | |
1016 | ary[n++] = p; | |
1017 | ||
1018 | qsort(ary, n, sizeof(struct packed_git *), sort_pack); | |
1019 | ||
1020 | /* link them back again */ | |
1021 | for (i = 0; i < n - 1; i++) | |
1022 | ary[i]->next = ary[i + 1]; | |
1023 | ary[n - 1]->next = NULL; | |
1024 | packed_git = ary[0]; | |
1025 | ||
1026 | free(ary); | |
1027 | } | |
1028 | ||
637cdd9d | 1029 | static int prepare_packed_git_run_once = 0; |
9a217f2a | 1030 | void prepare_packed_git(void) |
1f688557 | 1031 | { |
d5a63b99 | 1032 | struct alternate_object_database *alt; |
1f688557 | 1033 | |
637cdd9d | 1034 | if (prepare_packed_git_run_once) |
1f688557 | 1035 | return; |
9d835df2 | 1036 | prepare_packed_git_one(get_object_directory(), 1); |
9a217f2a | 1037 | prepare_alt_odb(); |
d5a63b99 | 1038 | for (alt = alt_odb_list; alt; alt = alt->next) { |
1494e038 | 1039 | alt->name[-1] = 0; |
9d835df2 | 1040 | prepare_packed_git_one(alt->base, 0); |
1494e038 | 1041 | alt->name[-1] = '/'; |
1f688557 | 1042 | } |
b867092f | 1043 | rearrange_packed_git(); |
637cdd9d JK |
1044 | prepare_packed_git_run_once = 1; |
1045 | } | |
1046 | ||
fc04c412 | 1047 | void reprepare_packed_git(void) |
637cdd9d | 1048 | { |
4b480c67 | 1049 | discard_revindex(); |
637cdd9d JK |
1050 | prepare_packed_git_run_once = 0; |
1051 | prepare_packed_git(); | |
1f688557 JH |
1052 | } |
1053 | ||
8eca0b47 NP |
1054 | static void mark_bad_packed_object(struct packed_git *p, |
1055 | const unsigned char *sha1) | |
1056 | { | |
1057 | unsigned i; | |
1058 | for (i = 0; i < p->num_bad_objects; i++) | |
1059 | if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i)) | |
1060 | return; | |
1061 | p->bad_object_sha1 = xrealloc(p->bad_object_sha1, 20 * (p->num_bad_objects + 1)); | |
1062 | hashcpy(p->bad_object_sha1 + 20 * p->num_bad_objects, sha1); | |
1063 | p->num_bad_objects++; | |
1064 | } | |
1065 | ||
ac939109 NP |
1066 | static int has_packed_and_bad(const unsigned char *sha1) |
1067 | { | |
1068 | struct packed_git *p; | |
1069 | unsigned i; | |
1070 | ||
1071 | for (p = packed_git; p; p = p->next) | |
1072 | for (i = 0; i < p->num_bad_objects; i++) | |
1073 | if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i)) | |
1074 | return 1; | |
1075 | return 0; | |
1076 | } | |
1077 | ||
5d6ccf5c | 1078 | int check_sha1_signature(const unsigned char *sha1, void *map, unsigned long size, const char *type) |
0fcfd160 LT |
1079 | { |
1080 | unsigned char real_sha1[20]; | |
7cfb5f36 | 1081 | hash_sha1_file(map, size, type, real_sha1); |
a89fccd2 | 1082 | return hashcmp(sha1, real_sha1) ? -1 : 0; |
0fcfd160 LT |
1083 | } |
1084 | ||
44d1c19e LT |
1085 | static int git_open_noatime(const char *name) |
1086 | { | |
1087 | static int sha1_file_open_flag = O_NOATIME; | |
1088 | int fd = open(name, O_RDONLY | sha1_file_open_flag); | |
1089 | ||
1090 | /* Might the failure be due to O_NOATIME? */ | |
1091 | if (fd < 0 && errno != ENOENT && sha1_file_open_flag) { | |
1092 | fd = open(name, O_RDONLY); | |
1093 | if (fd >= 0) | |
1094 | sha1_file_open_flag = 0; | |
1095 | } | |
1096 | return fd; | |
1097 | } | |
1098 | ||
1099 | static int open_sha1_file(const unsigned char *sha1) | |
1100 | { | |
1101 | int fd; | |
1102 | char *name = sha1_file_name(sha1); | |
1103 | struct alternate_object_database *alt; | |
1104 | ||
1105 | fd = git_open_noatime(name); | |
1106 | if (fd >= 0) | |
1107 | return fd; | |
1108 | ||
1109 | prepare_alt_odb(); | |
1110 | errno = ENOENT; | |
1111 | for (alt = alt_odb_list; alt; alt = alt->next) { | |
1112 | name = alt->name; | |
1113 | fill_sha1_path(name, sha1); | |
1114 | fd = git_open_noatime(alt->base); | |
1115 | if (fd >= 0) | |
1116 | return fd; | |
1117 | } | |
1118 | return -1; | |
1119 | } | |
1120 | ||
4175e9e3 | 1121 | static void *map_sha1_file(const unsigned char *sha1, unsigned long *size) |
0fcfd160 | 1122 | { |
0fcfd160 | 1123 | void *map; |
144bde78 | 1124 | int fd; |
ace1534d | 1125 | |
44d1c19e LT |
1126 | fd = open_sha1_file(sha1); |
1127 | map = NULL; | |
1128 | if (fd >= 0) { | |
1129 | struct stat st; | |
0fcfd160 | 1130 | |
44d1c19e LT |
1131 | if (!fstat(fd, &st)) { |
1132 | *size = xsize_t(st.st_size); | |
1133 | map = xmmap(NULL, *size, PROT_READ, MAP_PRIVATE, fd, 0); | |
144bde78 | 1134 | } |
44d1c19e | 1135 | close(fd); |
0fcfd160 | 1136 | } |
0fcfd160 LT |
1137 | return map; |
1138 | } | |
1139 | ||
726f852b | 1140 | static int legacy_loose_object(unsigned char *map) |
bb6b8e4f JH |
1141 | { |
1142 | unsigned int word; | |
1143 | ||
1144 | /* | |
1145 | * Is it a zlib-compressed buffer? If so, the first byte | |
1146 | * must be 0x78 (15-bit window size, deflated), and the | |
1147 | * first 16-bit word is evenly divisible by 31 | |
1148 | */ | |
1149 | word = (map[0] << 8) + map[1]; | |
1150 | if (map[0] == 0x78 && !(word % 31)) | |
1151 | return 1; | |
1152 | else | |
1153 | return 0; | |
1154 | } | |
1155 | ||
09ded04b NP |
1156 | unsigned long unpack_object_header_buffer(const unsigned char *buf, |
1157 | unsigned long len, enum object_type *type, unsigned long *sizep) | |
c4483576 | 1158 | { |
ad1ed5ee | 1159 | unsigned shift; |
93821bd9 | 1160 | unsigned char c; |
93821bd9 | 1161 | unsigned long size; |
ad1ed5ee JH |
1162 | unsigned long used = 0; |
1163 | ||
1164 | c = buf[used++]; | |
1165 | *type = (c >> 4) & 7; | |
1166 | size = c & 15; | |
1167 | shift = 4; | |
1168 | while (c & 0x80) { | |
09ded04b NP |
1169 | if (len <= used || sizeof(long) * 8 <= shift) { |
1170 | error("bad object header"); | |
ad1ed5ee | 1171 | return 0; |
09ded04b | 1172 | } |
ad1ed5ee JH |
1173 | c = buf[used++]; |
1174 | size += (c & 0x7f) << shift; | |
1175 | shift += 7; | |
1176 | } | |
1177 | *sizep = size; | |
1178 | return used; | |
1179 | } | |
1180 | ||
1181 | static int unpack_sha1_header(z_stream *stream, unsigned char *map, unsigned long mapsize, void *buffer, unsigned long bufsiz) | |
1182 | { | |
1183 | unsigned long size, used; | |
1184 | static const char valid_loose_object_type[8] = { | |
1185 | 0, /* OBJ_EXT */ | |
1186 | 1, 1, 1, 1, /* "commit", "tree", "blob", "tag" */ | |
1187 | 0, /* "delta" and others are invalid in a loose object */ | |
93821bd9 | 1188 | }; |
ad1ed5ee | 1189 | enum object_type type; |
93821bd9 | 1190 | |
c4483576 LT |
1191 | /* Get the data stream */ |
1192 | memset(stream, 0, sizeof(*stream)); | |
1193 | stream->next_in = map; | |
1194 | stream->avail_in = mapsize; | |
1195 | stream->next_out = buffer; | |
93821bd9 LT |
1196 | stream->avail_out = bufsiz; |
1197 | ||
bb6b8e4f | 1198 | if (legacy_loose_object(map)) { |
39c68542 LT |
1199 | git_inflate_init(stream); |
1200 | return git_inflate(stream, 0); | |
93821bd9 LT |
1201 | } |
1202 | ||
726f852b NP |
1203 | |
1204 | /* | |
1205 | * There used to be a second loose object header format which | |
1206 | * was meant to mimic the in-pack format, allowing for direct | |
1207 | * copy of the object data. This format turned up not to be | |
1208 | * really worth it and we don't write it any longer. But we | |
1209 | * can still read it. | |
1210 | */ | |
09ded04b | 1211 | used = unpack_object_header_buffer(map, mapsize, &type, &size); |
ad1ed5ee | 1212 | if (!used || !valid_loose_object_type[type]) |
93821bd9 | 1213 | return -1; |
ad1ed5ee JH |
1214 | map += used; |
1215 | mapsize -= used; | |
93821bd9 LT |
1216 | |
1217 | /* Set up the stream for the rest.. */ | |
1218 | stream->next_in = map; | |
1219 | stream->avail_in = mapsize; | |
39c68542 | 1220 | git_inflate_init(stream); |
93821bd9 LT |
1221 | |
1222 | /* And generate the fake traditional header */ | |
ad1ed5ee | 1223 | stream->total_out = 1 + snprintf(buffer, bufsiz, "%s %lu", |
df843662 | 1224 | typename(type), size); |
93821bd9 | 1225 | return 0; |
c4483576 LT |
1226 | } |
1227 | ||
7efbff75 | 1228 | static void *unpack_sha1_rest(z_stream *stream, void *buffer, unsigned long size, const unsigned char *sha1) |
5180cacc LT |
1229 | { |
1230 | int bytes = strlen(buffer) + 1; | |
d565b341 | 1231 | unsigned char *buf = xmalloc(1+size); |
93821bd9 | 1232 | unsigned long n; |
7efbff75 | 1233 | int status = Z_OK; |
5180cacc | 1234 | |
93821bd9 LT |
1235 | n = stream->total_out - bytes; |
1236 | if (n > size) | |
1237 | n = size; | |
1238 | memcpy(buf, (char *) buffer + bytes, n); | |
1239 | bytes = n; | |
456cdf6e LT |
1240 | if (bytes <= size) { |
1241 | /* | |
1242 | * The above condition must be (bytes <= size), not | |
1243 | * (bytes < size). In other words, even though we | |
1244 | * expect no more output and set avail_out to zer0, | |
1245 | * the input zlib stream may have bytes that express | |
1246 | * "this concludes the stream", and we *do* want to | |
1247 | * eat that input. | |
1248 | * | |
1249 | * Otherwise we would not be able to test that we | |
1250 | * consumed all the input to reach the expected size; | |
1251 | * we also want to check that zlib tells us that all | |
1252 | * went well with status == Z_STREAM_END at the end. | |
1253 | */ | |
5180cacc LT |
1254 | stream->next_out = buf + bytes; |
1255 | stream->avail_out = size - bytes; | |
7efbff75 | 1256 | while (status == Z_OK) |
39c68542 | 1257 | status = git_inflate(stream, Z_FINISH); |
5180cacc LT |
1258 | } |
1259 | buf[size] = 0; | |
456cdf6e | 1260 | if (status == Z_STREAM_END && !stream->avail_in) { |
39c68542 | 1261 | git_inflate_end(stream); |
7efbff75 JH |
1262 | return buf; |
1263 | } | |
1264 | ||
1265 | if (status < 0) | |
1266 | error("corrupt loose object '%s'", sha1_to_hex(sha1)); | |
1267 | else if (stream->avail_in) | |
1268 | error("garbage at end of loose object '%s'", | |
1269 | sha1_to_hex(sha1)); | |
1270 | free(buf); | |
1271 | return NULL; | |
5180cacc LT |
1272 | } |
1273 | ||
1274 | /* | |
1275 | * We used to just use "sscanf()", but that's actually way | |
1276 | * too permissive for what we want to check. So do an anal | |
1277 | * object header parse by hand. | |
1278 | */ | |
21666f1a | 1279 | static int parse_sha1_header(const char *hdr, unsigned long *sizep) |
5180cacc | 1280 | { |
21666f1a | 1281 | char type[10]; |
5180cacc LT |
1282 | int i; |
1283 | unsigned long size; | |
1284 | ||
1285 | /* | |
a6080a0a | 1286 | * The type can be at most ten bytes (including the |
5180cacc | 1287 | * terminating '\0' that we add), and is followed by |
21666f1a | 1288 | * a space. |
5180cacc | 1289 | */ |
21666f1a | 1290 | i = 0; |
5180cacc LT |
1291 | for (;;) { |
1292 | char c = *hdr++; | |
1293 | if (c == ' ') | |
1294 | break; | |
21666f1a NP |
1295 | type[i++] = c; |
1296 | if (i >= sizeof(type)) | |
5180cacc | 1297 | return -1; |
5180cacc | 1298 | } |
21666f1a | 1299 | type[i] = 0; |
5180cacc LT |
1300 | |
1301 | /* | |
1302 | * The length must follow immediately, and be in canonical | |
1303 | * decimal format (ie "010" is not valid). | |
1304 | */ | |
1305 | size = *hdr++ - '0'; | |
1306 | if (size > 9) | |
1307 | return -1; | |
1308 | if (size) { | |
1309 | for (;;) { | |
1310 | unsigned long c = *hdr - '0'; | |
1311 | if (c > 9) | |
1312 | break; | |
1313 | hdr++; | |
1314 | size = size * 10 + c; | |
1315 | } | |
1316 | } | |
1317 | *sizep = size; | |
1318 | ||
1319 | /* | |
1320 | * The length must be followed by a zero byte | |
1321 | */ | |
21666f1a | 1322 | return *hdr ? -1 : type_from_string(type); |
5180cacc LT |
1323 | } |
1324 | ||
7efbff75 | 1325 | static void *unpack_sha1_file(void *map, unsigned long mapsize, enum object_type *type, unsigned long *size, const unsigned char *sha1) |
0fcfd160 | 1326 | { |
5180cacc | 1327 | int ret; |
0fcfd160 | 1328 | z_stream stream; |
5180cacc | 1329 | char hdr[8192]; |
0fcfd160 | 1330 | |
5180cacc | 1331 | ret = unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)); |
21666f1a | 1332 | if (ret < Z_OK || (*type = parse_sha1_header(hdr, size)) < 0) |
0fcfd160 LT |
1333 | return NULL; |
1334 | ||
7efbff75 | 1335 | return unpack_sha1_rest(&stream, hdr, *size, sha1); |
0fcfd160 LT |
1336 | } |
1337 | ||
54dab52a NP |
1338 | unsigned long get_size_from_delta(struct packed_git *p, |
1339 | struct pack_window **w_curs, | |
1340 | off_t curpos) | |
1341 | { | |
1342 | const unsigned char *data; | |
1343 | unsigned char delta_head[20], *in; | |
1344 | z_stream stream; | |
1345 | int st; | |
1346 | ||
1347 | memset(&stream, 0, sizeof(stream)); | |
1348 | stream.next_out = delta_head; | |
1349 | stream.avail_out = sizeof(delta_head); | |
1350 | ||
39c68542 | 1351 | git_inflate_init(&stream); |
54dab52a NP |
1352 | do { |
1353 | in = use_pack(p, w_curs, curpos, &stream.avail_in); | |
1354 | stream.next_in = in; | |
39c68542 | 1355 | st = git_inflate(&stream, Z_FINISH); |
54dab52a NP |
1356 | curpos += stream.next_in - in; |
1357 | } while ((st == Z_OK || st == Z_BUF_ERROR) && | |
1358 | stream.total_out < sizeof(delta_head)); | |
39c68542 | 1359 | git_inflate_end(&stream); |
3d77d877 NP |
1360 | if ((st != Z_STREAM_END) && stream.total_out != sizeof(delta_head)) { |
1361 | error("delta data unpack-initial failed"); | |
1362 | return 0; | |
1363 | } | |
54dab52a NP |
1364 | |
1365 | /* Examine the initial part of the delta to figure out | |
1366 | * the result size. | |
1367 | */ | |
1368 | data = delta_head; | |
1369 | ||
1370 | /* ignore base size */ | |
1371 | get_delta_hdr_size(&data, delta_head+sizeof(delta_head)); | |
1372 | ||
1373 | /* Read the result size */ | |
1374 | return get_delta_hdr_size(&data, delta_head+sizeof(delta_head)); | |
1375 | } | |
1376 | ||
c4001d92 | 1377 | static off_t get_delta_base(struct packed_git *p, |
03e79c88 | 1378 | struct pack_window **w_curs, |
c4001d92 | 1379 | off_t *curpos, |
21666f1a | 1380 | enum object_type type, |
c4001d92 | 1381 | off_t delta_obj_offset) |
eb32d236 | 1382 | { |
2b87c45b | 1383 | unsigned char *base_info = use_pack(p, w_curs, *curpos, NULL); |
c4001d92 | 1384 | off_t base_offset; |
eb32d236 | 1385 | |
8d8a4ea5 SP |
1386 | /* use_pack() assured us we have [base_info, base_info + 20) |
1387 | * as a range that we can look at without walking off the | |
1388 | * end of the mapped window. Its actually the hash size | |
1389 | * that is assured. An OFS_DELTA longer than the hash size | |
1390 | * is stupid, as then a REF_DELTA would be smaller to store. | |
1391 | */ | |
21666f1a | 1392 | if (type == OBJ_OFS_DELTA) { |
eb32d236 NP |
1393 | unsigned used = 0; |
1394 | unsigned char c = base_info[used++]; | |
1395 | base_offset = c & 127; | |
1396 | while (c & 128) { | |
1397 | base_offset += 1; | |
8723f216 | 1398 | if (!base_offset || MSB(base_offset, 7)) |
8eca0b47 | 1399 | return 0; /* overflow */ |
eb32d236 NP |
1400 | c = base_info[used++]; |
1401 | base_offset = (base_offset << 7) + (c & 127); | |
1402 | } | |
1403 | base_offset = delta_obj_offset - base_offset; | |
d8f32556 | 1404 | if (base_offset <= 0 || base_offset >= delta_obj_offset) |
8eca0b47 | 1405 | return 0; /* out of bound */ |
2b87c45b | 1406 | *curpos += used; |
21666f1a | 1407 | } else if (type == OBJ_REF_DELTA) { |
eb32d236 NP |
1408 | /* The base entry _must_ be in the same pack */ |
1409 | base_offset = find_pack_entry_one(base_info, p); | |
2b87c45b | 1410 | *curpos += 20; |
eb32d236 NP |
1411 | } else |
1412 | die("I am totally screwed"); | |
2b87c45b | 1413 | return base_offset; |
eb32d236 NP |
1414 | } |
1415 | ||
f3bf9224 | 1416 | /* forward declaration for a mutually recursive function */ |
c4001d92 | 1417 | static int packed_object_info(struct packed_git *p, off_t offset, |
21666f1a | 1418 | unsigned long *sizep); |
f3bf9224 | 1419 | |
43057304 | 1420 | static int packed_delta_info(struct packed_git *p, |
03e79c88 | 1421 | struct pack_window **w_curs, |
c4001d92 | 1422 | off_t curpos, |
21666f1a | 1423 | enum object_type type, |
c4001d92 | 1424 | off_t obj_offset, |
43057304 | 1425 | unsigned long *sizep) |
5db47c2b | 1426 | { |
c4001d92 | 1427 | off_t base_offset; |
f3bf9224 | 1428 | |
21666f1a | 1429 | base_offset = get_delta_base(p, w_curs, &curpos, type, obj_offset); |
3d77d877 NP |
1430 | if (!base_offset) |
1431 | return OBJ_BAD; | |
21666f1a | 1432 | type = packed_object_info(p, base_offset, NULL); |
3d77d877 | 1433 | if (type <= OBJ_NONE) { |
08698b1e NP |
1434 | struct revindex_entry *revidx; |
1435 | const unsigned char *base_sha1; | |
1436 | revidx = find_pack_revindex(p, base_offset); | |
1437 | if (!revidx) | |
1438 | return OBJ_BAD; | |
1439 | base_sha1 = nth_packed_object_sha1(p, revidx->nr); | |
3d77d877 NP |
1440 | mark_bad_packed_object(p, base_sha1); |
1441 | type = sha1_object_info(base_sha1, NULL); | |
1442 | if (type <= OBJ_NONE) | |
1443 | return OBJ_BAD; | |
1444 | } | |
f3bf9224 | 1445 | |
c62266f3 JH |
1446 | /* We choose to only get the type of the base object and |
1447 | * ignore potentially corrupt pack file that expects the delta | |
1448 | * based on a base with a wrong size. This saves tons of | |
1449 | * inflate() calls. | |
1450 | */ | |
3d77d877 | 1451 | if (sizep) { |
54dab52a | 1452 | *sizep = get_size_from_delta(p, w_curs, curpos); |
3d77d877 NP |
1453 | if (*sizep == 0) |
1454 | type = OBJ_BAD; | |
1455 | } | |
21666f1a NP |
1456 | |
1457 | return type; | |
5db47c2b JH |
1458 | } |
1459 | ||
2b87c45b NP |
1460 | static int unpack_object_header(struct packed_git *p, |
1461 | struct pack_window **w_curs, | |
c4001d92 | 1462 | off_t *curpos, |
2b87c45b | 1463 | unsigned long *sizep) |
a733cb60 | 1464 | { |
03e79c88 SP |
1465 | unsigned char *base; |
1466 | unsigned int left; | |
ad1ed5ee | 1467 | unsigned long used; |
2b87c45b | 1468 | enum object_type type; |
a733cb60 | 1469 | |
8d8a4ea5 SP |
1470 | /* use_pack() assures us we have [base, base + 20) available |
1471 | * as a range that we can look at at. (Its actually the hash | |
3dff5379 | 1472 | * size that is assured.) With our object header encoding |
8d8a4ea5 SP |
1473 | * the maximum deflated object size is 2^137, which is just |
1474 | * insane, so we know won't exceed what we have been given. | |
1475 | */ | |
2b87c45b | 1476 | base = use_pack(p, w_curs, *curpos, &left); |
09ded04b NP |
1477 | used = unpack_object_header_buffer(base, left, &type, sizep); |
1478 | if (!used) { | |
1479 | type = OBJ_BAD; | |
1480 | } else | |
1481 | *curpos += used; | |
ad1ed5ee | 1482 | |
2b87c45b | 1483 | return type; |
a733cb60 LT |
1484 | } |
1485 | ||
21666f1a | 1486 | const char *packed_object_info_detail(struct packed_git *p, |
c4001d92 | 1487 | off_t obj_offset, |
21666f1a NP |
1488 | unsigned long *size, |
1489 | unsigned long *store_size, | |
1490 | unsigned int *delta_chain_length, | |
1491 | unsigned char *base_sha1) | |
ad8c80a5 | 1492 | { |
03e79c88 | 1493 | struct pack_window *w_curs = NULL; |
c4001d92 SP |
1494 | off_t curpos; |
1495 | unsigned long dummy; | |
43057304 | 1496 | unsigned char *next_sha1; |
21666f1a | 1497 | enum object_type type; |
70f5d5d3 | 1498 | struct revindex_entry *revidx; |
ad8c80a5 | 1499 | |
43057304 | 1500 | *delta_chain_length = 0; |
2b87c45b | 1501 | curpos = obj_offset; |
21666f1a | 1502 | type = unpack_object_header(p, &w_curs, &curpos, size); |
43057304 | 1503 | |
70f5d5d3 NP |
1504 | revidx = find_pack_revindex(p, obj_offset); |
1505 | *store_size = revidx[1].offset - obj_offset; | |
1506 | ||
43057304 | 1507 | for (;;) { |
21666f1a | 1508 | switch (type) { |
43057304 | 1509 | default: |
08a19d87 | 1510 | die("pack %s contains unknown object type %d", |
21666f1a | 1511 | p->pack_name, type); |
43057304 NP |
1512 | case OBJ_COMMIT: |
1513 | case OBJ_TREE: | |
1514 | case OBJ_BLOB: | |
1515 | case OBJ_TAG: | |
03e79c88 | 1516 | unuse_pack(&w_curs); |
21666f1a | 1517 | return typename(type); |
eb32d236 | 1518 | case OBJ_OFS_DELTA: |
21666f1a | 1519 | obj_offset = get_delta_base(p, &w_curs, &curpos, type, obj_offset); |
8eca0b47 NP |
1520 | if (!obj_offset) |
1521 | die("pack %s contains bad delta base reference of type %s", | |
1522 | p->pack_name, typename(type)); | |
eb32d236 | 1523 | if (*delta_chain_length == 0) { |
70f5d5d3 NP |
1524 | revidx = find_pack_revindex(p, obj_offset); |
1525 | hashcpy(base_sha1, nth_packed_object_sha1(p, revidx->nr)); | |
eb32d236 NP |
1526 | } |
1527 | break; | |
1528 | case OBJ_REF_DELTA: | |
2b87c45b | 1529 | next_sha1 = use_pack(p, &w_curs, curpos, NULL); |
43057304 NP |
1530 | if (*delta_chain_length == 0) |
1531 | hashcpy(base_sha1, next_sha1); | |
2b87c45b | 1532 | obj_offset = find_pack_entry_one(next_sha1, p); |
43057304 NP |
1533 | break; |
1534 | } | |
43057304 | 1535 | (*delta_chain_length)++; |
2b87c45b | 1536 | curpos = obj_offset; |
21666f1a | 1537 | type = unpack_object_header(p, &w_curs, &curpos, &dummy); |
ad8c80a5 | 1538 | } |
ad8c80a5 JH |
1539 | } |
1540 | ||
c4001d92 | 1541 | static int packed_object_info(struct packed_git *p, off_t obj_offset, |
21666f1a | 1542 | unsigned long *sizep) |
1f688557 | 1543 | { |
03e79c88 | 1544 | struct pack_window *w_curs = NULL; |
c4001d92 SP |
1545 | unsigned long size; |
1546 | off_t curpos = obj_offset; | |
21666f1a | 1547 | enum object_type type; |
5db47c2b | 1548 | |
21666f1a | 1549 | type = unpack_object_header(p, &w_curs, &curpos, &size); |
5db47c2b | 1550 | |
21666f1a | 1551 | switch (type) { |
eb32d236 NP |
1552 | case OBJ_OFS_DELTA: |
1553 | case OBJ_REF_DELTA: | |
21666f1a NP |
1554 | type = packed_delta_info(p, &w_curs, curpos, |
1555 | type, obj_offset, sizep); | |
2b87c45b | 1556 | break; |
a733cb60 | 1557 | case OBJ_COMMIT: |
a733cb60 | 1558 | case OBJ_TREE: |
a733cb60 | 1559 | case OBJ_BLOB: |
a733cb60 | 1560 | case OBJ_TAG: |
2b87c45b NP |
1561 | if (sizep) |
1562 | *sizep = size; | |
a69d0943 | 1563 | break; |
1f688557 | 1564 | default: |
3d77d877 NP |
1565 | error("unknown object type %i at offset %"PRIuMAX" in %s", |
1566 | type, (uintmax_t)obj_offset, p->pack_name); | |
1567 | type = OBJ_BAD; | |
1f688557 | 1568 | } |
2b87c45b | 1569 | unuse_pack(&w_curs); |
21666f1a | 1570 | return type; |
1f688557 JH |
1571 | } |
1572 | ||
eb950c19 | 1573 | static void *unpack_compressed_entry(struct packed_git *p, |
03e79c88 | 1574 | struct pack_window **w_curs, |
c4001d92 | 1575 | off_t curpos, |
eb950c19 | 1576 | unsigned long size) |
de530aaa SP |
1577 | { |
1578 | int st; | |
1579 | z_stream stream; | |
079afb18 | 1580 | unsigned char *buffer, *in; |
de530aaa SP |
1581 | |
1582 | buffer = xmalloc(size + 1); | |
1583 | buffer[size] = 0; | |
1584 | memset(&stream, 0, sizeof(stream)); | |
de530aaa SP |
1585 | stream.next_out = buffer; |
1586 | stream.avail_out = size; | |
1587 | ||
39c68542 | 1588 | git_inflate_init(&stream); |
079afb18 | 1589 | do { |
2b87c45b | 1590 | in = use_pack(p, w_curs, curpos, &stream.avail_in); |
079afb18 | 1591 | stream.next_in = in; |
39c68542 | 1592 | st = git_inflate(&stream, Z_FINISH); |
2b87c45b | 1593 | curpos += stream.next_in - in; |
079afb18 | 1594 | } while (st == Z_OK || st == Z_BUF_ERROR); |
39c68542 | 1595 | git_inflate_end(&stream); |
de530aaa SP |
1596 | if ((st != Z_STREAM_END) || stream.total_out != size) { |
1597 | free(buffer); | |
1598 | return NULL; | |
1599 | } | |
1600 | ||
1601 | return buffer; | |
1602 | } | |
1603 | ||
e5e01619 LT |
1604 | #define MAX_DELTA_CACHE (256) |
1605 | ||
18bdec11 | 1606 | static size_t delta_base_cached; |
5e08ecbf NP |
1607 | |
1608 | static struct delta_base_cache_lru_list { | |
1609 | struct delta_base_cache_lru_list *prev; | |
1610 | struct delta_base_cache_lru_list *next; | |
1611 | } delta_base_cache_lru = { &delta_base_cache_lru, &delta_base_cache_lru }; | |
1612 | ||
e5e01619 | 1613 | static struct delta_base_cache_entry { |
5e08ecbf NP |
1614 | struct delta_base_cache_lru_list lru; |
1615 | void *data; | |
e5e01619 LT |
1616 | struct packed_git *p; |
1617 | off_t base_offset; | |
1618 | unsigned long size; | |
e5e01619 LT |
1619 | enum object_type type; |
1620 | } delta_base_cache[MAX_DELTA_CACHE]; | |
1621 | ||
1622 | static unsigned long pack_entry_hash(struct packed_git *p, off_t base_offset) | |
1623 | { | |
1624 | unsigned long hash; | |
1625 | ||
1626 | hash = (unsigned long)p + (unsigned long)base_offset; | |
1627 | hash += (hash >> 8) + (hash >> 16); | |
3358004a | 1628 | return hash % MAX_DELTA_CACHE; |
e5e01619 LT |
1629 | } |
1630 | ||
62f255ad | 1631 | static void *cache_or_unpack_entry(struct packed_git *p, off_t base_offset, |
a0cba108 | 1632 | unsigned long *base_size, enum object_type *type, int keep_cache) |
62f255ad | 1633 | { |
e5e01619 LT |
1634 | void *ret; |
1635 | unsigned long hash = pack_entry_hash(p, base_offset); | |
1636 | struct delta_base_cache_entry *ent = delta_base_cache + hash; | |
1637 | ||
1638 | ret = ent->data; | |
749bc58c MV |
1639 | if (!ret || ent->p != p || ent->base_offset != base_offset) |
1640 | return unpack_entry(p, base_offset, type, base_size); | |
e5e01619 | 1641 | |
18bdec11 | 1642 | if (!keep_cache) { |
a0cba108 | 1643 | ent->data = NULL; |
5e08ecbf NP |
1644 | ent->lru.next->prev = ent->lru.prev; |
1645 | ent->lru.prev->next = ent->lru.next; | |
18bdec11 | 1646 | delta_base_cached -= ent->size; |
182af834 PH |
1647 | } else { |
1648 | ret = xmemdupz(ent->data, ent->size); | |
a0cba108 | 1649 | } |
e5e01619 LT |
1650 | *type = ent->type; |
1651 | *base_size = ent->size; | |
1652 | return ret; | |
62f255ad LT |
1653 | } |
1654 | ||
18bdec11 SP |
1655 | static inline void release_delta_base_cache(struct delta_base_cache_entry *ent) |
1656 | { | |
1657 | if (ent->data) { | |
1658 | free(ent->data); | |
1659 | ent->data = NULL; | |
5e08ecbf NP |
1660 | ent->lru.next->prev = ent->lru.prev; |
1661 | ent->lru.prev->next = ent->lru.next; | |
18bdec11 SP |
1662 | delta_base_cached -= ent->size; |
1663 | } | |
1664 | } | |
1665 | ||
62f255ad LT |
1666 | static void add_delta_base_cache(struct packed_git *p, off_t base_offset, |
1667 | void *base, unsigned long base_size, enum object_type type) | |
1668 | { | |
5e08ecbf | 1669 | unsigned long hash = pack_entry_hash(p, base_offset); |
e5e01619 | 1670 | struct delta_base_cache_entry *ent = delta_base_cache + hash; |
5e08ecbf | 1671 | struct delta_base_cache_lru_list *lru; |
e5e01619 | 1672 | |
18bdec11 SP |
1673 | release_delta_base_cache(ent); |
1674 | delta_base_cached += base_size; | |
5e08ecbf NP |
1675 | |
1676 | for (lru = delta_base_cache_lru.next; | |
1677 | delta_base_cached > delta_base_cache_limit | |
1678 | && lru != &delta_base_cache_lru; | |
1679 | lru = lru->next) { | |
1680 | struct delta_base_cache_entry *f = (void *)lru; | |
18bdec11 SP |
1681 | if (f->type == OBJ_BLOB) |
1682 | release_delta_base_cache(f); | |
1683 | } | |
5e08ecbf NP |
1684 | for (lru = delta_base_cache_lru.next; |
1685 | delta_base_cached > delta_base_cache_limit | |
1686 | && lru != &delta_base_cache_lru; | |
1687 | lru = lru->next) { | |
1688 | struct delta_base_cache_entry *f = (void *)lru; | |
1689 | release_delta_base_cache(f); | |
1690 | } | |
18bdec11 | 1691 | |
e5e01619 LT |
1692 | ent->p = p; |
1693 | ent->base_offset = base_offset; | |
1694 | ent->type = type; | |
1695 | ent->data = base; | |
1696 | ent->size = base_size; | |
5e08ecbf NP |
1697 | ent->lru.next = &delta_base_cache_lru; |
1698 | ent->lru.prev = delta_base_cache_lru.prev; | |
1699 | delta_base_cache_lru.prev->next = &ent->lru; | |
1700 | delta_base_cache_lru.prev = &ent->lru; | |
62f255ad LT |
1701 | } |
1702 | ||
c2c5b270 CC |
1703 | static void *read_object(const unsigned char *sha1, enum object_type *type, |
1704 | unsigned long *size); | |
1705 | ||
eb950c19 | 1706 | static void *unpack_delta_entry(struct packed_git *p, |
03e79c88 | 1707 | struct pack_window **w_curs, |
c4001d92 | 1708 | off_t curpos, |
1f688557 | 1709 | unsigned long delta_size, |
c4001d92 | 1710 | off_t obj_offset, |
21666f1a | 1711 | enum object_type *type, |
eb950c19 | 1712 | unsigned long *sizep) |
1f688557 | 1713 | { |
7c3e8be3 | 1714 | void *delta_data, *result, *base; |
c4001d92 SP |
1715 | unsigned long base_size; |
1716 | off_t base_offset; | |
43057304 | 1717 | |
21666f1a | 1718 | base_offset = get_delta_base(p, w_curs, &curpos, *type, obj_offset); |
8eca0b47 NP |
1719 | if (!base_offset) { |
1720 | error("failed to validate delta base reference " | |
1721 | "at offset %"PRIuMAX" from %s", | |
1722 | (uintmax_t)curpos, p->pack_name); | |
1723 | return NULL; | |
1724 | } | |
eac12e2d | 1725 | unuse_pack(w_curs); |
a0cba108 | 1726 | base = cache_or_unpack_entry(p, base_offset, &base_size, type, 0); |
8eca0b47 NP |
1727 | if (!base) { |
1728 | /* | |
1729 | * We're probably in deep shit, but let's try to fetch | |
1730 | * the required base anyway from another pack or loose. | |
1731 | * This is costly but should happen only in the presence | |
1732 | * of a corrupted pack, and is better than failing outright. | |
1733 | */ | |
08698b1e NP |
1734 | struct revindex_entry *revidx; |
1735 | const unsigned char *base_sha1; | |
1736 | revidx = find_pack_revindex(p, base_offset); | |
1737 | if (!revidx) | |
1738 | return NULL; | |
1739 | base_sha1 = nth_packed_object_sha1(p, revidx->nr); | |
8eca0b47 NP |
1740 | error("failed to read delta base object %s" |
1741 | " at offset %"PRIuMAX" from %s", | |
1742 | sha1_to_hex(base_sha1), (uintmax_t)base_offset, | |
1743 | p->pack_name); | |
1744 | mark_bad_packed_object(p, base_sha1); | |
ac939109 | 1745 | base = read_object(base_sha1, type, &base_size); |
8eca0b47 NP |
1746 | if (!base) |
1747 | return NULL; | |
1748 | } | |
67686d95 | 1749 | |
2b87c45b | 1750 | delta_data = unpack_compressed_entry(p, w_curs, curpos, delta_size); |
8eca0b47 NP |
1751 | if (!delta_data) { |
1752 | error("failed to unpack compressed delta " | |
1753 | "at offset %"PRIuMAX" from %s", | |
1754 | (uintmax_t)curpos, p->pack_name); | |
1755 | free(base); | |
1756 | return NULL; | |
1757 | } | |
1f688557 JH |
1758 | result = patch_delta(base, base_size, |
1759 | delta_data, delta_size, | |
21666f1a | 1760 | sizep); |
1f688557 JH |
1761 | if (!result) |
1762 | die("failed to apply delta"); | |
1763 | free(delta_data); | |
62f255ad | 1764 | add_delta_base_cache(p, base_offset, base, base_size, *type); |
1f688557 JH |
1765 | return result; |
1766 | } | |
1767 | ||
0e8189e2 NP |
1768 | int do_check_packed_object_crc; |
1769 | ||
c4001d92 | 1770 | void *unpack_entry(struct packed_git *p, off_t obj_offset, |
21666f1a | 1771 | enum object_type *type, unsigned long *sizep) |
1f688557 | 1772 | { |
03e79c88 | 1773 | struct pack_window *w_curs = NULL; |
c4001d92 | 1774 | off_t curpos = obj_offset; |
21666f1a | 1775 | void *data; |
1f688557 | 1776 | |
0e8189e2 NP |
1777 | if (do_check_packed_object_crc && p->index_version > 1) { |
1778 | struct revindex_entry *revidx = find_pack_revindex(p, obj_offset); | |
1779 | unsigned long len = revidx[1].offset - obj_offset; | |
1780 | if (check_pack_crc(p, &w_curs, obj_offset, len, revidx->nr)) { | |
1781 | const unsigned char *sha1 = | |
1782 | nth_packed_object_sha1(p, revidx->nr); | |
1783 | error("bad packed object CRC for %s", | |
1784 | sha1_to_hex(sha1)); | |
1785 | mark_bad_packed_object(p, sha1); | |
f755bb99 | 1786 | unuse_pack(&w_curs); |
0e8189e2 NP |
1787 | return NULL; |
1788 | } | |
1789 | } | |
1790 | ||
21666f1a NP |
1791 | *type = unpack_object_header(p, &w_curs, &curpos, sizep); |
1792 | switch (*type) { | |
eb32d236 NP |
1793 | case OBJ_OFS_DELTA: |
1794 | case OBJ_REF_DELTA: | |
21666f1a NP |
1795 | data = unpack_delta_entry(p, &w_curs, curpos, *sizep, |
1796 | obj_offset, type, sizep); | |
4d703a1a | 1797 | break; |
a733cb60 | 1798 | case OBJ_COMMIT: |
a733cb60 | 1799 | case OBJ_TREE: |
a733cb60 | 1800 | case OBJ_BLOB: |
a733cb60 | 1801 | case OBJ_TAG: |
21666f1a | 1802 | data = unpack_compressed_entry(p, &w_curs, curpos, *sizep); |
4d703a1a | 1803 | break; |
1f688557 | 1804 | default: |
8eca0b47 NP |
1805 | data = NULL; |
1806 | error("unknown object type %i at offset %"PRIuMAX" in %s", | |
1807 | *type, (uintmax_t)obj_offset, p->pack_name); | |
1f688557 | 1808 | } |
03e79c88 | 1809 | unuse_pack(&w_curs); |
21666f1a | 1810 | return data; |
1f688557 JH |
1811 | } |
1812 | ||
d079837e | 1813 | const unsigned char *nth_packed_object_sha1(struct packed_git *p, |
d72308e0 | 1814 | uint32_t n) |
9a217f2a | 1815 | { |
42873078 | 1816 | const unsigned char *index = p->index_data; |
d079837e SP |
1817 | if (!index) { |
1818 | if (open_pack_index(p)) | |
1819 | return NULL; | |
1820 | index = p->index_data; | |
1821 | } | |
57059091 | 1822 | if (n >= p->num_objects) |
d72308e0 | 1823 | return NULL; |
74e34e1f NP |
1824 | index += 4 * 256; |
1825 | if (p->index_version == 1) { | |
1826 | return index + 24 * n + 4; | |
1827 | } else { | |
1828 | index += 8; | |
1829 | return index + 20 * n; | |
1830 | } | |
1831 | } | |
1832 | ||
99093238 | 1833 | off_t nth_packed_object_offset(const struct packed_git *p, uint32_t n) |
74e34e1f NP |
1834 | { |
1835 | const unsigned char *index = p->index_data; | |
1836 | index += 4 * 256; | |
1837 | if (p->index_version == 1) { | |
1838 | return ntohl(*((uint32_t *)(index + 24 * n))); | |
1839 | } else { | |
1840 | uint32_t off; | |
1841 | index += 8 + p->num_objects * (20 + 4); | |
1842 | off = ntohl(*((uint32_t *)(index + 4 * n))); | |
1843 | if (!(off & 0x80000000)) | |
1844 | return off; | |
1845 | index += p->num_objects * 4 + (off & 0x7fffffff) * 8; | |
1846 | return (((uint64_t)ntohl(*((uint32_t *)(index + 0)))) << 32) | | |
1847 | ntohl(*((uint32_t *)(index + 4))); | |
1848 | } | |
9a217f2a JH |
1849 | } |
1850 | ||
c4001d92 | 1851 | off_t find_pack_entry_one(const unsigned char *sha1, |
43057304 | 1852 | struct packed_git *p) |
1f688557 | 1853 | { |
42873078 | 1854 | const uint32_t *level1_ofs = p->index_data; |
42873078 | 1855 | const unsigned char *index = p->index_data; |
628522ec JH |
1856 | unsigned hi, lo, stride; |
1857 | static int use_lookup = -1; | |
1858 | static int debug_lookup = -1; | |
1859 | ||
1860 | if (debug_lookup < 0) | |
1861 | debug_lookup = !!getenv("GIT_DEBUG_LOOKUP"); | |
42873078 | 1862 | |
d079837e SP |
1863 | if (!index) { |
1864 | if (open_pack_index(p)) | |
1865 | return 0; | |
1866 | level1_ofs = p->index_data; | |
1867 | index = p->index_data; | |
1868 | } | |
74e34e1f NP |
1869 | if (p->index_version > 1) { |
1870 | level1_ofs += 2; | |
1871 | index += 8; | |
1872 | } | |
42873078 | 1873 | index += 4 * 256; |
74e34e1f NP |
1874 | hi = ntohl(level1_ofs[*sha1]); |
1875 | lo = ((*sha1 == 0x0) ? 0 : ntohl(level1_ofs[*sha1 - 1])); | |
628522ec JH |
1876 | if (p->index_version > 1) { |
1877 | stride = 20; | |
1878 | } else { | |
1879 | stride = 24; | |
1880 | index += 4; | |
1881 | } | |
1882 | ||
1883 | if (debug_lookup) | |
6e1c2344 | 1884 | printf("%02x%02x%02x... lo %u hi %u nr %"PRIu32"\n", |
628522ec JH |
1885 | sha1[0], sha1[1], sha1[2], lo, hi, p->num_objects); |
1886 | ||
1887 | if (use_lookup < 0) | |
1888 | use_lookup = !!getenv("GIT_USE_LOOKUP"); | |
1889 | if (use_lookup) { | |
1890 | int pos = sha1_entry_pos(index, stride, 0, | |
1891 | lo, hi, p->num_objects, sha1); | |
1892 | if (pos < 0) | |
1893 | return 0; | |
1894 | return nth_packed_object_offset(p, pos); | |
1895 | } | |
1f688557 JH |
1896 | |
1897 | do { | |
74e34e1f | 1898 | unsigned mi = (lo + hi) / 2; |
628522ec JH |
1899 | int cmp = hashcmp(index + mi * stride, sha1); |
1900 | ||
1901 | if (debug_lookup) | |
1902 | printf("lo %u hi %u rg %u mi %u\n", | |
1903 | lo, hi, hi - lo, mi); | |
43057304 | 1904 | if (!cmp) |
74e34e1f | 1905 | return nth_packed_object_offset(p, mi); |
1f688557 JH |
1906 | if (cmp > 0) |
1907 | hi = mi; | |
1908 | else | |
1909 | lo = mi+1; | |
1910 | } while (lo < hi); | |
1911 | return 0; | |
1912 | } | |
1913 | ||
000dfd3f | 1914 | int matches_pack_name(struct packed_git *p, const char *name) |
d4ff6d92 SP |
1915 | { |
1916 | const char *last_c, *c; | |
1917 | ||
000dfd3f JH |
1918 | if (!strcmp(p->pack_name, name)) |
1919 | return 1; | |
d4ff6d92 SP |
1920 | |
1921 | for (c = p->pack_name, last_c = c; *c;) | |
1922 | if (*c == '/') | |
1923 | last_c = ++c; | |
1924 | else | |
1925 | ++c; | |
000dfd3f JH |
1926 | if (!strcmp(last_c, name)) |
1927 | return 1; | |
d4ff6d92 | 1928 | |
000dfd3f | 1929 | return 0; |
d4ff6d92 SP |
1930 | } |
1931 | ||
106d710b | 1932 | static int find_pack_entry(const unsigned char *sha1, struct pack_entry *e, const char **ignore_packed) |
1f688557 | 1933 | { |
f7c22cc6 | 1934 | static struct packed_git *last_found = (void *)1; |
1f688557 | 1935 | struct packed_git *p; |
c4001d92 | 1936 | off_t offset; |
43057304 | 1937 | |
1f688557 | 1938 | prepare_packed_git(); |
f7c22cc6 NP |
1939 | if (!packed_git) |
1940 | return 0; | |
1941 | p = (last_found == (void *)1) ? packed_git : last_found; | |
1f688557 | 1942 | |
f7c22cc6 | 1943 | do { |
106d710b JH |
1944 | if (ignore_packed) { |
1945 | const char **ig; | |
1946 | for (ig = ignore_packed; *ig; ig++) | |
000dfd3f | 1947 | if (matches_pack_name(p, *ig)) |
106d710b JH |
1948 | break; |
1949 | if (*ig) | |
f7c22cc6 | 1950 | goto next; |
106d710b | 1951 | } |
f7c22cc6 | 1952 | |
8eca0b47 NP |
1953 | if (p->num_bad_objects) { |
1954 | unsigned i; | |
1955 | for (i = 0; i < p->num_bad_objects; i++) | |
1956 | if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i)) | |
1957 | goto next; | |
1958 | } | |
1959 | ||
43057304 NP |
1960 | offset = find_pack_entry_one(sha1, p); |
1961 | if (offset) { | |
c715f783 SP |
1962 | /* |
1963 | * We are about to tell the caller where they can | |
1964 | * locate the requested object. We better make | |
1965 | * sure the packfile is still here and can be | |
1966 | * accessed before supplying that answer, as | |
1967 | * it may have been deleted since the index | |
1968 | * was loaded! | |
1969 | */ | |
1970 | if (p->pack_fd == -1 && open_packed_git(p)) { | |
1971 | error("packfile %s cannot be accessed", p->pack_name); | |
f7c22cc6 | 1972 | goto next; |
c715f783 | 1973 | } |
43057304 NP |
1974 | e->offset = offset; |
1975 | e->p = p; | |
1976 | hashcpy(e->sha1, sha1); | |
f7c22cc6 | 1977 | last_found = p; |
1f688557 | 1978 | return 1; |
43057304 | 1979 | } |
f7c22cc6 NP |
1980 | |
1981 | next: | |
1982 | if (p == last_found) | |
1983 | p = packed_git; | |
1984 | else | |
1985 | p = p->next; | |
1986 | if (p == last_found) | |
1987 | p = p->next; | |
1988 | } while (p); | |
1f688557 JH |
1989 | return 0; |
1990 | } | |
1991 | ||
a6080a0a | 1992 | struct packed_git *find_sha1_pack(const unsigned char *sha1, |
bf592c50 DB |
1993 | struct packed_git *packs) |
1994 | { | |
1995 | struct packed_git *p; | |
bf592c50 DB |
1996 | |
1997 | for (p = packs; p; p = p->next) { | |
43057304 | 1998 | if (find_pack_entry_one(sha1, p)) |
bf592c50 DB |
1999 | return p; |
2000 | } | |
2001 | return NULL; | |
21666f1a | 2002 | |
bf592c50 DB |
2003 | } |
2004 | ||
21666f1a | 2005 | static int sha1_loose_object_info(const unsigned char *sha1, unsigned long *sizep) |
65c2e0c3 | 2006 | { |
36e4d74a | 2007 | int status; |
65c2e0c3 JH |
2008 | unsigned long mapsize, size; |
2009 | void *map; | |
2010 | z_stream stream; | |
d65a16f6 | 2011 | char hdr[32]; |
65c2e0c3 | 2012 | |
bb6b8e4f | 2013 | map = map_sha1_file(sha1, &mapsize); |
f0df4ed5 JS |
2014 | if (!map) |
2015 | return error("unable to find %s", sha1_to_hex(sha1)); | |
36e4d74a JH |
2016 | if (unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)) < 0) |
2017 | status = error("unable to unpack %s header", | |
2018 | sha1_to_hex(sha1)); | |
21666f1a | 2019 | else if ((status = parse_sha1_header(hdr, &size)) < 0) |
36e4d74a | 2020 | status = error("unable to parse %s header", sha1_to_hex(sha1)); |
21666f1a NP |
2021 | else if (sizep) |
2022 | *sizep = size; | |
39c68542 | 2023 | git_inflate_end(&stream); |
65c2e0c3 JH |
2024 | munmap(map, mapsize); |
2025 | return status; | |
2026 | } | |
2027 | ||
21666f1a | 2028 | int sha1_object_info(const unsigned char *sha1, unsigned long *sizep) |
f0df4ed5 | 2029 | { |
f0df4ed5 | 2030 | struct pack_entry e; |
ddd63e64 | 2031 | int status; |
f0df4ed5 JS |
2032 | |
2033 | if (!find_pack_entry(sha1, &e, NULL)) { | |
ddd63e64 SG |
2034 | /* Most likely it's a loose object. */ |
2035 | status = sha1_loose_object_info(sha1, sizep); | |
2036 | if (status >= 0) | |
2037 | return status; | |
2038 | ||
2039 | /* Not a loose object; someone else may have just packed it. */ | |
f0df4ed5 JS |
2040 | reprepare_packed_git(); |
2041 | if (!find_pack_entry(sha1, &e, NULL)) | |
ddd63e64 | 2042 | return status; |
f0df4ed5 | 2043 | } |
3d77d877 NP |
2044 | |
2045 | status = packed_object_info(e.p, e.offset, sizep); | |
2046 | if (status < 0) { | |
2047 | mark_bad_packed_object(e.p, sha1); | |
2048 | status = sha1_object_info(sha1, sizep); | |
2049 | } | |
2050 | ||
2051 | return status; | |
f0df4ed5 JS |
2052 | } |
2053 | ||
21666f1a NP |
2054 | static void *read_packed_sha1(const unsigned char *sha1, |
2055 | enum object_type *type, unsigned long *size) | |
1f688557 JH |
2056 | { |
2057 | struct pack_entry e; | |
8eca0b47 | 2058 | void *data; |
1f688557 | 2059 | |
8276c007 | 2060 | if (!find_pack_entry(sha1, &e, NULL)) |
1f688557 | 2061 | return NULL; |
8eca0b47 NP |
2062 | data = cache_or_unpack_entry(e.p, e.offset, size, type, 1); |
2063 | if (!data) { | |
2064 | /* | |
2065 | * We're probably in deep shit, but let's try to fetch | |
2066 | * the required object anyway from another pack or loose. | |
2067 | * This should happen only in the presence of a corrupted | |
2068 | * pack, and is better than failing outright. | |
2069 | */ | |
2070 | error("failed to read object %s at offset %"PRIuMAX" from %s", | |
2071 | sha1_to_hex(sha1), (uintmax_t)e.offset, e.p->pack_name); | |
2072 | mark_bad_packed_object(e.p, sha1); | |
ac939109 | 2073 | data = read_object(sha1, type, size); |
8eca0b47 NP |
2074 | } |
2075 | return data; | |
1f688557 JH |
2076 | } |
2077 | ||
d66b37bb JH |
2078 | /* |
2079 | * This is meant to hold a *small* number of objects that you would | |
2080 | * want read_sha1_file() to be able to return, but yet you do not want | |
2081 | * to write them into the object store (e.g. a browse-only | |
2082 | * application). | |
2083 | */ | |
2084 | static struct cached_object { | |
2085 | unsigned char sha1[20]; | |
21666f1a | 2086 | enum object_type type; |
d66b37bb JH |
2087 | void *buf; |
2088 | unsigned long size; | |
2089 | } *cached_objects; | |
2090 | static int cached_object_nr, cached_object_alloc; | |
2091 | ||
346245a1 | 2092 | static struct cached_object empty_tree = { |
14d9c578 | 2093 | EMPTY_TREE_SHA1_BIN, |
346245a1 JK |
2094 | OBJ_TREE, |
2095 | "", | |
2096 | 0 | |
2097 | }; | |
2098 | ||
d66b37bb JH |
2099 | static struct cached_object *find_cached_object(const unsigned char *sha1) |
2100 | { | |
2101 | int i; | |
2102 | struct cached_object *co = cached_objects; | |
2103 | ||
2104 | for (i = 0; i < cached_object_nr; i++, co++) { | |
2105 | if (!hashcmp(co->sha1, sha1)) | |
2106 | return co; | |
2107 | } | |
346245a1 JK |
2108 | if (!hashcmp(sha1, empty_tree.sha1)) |
2109 | return &empty_tree; | |
d66b37bb JH |
2110 | return NULL; |
2111 | } | |
2112 | ||
21666f1a NP |
2113 | int pretend_sha1_file(void *buf, unsigned long len, enum object_type type, |
2114 | unsigned char *sha1) | |
d66b37bb JH |
2115 | { |
2116 | struct cached_object *co; | |
2117 | ||
21666f1a | 2118 | hash_sha1_file(buf, len, typename(type), sha1); |
d66b37bb JH |
2119 | if (has_sha1_file(sha1) || find_cached_object(sha1)) |
2120 | return 0; | |
2121 | if (cached_object_alloc <= cached_object_nr) { | |
2122 | cached_object_alloc = alloc_nr(cached_object_alloc); | |
2123 | cached_objects = xrealloc(cached_objects, | |
2124 | sizeof(*cached_objects) * | |
2125 | cached_object_alloc); | |
2126 | } | |
2127 | co = &cached_objects[cached_object_nr++]; | |
2128 | co->size = len; | |
21666f1a | 2129 | co->type = type; |
efa13f7b JH |
2130 | co->buf = xmalloc(len); |
2131 | memcpy(co->buf, buf, len); | |
d66b37bb JH |
2132 | hashcpy(co->sha1, sha1); |
2133 | return 0; | |
2134 | } | |
2135 | ||
c2c5b270 CC |
2136 | static void *read_object(const unsigned char *sha1, enum object_type *type, |
2137 | unsigned long *size) | |
0fcfd160 LT |
2138 | { |
2139 | unsigned long mapsize; | |
2140 | void *map, *buf; | |
d66b37bb JH |
2141 | struct cached_object *co; |
2142 | ||
2143 | co = find_cached_object(sha1); | |
2144 | if (co) { | |
21666f1a | 2145 | *type = co->type; |
d66b37bb | 2146 | *size = co->size; |
182af834 | 2147 | return xmemdupz(co->buf, co->size); |
d66b37bb | 2148 | } |
0fcfd160 | 2149 | |
8276c007 PE |
2150 | buf = read_packed_sha1(sha1, type, size); |
2151 | if (buf) | |
2152 | return buf; | |
bb6b8e4f | 2153 | map = map_sha1_file(sha1, &mapsize); |
0fcfd160 | 2154 | if (map) { |
7efbff75 | 2155 | buf = unpack_sha1_file(map, mapsize, type, size, sha1); |
0fcfd160 LT |
2156 | munmap(map, mapsize); |
2157 | return buf; | |
2158 | } | |
637cdd9d | 2159 | reprepare_packed_git(); |
8276c007 | 2160 | return read_packed_sha1(sha1, type, size); |
0fcfd160 LT |
2161 | } |
2162 | ||
ac939109 NP |
2163 | void *read_sha1_file(const unsigned char *sha1, enum object_type *type, |
2164 | unsigned long *size) | |
2165 | { | |
2166 | void *data = read_object(sha1, type, size); | |
2167 | /* legacy behavior is to die on corrupted objects */ | |
2168 | if (!data && (has_loose_object(sha1) || has_packed_and_bad(sha1))) | |
2169 | die("object %s is corrupted", sha1_to_hex(sha1)); | |
2170 | return data; | |
2171 | } | |
2172 | ||
40469ee9 | 2173 | void *read_object_with_reference(const unsigned char *sha1, |
21666f1a | 2174 | const char *required_type_name, |
40469ee9 JH |
2175 | unsigned long *size, |
2176 | unsigned char *actual_sha1_return) | |
f4913f91 | 2177 | { |
21666f1a | 2178 | enum object_type type, required_type; |
f4913f91 JH |
2179 | void *buffer; |
2180 | unsigned long isize; | |
40469ee9 | 2181 | unsigned char actual_sha1[20]; |
f4913f91 | 2182 | |
21666f1a | 2183 | required_type = type_from_string(required_type_name); |
e702496e | 2184 | hashcpy(actual_sha1, sha1); |
40469ee9 JH |
2185 | while (1) { |
2186 | int ref_length = -1; | |
2187 | const char *ref_type = NULL; | |
f4913f91 | 2188 | |
21666f1a | 2189 | buffer = read_sha1_file(actual_sha1, &type, &isize); |
40469ee9 JH |
2190 | if (!buffer) |
2191 | return NULL; | |
21666f1a | 2192 | if (type == required_type) { |
40469ee9 JH |
2193 | *size = isize; |
2194 | if (actual_sha1_return) | |
e702496e | 2195 | hashcpy(actual_sha1_return, actual_sha1); |
40469ee9 JH |
2196 | return buffer; |
2197 | } | |
2198 | /* Handle references */ | |
21666f1a | 2199 | else if (type == OBJ_COMMIT) |
40469ee9 | 2200 | ref_type = "tree "; |
21666f1a | 2201 | else if (type == OBJ_TAG) |
40469ee9 JH |
2202 | ref_type = "object "; |
2203 | else { | |
2204 | free(buffer); | |
2205 | return NULL; | |
2206 | } | |
2207 | ref_length = strlen(ref_type); | |
f4913f91 | 2208 | |
50974ec9 MK |
2209 | if (ref_length + 40 > isize || |
2210 | memcmp(buffer, ref_type, ref_length) || | |
1d7f171c | 2211 | get_sha1_hex((char *) buffer + ref_length, actual_sha1)) { |
40469ee9 JH |
2212 | free(buffer); |
2213 | return NULL; | |
2214 | } | |
1cf58e72 | 2215 | free(buffer); |
40469ee9 JH |
2216 | /* Now we have the ID of the referred-to object in |
2217 | * actual_sha1. Check again. */ | |
f4913f91 | 2218 | } |
f4913f91 JH |
2219 | } |
2220 | ||
ce9fbf16 | 2221 | static void write_sha1_file_prepare(const void *buf, unsigned long len, |
972a9155 | 2222 | const char *type, unsigned char *sha1, |
d65a16f6 | 2223 | char *hdr, int *hdrlen) |
d410c0f5 | 2224 | { |
9126f009 | 2225 | git_SHA_CTX c; |
d410c0f5 JH |
2226 | |
2227 | /* Generate the header */ | |
d65a16f6 | 2228 | *hdrlen = sprintf(hdr, "%s %lu", type, len)+1; |
d410c0f5 JH |
2229 | |
2230 | /* Sha1.. */ | |
9126f009 NP |
2231 | git_SHA1_Init(&c); |
2232 | git_SHA1_Update(&c, hdr, *hdrlen); | |
2233 | git_SHA1_Update(&c, buf, len); | |
2234 | git_SHA1_Final(sha1, &c); | |
d410c0f5 JH |
2235 | } |
2236 | ||
230f1322 LT |
2237 | /* |
2238 | * Move the just written object into its final resting place | |
2239 | */ | |
839837b9 | 2240 | int move_temp_to_file(const char *tmpfile, const char *filename) |
230f1322 | 2241 | { |
e32c0a9c TR |
2242 | int ret = 0; |
2243 | if (link(tmpfile, filename)) | |
2244 | ret = errno; | |
7ebb6fca LT |
2245 | |
2246 | /* | |
2247 | * Coda hack - coda doesn't like cross-directory links, | |
2248 | * so we fall back to a rename, which will mean that it | |
2249 | * won't be able to check collisions, but that's not a | |
2250 | * big deal. | |
2251 | * | |
2252 | * The same holds for FAT formatted media. | |
2253 | * | |
2254 | * When this succeeds, we just return 0. We have nothing | |
2255 | * left to unlink. | |
2256 | */ | |
2257 | if (ret && ret != EEXIST) { | |
2258 | if (!rename(tmpfile, filename)) | |
230f1322 | 2259 | return 0; |
9e48b389 | 2260 | ret = errno; |
230f1322 LT |
2261 | } |
2262 | unlink(tmpfile); | |
2263 | if (ret) { | |
2264 | if (ret != EEXIST) { | |
916d081b | 2265 | return error("unable to write sha1 filename %s: %s\n", filename, strerror(ret)); |
230f1322 LT |
2266 | } |
2267 | /* FIXME!!! Collision check here ? */ | |
2268 | } | |
2269 | ||
2270 | return 0; | |
2271 | } | |
2272 | ||
4d548150 LT |
2273 | static int write_buffer(int fd, const void *buf, size_t len) |
2274 | { | |
d34cf19b | 2275 | if (write_in_full(fd, buf, len) < 0) |
93822c22 | 2276 | return error("file write error (%s)", strerror(errno)); |
4d548150 LT |
2277 | return 0; |
2278 | } | |
2279 | ||
ce9fbf16 | 2280 | int hash_sha1_file(const void *buf, unsigned long len, const char *type, |
abdc3fc8 RS |
2281 | unsigned char *sha1) |
2282 | { | |
d65a16f6 | 2283 | char hdr[32]; |
abdc3fc8 RS |
2284 | int hdrlen; |
2285 | write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen); | |
2286 | return 0; | |
2287 | } | |
2288 | ||
e9039dd3 LT |
2289 | /* Finalize a file on disk, and close it. */ |
2290 | static void close_sha1_file(int fd) | |
2291 | { | |
aafe9fba LT |
2292 | if (fsync_object_files) |
2293 | fsync_or_die(fd, "sha1 file"); | |
e9039dd3 LT |
2294 | fchmod(fd, 0444); |
2295 | if (close(fd) != 0) | |
2296 | die("unable to write sha1 file"); | |
2297 | } | |
2298 | ||
5723fe7e LT |
2299 | /* Size of directory component, including the ending '/' */ |
2300 | static inline int directory_size(const char *filename) | |
2301 | { | |
2302 | const char *s = strrchr(filename, '/'); | |
2303 | if (!s) | |
2304 | return 0; | |
2305 | return s - filename + 1; | |
2306 | } | |
2307 | ||
2308 | /* | |
2309 | * This creates a temporary file in the same directory as the final | |
2310 | * 'filename' | |
2311 | * | |
2312 | * We want to avoid cross-directory filename renames, because those | |
2313 | * can have problems on various filesystems (FAT, NFS, Coda). | |
2314 | */ | |
2315 | static int create_tmpfile(char *buffer, size_t bufsiz, const char *filename) | |
2316 | { | |
2317 | int fd, dirlen = directory_size(filename); | |
2318 | ||
2319 | if (dirlen + 20 > bufsiz) { | |
2320 | errno = ENAMETOOLONG; | |
2321 | return -1; | |
2322 | } | |
2323 | memcpy(buffer, filename, dirlen); | |
2324 | strcpy(buffer + dirlen, "tmp_obj_XXXXXX"); | |
2325 | fd = mkstemp(buffer); | |
cbacbf4e | 2326 | if (fd < 0 && dirlen && errno == ENOENT) { |
5723fe7e | 2327 | /* Make sure the directory exists */ |
6ff6af62 | 2328 | memcpy(buffer, filename, dirlen); |
5723fe7e | 2329 | buffer[dirlen-1] = 0; |
3bfaf018 | 2330 | if (mkdir(buffer, 0777) || adjust_shared_perm(buffer)) |
5723fe7e LT |
2331 | return -1; |
2332 | ||
2333 | /* Try again */ | |
2334 | strcpy(buffer + dirlen - 1, "/tmp_obj_XXXXXX"); | |
2335 | fd = mkstemp(buffer); | |
2336 | } | |
2337 | return fd; | |
2338 | } | |
2339 | ||
bbac7311 NP |
2340 | static int write_loose_object(const unsigned char *sha1, char *hdr, int hdrlen, |
2341 | void *buf, unsigned long len, time_t mtime) | |
0fcfd160 | 2342 | { |
915308b1 JK |
2343 | int fd, ret; |
2344 | size_t size; | |
bf0f910d | 2345 | unsigned char *compressed; |
0fcfd160 | 2346 | z_stream stream; |
706bc531 | 2347 | char *filename; |
aac17941 | 2348 | static char tmpfile[PATH_MAX]; |
a44c9a5e | 2349 | |
972a9155 | 2350 | filename = sha1_file_name(sha1); |
5723fe7e | 2351 | fd = create_tmpfile(tmpfile, sizeof(tmpfile), filename); |
aac17941 | 2352 | if (fd < 0) { |
35243577 | 2353 | if (errno == EACCES) |
916d081b PB |
2354 | return error("insufficient permission for adding an object to repository database %s\n", get_object_directory()); |
2355 | else | |
2356 | return error("unable to create temporary sha1 filename %s: %s\n", tmpfile, strerror(errno)); | |
aac17941 LT |
2357 | } |
2358 | ||
0fcfd160 LT |
2359 | /* Set it up */ |
2360 | memset(&stream, 0, sizeof(stream)); | |
12f6c308 | 2361 | deflateInit(&stream, zlib_compression_level); |
93821bd9 | 2362 | size = 8 + deflateBound(&stream, len+hdrlen); |
812666c8 | 2363 | compressed = xmalloc(size); |
0fcfd160 LT |
2364 | |
2365 | /* Compress it */ | |
0fcfd160 LT |
2366 | stream.next_out = compressed; |
2367 | stream.avail_out = size; | |
a44c9a5e LT |
2368 | |
2369 | /* First header.. */ | |
d65a16f6 | 2370 | stream.next_in = (unsigned char *)hdr; |
a44c9a5e | 2371 | stream.avail_in = hdrlen; |
726f852b NP |
2372 | while (deflate(&stream, 0) == Z_OK) |
2373 | /* nothing */; | |
a44c9a5e LT |
2374 | |
2375 | /* Then the data itself.. */ | |
2376 | stream.next_in = buf; | |
2377 | stream.avail_in = len; | |
ac54c277 LT |
2378 | ret = deflate(&stream, Z_FINISH); |
2379 | if (ret != Z_STREAM_END) | |
2380 | die("unable to deflate new object %s (%d)", sha1_to_hex(sha1), ret); | |
2381 | ||
2382 | ret = deflateEnd(&stream); | |
2383 | if (ret != Z_OK) | |
2384 | die("deflateEnd on object %s failed (%d)", sha1_to_hex(sha1), ret); | |
2385 | ||
0fcfd160 LT |
2386 | size = stream.total_out; |
2387 | ||
4d548150 LT |
2388 | if (write_buffer(fd, compressed, size) < 0) |
2389 | die("unable to write sha1 file"); | |
e9039dd3 | 2390 | close_sha1_file(fd); |
383f85b7 | 2391 | free(compressed); |
0fcfd160 | 2392 | |
bbac7311 NP |
2393 | if (mtime) { |
2394 | struct utimbuf utb; | |
2395 | utb.actime = mtime; | |
2396 | utb.modtime = mtime; | |
2397 | if (utime(tmpfile, &utb) < 0) | |
2398 | warning("failed utime() on %s: %s", | |
2399 | tmpfile, strerror(errno)); | |
2400 | } | |
2401 | ||
230f1322 | 2402 | return move_temp_to_file(tmpfile, filename); |
0fcfd160 | 2403 | } |
8237b185 | 2404 | |
bbac7311 NP |
2405 | int write_sha1_file(void *buf, unsigned long len, const char *type, unsigned char *returnsha1) |
2406 | { | |
2407 | unsigned char sha1[20]; | |
2408 | char hdr[32]; | |
2409 | int hdrlen; | |
2410 | ||
2411 | /* Normally if we have it in the pack then we do not bother writing | |
2412 | * it out into .git/objects/??/?{38} file. | |
2413 | */ | |
2414 | write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen); | |
2415 | if (returnsha1) | |
2416 | hashcpy(returnsha1, sha1); | |
2417 | if (has_sha1_file(sha1)) | |
2418 | return 0; | |
2419 | return write_loose_object(sha1, hdr, hdrlen, buf, len, 0); | |
2420 | } | |
2421 | ||
2422 | int force_object_loose(const unsigned char *sha1, time_t mtime) | |
2423 | { | |
bbac7311 NP |
2424 | void *buf; |
2425 | unsigned long len; | |
2426 | enum object_type type; | |
2427 | char hdr[32]; | |
2428 | int hdrlen; | |
1fb23e65 | 2429 | int ret; |
bbac7311 | 2430 | |
c529d75a | 2431 | if (has_loose_object(sha1)) |
bbac7311 NP |
2432 | return 0; |
2433 | buf = read_packed_sha1(sha1, &type, &len); | |
2434 | if (!buf) | |
2435 | return error("cannot read sha1_file for %s", sha1_to_hex(sha1)); | |
2436 | hdrlen = sprintf(hdr, "%s %lu", typename(type), len) + 1; | |
1fb23e65 BS |
2437 | ret = write_loose_object(sha1, hdr, hdrlen, buf, len, mtime); |
2438 | free(buf); | |
2439 | ||
2440 | return ret; | |
bbac7311 NP |
2441 | } |
2442 | ||
bf592c50 DB |
2443 | int has_pack_index(const unsigned char *sha1) |
2444 | { | |
2445 | struct stat st; | |
2446 | if (stat(sha1_pack_index_name(sha1), &st)) | |
2447 | return 0; | |
2448 | return 1; | |
2449 | } | |
2450 | ||
2451 | int has_pack_file(const unsigned char *sha1) | |
2452 | { | |
2453 | struct stat st; | |
2454 | if (stat(sha1_pack_name(sha1), &st)) | |
2455 | return 0; | |
2456 | return 1; | |
2457 | } | |
2458 | ||
106d710b | 2459 | int has_sha1_pack(const unsigned char *sha1, const char **ignore_packed) |
dade09c2 LT |
2460 | { |
2461 | struct pack_entry e; | |
106d710b | 2462 | return find_pack_entry(sha1, &e, ignore_packed); |
dade09c2 LT |
2463 | } |
2464 | ||
8237b185 DB |
2465 | int has_sha1_file(const unsigned char *sha1) |
2466 | { | |
1f688557 JH |
2467 | struct pack_entry e; |
2468 | ||
106d710b | 2469 | if (find_pack_entry(sha1, &e, NULL)) |
1f688557 | 2470 | return 1; |
c529d75a | 2471 | return has_loose_object(sha1); |
8237b185 | 2472 | } |
74400e71 | 2473 | |
43df4f86 DP |
2474 | static int index_mem(unsigned char *sha1, void *buf, size_t size, |
2475 | int write_object, enum object_type type, const char *path) | |
e7332f96 | 2476 | { |
6c510bee | 2477 | int ret, re_allocated = 0; |
74400e71 | 2478 | |
7672db20 | 2479 | if (!type) |
edaec3fb | 2480 | type = OBJ_BLOB; |
6c510bee LT |
2481 | |
2482 | /* | |
2483 | * Convert blobs to git internal format | |
2484 | */ | |
43df4f86 | 2485 | if ((type == OBJ_BLOB) && path) { |
f285a2d7 | 2486 | struct strbuf nbuf = STRBUF_INIT; |
21e5ad50 SP |
2487 | if (convert_to_git(path, buf, size, &nbuf, |
2488 | write_object ? safe_crlf : 0)) { | |
b315c5c0 | 2489 | buf = strbuf_detach(&nbuf, &size); |
6c510bee LT |
2490 | re_allocated = 1; |
2491 | } | |
2492 | } | |
2493 | ||
7672db20 | 2494 | if (write_object) |
edaec3fb | 2495 | ret = write_sha1_file(buf, size, typename(type), sha1); |
abdc3fc8 | 2496 | else |
edaec3fb | 2497 | ret = hash_sha1_file(buf, size, typename(type), sha1); |
43df4f86 | 2498 | if (re_allocated) |
6c510bee | 2499 | free(buf); |
43df4f86 DP |
2500 | return ret; |
2501 | } | |
2502 | ||
2503 | int index_fd(unsigned char *sha1, int fd, struct stat *st, int write_object, | |
2504 | enum object_type type, const char *path) | |
2505 | { | |
2506 | int ret; | |
2507 | size_t size = xsize_t(st->st_size); | |
2508 | ||
2509 | if (!S_ISREG(st->st_mode)) { | |
f285a2d7 | 2510 | struct strbuf sbuf = STRBUF_INIT; |
43df4f86 DP |
2511 | if (strbuf_read(&sbuf, fd, 4096) >= 0) |
2512 | ret = index_mem(sha1, sbuf.buf, sbuf.len, write_object, | |
2513 | type, path); | |
2514 | else | |
2515 | ret = -1; | |
2516 | strbuf_release(&sbuf); | |
2517 | } else if (size) { | |
2518 | void *buf = xmmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0); | |
2519 | ret = index_mem(sha1, buf, size, write_object, type, path); | |
aac17941 | 2520 | munmap(buf, size); |
43df4f86 DP |
2521 | } else |
2522 | ret = index_mem(sha1, NULL, size, write_object, type, path); | |
2523 | close(fd); | |
aac17941 | 2524 | return ret; |
74400e71 | 2525 | } |
ec1fcc16 JH |
2526 | |
2527 | int index_path(unsigned char *sha1, const char *path, struct stat *st, int write_object) | |
2528 | { | |
2529 | int fd; | |
b760d3aa | 2530 | struct strbuf sb = STRBUF_INIT; |
ec1fcc16 JH |
2531 | |
2532 | switch (st->st_mode & S_IFMT) { | |
2533 | case S_IFREG: | |
2534 | fd = open(path, O_RDONLY); | |
2535 | if (fd < 0) | |
2536 | return error("open(\"%s\"): %s", path, | |
2537 | strerror(errno)); | |
53bca91a | 2538 | if (index_fd(sha1, fd, st, write_object, OBJ_BLOB, path) < 0) |
ec1fcc16 JH |
2539 | return error("%s: failed to insert into database", |
2540 | path); | |
2541 | break; | |
2542 | case S_IFLNK: | |
b760d3aa | 2543 | if (strbuf_readlink(&sb, path, st->st_size)) { |
ec1fcc16 | 2544 | char *errstr = strerror(errno); |
ec1fcc16 JH |
2545 | return error("readlink(\"%s\"): %s", path, |
2546 | errstr); | |
2547 | } | |
abdc3fc8 | 2548 | if (!write_object) |
b760d3aa LT |
2549 | hash_sha1_file(sb.buf, sb.len, blob_type, sha1); |
2550 | else if (write_sha1_file(sb.buf, sb.len, blob_type, sha1)) | |
ec1fcc16 JH |
2551 | return error("%s: failed to insert into database", |
2552 | path); | |
b760d3aa | 2553 | strbuf_release(&sb); |
ec1fcc16 | 2554 | break; |
f35a6d3b LT |
2555 | case S_IFDIR: |
2556 | return resolve_gitlink_ref(path, "HEAD", sha1); | |
ec1fcc16 JH |
2557 | default: |
2558 | return error("%s: unsupported file type", path); | |
2559 | } | |
2560 | return 0; | |
2561 | } | |
a69e5429 JH |
2562 | |
2563 | int read_pack_header(int fd, struct pack_header *header) | |
2564 | { | |
c697ad14 HO |
2565 | if (read_in_full(fd, header, sizeof(*header)) < sizeof(*header)) |
2566 | /* "eof before pack header was fully read" */ | |
2567 | return PH_ERROR_EOF; | |
2568 | ||
a69e5429 JH |
2569 | if (header->hdr_signature != htonl(PACK_SIGNATURE)) |
2570 | /* "protocol error (pack signature mismatch detected)" */ | |
2571 | return PH_ERROR_PACK_SIGNATURE; | |
2572 | if (!pack_version_ok(header->hdr_version)) | |
2573 | /* "protocol error (pack version unsupported)" */ | |
2574 | return PH_ERROR_PROTOCOL; | |
2575 | return 0; | |
2576 | } |