]>
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" |
0fcfd160 | 17 | |
144bde78 LT |
18 | #ifndef O_NOATIME |
19 | #if defined(__linux__) && (defined(__i386__) || defined(__PPC__)) | |
20 | #define O_NOATIME 01000000 | |
21 | #else | |
22 | #define O_NOATIME 0 | |
23 | #endif | |
24 | #endif | |
25 | ||
e05db0fd PR |
26 | #ifdef NO_C99_FORMAT |
27 | #define SZ_FMT "lu" | |
28 | #else | |
29 | #define SZ_FMT "zu" | |
30 | #endif | |
31 | ||
96f1e58f | 32 | const unsigned char null_sha1[20]; |
88cd621d | 33 | |
144bde78 LT |
34 | static unsigned int sha1_file_open_flag = O_NOATIME; |
35 | ||
192a6be2 | 36 | const signed char hexval_table[256] = { |
e49521b5 JH |
37 | -1, -1, -1, -1, -1, -1, -1, -1, /* 00-07 */ |
38 | -1, -1, -1, -1, -1, -1, -1, -1, /* 08-0f */ | |
39 | -1, -1, -1, -1, -1, -1, -1, -1, /* 10-17 */ | |
40 | -1, -1, -1, -1, -1, -1, -1, -1, /* 18-1f */ | |
41 | -1, -1, -1, -1, -1, -1, -1, -1, /* 20-27 */ | |
42 | -1, -1, -1, -1, -1, -1, -1, -1, /* 28-2f */ | |
43 | 0, 1, 2, 3, 4, 5, 6, 7, /* 30-37 */ | |
44 | 8, 9, -1, -1, -1, -1, -1, -1, /* 38-3f */ | |
45 | -1, 10, 11, 12, 13, 14, 15, -1, /* 40-47 */ | |
46 | -1, -1, -1, -1, -1, -1, -1, -1, /* 48-4f */ | |
47 | -1, -1, -1, -1, -1, -1, -1, -1, /* 50-57 */ | |
48 | -1, -1, -1, -1, -1, -1, -1, -1, /* 58-5f */ | |
49 | -1, 10, 11, 12, 13, 14, 15, -1, /* 60-67 */ | |
50 | -1, -1, -1, -1, -1, -1, -1, -1, /* 68-67 */ | |
51 | -1, -1, -1, -1, -1, -1, -1, -1, /* 70-77 */ | |
52 | -1, -1, -1, -1, -1, -1, -1, -1, /* 78-7f */ | |
53 | -1, -1, -1, -1, -1, -1, -1, -1, /* 80-87 */ | |
54 | -1, -1, -1, -1, -1, -1, -1, -1, /* 88-8f */ | |
55 | -1, -1, -1, -1, -1, -1, -1, -1, /* 90-97 */ | |
56 | -1, -1, -1, -1, -1, -1, -1, -1, /* 98-9f */ | |
57 | -1, -1, -1, -1, -1, -1, -1, -1, /* a0-a7 */ | |
58 | -1, -1, -1, -1, -1, -1, -1, -1, /* a8-af */ | |
59 | -1, -1, -1, -1, -1, -1, -1, -1, /* b0-b7 */ | |
60 | -1, -1, -1, -1, -1, -1, -1, -1, /* b8-bf */ | |
61 | -1, -1, -1, -1, -1, -1, -1, -1, /* c0-c7 */ | |
62 | -1, -1, -1, -1, -1, -1, -1, -1, /* c8-cf */ | |
63 | -1, -1, -1, -1, -1, -1, -1, -1, /* d0-d7 */ | |
64 | -1, -1, -1, -1, -1, -1, -1, -1, /* d8-df */ | |
65 | -1, -1, -1, -1, -1, -1, -1, -1, /* e0-e7 */ | |
66 | -1, -1, -1, -1, -1, -1, -1, -1, /* e8-ef */ | |
67 | -1, -1, -1, -1, -1, -1, -1, -1, /* f0-f7 */ | |
68 | -1, -1, -1, -1, -1, -1, -1, -1, /* f8-ff */ | |
69 | }; | |
0fcfd160 LT |
70 | |
71 | int get_sha1_hex(const char *hex, unsigned char *sha1) | |
72 | { | |
73 | int i; | |
74 | for (i = 0; i < 20; i++) { | |
75 | unsigned int val = (hexval(hex[0]) << 4) | hexval(hex[1]); | |
76 | if (val & ~0xff) | |
77 | return -1; | |
78 | *sha1++ = val; | |
79 | hex += 2; | |
80 | } | |
81 | return 0; | |
82 | } | |
83 | ||
b2cb9425 JH |
84 | int safe_create_leading_directories(char *path) |
85 | { | |
86 | char *pos = path; | |
67d42212 JR |
87 | struct stat st; |
88 | ||
67ffdf4c JS |
89 | if (*pos == '/') |
90 | pos++; | |
b2cb9425 JH |
91 | |
92 | while (pos) { | |
93 | pos = strchr(pos, '/'); | |
94 | if (!pos) | |
95 | break; | |
96 | *pos = 0; | |
67d42212 JR |
97 | if (!stat(path, &st)) { |
98 | /* path exists */ | |
99 | if (!S_ISDIR(st.st_mode)) { | |
b2cb9425 | 100 | *pos = '/'; |
67d42212 | 101 | return -3; |
b2cb9425 | 102 | } |
457f06d6 | 103 | } |
67d42212 JR |
104 | else if (mkdir(path, 0777)) { |
105 | *pos = '/'; | |
106 | return -1; | |
107 | } | |
457f06d6 JS |
108 | else if (adjust_shared_perm(path)) { |
109 | *pos = '/'; | |
110 | return -2; | |
111 | } | |
b2cb9425 JH |
112 | *pos++ = '/'; |
113 | } | |
114 | return 0; | |
115 | } | |
723c31fe | 116 | |
0fcfd160 LT |
117 | char * sha1_to_hex(const unsigned char *sha1) |
118 | { | |
dcb3450f LT |
119 | static int bufno; |
120 | static char hexbuffer[4][50]; | |
0fcfd160 | 121 | static const char hex[] = "0123456789abcdef"; |
dcb3450f | 122 | char *buffer = hexbuffer[3 & ++bufno], *buf = buffer; |
0fcfd160 LT |
123 | int i; |
124 | ||
125 | for (i = 0; i < 20; i++) { | |
126 | unsigned int val = *sha1++; | |
127 | *buf++ = hex[val >> 4]; | |
128 | *buf++ = hex[val & 0xf]; | |
129 | } | |
1e80e044 JS |
130 | *buf = '\0'; |
131 | ||
0fcfd160 LT |
132 | return buffer; |
133 | } | |
134 | ||
ace1534d JH |
135 | static void fill_sha1_path(char *pathbuf, const unsigned char *sha1) |
136 | { | |
137 | int i; | |
138 | for (i = 0; i < 20; i++) { | |
139 | static char hex[] = "0123456789abcdef"; | |
140 | unsigned int val = sha1[i]; | |
141 | char *pos = pathbuf + i*2 + (i > 0); | |
142 | *pos++ = hex[val >> 4]; | |
143 | *pos = hex[val & 0xf]; | |
144 | } | |
145 | } | |
146 | ||
0fcfd160 LT |
147 | /* |
148 | * NOTE! This returns a statically allocated buffer, so you have to be | |
9befac47 | 149 | * careful about using it. Do a "xstrdup()" if you need to save the |
0fcfd160 | 150 | * filename. |
ace1534d JH |
151 | * |
152 | * Also note that this returns the location for creating. Reading | |
153 | * SHA1 file can happen from any alternate directory listed in the | |
d19938ab | 154 | * DB_ENVIRONMENT environment variable if it is not found in |
ace1534d | 155 | * the primary object database. |
0fcfd160 LT |
156 | */ |
157 | char *sha1_file_name(const unsigned char *sha1) | |
158 | { | |
0fcfd160 LT |
159 | static char *name, *base; |
160 | ||
161 | if (!base) { | |
d19938ab | 162 | const char *sha1_file_directory = get_object_directory(); |
0fcfd160 | 163 | int len = strlen(sha1_file_directory); |
812666c8 | 164 | base = xmalloc(len + 60); |
0fcfd160 LT |
165 | memcpy(base, sha1_file_directory, len); |
166 | memset(base+len, 0, 60); | |
167 | base[len] = '/'; | |
168 | base[len+3] = '/'; | |
169 | name = base + len + 1; | |
170 | } | |
ace1534d | 171 | fill_sha1_path(name, sha1); |
0fcfd160 LT |
172 | return base; |
173 | } | |
174 | ||
bf592c50 DB |
175 | char *sha1_pack_name(const unsigned char *sha1) |
176 | { | |
177 | static const char hex[] = "0123456789abcdef"; | |
178 | static char *name, *base, *buf; | |
179 | int i; | |
180 | ||
181 | if (!base) { | |
182 | const char *sha1_file_directory = get_object_directory(); | |
183 | int len = strlen(sha1_file_directory); | |
184 | base = xmalloc(len + 60); | |
185 | sprintf(base, "%s/pack/pack-1234567890123456789012345678901234567890.pack", sha1_file_directory); | |
186 | name = base + len + 11; | |
187 | } | |
188 | ||
189 | buf = name; | |
190 | ||
191 | for (i = 0; i < 20; i++) { | |
192 | unsigned int val = *sha1++; | |
193 | *buf++ = hex[val >> 4]; | |
194 | *buf++ = hex[val & 0xf]; | |
195 | } | |
a6080a0a | 196 | |
bf592c50 DB |
197 | return base; |
198 | } | |
199 | ||
200 | char *sha1_pack_index_name(const unsigned char *sha1) | |
201 | { | |
202 | static const char hex[] = "0123456789abcdef"; | |
203 | static char *name, *base, *buf; | |
204 | int i; | |
205 | ||
206 | if (!base) { | |
207 | const char *sha1_file_directory = get_object_directory(); | |
208 | int len = strlen(sha1_file_directory); | |
209 | base = xmalloc(len + 60); | |
210 | sprintf(base, "%s/pack/pack-1234567890123456789012345678901234567890.idx", sha1_file_directory); | |
211 | name = base + len + 11; | |
212 | } | |
213 | ||
214 | buf = name; | |
215 | ||
216 | for (i = 0; i < 20; i++) { | |
217 | unsigned int val = *sha1++; | |
218 | *buf++ = hex[val >> 4]; | |
219 | *buf++ = hex[val & 0xf]; | |
220 | } | |
a6080a0a | 221 | |
bf592c50 DB |
222 | return base; |
223 | } | |
224 | ||
d5a63b99 JH |
225 | struct alternate_object_database *alt_odb_list; |
226 | static struct alternate_object_database **alt_odb_tail; | |
ace1534d | 227 | |
c2f493a4 MW |
228 | static void read_info_alternates(const char * alternates, int depth); |
229 | ||
ddd5d056 JH |
230 | /* |
231 | * Prepare alternate object database registry. | |
d5a63b99 JH |
232 | * |
233 | * The variable alt_odb_list points at the list of struct | |
234 | * alternate_object_database. The elements on this list come from | |
235 | * non-empty elements from colon separated ALTERNATE_DB_ENVIRONMENT | |
236 | * environment variable, and $GIT_OBJECT_DIRECTORY/info/alternates, | |
1494e038 JH |
237 | * whose contents is similar to that environment variable but can be |
238 | * LF separated. Its base points at a statically allocated buffer that | |
d5a63b99 JH |
239 | * contains "/the/directory/corresponding/to/.git/objects/...", while |
240 | * its name points just after the slash at the end of ".git/objects/" | |
241 | * in the example above, and has enough space to hold 40-byte hex | |
242 | * SHA1, an extra slash for the first level indirection, and the | |
243 | * terminating NUL. | |
ddd5d056 | 244 | */ |
c2f493a4 | 245 | static int link_alt_odb_entry(const char * entry, int len, const char * relative_base, int depth) |
ace1534d | 246 | { |
c2f493a4 | 247 | struct stat st; |
1494e038 | 248 | const char *objdir = get_object_directory(); |
c2f493a4 MW |
249 | struct alternate_object_database *ent; |
250 | struct alternate_object_database *alt; | |
251 | /* 43 = 40-byte + 2 '/' + terminating NUL */ | |
252 | int pfxlen = len; | |
253 | int entlen = pfxlen + 43; | |
ccfd3e99 | 254 | int base_len = -1; |
d5a63b99 | 255 | |
c2f493a4 MW |
256 | if (*entry != '/' && relative_base) { |
257 | /* Relative alt-odb */ | |
258 | if (base_len < 0) | |
259 | base_len = strlen(relative_base) + 1; | |
260 | entlen += base_len; | |
261 | pfxlen += base_len; | |
262 | } | |
263 | ent = xmalloc(sizeof(*ent) + entlen); | |
264 | ||
265 | if (*entry != '/' && relative_base) { | |
266 | memcpy(ent->base, relative_base, base_len - 1); | |
267 | ent->base[base_len - 1] = '/'; | |
268 | memcpy(ent->base + base_len, entry, len); | |
269 | } | |
270 | else | |
271 | memcpy(ent->base, entry, pfxlen); | |
272 | ||
273 | ent->name = ent->base + pfxlen + 1; | |
274 | ent->base[pfxlen + 3] = '/'; | |
275 | ent->base[pfxlen] = ent->base[entlen-1] = 0; | |
276 | ||
277 | /* Detect cases where alternate disappeared */ | |
278 | if (stat(ent->base, &st) || !S_ISDIR(st.st_mode)) { | |
279 | error("object directory %s does not exist; " | |
280 | "check .git/objects/info/alternates.", | |
281 | ent->base); | |
282 | free(ent); | |
283 | return -1; | |
284 | } | |
285 | ||
286 | /* Prevent the common mistake of listing the same | |
287 | * thing twice, or object directory itself. | |
288 | */ | |
289 | for (alt = alt_odb_list; alt; alt = alt->next) { | |
290 | if (!memcmp(ent->base, alt->base, pfxlen)) { | |
291 | free(ent); | |
292 | return -1; | |
293 | } | |
294 | } | |
295 | if (!memcmp(ent->base, objdir, pfxlen)) { | |
296 | free(ent); | |
297 | return -1; | |
298 | } | |
299 | ||
300 | /* add the alternate entry */ | |
301 | *alt_odb_tail = ent; | |
302 | alt_odb_tail = &(ent->next); | |
303 | ent->next = NULL; | |
304 | ||
305 | /* recursively add alternates */ | |
306 | read_info_alternates(ent->base, depth + 1); | |
307 | ||
308 | ent->base[pfxlen] = '/'; | |
309 | ||
310 | return 0; | |
311 | } | |
312 | ||
313 | static void link_alt_odb_entries(const char *alt, const char *ep, int sep, | |
314 | const char *relative_base, int depth) | |
315 | { | |
316 | const char *cp, *last; | |
317 | ||
318 | if (depth > 5) { | |
319 | error("%s: ignoring alternate object stores, nesting too deep.", | |
320 | relative_base); | |
321 | return; | |
322 | } | |
323 | ||
d5a63b99 | 324 | last = alt; |
9577e7e3 JH |
325 | while (last < ep) { |
326 | cp = last; | |
327 | if (cp < ep && *cp == '#') { | |
328 | while (cp < ep && *cp != sep) | |
329 | cp++; | |
330 | last = cp + 1; | |
331 | continue; | |
332 | } | |
c2f493a4 MW |
333 | while (cp < ep && *cp != sep) |
334 | cp++; | |
d5a63b99 | 335 | if (last != cp) { |
c2f493a4 MW |
336 | if ((*last != '/') && depth) { |
337 | error("%s: ignoring relative alternate object store %s", | |
338 | relative_base, last); | |
339 | } else { | |
340 | link_alt_odb_entry(last, cp - last, | |
341 | relative_base, depth); | |
1494e038 | 342 | } |
d5a63b99 | 343 | } |
9577e7e3 | 344 | while (cp < ep && *cp == sep) |
d5a63b99 JH |
345 | cp++; |
346 | last = cp; | |
9577e7e3 | 347 | } |
d5a63b99 JH |
348 | } |
349 | ||
c2f493a4 | 350 | static void read_info_alternates(const char * relative_base, int depth) |
d5a63b99 | 351 | { |
9577e7e3 | 352 | char *map; |
dc49cd76 | 353 | size_t mapsz; |
d5a63b99 | 354 | struct stat st; |
9cb18f56 JM |
355 | const char alt_file_name[] = "info/alternates"; |
356 | /* Given that relative_base is no longer than PATH_MAX, | |
357 | ensure that "path" has enough space to append "/", the | |
358 | file name, "info/alternates", and a trailing NUL. */ | |
359 | char path[PATH_MAX + 1 + sizeof alt_file_name]; | |
c2f493a4 | 360 | int fd; |
d5a63b99 | 361 | |
9cb18f56 | 362 | sprintf(path, "%s/%s", relative_base, alt_file_name); |
d5a63b99 JH |
363 | fd = open(path, O_RDONLY); |
364 | if (fd < 0) | |
365 | return; | |
366 | if (fstat(fd, &st) || (st.st_size == 0)) { | |
367 | close(fd); | |
9a217f2a | 368 | return; |
ace1534d | 369 | } |
dc49cd76 SP |
370 | mapsz = xsize_t(st.st_size); |
371 | map = xmmap(NULL, mapsz, PROT_READ, MAP_PRIVATE, fd, 0); | |
d5a63b99 | 372 | close(fd); |
d5a63b99 | 373 | |
dc49cd76 | 374 | link_alt_odb_entries(map, map + mapsz, '\n', relative_base, depth); |
c2f493a4 | 375 | |
dc49cd76 | 376 | munmap(map, mapsz); |
ace1534d JH |
377 | } |
378 | ||
c2f493a4 MW |
379 | void prepare_alt_odb(void) |
380 | { | |
554fe20d | 381 | const char *alt; |
c2f493a4 | 382 | |
7dc24aa5 SP |
383 | if (alt_odb_tail) |
384 | return; | |
385 | ||
c2f493a4 MW |
386 | alt = getenv(ALTERNATE_DB_ENVIRONMENT); |
387 | if (!alt) alt = ""; | |
388 | ||
c2f493a4 MW |
389 | alt_odb_tail = &alt_odb_list; |
390 | link_alt_odb_entries(alt, alt + strlen(alt), ':', NULL, 0); | |
391 | ||
392 | read_info_alternates(get_object_directory(), 0); | |
393 | } | |
394 | ||
ace1534d JH |
395 | static char *find_sha1_file(const unsigned char *sha1, struct stat *st) |
396 | { | |
ace1534d | 397 | char *name = sha1_file_name(sha1); |
d5a63b99 | 398 | struct alternate_object_database *alt; |
ace1534d JH |
399 | |
400 | if (!stat(name, st)) | |
401 | return name; | |
9a217f2a | 402 | prepare_alt_odb(); |
d5a63b99 JH |
403 | for (alt = alt_odb_list; alt; alt = alt->next) { |
404 | name = alt->name; | |
ace1534d | 405 | fill_sha1_path(name, sha1); |
d5a63b99 JH |
406 | if (!stat(alt->base, st)) |
407 | return alt->base; | |
ace1534d JH |
408 | } |
409 | return NULL; | |
410 | } | |
411 | ||
60bb8b14 | 412 | static unsigned int pack_used_ctr; |
a53128b6 SP |
413 | static unsigned int pack_mmap_calls; |
414 | static unsigned int peak_pack_open_windows; | |
415 | static unsigned int pack_open_windows; | |
416 | static size_t peak_pack_mapped; | |
60bb8b14 | 417 | static size_t pack_mapped; |
9a217f2a | 418 | struct packed_git *packed_git; |
1f688557 | 419 | |
b79d18c9 | 420 | void pack_report(void) |
a53128b6 SP |
421 | { |
422 | fprintf(stderr, | |
e05db0fd PR |
423 | "pack_report: getpagesize() = %10" SZ_FMT "\n" |
424 | "pack_report: core.packedGitWindowSize = %10" SZ_FMT "\n" | |
425 | "pack_report: core.packedGitLimit = %10" SZ_FMT "\n", | |
78a28df9 | 426 | (size_t) getpagesize(), |
a53128b6 SP |
427 | packed_git_window_size, |
428 | packed_git_limit); | |
429 | fprintf(stderr, | |
430 | "pack_report: pack_used_ctr = %10u\n" | |
431 | "pack_report: pack_mmap_calls = %10u\n" | |
432 | "pack_report: pack_open_windows = %10u / %10u\n" | |
e05db0fd PR |
433 | "pack_report: pack_mapped = " |
434 | "%10" SZ_FMT " / %10" SZ_FMT "\n", | |
a53128b6 SP |
435 | pack_used_ctr, |
436 | pack_mmap_calls, | |
437 | pack_open_windows, peak_pack_open_windows, | |
438 | pack_mapped, peak_pack_mapped); | |
439 | } | |
440 | ||
42873078 | 441 | static int check_packed_git_idx(const char *path, struct packed_git *p) |
1f688557 JH |
442 | { |
443 | void *idx_map; | |
42873078 | 444 | struct pack_idx_header *hdr; |
dc49cd76 | 445 | size_t idx_size; |
74e34e1f | 446 | uint32_t version, nr, i, *index; |
e93ec6f9 | 447 | int fd = open(path, O_RDONLY); |
1f688557 | 448 | struct stat st; |
42873078 | 449 | |
1f688557 JH |
450 | if (fd < 0) |
451 | return -1; | |
452 | if (fstat(fd, &st)) { | |
453 | close(fd); | |
454 | return -1; | |
455 | } | |
dc49cd76 | 456 | idx_size = xsize_t(st.st_size); |
2d88451b SP |
457 | if (idx_size < 4 * 256 + 20 + 20) { |
458 | close(fd); | |
459 | return error("index file %s is too small", path); | |
460 | } | |
c4712e45 | 461 | idx_map = xmmap(NULL, idx_size, PROT_READ, MAP_PRIVATE, fd, 0); |
1f688557 | 462 | close(fd); |
1f688557 | 463 | |
42873078 NP |
464 | hdr = idx_map; |
465 | if (hdr->idx_signature == htonl(PACK_IDX_SIGNATURE)) { | |
74e34e1f NP |
466 | version = ntohl(hdr->idx_version); |
467 | if (version < 2 || version > 2) { | |
468 | munmap(idx_map, idx_size); | |
469 | return error("index file %s is version %d" | |
470 | " and is not supported by this binary" | |
471 | " (try upgrading GIT to a newer version)", | |
472 | path, version); | |
473 | } | |
474 | } else | |
475 | version = 1; | |
df1b059d | 476 | |
1f688557 | 477 | nr = 0; |
42873078 | 478 | index = idx_map; |
74e34e1f NP |
479 | if (version > 1) |
480 | index += 2; /* skip index header */ | |
1f688557 | 481 | for (i = 0; i < 256; i++) { |
326bf396 | 482 | uint32_t n = ntohl(index[i]); |
2d88451b SP |
483 | if (n < nr) { |
484 | munmap(idx_map, idx_size); | |
df1b059d | 485 | return error("non-monotonic index %s", path); |
2d88451b | 486 | } |
1f688557 JH |
487 | nr = n; |
488 | } | |
489 | ||
74e34e1f NP |
490 | if (version == 1) { |
491 | /* | |
492 | * Total size: | |
493 | * - 256 index entries 4 bytes each | |
494 | * - 24-byte entries * nr (20-byte sha1 + 4-byte offset) | |
495 | * - 20-byte SHA1 of the packfile | |
496 | * - 20-byte SHA1 file checksum | |
497 | */ | |
498 | if (idx_size != 4*256 + nr * 24 + 20 + 20) { | |
499 | munmap(idx_map, idx_size); | |
eef427a0 | 500 | return error("wrong index v1 file size in %s", path); |
74e34e1f NP |
501 | } |
502 | } else if (version == 2) { | |
503 | /* | |
504 | * Minimum size: | |
505 | * - 8 bytes of header | |
506 | * - 256 index entries 4 bytes each | |
507 | * - 20-byte sha1 entry * nr | |
508 | * - 4-byte crc entry * nr | |
509 | * - 4-byte offset entry * nr | |
510 | * - 20-byte SHA1 of the packfile | |
511 | * - 20-byte SHA1 file checksum | |
512 | * And after the 4-byte offset table might be a | |
513 | * variable sized table containing 8-byte entries | |
514 | * for offsets larger than 2^31. | |
515 | */ | |
516 | unsigned long min_size = 8 + 4*256 + nr*(20 + 4 + 4) + 20 + 20; | |
1164f1e4 LT |
517 | unsigned long max_size = min_size; |
518 | if (nr) | |
519 | max_size += (nr - 1)*8; | |
520 | if (idx_size < min_size || idx_size > max_size) { | |
74e34e1f | 521 | munmap(idx_map, idx_size); |
eef427a0 | 522 | return error("wrong index v2 file size in %s", path); |
74e34e1f NP |
523 | } |
524 | if (idx_size != min_size) { | |
525 | /* make sure we can deal with large pack offsets */ | |
526 | off_t x = 0x7fffffffUL, y = 0xffffffffUL; | |
527 | if (x > (x + 1) || y > (y + 1)) { | |
528 | munmap(idx_map, idx_size); | |
529 | return error("pack too large for current definition of off_t in %s", path); | |
530 | } | |
531 | } | |
2d88451b | 532 | } |
1f688557 | 533 | |
74e34e1f | 534 | p->index_version = version; |
42873078 NP |
535 | p->index_data = idx_map; |
536 | p->index_size = idx_size; | |
57059091 | 537 | p->num_objects = nr; |
1f688557 JH |
538 | return 0; |
539 | } | |
540 | ||
bc8e478a | 541 | int open_pack_index(struct packed_git *p) |
d079837e SP |
542 | { |
543 | char *idx_name; | |
544 | int ret; | |
545 | ||
546 | if (p->index_data) | |
547 | return 0; | |
548 | ||
549 | idx_name = xstrdup(p->pack_name); | |
550 | strcpy(idx_name + strlen(idx_name) - strlen(".pack"), ".idx"); | |
551 | ret = check_packed_git_idx(idx_name, p); | |
552 | free(idx_name); | |
553 | return ret; | |
554 | } | |
555 | ||
11daf39b SP |
556 | static void scan_windows(struct packed_git *p, |
557 | struct packed_git **lru_p, | |
558 | struct pack_window **lru_w, | |
559 | struct pack_window **lru_l) | |
1f688557 | 560 | { |
11daf39b SP |
561 | struct pack_window *w, *w_l; |
562 | ||
563 | for (w_l = NULL, w = p->windows; w; w = w->next) { | |
564 | if (!w->inuse_cnt) { | |
565 | if (!*lru_w || w->last_used < (*lru_w)->last_used) { | |
566 | *lru_p = p; | |
567 | *lru_w = w; | |
568 | *lru_l = w_l; | |
54044bf8 | 569 | } |
54044bf8 | 570 | } |
11daf39b | 571 | w_l = w; |
f9253394 | 572 | } |
11daf39b SP |
573 | } |
574 | ||
d1efefa4 | 575 | static int unuse_one_window(struct packed_git *current, int keep_fd) |
11daf39b SP |
576 | { |
577 | struct packed_git *p, *lru_p = NULL; | |
578 | struct pack_window *lru_w = NULL, *lru_l = NULL; | |
579 | ||
580 | if (current) | |
581 | scan_windows(current, &lru_p, &lru_w, &lru_l); | |
582 | for (p = packed_git; p; p = p->next) | |
583 | scan_windows(p, &lru_p, &lru_w, &lru_l); | |
54044bf8 SP |
584 | if (lru_p) { |
585 | munmap(lru_w->base, lru_w->len); | |
586 | pack_mapped -= lru_w->len; | |
587 | if (lru_l) | |
588 | lru_l->next = lru_w->next; | |
589 | else { | |
590 | lru_p->windows = lru_w->next; | |
d1efefa4 | 591 | if (!lru_p->windows && lru_p->pack_fd != keep_fd) { |
54044bf8 SP |
592 | close(lru_p->pack_fd); |
593 | lru_p->pack_fd = -1; | |
594 | } | |
595 | } | |
596 | free(lru_w); | |
a53128b6 | 597 | pack_open_windows--; |
54044bf8 SP |
598 | return 1; |
599 | } | |
600 | return 0; | |
f9253394 JH |
601 | } |
602 | ||
d1efefa4 | 603 | void release_pack_memory(size_t need, int fd) |
97bfeb34 SP |
604 | { |
605 | size_t cur = pack_mapped; | |
d1efefa4 | 606 | while (need >= (cur - pack_mapped) && unuse_one_window(NULL, fd)) |
97bfeb34 SP |
607 | ; /* nothing */ |
608 | } | |
609 | ||
03e79c88 | 610 | void unuse_pack(struct pack_window **w_cursor) |
f9253394 | 611 | { |
03e79c88 SP |
612 | struct pack_window *w = *w_cursor; |
613 | if (w) { | |
614 | w->inuse_cnt--; | |
615 | *w_cursor = NULL; | |
616 | } | |
1f688557 JH |
617 | } |
618 | ||
3cf8b462 SP |
619 | /* |
620 | * Do not call this directly as this leaks p->pack_fd on error return; | |
621 | * call open_packed_git() instead. | |
622 | */ | |
623 | static int open_packed_git_1(struct packed_git *p) | |
1f688557 | 624 | { |
9bc879c1 SP |
625 | struct stat st; |
626 | struct pack_header hdr; | |
627 | unsigned char sha1[20]; | |
628 | unsigned char *idx_sha1; | |
2c039da8 | 629 | long fd_flag; |
9bc879c1 | 630 | |
d079837e SP |
631 | if (!p->index_data && open_pack_index(p)) |
632 | return error("packfile %s index unavailable", p->pack_name); | |
633 | ||
9bc879c1 SP |
634 | p->pack_fd = open(p->pack_name, O_RDONLY); |
635 | if (p->pack_fd < 0 || fstat(p->pack_fd, &st)) | |
072db278 | 636 | return -1; |
9bc879c1 SP |
637 | |
638 | /* If we created the struct before we had the pack we lack size. */ | |
bf592c50 | 639 | if (!p->pack_size) { |
bf592c50 | 640 | if (!S_ISREG(st.st_mode)) |
072db278 | 641 | return error("packfile %s not a regular file", p->pack_name); |
bf592c50 | 642 | p->pack_size = st.st_size; |
9bc879c1 | 643 | } else if (p->pack_size != st.st_size) |
072db278 | 644 | return error("packfile %s size changed", p->pack_name); |
9bc879c1 | 645 | |
2c039da8 JH |
646 | /* We leave these file descriptors open with sliding mmap; |
647 | * there is no point keeping them open across exec(), though. | |
648 | */ | |
649 | fd_flag = fcntl(p->pack_fd, F_GETFD, 0); | |
650 | if (fd_flag < 0) | |
072db278 | 651 | return error("cannot determine file descriptor flags"); |
2c039da8 JH |
652 | fd_flag |= FD_CLOEXEC; |
653 | if (fcntl(p->pack_fd, F_SETFD, fd_flag) == -1) | |
072db278 | 654 | return error("cannot set FD_CLOEXEC"); |
2c039da8 | 655 | |
9bc879c1 | 656 | /* Verify we recognize this pack file format. */ |
e6e2bd62 | 657 | if (read_in_full(p->pack_fd, &hdr, sizeof(hdr)) != sizeof(hdr)) |
072db278 | 658 | return error("file %s is far too short to be a packfile", p->pack_name); |
9bc879c1 | 659 | if (hdr.hdr_signature != htonl(PACK_SIGNATURE)) |
072db278 | 660 | return error("file %s is not a GIT packfile", p->pack_name); |
9bc879c1 | 661 | if (!pack_version_ok(hdr.hdr_version)) |
072db278 | 662 | return error("packfile %s is version %u and not supported" |
9bc879c1 SP |
663 | " (try upgrading GIT to a newer version)", |
664 | p->pack_name, ntohl(hdr.hdr_version)); | |
665 | ||
666 | /* Verify the pack matches its index. */ | |
57059091 | 667 | if (p->num_objects != ntohl(hdr.hdr_entries)) |
072db278 | 668 | return error("packfile %s claims to have %u objects" |
57059091 NP |
669 | " while index indicates %u objects", |
670 | p->pack_name, ntohl(hdr.hdr_entries), | |
671 | p->num_objects); | |
9bc879c1 | 672 | if (lseek(p->pack_fd, p->pack_size - sizeof(sha1), SEEK_SET) == -1) |
072db278 | 673 | return error("end of packfile %s is unavailable", p->pack_name); |
e6e2bd62 | 674 | if (read_in_full(p->pack_fd, sha1, sizeof(sha1)) != sizeof(sha1)) |
072db278 | 675 | return error("packfile %s signature is unavailable", p->pack_name); |
42873078 | 676 | idx_sha1 = ((unsigned char *)p->index_data) + p->index_size - 40; |
9bc879c1 | 677 | if (hashcmp(sha1, idx_sha1)) |
072db278 SP |
678 | return error("packfile %s does not match index", p->pack_name); |
679 | return 0; | |
9bc879c1 SP |
680 | } |
681 | ||
3cf8b462 SP |
682 | static int open_packed_git(struct packed_git *p) |
683 | { | |
684 | if (!open_packed_git_1(p)) | |
685 | return 0; | |
686 | if (p->pack_fd != -1) { | |
687 | close(p->pack_fd); | |
688 | p->pack_fd = -1; | |
689 | } | |
690 | return -1; | |
691 | } | |
692 | ||
c4001d92 | 693 | static int in_window(struct pack_window *win, off_t offset) |
60bb8b14 SP |
694 | { |
695 | /* We must promise at least 20 bytes (one hash) after the | |
696 | * offset is available from this window, otherwise the offset | |
697 | * is not actually in this window and a different window (which | |
698 | * has that one hash excess) must be used. This is to support | |
699 | * the object header and delta base parsing routines below. | |
700 | */ | |
701 | off_t win_off = win->offset; | |
702 | return win_off <= offset | |
703 | && (offset + 20) <= (win_off + win->len); | |
704 | } | |
705 | ||
03e79c88 SP |
706 | unsigned char* use_pack(struct packed_git *p, |
707 | struct pack_window **w_cursor, | |
c4001d92 | 708 | off_t offset, |
03e79c88 | 709 | unsigned int *left) |
9bc879c1 | 710 | { |
60bb8b14 | 711 | struct pack_window *win = *w_cursor; |
03e79c88 | 712 | |
072db278 SP |
713 | if (p->pack_fd == -1 && open_packed_git(p)) |
714 | die("packfile %s cannot be accessed", p->pack_name); | |
60bb8b14 SP |
715 | |
716 | /* Since packfiles end in a hash of their content and its | |
717 | * pointless to ask for an offset into the middle of that | |
718 | * hash, and the in_window function above wouldn't match | |
719 | * don't allow an offset too close to the end of the file. | |
720 | */ | |
721 | if (offset > (p->pack_size - 20)) | |
722 | die("offset beyond end of packfile (truncated pack?)"); | |
723 | ||
724 | if (!win || !in_window(win, offset)) { | |
725 | if (win) | |
726 | win->inuse_cnt--; | |
727 | for (win = p->windows; win; win = win->next) { | |
728 | if (in_window(win, offset)) | |
729 | break; | |
730 | } | |
731 | if (!win) { | |
78a28df9 | 732 | size_t window_align = packed_git_window_size / 2; |
dc49cd76 | 733 | off_t len; |
60bb8b14 | 734 | win = xcalloc(1, sizeof(*win)); |
78a28df9 | 735 | win->offset = (offset / window_align) * window_align; |
dc49cd76 SP |
736 | len = p->pack_size - win->offset; |
737 | if (len > packed_git_window_size) | |
738 | len = packed_git_window_size; | |
739 | win->len = (size_t)len; | |
60bb8b14 | 740 | pack_mapped += win->len; |
11daf39b | 741 | while (packed_git_limit < pack_mapped |
d1efefa4 | 742 | && unuse_one_window(p, p->pack_fd)) |
60bb8b14 | 743 | ; /* nothing */ |
c4712e45 | 744 | win->base = xmmap(NULL, win->len, |
60bb8b14 SP |
745 | PROT_READ, MAP_PRIVATE, |
746 | p->pack_fd, win->offset); | |
747 | if (win->base == MAP_FAILED) | |
73b4e4be SP |
748 | die("packfile %s cannot be mapped: %s", |
749 | p->pack_name, | |
750 | strerror(errno)); | |
a53128b6 SP |
751 | pack_mmap_calls++; |
752 | pack_open_windows++; | |
753 | if (pack_mapped > peak_pack_mapped) | |
754 | peak_pack_mapped = pack_mapped; | |
755 | if (pack_open_windows > peak_pack_open_windows) | |
756 | peak_pack_open_windows = pack_open_windows; | |
60bb8b14 SP |
757 | win->next = p->windows; |
758 | p->windows = win; | |
759 | } | |
1f688557 | 760 | } |
03e79c88 SP |
761 | if (win != *w_cursor) { |
762 | win->last_used = pack_used_ctr++; | |
763 | win->inuse_cnt++; | |
764 | *w_cursor = win; | |
765 | } | |
60bb8b14 | 766 | offset -= win->offset; |
03e79c88 | 767 | if (left) |
dc49cd76 | 768 | *left = win->len - xsize_t(offset); |
03e79c88 | 769 | return win->base + offset; |
1f688557 JH |
770 | } |
771 | ||
42873078 | 772 | struct packed_git *add_packed_git(const char *path, int path_len, int local) |
1f688557 JH |
773 | { |
774 | struct stat st; | |
42873078 | 775 | struct packed_git *p = xmalloc(sizeof(*p) + path_len + 2); |
1f688557 | 776 | |
42873078 NP |
777 | /* |
778 | * Make sure a corresponding .pack file exists and that | |
779 | * the index looks sane. | |
780 | */ | |
781 | path_len -= strlen(".idx"); | |
782 | if (path_len < 1) | |
1f688557 | 783 | return NULL; |
42873078 NP |
784 | memcpy(p->pack_name, path, path_len); |
785 | strcpy(p->pack_name + path_len, ".pack"); | |
d079837e | 786 | if (stat(p->pack_name, &st) || !S_ISREG(st.st_mode)) { |
42873078 | 787 | free(p); |
1f688557 JH |
788 | return NULL; |
789 | } | |
42873078 | 790 | |
1f688557 JH |
791 | /* ok, it looks sane as far as we can check without |
792 | * actually mapping the pack file. | |
793 | */ | |
d079837e SP |
794 | p->index_version = 0; |
795 | p->index_data = NULL; | |
796 | p->index_size = 0; | |
797 | p->num_objects = 0; | |
1f688557 | 798 | p->pack_size = st.st_size; |
1f688557 | 799 | p->next = NULL; |
c41ee586 | 800 | p->windows = NULL; |
9bc879c1 | 801 | p->pack_fd = -1; |
9d835df2 | 802 | p->pack_local = local; |
b867092f | 803 | p->mtime = st.st_mtime; |
42873078 NP |
804 | if (path_len < 40 || get_sha1_hex(path + path_len - 40, p->sha1)) |
805 | hashclr(p->sha1); | |
1f688557 JH |
806 | return p; |
807 | } | |
808 | ||
bf592c50 | 809 | struct packed_git *parse_pack_index(unsigned char *sha1) |
c508df5e DB |
810 | { |
811 | char *path = sha1_pack_index_name(sha1); | |
812 | return parse_pack_index_file(sha1, path); | |
813 | } | |
814 | ||
42873078 NP |
815 | struct packed_git *parse_pack_index_file(const unsigned char *sha1, |
816 | const char *idx_path) | |
bf592c50 | 817 | { |
42873078 NP |
818 | const char *path = sha1_pack_name(sha1); |
819 | struct packed_git *p = xmalloc(sizeof(*p) + strlen(path) + 2); | |
bf592c50 | 820 | |
42873078 NP |
821 | if (check_packed_git_idx(idx_path, p)) { |
822 | free(p); | |
bf592c50 | 823 | return NULL; |
42873078 | 824 | } |
bf592c50 | 825 | |
bf592c50 | 826 | strcpy(p->pack_name, path); |
bf592c50 | 827 | p->pack_size = 0; |
bf592c50 | 828 | p->next = NULL; |
c41ee586 | 829 | p->windows = NULL; |
9bc879c1 | 830 | p->pack_fd = -1; |
e702496e | 831 | hashcpy(p->sha1, sha1); |
bf592c50 DB |
832 | return p; |
833 | } | |
834 | ||
835 | void install_packed_git(struct packed_git *pack) | |
836 | { | |
837 | pack->next = packed_git; | |
838 | packed_git = pack; | |
839 | } | |
840 | ||
9d835df2 | 841 | static void prepare_packed_git_one(char *objdir, int local) |
1f688557 | 842 | { |
9cb18f56 JM |
843 | /* Ensure that this buffer is large enough so that we can |
844 | append "/pack/" without clobbering the stack even if | |
845 | strlen(objdir) were PATH_MAX. */ | |
846 | char path[PATH_MAX + 1 + 4 + 1 + 1]; | |
1f688557 JH |
847 | int len; |
848 | DIR *dir; | |
849 | struct dirent *de; | |
850 | ||
851 | sprintf(path, "%s/pack", objdir); | |
852 | len = strlen(path); | |
853 | dir = opendir(path); | |
b5b16990 | 854 | if (!dir) { |
26125f6b | 855 | if (errno != ENOENT) |
bd2afde8 | 856 | error("unable to open object pack directory: %s: %s", |
26125f6b | 857 | path, strerror(errno)); |
1f688557 | 858 | return; |
b5b16990 | 859 | } |
1f688557 JH |
860 | path[len++] = '/'; |
861 | while ((de = readdir(dir)) != NULL) { | |
862 | int namelen = strlen(de->d_name); | |
863 | struct packed_git *p; | |
864 | ||
5bb1cda5 | 865 | if (!has_extension(de->d_name, ".idx")) |
1f688557 JH |
866 | continue; |
867 | ||
9cb18f56 JM |
868 | if (len + namelen + 1 > sizeof(path)) |
869 | continue; | |
870 | ||
54a15a8d | 871 | /* Don't reopen a pack we already have. */ |
1f688557 | 872 | strcpy(path + len, de->d_name); |
86f7780c JK |
873 | for (p = packed_git; p; p = p->next) { |
874 | if (!memcmp(path, p->pack_name, len + namelen - 4)) | |
875 | break; | |
876 | } | |
877 | if (p) | |
878 | continue; | |
54a15a8d SP |
879 | /* See if it really is a valid .idx file with corresponding |
880 | * .pack file that we can map. | |
881 | */ | |
9d835df2 | 882 | p = add_packed_git(path, len + namelen, local); |
1f688557 JH |
883 | if (!p) |
884 | continue; | |
625e9421 | 885 | install_packed_git(p); |
1f688557 | 886 | } |
5b35bcd5 | 887 | closedir(dir); |
1f688557 JH |
888 | } |
889 | ||
b867092f JH |
890 | static int sort_pack(const void *a_, const void *b_) |
891 | { | |
892 | struct packed_git *a = *((struct packed_git **)a_); | |
893 | struct packed_git *b = *((struct packed_git **)b_); | |
894 | int st; | |
895 | ||
896 | /* | |
897 | * Local packs tend to contain objects specific to our | |
898 | * variant of the project than remote ones. In addition, | |
899 | * remote ones could be on a network mounted filesystem. | |
900 | * Favor local ones for these reasons. | |
901 | */ | |
902 | st = a->pack_local - b->pack_local; | |
903 | if (st) | |
904 | return -st; | |
905 | ||
906 | /* | |
907 | * Younger packs tend to contain more recent objects, | |
908 | * and more recent objects tend to get accessed more | |
909 | * often. | |
910 | */ | |
911 | if (a->mtime < b->mtime) | |
912 | return 1; | |
913 | else if (a->mtime == b->mtime) | |
914 | return 0; | |
915 | return -1; | |
916 | } | |
917 | ||
918 | static void rearrange_packed_git(void) | |
919 | { | |
920 | struct packed_git **ary, *p; | |
921 | int i, n; | |
922 | ||
923 | for (n = 0, p = packed_git; p; p = p->next) | |
924 | n++; | |
925 | if (n < 2) | |
926 | return; | |
927 | ||
928 | /* prepare an array of packed_git for easier sorting */ | |
929 | ary = xcalloc(n, sizeof(struct packed_git *)); | |
930 | for (n = 0, p = packed_git; p; p = p->next) | |
931 | ary[n++] = p; | |
932 | ||
933 | qsort(ary, n, sizeof(struct packed_git *), sort_pack); | |
934 | ||
935 | /* link them back again */ | |
936 | for (i = 0; i < n - 1; i++) | |
937 | ary[i]->next = ary[i + 1]; | |
938 | ary[n - 1]->next = NULL; | |
939 | packed_git = ary[0]; | |
940 | ||
941 | free(ary); | |
942 | } | |
943 | ||
637cdd9d | 944 | static int prepare_packed_git_run_once = 0; |
9a217f2a | 945 | void prepare_packed_git(void) |
1f688557 | 946 | { |
d5a63b99 | 947 | struct alternate_object_database *alt; |
1f688557 | 948 | |
637cdd9d | 949 | if (prepare_packed_git_run_once) |
1f688557 | 950 | return; |
9d835df2 | 951 | prepare_packed_git_one(get_object_directory(), 1); |
9a217f2a | 952 | prepare_alt_odb(); |
d5a63b99 | 953 | for (alt = alt_odb_list; alt; alt = alt->next) { |
1494e038 | 954 | alt->name[-1] = 0; |
9d835df2 | 955 | prepare_packed_git_one(alt->base, 0); |
1494e038 | 956 | alt->name[-1] = '/'; |
1f688557 | 957 | } |
b867092f | 958 | rearrange_packed_git(); |
637cdd9d JK |
959 | prepare_packed_git_run_once = 1; |
960 | } | |
961 | ||
fc04c412 | 962 | void reprepare_packed_git(void) |
637cdd9d JK |
963 | { |
964 | prepare_packed_git_run_once = 0; | |
965 | prepare_packed_git(); | |
1f688557 JH |
966 | } |
967 | ||
5d6ccf5c | 968 | int check_sha1_signature(const unsigned char *sha1, void *map, unsigned long size, const char *type) |
0fcfd160 LT |
969 | { |
970 | unsigned char real_sha1[20]; | |
7cfb5f36 | 971 | hash_sha1_file(map, size, type, real_sha1); |
a89fccd2 | 972 | return hashcmp(sha1, real_sha1) ? -1 : 0; |
0fcfd160 LT |
973 | } |
974 | ||
4175e9e3 | 975 | static void *map_sha1_file(const unsigned char *sha1, unsigned long *size) |
0fcfd160 | 976 | { |
0fcfd160 LT |
977 | struct stat st; |
978 | void *map; | |
144bde78 | 979 | int fd; |
ace1534d JH |
980 | char *filename = find_sha1_file(sha1, &st); |
981 | ||
982 | if (!filename) { | |
ace1534d JH |
983 | return NULL; |
984 | } | |
0fcfd160 | 985 | |
144bde78 | 986 | fd = open(filename, O_RDONLY | sha1_file_open_flag); |
0fcfd160 | 987 | if (fd < 0) { |
144bde78 LT |
988 | /* See if it works without O_NOATIME */ |
989 | switch (sha1_file_open_flag) { | |
990 | default: | |
991 | fd = open(filename, O_RDONLY); | |
992 | if (fd >= 0) | |
993 | break; | |
994 | /* Fallthrough */ | |
995 | case 0: | |
144bde78 LT |
996 | return NULL; |
997 | } | |
998 | ||
1f688557 JH |
999 | /* If it failed once, it will probably fail again. |
1000 | * Stop using O_NOATIME | |
1001 | */ | |
144bde78 | 1002 | sha1_file_open_flag = 0; |
0fcfd160 | 1003 | } |
dc49cd76 SP |
1004 | *size = xsize_t(st.st_size); |
1005 | map = xmmap(NULL, *size, PROT_READ, MAP_PRIVATE, fd, 0); | |
0fcfd160 | 1006 | close(fd); |
0fcfd160 LT |
1007 | return map; |
1008 | } | |
1009 | ||
726f852b | 1010 | static int legacy_loose_object(unsigned char *map) |
bb6b8e4f JH |
1011 | { |
1012 | unsigned int word; | |
1013 | ||
1014 | /* | |
1015 | * Is it a zlib-compressed buffer? If so, the first byte | |
1016 | * must be 0x78 (15-bit window size, deflated), and the | |
1017 | * first 16-bit word is evenly divisible by 31 | |
1018 | */ | |
1019 | word = (map[0] << 8) + map[1]; | |
1020 | if (map[0] == 0x78 && !(word % 31)) | |
1021 | return 1; | |
1022 | else | |
1023 | return 0; | |
1024 | } | |
1025 | ||
72518e9c | 1026 | unsigned long unpack_object_header_gently(const unsigned char *buf, unsigned long len, enum object_type *type, unsigned long *sizep) |
c4483576 | 1027 | { |
ad1ed5ee | 1028 | unsigned shift; |
93821bd9 | 1029 | unsigned char c; |
93821bd9 | 1030 | unsigned long size; |
ad1ed5ee JH |
1031 | unsigned long used = 0; |
1032 | ||
1033 | c = buf[used++]; | |
1034 | *type = (c >> 4) & 7; | |
1035 | size = c & 15; | |
1036 | shift = 4; | |
1037 | while (c & 0x80) { | |
1038 | if (len <= used) | |
1039 | return 0; | |
1040 | if (sizeof(long) * 8 <= shift) | |
1041 | return 0; | |
1042 | c = buf[used++]; | |
1043 | size += (c & 0x7f) << shift; | |
1044 | shift += 7; | |
1045 | } | |
1046 | *sizep = size; | |
1047 | return used; | |
1048 | } | |
1049 | ||
1050 | static int unpack_sha1_header(z_stream *stream, unsigned char *map, unsigned long mapsize, void *buffer, unsigned long bufsiz) | |
1051 | { | |
1052 | unsigned long size, used; | |
1053 | static const char valid_loose_object_type[8] = { | |
1054 | 0, /* OBJ_EXT */ | |
1055 | 1, 1, 1, 1, /* "commit", "tree", "blob", "tag" */ | |
1056 | 0, /* "delta" and others are invalid in a loose object */ | |
93821bd9 | 1057 | }; |
ad1ed5ee | 1058 | enum object_type type; |
93821bd9 | 1059 | |
c4483576 LT |
1060 | /* Get the data stream */ |
1061 | memset(stream, 0, sizeof(*stream)); | |
1062 | stream->next_in = map; | |
1063 | stream->avail_in = mapsize; | |
1064 | stream->next_out = buffer; | |
93821bd9 LT |
1065 | stream->avail_out = bufsiz; |
1066 | ||
bb6b8e4f | 1067 | if (legacy_loose_object(map)) { |
93821bd9 LT |
1068 | inflateInit(stream); |
1069 | return inflate(stream, 0); | |
1070 | } | |
1071 | ||
726f852b NP |
1072 | |
1073 | /* | |
1074 | * There used to be a second loose object header format which | |
1075 | * was meant to mimic the in-pack format, allowing for direct | |
1076 | * copy of the object data. This format turned up not to be | |
1077 | * really worth it and we don't write it any longer. But we | |
1078 | * can still read it. | |
1079 | */ | |
ad1ed5ee JH |
1080 | used = unpack_object_header_gently(map, mapsize, &type, &size); |
1081 | if (!used || !valid_loose_object_type[type]) | |
93821bd9 | 1082 | return -1; |
ad1ed5ee JH |
1083 | map += used; |
1084 | mapsize -= used; | |
93821bd9 LT |
1085 | |
1086 | /* Set up the stream for the rest.. */ | |
1087 | stream->next_in = map; | |
1088 | stream->avail_in = mapsize; | |
c4483576 | 1089 | inflateInit(stream); |
93821bd9 LT |
1090 | |
1091 | /* And generate the fake traditional header */ | |
ad1ed5ee | 1092 | stream->total_out = 1 + snprintf(buffer, bufsiz, "%s %lu", |
df843662 | 1093 | typename(type), size); |
93821bd9 | 1094 | return 0; |
c4483576 LT |
1095 | } |
1096 | ||
7efbff75 | 1097 | static void *unpack_sha1_rest(z_stream *stream, void *buffer, unsigned long size, const unsigned char *sha1) |
5180cacc LT |
1098 | { |
1099 | int bytes = strlen(buffer) + 1; | |
d565b341 | 1100 | unsigned char *buf = xmalloc(1+size); |
93821bd9 | 1101 | unsigned long n; |
7efbff75 | 1102 | int status = Z_OK; |
5180cacc | 1103 | |
93821bd9 LT |
1104 | n = stream->total_out - bytes; |
1105 | if (n > size) | |
1106 | n = size; | |
1107 | memcpy(buf, (char *) buffer + bytes, n); | |
1108 | bytes = n; | |
456cdf6e LT |
1109 | if (bytes <= size) { |
1110 | /* | |
1111 | * The above condition must be (bytes <= size), not | |
1112 | * (bytes < size). In other words, even though we | |
1113 | * expect no more output and set avail_out to zer0, | |
1114 | * the input zlib stream may have bytes that express | |
1115 | * "this concludes the stream", and we *do* want to | |
1116 | * eat that input. | |
1117 | * | |
1118 | * Otherwise we would not be able to test that we | |
1119 | * consumed all the input to reach the expected size; | |
1120 | * we also want to check that zlib tells us that all | |
1121 | * went well with status == Z_STREAM_END at the end. | |
1122 | */ | |
5180cacc LT |
1123 | stream->next_out = buf + bytes; |
1124 | stream->avail_out = size - bytes; | |
7efbff75 JH |
1125 | while (status == Z_OK) |
1126 | status = inflate(stream, Z_FINISH); | |
5180cacc LT |
1127 | } |
1128 | buf[size] = 0; | |
456cdf6e | 1129 | if (status == Z_STREAM_END && !stream->avail_in) { |
7efbff75 JH |
1130 | inflateEnd(stream); |
1131 | return buf; | |
1132 | } | |
1133 | ||
1134 | if (status < 0) | |
1135 | error("corrupt loose object '%s'", sha1_to_hex(sha1)); | |
1136 | else if (stream->avail_in) | |
1137 | error("garbage at end of loose object '%s'", | |
1138 | sha1_to_hex(sha1)); | |
1139 | free(buf); | |
1140 | return NULL; | |
5180cacc LT |
1141 | } |
1142 | ||
1143 | /* | |
1144 | * We used to just use "sscanf()", but that's actually way | |
1145 | * too permissive for what we want to check. So do an anal | |
1146 | * object header parse by hand. | |
1147 | */ | |
21666f1a | 1148 | static int parse_sha1_header(const char *hdr, unsigned long *sizep) |
5180cacc | 1149 | { |
21666f1a | 1150 | char type[10]; |
5180cacc LT |
1151 | int i; |
1152 | unsigned long size; | |
1153 | ||
1154 | /* | |
a6080a0a | 1155 | * The type can be at most ten bytes (including the |
5180cacc | 1156 | * terminating '\0' that we add), and is followed by |
21666f1a | 1157 | * a space. |
5180cacc | 1158 | */ |
21666f1a | 1159 | i = 0; |
5180cacc LT |
1160 | for (;;) { |
1161 | char c = *hdr++; | |
1162 | if (c == ' ') | |
1163 | break; | |
21666f1a NP |
1164 | type[i++] = c; |
1165 | if (i >= sizeof(type)) | |
5180cacc | 1166 | return -1; |
5180cacc | 1167 | } |
21666f1a | 1168 | type[i] = 0; |
5180cacc LT |
1169 | |
1170 | /* | |
1171 | * The length must follow immediately, and be in canonical | |
1172 | * decimal format (ie "010" is not valid). | |
1173 | */ | |
1174 | size = *hdr++ - '0'; | |
1175 | if (size > 9) | |
1176 | return -1; | |
1177 | if (size) { | |
1178 | for (;;) { | |
1179 | unsigned long c = *hdr - '0'; | |
1180 | if (c > 9) | |
1181 | break; | |
1182 | hdr++; | |
1183 | size = size * 10 + c; | |
1184 | } | |
1185 | } | |
1186 | *sizep = size; | |
1187 | ||
1188 | /* | |
1189 | * The length must be followed by a zero byte | |
1190 | */ | |
21666f1a | 1191 | return *hdr ? -1 : type_from_string(type); |
5180cacc LT |
1192 | } |
1193 | ||
7efbff75 | 1194 | static void *unpack_sha1_file(void *map, unsigned long mapsize, enum object_type *type, unsigned long *size, const unsigned char *sha1) |
0fcfd160 | 1195 | { |
5180cacc | 1196 | int ret; |
0fcfd160 | 1197 | z_stream stream; |
5180cacc | 1198 | char hdr[8192]; |
0fcfd160 | 1199 | |
5180cacc | 1200 | ret = unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)); |
21666f1a | 1201 | if (ret < Z_OK || (*type = parse_sha1_header(hdr, size)) < 0) |
0fcfd160 LT |
1202 | return NULL; |
1203 | ||
7efbff75 | 1204 | return unpack_sha1_rest(&stream, hdr, *size, sha1); |
0fcfd160 LT |
1205 | } |
1206 | ||
54dab52a NP |
1207 | unsigned long get_size_from_delta(struct packed_git *p, |
1208 | struct pack_window **w_curs, | |
1209 | off_t curpos) | |
1210 | { | |
1211 | const unsigned char *data; | |
1212 | unsigned char delta_head[20], *in; | |
1213 | z_stream stream; | |
1214 | int st; | |
1215 | ||
1216 | memset(&stream, 0, sizeof(stream)); | |
1217 | stream.next_out = delta_head; | |
1218 | stream.avail_out = sizeof(delta_head); | |
1219 | ||
1220 | inflateInit(&stream); | |
1221 | do { | |
1222 | in = use_pack(p, w_curs, curpos, &stream.avail_in); | |
1223 | stream.next_in = in; | |
1224 | st = inflate(&stream, Z_FINISH); | |
1225 | curpos += stream.next_in - in; | |
1226 | } while ((st == Z_OK || st == Z_BUF_ERROR) && | |
1227 | stream.total_out < sizeof(delta_head)); | |
1228 | inflateEnd(&stream); | |
1229 | if ((st != Z_STREAM_END) && stream.total_out != sizeof(delta_head)) | |
1230 | die("delta data unpack-initial failed"); | |
1231 | ||
1232 | /* Examine the initial part of the delta to figure out | |
1233 | * the result size. | |
1234 | */ | |
1235 | data = delta_head; | |
1236 | ||
1237 | /* ignore base size */ | |
1238 | get_delta_hdr_size(&data, delta_head+sizeof(delta_head)); | |
1239 | ||
1240 | /* Read the result size */ | |
1241 | return get_delta_hdr_size(&data, delta_head+sizeof(delta_head)); | |
1242 | } | |
1243 | ||
c4001d92 | 1244 | static off_t get_delta_base(struct packed_git *p, |
03e79c88 | 1245 | struct pack_window **w_curs, |
c4001d92 | 1246 | off_t *curpos, |
21666f1a | 1247 | enum object_type type, |
c4001d92 | 1248 | off_t delta_obj_offset) |
eb32d236 | 1249 | { |
2b87c45b | 1250 | unsigned char *base_info = use_pack(p, w_curs, *curpos, NULL); |
c4001d92 | 1251 | off_t base_offset; |
eb32d236 | 1252 | |
8d8a4ea5 SP |
1253 | /* use_pack() assured us we have [base_info, base_info + 20) |
1254 | * as a range that we can look at without walking off the | |
1255 | * end of the mapped window. Its actually the hash size | |
1256 | * that is assured. An OFS_DELTA longer than the hash size | |
1257 | * is stupid, as then a REF_DELTA would be smaller to store. | |
1258 | */ | |
21666f1a | 1259 | if (type == OBJ_OFS_DELTA) { |
eb32d236 NP |
1260 | unsigned used = 0; |
1261 | unsigned char c = base_info[used++]; | |
1262 | base_offset = c & 127; | |
1263 | while (c & 128) { | |
1264 | base_offset += 1; | |
8723f216 | 1265 | if (!base_offset || MSB(base_offset, 7)) |
eb32d236 NP |
1266 | die("offset value overflow for delta base object"); |
1267 | c = base_info[used++]; | |
1268 | base_offset = (base_offset << 7) + (c & 127); | |
1269 | } | |
1270 | base_offset = delta_obj_offset - base_offset; | |
1271 | if (base_offset >= delta_obj_offset) | |
1272 | die("delta base offset out of bound"); | |
2b87c45b | 1273 | *curpos += used; |
21666f1a | 1274 | } else if (type == OBJ_REF_DELTA) { |
eb32d236 NP |
1275 | /* The base entry _must_ be in the same pack */ |
1276 | base_offset = find_pack_entry_one(base_info, p); | |
1277 | if (!base_offset) | |
1278 | die("failed to find delta-pack base object %s", | |
1279 | sha1_to_hex(base_info)); | |
2b87c45b | 1280 | *curpos += 20; |
eb32d236 NP |
1281 | } else |
1282 | die("I am totally screwed"); | |
2b87c45b | 1283 | return base_offset; |
eb32d236 NP |
1284 | } |
1285 | ||
f3bf9224 | 1286 | /* forward declaration for a mutually recursive function */ |
c4001d92 | 1287 | static int packed_object_info(struct packed_git *p, off_t offset, |
21666f1a | 1288 | unsigned long *sizep); |
f3bf9224 | 1289 | |
43057304 | 1290 | static int packed_delta_info(struct packed_git *p, |
03e79c88 | 1291 | struct pack_window **w_curs, |
c4001d92 | 1292 | off_t curpos, |
21666f1a | 1293 | enum object_type type, |
c4001d92 | 1294 | off_t obj_offset, |
43057304 | 1295 | unsigned long *sizep) |
5db47c2b | 1296 | { |
c4001d92 | 1297 | off_t base_offset; |
f3bf9224 | 1298 | |
21666f1a NP |
1299 | base_offset = get_delta_base(p, w_curs, &curpos, type, obj_offset); |
1300 | type = packed_object_info(p, base_offset, NULL); | |
f3bf9224 | 1301 | |
c62266f3 JH |
1302 | /* We choose to only get the type of the base object and |
1303 | * ignore potentially corrupt pack file that expects the delta | |
1304 | * based on a base with a wrong size. This saves tons of | |
1305 | * inflate() calls. | |
1306 | */ | |
54dab52a NP |
1307 | if (sizep) |
1308 | *sizep = get_size_from_delta(p, w_curs, curpos); | |
21666f1a NP |
1309 | |
1310 | return type; | |
5db47c2b JH |
1311 | } |
1312 | ||
2b87c45b NP |
1313 | static int unpack_object_header(struct packed_git *p, |
1314 | struct pack_window **w_curs, | |
c4001d92 | 1315 | off_t *curpos, |
2b87c45b | 1316 | unsigned long *sizep) |
a733cb60 | 1317 | { |
03e79c88 SP |
1318 | unsigned char *base; |
1319 | unsigned int left; | |
ad1ed5ee | 1320 | unsigned long used; |
2b87c45b | 1321 | enum object_type type; |
a733cb60 | 1322 | |
8d8a4ea5 SP |
1323 | /* use_pack() assures us we have [base, base + 20) available |
1324 | * as a range that we can look at at. (Its actually the hash | |
3dff5379 | 1325 | * size that is assured.) With our object header encoding |
8d8a4ea5 SP |
1326 | * the maximum deflated object size is 2^137, which is just |
1327 | * insane, so we know won't exceed what we have been given. | |
1328 | */ | |
2b87c45b NP |
1329 | base = use_pack(p, w_curs, *curpos, &left); |
1330 | used = unpack_object_header_gently(base, left, &type, sizep); | |
ad1ed5ee JH |
1331 | if (!used) |
1332 | die("object offset outside of pack file"); | |
2b87c45b | 1333 | *curpos += used; |
ad1ed5ee | 1334 | |
2b87c45b | 1335 | return type; |
a733cb60 LT |
1336 | } |
1337 | ||
21666f1a | 1338 | const char *packed_object_info_detail(struct packed_git *p, |
c4001d92 | 1339 | off_t obj_offset, |
21666f1a NP |
1340 | unsigned long *size, |
1341 | unsigned long *store_size, | |
1342 | unsigned int *delta_chain_length, | |
1343 | unsigned char *base_sha1) | |
ad8c80a5 | 1344 | { |
03e79c88 | 1345 | struct pack_window *w_curs = NULL; |
c4001d92 SP |
1346 | off_t curpos; |
1347 | unsigned long dummy; | |
43057304 | 1348 | unsigned char *next_sha1; |
21666f1a | 1349 | enum object_type type; |
ad8c80a5 | 1350 | |
43057304 | 1351 | *delta_chain_length = 0; |
2b87c45b | 1352 | curpos = obj_offset; |
21666f1a | 1353 | type = unpack_object_header(p, &w_curs, &curpos, size); |
43057304 NP |
1354 | |
1355 | for (;;) { | |
21666f1a | 1356 | switch (type) { |
43057304 | 1357 | default: |
08a19d87 | 1358 | die("pack %s contains unknown object type %d", |
21666f1a | 1359 | p->pack_name, type); |
43057304 NP |
1360 | case OBJ_COMMIT: |
1361 | case OBJ_TREE: | |
1362 | case OBJ_BLOB: | |
1363 | case OBJ_TAG: | |
43057304 | 1364 | *store_size = 0; /* notyet */ |
03e79c88 | 1365 | unuse_pack(&w_curs); |
21666f1a | 1366 | return typename(type); |
eb32d236 | 1367 | case OBJ_OFS_DELTA: |
21666f1a | 1368 | obj_offset = get_delta_base(p, &w_curs, &curpos, type, obj_offset); |
eb32d236 | 1369 | if (*delta_chain_length == 0) { |
2b87c45b | 1370 | /* TODO: find base_sha1 as pointed by curpos */ |
30fee062 | 1371 | hashclr(base_sha1); |
eb32d236 NP |
1372 | } |
1373 | break; | |
1374 | case OBJ_REF_DELTA: | |
2b87c45b | 1375 | next_sha1 = use_pack(p, &w_curs, curpos, NULL); |
43057304 NP |
1376 | if (*delta_chain_length == 0) |
1377 | hashcpy(base_sha1, next_sha1); | |
2b87c45b | 1378 | obj_offset = find_pack_entry_one(next_sha1, p); |
43057304 NP |
1379 | break; |
1380 | } | |
43057304 | 1381 | (*delta_chain_length)++; |
2b87c45b | 1382 | curpos = obj_offset; |
21666f1a | 1383 | type = unpack_object_header(p, &w_curs, &curpos, &dummy); |
ad8c80a5 | 1384 | } |
ad8c80a5 JH |
1385 | } |
1386 | ||
c4001d92 | 1387 | static int packed_object_info(struct packed_git *p, off_t obj_offset, |
21666f1a | 1388 | unsigned long *sizep) |
1f688557 | 1389 | { |
03e79c88 | 1390 | struct pack_window *w_curs = NULL; |
c4001d92 SP |
1391 | unsigned long size; |
1392 | off_t curpos = obj_offset; | |
21666f1a | 1393 | enum object_type type; |
5db47c2b | 1394 | |
21666f1a | 1395 | type = unpack_object_header(p, &w_curs, &curpos, &size); |
5db47c2b | 1396 | |
21666f1a | 1397 | switch (type) { |
eb32d236 NP |
1398 | case OBJ_OFS_DELTA: |
1399 | case OBJ_REF_DELTA: | |
21666f1a NP |
1400 | type = packed_delta_info(p, &w_curs, curpos, |
1401 | type, obj_offset, sizep); | |
2b87c45b | 1402 | break; |
a733cb60 | 1403 | case OBJ_COMMIT: |
a733cb60 | 1404 | case OBJ_TREE: |
a733cb60 | 1405 | case OBJ_BLOB: |
a733cb60 | 1406 | case OBJ_TAG: |
2b87c45b NP |
1407 | if (sizep) |
1408 | *sizep = size; | |
a69d0943 | 1409 | break; |
1f688557 | 1410 | default: |
08a19d87 | 1411 | die("pack %s contains unknown object type %d", |
21666f1a | 1412 | p->pack_name, type); |
1f688557 | 1413 | } |
2b87c45b | 1414 | unuse_pack(&w_curs); |
21666f1a | 1415 | return type; |
1f688557 JH |
1416 | } |
1417 | ||
eb950c19 | 1418 | static void *unpack_compressed_entry(struct packed_git *p, |
03e79c88 | 1419 | struct pack_window **w_curs, |
c4001d92 | 1420 | off_t curpos, |
eb950c19 | 1421 | unsigned long size) |
de530aaa SP |
1422 | { |
1423 | int st; | |
1424 | z_stream stream; | |
079afb18 | 1425 | unsigned char *buffer, *in; |
de530aaa SP |
1426 | |
1427 | buffer = xmalloc(size + 1); | |
1428 | buffer[size] = 0; | |
1429 | memset(&stream, 0, sizeof(stream)); | |
de530aaa SP |
1430 | stream.next_out = buffer; |
1431 | stream.avail_out = size; | |
1432 | ||
1433 | inflateInit(&stream); | |
079afb18 | 1434 | do { |
2b87c45b | 1435 | in = use_pack(p, w_curs, curpos, &stream.avail_in); |
079afb18 SP |
1436 | stream.next_in = in; |
1437 | st = inflate(&stream, Z_FINISH); | |
2b87c45b | 1438 | curpos += stream.next_in - in; |
079afb18 | 1439 | } while (st == Z_OK || st == Z_BUF_ERROR); |
de530aaa SP |
1440 | inflateEnd(&stream); |
1441 | if ((st != Z_STREAM_END) || stream.total_out != size) { | |
1442 | free(buffer); | |
1443 | return NULL; | |
1444 | } | |
1445 | ||
1446 | return buffer; | |
1447 | } | |
1448 | ||
e5e01619 LT |
1449 | #define MAX_DELTA_CACHE (256) |
1450 | ||
18bdec11 | 1451 | static size_t delta_base_cached; |
5e08ecbf NP |
1452 | |
1453 | static struct delta_base_cache_lru_list { | |
1454 | struct delta_base_cache_lru_list *prev; | |
1455 | struct delta_base_cache_lru_list *next; | |
1456 | } delta_base_cache_lru = { &delta_base_cache_lru, &delta_base_cache_lru }; | |
1457 | ||
e5e01619 | 1458 | static struct delta_base_cache_entry { |
5e08ecbf NP |
1459 | struct delta_base_cache_lru_list lru; |
1460 | void *data; | |
e5e01619 LT |
1461 | struct packed_git *p; |
1462 | off_t base_offset; | |
1463 | unsigned long size; | |
e5e01619 LT |
1464 | enum object_type type; |
1465 | } delta_base_cache[MAX_DELTA_CACHE]; | |
1466 | ||
1467 | static unsigned long pack_entry_hash(struct packed_git *p, off_t base_offset) | |
1468 | { | |
1469 | unsigned long hash; | |
1470 | ||
1471 | hash = (unsigned long)p + (unsigned long)base_offset; | |
1472 | hash += (hash >> 8) + (hash >> 16); | |
3358004a | 1473 | return hash % MAX_DELTA_CACHE; |
e5e01619 LT |
1474 | } |
1475 | ||
62f255ad | 1476 | static void *cache_or_unpack_entry(struct packed_git *p, off_t base_offset, |
a0cba108 | 1477 | unsigned long *base_size, enum object_type *type, int keep_cache) |
62f255ad | 1478 | { |
e5e01619 LT |
1479 | void *ret; |
1480 | unsigned long hash = pack_entry_hash(p, base_offset); | |
1481 | struct delta_base_cache_entry *ent = delta_base_cache + hash; | |
1482 | ||
1483 | ret = ent->data; | |
1484 | if (ret && ent->p == p && ent->base_offset == base_offset) | |
1485 | goto found_cache_entry; | |
62f255ad | 1486 | return unpack_entry(p, base_offset, type, base_size); |
e5e01619 LT |
1487 | |
1488 | found_cache_entry: | |
18bdec11 | 1489 | if (!keep_cache) { |
a0cba108 | 1490 | ent->data = NULL; |
5e08ecbf NP |
1491 | ent->lru.next->prev = ent->lru.prev; |
1492 | ent->lru.prev->next = ent->lru.next; | |
18bdec11 SP |
1493 | delta_base_cached -= ent->size; |
1494 | } | |
a0cba108 NP |
1495 | else { |
1496 | ret = xmalloc(ent->size + 1); | |
1497 | memcpy(ret, ent->data, ent->size); | |
1498 | ((char *)ret)[ent->size] = 0; | |
1499 | } | |
e5e01619 LT |
1500 | *type = ent->type; |
1501 | *base_size = ent->size; | |
1502 | return ret; | |
62f255ad LT |
1503 | } |
1504 | ||
18bdec11 SP |
1505 | static inline void release_delta_base_cache(struct delta_base_cache_entry *ent) |
1506 | { | |
1507 | if (ent->data) { | |
1508 | free(ent->data); | |
1509 | ent->data = NULL; | |
5e08ecbf NP |
1510 | ent->lru.next->prev = ent->lru.prev; |
1511 | ent->lru.prev->next = ent->lru.next; | |
18bdec11 SP |
1512 | delta_base_cached -= ent->size; |
1513 | } | |
1514 | } | |
1515 | ||
62f255ad LT |
1516 | static void add_delta_base_cache(struct packed_git *p, off_t base_offset, |
1517 | void *base, unsigned long base_size, enum object_type type) | |
1518 | { | |
5e08ecbf | 1519 | unsigned long hash = pack_entry_hash(p, base_offset); |
e5e01619 | 1520 | struct delta_base_cache_entry *ent = delta_base_cache + hash; |
5e08ecbf | 1521 | struct delta_base_cache_lru_list *lru; |
e5e01619 | 1522 | |
18bdec11 SP |
1523 | release_delta_base_cache(ent); |
1524 | delta_base_cached += base_size; | |
5e08ecbf NP |
1525 | |
1526 | for (lru = delta_base_cache_lru.next; | |
1527 | delta_base_cached > delta_base_cache_limit | |
1528 | && lru != &delta_base_cache_lru; | |
1529 | lru = lru->next) { | |
1530 | struct delta_base_cache_entry *f = (void *)lru; | |
18bdec11 SP |
1531 | if (f->type == OBJ_BLOB) |
1532 | release_delta_base_cache(f); | |
1533 | } | |
5e08ecbf NP |
1534 | for (lru = delta_base_cache_lru.next; |
1535 | delta_base_cached > delta_base_cache_limit | |
1536 | && lru != &delta_base_cache_lru; | |
1537 | lru = lru->next) { | |
1538 | struct delta_base_cache_entry *f = (void *)lru; | |
1539 | release_delta_base_cache(f); | |
1540 | } | |
18bdec11 | 1541 | |
e5e01619 LT |
1542 | ent->p = p; |
1543 | ent->base_offset = base_offset; | |
1544 | ent->type = type; | |
1545 | ent->data = base; | |
1546 | ent->size = base_size; | |
5e08ecbf NP |
1547 | ent->lru.next = &delta_base_cache_lru; |
1548 | ent->lru.prev = delta_base_cache_lru.prev; | |
1549 | delta_base_cache_lru.prev->next = &ent->lru; | |
1550 | delta_base_cache_lru.prev = &ent->lru; | |
62f255ad LT |
1551 | } |
1552 | ||
eb950c19 | 1553 | static void *unpack_delta_entry(struct packed_git *p, |
03e79c88 | 1554 | struct pack_window **w_curs, |
c4001d92 | 1555 | off_t curpos, |
1f688557 | 1556 | unsigned long delta_size, |
c4001d92 | 1557 | off_t obj_offset, |
21666f1a | 1558 | enum object_type *type, |
eb950c19 | 1559 | unsigned long *sizep) |
1f688557 | 1560 | { |
7c3e8be3 | 1561 | void *delta_data, *result, *base; |
c4001d92 SP |
1562 | unsigned long base_size; |
1563 | off_t base_offset; | |
43057304 | 1564 | |
21666f1a | 1565 | base_offset = get_delta_base(p, w_curs, &curpos, *type, obj_offset); |
a0cba108 | 1566 | base = cache_or_unpack_entry(p, base_offset, &base_size, type, 0); |
67686d95 | 1567 | if (!base) |
c4001d92 SP |
1568 | die("failed to read delta base object" |
1569 | " at %"PRIuMAX" from %s", | |
1570 | (uintmax_t)base_offset, p->pack_name); | |
67686d95 | 1571 | |
2b87c45b | 1572 | delta_data = unpack_compressed_entry(p, w_curs, curpos, delta_size); |
1f688557 JH |
1573 | result = patch_delta(base, base_size, |
1574 | delta_data, delta_size, | |
21666f1a | 1575 | sizep); |
1f688557 JH |
1576 | if (!result) |
1577 | die("failed to apply delta"); | |
1578 | free(delta_data); | |
62f255ad | 1579 | add_delta_base_cache(p, base_offset, base, base_size, *type); |
1f688557 JH |
1580 | return result; |
1581 | } | |
1582 | ||
c4001d92 | 1583 | void *unpack_entry(struct packed_git *p, off_t obj_offset, |
21666f1a | 1584 | enum object_type *type, unsigned long *sizep) |
1f688557 | 1585 | { |
03e79c88 | 1586 | struct pack_window *w_curs = NULL; |
c4001d92 | 1587 | off_t curpos = obj_offset; |
21666f1a | 1588 | void *data; |
1f688557 | 1589 | |
21666f1a NP |
1590 | *type = unpack_object_header(p, &w_curs, &curpos, sizep); |
1591 | switch (*type) { | |
eb32d236 NP |
1592 | case OBJ_OFS_DELTA: |
1593 | case OBJ_REF_DELTA: | |
21666f1a NP |
1594 | data = unpack_delta_entry(p, &w_curs, curpos, *sizep, |
1595 | obj_offset, type, sizep); | |
4d703a1a | 1596 | break; |
a733cb60 | 1597 | case OBJ_COMMIT: |
a733cb60 | 1598 | case OBJ_TREE: |
a733cb60 | 1599 | case OBJ_BLOB: |
a733cb60 | 1600 | case OBJ_TAG: |
21666f1a | 1601 | data = unpack_compressed_entry(p, &w_curs, curpos, *sizep); |
4d703a1a | 1602 | break; |
1f688557 | 1603 | default: |
21666f1a | 1604 | die("unknown object type %i in %s", *type, p->pack_name); |
1f688557 | 1605 | } |
03e79c88 | 1606 | unuse_pack(&w_curs); |
21666f1a | 1607 | return data; |
1f688557 JH |
1608 | } |
1609 | ||
d079837e | 1610 | const unsigned char *nth_packed_object_sha1(struct packed_git *p, |
d72308e0 | 1611 | uint32_t n) |
9a217f2a | 1612 | { |
42873078 | 1613 | const unsigned char *index = p->index_data; |
d079837e SP |
1614 | if (!index) { |
1615 | if (open_pack_index(p)) | |
1616 | return NULL; | |
1617 | index = p->index_data; | |
1618 | } | |
57059091 | 1619 | if (n >= p->num_objects) |
d72308e0 | 1620 | return NULL; |
74e34e1f NP |
1621 | index += 4 * 256; |
1622 | if (p->index_version == 1) { | |
1623 | return index + 24 * n + 4; | |
1624 | } else { | |
1625 | index += 8; | |
1626 | return index + 20 * n; | |
1627 | } | |
1628 | } | |
1629 | ||
1630 | static off_t nth_packed_object_offset(const struct packed_git *p, uint32_t n) | |
1631 | { | |
1632 | const unsigned char *index = p->index_data; | |
1633 | index += 4 * 256; | |
1634 | if (p->index_version == 1) { | |
1635 | return ntohl(*((uint32_t *)(index + 24 * n))); | |
1636 | } else { | |
1637 | uint32_t off; | |
1638 | index += 8 + p->num_objects * (20 + 4); | |
1639 | off = ntohl(*((uint32_t *)(index + 4 * n))); | |
1640 | if (!(off & 0x80000000)) | |
1641 | return off; | |
1642 | index += p->num_objects * 4 + (off & 0x7fffffff) * 8; | |
1643 | return (((uint64_t)ntohl(*((uint32_t *)(index + 0)))) << 32) | | |
1644 | ntohl(*((uint32_t *)(index + 4))); | |
1645 | } | |
9a217f2a JH |
1646 | } |
1647 | ||
c4001d92 | 1648 | off_t find_pack_entry_one(const unsigned char *sha1, |
43057304 | 1649 | struct packed_git *p) |
1f688557 | 1650 | { |
42873078 | 1651 | const uint32_t *level1_ofs = p->index_data; |
42873078 | 1652 | const unsigned char *index = p->index_data; |
74e34e1f | 1653 | unsigned hi, lo; |
42873078 | 1654 | |
d079837e SP |
1655 | if (!index) { |
1656 | if (open_pack_index(p)) | |
1657 | return 0; | |
1658 | level1_ofs = p->index_data; | |
1659 | index = p->index_data; | |
1660 | } | |
74e34e1f NP |
1661 | if (p->index_version > 1) { |
1662 | level1_ofs += 2; | |
1663 | index += 8; | |
1664 | } | |
42873078 | 1665 | index += 4 * 256; |
74e34e1f NP |
1666 | hi = ntohl(level1_ofs[*sha1]); |
1667 | lo = ((*sha1 == 0x0) ? 0 : ntohl(level1_ofs[*sha1 - 1])); | |
1f688557 JH |
1668 | |
1669 | do { | |
74e34e1f NP |
1670 | unsigned mi = (lo + hi) / 2; |
1671 | unsigned x = (p->index_version > 1) ? (mi * 20) : (mi * 24 + 4); | |
1672 | int cmp = hashcmp(index + x, sha1); | |
43057304 | 1673 | if (!cmp) |
74e34e1f | 1674 | return nth_packed_object_offset(p, mi); |
1f688557 JH |
1675 | if (cmp > 0) |
1676 | hi = mi; | |
1677 | else | |
1678 | lo = mi+1; | |
1679 | } while (lo < hi); | |
1680 | return 0; | |
1681 | } | |
1682 | ||
d4ff6d92 SP |
1683 | static int matches_pack_name(struct packed_git *p, const char *ig) |
1684 | { | |
1685 | const char *last_c, *c; | |
1686 | ||
1687 | if (!strcmp(p->pack_name, ig)) | |
1688 | return 0; | |
1689 | ||
1690 | for (c = p->pack_name, last_c = c; *c;) | |
1691 | if (*c == '/') | |
1692 | last_c = ++c; | |
1693 | else | |
1694 | ++c; | |
1695 | if (!strcmp(last_c, ig)) | |
1696 | return 0; | |
1697 | ||
1698 | return 1; | |
1699 | } | |
1700 | ||
106d710b | 1701 | static int find_pack_entry(const unsigned char *sha1, struct pack_entry *e, const char **ignore_packed) |
1f688557 | 1702 | { |
f7c22cc6 | 1703 | static struct packed_git *last_found = (void *)1; |
1f688557 | 1704 | struct packed_git *p; |
c4001d92 | 1705 | off_t offset; |
43057304 | 1706 | |
1f688557 | 1707 | prepare_packed_git(); |
f7c22cc6 NP |
1708 | if (!packed_git) |
1709 | return 0; | |
1710 | p = (last_found == (void *)1) ? packed_git : last_found; | |
1f688557 | 1711 | |
f7c22cc6 | 1712 | do { |
106d710b JH |
1713 | if (ignore_packed) { |
1714 | const char **ig; | |
1715 | for (ig = ignore_packed; *ig; ig++) | |
d4ff6d92 | 1716 | if (!matches_pack_name(p, *ig)) |
106d710b JH |
1717 | break; |
1718 | if (*ig) | |
f7c22cc6 | 1719 | goto next; |
106d710b | 1720 | } |
f7c22cc6 | 1721 | |
43057304 NP |
1722 | offset = find_pack_entry_one(sha1, p); |
1723 | if (offset) { | |
c715f783 SP |
1724 | /* |
1725 | * We are about to tell the caller where they can | |
1726 | * locate the requested object. We better make | |
1727 | * sure the packfile is still here and can be | |
1728 | * accessed before supplying that answer, as | |
1729 | * it may have been deleted since the index | |
1730 | * was loaded! | |
1731 | */ | |
1732 | if (p->pack_fd == -1 && open_packed_git(p)) { | |
1733 | error("packfile %s cannot be accessed", p->pack_name); | |
f7c22cc6 | 1734 | goto next; |
c715f783 | 1735 | } |
43057304 NP |
1736 | e->offset = offset; |
1737 | e->p = p; | |
1738 | hashcpy(e->sha1, sha1); | |
f7c22cc6 | 1739 | last_found = p; |
1f688557 | 1740 | return 1; |
43057304 | 1741 | } |
f7c22cc6 NP |
1742 | |
1743 | next: | |
1744 | if (p == last_found) | |
1745 | p = packed_git; | |
1746 | else | |
1747 | p = p->next; | |
1748 | if (p == last_found) | |
1749 | p = p->next; | |
1750 | } while (p); | |
1f688557 JH |
1751 | return 0; |
1752 | } | |
1753 | ||
a6080a0a | 1754 | struct packed_git *find_sha1_pack(const unsigned char *sha1, |
bf592c50 DB |
1755 | struct packed_git *packs) |
1756 | { | |
1757 | struct packed_git *p; | |
bf592c50 DB |
1758 | |
1759 | for (p = packs; p; p = p->next) { | |
43057304 | 1760 | if (find_pack_entry_one(sha1, p)) |
bf592c50 DB |
1761 | return p; |
1762 | } | |
1763 | return NULL; | |
21666f1a | 1764 | |
bf592c50 DB |
1765 | } |
1766 | ||
21666f1a | 1767 | static int sha1_loose_object_info(const unsigned char *sha1, unsigned long *sizep) |
65c2e0c3 | 1768 | { |
36e4d74a | 1769 | int status; |
65c2e0c3 JH |
1770 | unsigned long mapsize, size; |
1771 | void *map; | |
1772 | z_stream stream; | |
d65a16f6 | 1773 | char hdr[32]; |
65c2e0c3 | 1774 | |
bb6b8e4f | 1775 | map = map_sha1_file(sha1, &mapsize); |
f0df4ed5 JS |
1776 | if (!map) |
1777 | return error("unable to find %s", sha1_to_hex(sha1)); | |
36e4d74a JH |
1778 | if (unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)) < 0) |
1779 | status = error("unable to unpack %s header", | |
1780 | sha1_to_hex(sha1)); | |
21666f1a | 1781 | else if ((status = parse_sha1_header(hdr, &size)) < 0) |
36e4d74a | 1782 | status = error("unable to parse %s header", sha1_to_hex(sha1)); |
21666f1a NP |
1783 | else if (sizep) |
1784 | *sizep = size; | |
65c2e0c3 JH |
1785 | inflateEnd(&stream); |
1786 | munmap(map, mapsize); | |
1787 | return status; | |
1788 | } | |
1789 | ||
21666f1a | 1790 | int sha1_object_info(const unsigned char *sha1, unsigned long *sizep) |
f0df4ed5 | 1791 | { |
f0df4ed5 JS |
1792 | struct pack_entry e; |
1793 | ||
1794 | if (!find_pack_entry(sha1, &e, NULL)) { | |
1795 | reprepare_packed_git(); | |
1796 | if (!find_pack_entry(sha1, &e, NULL)) | |
21666f1a | 1797 | return sha1_loose_object_info(sha1, sizep); |
f0df4ed5 | 1798 | } |
21666f1a | 1799 | return packed_object_info(e.p, e.offset, sizep); |
f0df4ed5 JS |
1800 | } |
1801 | ||
21666f1a NP |
1802 | static void *read_packed_sha1(const unsigned char *sha1, |
1803 | enum object_type *type, unsigned long *size) | |
1f688557 JH |
1804 | { |
1805 | struct pack_entry e; | |
1806 | ||
8276c007 | 1807 | if (!find_pack_entry(sha1, &e, NULL)) |
1f688557 | 1808 | return NULL; |
8276c007 | 1809 | else |
a0cba108 | 1810 | return cache_or_unpack_entry(e.p, e.offset, size, type, 1); |
1f688557 JH |
1811 | } |
1812 | ||
d66b37bb JH |
1813 | /* |
1814 | * This is meant to hold a *small* number of objects that you would | |
1815 | * want read_sha1_file() to be able to return, but yet you do not want | |
1816 | * to write them into the object store (e.g. a browse-only | |
1817 | * application). | |
1818 | */ | |
1819 | static struct cached_object { | |
1820 | unsigned char sha1[20]; | |
21666f1a | 1821 | enum object_type type; |
d66b37bb JH |
1822 | void *buf; |
1823 | unsigned long size; | |
1824 | } *cached_objects; | |
1825 | static int cached_object_nr, cached_object_alloc; | |
1826 | ||
1827 | static struct cached_object *find_cached_object(const unsigned char *sha1) | |
1828 | { | |
1829 | int i; | |
1830 | struct cached_object *co = cached_objects; | |
1831 | ||
1832 | for (i = 0; i < cached_object_nr; i++, co++) { | |
1833 | if (!hashcmp(co->sha1, sha1)) | |
1834 | return co; | |
1835 | } | |
1836 | return NULL; | |
1837 | } | |
1838 | ||
21666f1a NP |
1839 | int pretend_sha1_file(void *buf, unsigned long len, enum object_type type, |
1840 | unsigned char *sha1) | |
d66b37bb JH |
1841 | { |
1842 | struct cached_object *co; | |
1843 | ||
21666f1a | 1844 | hash_sha1_file(buf, len, typename(type), sha1); |
d66b37bb JH |
1845 | if (has_sha1_file(sha1) || find_cached_object(sha1)) |
1846 | return 0; | |
1847 | if (cached_object_alloc <= cached_object_nr) { | |
1848 | cached_object_alloc = alloc_nr(cached_object_alloc); | |
1849 | cached_objects = xrealloc(cached_objects, | |
1850 | sizeof(*cached_objects) * | |
1851 | cached_object_alloc); | |
1852 | } | |
1853 | co = &cached_objects[cached_object_nr++]; | |
1854 | co->size = len; | |
21666f1a | 1855 | co->type = type; |
efa13f7b JH |
1856 | co->buf = xmalloc(len); |
1857 | memcpy(co->buf, buf, len); | |
d66b37bb JH |
1858 | hashcpy(co->sha1, sha1); |
1859 | return 0; | |
1860 | } | |
1861 | ||
21666f1a NP |
1862 | void *read_sha1_file(const unsigned char *sha1, enum object_type *type, |
1863 | unsigned long *size) | |
0fcfd160 LT |
1864 | { |
1865 | unsigned long mapsize; | |
1866 | void *map, *buf; | |
d66b37bb JH |
1867 | struct cached_object *co; |
1868 | ||
1869 | co = find_cached_object(sha1); | |
1870 | if (co) { | |
1871 | buf = xmalloc(co->size + 1); | |
1872 | memcpy(buf, co->buf, co->size); | |
1873 | ((char*)buf)[co->size] = 0; | |
21666f1a | 1874 | *type = co->type; |
d66b37bb JH |
1875 | *size = co->size; |
1876 | return buf; | |
1877 | } | |
0fcfd160 | 1878 | |
8276c007 PE |
1879 | buf = read_packed_sha1(sha1, type, size); |
1880 | if (buf) | |
1881 | return buf; | |
bb6b8e4f | 1882 | map = map_sha1_file(sha1, &mapsize); |
0fcfd160 | 1883 | if (map) { |
7efbff75 | 1884 | buf = unpack_sha1_file(map, mapsize, type, size, sha1); |
0fcfd160 LT |
1885 | munmap(map, mapsize); |
1886 | return buf; | |
1887 | } | |
637cdd9d | 1888 | reprepare_packed_git(); |
8276c007 | 1889 | return read_packed_sha1(sha1, type, size); |
0fcfd160 LT |
1890 | } |
1891 | ||
40469ee9 | 1892 | void *read_object_with_reference(const unsigned char *sha1, |
21666f1a | 1893 | const char *required_type_name, |
40469ee9 JH |
1894 | unsigned long *size, |
1895 | unsigned char *actual_sha1_return) | |
f4913f91 | 1896 | { |
21666f1a | 1897 | enum object_type type, required_type; |
f4913f91 JH |
1898 | void *buffer; |
1899 | unsigned long isize; | |
40469ee9 | 1900 | unsigned char actual_sha1[20]; |
f4913f91 | 1901 | |
21666f1a | 1902 | required_type = type_from_string(required_type_name); |
e702496e | 1903 | hashcpy(actual_sha1, sha1); |
40469ee9 JH |
1904 | while (1) { |
1905 | int ref_length = -1; | |
1906 | const char *ref_type = NULL; | |
f4913f91 | 1907 | |
21666f1a | 1908 | buffer = read_sha1_file(actual_sha1, &type, &isize); |
40469ee9 JH |
1909 | if (!buffer) |
1910 | return NULL; | |
21666f1a | 1911 | if (type == required_type) { |
40469ee9 JH |
1912 | *size = isize; |
1913 | if (actual_sha1_return) | |
e702496e | 1914 | hashcpy(actual_sha1_return, actual_sha1); |
40469ee9 JH |
1915 | return buffer; |
1916 | } | |
1917 | /* Handle references */ | |
21666f1a | 1918 | else if (type == OBJ_COMMIT) |
40469ee9 | 1919 | ref_type = "tree "; |
21666f1a | 1920 | else if (type == OBJ_TAG) |
40469ee9 JH |
1921 | ref_type = "object "; |
1922 | else { | |
1923 | free(buffer); | |
1924 | return NULL; | |
1925 | } | |
1926 | ref_length = strlen(ref_type); | |
f4913f91 | 1927 | |
40469ee9 | 1928 | if (memcmp(buffer, ref_type, ref_length) || |
1d7f171c | 1929 | get_sha1_hex((char *) buffer + ref_length, actual_sha1)) { |
40469ee9 JH |
1930 | free(buffer); |
1931 | return NULL; | |
1932 | } | |
1cf58e72 | 1933 | free(buffer); |
40469ee9 JH |
1934 | /* Now we have the ID of the referred-to object in |
1935 | * actual_sha1. Check again. */ | |
f4913f91 | 1936 | } |
f4913f91 JH |
1937 | } |
1938 | ||
ce9fbf16 | 1939 | static void write_sha1_file_prepare(const void *buf, unsigned long len, |
972a9155 | 1940 | const char *type, unsigned char *sha1, |
d65a16f6 | 1941 | char *hdr, int *hdrlen) |
d410c0f5 JH |
1942 | { |
1943 | SHA_CTX c; | |
1944 | ||
1945 | /* Generate the header */ | |
d65a16f6 | 1946 | *hdrlen = sprintf(hdr, "%s %lu", type, len)+1; |
d410c0f5 JH |
1947 | |
1948 | /* Sha1.. */ | |
1949 | SHA1_Init(&c); | |
1950 | SHA1_Update(&c, hdr, *hdrlen); | |
1951 | SHA1_Update(&c, buf, len); | |
1952 | SHA1_Final(sha1, &c); | |
d410c0f5 JH |
1953 | } |
1954 | ||
230f1322 LT |
1955 | /* |
1956 | * Link the tempfile to the final place, possibly creating the | |
1957 | * last directory level as you do so. | |
1958 | * | |
1959 | * Returns the errno on failure, 0 on success. | |
1960 | */ | |
839837b9 | 1961 | static int link_temp_to_file(const char *tmpfile, const char *filename) |
230f1322 LT |
1962 | { |
1963 | int ret; | |
756aaf4a | 1964 | char *dir; |
230f1322 LT |
1965 | |
1966 | if (!link(tmpfile, filename)) | |
1967 | return 0; | |
1968 | ||
1969 | /* | |
756aaf4a | 1970 | * Try to mkdir the last path component if that failed. |
230f1322 LT |
1971 | * |
1972 | * Re-try the "link()" regardless of whether the mkdir | |
1973 | * succeeds, since a race might mean that somebody | |
1974 | * else succeeded. | |
1975 | */ | |
1976 | ret = errno; | |
756aaf4a SP |
1977 | dir = strrchr(filename, '/'); |
1978 | if (dir) { | |
1979 | *dir = 0; | |
866cae0d | 1980 | if (!mkdir(filename, 0777) && adjust_shared_perm(filename)) { |
91c23e48 | 1981 | *dir = '/'; |
756aaf4a | 1982 | return -2; |
91c23e48 | 1983 | } |
756aaf4a SP |
1984 | *dir = '/'; |
1985 | if (!link(tmpfile, filename)) | |
1986 | return 0; | |
1987 | ret = errno; | |
230f1322 LT |
1988 | } |
1989 | return ret; | |
1990 | } | |
1991 | ||
1992 | /* | |
1993 | * Move the just written object into its final resting place | |
1994 | */ | |
839837b9 | 1995 | int move_temp_to_file(const char *tmpfile, const char *filename) |
230f1322 LT |
1996 | { |
1997 | int ret = link_temp_to_file(tmpfile, filename); | |
7ebb6fca LT |
1998 | |
1999 | /* | |
2000 | * Coda hack - coda doesn't like cross-directory links, | |
2001 | * so we fall back to a rename, which will mean that it | |
2002 | * won't be able to check collisions, but that's not a | |
2003 | * big deal. | |
2004 | * | |
2005 | * The same holds for FAT formatted media. | |
2006 | * | |
2007 | * When this succeeds, we just return 0. We have nothing | |
2008 | * left to unlink. | |
2009 | */ | |
2010 | if (ret && ret != EEXIST) { | |
2011 | if (!rename(tmpfile, filename)) | |
230f1322 | 2012 | return 0; |
9e48b389 | 2013 | ret = errno; |
230f1322 LT |
2014 | } |
2015 | unlink(tmpfile); | |
2016 | if (ret) { | |
2017 | if (ret != EEXIST) { | |
916d081b | 2018 | return error("unable to write sha1 filename %s: %s\n", filename, strerror(ret)); |
230f1322 LT |
2019 | } |
2020 | /* FIXME!!! Collision check here ? */ | |
2021 | } | |
2022 | ||
2023 | return 0; | |
2024 | } | |
2025 | ||
4d548150 LT |
2026 | static int write_buffer(int fd, const void *buf, size_t len) |
2027 | { | |
d34cf19b | 2028 | if (write_in_full(fd, buf, len) < 0) |
93822c22 | 2029 | return error("file write error (%s)", strerror(errno)); |
4d548150 LT |
2030 | return 0; |
2031 | } | |
2032 | ||
ce9fbf16 | 2033 | int hash_sha1_file(const void *buf, unsigned long len, const char *type, |
abdc3fc8 RS |
2034 | unsigned char *sha1) |
2035 | { | |
d65a16f6 | 2036 | char hdr[32]; |
abdc3fc8 RS |
2037 | int hdrlen; |
2038 | write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen); | |
2039 | return 0; | |
2040 | } | |
2041 | ||
bf0f910d | 2042 | int write_sha1_file(void *buf, unsigned long len, const char *type, unsigned char *returnsha1) |
0fcfd160 | 2043 | { |
ac54c277 | 2044 | int size, ret; |
bf0f910d | 2045 | unsigned char *compressed; |
0fcfd160 LT |
2046 | z_stream stream; |
2047 | unsigned char sha1[20]; | |
706bc531 | 2048 | char *filename; |
aac17941 | 2049 | static char tmpfile[PATH_MAX]; |
d65a16f6 | 2050 | char hdr[32]; |
230f1322 | 2051 | int fd, hdrlen; |
a44c9a5e | 2052 | |
d410c0f5 JH |
2053 | /* Normally if we have it in the pack then we do not bother writing |
2054 | * it out into .git/objects/??/?{38} file. | |
2055 | */ | |
972a9155 RS |
2056 | write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen); |
2057 | filename = sha1_file_name(sha1); | |
706bc531 | 2058 | if (returnsha1) |
e702496e | 2059 | hashcpy(returnsha1, sha1); |
d410c0f5 JH |
2060 | if (has_sha1_file(sha1)) |
2061 | return 0; | |
aac17941 LT |
2062 | fd = open(filename, O_RDONLY); |
2063 | if (fd >= 0) { | |
706bc531 | 2064 | /* |
aac17941 LT |
2065 | * FIXME!!! We might do collision checking here, but we'd |
2066 | * need to uncompress the old file and check it. Later. | |
706bc531 | 2067 | */ |
aac17941 | 2068 | close(fd); |
706bc531 LT |
2069 | return 0; |
2070 | } | |
2071 | ||
aac17941 | 2072 | if (errno != ENOENT) { |
916d081b | 2073 | return error("sha1 file %s: %s\n", filename, strerror(errno)); |
aac17941 LT |
2074 | } |
2075 | ||
0e55181f | 2076 | snprintf(tmpfile, sizeof(tmpfile), "%s/tmp_obj_XXXXXX", get_object_directory()); |
ace1534d | 2077 | |
aac17941 LT |
2078 | fd = mkstemp(tmpfile); |
2079 | if (fd < 0) { | |
916d081b PB |
2080 | if (errno == EPERM) |
2081 | return error("insufficient permission for adding an object to repository database %s\n", get_object_directory()); | |
2082 | else | |
2083 | return error("unable to create temporary sha1 filename %s: %s\n", tmpfile, strerror(errno)); | |
aac17941 LT |
2084 | } |
2085 | ||
0fcfd160 LT |
2086 | /* Set it up */ |
2087 | memset(&stream, 0, sizeof(stream)); | |
12f6c308 | 2088 | deflateInit(&stream, zlib_compression_level); |
93821bd9 | 2089 | size = 8 + deflateBound(&stream, len+hdrlen); |
812666c8 | 2090 | compressed = xmalloc(size); |
0fcfd160 LT |
2091 | |
2092 | /* Compress it */ | |
0fcfd160 LT |
2093 | stream.next_out = compressed; |
2094 | stream.avail_out = size; | |
a44c9a5e LT |
2095 | |
2096 | /* First header.. */ | |
d65a16f6 | 2097 | stream.next_in = (unsigned char *)hdr; |
a44c9a5e | 2098 | stream.avail_in = hdrlen; |
726f852b NP |
2099 | while (deflate(&stream, 0) == Z_OK) |
2100 | /* nothing */; | |
a44c9a5e LT |
2101 | |
2102 | /* Then the data itself.. */ | |
2103 | stream.next_in = buf; | |
2104 | stream.avail_in = len; | |
ac54c277 LT |
2105 | ret = deflate(&stream, Z_FINISH); |
2106 | if (ret != Z_STREAM_END) | |
2107 | die("unable to deflate new object %s (%d)", sha1_to_hex(sha1), ret); | |
2108 | ||
2109 | ret = deflateEnd(&stream); | |
2110 | if (ret != Z_OK) | |
2111 | die("deflateEnd on object %s failed (%d)", sha1_to_hex(sha1), ret); | |
2112 | ||
0fcfd160 LT |
2113 | size = stream.total_out; |
2114 | ||
4d548150 LT |
2115 | if (write_buffer(fd, compressed, size) < 0) |
2116 | die("unable to write sha1 file"); | |
aac17941 | 2117 | fchmod(fd, 0444); |
e82973cf JH |
2118 | if (close(fd)) |
2119 | die("unable to write sha1 file"); | |
383f85b7 | 2120 | free(compressed); |
0fcfd160 | 2121 | |
230f1322 | 2122 | return move_temp_to_file(tmpfile, filename); |
0fcfd160 | 2123 | } |
8237b185 | 2124 | |
4d548150 LT |
2125 | /* |
2126 | * We need to unpack and recompress the object for writing | |
2127 | * it out to a different file. | |
2128 | */ | |
2129 | static void *repack_object(const unsigned char *sha1, unsigned long *objsize) | |
a5eda52b | 2130 | { |
4d548150 | 2131 | size_t size; |
a5eda52b | 2132 | z_stream stream; |
4d548150 LT |
2133 | unsigned char *unpacked; |
2134 | unsigned long len; | |
21666f1a | 2135 | enum object_type type; |
d65a16f6 | 2136 | char hdr[32]; |
4d548150 LT |
2137 | int hdrlen; |
2138 | void *buf; | |
bfc66daf | 2139 | |
a9486b02 | 2140 | /* need to unpack and recompress it by itself */ |
21666f1a | 2141 | unpacked = read_packed_sha1(sha1, &type, &len); |
8276c007 PE |
2142 | if (!unpacked) |
2143 | error("cannot read sha1_file for %s", sha1_to_hex(sha1)); | |
a5eda52b | 2144 | |
21666f1a | 2145 | hdrlen = sprintf(hdr, "%s %lu", typename(type), len) + 1; |
a5eda52b | 2146 | |
4d548150 LT |
2147 | /* Set it up */ |
2148 | memset(&stream, 0, sizeof(stream)); | |
12f6c308 | 2149 | deflateInit(&stream, zlib_compression_level); |
4d548150 LT |
2150 | size = deflateBound(&stream, len + hdrlen); |
2151 | buf = xmalloc(size); | |
a5eda52b | 2152 | |
4d548150 LT |
2153 | /* Compress it */ |
2154 | stream.next_out = buf; | |
2155 | stream.avail_out = size; | |
a5eda52b | 2156 | |
4d548150 LT |
2157 | /* First header.. */ |
2158 | stream.next_in = (void *)hdr; | |
2159 | stream.avail_in = hdrlen; | |
2160 | while (deflate(&stream, 0) == Z_OK) | |
2161 | /* nothing */; | |
bfc66daf | 2162 | |
4d548150 LT |
2163 | /* Then the data itself.. */ |
2164 | stream.next_in = unpacked; | |
2165 | stream.avail_in = len; | |
2166 | while (deflate(&stream, Z_FINISH) == Z_OK) | |
2167 | /* nothing */; | |
2168 | deflateEnd(&stream); | |
2169 | free(unpacked); | |
bfc66daf | 2170 | |
4d548150 LT |
2171 | *objsize = stream.total_out; |
2172 | return buf; | |
2173 | } | |
2174 | ||
2175 | int write_sha1_to_fd(int fd, const unsigned char *sha1) | |
2176 | { | |
2177 | int retval; | |
2178 | unsigned long objsize; | |
bb6b8e4f | 2179 | void *buf = map_sha1_file(sha1, &objsize); |
4d548150 LT |
2180 | |
2181 | if (buf) { | |
2182 | retval = write_buffer(fd, buf, objsize); | |
2183 | munmap(buf, objsize); | |
2184 | return retval; | |
2185 | } | |
2186 | ||
2187 | buf = repack_object(sha1, &objsize); | |
2188 | retval = write_buffer(fd, buf, objsize); | |
2189 | free(buf); | |
2190 | return retval; | |
a5eda52b DB |
2191 | } |
2192 | ||
70b9829e DB |
2193 | int write_sha1_from_fd(const unsigned char *sha1, int fd, char *buffer, |
2194 | size_t bufsize, size_t *bufposn) | |
8237b185 | 2195 | { |
230f1322 | 2196 | char tmpfile[PATH_MAX]; |
8237b185 DB |
2197 | int local; |
2198 | z_stream stream; | |
2199 | unsigned char real_sha1[20]; | |
bf0f910d | 2200 | unsigned char discard[4096]; |
8237b185 DB |
2201 | int ret; |
2202 | SHA_CTX c; | |
2203 | ||
0e55181f | 2204 | snprintf(tmpfile, sizeof(tmpfile), "%s/tmp_obj_XXXXXX", get_object_directory()); |
8237b185 | 2205 | |
230f1322 | 2206 | local = mkstemp(tmpfile); |
916d081b PB |
2207 | if (local < 0) { |
2208 | if (errno == EPERM) | |
2209 | return error("insufficient permission for adding an object to repository database %s\n", get_object_directory()); | |
2210 | else | |
2211 | return error("unable to create temporary sha1 filename %s: %s\n", tmpfile, strerror(errno)); | |
2212 | } | |
8237b185 DB |
2213 | |
2214 | memset(&stream, 0, sizeof(stream)); | |
2215 | ||
2216 | inflateInit(&stream); | |
2217 | ||
2218 | SHA1_Init(&c); | |
2219 | ||
2220 | do { | |
2221 | ssize_t size; | |
70b9829e DB |
2222 | if (*bufposn) { |
2223 | stream.avail_in = *bufposn; | |
96ad15ae | 2224 | stream.next_in = (unsigned char *) buffer; |
70b9829e DB |
2225 | do { |
2226 | stream.next_out = discard; | |
2227 | stream.avail_out = sizeof(discard); | |
2228 | ret = inflate(&stream, Z_SYNC_FLUSH); | |
2229 | SHA1_Update(&c, discard, sizeof(discard) - | |
2230 | stream.avail_out); | |
2231 | } while (stream.avail_in && ret == Z_OK); | |
4d548150 LT |
2232 | if (write_buffer(local, buffer, *bufposn - stream.avail_in) < 0) |
2233 | die("unable to write sha1 file"); | |
70b9829e DB |
2234 | memmove(buffer, buffer + *bufposn - stream.avail_in, |
2235 | stream.avail_in); | |
2236 | *bufposn = stream.avail_in; | |
2237 | if (ret != Z_OK) | |
2238 | break; | |
2239 | } | |
93d26e4c | 2240 | size = xread(fd, buffer + *bufposn, bufsize - *bufposn); |
8237b185 DB |
2241 | if (size <= 0) { |
2242 | close(local); | |
230f1322 | 2243 | unlink(tmpfile); |
8237b185 DB |
2244 | if (!size) |
2245 | return error("Connection closed?"); | |
2246 | perror("Reading from connection"); | |
2247 | return -1; | |
2248 | } | |
70b9829e DB |
2249 | *bufposn += size; |
2250 | } while (1); | |
8237b185 DB |
2251 | inflateEnd(&stream); |
2252 | ||
b5b8d814 | 2253 | fchmod(local, 0444); |
0d315468 JM |
2254 | if (close(local) != 0) |
2255 | die("unable to write sha1 file"); | |
8237b185 DB |
2256 | SHA1_Final(real_sha1, &c); |
2257 | if (ret != Z_STREAM_END) { | |
230f1322 | 2258 | unlink(tmpfile); |
8237b185 DB |
2259 | return error("File %s corrupted", sha1_to_hex(sha1)); |
2260 | } | |
a89fccd2 | 2261 | if (hashcmp(sha1, real_sha1)) { |
230f1322 | 2262 | unlink(tmpfile); |
bd2afde8 | 2263 | return error("File %s has bad hash", sha1_to_hex(sha1)); |
8237b185 | 2264 | } |
230f1322 LT |
2265 | |
2266 | return move_temp_to_file(tmpfile, sha1_file_name(sha1)); | |
8237b185 DB |
2267 | } |
2268 | ||
bf592c50 DB |
2269 | int has_pack_index(const unsigned char *sha1) |
2270 | { | |
2271 | struct stat st; | |
2272 | if (stat(sha1_pack_index_name(sha1), &st)) | |
2273 | return 0; | |
2274 | return 1; | |
2275 | } | |
2276 | ||
2277 | int has_pack_file(const unsigned char *sha1) | |
2278 | { | |
2279 | struct stat st; | |
2280 | if (stat(sha1_pack_name(sha1), &st)) | |
2281 | return 0; | |
2282 | return 1; | |
2283 | } | |
2284 | ||
106d710b | 2285 | int has_sha1_pack(const unsigned char *sha1, const char **ignore_packed) |
dade09c2 LT |
2286 | { |
2287 | struct pack_entry e; | |
106d710b | 2288 | return find_pack_entry(sha1, &e, ignore_packed); |
dade09c2 LT |
2289 | } |
2290 | ||
8237b185 DB |
2291 | int has_sha1_file(const unsigned char *sha1) |
2292 | { | |
8237b185 | 2293 | struct stat st; |
1f688557 JH |
2294 | struct pack_entry e; |
2295 | ||
106d710b | 2296 | if (find_pack_entry(sha1, &e, NULL)) |
1f688557 | 2297 | return 1; |
ab90ea5d | 2298 | return find_sha1_file(sha1, &st) ? 1 : 0; |
8237b185 | 2299 | } |
74400e71 | 2300 | |
e7332f96 BE |
2301 | /* |
2302 | * reads from fd as long as possible into a supplied buffer of size bytes. | |
82e5a82f | 2303 | * If necessary the buffer's size is increased using realloc() |
e7332f96 BE |
2304 | * |
2305 | * returns 0 if anything went fine and -1 otherwise | |
2306 | * | |
c4fba0a3 CR |
2307 | * The buffer is always NUL-terminated, not including it in returned size. |
2308 | * | |
e7332f96 BE |
2309 | * NOTE: both buf and size may change, but even when -1 is returned |
2310 | * you still have to free() it yourself. | |
2311 | */ | |
c4fba0a3 | 2312 | int read_fd(int fd, char **return_buf, unsigned long *return_size) |
024510c8 | 2313 | { |
c4fba0a3 | 2314 | char *buf = *return_buf; |
e7332f96 | 2315 | unsigned long size = *return_size; |
8a912bcb | 2316 | ssize_t iret; |
024510c8 | 2317 | unsigned long off = 0; |
e7332f96 | 2318 | |
c4fba0a3 CR |
2319 | if (!buf || size <= 1) { |
2320 | size = 1024; | |
2321 | buf = xrealloc(buf, size); | |
2322 | } | |
2323 | ||
024510c8 | 2324 | do { |
c4fba0a3 | 2325 | iret = xread(fd, buf + off, (size - 1) - off); |
024510c8 DB |
2326 | if (iret > 0) { |
2327 | off += iret; | |
c4fba0a3 CR |
2328 | if (off == size - 1) { |
2329 | size = alloc_nr(size); | |
83572c1a | 2330 | buf = xrealloc(buf, size); |
024510c8 DB |
2331 | } |
2332 | } | |
2333 | } while (iret > 0); | |
e7332f96 | 2334 | |
c4fba0a3 CR |
2335 | buf[off] = '\0'; |
2336 | ||
e7332f96 BE |
2337 | *return_buf = buf; |
2338 | *return_size = off; | |
2339 | ||
2340 | if (iret < 0) | |
2341 | return -1; | |
2342 | return 0; | |
2343 | } | |
2344 | ||
2345 | int index_pipe(unsigned char *sha1, int fd, const char *type, int write_object) | |
2346 | { | |
2347 | unsigned long size = 4096; | |
2d7320d0 | 2348 | char *buf = xmalloc(size); |
e7332f96 | 2349 | int ret; |
e7332f96 | 2350 | |
c4fba0a3 | 2351 | if (read_fd(fd, &buf, &size)) { |
024510c8 DB |
2352 | free(buf); |
2353 | return -1; | |
2354 | } | |
e7332f96 | 2355 | |
024510c8 | 2356 | if (!type) |
8e440259 | 2357 | type = blob_type; |
024510c8 | 2358 | if (write_object) |
e7332f96 | 2359 | ret = write_sha1_file(buf, size, type, sha1); |
abdc3fc8 RS |
2360 | else |
2361 | ret = hash_sha1_file(buf, size, type, sha1); | |
024510c8 DB |
2362 | free(buf); |
2363 | return ret; | |
2364 | } | |
2365 | ||
edaec3fb | 2366 | int index_fd(unsigned char *sha1, int fd, struct stat *st, int write_object, |
53bca91a | 2367 | enum object_type type, const char *path) |
74400e71 | 2368 | { |
dc49cd76 | 2369 | size_t size = xsize_t(st->st_size); |
3a55602e | 2370 | void *buf = NULL; |
6c510bee | 2371 | int ret, re_allocated = 0; |
74400e71 | 2372 | |
74400e71 | 2373 | if (size) |
c4712e45 | 2374 | buf = xmmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0); |
74400e71 | 2375 | close(fd); |
74400e71 | 2376 | |
7672db20 | 2377 | if (!type) |
edaec3fb | 2378 | type = OBJ_BLOB; |
6c510bee LT |
2379 | |
2380 | /* | |
2381 | * Convert blobs to git internal format | |
2382 | */ | |
d0d8e14d | 2383 | if ((type == OBJ_BLOB) && S_ISREG(st->st_mode)) { |
6c510bee | 2384 | unsigned long nsize = size; |
ac78e548 AR |
2385 | char *nbuf = convert_to_git(path, buf, &nsize); |
2386 | if (nbuf) { | |
2387 | munmap(buf, size); | |
6c510bee LT |
2388 | size = nsize; |
2389 | buf = nbuf; | |
2390 | re_allocated = 1; | |
2391 | } | |
2392 | } | |
2393 | ||
7672db20 | 2394 | if (write_object) |
edaec3fb | 2395 | ret = write_sha1_file(buf, size, typename(type), sha1); |
abdc3fc8 | 2396 | else |
edaec3fb | 2397 | ret = hash_sha1_file(buf, size, typename(type), sha1); |
6c510bee LT |
2398 | if (re_allocated) { |
2399 | free(buf); | |
2400 | return ret; | |
2401 | } | |
aac17941 LT |
2402 | if (size) |
2403 | munmap(buf, size); | |
2404 | return ret; | |
74400e71 | 2405 | } |
ec1fcc16 JH |
2406 | |
2407 | int index_path(unsigned char *sha1, const char *path, struct stat *st, int write_object) | |
2408 | { | |
2409 | int fd; | |
2410 | char *target; | |
dc49cd76 | 2411 | size_t len; |
ec1fcc16 JH |
2412 | |
2413 | switch (st->st_mode & S_IFMT) { | |
2414 | case S_IFREG: | |
2415 | fd = open(path, O_RDONLY); | |
2416 | if (fd < 0) | |
2417 | return error("open(\"%s\"): %s", path, | |
2418 | strerror(errno)); | |
53bca91a | 2419 | if (index_fd(sha1, fd, st, write_object, OBJ_BLOB, path) < 0) |
ec1fcc16 JH |
2420 | return error("%s: failed to insert into database", |
2421 | path); | |
2422 | break; | |
2423 | case S_IFLNK: | |
dc49cd76 SP |
2424 | len = xsize_t(st->st_size); |
2425 | target = xmalloc(len + 1); | |
2426 | if (readlink(path, target, len + 1) != st->st_size) { | |
ec1fcc16 JH |
2427 | char *errstr = strerror(errno); |
2428 | free(target); | |
2429 | return error("readlink(\"%s\"): %s", path, | |
2430 | errstr); | |
2431 | } | |
abdc3fc8 | 2432 | if (!write_object) |
dc49cd76 SP |
2433 | hash_sha1_file(target, len, blob_type, sha1); |
2434 | else if (write_sha1_file(target, len, blob_type, sha1)) | |
ec1fcc16 JH |
2435 | return error("%s: failed to insert into database", |
2436 | path); | |
2437 | free(target); | |
2438 | break; | |
f35a6d3b LT |
2439 | case S_IFDIR: |
2440 | return resolve_gitlink_ref(path, "HEAD", sha1); | |
ec1fcc16 JH |
2441 | default: |
2442 | return error("%s: unsupported file type", path); | |
2443 | } | |
2444 | return 0; | |
2445 | } | |
a69e5429 JH |
2446 | |
2447 | int read_pack_header(int fd, struct pack_header *header) | |
2448 | { | |
2449 | char *c = (char*)header; | |
2450 | ssize_t remaining = sizeof(struct pack_header); | |
2451 | do { | |
2452 | ssize_t r = xread(fd, c, remaining); | |
2453 | if (r <= 0) | |
2454 | /* "eof before pack header was fully read" */ | |
2455 | return PH_ERROR_EOF; | |
2456 | remaining -= r; | |
2457 | c += r; | |
2458 | } while (remaining > 0); | |
2459 | if (header->hdr_signature != htonl(PACK_SIGNATURE)) | |
2460 | /* "protocol error (pack signature mismatch detected)" */ | |
2461 | return PH_ERROR_PACK_SIGNATURE; | |
2462 | if (!pack_version_ok(header->hdr_version)) | |
2463 | /* "protocol error (pack version unsupported)" */ | |
2464 | return PH_ERROR_PROTOCOL; | |
2465 | return 0; | |
2466 | } |