sha1_file.c

   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  */
   9 #include <stdarg.h>
  10 #include <limits.h>
  11 #include "cache.h"
  12
  13 #ifndef O_NOATIME
  14 #if defined(__linux__) && (defined(__i386__) || defined(__PPC__))
  15 #define O_NOATIME 01000000
  16 #else
  17 #define O_NOATIME 0
  18 #endif
  19 #endif
  20
  21 static unsigned int sha1_file_open_flag = O_NOATIME;
  22
  23 static unsigned hexval(char c)
  24 {
  25         if (c >= '0' && c <= '9')
  26                 return c - '0';
  27         if (c >= 'a' && c <= 'f')
  28                 return c - 'a' + 10;
  29         if (c >= 'A' && c <= 'F')
  30                 return c - 'A' + 10;
  31         return ~0;
  32 }
  33
  34 int get_sha1_hex(const char *hex, unsigned char *sha1)
  35 {
  36         int i;
  37         for (i = 0; i < 20; i++) {
  38                 unsigned int val = (hexval(hex[0]) << 4) | hexval(hex[1]);
  39                 if (val & ~0xff)
  40                         return -1;
  41                 *sha1++ = val;
  42                 hex += 2;
  43         }
  44         return 0;
  45 }
  46
  47 int get_sha1_file(const char *path, unsigned char *result)
  48 {
  49         char buffer[60];
  50         int fd = open(path, O_RDONLY);
  51         int len;
  52
  53         if (fd < 0)
  54                 return -1;
  55         len = read(fd, buffer, sizeof(buffer));
  56         close(fd);
  57         if (len < 40)
  58                 return -1;
  59         return get_sha1_hex(buffer, result);
  60 }
  61
  62 int get_sha1(const char *str, unsigned char *sha1)
  63 {
  64         static char pathname[PATH_MAX];
  65         static const char *prefix[] = {
  66                 "",
  67                 "refs",
  68                 "refs/tags",
  69                 "refs/heads",
  70                 "refs/snap",
  71                 NULL
  72         };
  73         const char *gitdir;
  74         const char **p;
  75
  76         if (!get_sha1_hex(str, sha1))
  77                 return 0;
  78
  79         gitdir = ".git";
  80         for (p = prefix; *p; p++) {
  81                 snprintf(pathname, sizeof(pathname), "%s/%s/%s", gitdir, *p, str);
  82                 if (!get_sha1_file(pathname, sha1))
  83                         return 0;
  84         }
  85
  86         return -1;
  87 }
  88
  89 char * sha1_to_hex(const unsigned char *sha1)
  90 {
  91         static char buffer[50];
  92         static const char hex[] = "0123456789abcdef";
  93         char *buf = buffer;
  94         int i;
  95
  96         for (i = 0; i < 20; i++) {
  97                 unsigned int val = *sha1++;
  98                 *buf++ = hex[val >> 4];
  99                 *buf++ = hex[val & 0xf];
 100         }
 101         return buffer;
 102 }
 103
 104 static void fill_sha1_path(char *pathbuf, const unsigned char *sha1)
 105 {
 106         int i;
 107         for (i = 0; i < 20; i++) {
 108                 static char hex[] = "0123456789abcdef";
 109                 unsigned int val = sha1[i];
 110                 char *pos = pathbuf + i*2 + (i > 0);
 111                 *pos++ = hex[val >> 4];
 112                 *pos = hex[val & 0xf];
 113         }
 114 }
 115
 116 /*
 117  * NOTE! This returns a statically allocated buffer, so you have to be
 118  * careful about using it. Do a "strdup()" if you need to save the
 119  * filename.
 120  *
 121  * Also note that this returns the location for creating.  Reading
 122  * SHA1 file can happen from any alternate directory listed in the
 123  * SHA1_FILE_DIRECTORIES environment variable if it is not found in
 124  * the primary object database.
 125  */
 126 char *sha1_file_name(const unsigned char *sha1)
 127 {
 128         static char *name, *base;
 129
 130         if (!base) {
 131                 char *sha1_file_directory = get_object_directory();
 132                 int len = strlen(sha1_file_directory);
 133                 base = xmalloc(len + 60);
 134                 memcpy(base, sha1_file_directory, len);
 135                 memset(base+len, 0, 60);
 136                 base[len] = '/';
 137                 base[len+3] = '/';
 138                 name = base + len + 1;
 139         }
 140         fill_sha1_path(name, sha1);
 141         return base;
 142 }
 143
 144 static struct alternate_object_database {
 145         char *base;
 146         char *name;
 147 } *alt_odb;
 148
 149 /*
 150  * Prepare alternate object database registry.
 151  * alt_odb points at an array of struct alternate_object_database.
 152  * This array is terminated with an element that has both its base
 153  * and name set to NULL.  alt_odb[n] comes from n'th non-empty
 154  * element from colon separated $SHA1_FILE_DIRECTORIES environment
 155  * variable, and its base points at a statically allocated buffer
 156  * that contains "/the/directory/corresponding/to/.git/objects/...",
 157  * while its name points just after the slash at the end of
 158  * ".git/objects/" in the example above, and has enough space to hold
 159  * 40-byte hex SHA1, an extra slash for the first level indirection,
 160  * and the terminating NUL.
 161  * This function allocates the alt_odb array and all the strings
 162  * pointed by base fields of the array elements with one xmalloc();
 163  * the string pool immediately follows the array.
 164  */
 165 static void prepare_alt_odb(void)
 166 {
 167         int pass, totlen, i;
 168         const char *cp, *last;
 169         char *op = 0;
 170         const char *alt = getenv(ALTERNATE_DB_ENVIRONMENT) ? : "";
 171
 172         /* The first pass counts how large an area to allocate to
 173          * hold the entire alt_odb structure, including array of
 174          * structs and path buffers for them.  The second pass fills
 175          * the structure and prepares the path buffers for use by
 176          * fill_sha1_path().
 177          */
 178         for (totlen = pass = 0; pass < 2; pass++) {
 179                 last = alt;
 180                 i = 0;
 181                 do {
 182                         cp = strchr(last, ':') ? : last + strlen(last);
 183                         if (last != cp) {
 184                                 /* 43 = 40-byte + 2 '/' + terminating NUL */
 185                                 int pfxlen = cp - last;
 186                                 int entlen = pfxlen + 43;
 187                                 if (pass == 0)
 188                                         totlen += entlen;
 189                                 else {
 190                                         alt_odb[i].base = op;
 191                                         alt_odb[i].name = op + pfxlen + 1;
 192                                         memcpy(op, last, pfxlen);
 193                                         op[pfxlen] = op[pfxlen + 3] = '/';
 194                                         op[entlen-1] = 0;
 195                                         op += entlen;
 196                                 }
 197                                 i++;
 198                         }
 199                         while (*cp && *cp == ':')
 200                                 cp++;
 201                         last = cp;
 202                 } while (*cp);
 203                 if (pass)
 204                         break;
 205                 alt_odb = xmalloc(sizeof(*alt_odb) * (i + 1) + totlen);
 206                 alt_odb[i].base = alt_odb[i].name = 0;
 207                 op = (char*)(&alt_odb[i+1]);
 208         }
 209 }
 210
 211 static char *find_sha1_file(const unsigned char *sha1, struct stat *st)
 212 {
 213         int i;
 214         char *name = sha1_file_name(sha1);
 215
 216         if (!stat(name, st))
 217                 return name;
 218         if (!alt_odb)
 219                 prepare_alt_odb();
 220         for (i = 0; (name = alt_odb[i].name) != NULL; i++) {
 221                 fill_sha1_path(name, sha1);
 222                 if (!stat(alt_odb[i].base, st))
 223                         return alt_odb[i].base;
 224         }
 225         return NULL;
 226 }
 227
 228 int check_sha1_signature(unsigned char *sha1, void *map, unsigned long size, const char *type)
 229 {
 230         char header[100];
 231         unsigned char real_sha1[20];
 232         SHA_CTX c;
 233
 234         SHA1_Init(&c);
 235         SHA1_Update(&c, header, 1+sprintf(header, "%s %lu", type, size));
 236         SHA1_Update(&c, map, size);
 237         SHA1_Final(real_sha1, &c);
 238         return memcmp(sha1, real_sha1, 20) ? -1 : 0;
 239 }
 240
 241 void *map_sha1_file(const unsigned char *sha1, unsigned long *size)
 242 {
 243         struct stat st;
 244         void *map;
 245         int fd;
 246         char *filename = find_sha1_file(sha1, &st);
 247
 248         if (!filename) {
 249                 error("cannot map sha1 file %s", sha1_to_hex(sha1));
 250                 return NULL;
 251         }
 252
 253         fd = open(filename, O_RDONLY | sha1_file_open_flag);
 254         if (fd < 0) {
 255                 /* See if it works without O_NOATIME */
 256                 switch (sha1_file_open_flag) {
 257                 default:
 258                         fd = open(filename, O_RDONLY);
 259                         if (fd >= 0)
 260                                 break;
 261                 /* Fallthrough */
 262                 case 0:
 263                         perror(filename);
 264                         return NULL;
 265                 }
 266
 267                 /* If it failed once, it will probably fail again. Stop using O_NOATIME */
 268                 sha1_file_open_flag = 0;
 269         }
 270         map = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
 271         close(fd);
 272         if (-1 == (int)(long)map)
 273                 return NULL;
 274         *size = st.st_size;
 275         return map;
 276 }
 277
 278 void * unpack_sha1_file(void *map, unsigned long mapsize, char *type, unsigned long *size)
 279 {
 280         int ret, bytes;
 281         z_stream stream;
 282         char buffer[8192];
 283         char *buf;
 284
 285         /* Get the data stream */
 286         memset(&stream, 0, sizeof(stream));
 287         stream.next_in = map;
 288         stream.avail_in = mapsize;
 289         stream.next_out = buffer;
 290         stream.avail_out = sizeof(buffer);
 291
 292         inflateInit(&stream);
 293         ret = inflate(&stream, 0);
 294         if (ret < Z_OK)
 295                 return NULL;
 296         if (sscanf(buffer, "%10s %lu", type, size) != 2)
 297                 return NULL;
 298
 299         bytes = strlen(buffer) + 1;
 300         buf = xmalloc(*size);
 301
 302         memcpy(buf, buffer + bytes, stream.total_out - bytes);
 303         bytes = stream.total_out - bytes;
 304         if (bytes < *size && ret == Z_OK) {
 305                 stream.next_out = buf + bytes;
 306                 stream.avail_out = *size - bytes;
 307                 while (inflate(&stream, Z_FINISH) == Z_OK)
 308                         /* nothing */;
 309         }
 310         inflateEnd(&stream);
 311         return buf;
 312 }
 313
 314 void * read_sha1_file(const unsigned char *sha1, char *type, unsigned long *size)
 315 {
 316         unsigned long mapsize;
 317         void *map, *buf;
 318
 319         map = map_sha1_file(sha1, &mapsize);
 320         if (map) {
 321                 buf = unpack_sha1_file(map, mapsize, type, size);
 322                 munmap(map, mapsize);
 323                 return buf;
 324         }
 325         return NULL;
 326 }
 327
 328 void *read_object_with_reference(const unsigned char *sha1,
 329                                  const unsigned char *required_type,
 330                                  unsigned long *size,
 331                                  unsigned char *actual_sha1_return)
 332 {
 333         char type[20];
 334         void *buffer;
 335         unsigned long isize;
 336         unsigned char actual_sha1[20];
 337
 338         memcpy(actual_sha1, sha1, 20);
 339         while (1) {
 340                 int ref_length = -1;
 341                 const char *ref_type = NULL;
 342
 343                 buffer = read_sha1_file(actual_sha1, type, &isize);
 344                 if (!buffer)
 345                         return NULL;
 346                 if (!strcmp(type, required_type)) {
 347                         *size = isize;
 348                         if (actual_sha1_return)
 349                                 memcpy(actual_sha1_return, actual_sha1, 20);
 350                         return buffer;
 351                 }
 352                 /* Handle references */
 353                 else if (!strcmp(type, "commit"))
 354                         ref_type = "tree ";
 355                 else if (!strcmp(type, "tag"))
 356                         ref_type = "object ";
 357                 else {
 358                         free(buffer);
 359                         return NULL;
 360                 }
 361                 ref_length = strlen(ref_type);
 362
 363                 if (memcmp(buffer, ref_type, ref_length) ||
 364                     get_sha1_hex(buffer + ref_length, actual_sha1)) {
 365                         free(buffer);
 366                         return NULL;
 367                 }
 368                 /* Now we have the ID of the referred-to object in
 369                  * actual_sha1.  Check again. */
 370         }
 371 }
 372
 373 int write_sha1_file(char *buf, unsigned long len, const char *type, unsigned char *returnsha1)
 374 {
 375         int size;
 376         char *compressed;
 377         z_stream stream;
 378         unsigned char sha1[20];
 379         SHA_CTX c;
 380         char *filename;
 381         static char tmpfile[PATH_MAX];
 382         char hdr[50];
 383         int fd, hdrlen, ret;
 384
 385         /* Generate the header */
 386         hdrlen = sprintf(hdr, "%s %lu", type, len)+1;
 387
 388         /* Sha1.. */
 389         SHA1_Init(&c);
 390         SHA1_Update(&c, hdr, hdrlen);
 391         SHA1_Update(&c, buf, len);
 392         SHA1_Final(sha1, &c);
 393
 394         if (returnsha1)
 395                 memcpy(returnsha1, sha1, 20);
 396
 397         filename = sha1_file_name(sha1);
 398         fd = open(filename, O_RDONLY);
 399         if (fd >= 0) {
 400                 /*
 401                  * FIXME!!! We might do collision checking here, but we'd
 402                  * need to uncompress the old file and check it. Later.
 403                  */
 404                 close(fd);
 405                 return 0;
 406         }
 407
 408         if (errno != ENOENT) {
 409                 fprintf(stderr, "sha1 file %s: %s", filename, strerror(errno));
 410                 return -1;
 411         }
 412
 413         snprintf(tmpfile, sizeof(tmpfile), "%s/obj_XXXXXX", get_object_directory());
 414
 415         fd = mkstemp(tmpfile);
 416         if (fd < 0) {
 417                 fprintf(stderr, "unable to create temporary sha1 filename %s: %s", tmpfile, strerror(errno));
 418                 return -1;
 419         }
 420
 421         /* Set it up */
 422         memset(&stream, 0, sizeof(stream));
 423         deflateInit(&stream, Z_BEST_COMPRESSION);
 424         size = deflateBound(&stream, len+hdrlen);
 425         compressed = xmalloc(size);
 426
 427         /* Compress it */
 428         stream.next_out = compressed;
 429         stream.avail_out = size;
 430
 431         /* First header.. */
 432         stream.next_in = hdr;
 433         stream.avail_in = hdrlen;
 434         while (deflate(&stream, 0) == Z_OK)
 435                 /* nothing */
 436
 437         /* Then the data itself.. */
 438         stream.next_in = buf;
 439         stream.avail_in = len;
 440         while (deflate(&stream, Z_FINISH) == Z_OK)
 441                 /* nothing */;
 442         deflateEnd(&stream);
 443         size = stream.total_out;
 444
 445         if (write(fd, compressed, size) != size)
 446                 die("unable to write file");
 447         fchmod(fd, 0444);
 448         close(fd);
 449         free(compressed);
 450
 451         ret = link(tmpfile, filename);
 452         if (ret < 0) {
 453                 ret = errno;
 454
 455                 /*
 456                  * Coda hack - coda doesn't like cross-directory links,
 457                  * so we fall back to a rename, which will mean that it
 458                  * won't be able to check collisions, but that's not a
 459                  * big deal.
 460                  *
 461                  * When this succeeds, we just return 0. We have nothing
 462                  * left to unlink.
 463                  */
 464                 if (ret == EXDEV && !rename(tmpfile, filename))
 465                         return 0;
 466         }
 467         unlink(tmpfile);
 468         if (ret) {
 469                 if (ret != EEXIST) {
 470                         fprintf(stderr, "unable to write sha1 filename %s: %s", filename, strerror(ret));
 471                         return -1;
 472                 }
 473                 /* FIXME!!! Collision check here ? */
 474         }
 475
 476         return 0;
 477 }
 478
 479 int write_sha1_from_fd(const unsigned char *sha1, int fd)
 480 {
 481         char *filename = sha1_file_name(sha1);
 482
 483         int local;
 484         z_stream stream;
 485         unsigned char real_sha1[20];
 486         char buf[4096];
 487         char discard[4096];
 488         int ret;
 489         SHA_CTX c;
 490
 491         local = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0666);
 492
 493         if (local < 0)
 494                 return error("Couldn't open %s\n", filename);
 495
 496         memset(&stream, 0, sizeof(stream));
 497
 498         inflateInit(&stream);
 499
 500         SHA1_Init(&c);
 501
 502         do {
 503                 ssize_t size;
 504                 size = read(fd, buf, 4096);
 505                 if (size <= 0) {
 506                         close(local);
 507                         unlink(filename);
 508                         if (!size)
 509                                 return error("Connection closed?");
 510                         perror("Reading from connection");
 511                         return -1;
 512                 }
 513                 write(local, buf, size);
 514                 stream.avail_in = size;
 515                 stream.next_in = buf;
 516                 do {
 517                         stream.next_out = discard;
 518                         stream.avail_out = sizeof(discard);
 519                         ret = inflate(&stream, Z_SYNC_FLUSH);
 520                         SHA1_Update(&c, discard, sizeof(discard) -
 521                                     stream.avail_out);
 522                 } while (stream.avail_in && ret == Z_OK);
 523
 524         } while (ret == Z_OK);
 525         inflateEnd(&stream);
 526
 527         close(local);
 528         SHA1_Final(real_sha1, &c);
 529         if (ret != Z_STREAM_END) {
 530                 unlink(filename);
 531                 return error("File %s corrupted", sha1_to_hex(sha1));
 532         }
 533         if (memcmp(sha1, real_sha1, 20)) {
 534                 unlink(filename);
 535                 return error("File %s has bad hash\n", sha1_to_hex(sha1));
 536         }
 537
 538         return 0;
 539 }
 540
 541 int has_sha1_file(const unsigned char *sha1)
 542 {
 543         struct stat st;
 544         return !!find_sha1_file(sha1, &st);
 545 }
 546
 547 int index_fd(unsigned char *sha1, int fd, struct stat *st)
 548 {
 549         unsigned long size = st->st_size;
 550         void *buf;
 551         int ret;
 552
 553         buf = "";
 554         if (size)
 555                 buf = mmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
 556         close(fd);
 557         if ((int)(long)buf == -1)
 558                 return -1;
 559
 560         ret = write_sha1_file(buf, size, "blob", sha1);
 561         if (size)
 562                 munmap(buf, size);
 563         return ret;
 564 }