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 "cache.h"
  11
  12 #ifndef O_NOATIME
  13 #if defined(__linux__) && (defined(__i386__) || defined(__PPC__))
  14 #define O_NOATIME 01000000
  15 #else
  16 #define O_NOATIME 0
  17 #endif
  18 #endif
  19
  20 static unsigned int sha1_file_open_flag = O_NOATIME;
  21
  22 static unsigned hexval(char c)
  23 {
  24         if (c >= '0' && c <= '9')
  25                 return c - '0';
  26         if (c >= 'a' && c <= 'f')
  27                 return c - 'a' + 10;
  28         if (c >= 'A' && c <= 'F')
  29                 return c - 'A' + 10;
  30         return ~0;
  31 }
  32
  33 int get_sha1_hex(const char *hex, unsigned char *sha1)
  34 {
  35         int i;
  36         for (i = 0; i < 20; i++) {
  37                 unsigned int val = (hexval(hex[0]) << 4) | hexval(hex[1]);
  38                 if (val & ~0xff)
  39                         return -1;
  40                 *sha1++ = val;
  41                 hex += 2;
  42         }
  43         return 0;
  44 }
  45
  46 int get_sha1_file(const char *path, unsigned char *result)
  47 {
  48         char buffer[60];
  49         int fd = open(path, O_RDONLY);
  50         int len;
  51
  52         if (fd < 0)
  53                 return -1;
  54         len = read(fd, buffer, sizeof(buffer));
  55         close(fd);
  56         if (len < 40)
  57                 return -1;
  58         return get_sha1_hex(buffer, result);
  59 }
  60
  61 int get_sha1(const char *str, unsigned char *sha1)
  62 {
  63         static char pathname[PATH_MAX];
  64         static const char *prefix[] = {
  65                 "",
  66                 "refs",
  67                 "refs/tags",
  68                 "refs/heads",
  69                 "refs/snap",
  70                 NULL
  71         };
  72         const char *gitdir;
  73         const char **p;
  74
  75         if (!get_sha1_hex(str, sha1))
  76                 return 0;
  77
  78         gitdir = ".git";
  79         for (p = prefix; *p; p++) {
  80                 snprintf(pathname, sizeof(pathname), "%s/%s/%s", gitdir, *p, str);
  81                 if (!get_sha1_file(pathname, sha1))
  82                         return 0;
  83         }
  84
  85         return -1;
  86 }
  87
  88 char * sha1_to_hex(const unsigned char *sha1)
  89 {
  90         static char buffer[50];
  91         static const char hex[] = "0123456789abcdef";
  92         char *buf = buffer;
  93         int i;
  94
  95         for (i = 0; i < 20; i++) {
  96                 unsigned int val = *sha1++;
  97                 *buf++ = hex[val >> 4];
  98                 *buf++ = hex[val & 0xf];
  99         }
 100         return buffer;
 101 }
 102
 103 /*
 104  * NOTE! This returns a statically allocated buffer, so you have to be
 105  * careful about using it. Do a "strdup()" if you need to save the
 106  * filename.
 107  */
 108 char *sha1_file_name(const unsigned char *sha1)
 109 {
 110         int i;
 111         static char *name, *base;
 112
 113         if (!base) {
 114                 char *sha1_file_directory = getenv(DB_ENVIRONMENT) ? : DEFAULT_DB_ENVIRONMENT;
 115                 int len = strlen(sha1_file_directory);
 116                 base = xmalloc(len + 60);
 117                 memcpy(base, sha1_file_directory, len);
 118                 memset(base+len, 0, 60);
 119                 base[len] = '/';
 120                 base[len+3] = '/';
 121                 name = base + len + 1;
 122         }
 123         for (i = 0; i < 20; i++) {
 124                 static char hex[] = "0123456789abcdef";
 125                 unsigned int val = sha1[i];
 126                 char *pos = name + i*2 + (i > 0);
 127                 *pos++ = hex[val >> 4];
 128                 *pos = hex[val & 0xf];
 129         }
 130         return base;
 131 }
 132
 133 int check_sha1_signature(unsigned char *sha1, void *map, unsigned long size, const char *type)
 134 {
 135         char header[100];
 136         unsigned char real_sha1[20];
 137         SHA_CTX c;
 138
 139         SHA1_Init(&c);
 140         SHA1_Update(&c, header, 1+sprintf(header, "%s %lu", type, size));
 141         SHA1_Update(&c, map, size);
 142         SHA1_Final(real_sha1, &c);
 143         return memcmp(sha1, real_sha1, 20) ? -1 : 0;
 144 }
 145
 146 void *map_sha1_file(const unsigned char *sha1, unsigned long *size)
 147 {
 148         char *filename = sha1_file_name(sha1);
 149         struct stat st;
 150         void *map;
 151         int fd;
 152
 153         fd = open(filename, O_RDONLY | sha1_file_open_flag);
 154         if (fd < 0) {
 155                 /* See if it works without O_NOATIME */
 156                 switch (sha1_file_open_flag) {
 157                 default:
 158                         fd = open(filename, O_RDONLY);
 159                         if (fd >= 0)
 160                                 break;
 161                 /* Fallthrough */
 162                 case 0:
 163                         perror(filename);
 164                         return NULL;
 165                 }
 166
 167                 /* If it failed once, it will probably fail again. Stop using O_NOATIME */
 168                 sha1_file_open_flag = 0;
 169         }
 170         if (fstat(fd, &st) < 0) {
 171                 close(fd);
 172                 return NULL;
 173         }
 174         map = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
 175         close(fd);
 176         if (-1 == (int)(long)map)
 177                 return NULL;
 178         *size = st.st_size;
 179         return map;
 180 }
 181
 182 void * unpack_sha1_file(void *map, unsigned long mapsize, char *type, unsigned long *size)
 183 {
 184         int ret, bytes;
 185         z_stream stream;
 186         char buffer[8192];
 187         char *buf;
 188
 189         /* Get the data stream */
 190         memset(&stream, 0, sizeof(stream));
 191         stream.next_in = map;
 192         stream.avail_in = mapsize;
 193         stream.next_out = buffer;
 194         stream.avail_out = sizeof(buffer);
 195
 196         inflateInit(&stream);
 197         ret = inflate(&stream, 0);
 198         if (ret < Z_OK)
 199                 return NULL;
 200         if (sscanf(buffer, "%10s %lu", type, size) != 2)
 201                 return NULL;
 202
 203         bytes = strlen(buffer) + 1;
 204         buf = xmalloc(*size);
 205
 206         memcpy(buf, buffer + bytes, stream.total_out - bytes);
 207         bytes = stream.total_out - bytes;
 208         if (bytes < *size && ret == Z_OK) {
 209                 stream.next_out = buf + bytes;
 210                 stream.avail_out = *size - bytes;
 211                 while (inflate(&stream, Z_FINISH) == Z_OK)
 212                         /* nothing */;
 213         }
 214         inflateEnd(&stream);
 215         return buf;
 216 }
 217
 218 void * read_sha1_file(const unsigned char *sha1, char *type, unsigned long *size)
 219 {
 220         unsigned long mapsize;
 221         void *map, *buf;
 222
 223         map = map_sha1_file(sha1, &mapsize);
 224         if (map) {
 225                 buf = unpack_sha1_file(map, mapsize, type, size);
 226                 munmap(map, mapsize);
 227                 return buf;
 228         }
 229         return NULL;
 230 }
 231
 232 void *read_object_with_reference(const unsigned char *sha1,
 233                                  const unsigned char *required_type,
 234                                  unsigned long *size,
 235                                  unsigned char *actual_sha1_return)
 236 {
 237         char type[20];
 238         void *buffer;
 239         unsigned long isize;
 240         unsigned char actual_sha1[20];
 241
 242         memcpy(actual_sha1, sha1, 20);
 243         while (1) {
 244                 int ref_length = -1;
 245                 const char *ref_type = NULL;
 246
 247                 buffer = read_sha1_file(actual_sha1, type, &isize);
 248                 if (!buffer)
 249                         return NULL;
 250                 if (!strcmp(type, required_type)) {
 251                         *size = isize;
 252                         if (actual_sha1_return)
 253                                 memcpy(actual_sha1_return, actual_sha1, 20);
 254                         return buffer;
 255                 }
 256                 /* Handle references */
 257                 else if (!strcmp(type, "commit"))
 258                         ref_type = "tree ";
 259                 else if (!strcmp(type, "tag"))
 260                         ref_type = "object ";
 261                 else {
 262                         free(buffer);
 263                         return NULL;
 264                 }
 265                 ref_length = strlen(ref_type);
 266
 267                 if (memcmp(buffer, ref_type, ref_length) ||
 268                     get_sha1_hex(buffer + ref_length, actual_sha1)) {
 269                         free(buffer);
 270                         return NULL;
 271                 }
 272                 /* Now we have the ID of the referred-to object in
 273                  * actual_sha1.  Check again. */
 274         }
 275 }
 276
 277 int write_sha1_file(char *buf, unsigned long len, const char *type, unsigned char *returnsha1)
 278 {
 279         int size;
 280         char *compressed;
 281         z_stream stream;
 282         unsigned char sha1[20];
 283         SHA_CTX c;
 284         char *filename;
 285         static char tmpfile[PATH_MAX];
 286         char hdr[50];
 287         int fd, hdrlen, ret;
 288
 289         /* Generate the header */
 290         hdrlen = sprintf(hdr, "%s %lu", type, len)+1;
 291
 292         /* Sha1.. */
 293         SHA1_Init(&c);
 294         SHA1_Update(&c, hdr, hdrlen);
 295         SHA1_Update(&c, buf, len);
 296         SHA1_Final(sha1, &c);
 297
 298         if (returnsha1)
 299                 memcpy(returnsha1, sha1, 20);
 300
 301         filename = sha1_file_name(sha1);
 302         fd = open(filename, O_RDONLY);
 303         if (fd >= 0) {
 304                 /*
 305                  * FIXME!!! We might do collision checking here, but we'd
 306                  * need to uncompress the old file and check it. Later.
 307                  */
 308                 close(fd);
 309                 return 0;
 310         }
 311
 312         if (errno != ENOENT) {
 313                 fprintf(stderr, "sha1 file %s: %s", filename, strerror(errno));
 314                 return -1;
 315         }
 316
 317         snprintf(tmpfile, sizeof(tmpfile), "%s/obj_XXXXXX", get_object_directory());
 318         fd = mkstemp(tmpfile);
 319         if (fd < 0) {
 320                 fprintf(stderr, "unable to create temporary sha1 filename %s: %s", tmpfile, strerror(errno));
 321                 return -1;
 322         }
 323
 324         /* Set it up */
 325         memset(&stream, 0, sizeof(stream));
 326         deflateInit(&stream, Z_BEST_COMPRESSION);
 327         size = deflateBound(&stream, len+hdrlen);
 328         compressed = xmalloc(size);
 329
 330         /* Compress it */
 331         stream.next_out = compressed;
 332         stream.avail_out = size;
 333
 334         /* First header.. */
 335         stream.next_in = hdr;
 336         stream.avail_in = hdrlen;
 337         while (deflate(&stream, 0) == Z_OK)
 338                 /* nothing */;
 339
 340         /* Then the data itself.. */
 341         stream.next_in = buf;
 342         stream.avail_in = len;
 343         while (deflate(&stream, Z_FINISH) == Z_OK)
 344                 /* nothing */;
 345         deflateEnd(&stream);
 346         size = stream.total_out;
 347
 348         if (write(fd, compressed, size) != size)
 349                 die("unable to write file");
 350         fchmod(fd, 0444);
 351         close(fd);
 352
 353         ret = link(tmpfile, filename);
 354         if (ret < 0) {
 355                 ret = errno;
 356
 357                 /*
 358                  * Coda hack - coda doesn't like cross-directory links,
 359                  * so we fall back to a rename, which will mean that it
 360                  * won't be able to check collisions, but that's not a
 361                  * big deal.
 362                  *
 363                  * When this succeeds, we just return 0. We have nothing
 364                  * left to unlink.
 365                  */
 366                 if (ret == EXDEV && !rename(tmpfile, filename))
 367                         return 0;
 368         }
 369         unlink(tmpfile);
 370         if (ret) {
 371                 if (ret != EEXIST) {
 372                         fprintf(stderr, "unable to write sha1 filename %s: %s", filename, strerror(ret));
 373                         return -1;
 374                 }
 375                 /* FIXME!!! Collision check here ? */
 376         }
 377
 378         return 0;
 379 }
 380
 381 int write_sha1_from_fd(const unsigned char *sha1, int fd)
 382 {
 383         char *filename = sha1_file_name(sha1);
 384
 385         int local;
 386         z_stream stream;
 387         unsigned char real_sha1[20];
 388         char buf[4096];
 389         char discard[4096];
 390         int ret;
 391         SHA_CTX c;
 392
 393         local = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0666);
 394
 395         if (local < 0)
 396                 return error("Couldn't open %s\n", filename);
 397
 398         memset(&stream, 0, sizeof(stream));
 399
 400         inflateInit(&stream);
 401
 402         SHA1_Init(&c);
 403
 404         do {
 405                 ssize_t size;
 406                 size = read(fd, buf, 4096);
 407                 if (size <= 0) {
 408                         close(local);
 409                         unlink(filename);
 410                         if (!size)
 411                                 return error("Connection closed?");
 412                         perror("Reading from connection");
 413                         return -1;
 414                 }
 415                 write(local, buf, size);
 416                 stream.avail_in = size;
 417                 stream.next_in = buf;
 418                 do {
 419                         stream.next_out = discard;
 420                         stream.avail_out = sizeof(discard);
 421                         ret = inflate(&stream, Z_SYNC_FLUSH);
 422                         SHA1_Update(&c, discard, sizeof(discard) -
 423                                     stream.avail_out);
 424                 } while (stream.avail_in && ret == Z_OK);
 425
 426         } while (ret == Z_OK);
 427         inflateEnd(&stream);
 428
 429         close(local);
 430         SHA1_Final(real_sha1, &c);
 431         if (ret != Z_STREAM_END) {
 432                 unlink(filename);
 433                 return error("File %s corrupted", sha1_to_hex(sha1));
 434         }
 435         if (memcmp(sha1, real_sha1, 20)) {
 436                 unlink(filename);
 437                 return error("File %s has bad hash\n", sha1_to_hex(sha1));
 438         }
 439
 440         return 0;
 441 }
 442
 443 int has_sha1_file(const unsigned char *sha1)
 444 {
 445         char *filename = sha1_file_name(sha1);
 446         struct stat st;
 447
 448         if (!stat(filename, &st))
 449                 return 1;
 450         return 0;
 451 }
 452
 453 int index_fd(unsigned char *sha1, int fd, struct stat *st)
 454 {
 455         unsigned long size = st->st_size;
 456         void *buf;
 457         int ret;
 458
 459         buf = "";
 460         if (size)
 461                 buf = mmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
 462         close(fd);
 463         if ((int)(long)buf == -1)
 464                 return -1;
 465
 466         ret = write_sha1_file(buf, size, "blob", sha1);
 467         if (size)
 468                 munmap(buf, size);
 469         return ret;
 470 }