archive-tar.c

   1 /*
   2  * Copyright (c) 2005, 2006 Rene Scharfe
   3  */
   4 #include "cache.h"
   5 #include "config.h"
   6 #include "tar.h"
   7 #include "archive.h"
   8 #include "object-store.h"
   9 #include "streaming.h"
  10 #include "run-command.h"
  11
  12 #define RECORDSIZE      (512)
  13 #define BLOCKSIZE       (RECORDSIZE * 20)
  14
  15 static char block[BLOCKSIZE];
  16 static unsigned long offset;
  17
  18 static int tar_umask = 002;
  19
  20 static int write_tar_filter_archive(const struct archiver *ar,
  21                                     struct archiver_args *args);
  22
  23 /*
  24  * This is the max value that a ustar size header can specify, as it is fixed
  25  * at 11 octal digits. POSIX specifies that we switch to extended headers at
  26  * this size.
  27  *
  28  * Likewise for the mtime (which happens to use a buffer of the same size).
  29  */
  30 #if ULONG_MAX == 0xFFFFFFFF
  31 #define USTAR_MAX_SIZE ULONG_MAX
  32 #else
  33 #define USTAR_MAX_SIZE 077777777777UL
  34 #endif
  35 #if TIME_MAX == 0xFFFFFFFF
  36 #define USTAR_MAX_MTIME TIME_MAX
  37 #else
  38 #define USTAR_MAX_MTIME 077777777777ULL
  39 #endif
  40
  41 static void tar_write_block(const void *buf)
  42 {
  43         write_or_die(1, buf, BLOCKSIZE);
  44 }
  45
  46 static void (*write_block)(const void *) = tar_write_block;
  47
  48 /* writes out the whole block, but only if it is full */
  49 static void write_if_needed(void)
  50 {
  51         if (offset == BLOCKSIZE) {
  52                 write_block(block);
  53                 offset = 0;
  54         }
  55 }
  56
  57 /*
  58  * queues up writes, so that all our write(2) calls write exactly one
  59  * full block; pads writes to RECORDSIZE
  60  */
  61 static void do_write_blocked(const void *data, unsigned long size)
  62 {
  63         const char *buf = data;
  64
  65         if (offset) {
  66                 unsigned long chunk = BLOCKSIZE - offset;
  67                 if (size < chunk)
  68                         chunk = size;
  69                 memcpy(block + offset, buf, chunk);
  70                 size -= chunk;
  71                 offset += chunk;
  72                 buf += chunk;
  73                 write_if_needed();
  74         }
  75         while (size >= BLOCKSIZE) {
  76                 write_block(buf);
  77                 size -= BLOCKSIZE;
  78                 buf += BLOCKSIZE;
  79         }
  80         if (size) {
  81                 memcpy(block + offset, buf, size);
  82                 offset += size;
  83         }
  84 }
  85
  86 static void finish_record(void)
  87 {
  88         unsigned long tail;
  89         tail = offset % RECORDSIZE;
  90         if (tail)  {
  91                 memset(block + offset, 0, RECORDSIZE - tail);
  92                 offset += RECORDSIZE - tail;
  93         }
  94         write_if_needed();
  95 }
  96
  97 static void write_blocked(const void *data, unsigned long size)
  98 {
  99         do_write_blocked(data, size);
 100         finish_record();
 101 }
 102
 103 /*
 104  * The end of tar archives is marked by 2*512 nul bytes and after that
 105  * follows the rest of the block (if any).
 106  */
 107 static void write_trailer(void)
 108 {
 109         int tail = BLOCKSIZE - offset;
 110         memset(block + offset, 0, tail);
 111         write_block(block);
 112         if (tail < 2 * RECORDSIZE) {
 113                 memset(block, 0, offset);
 114                 write_block(block);
 115         }
 116 }
 117
 118 /*
 119  * queues up writes, so that all our write(2) calls write exactly one
 120  * full block; pads writes to RECORDSIZE
 121  */
 122 static int stream_blocked(struct repository *r, const struct object_id *oid)
 123 {
 124         struct git_istream *st;
 125         enum object_type type;
 126         unsigned long sz;
 127         char buf[BLOCKSIZE];
 128         ssize_t readlen;
 129
 130         st = open_istream(r, oid, &type, &sz, NULL);
 131         if (!st)
 132                 return error(_("cannot stream blob %s"), oid_to_hex(oid));
 133         for (;;) {
 134                 readlen = read_istream(st, buf, sizeof(buf));
 135                 if (readlen <= 0)
 136                         break;
 137                 do_write_blocked(buf, readlen);
 138         }
 139         close_istream(st);
 140         if (!readlen)
 141                 finish_record();
 142         return readlen;
 143 }
 144
 145 /*
 146  * pax extended header records have the format "%u %s=%s\n".  %u contains
 147  * the size of the whole string (including the %u), the first %s is the
 148  * keyword, the second one is the value.  This function constructs such a
 149  * string and appends it to a struct strbuf.
 150  */
 151 static void strbuf_append_ext_header(struct strbuf *sb, const char *keyword,
 152                                      const char *value, size_t valuelen)
 153 {
 154         size_t orig_len = sb->len;
 155         size_t len, tmp;
 156
 157         /* "%u %s=%s\n" */
 158         len = 1 + 1 + strlen(keyword) + 1 + valuelen + 1;
 159         for (tmp = 1; len / 10 >= tmp; tmp *= 10)
 160                 len++;
 161
 162         strbuf_grow(sb, len);
 163         strbuf_addf(sb, "%"PRIuMAX" %s=", (uintmax_t)len, keyword);
 164         strbuf_add(sb, value, valuelen);
 165         strbuf_addch(sb, '\n');
 166
 167         if (len != sb->len - orig_len)
 168                 BUG("pax extended header length miscalculated as %"PRIuMAX
 169                     ", should be %"PRIuMAX,
 170                     (uintmax_t)len, (uintmax_t)(sb->len - orig_len));
 171 }
 172
 173 /*
 174  * Like strbuf_append_ext_header, but for numeric values.
 175  */
 176 static void strbuf_append_ext_header_uint(struct strbuf *sb,
 177                                           const char *keyword,
 178                                           uintmax_t value)
 179 {
 180         char buf[40]; /* big enough for 2^128 in decimal, plus NUL */
 181         int len;
 182
 183         len = xsnprintf(buf, sizeof(buf), "%"PRIuMAX, value);
 184         strbuf_append_ext_header(sb, keyword, buf, len);
 185 }
 186
 187 static unsigned int ustar_header_chksum(const struct ustar_header *header)
 188 {
 189         const unsigned char *p = (const unsigned char *)header;
 190         unsigned int chksum = 0;
 191         while (p < (const unsigned char *)header->chksum)
 192                 chksum += *p++;
 193         chksum += sizeof(header->chksum) * ' ';
 194         p += sizeof(header->chksum);
 195         while (p < (const unsigned char *)header + sizeof(struct ustar_header))
 196                 chksum += *p++;
 197         return chksum;
 198 }
 199
 200 static size_t get_path_prefix(const char *path, size_t pathlen, size_t maxlen)
 201 {
 202         size_t i = pathlen;
 203         if (i > 1 && path[i - 1] == '/')
 204                 i--;
 205         if (i > maxlen)
 206                 i = maxlen;
 207         do {
 208                 i--;
 209         } while (i > 0 && path[i] != '/');
 210         return i;
 211 }
 212
 213 static void prepare_header(struct archiver_args *args,
 214                            struct ustar_header *header,
 215                            unsigned int mode, unsigned long size)
 216 {
 217         xsnprintf(header->mode, sizeof(header->mode), "%07o", mode & 07777);
 218         xsnprintf(header->size, sizeof(header->size), "%011"PRIoMAX , S_ISREG(mode) ? (uintmax_t)size : (uintmax_t)0);
 219         xsnprintf(header->mtime, sizeof(header->mtime), "%011lo", (unsigned long) args->time);
 220
 221         xsnprintf(header->uid, sizeof(header->uid), "%07o", 0);
 222         xsnprintf(header->gid, sizeof(header->gid), "%07o", 0);
 223         strlcpy(header->uname, "root", sizeof(header->uname));
 224         strlcpy(header->gname, "root", sizeof(header->gname));
 225         xsnprintf(header->devmajor, sizeof(header->devmajor), "%07o", 0);
 226         xsnprintf(header->devminor, sizeof(header->devminor), "%07o", 0);
 227
 228         memcpy(header->magic, "ustar", 6);
 229         memcpy(header->version, "00", 2);
 230
 231         xsnprintf(header->chksum, sizeof(header->chksum), "%07o", ustar_header_chksum(header));
 232 }
 233
 234 static void write_extended_header(struct archiver_args *args,
 235                                   const struct object_id *oid,
 236                                   const void *buffer, unsigned long size)
 237 {
 238         struct ustar_header header;
 239         unsigned int mode;
 240         memset(&header, 0, sizeof(header));
 241         *header.typeflag = TYPEFLAG_EXT_HEADER;
 242         mode = 0100666;
 243         xsnprintf(header.name, sizeof(header.name), "%s.paxheader", oid_to_hex(oid));
 244         prepare_header(args, &header, mode, size);
 245         write_blocked(&header, sizeof(header));
 246         write_blocked(buffer, size);
 247 }
 248
 249 static int write_tar_entry(struct archiver_args *args,
 250                            const struct object_id *oid,
 251                            const char *path, size_t pathlen,
 252                            unsigned int mode,
 253                            void *buffer, unsigned long size)
 254 {
 255         struct ustar_header header;
 256         struct strbuf ext_header = STRBUF_INIT;
 257         unsigned long size_in_header;
 258         int err = 0;
 259
 260         memset(&header, 0, sizeof(header));
 261
 262         if (S_ISDIR(mode) || S_ISGITLINK(mode)) {
 263                 *header.typeflag = TYPEFLAG_DIR;
 264                 mode = (mode | 0777) & ~tar_umask;
 265         } else if (S_ISLNK(mode)) {
 266                 *header.typeflag = TYPEFLAG_LNK;
 267                 mode |= 0777;
 268         } else if (S_ISREG(mode)) {
 269                 *header.typeflag = TYPEFLAG_REG;
 270                 mode = (mode | ((mode & 0100) ? 0777 : 0666)) & ~tar_umask;
 271         } else {
 272                 return error(_("unsupported file mode: 0%o (SHA1: %s)"),
 273                              mode, oid_to_hex(oid));
 274         }
 275         if (pathlen > sizeof(header.name)) {
 276                 size_t plen = get_path_prefix(path, pathlen,
 277                                               sizeof(header.prefix));
 278                 size_t rest = pathlen - plen - 1;
 279                 if (plen > 0 && rest <= sizeof(header.name)) {
 280                         memcpy(header.prefix, path, plen);
 281                         memcpy(header.name, path + plen + 1, rest);
 282                 } else {
 283                         xsnprintf(header.name, sizeof(header.name), "%s.data",
 284                                   oid_to_hex(oid));
 285                         strbuf_append_ext_header(&ext_header, "path",
 286                                                  path, pathlen);
 287                 }
 288         } else
 289                 memcpy(header.name, path, pathlen);
 290
 291         if (S_ISLNK(mode)) {
 292                 if (size > sizeof(header.linkname)) {
 293                         xsnprintf(header.linkname, sizeof(header.linkname),
 294                                   "see %s.paxheader", oid_to_hex(oid));
 295                         strbuf_append_ext_header(&ext_header, "linkpath",
 296                                                  buffer, size);
 297                 } else
 298                         memcpy(header.linkname, buffer, size);
 299         }
 300
 301         size_in_header = size;
 302         if (S_ISREG(mode) && size > USTAR_MAX_SIZE) {
 303                 size_in_header = 0;
 304                 strbuf_append_ext_header_uint(&ext_header, "size", size);
 305         }
 306
 307         prepare_header(args, &header, mode, size_in_header);
 308
 309         if (ext_header.len > 0) {
 310                 write_extended_header(args, oid, ext_header.buf,
 311                                       ext_header.len);
 312         }
 313         strbuf_release(&ext_header);
 314         write_blocked(&header, sizeof(header));
 315         if (S_ISREG(mode) && size > 0) {
 316                 if (buffer)
 317                         write_blocked(buffer, size);
 318                 else
 319                         err = stream_blocked(args->repo, oid);
 320         }
 321         return err;
 322 }
 323
 324 static void write_global_extended_header(struct archiver_args *args)
 325 {
 326         const struct object_id *oid = args->commit_oid;
 327         struct strbuf ext_header = STRBUF_INIT;
 328         struct ustar_header header;
 329         unsigned int mode;
 330
 331         if (oid)
 332                 strbuf_append_ext_header(&ext_header, "comment",
 333                                          oid_to_hex(oid),
 334                                          the_hash_algo->hexsz);
 335         if (args->time > USTAR_MAX_MTIME) {
 336                 strbuf_append_ext_header_uint(&ext_header, "mtime",
 337                                               args->time);
 338                 args->time = USTAR_MAX_MTIME;
 339         }
 340
 341         if (!ext_header.len)
 342                 return;
 343
 344         memset(&header, 0, sizeof(header));
 345         *header.typeflag = TYPEFLAG_GLOBAL_HEADER;
 346         mode = 0100666;
 347         xsnprintf(header.name, sizeof(header.name), "pax_global_header");
 348         prepare_header(args, &header, mode, ext_header.len);
 349         write_blocked(&header, sizeof(header));
 350         write_blocked(ext_header.buf, ext_header.len);
 351         strbuf_release(&ext_header);
 352 }
 353
 354 static struct archiver **tar_filters;
 355 static int nr_tar_filters;
 356 static int alloc_tar_filters;
 357
 358 static struct archiver *find_tar_filter(const char *name, size_t len)
 359 {
 360         int i;
 361         for (i = 0; i < nr_tar_filters; i++) {
 362                 struct archiver *ar = tar_filters[i];
 363                 if (!strncmp(ar->name, name, len) && !ar->name[len])
 364                         return ar;
 365         }
 366         return NULL;
 367 }
 368
 369 static int tar_filter_config(const char *var, const char *value,
 370                              void *data UNUSED)
 371 {
 372         struct archiver *ar;
 373         const char *name;
 374         const char *type;
 375         size_t namelen;
 376
 377         if (parse_config_key(var, "tar", &name, &namelen, &type) < 0 || !name)
 378                 return 0;
 379
 380         ar = find_tar_filter(name, namelen);
 381         if (!ar) {
 382                 CALLOC_ARRAY(ar, 1);
 383                 ar->name = xmemdupz(name, namelen);
 384                 ar->write_archive = write_tar_filter_archive;
 385                 ar->flags = ARCHIVER_WANT_COMPRESSION_LEVELS |
 386                             ARCHIVER_HIGH_COMPRESSION_LEVELS;
 387                 ALLOC_GROW(tar_filters, nr_tar_filters + 1, alloc_tar_filters);
 388                 tar_filters[nr_tar_filters++] = ar;
 389         }
 390
 391         if (!strcmp(type, "command")) {
 392                 if (!value)
 393                         return config_error_nonbool(var);
 394                 free(ar->filter_command);
 395                 ar->filter_command = xstrdup(value);
 396                 return 0;
 397         }
 398         if (!strcmp(type, "remote")) {
 399                 if (git_config_bool(var, value))
 400                         ar->flags |= ARCHIVER_REMOTE;
 401                 else
 402                         ar->flags &= ~ARCHIVER_REMOTE;
 403                 return 0;
 404         }
 405
 406         return 0;
 407 }
 408
 409 static int git_tar_config(const char *var, const char *value, void *cb)
 410 {
 411         if (!strcmp(var, "tar.umask")) {
 412                 if (value && !strcmp(value, "user")) {
 413                         tar_umask = umask(0);
 414                         umask(tar_umask);
 415                 } else {
 416                         tar_umask = git_config_int(var, value);
 417                 }
 418                 return 0;
 419         }
 420
 421         return tar_filter_config(var, value, cb);
 422 }
 423
 424 static int write_tar_archive(const struct archiver *ar UNUSED,
 425                              struct archiver_args *args)
 426 {
 427         int err = 0;
 428
 429         write_global_extended_header(args);
 430         err = write_archive_entries(args, write_tar_entry);
 431         if (!err)
 432                 write_trailer();
 433         return err;
 434 }
 435
 436 static git_zstream gzstream;
 437 static unsigned char outbuf[16384];
 438
 439 static void tgz_deflate(int flush)
 440 {
 441         while (gzstream.avail_in || flush == Z_FINISH) {
 442                 int status = git_deflate(&gzstream, flush);
 443                 if (!gzstream.avail_out || status == Z_STREAM_END) {
 444                         write_or_die(1, outbuf, gzstream.next_out - outbuf);
 445                         gzstream.next_out = outbuf;
 446                         gzstream.avail_out = sizeof(outbuf);
 447                         if (status == Z_STREAM_END)
 448                                 break;
 449                 }
 450                 if (status != Z_OK && status != Z_BUF_ERROR)
 451                         die(_("deflate error (%d)"), status);
 452         }
 453 }
 454
 455 static void tgz_write_block(const void *data)
 456 {
 457         gzstream.next_in = (void *)data;
 458         gzstream.avail_in = BLOCKSIZE;
 459         tgz_deflate(Z_NO_FLUSH);
 460 }
 461
 462 static const char internal_gzip_command[] = "git archive gzip";
 463
 464 static int write_tar_filter_archive(const struct archiver *ar,
 465                                     struct archiver_args *args)
 466 {
 467 #if ZLIB_VERNUM >= 0x1221
 468         struct gz_header_s gzhead = { .os = 3 }; /* Unix, for reproducibility */
 469 #endif
 470         struct strbuf cmd = STRBUF_INIT;
 471         struct child_process filter = CHILD_PROCESS_INIT;
 472         int r;
 473
 474         if (!ar->filter_command)
 475                 BUG("tar-filter archiver called with no filter defined");
 476
 477         if (!strcmp(ar->filter_command, internal_gzip_command)) {
 478                 write_block = tgz_write_block;
 479                 git_deflate_init_gzip(&gzstream, args->compression_level);
 480 #if ZLIB_VERNUM >= 0x1221
 481                 if (deflateSetHeader(&gzstream.z, &gzhead) != Z_OK)
 482                         BUG("deflateSetHeader() called too late");
 483 #endif
 484                 gzstream.next_out = outbuf;
 485                 gzstream.avail_out = sizeof(outbuf);
 486
 487                 r = write_tar_archive(ar, args);
 488
 489                 tgz_deflate(Z_FINISH);
 490                 git_deflate_end(&gzstream);
 491                 return r;
 492         }
 493
 494         strbuf_addstr(&cmd, ar->filter_command);
 495         if (args->compression_level >= 0)
 496                 strbuf_addf(&cmd, " -%d", args->compression_level);
 497
 498         strvec_push(&filter.args, cmd.buf);
 499         filter.use_shell = 1;
 500         filter.in = -1;
 501
 502         if (start_command(&filter) < 0)
 503                 die_errno(_("unable to start '%s' filter"), cmd.buf);
 504         close(1);
 505         if (dup2(filter.in, 1) < 0)
 506                 die_errno(_("unable to redirect descriptor"));
 507         close(filter.in);
 508
 509         r = write_tar_archive(ar, args);
 510
 511         close(1);
 512         if (finish_command(&filter) != 0)
 513                 die(_("'%s' filter reported error"), cmd.buf);
 514
 515         strbuf_release(&cmd);
 516         return r;
 517 }
 518
 519 static struct archiver tar_archiver = {
 520         .name = "tar",
 521         .write_archive = write_tar_archive,
 522         .flags = ARCHIVER_REMOTE,
 523 };
 524
 525 void init_tar_archiver(void)
 526 {
 527         int i;
 528         register_archiver(&tar_archiver);
 529
 530         tar_filter_config("tar.tgz.command", internal_gzip_command, NULL);
 531         tar_filter_config("tar.tgz.remote", "true", NULL);
 532         tar_filter_config("tar.tar.gz.command", internal_gzip_command, NULL);
 533         tar_filter_config("tar.tar.gz.remote", "true", NULL);
 534         git_config(git_tar_config, NULL);
 535         for (i = 0; i < nr_tar_filters; i++) {
 536                 /* omit any filters that never had a command configured */
 537                 if (tar_filters[i]->filter_command)
 538                         register_archiver(tar_filters[i]);
 539         }
 540 }