archive-tar.c

   1 /*
   2  * Copyright (c) 2005, 2006 Rene Scharfe
   3  */
   4 #include "git-compat-util.h"
   5 #include "alloc.h"
   6 #include "config.h"
   7 #include "tar.h"
   8 #include "archive.h"
   9 #include "object-store.h"
  10 #include "streaming.h"
  11 #include "run-command.h"
  12
  13 #define RECORDSIZE      (512)
  14 #define BLOCKSIZE       (RECORDSIZE * 20)
  15
  16 static char block[BLOCKSIZE];
  17 static unsigned long offset;
  18
  19 static int tar_umask = 002;
  20
  21 static int write_tar_filter_archive(const struct archiver *ar,
  22                                     struct archiver_args *args);
  23
  24 /*
  25  * This is the max value that a ustar size header can specify, as it is fixed
  26  * at 11 octal digits. POSIX specifies that we switch to extended headers at
  27  * this size.
  28  *
  29  * Likewise for the mtime (which happens to use a buffer of the same size).
  30  */
  31 #if ULONG_MAX == 0xFFFFFFFF
  32 #define USTAR_MAX_SIZE ULONG_MAX
  33 #else
  34 #define USTAR_MAX_SIZE 077777777777UL
  35 #endif
  36 #if TIME_MAX == 0xFFFFFFFF
  37 #define USTAR_MAX_MTIME TIME_MAX
  38 #else
  39 #define USTAR_MAX_MTIME 077777777777ULL
  40 #endif
  41
  42 static void tar_write_block(const void *buf)
  43 {
  44         write_or_die(1, buf, BLOCKSIZE);
  45 }
  46
  47 static void (*write_block)(const void *) = tar_write_block;
  48
  49 /* writes out the whole block, but only if it is full */
  50 static void write_if_needed(void)
  51 {
  52         if (offset == BLOCKSIZE) {
  53                 write_block(block);
  54                 offset = 0;
  55         }
  56 }
  57
  58 /*
  59  * queues up writes, so that all our write(2) calls write exactly one
  60  * full block; pads writes to RECORDSIZE
  61  */
  62 static void do_write_blocked(const void *data, unsigned long size)
  63 {
  64         const char *buf = data;
  65
  66         if (offset) {
  67                 unsigned long chunk = BLOCKSIZE - offset;
  68                 if (size < chunk)
  69                         chunk = size;
  70                 memcpy(block + offset, buf, chunk);
  71                 size -= chunk;
  72                 offset += chunk;
  73                 buf += chunk;
  74                 write_if_needed();
  75         }
  76         while (size >= BLOCKSIZE) {
  77                 write_block(buf);
  78                 size -= BLOCKSIZE;
  79                 buf += BLOCKSIZE;
  80         }
  81         if (size) {
  82                 memcpy(block + offset, buf, size);
  83                 offset += size;
  84         }
  85 }
  86
  87 static void finish_record(void)
  88 {
  89         unsigned long tail;
  90         tail = offset % RECORDSIZE;
  91         if (tail)  {
  92                 memset(block + offset, 0, RECORDSIZE - tail);
  93                 offset += RECORDSIZE - tail;
  94         }
  95         write_if_needed();
  96 }
  97
  98 static void write_blocked(const void *data, unsigned long size)
  99 {
 100         do_write_blocked(data, size);
 101         finish_record();
 102 }
 103
 104 /*
 105  * The end of tar archives is marked by 2*512 nul bytes and after that
 106  * follows the rest of the block (if any).
 107  */
 108 static void write_trailer(void)
 109 {
 110         int tail = BLOCKSIZE - offset;
 111         memset(block + offset, 0, tail);
 112         write_block(block);
 113         if (tail < 2 * RECORDSIZE) {
 114                 memset(block, 0, offset);
 115                 write_block(block);
 116         }
 117 }
 118
 119 /*
 120  * queues up writes, so that all our write(2) calls write exactly one
 121  * full block; pads writes to RECORDSIZE
 122  */
 123 static int stream_blocked(struct repository *r, const struct object_id *oid)
 124 {
 125         struct git_istream *st;
 126         enum object_type type;
 127         unsigned long sz;
 128         char buf[BLOCKSIZE];
 129         ssize_t readlen;
 130
 131         st = open_istream(r, oid, &type, &sz, NULL);
 132         if (!st)
 133                 return error(_("cannot stream blob %s"), oid_to_hex(oid));
 134         for (;;) {
 135                 readlen = read_istream(st, buf, sizeof(buf));
 136                 if (readlen <= 0)
 137                         break;
 138                 do_write_blocked(buf, readlen);
 139         }
 140         close_istream(st);
 141         if (!readlen)
 142                 finish_record();
 143         return readlen;
 144 }
 145
 146 /*
 147  * pax extended header records have the format "%u %s=%s\n".  %u contains
 148  * the size of the whole string (including the %u), the first %s is the
 149  * keyword, the second one is the value.  This function constructs such a
 150  * string and appends it to a struct strbuf.
 151  */
 152 static void strbuf_append_ext_header(struct strbuf *sb, const char *keyword,
 153                                      const char *value, size_t valuelen)
 154 {
 155         size_t orig_len = sb->len;
 156         size_t len, tmp;
 157
 158         /* "%u %s=%s\n" */
 159         len = 1 + 1 + strlen(keyword) + 1 + valuelen + 1;
 160         for (tmp = 1; len / 10 >= tmp; tmp *= 10)
 161                 len++;
 162
 163         strbuf_grow(sb, len);
 164         strbuf_addf(sb, "%"PRIuMAX" %s=", (uintmax_t)len, keyword);
 165         strbuf_add(sb, value, valuelen);
 166         strbuf_addch(sb, '\n');
 167
 168         if (len != sb->len - orig_len)
 169                 BUG("pax extended header length miscalculated as %"PRIuMAX
 170                     ", should be %"PRIuMAX,
 171                     (uintmax_t)len, (uintmax_t)(sb->len - orig_len));
 172 }
 173
 174 /*
 175  * Like strbuf_append_ext_header, but for numeric values.
 176  */
 177 static void strbuf_append_ext_header_uint(struct strbuf *sb,
 178                                           const char *keyword,
 179                                           uintmax_t value)
 180 {
 181         char buf[40]; /* big enough for 2^128 in decimal, plus NUL */
 182         int len;
 183
 184         len = xsnprintf(buf, sizeof(buf), "%"PRIuMAX, value);
 185         strbuf_append_ext_header(sb, keyword, buf, len);
 186 }
 187
 188 static unsigned int ustar_header_chksum(const struct ustar_header *header)
 189 {
 190         const unsigned char *p = (const unsigned char *)header;
 191         unsigned int chksum = 0;
 192         while (p < (const unsigned char *)header->chksum)
 193                 chksum += *p++;
 194         chksum += sizeof(header->chksum) * ' ';
 195         p += sizeof(header->chksum);
 196         while (p < (const unsigned char *)header + sizeof(struct ustar_header))
 197                 chksum += *p++;
 198         return chksum;
 199 }
 200
 201 static size_t get_path_prefix(const char *path, size_t pathlen, size_t maxlen)
 202 {
 203         size_t i = pathlen;
 204         if (i > 1 && path[i - 1] == '/')
 205                 i--;
 206         if (i > maxlen)
 207                 i = maxlen;
 208         do {
 209                 i--;
 210         } while (i > 0 && path[i] != '/');
 211         return i;
 212 }
 213
 214 static void prepare_header(struct archiver_args *args,
 215                            struct ustar_header *header,
 216                            unsigned int mode, unsigned long size)
 217 {
 218         xsnprintf(header->mode, sizeof(header->mode), "%07o", mode & 07777);
 219         xsnprintf(header->size, sizeof(header->size), "%011"PRIoMAX , S_ISREG(mode) ? (uintmax_t)size : (uintmax_t)0);
 220         xsnprintf(header->mtime, sizeof(header->mtime), "%011lo", (unsigned long) args->time);
 221
 222         xsnprintf(header->uid, sizeof(header->uid), "%07o", 0);
 223         xsnprintf(header->gid, sizeof(header->gid), "%07o", 0);
 224         strlcpy(header->uname, "root", sizeof(header->uname));
 225         strlcpy(header->gname, "root", sizeof(header->gname));
 226         xsnprintf(header->devmajor, sizeof(header->devmajor), "%07o", 0);
 227         xsnprintf(header->devminor, sizeof(header->devminor), "%07o", 0);
 228
 229         memcpy(header->magic, "ustar", 6);
 230         memcpy(header->version, "00", 2);
 231
 232         xsnprintf(header->chksum, sizeof(header->chksum), "%07o", ustar_header_chksum(header));
 233 }
 234
 235 static void write_extended_header(struct archiver_args *args,
 236                                   const struct object_id *oid,
 237                                   const void *buffer, unsigned long size)
 238 {
 239         struct ustar_header header;
 240         unsigned int mode;
 241         memset(&header, 0, sizeof(header));
 242         *header.typeflag = TYPEFLAG_EXT_HEADER;
 243         mode = 0100666;
 244         xsnprintf(header.name, sizeof(header.name), "%s.paxheader", oid_to_hex(oid));
 245         prepare_header(args, &header, mode, size);
 246         write_blocked(&header, sizeof(header));
 247         write_blocked(buffer, size);
 248 }
 249
 250 static int write_tar_entry(struct archiver_args *args,
 251                            const struct object_id *oid,
 252                            const char *path, size_t pathlen,
 253                            unsigned int mode,
 254                            void *buffer, unsigned long size)
 255 {
 256         struct ustar_header header;
 257         struct strbuf ext_header = STRBUF_INIT;
 258         unsigned long size_in_header;
 259         int err = 0;
 260
 261         memset(&header, 0, sizeof(header));
 262
 263         if (S_ISDIR(mode) || S_ISGITLINK(mode)) {
 264                 *header.typeflag = TYPEFLAG_DIR;
 265                 mode = (mode | 0777) & ~tar_umask;
 266         } else if (S_ISLNK(mode)) {
 267                 *header.typeflag = TYPEFLAG_LNK;
 268                 mode |= 0777;
 269         } else if (S_ISREG(mode)) {
 270                 *header.typeflag = TYPEFLAG_REG;
 271                 mode = (mode | ((mode & 0100) ? 0777 : 0666)) & ~tar_umask;
 272         } else {
 273                 return error(_("unsupported file mode: 0%o (SHA1: %s)"),
 274                              mode, oid_to_hex(oid));
 275         }
 276         if (pathlen > sizeof(header.name)) {
 277                 size_t plen = get_path_prefix(path, pathlen,
 278                                               sizeof(header.prefix));
 279                 size_t rest = pathlen - plen - 1;
 280                 if (plen > 0 && rest <= sizeof(header.name)) {
 281                         memcpy(header.prefix, path, plen);
 282                         memcpy(header.name, path + plen + 1, rest);
 283                 } else {
 284                         xsnprintf(header.name, sizeof(header.name), "%s.data",
 285                                   oid_to_hex(oid));
 286                         strbuf_append_ext_header(&ext_header, "path",
 287                                                  path, pathlen);
 288                 }
 289         } else
 290                 memcpy(header.name, path, pathlen);
 291
 292         if (S_ISLNK(mode)) {
 293                 if (size > sizeof(header.linkname)) {
 294                         xsnprintf(header.linkname, sizeof(header.linkname),
 295                                   "see %s.paxheader", oid_to_hex(oid));
 296                         strbuf_append_ext_header(&ext_header, "linkpath",
 297                                                  buffer, size);
 298                 } else
 299                         memcpy(header.linkname, buffer, size);
 300         }
 301
 302         size_in_header = size;
 303         if (S_ISREG(mode) && size > USTAR_MAX_SIZE) {
 304                 size_in_header = 0;
 305                 strbuf_append_ext_header_uint(&ext_header, "size", size);
 306         }
 307
 308         prepare_header(args, &header, mode, size_in_header);
 309
 310         if (ext_header.len > 0) {
 311                 write_extended_header(args, oid, ext_header.buf,
 312                                       ext_header.len);
 313         }
 314         strbuf_release(&ext_header);
 315         write_blocked(&header, sizeof(header));
 316         if (S_ISREG(mode) && size > 0) {
 317                 if (buffer)
 318                         write_blocked(buffer, size);
 319                 else
 320                         err = stream_blocked(args->repo, oid);
 321         }
 322         return err;
 323 }
 324
 325 static void write_global_extended_header(struct archiver_args *args)
 326 {
 327         const struct object_id *oid = args->commit_oid;
 328         struct strbuf ext_header = STRBUF_INIT;
 329         struct ustar_header header;
 330         unsigned int mode;
 331
 332         if (oid)
 333                 strbuf_append_ext_header(&ext_header, "comment",
 334                                          oid_to_hex(oid),
 335                                          the_hash_algo->hexsz);
 336         if (args->time > USTAR_MAX_MTIME) {
 337                 strbuf_append_ext_header_uint(&ext_header, "mtime",
 338                                               args->time);
 339                 args->time = USTAR_MAX_MTIME;
 340         }
 341
 342         if (!ext_header.len)
 343                 return;
 344
 345         memset(&header, 0, sizeof(header));
 346         *header.typeflag = TYPEFLAG_GLOBAL_HEADER;
 347         mode = 0100666;
 348         xsnprintf(header.name, sizeof(header.name), "pax_global_header");
 349         prepare_header(args, &header, mode, ext_header.len);
 350         write_blocked(&header, sizeof(header));
 351         write_blocked(ext_header.buf, ext_header.len);
 352         strbuf_release(&ext_header);
 353 }
 354
 355 static struct archiver **tar_filters;
 356 static int nr_tar_filters;
 357 static int alloc_tar_filters;
 358
 359 static struct archiver *find_tar_filter(const char *name, size_t len)
 360 {
 361         int i;
 362         for (i = 0; i < nr_tar_filters; i++) {
 363                 struct archiver *ar = tar_filters[i];
 364                 if (!strncmp(ar->name, name, len) && !ar->name[len])
 365                         return ar;
 366         }
 367         return NULL;
 368 }
 369
 370 static int tar_filter_config(const char *var, const char *value,
 371                              void *data UNUSED)
 372 {
 373         struct archiver *ar;
 374         const char *name;
 375         const char *type;
 376         size_t namelen;
 377
 378         if (parse_config_key(var, "tar", &name, &namelen, &type) < 0 || !name)
 379                 return 0;
 380
 381         ar = find_tar_filter(name, namelen);
 382         if (!ar) {
 383                 CALLOC_ARRAY(ar, 1);
 384                 ar->name = xmemdupz(name, namelen);
 385                 ar->write_archive = write_tar_filter_archive;
 386                 ar->flags = ARCHIVER_WANT_COMPRESSION_LEVELS |
 387                             ARCHIVER_HIGH_COMPRESSION_LEVELS;
 388                 ALLOC_GROW(tar_filters, nr_tar_filters + 1, alloc_tar_filters);
 389                 tar_filters[nr_tar_filters++] = ar;
 390         }
 391
 392         if (!strcmp(type, "command")) {
 393                 if (!value)
 394                         return config_error_nonbool(var);
 395                 free(ar->filter_command);
 396                 ar->filter_command = xstrdup(value);
 397                 return 0;
 398         }
 399         if (!strcmp(type, "remote")) {
 400                 if (git_config_bool(var, value))
 401                         ar->flags |= ARCHIVER_REMOTE;
 402                 else
 403                         ar->flags &= ~ARCHIVER_REMOTE;
 404                 return 0;
 405         }
 406
 407         return 0;
 408 }
 409
 410 static int git_tar_config(const char *var, const char *value, void *cb)
 411 {
 412         if (!strcmp(var, "tar.umask")) {
 413                 if (value && !strcmp(value, "user")) {
 414                         tar_umask = umask(0);
 415                         umask(tar_umask);
 416                 } else {
 417                         tar_umask = git_config_int(var, value);
 418                 }
 419                 return 0;
 420         }
 421
 422         return tar_filter_config(var, value, cb);
 423 }
 424
 425 static int write_tar_archive(const struct archiver *ar UNUSED,
 426                              struct archiver_args *args)
 427 {
 428         int err = 0;
 429
 430         write_global_extended_header(args);
 431         err = write_archive_entries(args, write_tar_entry);
 432         if (!err)
 433                 write_trailer();
 434         return err;
 435 }
 436
 437 static git_zstream gzstream;
 438 static unsigned char outbuf[16384];
 439
 440 static void tgz_deflate(int flush)
 441 {
 442         while (gzstream.avail_in || flush == Z_FINISH) {
 443                 int status = git_deflate(&gzstream, flush);
 444                 if (!gzstream.avail_out || status == Z_STREAM_END) {
 445                         write_or_die(1, outbuf, gzstream.next_out - outbuf);
 446                         gzstream.next_out = outbuf;
 447                         gzstream.avail_out = sizeof(outbuf);
 448                         if (status == Z_STREAM_END)
 449                                 break;
 450                 }
 451                 if (status != Z_OK && status != Z_BUF_ERROR)
 452                         die(_("deflate error (%d)"), status);
 453         }
 454 }
 455
 456 static void tgz_write_block(const void *data)
 457 {
 458         gzstream.next_in = (void *)data;
 459         gzstream.avail_in = BLOCKSIZE;
 460         tgz_deflate(Z_NO_FLUSH);
 461 }
 462
 463 static const char internal_gzip_command[] = "git archive gzip";
 464
 465 static int write_tar_filter_archive(const struct archiver *ar,
 466                                     struct archiver_args *args)
 467 {
 468 #if ZLIB_VERNUM >= 0x1221
 469         struct gz_header_s gzhead = { .os = 3 }; /* Unix, for reproducibility */
 470 #endif
 471         struct strbuf cmd = STRBUF_INIT;
 472         struct child_process filter = CHILD_PROCESS_INIT;
 473         int r;
 474
 475         if (!ar->filter_command)
 476                 BUG("tar-filter archiver called with no filter defined");
 477
 478         if (!strcmp(ar->filter_command, internal_gzip_command)) {
 479                 write_block = tgz_write_block;
 480                 git_deflate_init_gzip(&gzstream, args->compression_level);
 481 #if ZLIB_VERNUM >= 0x1221
 482                 if (deflateSetHeader(&gzstream.z, &gzhead) != Z_OK)
 483                         BUG("deflateSetHeader() called too late");
 484 #endif
 485                 gzstream.next_out = outbuf;
 486                 gzstream.avail_out = sizeof(outbuf);
 487
 488                 r = write_tar_archive(ar, args);
 489
 490                 tgz_deflate(Z_FINISH);
 491                 git_deflate_end(&gzstream);
 492                 return r;
 493         }
 494
 495         strbuf_addstr(&cmd, ar->filter_command);
 496         if (args->compression_level >= 0)
 497                 strbuf_addf(&cmd, " -%d", args->compression_level);
 498
 499         strvec_push(&filter.args, cmd.buf);
 500         filter.use_shell = 1;
 501         filter.in = -1;
 502         filter.silent_exec_failure = 1;
 503
 504         if (start_command(&filter) < 0)
 505                 die_errno(_("unable to start '%s' filter"), cmd.buf);
 506         close(1);
 507         if (dup2(filter.in, 1) < 0)
 508                 die_errno(_("unable to redirect descriptor"));
 509         close(filter.in);
 510
 511         r = write_tar_archive(ar, args);
 512
 513         close(1);
 514         if (finish_command(&filter) != 0)
 515                 die(_("'%s' filter reported error"), cmd.buf);
 516
 517         strbuf_release(&cmd);
 518         return r;
 519 }
 520
 521 static struct archiver tar_archiver = {
 522         .name = "tar",
 523         .write_archive = write_tar_archive,
 524         .flags = ARCHIVER_REMOTE,
 525 };
 526
 527 void init_tar_archiver(void)
 528 {
 529         int i;
 530         register_archiver(&tar_archiver);
 531
 532         tar_filter_config("tar.tgz.command", internal_gzip_command, NULL);
 533         tar_filter_config("tar.tgz.remote", "true", NULL);
 534         tar_filter_config("tar.tar.gz.command", internal_gzip_command, NULL);
 535         tar_filter_config("tar.tar.gz.remote", "true", NULL);
 536         git_config(git_tar_config, NULL);
 537         for (i = 0; i < nr_tar_filters; i++) {
 538                 /* omit any filters that never had a command configured */
 539                 if (tar_filters[i]->filter_command)
 540                         register_archiver(tar_filters[i]);
 541         }
 542 }