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 "hex.h"
   8 #include "tar.h"
   9 #include "archive.h"
  10 #include "object-store.h"
  11 #include "streaming.h"
  12 #include "run-command.h"
  13
  14 #define RECORDSIZE      (512)
  15 #define BLOCKSIZE       (RECORDSIZE * 20)
  16
  17 static char block[BLOCKSIZE];
  18 static unsigned long offset;
  19
  20 static int tar_umask = 002;
  21
  22 static int write_tar_filter_archive(const struct archiver *ar,
  23                                     struct archiver_args *args);
  24
  25 /*
  26  * This is the max value that a ustar size header can specify, as it is fixed
  27  * at 11 octal digits. POSIX specifies that we switch to extended headers at
  28  * this size.
  29  *
  30  * Likewise for the mtime (which happens to use a buffer of the same size).
  31  */
  32 #if ULONG_MAX == 0xFFFFFFFF
  33 #define USTAR_MAX_SIZE ULONG_MAX
  34 #else
  35 #define USTAR_MAX_SIZE 077777777777UL
  36 #endif
  37 #if TIME_MAX == 0xFFFFFFFF
  38 #define USTAR_MAX_MTIME TIME_MAX
  39 #else
  40 #define USTAR_MAX_MTIME 077777777777ULL
  41 #endif
  42
  43 static void tar_write_block(const void *buf)
  44 {
  45         write_or_die(1, buf, BLOCKSIZE);
  46 }
  47
  48 static void (*write_block)(const void *) = tar_write_block;
  49
  50 /* writes out the whole block, but only if it is full */
  51 static void write_if_needed(void)
  52 {
  53         if (offset == BLOCKSIZE) {
  54                 write_block(block);
  55                 offset = 0;
  56         }
  57 }
  58
  59 /*
  60  * queues up writes, so that all our write(2) calls write exactly one
  61  * full block; pads writes to RECORDSIZE
  62  */
  63 static void do_write_blocked(const void *data, unsigned long size)
  64 {
  65         const char *buf = data;
  66
  67         if (offset) {
  68                 unsigned long chunk = BLOCKSIZE - offset;
  69                 if (size < chunk)
  70                         chunk = size;
  71                 memcpy(block + offset, buf, chunk);
  72                 size -= chunk;
  73                 offset += chunk;
  74                 buf += chunk;
  75                 write_if_needed();
  76         }
  77         while (size >= BLOCKSIZE) {
  78                 write_block(buf);
  79                 size -= BLOCKSIZE;
  80                 buf += BLOCKSIZE;
  81         }
  82         if (size) {
  83                 memcpy(block + offset, buf, size);
  84                 offset += size;
  85         }
  86 }
  87
  88 static void finish_record(void)
  89 {
  90         unsigned long tail;
  91         tail = offset % RECORDSIZE;
  92         if (tail)  {
  93                 memset(block + offset, 0, RECORDSIZE - tail);
  94                 offset += RECORDSIZE - tail;
  95         }
  96         write_if_needed();
  97 }
  98
  99 static void write_blocked(const void *data, unsigned long size)
 100 {
 101         do_write_blocked(data, size);
 102         finish_record();
 103 }
 104
 105 /*
 106  * The end of tar archives is marked by 2*512 nul bytes and after that
 107  * follows the rest of the block (if any).
 108  */
 109 static void write_trailer(void)
 110 {
 111         int tail = BLOCKSIZE - offset;
 112         memset(block + offset, 0, tail);
 113         write_block(block);
 114         if (tail < 2 * RECORDSIZE) {
 115                 memset(block, 0, offset);
 116                 write_block(block);
 117         }
 118 }
 119
 120 /*
 121  * queues up writes, so that all our write(2) calls write exactly one
 122  * full block; pads writes to RECORDSIZE
 123  */
 124 static int stream_blocked(struct repository *r, const struct object_id *oid)
 125 {
 126         struct git_istream *st;
 127         enum object_type type;
 128         unsigned long sz;
 129         char buf[BLOCKSIZE];
 130         ssize_t readlen;
 131
 132         st = open_istream(r, oid, &type, &sz, NULL);
 133         if (!st)
 134                 return error(_("cannot stream blob %s"), oid_to_hex(oid));
 135         for (;;) {
 136                 readlen = read_istream(st, buf, sizeof(buf));
 137                 if (readlen <= 0)
 138                         break;
 139                 do_write_blocked(buf, readlen);
 140         }
 141         close_istream(st);
 142         if (!readlen)
 143                 finish_record();
 144         return readlen;
 145 }
 146
 147 /*
 148  * pax extended header records have the format "%u %s=%s\n".  %u contains
 149  * the size of the whole string (including the %u), the first %s is the
 150  * keyword, the second one is the value.  This function constructs such a
 151  * string and appends it to a struct strbuf.
 152  */
 153 static void strbuf_append_ext_header(struct strbuf *sb, const char *keyword,
 154                                      const char *value, size_t valuelen)
 155 {
 156         size_t orig_len = sb->len;
 157         size_t len, tmp;
 158
 159         /* "%u %s=%s\n" */
 160         len = 1 + 1 + strlen(keyword) + 1 + valuelen + 1;
 161         for (tmp = 1; len / 10 >= tmp; tmp *= 10)
 162                 len++;
 163
 164         strbuf_grow(sb, len);
 165         strbuf_addf(sb, "%"PRIuMAX" %s=", (uintmax_t)len, keyword);
 166         strbuf_add(sb, value, valuelen);
 167         strbuf_addch(sb, '\n');
 168
 169         if (len != sb->len - orig_len)
 170                 BUG("pax extended header length miscalculated as %"PRIuMAX
 171                     ", should be %"PRIuMAX,
 172                     (uintmax_t)len, (uintmax_t)(sb->len - orig_len));
 173 }
 174
 175 /*
 176  * Like strbuf_append_ext_header, but for numeric values.
 177  */
 178 static void strbuf_append_ext_header_uint(struct strbuf *sb,
 179                                           const char *keyword,
 180                                           uintmax_t value)
 181 {
 182         char buf[40]; /* big enough for 2^128 in decimal, plus NUL */
 183         int len;
 184
 185         len = xsnprintf(buf, sizeof(buf), "%"PRIuMAX, value);
 186         strbuf_append_ext_header(sb, keyword, buf, len);
 187 }
 188
 189 static unsigned int ustar_header_chksum(const struct ustar_header *header)
 190 {
 191         const unsigned char *p = (const unsigned char *)header;
 192         unsigned int chksum = 0;
 193         while (p < (const unsigned char *)header->chksum)
 194                 chksum += *p++;
 195         chksum += sizeof(header->chksum) * ' ';
 196         p += sizeof(header->chksum);
 197         while (p < (const unsigned char *)header + sizeof(struct ustar_header))
 198                 chksum += *p++;
 199         return chksum;
 200 }
 201
 202 static size_t get_path_prefix(const char *path, size_t pathlen, size_t maxlen)
 203 {
 204         size_t i = pathlen;
 205         if (i > 1 && path[i - 1] == '/')
 206                 i--;
 207         if (i > maxlen)
 208                 i = maxlen;
 209         do {
 210                 i--;
 211         } while (i > 0 && path[i] != '/');
 212         return i;
 213 }
 214
 215 static void prepare_header(struct archiver_args *args,
 216                            struct ustar_header *header,
 217                            unsigned int mode, unsigned long size)
 218 {
 219         xsnprintf(header->mode, sizeof(header->mode), "%07o", mode & 07777);
 220         xsnprintf(header->size, sizeof(header->size), "%011"PRIoMAX , S_ISREG(mode) ? (uintmax_t)size : (uintmax_t)0);
 221         xsnprintf(header->mtime, sizeof(header->mtime), "%011lo", (unsigned long) args->time);
 222
 223         xsnprintf(header->uid, sizeof(header->uid), "%07o", 0);
 224         xsnprintf(header->gid, sizeof(header->gid), "%07o", 0);
 225         strlcpy(header->uname, "root", sizeof(header->uname));
 226         strlcpy(header->gname, "root", sizeof(header->gname));
 227         xsnprintf(header->devmajor, sizeof(header->devmajor), "%07o", 0);
 228         xsnprintf(header->devminor, sizeof(header->devminor), "%07o", 0);
 229
 230         memcpy(header->magic, "ustar", 6);
 231         memcpy(header->version, "00", 2);
 232
 233         xsnprintf(header->chksum, sizeof(header->chksum), "%07o", ustar_header_chksum(header));
 234 }
 235
 236 static void write_extended_header(struct archiver_args *args,
 237                                   const struct object_id *oid,
 238                                   const void *buffer, unsigned long size)
 239 {
 240         struct ustar_header header;
 241         unsigned int mode;
 242         memset(&header, 0, sizeof(header));
 243         *header.typeflag = TYPEFLAG_EXT_HEADER;
 244         mode = 0100666;
 245         xsnprintf(header.name, sizeof(header.name), "%s.paxheader", oid_to_hex(oid));
 246         prepare_header(args, &header, mode, size);
 247         write_blocked(&header, sizeof(header));
 248         write_blocked(buffer, size);
 249 }
 250
 251 static int write_tar_entry(struct archiver_args *args,
 252                            const struct object_id *oid,
 253                            const char *path, size_t pathlen,
 254                            unsigned int mode,
 255                            void *buffer, unsigned long size)
 256 {
 257         struct ustar_header header;
 258         struct strbuf ext_header = STRBUF_INIT;
 259         unsigned long size_in_header;
 260         int err = 0;
 261
 262         memset(&header, 0, sizeof(header));
 263
 264         if (S_ISDIR(mode) || S_ISGITLINK(mode)) {
 265                 *header.typeflag = TYPEFLAG_DIR;
 266                 mode = (mode | 0777) & ~tar_umask;
 267         } else if (S_ISLNK(mode)) {
 268                 *header.typeflag = TYPEFLAG_LNK;
 269                 mode |= 0777;
 270         } else if (S_ISREG(mode)) {
 271                 *header.typeflag = TYPEFLAG_REG;
 272                 mode = (mode | ((mode & 0100) ? 0777 : 0666)) & ~tar_umask;
 273         } else {
 274                 return error(_("unsupported file mode: 0%o (SHA1: %s)"),
 275                              mode, oid_to_hex(oid));
 276         }
 277         if (pathlen > sizeof(header.name)) {
 278                 size_t plen = get_path_prefix(path, pathlen,
 279                                               sizeof(header.prefix));
 280                 size_t rest = pathlen - plen - 1;
 281                 if (plen > 0 && rest <= sizeof(header.name)) {
 282                         memcpy(header.prefix, path, plen);
 283                         memcpy(header.name, path + plen + 1, rest);
 284                 } else {
 285                         xsnprintf(header.name, sizeof(header.name), "%s.data",
 286                                   oid_to_hex(oid));
 287                         strbuf_append_ext_header(&ext_header, "path",
 288                                                  path, pathlen);
 289                 }
 290         } else
 291                 memcpy(header.name, path, pathlen);
 292
 293         if (S_ISLNK(mode)) {
 294                 if (size > sizeof(header.linkname)) {
 295                         xsnprintf(header.linkname, sizeof(header.linkname),
 296                                   "see %s.paxheader", oid_to_hex(oid));
 297                         strbuf_append_ext_header(&ext_header, "linkpath",
 298                                                  buffer, size);
 299                 } else
 300                         memcpy(header.linkname, buffer, size);
 301         }
 302
 303         size_in_header = size;
 304         if (S_ISREG(mode) && size > USTAR_MAX_SIZE) {
 305                 size_in_header = 0;
 306                 strbuf_append_ext_header_uint(&ext_header, "size", size);
 307         }
 308
 309         prepare_header(args, &header, mode, size_in_header);
 310
 311         if (ext_header.len > 0) {
 312                 write_extended_header(args, oid, ext_header.buf,
 313                                       ext_header.len);
 314         }
 315         strbuf_release(&ext_header);
 316         write_blocked(&header, sizeof(header));
 317         if (S_ISREG(mode) && size > 0) {
 318                 if (buffer)
 319                         write_blocked(buffer, size);
 320                 else
 321                         err = stream_blocked(args->repo, oid);
 322         }
 323         return err;
 324 }
 325
 326 static void write_global_extended_header(struct archiver_args *args)
 327 {
 328         const struct object_id *oid = args->commit_oid;
 329         struct strbuf ext_header = STRBUF_INIT;
 330         struct ustar_header header;
 331         unsigned int mode;
 332
 333         if (oid)
 334                 strbuf_append_ext_header(&ext_header, "comment",
 335                                          oid_to_hex(oid),
 336                                          the_hash_algo->hexsz);
 337         if (args->time > USTAR_MAX_MTIME) {
 338                 strbuf_append_ext_header_uint(&ext_header, "mtime",
 339                                               args->time);
 340                 args->time = USTAR_MAX_MTIME;
 341         }
 342
 343         if (!ext_header.len)
 344                 return;
 345
 346         memset(&header, 0, sizeof(header));
 347         *header.typeflag = TYPEFLAG_GLOBAL_HEADER;
 348         mode = 0100666;
 349         xsnprintf(header.name, sizeof(header.name), "pax_global_header");
 350         prepare_header(args, &header, mode, ext_header.len);
 351         write_blocked(&header, sizeof(header));
 352         write_blocked(ext_header.buf, ext_header.len);
 353         strbuf_release(&ext_header);
 354 }
 355
 356 static struct archiver **tar_filters;
 357 static int nr_tar_filters;
 358 static int alloc_tar_filters;
 359
 360 static struct archiver *find_tar_filter(const char *name, size_t len)
 361 {
 362         int i;
 363         for (i = 0; i < nr_tar_filters; i++) {
 364                 struct archiver *ar = tar_filters[i];
 365                 if (!strncmp(ar->name, name, len) && !ar->name[len])
 366                         return ar;
 367         }
 368         return NULL;
 369 }
 370
 371 static int tar_filter_config(const char *var, const char *value,
 372                              void *data UNUSED)
 373 {
 374         struct archiver *ar;
 375         const char *name;
 376         const char *type;
 377         size_t namelen;
 378
 379         if (parse_config_key(var, "tar", &name, &namelen, &type) < 0 || !name)
 380                 return 0;
 381
 382         ar = find_tar_filter(name, namelen);
 383         if (!ar) {
 384                 CALLOC_ARRAY(ar, 1);
 385                 ar->name = xmemdupz(name, namelen);
 386                 ar->write_archive = write_tar_filter_archive;
 387                 ar->flags = ARCHIVER_WANT_COMPRESSION_LEVELS |
 388                             ARCHIVER_HIGH_COMPRESSION_LEVELS;
 389                 ALLOC_GROW(tar_filters, nr_tar_filters + 1, alloc_tar_filters);
 390                 tar_filters[nr_tar_filters++] = ar;
 391         }
 392
 393         if (!strcmp(type, "command")) {
 394                 if (!value)
 395                         return config_error_nonbool(var);
 396                 free(ar->filter_command);
 397                 ar->filter_command = xstrdup(value);
 398                 return 0;
 399         }
 400         if (!strcmp(type, "remote")) {
 401                 if (git_config_bool(var, value))
 402                         ar->flags |= ARCHIVER_REMOTE;
 403                 else
 404                         ar->flags &= ~ARCHIVER_REMOTE;
 405                 return 0;
 406         }
 407
 408         return 0;
 409 }
 410
 411 static int git_tar_config(const char *var, const char *value, void *cb)
 412 {
 413         if (!strcmp(var, "tar.umask")) {
 414                 if (value && !strcmp(value, "user")) {
 415                         tar_umask = umask(0);
 416                         umask(tar_umask);
 417                 } else {
 418                         tar_umask = git_config_int(var, value);
 419                 }
 420                 return 0;
 421         }
 422
 423         return tar_filter_config(var, value, cb);
 424 }
 425
 426 static int write_tar_archive(const struct archiver *ar UNUSED,
 427                              struct archiver_args *args)
 428 {
 429         int err = 0;
 430
 431         write_global_extended_header(args);
 432         err = write_archive_entries(args, write_tar_entry);
 433         if (!err)
 434                 write_trailer();
 435         return err;
 436 }
 437
 438 static git_zstream gzstream;
 439 static unsigned char outbuf[16384];
 440
 441 static void tgz_deflate(int flush)
 442 {
 443         while (gzstream.avail_in || flush == Z_FINISH) {
 444                 int status = git_deflate(&gzstream, flush);
 445                 if (!gzstream.avail_out || status == Z_STREAM_END) {
 446                         write_or_die(1, outbuf, gzstream.next_out - outbuf);
 447                         gzstream.next_out = outbuf;
 448                         gzstream.avail_out = sizeof(outbuf);
 449                         if (status == Z_STREAM_END)
 450                                 break;
 451                 }
 452                 if (status != Z_OK && status != Z_BUF_ERROR)
 453                         die(_("deflate error (%d)"), status);
 454         }
 455 }
 456
 457 static void tgz_write_block(const void *data)
 458 {
 459         gzstream.next_in = (void *)data;
 460         gzstream.avail_in = BLOCKSIZE;
 461         tgz_deflate(Z_NO_FLUSH);
 462 }
 463
 464 static const char internal_gzip_command[] = "git archive gzip";
 465
 466 static int write_tar_filter_archive(const struct archiver *ar,
 467                                     struct archiver_args *args)
 468 {
 469 #if ZLIB_VERNUM >= 0x1221
 470         struct gz_header_s gzhead = { .os = 3 }; /* Unix, for reproducibility */
 471 #endif
 472         struct strbuf cmd = STRBUF_INIT;
 473         struct child_process filter = CHILD_PROCESS_INIT;
 474         int r;
 475
 476         if (!ar->filter_command)
 477                 BUG("tar-filter archiver called with no filter defined");
 478
 479         if (!strcmp(ar->filter_command, internal_gzip_command)) {
 480                 write_block = tgz_write_block;
 481                 git_deflate_init_gzip(&gzstream, args->compression_level);
 482 #if ZLIB_VERNUM >= 0x1221
 483                 if (deflateSetHeader(&gzstream.z, &gzhead) != Z_OK)
 484                         BUG("deflateSetHeader() called too late");
 485 #endif
 486                 gzstream.next_out = outbuf;
 487                 gzstream.avail_out = sizeof(outbuf);
 488
 489                 r = write_tar_archive(ar, args);
 490
 491                 tgz_deflate(Z_FINISH);
 492                 git_deflate_end(&gzstream);
 493                 return r;
 494         }
 495
 496         strbuf_addstr(&cmd, ar->filter_command);
 497         if (args->compression_level >= 0)
 498                 strbuf_addf(&cmd, " -%d", args->compression_level);
 499
 500         strvec_push(&filter.args, cmd.buf);
 501         filter.use_shell = 1;
 502         filter.in = -1;
 503         filter.silent_exec_failure = 1;
 504
 505         if (start_command(&filter) < 0)
 506                 die_errno(_("unable to start '%s' filter"), cmd.buf);
 507         close(1);
 508         if (dup2(filter.in, 1) < 0)
 509                 die_errno(_("unable to redirect descriptor"));
 510         close(filter.in);
 511
 512         r = write_tar_archive(ar, args);
 513
 514         close(1);
 515         if (finish_command(&filter) != 0)
 516                 die(_("'%s' filter reported error"), cmd.buf);
 517
 518         strbuf_release(&cmd);
 519         return r;
 520 }
 521
 522 static struct archiver tar_archiver = {
 523         .name = "tar",
 524         .write_archive = write_tar_archive,
 525         .flags = ARCHIVER_REMOTE,
 526 };
 527
 528 void init_tar_archiver(void)
 529 {
 530         int i;
 531         register_archiver(&tar_archiver);
 532
 533         tar_filter_config("tar.tgz.command", internal_gzip_command, NULL);
 534         tar_filter_config("tar.tgz.remote", "true", NULL);
 535         tar_filter_config("tar.tar.gz.command", internal_gzip_command, NULL);
 536         tar_filter_config("tar.tar.gz.remote", "true", NULL);
 537         git_config(git_tar_config, NULL);
 538         for (i = 0; i < nr_tar_filters; i++) {
 539                 /* omit any filters that never had a command configured */
 540                 if (tar_filters[i]->filter_command)
 541                         register_archiver(tar_filters[i]);
 542         }
 543 }