archive-tar.c

   1 /*
   2  * Copyright (c) 2005, 2006 Rene Scharfe
   3  */
   4
   5 #define USE_THE_REPOSITORY_VARIABLE
   6
   7 #include "git-compat-util.h"
   8 #include "config.h"
   9 #include "gettext.h"
  10 #include "git-zlib.h"
  11 #include "hex.h"
  12 #include "tar.h"
  13 #include "archive.h"
  14 #include "object-store-ll.h"
  15 #include "strbuf.h"
  16 #include "streaming.h"
  17 #include "run-command.h"
  18 #include "write-or-die.h"
  19
  20 #define RECORDSIZE      (512)
  21 #define BLOCKSIZE       (RECORDSIZE * 20)
  22
  23 static char block[BLOCKSIZE];
  24 static unsigned long offset;
  25
  26 static int tar_umask = 002;
  27
  28 static int write_tar_filter_archive(const struct archiver *ar,
  29                                     struct archiver_args *args);
  30
  31 /*
  32  * This is the max value that a ustar size header can specify, as it is fixed
  33  * at 11 octal digits. POSIX specifies that we switch to extended headers at
  34  * this size.
  35  *
  36  * Likewise for the mtime (which happens to use a buffer of the same size).
  37  */
  38 #if ULONG_MAX == 0xFFFFFFFF
  39 #define USTAR_MAX_SIZE ULONG_MAX
  40 #else
  41 #define USTAR_MAX_SIZE 077777777777UL
  42 #endif
  43 #if TIME_MAX == 0xFFFFFFFF
  44 #define USTAR_MAX_MTIME TIME_MAX
  45 #else
  46 #define USTAR_MAX_MTIME 077777777777ULL
  47 #endif
  48
  49 static void tar_write_block(const void *buf)
  50 {
  51         write_or_die(1, buf, BLOCKSIZE);
  52 }
  53
  54 static void (*write_block)(const void *) = tar_write_block;
  55
  56 /* writes out the whole block, but only if it is full */
  57 static void write_if_needed(void)
  58 {
  59         if (offset == BLOCKSIZE) {
  60                 write_block(block);
  61                 offset = 0;
  62         }
  63 }
  64
  65 /*
  66  * queues up writes, so that all our write(2) calls write exactly one
  67  * full block; pads writes to RECORDSIZE
  68  */
  69 static void do_write_blocked(const void *data, unsigned long size)
  70 {
  71         const char *buf = data;
  72
  73         if (offset) {
  74                 unsigned long chunk = BLOCKSIZE - offset;
  75                 if (size < chunk)
  76                         chunk = size;
  77                 memcpy(block + offset, buf, chunk);
  78                 size -= chunk;
  79                 offset += chunk;
  80                 buf += chunk;
  81                 write_if_needed();
  82         }
  83         while (size >= BLOCKSIZE) {
  84                 write_block(buf);
  85                 size -= BLOCKSIZE;
  86                 buf += BLOCKSIZE;
  87         }
  88         if (size) {
  89                 memcpy(block + offset, buf, size);
  90                 offset += size;
  91         }
  92 }
  93
  94 static void finish_record(void)
  95 {
  96         unsigned long tail;
  97         tail = offset % RECORDSIZE;
  98         if (tail)  {
  99                 memset(block + offset, 0, RECORDSIZE - tail);
 100                 offset += RECORDSIZE - tail;
 101         }
 102         write_if_needed();
 103 }
 104
 105 static void write_blocked(const void *data, unsigned long size)
 106 {
 107         do_write_blocked(data, size);
 108         finish_record();
 109 }
 110
 111 /*
 112  * The end of tar archives is marked by 2*512 nul bytes and after that
 113  * follows the rest of the block (if any).
 114  */
 115 static void write_trailer(void)
 116 {
 117         int tail = BLOCKSIZE - offset;
 118         memset(block + offset, 0, tail);
 119         write_block(block);
 120         if (tail < 2 * RECORDSIZE) {
 121                 memset(block, 0, offset);
 122                 write_block(block);
 123         }
 124 }
 125
 126 /*
 127  * queues up writes, so that all our write(2) calls write exactly one
 128  * full block; pads writes to RECORDSIZE
 129  */
 130 static int stream_blocked(struct repository *r, const struct object_id *oid)
 131 {
 132         struct git_istream *st;
 133         enum object_type type;
 134         unsigned long sz;
 135         char buf[BLOCKSIZE];
 136         ssize_t readlen;
 137
 138         st = open_istream(r, oid, &type, &sz, NULL);
 139         if (!st)
 140                 return error(_("cannot stream blob %s"), oid_to_hex(oid));
 141         for (;;) {
 142                 readlen = read_istream(st, buf, sizeof(buf));
 143                 if (readlen <= 0)
 144                         break;
 145                 do_write_blocked(buf, readlen);
 146         }
 147         close_istream(st);
 148         if (!readlen)
 149                 finish_record();
 150         return readlen;
 151 }
 152
 153 /*
 154  * pax extended header records have the format "%u %s=%s\n".  %u contains
 155  * the size of the whole string (including the %u), the first %s is the
 156  * keyword, the second one is the value.  This function constructs such a
 157  * string and appends it to a struct strbuf.
 158  */
 159 static void strbuf_append_ext_header(struct strbuf *sb, const char *keyword,
 160                                      const char *value, size_t valuelen)
 161 {
 162         size_t orig_len = sb->len;
 163         size_t len, tmp;
 164
 165         /* "%u %s=%s\n" */
 166         len = 1 + 1 + strlen(keyword) + 1 + valuelen + 1;
 167         for (tmp = 1; len / 10 >= tmp; tmp *= 10)
 168                 len++;
 169
 170         strbuf_grow(sb, len);
 171         strbuf_addf(sb, "%"PRIuMAX" %s=", (uintmax_t)len, keyword);
 172         strbuf_add(sb, value, valuelen);
 173         strbuf_addch(sb, '\n');
 174
 175         if (len != sb->len - orig_len)
 176                 BUG("pax extended header length miscalculated as %"PRIuMAX
 177                     ", should be %"PRIuMAX,
 178                     (uintmax_t)len, (uintmax_t)(sb->len - orig_len));
 179 }
 180
 181 /*
 182  * Like strbuf_append_ext_header, but for numeric values.
 183  */
 184 static void strbuf_append_ext_header_uint(struct strbuf *sb,
 185                                           const char *keyword,
 186                                           uintmax_t value)
 187 {
 188         char buf[40]; /* big enough for 2^128 in decimal, plus NUL */
 189         int len;
 190
 191         len = xsnprintf(buf, sizeof(buf), "%"PRIuMAX, value);
 192         strbuf_append_ext_header(sb, keyword, buf, len);
 193 }
 194
 195 static unsigned int ustar_header_chksum(const struct ustar_header *header)
 196 {
 197         const unsigned char *p = (const unsigned char *)header;
 198         unsigned int chksum = 0;
 199         while (p < (const unsigned char *)header->chksum)
 200                 chksum += *p++;
 201         chksum += sizeof(header->chksum) * ' ';
 202         p += sizeof(header->chksum);
 203         while (p < (const unsigned char *)header + sizeof(struct ustar_header))
 204                 chksum += *p++;
 205         return chksum;
 206 }
 207
 208 static size_t get_path_prefix(const char *path, size_t pathlen, size_t maxlen)
 209 {
 210         size_t i = pathlen;
 211         if (i > 1 && path[i - 1] == '/')
 212                 i--;
 213         if (i > maxlen)
 214                 i = maxlen;
 215         do {
 216                 i--;
 217         } while (i > 0 && path[i] != '/');
 218         return i;
 219 }
 220
 221 static void prepare_header(struct archiver_args *args,
 222                            struct ustar_header *header,
 223                            unsigned int mode, unsigned long size)
 224 {
 225         xsnprintf(header->mode, sizeof(header->mode), "%07o", mode & 07777);
 226         xsnprintf(header->size, sizeof(header->size), "%011"PRIoMAX , S_ISREG(mode) ? (uintmax_t)size : (uintmax_t)0);
 227         xsnprintf(header->mtime, sizeof(header->mtime), "%011lo", (unsigned long) args->time);
 228
 229         xsnprintf(header->uid, sizeof(header->uid), "%07o", 0);
 230         xsnprintf(header->gid, sizeof(header->gid), "%07o", 0);
 231         strlcpy(header->uname, "root", sizeof(header->uname));
 232         strlcpy(header->gname, "root", sizeof(header->gname));
 233         xsnprintf(header->devmajor, sizeof(header->devmajor), "%07o", 0);
 234         xsnprintf(header->devminor, sizeof(header->devminor), "%07o", 0);
 235
 236         memcpy(header->magic, "ustar", 6);
 237         memcpy(header->version, "00", 2);
 238
 239         xsnprintf(header->chksum, sizeof(header->chksum), "%07o", ustar_header_chksum(header));
 240 }
 241
 242 static void write_extended_header(struct archiver_args *args,
 243                                   const struct object_id *oid,
 244                                   const void *buffer, unsigned long size)
 245 {
 246         struct ustar_header header;
 247         unsigned int mode;
 248         memset(&header, 0, sizeof(header));
 249         *header.typeflag = TYPEFLAG_EXT_HEADER;
 250         mode = 0100666;
 251         xsnprintf(header.name, sizeof(header.name), "%s.paxheader", oid_to_hex(oid));
 252         prepare_header(args, &header, mode, size);
 253         write_blocked(&header, sizeof(header));
 254         write_blocked(buffer, size);
 255 }
 256
 257 static int write_tar_entry(struct archiver_args *args,
 258                            const struct object_id *oid,
 259                            const char *path, size_t pathlen,
 260                            unsigned int mode,
 261                            void *buffer, unsigned long size)
 262 {
 263         struct ustar_header header;
 264         struct strbuf ext_header = STRBUF_INIT;
 265         unsigned long size_in_header;
 266         int err = 0;
 267
 268         memset(&header, 0, sizeof(header));
 269
 270         if (S_ISDIR(mode) || S_ISGITLINK(mode)) {
 271                 *header.typeflag = TYPEFLAG_DIR;
 272                 mode = (mode | 0777) & ~tar_umask;
 273         } else if (S_ISLNK(mode)) {
 274                 *header.typeflag = TYPEFLAG_LNK;
 275                 mode |= 0777;
 276         } else if (S_ISREG(mode)) {
 277                 *header.typeflag = TYPEFLAG_REG;
 278                 mode = (mode | ((mode & 0100) ? 0777 : 0666)) & ~tar_umask;
 279         } else {
 280                 return error(_("unsupported file mode: 0%o (SHA1: %s)"),
 281                              mode, oid_to_hex(oid));
 282         }
 283         if (pathlen > sizeof(header.name)) {
 284                 size_t plen = get_path_prefix(path, pathlen,
 285                                               sizeof(header.prefix));
 286                 size_t rest = pathlen - plen - 1;
 287                 if (plen > 0 && rest <= sizeof(header.name)) {
 288                         memcpy(header.prefix, path, plen);
 289                         memcpy(header.name, path + plen + 1, rest);
 290                 } else {
 291                         xsnprintf(header.name, sizeof(header.name), "%s.data",
 292                                   oid_to_hex(oid));
 293                         strbuf_append_ext_header(&ext_header, "path",
 294                                                  path, pathlen);
 295                 }
 296         } else
 297                 memcpy(header.name, path, pathlen);
 298
 299         if (S_ISLNK(mode)) {
 300                 if (size > sizeof(header.linkname)) {
 301                         xsnprintf(header.linkname, sizeof(header.linkname),
 302                                   "see %s.paxheader", oid_to_hex(oid));
 303                         strbuf_append_ext_header(&ext_header, "linkpath",
 304                                                  buffer, size);
 305                 } else
 306                         memcpy(header.linkname, buffer, size);
 307         }
 308
 309         size_in_header = size;
 310         if (S_ISREG(mode) && size > USTAR_MAX_SIZE) {
 311                 size_in_header = 0;
 312                 strbuf_append_ext_header_uint(&ext_header, "size", size);
 313         }
 314
 315         prepare_header(args, &header, mode, size_in_header);
 316
 317         if (ext_header.len > 0) {
 318                 write_extended_header(args, oid, ext_header.buf,
 319                                       ext_header.len);
 320         }
 321         strbuf_release(&ext_header);
 322         write_blocked(&header, sizeof(header));
 323         if (S_ISREG(mode) && size > 0) {
 324                 if (buffer)
 325                         write_blocked(buffer, size);
 326                 else
 327                         err = stream_blocked(args->repo, oid);
 328         }
 329         return err;
 330 }
 331
 332 static void write_global_extended_header(struct archiver_args *args)
 333 {
 334         const struct object_id *oid = args->commit_oid;
 335         struct strbuf ext_header = STRBUF_INIT;
 336         struct ustar_header header;
 337         unsigned int mode;
 338
 339         if (oid)
 340                 strbuf_append_ext_header(&ext_header, "comment",
 341                                          oid_to_hex(oid),
 342                                          the_hash_algo->hexsz);
 343         if (args->time > USTAR_MAX_MTIME) {
 344                 strbuf_append_ext_header_uint(&ext_header, "mtime",
 345                                               args->time);
 346                 args->time = USTAR_MAX_MTIME;
 347         }
 348
 349         if (!ext_header.len)
 350                 return;
 351
 352         memset(&header, 0, sizeof(header));
 353         *header.typeflag = TYPEFLAG_GLOBAL_HEADER;
 354         mode = 0100666;
 355         xsnprintf(header.name, sizeof(header.name), "pax_global_header");
 356         prepare_header(args, &header, mode, ext_header.len);
 357         write_blocked(&header, sizeof(header));
 358         write_blocked(ext_header.buf, ext_header.len);
 359         strbuf_release(&ext_header);
 360 }
 361
 362 static struct archiver **tar_filters;
 363 static int nr_tar_filters;
 364 static int alloc_tar_filters;
 365
 366 static struct archiver *find_tar_filter(const char *name, size_t len)
 367 {
 368         int i;
 369         for (i = 0; i < nr_tar_filters; i++) {
 370                 struct archiver *ar = tar_filters[i];
 371                 if (!xstrncmpz(ar->name, name, len))
 372                         return ar;
 373         }
 374         return NULL;
 375 }
 376
 377 static int tar_filter_config(const char *var, const char *value,
 378                              void *data UNUSED)
 379 {
 380         struct archiver *ar;
 381         const char *name;
 382         const char *type;
 383         size_t namelen;
 384
 385         if (parse_config_key(var, "tar", &name, &namelen, &type) < 0 || !name)
 386                 return 0;
 387
 388         ar = find_tar_filter(name, namelen);
 389         if (!ar) {
 390                 CALLOC_ARRAY(ar, 1);
 391                 ar->name = xmemdupz(name, namelen);
 392                 ar->write_archive = write_tar_filter_archive;
 393                 ar->flags = ARCHIVER_WANT_COMPRESSION_LEVELS |
 394                             ARCHIVER_HIGH_COMPRESSION_LEVELS;
 395                 ALLOC_GROW(tar_filters, nr_tar_filters + 1, alloc_tar_filters);
 396                 tar_filters[nr_tar_filters++] = ar;
 397         }
 398
 399         if (!strcmp(type, "command")) {
 400                 if (!value)
 401                         return config_error_nonbool(var);
 402                 free(ar->filter_command);
 403                 ar->filter_command = xstrdup(value);
 404                 return 0;
 405         }
 406         if (!strcmp(type, "remote")) {
 407                 if (git_config_bool(var, value))
 408                         ar->flags |= ARCHIVER_REMOTE;
 409                 else
 410                         ar->flags &= ~ARCHIVER_REMOTE;
 411                 return 0;
 412         }
 413
 414         return 0;
 415 }
 416
 417 static int git_tar_config(const char *var, const char *value,
 418                           const struct config_context *ctx, void *cb)
 419 {
 420         if (!strcmp(var, "tar.umask")) {
 421                 if (value && !strcmp(value, "user")) {
 422                         tar_umask = umask(0);
 423                         umask(tar_umask);
 424                 } else {
 425                         tar_umask = git_config_int(var, value, ctx->kvi);
 426                 }
 427                 return 0;
 428         }
 429
 430         return tar_filter_config(var, value, cb);
 431 }
 432
 433 static int write_tar_archive(const struct archiver *ar UNUSED,
 434                              struct archiver_args *args)
 435 {
 436         int err = 0;
 437
 438         write_global_extended_header(args);
 439         err = write_archive_entries(args, write_tar_entry);
 440         if (!err)
 441                 write_trailer();
 442         return err;
 443 }
 444
 445 static git_zstream gzstream;
 446 static unsigned char outbuf[16384];
 447
 448 static void tgz_deflate(int flush)
 449 {
 450         while (gzstream.avail_in || flush == Z_FINISH) {
 451                 int status = git_deflate(&gzstream, flush);
 452                 if (!gzstream.avail_out || status == Z_STREAM_END) {
 453                         write_or_die(1, outbuf, gzstream.next_out - outbuf);
 454                         gzstream.next_out = outbuf;
 455                         gzstream.avail_out = sizeof(outbuf);
 456                         if (status == Z_STREAM_END)
 457                                 break;
 458                 }
 459                 if (status != Z_OK && status != Z_BUF_ERROR)
 460                         die(_("deflate error (%d)"), status);
 461         }
 462 }
 463
 464 static void tgz_write_block(const void *data)
 465 {
 466         gzstream.next_in = (void *)data;
 467         gzstream.avail_in = BLOCKSIZE;
 468         tgz_deflate(Z_NO_FLUSH);
 469 }
 470
 471 static const char internal_gzip_command[] = "git archive gzip";
 472
 473 static int write_tar_filter_archive(const struct archiver *ar,
 474                                     struct archiver_args *args)
 475 {
 476 #if ZLIB_VERNUM >= 0x1221
 477         struct gz_header_s gzhead = { .os = 3 }; /* Unix, for reproducibility */
 478 #endif
 479         struct strbuf cmd = STRBUF_INIT;
 480         struct child_process filter = CHILD_PROCESS_INIT;
 481         int r;
 482
 483         if (!ar->filter_command)
 484                 BUG("tar-filter archiver called with no filter defined");
 485
 486         if (!strcmp(ar->filter_command, internal_gzip_command)) {
 487                 write_block = tgz_write_block;
 488                 git_deflate_init_gzip(&gzstream, args->compression_level);
 489 #if ZLIB_VERNUM >= 0x1221
 490                 if (deflateSetHeader(&gzstream.z, &gzhead) != Z_OK)
 491                         BUG("deflateSetHeader() called too late");
 492 #endif
 493                 gzstream.next_out = outbuf;
 494                 gzstream.avail_out = sizeof(outbuf);
 495
 496                 r = write_tar_archive(ar, args);
 497
 498                 tgz_deflate(Z_FINISH);
 499                 git_deflate_end(&gzstream);
 500                 return r;
 501         }
 502
 503         strbuf_addstr(&cmd, ar->filter_command);
 504         if (args->compression_level >= 0)
 505                 strbuf_addf(&cmd, " -%d", args->compression_level);
 506
 507         strvec_push(&filter.args, cmd.buf);
 508         filter.use_shell = 1;
 509         filter.in = -1;
 510         filter.silent_exec_failure = 1;
 511
 512         if (start_command(&filter) < 0)
 513                 die_errno(_("unable to start '%s' filter"), cmd.buf);
 514         close(1);
 515         if (dup2(filter.in, 1) < 0)
 516                 die_errno(_("unable to redirect descriptor"));
 517         close(filter.in);
 518
 519         r = write_tar_archive(ar, args);
 520
 521         close(1);
 522         if (finish_command(&filter) != 0)
 523                 die(_("'%s' filter reported error"), cmd.buf);
 524
 525         strbuf_release(&cmd);
 526         return r;
 527 }
 528
 529 static struct archiver tar_archiver = {
 530         .name = "tar",
 531         .write_archive = write_tar_archive,
 532         .flags = ARCHIVER_REMOTE,
 533 };
 534
 535 void init_tar_archiver(void)
 536 {
 537         int i;
 538         register_archiver(&tar_archiver);
 539
 540         tar_filter_config("tar.tgz.command", internal_gzip_command, NULL);
 541         tar_filter_config("tar.tgz.remote", "true", NULL);
 542         tar_filter_config("tar.tar.gz.command", internal_gzip_command, NULL);
 543         tar_filter_config("tar.tar.gz.remote", "true", NULL);
 544         git_config(git_tar_config, NULL);
 545         for (i = 0; i < nr_tar_filters; i++) {
 546                 /* omit any filters that never had a command configured */
 547                 if (tar_filters[i]->filter_command)
 548                         register_archiver(tar_filters[i]);
 549         }
 550 }