archive-tar.c

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