archive-tar.c

   1 /*
   2  * Copyright (c) 2005, 2006 Rene Scharfe
   3  */
   4 #include "cache.h"
   5 #include "tar.h"
   6 #include "archive.h"
   7 #include "streaming.h"
   8 #include "run-command.h"
   9
  10 #define RECORDSIZE      (512)
  11 #define BLOCKSIZE       (RECORDSIZE * 20)
  12
  13 static char block[BLOCKSIZE];
  14 static unsigned long offset;
  15
  16 static int tar_umask = 002;
  17
  18 static int write_tar_filter_archive(const struct archiver *ar,
  19                                     struct archiver_args *args);
  20
  21 /*
  22  * This is the max value that a ustar size header can specify, as it is fixed
  23  * at 11 octal digits. POSIX specifies that we switch to extended headers at
  24  * this size.
  25  *
  26  * Likewise for the mtime (which happens to use a buffer of the same size).
  27  */
  28 #define USTAR_MAX_SIZE 077777777777UL
  29 #define USTAR_MAX_MTIME 077777777777UL
  30
  31 /* writes out the whole block, but only if it is full */
  32 static void write_if_needed(void)
  33 {
  34         if (offset == BLOCKSIZE) {
  35                 write_or_die(1, block, BLOCKSIZE);
  36                 offset = 0;
  37         }
  38 }
  39
  40 /*
  41  * queues up writes, so that all our write(2) calls write exactly one
  42  * full block; pads writes to RECORDSIZE
  43  */
  44 static void do_write_blocked(const void *data, unsigned long size)
  45 {
  46         const char *buf = data;
  47
  48         if (offset) {
  49                 unsigned long chunk = BLOCKSIZE - offset;
  50                 if (size < chunk)
  51                         chunk = size;
  52                 memcpy(block + offset, buf, chunk);
  53                 size -= chunk;
  54                 offset += chunk;
  55                 buf += chunk;
  56                 write_if_needed();
  57         }
  58         while (size >= BLOCKSIZE) {
  59                 write_or_die(1, buf, BLOCKSIZE);
  60                 size -= BLOCKSIZE;
  61                 buf += BLOCKSIZE;
  62         }
  63         if (size) {
  64                 memcpy(block + offset, buf, size);
  65                 offset += size;
  66         }
  67 }
  68
  69 static void finish_record(void)
  70 {
  71         unsigned long tail;
  72         tail = offset % RECORDSIZE;
  73         if (tail)  {
  74                 memset(block + offset, 0, RECORDSIZE - tail);
  75                 offset += RECORDSIZE - tail;
  76         }
  77         write_if_needed();
  78 }
  79
  80 static void write_blocked(const void *data, unsigned long size)
  81 {
  82         do_write_blocked(data, size);
  83         finish_record();
  84 }
  85
  86 /*
  87  * The end of tar archives is marked by 2*512 nul bytes and after that
  88  * follows the rest of the block (if any).
  89  */
  90 static void write_trailer(void)
  91 {
  92         int tail = BLOCKSIZE - offset;
  93         memset(block + offset, 0, tail);
  94         write_or_die(1, block, BLOCKSIZE);
  95         if (tail < 2 * RECORDSIZE) {
  96                 memset(block, 0, offset);
  97                 write_or_die(1, block, BLOCKSIZE);
  98         }
  99 }
 100
 101 /*
 102  * queues up writes, so that all our write(2) calls write exactly one
 103  * full block; pads writes to RECORDSIZE
 104  */
 105 static int stream_blocked(const unsigned char *sha1)
 106 {
 107         struct git_istream *st;
 108         enum object_type type;
 109         unsigned long sz;
 110         char buf[BLOCKSIZE];
 111         ssize_t readlen;
 112
 113         st = open_istream(sha1, &type, &sz, NULL);
 114         if (!st)
 115                 return error("cannot stream blob %s", sha1_to_hex(sha1));
 116         for (;;) {
 117                 readlen = read_istream(st, buf, sizeof(buf));
 118                 if (readlen <= 0)
 119                         break;
 120                 do_write_blocked(buf, readlen);
 121         }
 122         close_istream(st);
 123         if (!readlen)
 124                 finish_record();
 125         return readlen;
 126 }
 127
 128 /*
 129  * pax extended header records have the format "%u %s=%s\n".  %u contains
 130  * the size of the whole string (including the %u), the first %s is the
 131  * keyword, the second one is the value.  This function constructs such a
 132  * string and appends it to a struct strbuf.
 133  */
 134 static void strbuf_append_ext_header(struct strbuf *sb, const char *keyword,
 135                                      const char *value, unsigned int valuelen)
 136 {
 137         int len, tmp;
 138
 139         /* "%u %s=%s\n" */
 140         len = 1 + 1 + strlen(keyword) + 1 + valuelen + 1;
 141         for (tmp = len; tmp > 9; tmp /= 10)
 142                 len++;
 143
 144         strbuf_grow(sb, len);
 145         strbuf_addf(sb, "%u %s=", len, keyword);
 146         strbuf_add(sb, value, valuelen);
 147         strbuf_addch(sb, '\n');
 148 }
 149
 150 /*
 151  * Like strbuf_append_ext_header, but for numeric values.
 152  */
 153 static void strbuf_append_ext_header_uint(struct strbuf *sb,
 154                                           const char *keyword,
 155                                           uintmax_t value)
 156 {
 157         char buf[40]; /* big enough for 2^128 in decimal, plus NUL */
 158         int len;
 159
 160         len = xsnprintf(buf, sizeof(buf), "%"PRIuMAX, value);
 161         strbuf_append_ext_header(sb, keyword, buf, len);
 162 }
 163
 164 static unsigned int ustar_header_chksum(const struct ustar_header *header)
 165 {
 166         const unsigned char *p = (const unsigned char *)header;
 167         unsigned int chksum = 0;
 168         while (p < (const unsigned char *)header->chksum)
 169                 chksum += *p++;
 170         chksum += sizeof(header->chksum) * ' ';
 171         p += sizeof(header->chksum);
 172         while (p < (const unsigned char *)header + sizeof(struct ustar_header))
 173                 chksum += *p++;
 174         return chksum;
 175 }
 176
 177 static size_t get_path_prefix(const char *path, size_t pathlen, size_t maxlen)
 178 {
 179         size_t i = pathlen;
 180         if (i > 1 && path[i - 1] == '/')
 181                 i--;
 182         if (i > maxlen)
 183                 i = maxlen;
 184         do {
 185                 i--;
 186         } while (i > 0 && path[i] != '/');
 187         return i;
 188 }
 189
 190 static void prepare_header(struct archiver_args *args,
 191                            struct ustar_header *header,
 192                            unsigned int mode, unsigned long size)
 193 {
 194         xsnprintf(header->mode, sizeof(header->mode), "%07o", mode & 07777);
 195         xsnprintf(header->size, sizeof(header->size), "%011lo", S_ISREG(mode) ? size : 0);
 196         xsnprintf(header->mtime, sizeof(header->mtime), "%011lo", (unsigned long) args->time);
 197
 198         xsnprintf(header->uid, sizeof(header->uid), "%07o", 0);
 199         xsnprintf(header->gid, sizeof(header->gid), "%07o", 0);
 200         strlcpy(header->uname, "root", sizeof(header->uname));
 201         strlcpy(header->gname, "root", sizeof(header->gname));
 202         xsnprintf(header->devmajor, sizeof(header->devmajor), "%07o", 0);
 203         xsnprintf(header->devminor, sizeof(header->devminor), "%07o", 0);
 204
 205         memcpy(header->magic, "ustar", 6);
 206         memcpy(header->version, "00", 2);
 207
 208         xsnprintf(header->chksum, sizeof(header->chksum), "%07o", ustar_header_chksum(header));
 209 }
 210
 211 static int write_extended_header(struct archiver_args *args,
 212                                  const unsigned char *sha1,
 213                                  const void *buffer, unsigned long size)
 214 {
 215         struct ustar_header header;
 216         unsigned int mode;
 217         memset(&header, 0, sizeof(header));
 218         *header.typeflag = TYPEFLAG_EXT_HEADER;
 219         mode = 0100666;
 220         xsnprintf(header.name, sizeof(header.name), "%s.paxheader", sha1_to_hex(sha1));
 221         prepare_header(args, &header, mode, size);
 222         write_blocked(&header, sizeof(header));
 223         write_blocked(buffer, size);
 224         return 0;
 225 }
 226
 227 static int write_tar_entry(struct archiver_args *args,
 228                            const unsigned char *sha1,
 229                            const char *path, size_t pathlen,
 230                            unsigned int mode)
 231 {
 232         struct ustar_header header;
 233         struct strbuf ext_header = STRBUF_INIT;
 234         unsigned int old_mode = mode;
 235         unsigned long size, size_in_header;
 236         void *buffer;
 237         int err = 0;
 238
 239         memset(&header, 0, sizeof(header));
 240
 241         if (S_ISDIR(mode) || S_ISGITLINK(mode)) {
 242                 *header.typeflag = TYPEFLAG_DIR;
 243                 mode = (mode | 0777) & ~tar_umask;
 244         } else if (S_ISLNK(mode)) {
 245                 *header.typeflag = TYPEFLAG_LNK;
 246                 mode |= 0777;
 247         } else if (S_ISREG(mode)) {
 248                 *header.typeflag = TYPEFLAG_REG;
 249                 mode = (mode | ((mode & 0100) ? 0777 : 0666)) & ~tar_umask;
 250         } else {
 251                 return error("unsupported file mode: 0%o (SHA1: %s)",
 252                              mode, sha1_to_hex(sha1));
 253         }
 254         if (pathlen > sizeof(header.name)) {
 255                 size_t plen = get_path_prefix(path, pathlen,
 256                                               sizeof(header.prefix));
 257                 size_t rest = pathlen - plen - 1;
 258                 if (plen > 0 && rest <= sizeof(header.name)) {
 259                         memcpy(header.prefix, path, plen);
 260                         memcpy(header.name, path + plen + 1, rest);
 261                 } else {
 262                         xsnprintf(header.name, sizeof(header.name), "%s.data",
 263                                   sha1_to_hex(sha1));
 264                         strbuf_append_ext_header(&ext_header, "path",
 265                                                  path, pathlen);
 266                 }
 267         } else
 268                 memcpy(header.name, path, pathlen);
 269
 270         if (S_ISREG(mode) && !args->convert &&
 271             sha1_object_info(sha1, &size) == OBJ_BLOB &&
 272             size > big_file_threshold)
 273                 buffer = NULL;
 274         else if (S_ISLNK(mode) || S_ISREG(mode)) {
 275                 enum object_type type;
 276                 buffer = sha1_file_to_archive(args, path, sha1, old_mode, &type, &size);
 277                 if (!buffer)
 278                         return error("cannot read %s", sha1_to_hex(sha1));
 279         } else {
 280                 buffer = NULL;
 281                 size = 0;
 282         }
 283
 284         if (S_ISLNK(mode)) {
 285                 if (size > sizeof(header.linkname)) {
 286                         xsnprintf(header.linkname, sizeof(header.linkname),
 287                                   "see %s.paxheader", sha1_to_hex(sha1));
 288                         strbuf_append_ext_header(&ext_header, "linkpath",
 289                                                  buffer, size);
 290                 } else
 291                         memcpy(header.linkname, buffer, size);
 292         }
 293
 294         size_in_header = size;
 295         if (S_ISREG(mode) && size > USTAR_MAX_SIZE) {
 296                 size_in_header = 0;
 297                 strbuf_append_ext_header_uint(&ext_header, "size", size);
 298         }
 299
 300         prepare_header(args, &header, mode, size_in_header);
 301
 302         if (ext_header.len > 0) {
 303                 err = write_extended_header(args, sha1, ext_header.buf,
 304                                             ext_header.len);
 305                 if (err) {
 306                         free(buffer);
 307                         return err;
 308                 }
 309         }
 310         strbuf_release(&ext_header);
 311         write_blocked(&header, sizeof(header));
 312         if (S_ISREG(mode) && size > 0) {
 313                 if (buffer)
 314                         write_blocked(buffer, size);
 315                 else
 316                         err = stream_blocked(sha1);
 317         }
 318         free(buffer);
 319         return err;
 320 }
 321
 322 static void write_global_extended_header(struct archiver_args *args)
 323 {
 324         const unsigned char *sha1 = args->commit_sha1;
 325         struct strbuf ext_header = STRBUF_INIT;
 326         struct ustar_header header;
 327         unsigned int mode;
 328
 329         if (sha1)
 330                 strbuf_append_ext_header(&ext_header, "comment",
 331                                          sha1_to_hex(sha1), 40);
 332         if (args->time > USTAR_MAX_MTIME) {
 333                 strbuf_append_ext_header_uint(&ext_header, "mtime",
 334                                               args->time);
 335                 args->time = USTAR_MAX_MTIME;
 336         }
 337
 338         if (!ext_header.len)
 339                 return;
 340
 341         memset(&header, 0, sizeof(header));
 342         *header.typeflag = TYPEFLAG_GLOBAL_HEADER;
 343         mode = 0100666;
 344         xsnprintf(header.name, sizeof(header.name), "pax_global_header");
 345         prepare_header(args, &header, mode, ext_header.len);
 346         write_blocked(&header, sizeof(header));
 347         write_blocked(ext_header.buf, ext_header.len);
 348         strbuf_release(&ext_header);
 349 }
 350
 351 static struct archiver **tar_filters;
 352 static int nr_tar_filters;
 353 static int alloc_tar_filters;
 354
 355 static struct archiver *find_tar_filter(const char *name, int len)
 356 {
 357         int i;
 358         for (i = 0; i < nr_tar_filters; i++) {
 359                 struct archiver *ar = tar_filters[i];
 360                 if (!strncmp(ar->name, name, len) && !ar->name[len])
 361                         return ar;
 362         }
 363         return NULL;
 364 }
 365
 366 static int tar_filter_config(const char *var, const char *value, void *data)
 367 {
 368         struct archiver *ar;
 369         const char *name;
 370         const char *type;
 371         int namelen;
 372
 373         if (parse_config_key(var, "tar", &name, &namelen, &type) < 0 || !name)
 374                 return 0;
 375
 376         ar = find_tar_filter(name, namelen);
 377         if (!ar) {
 378                 ar = xcalloc(1, sizeof(*ar));
 379                 ar->name = xmemdupz(name, namelen);
 380                 ar->write_archive = write_tar_filter_archive;
 381                 ar->flags = ARCHIVER_WANT_COMPRESSION_LEVELS;
 382                 ALLOC_GROW(tar_filters, nr_tar_filters + 1, alloc_tar_filters);
 383                 tar_filters[nr_tar_filters++] = ar;
 384         }
 385
 386         if (!strcmp(type, "command")) {
 387                 if (!value)
 388                         return config_error_nonbool(var);
 389                 free(ar->data);
 390                 ar->data = xstrdup(value);
 391                 return 0;
 392         }
 393         if (!strcmp(type, "remote")) {
 394                 if (git_config_bool(var, value))
 395                         ar->flags |= ARCHIVER_REMOTE;
 396                 else
 397                         ar->flags &= ~ARCHIVER_REMOTE;
 398                 return 0;
 399         }
 400
 401         return 0;
 402 }
 403
 404 static int git_tar_config(const char *var, const char *value, void *cb)
 405 {
 406         if (!strcmp(var, "tar.umask")) {
 407                 if (value && !strcmp(value, "user")) {
 408                         tar_umask = umask(0);
 409                         umask(tar_umask);
 410                 } else {
 411                         tar_umask = git_config_int(var, value);
 412                 }
 413                 return 0;
 414         }
 415
 416         return tar_filter_config(var, value, cb);
 417 }
 418
 419 static int write_tar_archive(const struct archiver *ar,
 420                              struct archiver_args *args)
 421 {
 422         int err = 0;
 423
 424         write_global_extended_header(args);
 425         err = write_archive_entries(args, write_tar_entry);
 426         if (!err)
 427                 write_trailer();
 428         return err;
 429 }
 430
 431 static int write_tar_filter_archive(const struct archiver *ar,
 432                                     struct archiver_args *args)
 433 {
 434         struct strbuf cmd = STRBUF_INIT;
 435         struct child_process filter = CHILD_PROCESS_INIT;
 436         const char *argv[2];
 437         int r;
 438
 439         if (!ar->data)
 440                 die("BUG: tar-filter archiver called with no filter defined");
 441
 442         strbuf_addstr(&cmd, ar->data);
 443         if (args->compression_level >= 0)
 444                 strbuf_addf(&cmd, " -%d", args->compression_level);
 445
 446         argv[0] = cmd.buf;
 447         argv[1] = NULL;
 448         filter.argv = argv;
 449         filter.use_shell = 1;
 450         filter.in = -1;
 451
 452         if (start_command(&filter) < 0)
 453                 die_errno("unable to start '%s' filter", argv[0]);
 454         close(1);
 455         if (dup2(filter.in, 1) < 0)
 456                 die_errno("unable to redirect descriptor");
 457         close(filter.in);
 458
 459         r = write_tar_archive(ar, args);
 460
 461         close(1);
 462         if (finish_command(&filter) != 0)
 463                 die("'%s' filter reported error", argv[0]);
 464
 465         strbuf_release(&cmd);
 466         return r;
 467 }
 468
 469 static struct archiver tar_archiver = {
 470         "tar",
 471         write_tar_archive,
 472         ARCHIVER_REMOTE
 473 };
 474
 475 void init_tar_archiver(void)
 476 {
 477         int i;
 478         register_archiver(&tar_archiver);
 479
 480         tar_filter_config("tar.tgz.command", "gzip -cn", NULL);
 481         tar_filter_config("tar.tgz.remote", "true", NULL);
 482         tar_filter_config("tar.tar.gz.command", "gzip -cn", NULL);
 483         tar_filter_config("tar.tar.gz.remote", "true", NULL);
 484         git_config(git_tar_config, NULL);
 485         for (i = 0; i < nr_tar_filters; i++) {
 486                 /* omit any filters that never had a command configured */
 487                 if (tar_filters[i]->data)
 488                         register_archiver(tar_filters[i]);
 489         }
 490 }