archive-tar.c

   1 /*
   2  * Copyright (c) 2005, 2006 Rene Scharfe
   3  */
   4 #include "cache.h"
   5 #include "config.h"
   6 #include "tar.h"
   7 #include "archive.h"
   8 #include "object-store.h"
   9 #include "streaming.h"
  10 #include "run-command.h"
  11
  12 #define RECORDSIZE      (512)
  13 #define BLOCKSIZE       (RECORDSIZE * 20)
  14
  15 static char block[BLOCKSIZE];
  16 static unsigned long offset;
  17
  18 static int tar_umask = 002;
  19
  20 static int write_tar_filter_archive(const struct archiver *ar,
  21                                     struct archiver_args *args);
  22
  23 /*
  24  * This is the max value that a ustar size header can specify, as it is fixed
  25  * at 11 octal digits. POSIX specifies that we switch to extended headers at
  26  * this size.
  27  *
  28  * Likewise for the mtime (which happens to use a buffer of the same size).
  29  */
  30 #if ULONG_MAX == 0xFFFFFFFF
  31 #define USTAR_MAX_SIZE ULONG_MAX
  32 #else
  33 #define USTAR_MAX_SIZE 077777777777UL
  34 #endif
  35 #if TIME_MAX == 0xFFFFFFFF
  36 #define USTAR_MAX_MTIME TIME_MAX
  37 #else
  38 #define USTAR_MAX_MTIME 077777777777ULL
  39 #endif
  40
  41 /* writes out the whole block, but only if it is full */
  42 static void write_if_needed(void)
  43 {
  44         if (offset == BLOCKSIZE) {
  45                 write_or_die(1, block, BLOCKSIZE);
  46                 offset = 0;
  47         }
  48 }
  49
  50 /*
  51  * queues up writes, so that all our write(2) calls write exactly one
  52  * full block; pads writes to RECORDSIZE
  53  */
  54 static void do_write_blocked(const void *data, unsigned long size)
  55 {
  56         const char *buf = data;
  57
  58         if (offset) {
  59                 unsigned long chunk = BLOCKSIZE - offset;
  60                 if (size < chunk)
  61                         chunk = size;
  62                 memcpy(block + offset, buf, chunk);
  63                 size -= chunk;
  64                 offset += chunk;
  65                 buf += chunk;
  66                 write_if_needed();
  67         }
  68         while (size >= BLOCKSIZE) {
  69                 write_or_die(1, buf, BLOCKSIZE);
  70                 size -= BLOCKSIZE;
  71                 buf += BLOCKSIZE;
  72         }
  73         if (size) {
  74                 memcpy(block + offset, buf, size);
  75                 offset += size;
  76         }
  77 }
  78
  79 static void finish_record(void)
  80 {
  81         unsigned long tail;
  82         tail = offset % RECORDSIZE;
  83         if (tail)  {
  84                 memset(block + offset, 0, RECORDSIZE - tail);
  85                 offset += RECORDSIZE - tail;
  86         }
  87         write_if_needed();
  88 }
  89
  90 static void write_blocked(const void *data, unsigned long size)
  91 {
  92         do_write_blocked(data, size);
  93         finish_record();
  94 }
  95
  96 /*
  97  * The end of tar archives is marked by 2*512 nul bytes and after that
  98  * follows the rest of the block (if any).
  99  */
 100 static void write_trailer(void)
 101 {
 102         int tail = BLOCKSIZE - offset;
 103         memset(block + offset, 0, tail);
 104         write_or_die(1, block, BLOCKSIZE);
 105         if (tail < 2 * RECORDSIZE) {
 106                 memset(block, 0, offset);
 107                 write_or_die(1, block, BLOCKSIZE);
 108         }
 109 }
 110
 111 /*
 112  * queues up writes, so that all our write(2) calls write exactly one
 113  * full block; pads writes to RECORDSIZE
 114  */
 115 static int stream_blocked(const struct object_id *oid)
 116 {
 117         struct git_istream *st;
 118         enum object_type type;
 119         unsigned long sz;
 120         char buf[BLOCKSIZE];
 121         ssize_t readlen;
 122
 123         st = open_istream(oid, &type, &sz, NULL);
 124         if (!st)
 125                 return error(_("cannot stream blob %s"), oid_to_hex(oid));
 126         for (;;) {
 127                 readlen = read_istream(st, buf, sizeof(buf));
 128                 if (readlen <= 0)
 129                         break;
 130                 do_write_blocked(buf, readlen);
 131         }
 132         close_istream(st);
 133         if (!readlen)
 134                 finish_record();
 135         return readlen;
 136 }
 137
 138 /*
 139  * pax extended header records have the format "%u %s=%s\n".  %u contains
 140  * the size of the whole string (including the %u), the first %s is the
 141  * keyword, the second one is the value.  This function constructs such a
 142  * string and appends it to a struct strbuf.
 143  */
 144 static void strbuf_append_ext_header(struct strbuf *sb, const char *keyword,
 145                                      const char *value, unsigned int valuelen)
 146 {
 147         int len, tmp;
 148
 149         /* "%u %s=%s\n" */
 150         len = 1 + 1 + strlen(keyword) + 1 + valuelen + 1;
 151         for (tmp = len; tmp > 9; tmp /= 10)
 152                 len++;
 153
 154         strbuf_grow(sb, len);
 155         strbuf_addf(sb, "%u %s=", len, keyword);
 156         strbuf_add(sb, value, valuelen);
 157         strbuf_addch(sb, '\n');
 158 }
 159
 160 /*
 161  * Like strbuf_append_ext_header, but for numeric values.
 162  */
 163 static void strbuf_append_ext_header_uint(struct strbuf *sb,
 164                                           const char *keyword,
 165                                           uintmax_t value)
 166 {
 167         char buf[40]; /* big enough for 2^128 in decimal, plus NUL */
 168         int len;
 169
 170         len = xsnprintf(buf, sizeof(buf), "%"PRIuMAX, value);
 171         strbuf_append_ext_header(sb, keyword, buf, len);
 172 }
 173
 174 static unsigned int ustar_header_chksum(const struct ustar_header *header)
 175 {
 176         const unsigned char *p = (const unsigned char *)header;
 177         unsigned int chksum = 0;
 178         while (p < (const unsigned char *)header->chksum)
 179                 chksum += *p++;
 180         chksum += sizeof(header->chksum) * ' ';
 181         p += sizeof(header->chksum);
 182         while (p < (const unsigned char *)header + sizeof(struct ustar_header))
 183                 chksum += *p++;
 184         return chksum;
 185 }
 186
 187 static size_t get_path_prefix(const char *path, size_t pathlen, size_t maxlen)
 188 {
 189         size_t i = pathlen;
 190         if (i > 1 && path[i - 1] == '/')
 191                 i--;
 192         if (i > maxlen)
 193                 i = maxlen;
 194         do {
 195                 i--;
 196         } while (i > 0 && path[i] != '/');
 197         return i;
 198 }
 199
 200 static void prepare_header(struct archiver_args *args,
 201                            struct ustar_header *header,
 202                            unsigned int mode, unsigned long size)
 203 {
 204         xsnprintf(header->mode, sizeof(header->mode), "%07o", mode & 07777);
 205         xsnprintf(header->size, sizeof(header->size), "%011"PRIoMAX , S_ISREG(mode) ? (uintmax_t)size : (uintmax_t)0);
 206         xsnprintf(header->mtime, sizeof(header->mtime), "%011lo", (unsigned long) args->time);
 207
 208         xsnprintf(header->uid, sizeof(header->uid), "%07o", 0);
 209         xsnprintf(header->gid, sizeof(header->gid), "%07o", 0);
 210         strlcpy(header->uname, "root", sizeof(header->uname));
 211         strlcpy(header->gname, "root", sizeof(header->gname));
 212         xsnprintf(header->devmajor, sizeof(header->devmajor), "%07o", 0);
 213         xsnprintf(header->devminor, sizeof(header->devminor), "%07o", 0);
 214
 215         memcpy(header->magic, "ustar", 6);
 216         memcpy(header->version, "00", 2);
 217
 218         xsnprintf(header->chksum, sizeof(header->chksum), "%07o", ustar_header_chksum(header));
 219 }
 220
 221 static void write_extended_header(struct archiver_args *args,
 222                                   const struct object_id *oid,
 223                                   const void *buffer, unsigned long size)
 224 {
 225         struct ustar_header header;
 226         unsigned int mode;
 227         memset(&header, 0, sizeof(header));
 228         *header.typeflag = TYPEFLAG_EXT_HEADER;
 229         mode = 0100666;
 230         xsnprintf(header.name, sizeof(header.name), "%s.paxheader", oid_to_hex(oid));
 231         prepare_header(args, &header, mode, size);
 232         write_blocked(&header, sizeof(header));
 233         write_blocked(buffer, size);
 234 }
 235
 236 static int write_tar_entry(struct archiver_args *args,
 237                            const struct object_id *oid,
 238                            const char *path, size_t pathlen,
 239                            unsigned int mode)
 240 {
 241         struct ustar_header header;
 242         struct strbuf ext_header = STRBUF_INIT;
 243         unsigned int old_mode = mode;
 244         unsigned long size, size_in_header;
 245         void *buffer;
 246         int err = 0;
 247
 248         memset(&header, 0, sizeof(header));
 249
 250         if (S_ISDIR(mode) || S_ISGITLINK(mode)) {
 251                 *header.typeflag = TYPEFLAG_DIR;
 252                 mode = (mode | 0777) & ~tar_umask;
 253         } else if (S_ISLNK(mode)) {
 254                 *header.typeflag = TYPEFLAG_LNK;
 255                 mode |= 0777;
 256         } else if (S_ISREG(mode)) {
 257                 *header.typeflag = TYPEFLAG_REG;
 258                 mode = (mode | ((mode & 0100) ? 0777 : 0666)) & ~tar_umask;
 259         } else {
 260                 return error(_("unsupported file mode: 0%o (SHA1: %s)"),
 261                              mode, oid_to_hex(oid));
 262         }
 263         if (pathlen > sizeof(header.name)) {
 264                 size_t plen = get_path_prefix(path, pathlen,
 265                                               sizeof(header.prefix));
 266                 size_t rest = pathlen - plen - 1;
 267                 if (plen > 0 && rest <= sizeof(header.name)) {
 268                         memcpy(header.prefix, path, plen);
 269                         memcpy(header.name, path + plen + 1, rest);
 270                 } else {
 271                         xsnprintf(header.name, sizeof(header.name), "%s.data",
 272                                   oid_to_hex(oid));
 273                         strbuf_append_ext_header(&ext_header, "path",
 274                                                  path, pathlen);
 275                 }
 276         } else
 277                 memcpy(header.name, path, pathlen);
 278
 279         if (S_ISREG(mode) && !args->convert &&
 280             oid_object_info(args->repo, oid, &size) == OBJ_BLOB &&
 281             size > big_file_threshold)
 282                 buffer = NULL;
 283         else if (S_ISLNK(mode) || S_ISREG(mode)) {
 284                 enum object_type type;
 285                 buffer = object_file_to_archive(args, path, oid, old_mode, &type, &size);
 286                 if (!buffer)
 287                         return error(_("cannot read %s"), oid_to_hex(oid));
 288         } else {
 289                 buffer = NULL;
 290                 size = 0;
 291         }
 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(oid);
 322         }
 323         free(buffer);
 324         return err;
 325 }
 326
 327 static void write_global_extended_header(struct archiver_args *args)
 328 {
 329         const struct object_id *oid = args->commit_oid;
 330         struct strbuf ext_header = STRBUF_INIT;
 331         struct ustar_header header;
 332         unsigned int mode;
 333
 334         if (oid)
 335                 strbuf_append_ext_header(&ext_header, "comment",
 336                                          oid_to_hex(oid),
 337                                          the_hash_algo->hexsz);
 338         if (args->time > USTAR_MAX_MTIME) {
 339                 strbuf_append_ext_header_uint(&ext_header, "mtime",
 340                                               args->time);
 341                 args->time = USTAR_MAX_MTIME;
 342         }
 343
 344         if (!ext_header.len)
 345                 return;
 346
 347         memset(&header, 0, sizeof(header));
 348         *header.typeflag = TYPEFLAG_GLOBAL_HEADER;
 349         mode = 0100666;
 350         xsnprintf(header.name, sizeof(header.name), "pax_global_header");
 351         prepare_header(args, &header, mode, ext_header.len);
 352         write_blocked(&header, sizeof(header));
 353         write_blocked(ext_header.buf, ext_header.len);
 354         strbuf_release(&ext_header);
 355 }
 356
 357 static struct archiver **tar_filters;
 358 static int nr_tar_filters;
 359 static int alloc_tar_filters;
 360
 361 static struct archiver *find_tar_filter(const char *name, int len)
 362 {
 363         int i;
 364         for (i = 0; i < nr_tar_filters; i++) {
 365                 struct archiver *ar = tar_filters[i];
 366                 if (!strncmp(ar->name, name, len) && !ar->name[len])
 367                         return ar;
 368         }
 369         return NULL;
 370 }
 371
 372 static int tar_filter_config(const char *var, const char *value, void *data)
 373 {
 374         struct archiver *ar;
 375         const char *name;
 376         const char *type;
 377         int 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                 ar = xcalloc(1, sizeof(*ar));
 385                 ar->name = xmemdupz(name, namelen);
 386                 ar->write_archive = write_tar_filter_archive;
 387                 ar->flags = ARCHIVER_WANT_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->data);
 396                 ar->data = 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,
 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 int write_tar_filter_archive(const struct archiver *ar,
 438                                     struct archiver_args *args)
 439 {
 440         struct strbuf cmd = STRBUF_INIT;
 441         struct child_process filter = CHILD_PROCESS_INIT;
 442         const char *argv[2];
 443         int r;
 444
 445         if (!ar->data)
 446                 BUG("tar-filter archiver called with no filter defined");
 447
 448         strbuf_addstr(&cmd, ar->data);
 449         if (args->compression_level >= 0)
 450                 strbuf_addf(&cmd, " -%d", args->compression_level);
 451
 452         argv[0] = cmd.buf;
 453         argv[1] = NULL;
 454         filter.argv = argv;
 455         filter.use_shell = 1;
 456         filter.in = -1;
 457
 458         if (start_command(&filter) < 0)
 459                 die_errno(_("unable to start '%s' filter"), argv[0]);
 460         close(1);
 461         if (dup2(filter.in, 1) < 0)
 462                 die_errno(_("unable to redirect descriptor"));
 463         close(filter.in);
 464
 465         r = write_tar_archive(ar, args);
 466
 467         close(1);
 468         if (finish_command(&filter) != 0)
 469                 die(_("'%s' filter reported error"), argv[0]);
 470
 471         strbuf_release(&cmd);
 472         return r;
 473 }
 474
 475 static struct archiver tar_archiver = {
 476         "tar",
 477         write_tar_archive,
 478         ARCHIVER_REMOTE
 479 };
 480
 481 void init_tar_archiver(void)
 482 {
 483         int i;
 484         register_archiver(&tar_archiver);
 485
 486         tar_filter_config("tar.tgz.command", "gzip -cn", NULL);
 487         tar_filter_config("tar.tgz.remote", "true", NULL);
 488         tar_filter_config("tar.tar.gz.command", "gzip -cn", NULL);
 489         tar_filter_config("tar.tar.gz.remote", "true", NULL);
 490         git_config(git_tar_config, NULL);
 491         for (i = 0; i < nr_tar_filters; i++) {
 492                 /* omit any filters that never had a command configured */
 493                 if (tar_filters[i]->data)
 494                         register_archiver(tar_filters[i]);
 495         }
 496 }