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 #if ULONG_MAX == 0xFFFFFFFF
  29 #define USTAR_MAX_SIZE ULONG_MAX
  30 #define USTAR_MAX_MTIME ULONG_MAX
  31 #else
  32 #define USTAR_MAX_SIZE 077777777777UL
  33 #define USTAR_MAX_MTIME 077777777777UL
  34 #endif
  35
  36 /* writes out the whole block, but only if it is full */
  37 static void write_if_needed(void)
  38 {
  39         if (offset == BLOCKSIZE) {
  40                 write_or_die(1, block, BLOCKSIZE);
  41                 offset = 0;
  42         }
  43 }
  44
  45 /*
  46  * queues up writes, so that all our write(2) calls write exactly one
  47  * full block; pads writes to RECORDSIZE
  48  */
  49 static void do_write_blocked(const void *data, unsigned long size)
  50 {
  51         const char *buf = data;
  52
  53         if (offset) {
  54                 unsigned long chunk = BLOCKSIZE - offset;
  55                 if (size < chunk)
  56                         chunk = size;
  57                 memcpy(block + offset, buf, chunk);
  58                 size -= chunk;
  59                 offset += chunk;
  60                 buf += chunk;
  61                 write_if_needed();
  62         }
  63         while (size >= BLOCKSIZE) {
  64                 write_or_die(1, buf, BLOCKSIZE);
  65                 size -= BLOCKSIZE;
  66                 buf += BLOCKSIZE;
  67         }
  68         if (size) {
  69                 memcpy(block + offset, buf, size);
  70                 offset += size;
  71         }
  72 }
  73
  74 static void finish_record(void)
  75 {
  76         unsigned long tail;
  77         tail = offset % RECORDSIZE;
  78         if (tail)  {
  79                 memset(block + offset, 0, RECORDSIZE - tail);
  80                 offset += RECORDSIZE - tail;
  81         }
  82         write_if_needed();
  83 }
  84
  85 static void write_blocked(const void *data, unsigned long size)
  86 {
  87         do_write_blocked(data, size);
  88         finish_record();
  89 }
  90
  91 /*
  92  * The end of tar archives is marked by 2*512 nul bytes and after that
  93  * follows the rest of the block (if any).
  94  */
  95 static void write_trailer(void)
  96 {
  97         int tail = BLOCKSIZE - offset;
  98         memset(block + offset, 0, tail);
  99         write_or_die(1, block, BLOCKSIZE);
 100         if (tail < 2 * RECORDSIZE) {
 101                 memset(block, 0, offset);
 102                 write_or_die(1, block, BLOCKSIZE);
 103         }
 104 }
 105
 106 /*
 107  * queues up writes, so that all our write(2) calls write exactly one
 108  * full block; pads writes to RECORDSIZE
 109  */
 110 static int stream_blocked(const unsigned char *sha1)
 111 {
 112         struct git_istream *st;
 113         enum object_type type;
 114         unsigned long sz;
 115         char buf[BLOCKSIZE];
 116         ssize_t readlen;
 117
 118         st = open_istream(sha1, &type, &sz, NULL);
 119         if (!st)
 120                 return error("cannot stream blob %s", sha1_to_hex(sha1));
 121         for (;;) {
 122                 readlen = read_istream(st, buf, sizeof(buf));
 123                 if (readlen <= 0)
 124                         break;
 125                 do_write_blocked(buf, readlen);
 126         }
 127         close_istream(st);
 128         if (!readlen)
 129                 finish_record();
 130         return readlen;
 131 }
 132
 133 /*
 134  * pax extended header records have the format "%u %s=%s\n".  %u contains
 135  * the size of the whole string (including the %u), the first %s is the
 136  * keyword, the second one is the value.  This function constructs such a
 137  * string and appends it to a struct strbuf.
 138  */
 139 static void strbuf_append_ext_header(struct strbuf *sb, const char *keyword,
 140                                      const char *value, unsigned int valuelen)
 141 {
 142         int len, tmp;
 143
 144         /* "%u %s=%s\n" */
 145         len = 1 + 1 + strlen(keyword) + 1 + valuelen + 1;
 146         for (tmp = len; tmp > 9; tmp /= 10)
 147                 len++;
 148
 149         strbuf_grow(sb, len);
 150         strbuf_addf(sb, "%u %s=", len, keyword);
 151         strbuf_add(sb, value, valuelen);
 152         strbuf_addch(sb, '\n');
 153 }
 154
 155 /*
 156  * Like strbuf_append_ext_header, but for numeric values.
 157  */
 158 static void strbuf_append_ext_header_uint(struct strbuf *sb,
 159                                           const char *keyword,
 160                                           uintmax_t value)
 161 {
 162         char buf[40]; /* big enough for 2^128 in decimal, plus NUL */
 163         int len;
 164
 165         len = xsnprintf(buf, sizeof(buf), "%"PRIuMAX, value);
 166         strbuf_append_ext_header(sb, keyword, buf, len);
 167 }
 168
 169 static unsigned int ustar_header_chksum(const struct ustar_header *header)
 170 {
 171         const unsigned char *p = (const unsigned char *)header;
 172         unsigned int chksum = 0;
 173         while (p < (const unsigned char *)header->chksum)
 174                 chksum += *p++;
 175         chksum += sizeof(header->chksum) * ' ';
 176         p += sizeof(header->chksum);
 177         while (p < (const unsigned char *)header + sizeof(struct ustar_header))
 178                 chksum += *p++;
 179         return chksum;
 180 }
 181
 182 static size_t get_path_prefix(const char *path, size_t pathlen, size_t maxlen)
 183 {
 184         size_t i = pathlen;
 185         if (i > 1 && path[i - 1] == '/')
 186                 i--;
 187         if (i > maxlen)
 188                 i = maxlen;
 189         do {
 190                 i--;
 191         } while (i > 0 && path[i] != '/');
 192         return i;
 193 }
 194
 195 static void prepare_header(struct archiver_args *args,
 196                            struct ustar_header *header,
 197                            unsigned int mode, unsigned long size)
 198 {
 199         xsnprintf(header->mode, sizeof(header->mode), "%07o", mode & 07777);
 200         xsnprintf(header->size, sizeof(header->size), "%011lo", S_ISREG(mode) ? size : 0);
 201         xsnprintf(header->mtime, sizeof(header->mtime), "%011lo", (unsigned long) args->time);
 202
 203         xsnprintf(header->uid, sizeof(header->uid), "%07o", 0);
 204         xsnprintf(header->gid, sizeof(header->gid), "%07o", 0);
 205         strlcpy(header->uname, "root", sizeof(header->uname));
 206         strlcpy(header->gname, "root", sizeof(header->gname));
 207         xsnprintf(header->devmajor, sizeof(header->devmajor), "%07o", 0);
 208         xsnprintf(header->devminor, sizeof(header->devminor), "%07o", 0);
 209
 210         memcpy(header->magic, "ustar", 6);
 211         memcpy(header->version, "00", 2);
 212
 213         xsnprintf(header->chksum, sizeof(header->chksum), "%07o", ustar_header_chksum(header));
 214 }
 215
 216 static void write_extended_header(struct archiver_args *args,
 217                                   const unsigned char *sha1,
 218                                   const void *buffer, unsigned long size)
 219 {
 220         struct ustar_header header;
 221         unsigned int mode;
 222         memset(&header, 0, sizeof(header));
 223         *header.typeflag = TYPEFLAG_EXT_HEADER;
 224         mode = 0100666;
 225         xsnprintf(header.name, sizeof(header.name), "%s.paxheader", sha1_to_hex(sha1));
 226         prepare_header(args, &header, mode, size);
 227         write_blocked(&header, sizeof(header));
 228         write_blocked(buffer, size);
 229 }
 230
 231 static int write_tar_entry(struct archiver_args *args,
 232                            const unsigned char *sha1,
 233                            const char *path, size_t pathlen,
 234                            unsigned int mode)
 235 {
 236         struct ustar_header header;
 237         struct strbuf ext_header = STRBUF_INIT;
 238         unsigned int old_mode = mode;
 239         unsigned long size, size_in_header;
 240         void *buffer;
 241         int err = 0;
 242
 243         memset(&header, 0, sizeof(header));
 244
 245         if (S_ISDIR(mode) || S_ISGITLINK(mode)) {
 246                 *header.typeflag = TYPEFLAG_DIR;
 247                 mode = (mode | 0777) & ~tar_umask;
 248         } else if (S_ISLNK(mode)) {
 249                 *header.typeflag = TYPEFLAG_LNK;
 250                 mode |= 0777;
 251         } else if (S_ISREG(mode)) {
 252                 *header.typeflag = TYPEFLAG_REG;
 253                 mode = (mode | ((mode & 0100) ? 0777 : 0666)) & ~tar_umask;
 254         } else {
 255                 return error("unsupported file mode: 0%o (SHA1: %s)",
 256                              mode, sha1_to_hex(sha1));
 257         }
 258         if (pathlen > sizeof(header.name)) {
 259                 size_t plen = get_path_prefix(path, pathlen,
 260                                               sizeof(header.prefix));
 261                 size_t rest = pathlen - plen - 1;
 262                 if (plen > 0 && rest <= sizeof(header.name)) {
 263                         memcpy(header.prefix, path, plen);
 264                         memcpy(header.name, path + plen + 1, rest);
 265                 } else {
 266                         xsnprintf(header.name, sizeof(header.name), "%s.data",
 267                                   sha1_to_hex(sha1));
 268                         strbuf_append_ext_header(&ext_header, "path",
 269                                                  path, pathlen);
 270                 }
 271         } else
 272                 memcpy(header.name, path, pathlen);
 273
 274         if (S_ISREG(mode) && !args->convert &&
 275             sha1_object_info(sha1, &size) == OBJ_BLOB &&
 276             size > big_file_threshold)
 277                 buffer = NULL;
 278         else if (S_ISLNK(mode) || S_ISREG(mode)) {
 279                 enum object_type type;
 280                 buffer = sha1_file_to_archive(args, path, sha1, old_mode, &type, &size);
 281                 if (!buffer)
 282                         return error("cannot read %s", sha1_to_hex(sha1));
 283         } else {
 284                 buffer = NULL;
 285                 size = 0;
 286         }
 287
 288         if (S_ISLNK(mode)) {
 289                 if (size > sizeof(header.linkname)) {
 290                         xsnprintf(header.linkname, sizeof(header.linkname),
 291                                   "see %s.paxheader", sha1_to_hex(sha1));
 292                         strbuf_append_ext_header(&ext_header, "linkpath",
 293                                                  buffer, size);
 294                 } else
 295                         memcpy(header.linkname, buffer, size);
 296         }
 297
 298         size_in_header = size;
 299         if (S_ISREG(mode) && size > USTAR_MAX_SIZE) {
 300                 size_in_header = 0;
 301                 strbuf_append_ext_header_uint(&ext_header, "size", size);
 302         }
 303
 304         prepare_header(args, &header, mode, size_in_header);
 305
 306         if (ext_header.len > 0) {
 307                 write_extended_header(args, sha1, ext_header.buf,
 308                                       ext_header.len);
 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 }