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(struct repository *r, 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(r, 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, size_t valuelen)
 146 {
 147         size_t orig_len = sb->len;
 148         size_t len, tmp;
 149
 150         /* "%u %s=%s\n" */
 151         len = 1 + 1 + strlen(keyword) + 1 + valuelen + 1;
 152         for (tmp = 1; len / 10 >= tmp; tmp *= 10)
 153                 len++;
 154
 155         strbuf_grow(sb, len);
 156         strbuf_addf(sb, "%"PRIuMAX" %s=", (uintmax_t)len, keyword);
 157         strbuf_add(sb, value, valuelen);
 158         strbuf_addch(sb, '\n');
 159
 160         if (len != sb->len - orig_len)
 161                 BUG("pax extended header length miscalculated as %"PRIuMAX
 162                     ", should be %"PRIuMAX,
 163                     (uintmax_t)len, (uintmax_t)(sb->len - orig_len));
 164 }
 165
 166 /*
 167  * Like strbuf_append_ext_header, but for numeric values.
 168  */
 169 static void strbuf_append_ext_header_uint(struct strbuf *sb,
 170                                           const char *keyword,
 171                                           uintmax_t value)
 172 {
 173         char buf[40]; /* big enough for 2^128 in decimal, plus NUL */
 174         int len;
 175
 176         len = xsnprintf(buf, sizeof(buf), "%"PRIuMAX, value);
 177         strbuf_append_ext_header(sb, keyword, buf, len);
 178 }
 179
 180 static unsigned int ustar_header_chksum(const struct ustar_header *header)
 181 {
 182         const unsigned char *p = (const unsigned char *)header;
 183         unsigned int chksum = 0;
 184         while (p < (const unsigned char *)header->chksum)
 185                 chksum += *p++;
 186         chksum += sizeof(header->chksum) * ' ';
 187         p += sizeof(header->chksum);
 188         while (p < (const unsigned char *)header + sizeof(struct ustar_header))
 189                 chksum += *p++;
 190         return chksum;
 191 }
 192
 193 static size_t get_path_prefix(const char *path, size_t pathlen, size_t maxlen)
 194 {
 195         size_t i = pathlen;
 196         if (i > 1 && path[i - 1] == '/')
 197                 i--;
 198         if (i > maxlen)
 199                 i = maxlen;
 200         do {
 201                 i--;
 202         } while (i > 0 && path[i] != '/');
 203         return i;
 204 }
 205
 206 static void prepare_header(struct archiver_args *args,
 207                            struct ustar_header *header,
 208                            unsigned int mode, unsigned long size)
 209 {
 210         xsnprintf(header->mode, sizeof(header->mode), "%07o", mode & 07777);
 211         xsnprintf(header->size, sizeof(header->size), "%011"PRIoMAX , S_ISREG(mode) ? (uintmax_t)size : (uintmax_t)0);
 212         xsnprintf(header->mtime, sizeof(header->mtime), "%011lo", (unsigned long) args->time);
 213
 214         xsnprintf(header->uid, sizeof(header->uid), "%07o", 0);
 215         xsnprintf(header->gid, sizeof(header->gid), "%07o", 0);
 216         strlcpy(header->uname, "root", sizeof(header->uname));
 217         strlcpy(header->gname, "root", sizeof(header->gname));
 218         xsnprintf(header->devmajor, sizeof(header->devmajor), "%07o", 0);
 219         xsnprintf(header->devminor, sizeof(header->devminor), "%07o", 0);
 220
 221         memcpy(header->magic, "ustar", 6);
 222         memcpy(header->version, "00", 2);
 223
 224         xsnprintf(header->chksum, sizeof(header->chksum), "%07o", ustar_header_chksum(header));
 225 }
 226
 227 static void write_extended_header(struct archiver_args *args,
 228                                   const struct object_id *oid,
 229                                   const void *buffer, unsigned long size)
 230 {
 231         struct ustar_header header;
 232         unsigned int mode;
 233         memset(&header, 0, sizeof(header));
 234         *header.typeflag = TYPEFLAG_EXT_HEADER;
 235         mode = 0100666;
 236         xsnprintf(header.name, sizeof(header.name), "%s.paxheader", oid_to_hex(oid));
 237         prepare_header(args, &header, mode, size);
 238         write_blocked(&header, sizeof(header));
 239         write_blocked(buffer, size);
 240 }
 241
 242 static int write_tar_entry(struct archiver_args *args,
 243                            const struct object_id *oid,
 244                            const char *path, size_t pathlen,
 245                            unsigned int mode,
 246                            void *buffer, unsigned long size)
 247 {
 248         struct ustar_header header;
 249         struct strbuf ext_header = STRBUF_INIT;
 250         unsigned long size_in_header;
 251         int err = 0;
 252
 253         memset(&header, 0, sizeof(header));
 254
 255         if (S_ISDIR(mode) || S_ISGITLINK(mode)) {
 256                 *header.typeflag = TYPEFLAG_DIR;
 257                 mode = (mode | 0777) & ~tar_umask;
 258         } else if (S_ISLNK(mode)) {
 259                 *header.typeflag = TYPEFLAG_LNK;
 260                 mode |= 0777;
 261         } else if (S_ISREG(mode)) {
 262                 *header.typeflag = TYPEFLAG_REG;
 263                 mode = (mode | ((mode & 0100) ? 0777 : 0666)) & ~tar_umask;
 264         } else {
 265                 return error(_("unsupported file mode: 0%o (SHA1: %s)"),
 266                              mode, oid_to_hex(oid));
 267         }
 268         if (pathlen > sizeof(header.name)) {
 269                 size_t plen = get_path_prefix(path, pathlen,
 270                                               sizeof(header.prefix));
 271                 size_t rest = pathlen - plen - 1;
 272                 if (plen > 0 && rest <= sizeof(header.name)) {
 273                         memcpy(header.prefix, path, plen);
 274                         memcpy(header.name, path + plen + 1, rest);
 275                 } else {
 276                         xsnprintf(header.name, sizeof(header.name), "%s.data",
 277                                   oid_to_hex(oid));
 278                         strbuf_append_ext_header(&ext_header, "path",
 279                                                  path, pathlen);
 280                 }
 281         } else
 282                 memcpy(header.name, path, pathlen);
 283
 284         if (S_ISLNK(mode)) {
 285                 if (size > sizeof(header.linkname)) {
 286                         xsnprintf(header.linkname, sizeof(header.linkname),
 287                                   "see %s.paxheader", oid_to_hex(oid));
 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                 write_extended_header(args, oid, ext_header.buf,
 304                                       ext_header.len);
 305         }
 306         strbuf_release(&ext_header);
 307         write_blocked(&header, sizeof(header));
 308         if (S_ISREG(mode) && size > 0) {
 309                 if (buffer)
 310                         write_blocked(buffer, size);
 311                 else
 312                         err = stream_blocked(args->repo, oid);
 313         }
 314         return err;
 315 }
 316
 317 static void write_global_extended_header(struct archiver_args *args)
 318 {
 319         const struct object_id *oid = args->commit_oid;
 320         struct strbuf ext_header = STRBUF_INIT;
 321         struct ustar_header header;
 322         unsigned int mode;
 323
 324         if (oid)
 325                 strbuf_append_ext_header(&ext_header, "comment",
 326                                          oid_to_hex(oid),
 327                                          the_hash_algo->hexsz);
 328         if (args->time > USTAR_MAX_MTIME) {
 329                 strbuf_append_ext_header_uint(&ext_header, "mtime",
 330                                               args->time);
 331                 args->time = USTAR_MAX_MTIME;
 332         }
 333
 334         if (!ext_header.len)
 335                 return;
 336
 337         memset(&header, 0, sizeof(header));
 338         *header.typeflag = TYPEFLAG_GLOBAL_HEADER;
 339         mode = 0100666;
 340         xsnprintf(header.name, sizeof(header.name), "pax_global_header");
 341         prepare_header(args, &header, mode, ext_header.len);
 342         write_blocked(&header, sizeof(header));
 343         write_blocked(ext_header.buf, ext_header.len);
 344         strbuf_release(&ext_header);
 345 }
 346
 347 static struct archiver **tar_filters;
 348 static int nr_tar_filters;
 349 static int alloc_tar_filters;
 350
 351 static struct archiver *find_tar_filter(const char *name, size_t len)
 352 {
 353         int i;
 354         for (i = 0; i < nr_tar_filters; i++) {
 355                 struct archiver *ar = tar_filters[i];
 356                 if (!strncmp(ar->name, name, len) && !ar->name[len])
 357                         return ar;
 358         }
 359         return NULL;
 360 }
 361
 362 static int tar_filter_config(const char *var, const char *value, void *data)
 363 {
 364         struct archiver *ar;
 365         const char *name;
 366         const char *type;
 367         size_t namelen;
 368
 369         if (parse_config_key(var, "tar", &name, &namelen, &type) < 0 || !name)
 370                 return 0;
 371
 372         ar = find_tar_filter(name, namelen);
 373         if (!ar) {
 374                 ar = xcalloc(1, sizeof(*ar));
 375                 ar->name = xmemdupz(name, namelen);
 376                 ar->write_archive = write_tar_filter_archive;
 377                 ar->flags = ARCHIVER_WANT_COMPRESSION_LEVELS |
 378                             ARCHIVER_HIGH_COMPRESSION_LEVELS;
 379                 ALLOC_GROW(tar_filters, nr_tar_filters + 1, alloc_tar_filters);
 380                 tar_filters[nr_tar_filters++] = ar;
 381         }
 382
 383         if (!strcmp(type, "command")) {
 384                 if (!value)
 385                         return config_error_nonbool(var);
 386                 free(ar->data);
 387                 ar->data = xstrdup(value);
 388                 return 0;
 389         }
 390         if (!strcmp(type, "remote")) {
 391                 if (git_config_bool(var, value))
 392                         ar->flags |= ARCHIVER_REMOTE;
 393                 else
 394                         ar->flags &= ~ARCHIVER_REMOTE;
 395                 return 0;
 396         }
 397
 398         return 0;
 399 }
 400
 401 static int git_tar_config(const char *var, const char *value, void *cb)
 402 {
 403         if (!strcmp(var, "tar.umask")) {
 404                 if (value && !strcmp(value, "user")) {
 405                         tar_umask = umask(0);
 406                         umask(tar_umask);
 407                 } else {
 408                         tar_umask = git_config_int(var, value);
 409                 }
 410                 return 0;
 411         }
 412
 413         return tar_filter_config(var, value, cb);
 414 }
 415
 416 static int write_tar_archive(const struct archiver *ar,
 417                              struct archiver_args *args)
 418 {
 419         int err = 0;
 420
 421         write_global_extended_header(args);
 422         err = write_archive_entries(args, write_tar_entry);
 423         if (!err)
 424                 write_trailer();
 425         return err;
 426 }
 427
 428 static int write_tar_filter_archive(const struct archiver *ar,
 429                                     struct archiver_args *args)
 430 {
 431         struct strbuf cmd = STRBUF_INIT;
 432         struct child_process filter = CHILD_PROCESS_INIT;
 433         const char *argv[2];
 434         int r;
 435
 436         if (!ar->data)
 437                 BUG("tar-filter archiver called with no filter defined");
 438
 439         strbuf_addstr(&cmd, ar->data);
 440         if (args->compression_level >= 0)
 441                 strbuf_addf(&cmd, " -%d", args->compression_level);
 442
 443         argv[0] = cmd.buf;
 444         argv[1] = NULL;
 445         filter.argv = argv;
 446         filter.use_shell = 1;
 447         filter.in = -1;
 448
 449         if (start_command(&filter) < 0)
 450                 die_errno(_("unable to start '%s' filter"), argv[0]);
 451         close(1);
 452         if (dup2(filter.in, 1) < 0)
 453                 die_errno(_("unable to redirect descriptor"));
 454         close(filter.in);
 455
 456         r = write_tar_archive(ar, args);
 457
 458         close(1);
 459         if (finish_command(&filter) != 0)
 460                 die(_("'%s' filter reported error"), argv[0]);
 461
 462         strbuf_release(&cmd);
 463         return r;
 464 }
 465
 466 static struct archiver tar_archiver = {
 467         "tar",
 468         write_tar_archive,
 469         ARCHIVER_REMOTE
 470 };
 471
 472 void init_tar_archiver(void)
 473 {
 474         int i;
 475         register_archiver(&tar_archiver);
 476
 477         tar_filter_config("tar.tgz.command", "gzip -cn", NULL);
 478         tar_filter_config("tar.tgz.remote", "true", NULL);
 479         tar_filter_config("tar.tar.gz.command", "gzip -cn", NULL);
 480         tar_filter_config("tar.tar.gz.remote", "true", NULL);
 481         git_config(git_tar_config, NULL);
 482         for (i = 0; i < nr_tar_filters; i++) {
 483                 /* omit any filters that never had a command configured */
 484                 if (tar_filters[i]->data)
 485                         register_archiver(tar_filters[i]);
 486         }
 487 }