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