uclient-fetch.c

   1 /*
   2  * uclient - ustream based protocol client library
   3  *
   4  * Copyright (C) 2014 Felix Fietkau <nbd@openwrt.org>
   5  *
   6  * Permission to use, copy, modify, and/or distribute this software for any
   7  * purpose with or without fee is hereby granted, provided that the above
   8  * copyright notice and this permission notice appear in all copies.
   9  *
  10  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  13  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  14  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  15  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  16  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  17  */
  18
  19 #define _GNU_SOURCE
  20 #include <sys/stat.h>
  21 #include <sys/socket.h>
  22 #include <unistd.h>
  23 #include <stdio.h>
  24 #include <dlfcn.h>
  25 #include <getopt.h>
  26 #include <fcntl.h>
  27 #include <glob.h>
  28 #include <stdint.h>
  29 #include <inttypes.h>
  30 #include <signal.h>
  31
  32 #include <libubox/blobmsg.h>
  33
  34 #include "progress.h"
  35 #include "uclient.h"
  36 #include "uclient-utils.h"
  37
  38 #ifdef __APPLE__
  39 #define LIB_EXT "dylib"
  40 #else
  41 #define LIB_EXT "so"
  42 #endif
  43
  44 static const char *user_agent = "uclient-fetch";
  45 static const char *post_data;
  46 static struct ustream_ssl_ctx *ssl_ctx;
  47 static const struct ustream_ssl_ops *ssl_ops;
  48 static int quiet = false;
  49 static bool verify = true;
  50 static bool proxy = true;
  51 static bool default_certs = false;
  52 static bool no_output;
  53 static const char *output_file;
  54 static int output_fd = -1;
  55 static int error_ret;
  56 static off_t out_offset;
  57 static off_t out_bytes;
  58 static off_t out_len;
  59 static char *auth_str;
  60 static char **urls;
  61 static int n_urls;
  62 static int timeout;
  63 static bool resume, cur_resume;
  64
  65 static struct progress pmt;
  66 static struct uloop_timeout pmt_timer;
  67
  68 static int init_request(struct uclient *cl);
  69 static void request_done(struct uclient *cl);
  70
  71 static void pmt_update(struct uloop_timeout *t)
  72 {
  73         progress_update(&pmt, out_offset, out_bytes, out_len);
  74         uloop_timeout_set(t, 1000);
  75 }
  76
  77 static const char *
  78 get_proxy_url(char *url)
  79 {
  80         char prefix[16];
  81         char *sep;
  82
  83         if (!proxy)
  84                 return NULL;
  85
  86         sep = strchr(url, ':');
  87         if (!sep)
  88                 return NULL;
  89
  90         if (sep - url > 5)
  91                 return NULL;
  92
  93         memcpy(prefix, url, sep - url);
  94         strcpy(prefix + (sep - url), "_proxy");
  95         return getenv(prefix);
  96 }
  97
  98 static int open_output_file(const char *path, uint64_t resume_offset)
  99 {
 100         char *filename = NULL;
 101         int flags;
 102         int ret;
 103
 104         if (cur_resume)
 105                 flags = O_RDWR;
 106         else
 107                 flags = O_WRONLY | O_TRUNC;
 108
 109         if (!cur_resume && !output_file)
 110                 flags |= O_EXCL;
 111
 112         flags |= O_CREAT;
 113
 114         if (output_file) {
 115                 if (!strcmp(output_file, "-")) {
 116                         if (!quiet)
 117                                 fprintf(stderr, "Writing to stdout\n");
 118
 119                         ret = STDOUT_FILENO;
 120                         goto done;
 121                 }
 122         } else {
 123                 filename = uclient_get_url_filename(path, "index.html");
 124                 output_file = filename;
 125         }
 126
 127         if (!quiet)
 128                 fprintf(stderr, "Writing to '%s'\n", output_file);
 129         ret = open(output_file, flags, 0644);
 130         if (ret < 0)
 131                 goto free;
 132
 133         if (resume_offset &&
 134             lseek(ret, resume_offset, SEEK_SET) < 0) {
 135                 if (!quiet)
 136                         fprintf(stderr, "Failed to seek %"PRIu64" bytes in output file\n", resume_offset);
 137                 close(ret);
 138                 ret = -1;
 139                 goto free;
 140         }
 141
 142         out_offset = resume_offset;
 143         out_bytes += resume_offset;
 144 done:
 145         if (!quiet) {
 146                 progress_init(&pmt, output_file);
 147                 pmt_timer.cb = pmt_update;
 148                 pmt_timer.cb(&pmt_timer);
 149         }
 150
 151 free:
 152         free(filename);
 153         return ret;
 154 }
 155
 156 static void header_done_cb(struct uclient *cl)
 157 {
 158         enum {
 159                 H_RANGE,
 160                 H_LEN,
 161                 __H_MAX
 162         };
 163         static const struct blobmsg_policy policy[__H_MAX] = {
 164                 [H_RANGE] = { .name = "content-range", .type = BLOBMSG_TYPE_STRING },
 165                 [H_LEN] = { .name = "content-length", .type = BLOBMSG_TYPE_STRING },
 166         };
 167         struct blob_attr *tb[__H_MAX];
 168         uint64_t resume_offset = 0, resume_end, resume_size;
 169         static int retries;
 170
 171         if (retries < 10) {
 172                 int ret = uclient_http_redirect(cl);
 173                 if (ret < 0) {
 174                         if (!quiet)
 175                                 fprintf(stderr, "Failed to redirect to %s on %s\n", cl->url->location, cl->url->host);
 176                         error_ret = 8;
 177                         request_done(cl);
 178                         return;
 179                 }
 180                 if (ret > 0) {
 181                         if (!quiet)
 182                                 fprintf(stderr, "Redirected to %s on %s\n", cl->url->location, cl->url->host);
 183
 184                         retries++;
 185                         return;
 186                 }
 187         }
 188
 189         if (cl->status_code == 204 && cur_resume) {
 190                 /* Resume attempt failed, try normal download */
 191                 cur_resume = false;
 192                 init_request(cl);
 193                 return;
 194         }
 195
 196         blobmsg_parse(policy, __H_MAX, tb, blob_data(cl->meta), blob_len(cl->meta));
 197
 198         switch (cl->status_code) {
 199         case 416:
 200                 if (!quiet)
 201                         fprintf(stderr, "File download already fully retrieved; nothing to do.\n");
 202                 request_done(cl);
 203                 break;
 204         case 206:
 205                 if (!cur_resume) {
 206                         if (!quiet)
 207                                 fprintf(stderr, "Error: Partial content received, full content requested\n");
 208                         error_ret = 8;
 209                         request_done(cl);
 210                         break;
 211                 }
 212
 213                 if (!tb[H_RANGE]) {
 214                         if (!quiet)
 215                                 fprintf(stderr, "Content-Range header is missing\n");
 216                         error_ret = 8;
 217                         break;
 218                 }
 219
 220                 if (sscanf(blobmsg_get_string(tb[H_RANGE]),
 221                            "bytes %"PRIu64"-%"PRIu64"/%"PRIu64,
 222                            &resume_offset, &resume_end, &resume_size) != 3) {
 223                         if (!quiet)
 224                                 fprintf(stderr, "Content-Range header is invalid\n");
 225                         error_ret = 8;
 226                         break;
 227                 }
 228         case 204:
 229         case 200:
 230                 if (no_output)
 231                         break;
 232
 233                 if (tb[H_LEN])
 234                         out_len = strtoul(blobmsg_get_string(tb[H_LEN]), NULL, 10);
 235
 236                 output_fd = open_output_file(cl->url->location, resume_offset);
 237                 if (output_fd < 0) {
 238                         if (!quiet)
 239                                 perror("Cannot open output file");
 240                         error_ret = 3;
 241                         request_done(cl);
 242                 }
 243                 break;
 244
 245         default:
 246                 if (!quiet)
 247                         fprintf(stderr, "HTTP error %d\n", cl->status_code);
 248                 request_done(cl);
 249                 error_ret = 8;
 250                 break;
 251         }
 252 }
 253
 254 static void read_data_cb(struct uclient *cl)
 255 {
 256         char buf[256];
 257         ssize_t n;
 258         int len;
 259
 260         if (!no_output && output_fd < 0)
 261                 return;
 262
 263         while (1) {
 264                 len = uclient_read(cl, buf, sizeof(buf));
 265                 if (len <= 0)
 266                         return;
 267
 268                 out_bytes += len;
 269                 if (!no_output) {
 270                         n = write(output_fd, buf, len);
 271                         if (n < 0)
 272                                 return;
 273                 }
 274         }
 275 }
 276
 277 static void msg_connecting(struct uclient *cl)
 278 {
 279         char addr[INET6_ADDRSTRLEN];
 280         int port;
 281
 282         if (quiet)
 283                 return;
 284
 285         uclient_get_addr(addr, &port, &cl->remote_addr);
 286         fprintf(stderr, "Connecting to %s:%d\n", addr, port);
 287 }
 288
 289 static void check_resume_offset(struct uclient *cl)
 290 {
 291         char range_str[64];
 292         struct stat st;
 293         char *file;
 294         int ret;
 295
 296         file = uclient_get_url_filename(cl->url->location, "index.html");
 297         if (!file)
 298                 return;
 299
 300         ret = stat(file, &st);
 301         free(file);
 302         if (ret)
 303                 return;
 304
 305         if (!st.st_size)
 306                 return;
 307
 308         snprintf(range_str, sizeof(range_str), "bytes=%"PRIu64"-", (uint64_t) st.st_size);
 309         uclient_http_set_header(cl, "Range", range_str);
 310 }
 311
 312 static int init_request(struct uclient *cl)
 313 {
 314         int rc;
 315
 316         out_offset = 0;
 317         out_bytes = 0;
 318         out_len = 0;
 319         uclient_http_set_ssl_ctx(cl, ssl_ops, ssl_ctx, verify);
 320
 321         if (timeout)
 322                 cl->timeout_msecs = timeout * 1000;
 323
 324         rc = uclient_connect(cl);
 325         if (rc)
 326                 return rc;
 327
 328         msg_connecting(cl);
 329
 330         rc = uclient_http_set_request_type(cl, post_data ? "POST" : "GET");
 331         if (rc)
 332                 return rc;
 333
 334         uclient_http_reset_headers(cl);
 335         uclient_http_set_header(cl, "User-Agent", user_agent);
 336         if (cur_resume)
 337                 check_resume_offset(cl);
 338
 339         if (post_data) {
 340                 uclient_http_set_header(cl, "Content-Type", "application/x-www-form-urlencoded");
 341                 uclient_write(cl, post_data, strlen(post_data));
 342         }
 343
 344         rc = uclient_request(cl);
 345         if (rc)
 346                 return rc;
 347
 348         return 0;
 349 }
 350
 351 static void request_done(struct uclient *cl)
 352 {
 353         const char *proxy_url;
 354
 355         if (n_urls) {
 356                 proxy_url = get_proxy_url(*urls);
 357                 if (proxy_url) {
 358                         uclient_set_url(cl, proxy_url, NULL);
 359                         uclient_set_proxy_url(cl, *urls, auth_str);
 360                 } else {
 361                         uclient_set_url(cl, *urls, auth_str);
 362                 }
 363                 n_urls--;
 364                 cur_resume = resume;
 365                 error_ret = init_request(cl);
 366                 if (error_ret == 0)
 367                         return;
 368         }
 369
 370         if (output_fd >= 0 && !output_file) {
 371                 close(output_fd);
 372                 output_fd = -1;
 373         }
 374         uclient_disconnect(cl);
 375         uloop_end();
 376 }
 377
 378
 379 static void eof_cb(struct uclient *cl)
 380 {
 381         if (!quiet) {
 382                 pmt_update(&pmt_timer);
 383                 uloop_timeout_cancel(&pmt_timer);
 384                 fprintf(stderr, "\n");
 385         }
 386
 387         if (!cl->data_eof) {
 388                 if (!quiet)
 389                         fprintf(stderr, "Connection reset prematurely\n");
 390                 error_ret = 4;
 391         } else if (!quiet) {
 392                 fprintf(stderr, "Download completed (%"PRIu64" bytes)\n", (uint64_t) out_bytes);
 393         }
 394         request_done(cl);
 395 }
 396
 397 static void handle_uclient_error(struct uclient *cl, int code)
 398 {
 399         const char *type = "Unknown error";
 400         bool ignore = false;
 401
 402         switch(code) {
 403         case UCLIENT_ERROR_CONNECT:
 404                 type = "Connection failed";
 405                 error_ret = 4;
 406                 break;
 407         case UCLIENT_ERROR_TIMEDOUT:
 408                 type = "Connection timed out";
 409                 error_ret = 4;
 410                 break;
 411         case UCLIENT_ERROR_SSL_INVALID_CERT:
 412                 type = "Invalid SSL certificate";
 413                 ignore = !verify;
 414                 error_ret = 5;
 415                 break;
 416         case UCLIENT_ERROR_SSL_CN_MISMATCH:
 417                 type = "Server hostname does not match SSL certificate";
 418                 ignore = !verify;
 419                 error_ret = 5;
 420                 break;
 421         default:
 422                 error_ret = 1;
 423                 break;
 424         }
 425
 426         if (!quiet)
 427                 fprintf(stderr, "Connection error: %s%s\n", type, ignore ? " (ignored)" : "");
 428
 429         if (ignore)
 430                 error_ret = 0;
 431         else
 432                 request_done(cl);
 433 }
 434
 435 static const struct uclient_cb cb = {
 436         .header_done = header_done_cb,
 437         .data_read = read_data_cb,
 438         .data_eof = eof_cb,
 439         .error = handle_uclient_error,
 440 };
 441
 442 static int usage(const char *progname)
 443 {
 444         fprintf(stderr,
 445                 "Usage: %s [options] <URL>\n"
 446                 "Options:\n"
 447                 "       -4                              Use IPv4 only\n"
 448                 "       -6                              Use IPv6 only\n"
 449                 "       -q                              Turn off status messages\n"
 450                 "       -O <file>                       Redirect output to file (use \"-\" for stdout)\n"
 451                 "       -P <dir>                        Set directory for output files\n"
 452                 "       --user=<user>                   HTTP authentication username\n"
 453                 "       --password=<password>           HTTP authentication password\n"
 454                 "       --user-agent|-U <str>           Set HTTP user agent\n"
 455                 "       --post-data=STRING              use the POST method; send STRING as the data\n"
 456                 "       --spider|-s                     Spider mode - only check file existence\n"
 457                 "       --timeout=N|-T N                Set connect/request timeout to N seconds\n"
 458                 "       --proxy=on|off|-Y on|off        Enable/disable env var configured proxy\n"
 459                 "\n"
 460                 "HTTPS options:\n"
 461                 "       --ca-certificate=<cert>         Load CA certificates from file <cert>\n"
 462                 "       --no-check-certificate          don't validate the server's certificate\n"
 463                 "\n", progname);
 464         return 1;
 465 }
 466
 467 static void init_ca_cert(void)
 468 {
 469         glob_t gl;
 470         int i;
 471
 472         glob("/etc/ssl/certs/*.crt", 0, NULL, &gl);
 473         for (i = 0; i < gl.gl_pathc; i++)
 474                 ssl_ops->context_add_ca_crt_file(ssl_ctx, gl.gl_pathv[i]);
 475 }
 476
 477 static void init_ustream_ssl(void)
 478 {
 479         void *dlh;
 480
 481         dlh = dlopen("libustream-ssl." LIB_EXT, RTLD_LAZY | RTLD_LOCAL);
 482         if (!dlh)
 483                 return;
 484
 485         ssl_ops = dlsym(dlh, "ustream_ssl_ops");
 486         if (!ssl_ops)
 487                 return;
 488
 489         ssl_ctx = ssl_ops->context_new(false);
 490 }
 491
 492 static int no_ssl(const char *progname)
 493 {
 494         fprintf(stderr,
 495                 "%s: SSL support not available, please install one of the "
 496                 "libustream-ssl-* libraries as well as the ca-bundle and "
 497                 "ca-certificates packages.\n",
 498                 progname);
 499
 500         return 1;
 501 }
 502
 503 enum {
 504         L_NO_CHECK_CERTIFICATE,
 505         L_CA_CERTIFICATE,
 506         L_USER,
 507         L_PASSWORD,
 508         L_USER_AGENT,
 509         L_POST_DATA,
 510         L_SPIDER,
 511         L_TIMEOUT,
 512         L_CONTINUE,
 513         L_PROXY,
 514         L_NO_PROXY,
 515         L_QUIET,
 516 };
 517
 518 static const struct option longopts[] = {
 519         [L_NO_CHECK_CERTIFICATE] = { "no-check-certificate", no_argument },
 520         [L_CA_CERTIFICATE] = { "ca-certificate", required_argument },
 521         [L_USER] = { "user", required_argument },
 522         [L_PASSWORD] = { "password", required_argument },
 523         [L_USER_AGENT] = { "user-agent", required_argument },
 524         [L_POST_DATA] = { "post-data", required_argument },
 525         [L_SPIDER] = { "spider", no_argument },
 526         [L_TIMEOUT] = { "timeout", required_argument },
 527         [L_CONTINUE] = { "continue", no_argument },
 528         [L_PROXY] = { "proxy", required_argument },
 529         [L_NO_PROXY] = { "no-proxy", no_argument },
 530         [L_QUIET] = { "quiet", no_argument },
 531         {}
 532 };
 533
 534
 535
 536 int main(int argc, char **argv)
 537 {
 538         const char *progname = argv[0];
 539         const char *proxy_url;
 540         char *username = NULL;
 541         char *password = NULL;
 542         struct uclient *cl;
 543         int longopt_idx = 0;
 544         bool has_cert = false;
 545         int i, ch;
 546         int rc;
 547         int af = -1;
 548
 549         signal(SIGPIPE, SIG_IGN);
 550         init_ustream_ssl();
 551
 552         while ((ch = getopt_long(argc, argv, "46cO:P:qsT:U:Y:", longopts, &longopt_idx)) != -1) {
 553                 switch(ch) {
 554                 case 0:
 555                         switch (longopt_idx) {
 556                         case L_NO_CHECK_CERTIFICATE:
 557                                 verify = false;
 558                                 break;
 559                         case L_CA_CERTIFICATE:
 560                                 has_cert = true;
 561                                 if (ssl_ctx)
 562                                         ssl_ops->context_add_ca_crt_file(ssl_ctx, optarg);
 563                                 break;
 564                         case L_USER:
 565                                 if (!strlen(optarg))
 566                                         break;
 567                                 username = strdup(optarg);
 568                                 memset(optarg, '*', strlen(optarg));
 569                                 break;
 570                         case L_PASSWORD:
 571                                 if (!strlen(optarg))
 572                                         break;
 573                                 password = strdup(optarg);
 574                                 memset(optarg, '*', strlen(optarg));
 575                                 break;
 576                         case L_USER_AGENT:
 577                                 user_agent = optarg;
 578                                 break;
 579                         case L_POST_DATA:
 580                                 post_data = optarg;
 581                                 break;
 582                         case L_SPIDER:
 583                                 no_output = true;
 584                                 break;
 585                         case L_TIMEOUT:
 586                                 timeout = atoi(optarg);
 587                                 break;
 588                         case L_CONTINUE:
 589                                 resume = true;
 590                                 break;
 591                         case L_PROXY:
 592                                 if (strcmp(optarg, "on") != 0)
 593                                         proxy = false;
 594                                 break;
 595                         case L_NO_PROXY:
 596                                 proxy = false;
 597                                 break;
 598                         case L_QUIET:
 599                                 quiet = true;
 600                                 break;
 601                         default:
 602                                 return usage(progname);
 603                         }
 604                         break;
 605                 case '4':
 606                         af = AF_INET;
 607                         break;
 608                 case '6':
 609                         af = AF_INET6;
 610                         break;
 611                 case 'c':
 612                         resume = true;
 613                         break;
 614                 case 'U':
 615                         user_agent = optarg;
 616                         break;
 617                 case 'O':
 618                         output_file = optarg;
 619                         break;
 620                 case 'P':
 621                         if (chdir(optarg)) {
 622                                 if (!quiet)
 623                                         perror("Change output directory");
 624                                 exit(1);
 625                         }
 626                         break;
 627                 case 'q':
 628                         quiet = true;
 629                         break;
 630                 case 's':
 631                         no_output = true;
 632                         break;
 633                 case 'T':
 634                         timeout = atoi(optarg);
 635                         break;
 636                 case 'Y':
 637                         if (strcmp(optarg, "on") != 0)
 638                                 proxy = false;
 639                         break;
 640                 default:
 641                         return usage(progname);
 642                 }
 643         }
 644
 645         argv += optind;
 646         argc -= optind;
 647
 648         if (verify && !has_cert)
 649                 default_certs = true;
 650
 651         if (argc < 1)
 652                 return usage(progname);
 653
 654         if (!ssl_ctx) {
 655                 for (i = 0; i < argc; i++) {
 656                         if (!strncmp(argv[i], "https", 5))
 657                                 return no_ssl(progname);
 658                 }
 659         }
 660
 661         urls = argv + 1;
 662         n_urls = argc - 1;
 663
 664         uloop_init();
 665
 666         if (username) {
 667                 if (password) {
 668                         rc = asprintf(&auth_str, "%s:%s", username, password);
 669                         if (rc < 0)
 670                                 return rc;
 671                 } else
 672                         auth_str = username;
 673         }
 674
 675         if (!quiet)
 676                 fprintf(stderr, "Downloading '%s'\n", argv[0]);
 677
 678         proxy_url = get_proxy_url(argv[0]);
 679         if (proxy_url) {
 680                 cl = uclient_new(proxy_url, auth_str, &cb);
 681                 if (cl)
 682                     uclient_set_proxy_url(cl, argv[0], NULL);
 683         } else {
 684                 cl = uclient_new(argv[0], auth_str, &cb);
 685         }
 686         if (!cl) {
 687                 fprintf(stderr, "Failed to allocate uclient context\n");
 688                 return 1;
 689         }
 690         if (af >= 0)
 691             uclient_http_set_address_family(cl, af);
 692
 693         if (ssl_ctx && default_certs)
 694                 init_ca_cert();
 695
 696         cur_resume = resume;
 697         rc = init_request(cl);
 698         if (!rc) {
 699                 /* no error received, we can enter main loop */
 700                 uloop_run();
 701         } else {
 702                 fprintf(stderr, "Failed to establish connection\n");
 703                 error_ret = 4;
 704         }
 705
 706         uloop_done();
 707
 708         uclient_free(cl);
 709
 710         if (output_fd >= 0 && output_fd != STDOUT_FILENO)
 711                 close(output_fd);
 712
 713         if (ssl_ctx)
 714                 ssl_ops->context_free(ssl_ctx);
 715
 716         return error_ret;
 717 }