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         int len;
 258
 259         if (!no_output && output_fd < 0)
 260                 return;
 261
 262         while (1) {
 263                 len = uclient_read(cl, buf, sizeof(buf));
 264                 if (!len)
 265                         return;
 266
 267                 out_bytes += len;
 268                 if (!no_output)
 269                         write(output_fd, buf, len);
 270         }
 271 }
 272
 273 static void msg_connecting(struct uclient *cl)
 274 {
 275         char addr[INET6_ADDRSTRLEN];
 276         int port;
 277
 278         if (quiet)
 279                 return;
 280
 281         uclient_get_addr(addr, &port, &cl->remote_addr);
 282         fprintf(stderr, "Connecting to %s:%d\n", addr, port);
 283 }
 284
 285 static void check_resume_offset(struct uclient *cl)
 286 {
 287         char range_str[64];
 288         struct stat st;
 289         char *file;
 290         int ret;
 291
 292         file = uclient_get_url_filename(cl->url->location, "index.html");
 293         if (!file)
 294                 return;
 295
 296         ret = stat(file, &st);
 297         free(file);
 298         if (ret)
 299                 return;
 300
 301         if (!st.st_size)
 302                 return;
 303
 304         snprintf(range_str, sizeof(range_str), "bytes=%"PRIu64"-", (uint64_t) st.st_size);
 305         uclient_http_set_header(cl, "Range", range_str);
 306 }
 307
 308 static int init_request(struct uclient *cl)
 309 {
 310         int rc;
 311
 312         out_offset = 0;
 313         out_bytes = 0;
 314         out_len = 0;
 315         uclient_http_set_ssl_ctx(cl, ssl_ops, ssl_ctx, verify);
 316
 317         if (timeout)
 318                 cl->timeout_msecs = timeout * 1000;
 319
 320         rc = uclient_connect(cl);
 321         if (rc)
 322                 return rc;
 323
 324         msg_connecting(cl);
 325
 326         rc = uclient_http_set_request_type(cl, post_data ? "POST" : "GET");
 327         if (rc)
 328                 return rc;
 329
 330         uclient_http_reset_headers(cl);
 331         uclient_http_set_header(cl, "User-Agent", user_agent);
 332         if (cur_resume)
 333                 check_resume_offset(cl);
 334
 335         if (post_data) {
 336                 uclient_http_set_header(cl, "Content-Type", "application/x-www-form-urlencoded");
 337                 uclient_write(cl, post_data, strlen(post_data));
 338         }
 339
 340         rc = uclient_request(cl);
 341         if (rc)
 342                 return rc;
 343
 344         return 0;
 345 }
 346
 347 static void request_done(struct uclient *cl)
 348 {
 349         const char *proxy_url;
 350
 351         if (n_urls) {
 352                 proxy_url = get_proxy_url(*urls);
 353                 if (proxy_url) {
 354                         uclient_set_url(cl, proxy_url, NULL);
 355                         uclient_set_proxy_url(cl, *urls, auth_str);
 356                 } else {
 357                         uclient_set_url(cl, *urls, auth_str);
 358                 }
 359                 n_urls--;
 360                 cur_resume = resume;
 361                 error_ret = init_request(cl);
 362                 if (error_ret == 0)
 363                         return;
 364         }
 365
 366         if (output_fd >= 0 && !output_file) {
 367                 close(output_fd);
 368                 output_fd = -1;
 369         }
 370         uclient_disconnect(cl);
 371         uloop_end();
 372 }
 373
 374
 375 static void eof_cb(struct uclient *cl)
 376 {
 377         if (!quiet) {
 378                 pmt_update(&pmt_timer);
 379                 uloop_timeout_cancel(&pmt_timer);
 380                 fprintf(stderr, "\n");
 381         }
 382
 383         if (!cl->data_eof) {
 384                 if (!quiet)
 385                         fprintf(stderr, "Connection reset prematurely\n");
 386                 error_ret = 4;
 387         } else if (!quiet) {
 388                 fprintf(stderr, "Download completed (%"PRIu64" bytes)\n", (uint64_t) out_bytes);
 389         }
 390         request_done(cl);
 391 }
 392
 393 static void handle_uclient_error(struct uclient *cl, int code)
 394 {
 395         const char *type = "Unknown error";
 396         bool ignore = false;
 397
 398         switch(code) {
 399         case UCLIENT_ERROR_CONNECT:
 400                 type = "Connection failed";
 401                 error_ret = 4;
 402                 break;
 403         case UCLIENT_ERROR_TIMEDOUT:
 404                 type = "Connection timed out";
 405                 error_ret = 4;
 406                 break;
 407         case UCLIENT_ERROR_SSL_INVALID_CERT:
 408                 type = "Invalid SSL certificate";
 409                 ignore = !verify;
 410                 error_ret = 5;
 411                 break;
 412         case UCLIENT_ERROR_SSL_CN_MISMATCH:
 413                 type = "Server hostname does not match SSL certificate";
 414                 ignore = !verify;
 415                 error_ret = 5;
 416                 break;
 417         default:
 418                 error_ret = 1;
 419                 break;
 420         }
 421
 422         if (!quiet)
 423                 fprintf(stderr, "Connection error: %s%s\n", type, ignore ? " (ignored)" : "");
 424
 425         if (ignore)
 426                 error_ret = 0;
 427         else
 428                 request_done(cl);
 429 }
 430
 431 static const struct uclient_cb cb = {
 432         .header_done = header_done_cb,
 433         .data_read = read_data_cb,
 434         .data_eof = eof_cb,
 435         .error = handle_uclient_error,
 436 };
 437
 438 static int usage(const char *progname)
 439 {
 440         fprintf(stderr,
 441                 "Usage: %s [options] <URL>\n"
 442                 "Options:\n"
 443                 "       -4                              Use IPv4 only\n"
 444                 "       -6                              Use IPv6 only\n"
 445                 "       -q                              Turn off status messages\n"
 446                 "       -O <file>                       Redirect output to file (use \"-\" for stdout)\n"
 447                 "       -P <dir>                        Set directory for output files\n"
 448                 "       --user=<user>                   HTTP authentication username\n"
 449                 "       --password=<password>           HTTP authentication password\n"
 450                 "       --user-agent|-U <str>           Set HTTP user agent\n"
 451                 "       --post-data=STRING              use the POST method; send STRING as the data\n"
 452                 "       --spider|-s                     Spider mode - only check file existence\n"
 453                 "       --timeout=N|-T N                Set connect/request timeout to N seconds\n"
 454                 "       --proxy=on|off|-Y on|off        Enable/disable env var configured proxy\n"
 455                 "\n"
 456                 "HTTPS options:\n"
 457                 "       --ca-certificate=<cert>         Load CA certificates from file <cert>\n"
 458                 "       --no-check-certificate          don't validate the server's certificate\n"
 459                 "\n", progname);
 460         return 1;
 461 }
 462
 463 static void init_ca_cert(void)
 464 {
 465         glob_t gl;
 466         int i;
 467
 468         glob("/etc/ssl/certs/*.crt", 0, NULL, &gl);
 469         for (i = 0; i < gl.gl_pathc; i++)
 470                 ssl_ops->context_add_ca_crt_file(ssl_ctx, gl.gl_pathv[i]);
 471 }
 472
 473 static void init_ustream_ssl(void)
 474 {
 475         void *dlh;
 476
 477         dlh = dlopen("libustream-ssl." LIB_EXT, RTLD_LAZY | RTLD_LOCAL);
 478         if (!dlh)
 479                 return;
 480
 481         ssl_ops = dlsym(dlh, "ustream_ssl_ops");
 482         if (!ssl_ops)
 483                 return;
 484
 485         ssl_ctx = ssl_ops->context_new(false);
 486 }
 487
 488 static int no_ssl(const char *progname)
 489 {
 490         fprintf(stderr, "%s: SSL support not available, please install ustream-ssl\n", progname);
 491         return 1;
 492 }
 493
 494 enum {
 495         L_NO_CHECK_CERTIFICATE,
 496         L_CA_CERTIFICATE,
 497         L_USER,
 498         L_PASSWORD,
 499         L_USER_AGENT,
 500         L_POST_DATA,
 501         L_SPIDER,
 502         L_TIMEOUT,
 503         L_CONTINUE,
 504         L_PROXY,
 505         L_NO_PROXY,
 506         L_QUIET,
 507 };
 508
 509 static const struct option longopts[] = {
 510         [L_NO_CHECK_CERTIFICATE] = { "no-check-certificate", no_argument },
 511         [L_CA_CERTIFICATE] = { "ca-certificate", required_argument },
 512         [L_USER] = { "user", required_argument },
 513         [L_PASSWORD] = { "password", required_argument },
 514         [L_USER_AGENT] = { "user-agent", required_argument },
 515         [L_POST_DATA] = { "post-data", required_argument },
 516         [L_SPIDER] = { "spider", no_argument },
 517         [L_TIMEOUT] = { "timeout", required_argument },
 518         [L_CONTINUE] = { "continue", no_argument },
 519         [L_PROXY] = { "proxy", required_argument },
 520         [L_NO_PROXY] = { "no-proxy", no_argument },
 521         [L_QUIET] = { "quiet", no_argument },
 522         {}
 523 };
 524
 525
 526
 527 int main(int argc, char **argv)
 528 {
 529         const char *progname = argv[0];
 530         const char *proxy_url;
 531         char *username = NULL;
 532         char *password = NULL;
 533         struct uclient *cl;
 534         int longopt_idx = 0;
 535         bool has_cert = false;
 536         int i, ch;
 537         int rc;
 538         int af = -1;
 539
 540         signal(SIGPIPE, SIG_IGN);
 541         init_ustream_ssl();
 542
 543         while ((ch = getopt_long(argc, argv, "46cO:P:qsT:U:Y:", longopts, &longopt_idx)) != -1) {
 544                 switch(ch) {
 545                 case 0:
 546                         switch (longopt_idx) {
 547                         case L_NO_CHECK_CERTIFICATE:
 548                                 verify = false;
 549                                 break;
 550                         case L_CA_CERTIFICATE:
 551                                 has_cert = true;
 552                                 if (ssl_ctx)
 553                                         ssl_ops->context_add_ca_crt_file(ssl_ctx, optarg);
 554                                 break;
 555                         case L_USER:
 556                                 if (!strlen(optarg))
 557                                         break;
 558                                 username = strdup(optarg);
 559                                 memset(optarg, '*', strlen(optarg));
 560                                 break;
 561                         case L_PASSWORD:
 562                                 if (!strlen(optarg))
 563                                         break;
 564                                 password = strdup(optarg);
 565                                 memset(optarg, '*', strlen(optarg));
 566                                 break;
 567                         case L_USER_AGENT:
 568                                 user_agent = optarg;
 569                                 break;
 570                         case L_POST_DATA:
 571                                 post_data = optarg;
 572                                 break;
 573                         case L_SPIDER:
 574                                 no_output = true;
 575                                 break;
 576                         case L_TIMEOUT:
 577                                 timeout = atoi(optarg);
 578                                 break;
 579                         case L_CONTINUE:
 580                                 resume = true;
 581                                 break;
 582                         case L_PROXY:
 583                                 if (strcmp(optarg, "on") != 0)
 584                                         proxy = false;
 585                                 break;
 586                         case L_NO_PROXY:
 587                                 proxy = false;
 588                                 break;
 589                         case L_QUIET:
 590                                 quiet = true;
 591                                 break;
 592                         default:
 593                                 return usage(progname);
 594                         }
 595                         break;
 596                 case '4':
 597                         af = AF_INET;
 598                         break;
 599                 case '6':
 600                         af = AF_INET6;
 601                         break;
 602                 case 'c':
 603                         resume = true;
 604                         break;
 605                 case 'U':
 606                         user_agent = optarg;
 607                         break;
 608                 case 'O':
 609                         output_file = optarg;
 610                         break;
 611                 case 'P':
 612                         if (chdir(optarg)) {
 613                                 if (!quiet)
 614                                         perror("Change output directory");
 615                                 exit(1);
 616                         }
 617                         break;
 618                 case 'q':
 619                         quiet = true;
 620                         break;
 621                 case 's':
 622                         no_output = true;
 623                         break;
 624                 case 'T':
 625                         timeout = atoi(optarg);
 626                         break;
 627                 case 'Y':
 628                         if (strcmp(optarg, "on") != 0)
 629                                 proxy = false;
 630                         break;
 631                 default:
 632                         return usage(progname);
 633                 }
 634         }
 635
 636         argv += optind;
 637         argc -= optind;
 638
 639         if (verify && !has_cert)
 640                 default_certs = true;
 641
 642         if (argc < 1)
 643                 return usage(progname);
 644
 645         if (!ssl_ctx) {
 646                 for (i = 0; i < argc; i++) {
 647                         if (!strncmp(argv[i], "https", 5))
 648                                 return no_ssl(progname);
 649                 }
 650         }
 651
 652         urls = argv + 1;
 653         n_urls = argc - 1;
 654
 655         uloop_init();
 656
 657         if (username) {
 658                 if (password)
 659                         asprintf(&auth_str, "%s:%s", username, password);
 660                 else
 661                         auth_str = username;
 662         }
 663
 664         if (!quiet)
 665                 fprintf(stderr, "Downloading '%s'\n", argv[0]);
 666
 667         proxy_url = get_proxy_url(argv[0]);
 668         if (proxy_url) {
 669                 cl = uclient_new(proxy_url, auth_str, &cb);
 670                 if (cl)
 671                     uclient_set_proxy_url(cl, argv[0], NULL);
 672         } else {
 673                 cl = uclient_new(argv[0], auth_str, &cb);
 674         }
 675         if (!cl) {
 676                 fprintf(stderr, "Failed to allocate uclient context\n");
 677                 return 1;
 678         }
 679         if (af >= 0)
 680             uclient_http_set_address_family(cl, af);
 681
 682         if (ssl_ctx && default_certs)
 683                 init_ca_cert();
 684
 685         cur_resume = resume;
 686         rc = init_request(cl);
 687         if (!rc) {
 688                 /* no error received, we can enter main loop */
 689                 uloop_run();
 690         } else {
 691                 fprintf(stderr, "Failed to establish connection\n");
 692                 error_ret = 4;
 693         }
 694
 695         uloop_done();
 696
 697         uclient_free(cl);
 698
 699         if (output_fd >= 0 && output_fd != STDOUT_FILENO)
 700                 close(output_fd);
 701
 702         if (ssl_ctx)
 703                 ssl_ops->context_free(ssl_ctx);
 704
 705         return error_ret;
 706 }