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