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