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