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