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