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