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