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 *auth_str;
  54 static char **urls;
  55 static int n_urls;
  56 static int timeout;
  57 static bool resume, cur_resume;
  58
  59 static int init_request(struct uclient *cl);
  60 static void request_done(struct uclient *cl);
  61
  62 static int open_output_file(const char *path, uint64_t resume_offset)
  63 {
  64         char *filename = NULL;
  65         int flags;
  66         int ret;
  67
  68         if (cur_resume)
  69                 flags = O_RDWR;
  70         else
  71                 flags = O_WRONLY | O_EXCL;
  72
  73         flags |= O_CREAT;
  74
  75         if (output_file) {
  76                 if (!strcmp(output_file, "-")) {
  77                         if (!quiet)
  78                                 fprintf(stderr, "Writing to stdout\n");
  79
  80                         return STDOUT_FILENO;
  81                 }
  82         } else {
  83                 filename = uclient_get_url_filename(path, "index.html");
  84                 output_file = filename;
  85         }
  86
  87         if (!quiet)
  88                 fprintf(stderr, "Writing to '%s'\n", output_file);
  89         ret = open(output_file, flags, 0644);
  90         free(filename);
  91
  92         if (ret < 0)
  93                 return ret;
  94
  95         if (resume_offset &&
  96             lseek(ret, resume_offset, SEEK_SET) < 0) {
  97                 if (!quiet)
  98                         fprintf(stderr, "Failed to seek %"PRIu64" bytes in output file\n", resume_offset);
  99                 close(ret);
 100                 return -1;
 101         }
 102
 103         out_bytes += resume_offset;
 104
 105         return ret;
 106 }
 107
 108 static void header_done_cb(struct uclient *cl)
 109 {
 110         static const struct blobmsg_policy policy = {
 111                 .name = "content-range",
 112                 .type = BLOBMSG_TYPE_STRING
 113         };
 114         struct blob_attr *attr;
 115         uint64_t resume_offset = 0, resume_end, resume_size;
 116         static int retries;
 117
 118         if (retries < 10 && uclient_http_redirect(cl)) {
 119                 if (!quiet)
 120                         fprintf(stderr, "Redirected to %s on %s\n", cl->url->location, cl->url->host);
 121
 122                 retries++;
 123                 return;
 124         }
 125
 126         if (cl->status_code == 204 && cur_resume) {
 127                 /* Resume attempt failed, try normal download */
 128                 cur_resume = false;
 129                 init_request(cl);
 130                 return;
 131         }
 132
 133         switch (cl->status_code) {
 134         case 416:
 135                 if (!quiet)
 136                         fprintf(stderr, "File download already fully retrieved; nothing to do.\n");
 137                 request_done(cl);
 138                 break;
 139         case 206:
 140                 if (!cur_resume) {
 141                         if (!quiet)
 142                                 fprintf(stderr, "Error: Partial content received, full content requested\n");
 143                         error_ret = 8;
 144                         request_done(cl);
 145                         break;
 146                 }
 147
 148                 blobmsg_parse(&policy, 1, &attr, blob_data(cl->meta), blob_len(cl->meta));
 149                 if (!attr) {
 150                         if (!quiet)
 151                                 fprintf(stderr, "Content-Range header is missing\n");
 152                         error_ret = 8;
 153                         break;
 154                 }
 155
 156                 if (sscanf(blobmsg_get_string(attr), "bytes %"PRIu64"-%"PRIu64"/%"PRIu64,
 157                            &resume_offset, &resume_end, &resume_size) != 3) {
 158                         if (!quiet)
 159                                 fprintf(stderr, "Content-Range header is invalid\n");
 160                         error_ret = 8;
 161                         break;
 162                 }
 163         case 204:
 164         case 200:
 165                 if (no_output)
 166                         break;
 167                 output_fd = open_output_file(cl->url->location, resume_offset);
 168                 if (output_fd < 0) {
 169                         if (!quiet)
 170                                 perror("Cannot open output file");
 171                         error_ret = 3;
 172                         request_done(cl);
 173                 }
 174                 break;
 175
 176         default:
 177                 if (!quiet)
 178                         fprintf(stderr, "HTTP error %d\n", cl->status_code);
 179                 request_done(cl);
 180                 error_ret = 8;
 181                 break;
 182         }
 183 }
 184
 185 static void read_data_cb(struct uclient *cl)
 186 {
 187         char buf[256];
 188         int len;
 189
 190         if (!no_output && output_fd < 0)
 191                 return;
 192
 193         while (1) {
 194                 len = uclient_read(cl, buf, sizeof(buf));
 195                 if (!len)
 196                         return;
 197
 198                 out_bytes += len;
 199                 if (!no_output)
 200                         write(output_fd, buf, len);
 201         }
 202 }
 203
 204 static void msg_connecting(struct uclient *cl)
 205 {
 206         char addr[INET6_ADDRSTRLEN];
 207         int port;
 208
 209         if (quiet)
 210                 return;
 211
 212         uclient_get_addr(addr, &port, &cl->remote_addr);
 213         fprintf(stderr, "Connecting to %s:%d\n", addr, port);
 214 }
 215
 216 static void check_resume_offset(struct uclient *cl)
 217 {
 218         char range_str[64];
 219         struct stat st;
 220         char *file;
 221         int ret;
 222
 223         file = uclient_get_url_filename(cl->url->location, "index.html");
 224         if (!file)
 225                 return;
 226
 227         ret = stat(file, &st);
 228         free(file);
 229         if (ret)
 230                 return;
 231
 232         if (!st.st_size)
 233                 return;
 234
 235         snprintf(range_str, sizeof(range_str), "bytes=%"PRIu64"-", (uint64_t) st.st_size);
 236         uclient_http_set_header(cl, "Range", range_str);
 237 }
 238
 239 static int init_request(struct uclient *cl)
 240 {
 241         int rc;
 242
 243         out_bytes = 0;
 244         uclient_http_set_ssl_ctx(cl, ssl_ops, ssl_ctx, verify);
 245
 246         if (timeout)
 247                 cl->timeout_msecs = timeout * 1000;
 248
 249         rc = uclient_connect(cl);
 250         if (rc)
 251                 return rc;
 252
 253         msg_connecting(cl);
 254
 255         rc = uclient_http_set_request_type(cl, post_data ? "POST" : "GET");
 256         if (rc)
 257                 return rc;
 258
 259         uclient_http_reset_headers(cl);
 260         uclient_http_set_header(cl, "User-Agent", user_agent);
 261         if (cur_resume)
 262                 check_resume_offset(cl);
 263
 264         if (post_data) {
 265                 uclient_http_set_header(cl, "Content-Type", "application/x-www-form-urlencoded");
 266                 uclient_write(cl, post_data, strlen(post_data));
 267         }
 268
 269         rc = uclient_request(cl);
 270         if (rc)
 271                 return rc;
 272
 273         return 0;
 274 }
 275
 276 static void request_done(struct uclient *cl)
 277 {
 278         if (n_urls) {
 279                 uclient_set_url(cl, *urls, auth_str);
 280                 n_urls--;
 281                 cur_resume = resume;
 282                 error_ret = init_request(cl);
 283                 if (error_ret == 0)
 284                         return;
 285         }
 286
 287         if (output_fd >= 0 && !output_file) {
 288                 close(output_fd);
 289                 output_fd = -1;
 290         }
 291         uclient_disconnect(cl);
 292         uloop_end();
 293 }
 294
 295
 296 static void eof_cb(struct uclient *cl)
 297 {
 298         if (!cl->data_eof) {
 299                 if (!quiet)
 300                         fprintf(stderr, "Connection reset prematurely\n");
 301                 error_ret = 4;
 302         } else if (!quiet) {
 303                 fprintf(stderr, "Download completed (%d bytes)\n", out_bytes);
 304         }
 305         request_done(cl);
 306 }
 307
 308 static void handle_uclient_error(struct uclient *cl, int code)
 309 {
 310         const char *type = "Unknown error";
 311         bool ignore = false;
 312
 313         switch(code) {
 314         case UCLIENT_ERROR_CONNECT:
 315                 type = "Connection failed";
 316                 error_ret = 4;
 317                 break;
 318         case UCLIENT_ERROR_TIMEDOUT:
 319                 type = "Connection timed out";
 320                 error_ret = 4;
 321                 break;
 322         case UCLIENT_ERROR_SSL_INVALID_CERT:
 323                 type = "Invalid SSL certificate";
 324                 ignore = !verify;
 325                 error_ret = 5;
 326                 break;
 327         case UCLIENT_ERROR_SSL_CN_MISMATCH:
 328                 type = "Server hostname does not match SSL certificate";
 329                 ignore = !verify;
 330                 error_ret = 5;
 331                 break;
 332         default:
 333                 error_ret = 1;
 334                 break;
 335         }
 336
 337         if (!quiet)
 338                 fprintf(stderr, "Connection error: %s%s\n", type, ignore ? " (ignored)" : "");
 339
 340         if (ignore)
 341                 error_ret = 0;
 342         else
 343                 request_done(cl);
 344 }
 345
 346 static const struct uclient_cb cb = {
 347         .header_done = header_done_cb,
 348         .data_read = read_data_cb,
 349         .data_eof = eof_cb,
 350         .error = handle_uclient_error,
 351 };
 352
 353 static int usage(const char *progname)
 354 {
 355         fprintf(stderr,
 356                 "Usage: %s [options] <URL>\n"
 357                 "Options:\n"
 358                 "       -q:                             Turn off status messages\n"
 359                 "       -O <file>:                      Redirect output to file (use \"-\" for stdout)\n"
 360                 "       --user=<user>                   HTTP authentication username\n"
 361                 "       --password=<password>           HTTP authentication password\n"
 362                 "       --user-agent|-U <str>           Set HTTP user agent\n"
 363                 "       --post-data=STRING              use the POST method; send STRING as the data\n"
 364                 "       --spider|-s                     Spider mode - only check file existence\n"
 365                 "       --timeout=N|-T N                Set connect/request timeout to N seconds\n"
 366                 "\n"
 367                 "HTTPS options:\n"
 368                 "       --ca-certificate=<cert>:        Load CA certificates from file <cert>\n"
 369                 "       --no-check-certificate:         don't validate the server's certificate\n"
 370                 "\n", progname);
 371         return 1;
 372 }
 373
 374 static void init_ca_cert(void)
 375 {
 376         glob_t gl;
 377         int i;
 378
 379         glob("/etc/ssl/certs/*.crt", 0, NULL, &gl);
 380         for (i = 0; i < gl.gl_pathc; i++)
 381                 ssl_ops->context_add_ca_crt_file(ssl_ctx, gl.gl_pathv[i]);
 382 }
 383
 384 static void init_ustream_ssl(void)
 385 {
 386         void *dlh;
 387
 388         dlh = dlopen("libustream-ssl." LIB_EXT, RTLD_LAZY | RTLD_LOCAL);
 389         if (!dlh)
 390                 return;
 391
 392         ssl_ops = dlsym(dlh, "ustream_ssl_ops");
 393         if (!ssl_ops)
 394                 return;
 395
 396         ssl_ctx = ssl_ops->context_new(false);
 397 }
 398
 399 static int no_ssl(const char *progname)
 400 {
 401         fprintf(stderr, "%s: SSL support not available, please install ustream-ssl\n", progname);
 402         return 1;
 403 }
 404
 405 enum {
 406         L_NO_CHECK_CERTIFICATE,
 407         L_CA_CERTIFICATE,
 408         L_USER,
 409         L_PASSWORD,
 410         L_USER_AGENT,
 411         L_POST_DATA,
 412         L_SPIDER,
 413         L_TIMEOUT,
 414         L_CONTINUE,
 415 };
 416
 417 static const struct option longopts[] = {
 418         [L_NO_CHECK_CERTIFICATE] = { "no-check-certificate", no_argument },
 419         [L_CA_CERTIFICATE] = { "ca-certificate", required_argument },
 420         [L_USER] = { "user", required_argument },
 421         [L_PASSWORD] = { "password", required_argument },
 422         [L_USER_AGENT] = { "user-agent", required_argument },
 423         [L_POST_DATA] = { "post-data", required_argument },
 424         [L_SPIDER] = { "spider", no_argument },
 425         [L_TIMEOUT] = { "timeout", required_argument },
 426         [L_CONTINUE] = { "continue", no_argument },
 427         {}
 428 };
 429
 430
 431
 432 int main(int argc, char **argv)
 433 {
 434         const char *progname = argv[0];
 435         char *username = NULL;
 436         char *password = NULL;
 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 }