1 // SPDX-License-Identifier: GPL-2.0 2 3 #define _GNU_SOURCE 4 5 #include <errno.h> 6 #include <limits.h> 7 #include <fcntl.h> 8 #include <string.h> 9 #include <stdarg.h> 10 #include <stdbool.h> 11 #include <stdint.h> 12 #include <stdio.h> 13 #include <stdlib.h> 14 #include <strings.h> 15 #include <signal.h> 16 #include <unistd.h> 17 #include <time.h> 18 19 #include <sys/ioctl.h> 20 #include <sys/poll.h> 21 #include <sys/sendfile.h> 22 #include <sys/stat.h> 23 #include <sys/socket.h> 24 #include <sys/types.h> 25 #include <sys/mman.h> 26 27 #include <netdb.h> 28 #include <netinet/in.h> 29 30 #include <linux/tcp.h> 31 #include <linux/time_types.h> 32 #include <linux/sockios.h> 33 34 extern int optind; 35 36 #ifndef IPPROTO_MPTCP 37 #define IPPROTO_MPTCP 262 38 #endif 39 #ifndef TCP_ULP 40 #define TCP_ULP 31 41 #endif 42 43 static int poll_timeout = 10 * 1000; 44 static bool listen_mode; 45 static bool quit; 46 47 enum cfg_mode { 48 CFG_MODE_POLL, 49 CFG_MODE_MMAP, 50 CFG_MODE_SENDFILE, 51 }; 52 53 enum cfg_peek { 54 CFG_NONE_PEEK, 55 CFG_WITH_PEEK, 56 CFG_AFTER_PEEK, 57 }; 58 59 static enum cfg_mode cfg_mode = CFG_MODE_POLL; 60 static enum cfg_peek cfg_peek = CFG_NONE_PEEK; 61 static const char *cfg_host; 62 static const char *cfg_port = "12000"; 63 static int cfg_sock_proto = IPPROTO_MPTCP; 64 static int pf = AF_INET; 65 static int cfg_sndbuf; 66 static int cfg_rcvbuf; 67 static bool cfg_join; 68 static bool cfg_remove; 69 static unsigned int cfg_time; 70 static unsigned int cfg_do_w; 71 static int cfg_wait; 72 static uint32_t cfg_mark; 73 static char *cfg_input; 74 static int cfg_repeat = 1; 75 76 struct cfg_cmsg_types { 77 unsigned int cmsg_enabled:1; 78 unsigned int timestampns:1; 79 unsigned int tcp_inq:1; 80 }; 81 82 struct cfg_sockopt_types { 83 unsigned int transparent:1; 84 }; 85 86 struct tcp_inq_state { 87 unsigned int last; 88 bool expect_eof; 89 }; 90 91 static struct tcp_inq_state tcp_inq; 92 93 static struct cfg_cmsg_types cfg_cmsg_types; 94 static struct cfg_sockopt_types cfg_sockopt_types; 95 96 static void die_usage(void) 97 { 98 fprintf(stderr, "Usage: mptcp_connect [-6] [-c cmsg] [-i file] [-I num] [-j] [-l] " 99 "[-m mode] [-M mark] [-o option] [-p port] [-P mode] [-j] [-l] [-r num] " 100 "[-s MPTCP|TCP] [-S num] [-r num] [-t num] [-T num] [-u] [-w sec] connect_address\n"); 101 fprintf(stderr, "\t-6 use ipv6\n"); 102 fprintf(stderr, "\t-c cmsg -- test cmsg type <cmsg>\n"); 103 fprintf(stderr, "\t-i file -- read the data to send from the given file instead of stdin"); 104 fprintf(stderr, "\t-I num -- repeat the transfer 'num' times. In listen mode accepts num " 105 "incoming connections, in client mode, disconnect and reconnect to the server\n"); 106 fprintf(stderr, "\t-j -- add additional sleep at connection start and tear down " 107 "-- for MPJ tests\n"); 108 fprintf(stderr, "\t-l -- listens mode, accepts incoming connection\n"); 109 fprintf(stderr, "\t-m [poll|mmap|sendfile] -- use poll(default)/mmap+write/sendfile\n"); 110 fprintf(stderr, "\t-M mark -- set socket packet mark\n"); 111 fprintf(stderr, "\t-o option -- test sockopt <option>\n"); 112 fprintf(stderr, "\t-p num -- use port num\n"); 113 fprintf(stderr, 114 "\t-P [saveWithPeek|saveAfterPeek] -- save data with/after MSG_PEEK form tcp socket\n"); 115 fprintf(stderr, "\t-t num -- set poll timeout to num\n"); 116 fprintf(stderr, "\t-T num -- set expected runtime to num ms\n"); 117 fprintf(stderr, "\t-r num -- enable slow mode, limiting each write to num bytes " 118 "-- for remove addr tests\n"); 119 fprintf(stderr, "\t-R num -- set SO_RCVBUF to num\n"); 120 fprintf(stderr, "\t-s [MPTCP|TCP] -- use mptcp(default) or tcp sockets\n"); 121 fprintf(stderr, "\t-S num -- set SO_SNDBUF to num\n"); 122 fprintf(stderr, "\t-w num -- wait num sec before closing the socket\n"); 123 exit(1); 124 } 125 126 static void xerror(const char *fmt, ...) 127 { 128 va_list ap; 129 130 va_start(ap, fmt); 131 vfprintf(stderr, fmt, ap); 132 va_end(ap); 133 exit(1); 134 } 135 136 static void handle_signal(int nr) 137 { 138 quit = true; 139 } 140 141 static const char *getxinfo_strerr(int err) 142 { 143 if (err == EAI_SYSTEM) 144 return strerror(errno); 145 146 return gai_strerror(err); 147 } 148 149 static void xgetnameinfo(const struct sockaddr *addr, socklen_t addrlen, 150 char *host, socklen_t hostlen, 151 char *serv, socklen_t servlen) 152 { 153 int flags = NI_NUMERICHOST | NI_NUMERICSERV; 154 int err = getnameinfo(addr, addrlen, host, hostlen, serv, servlen, 155 flags); 156 157 if (err) { 158 const char *errstr = getxinfo_strerr(err); 159 160 fprintf(stderr, "Fatal: getnameinfo: %s\n", errstr); 161 exit(1); 162 } 163 } 164 165 static void xgetaddrinfo(const char *node, const char *service, 166 const struct addrinfo *hints, 167 struct addrinfo **res) 168 { 169 int err = getaddrinfo(node, service, hints, res); 170 171 if (err) { 172 const char *errstr = getxinfo_strerr(err); 173 174 fprintf(stderr, "Fatal: getaddrinfo(%s:%s): %s\n", 175 node ? node : "", service ? service : "", errstr); 176 exit(1); 177 } 178 } 179 180 static void set_rcvbuf(int fd, unsigned int size) 181 { 182 int err; 183 184 err = setsockopt(fd, SOL_SOCKET, SO_RCVBUF, &size, sizeof(size)); 185 if (err) { 186 perror("set SO_RCVBUF"); 187 exit(1); 188 } 189 } 190 191 static void set_sndbuf(int fd, unsigned int size) 192 { 193 int err; 194 195 err = setsockopt(fd, SOL_SOCKET, SO_SNDBUF, &size, sizeof(size)); 196 if (err) { 197 perror("set SO_SNDBUF"); 198 exit(1); 199 } 200 } 201 202 static void set_mark(int fd, uint32_t mark) 203 { 204 int err; 205 206 err = setsockopt(fd, SOL_SOCKET, SO_MARK, &mark, sizeof(mark)); 207 if (err) { 208 perror("set SO_MARK"); 209 exit(1); 210 } 211 } 212 213 static void set_transparent(int fd, int pf) 214 { 215 int one = 1; 216 217 switch (pf) { 218 case AF_INET: 219 if (-1 == setsockopt(fd, SOL_IP, IP_TRANSPARENT, &one, sizeof(one))) 220 perror("IP_TRANSPARENT"); 221 break; 222 case AF_INET6: 223 if (-1 == setsockopt(fd, IPPROTO_IPV6, IPV6_TRANSPARENT, &one, sizeof(one))) 224 perror("IPV6_TRANSPARENT"); 225 break; 226 } 227 } 228 229 static int do_ulp_so(int sock, const char *name) 230 { 231 return setsockopt(sock, IPPROTO_TCP, TCP_ULP, name, strlen(name)); 232 } 233 234 #define X(m) xerror("%s:%u: %s: failed for proto %d at line %u", __FILE__, __LINE__, (m), proto, line) 235 static void sock_test_tcpulp(int sock, int proto, unsigned int line) 236 { 237 socklen_t buflen = 8; 238 char buf[8] = ""; 239 int ret = getsockopt(sock, IPPROTO_TCP, TCP_ULP, buf, &buflen); 240 241 if (ret != 0) 242 X("getsockopt"); 243 244 if (buflen > 0) { 245 if (strcmp(buf, "mptcp") != 0) 246 xerror("unexpected ULP '%s' for proto %d at line %u", buf, proto, line); 247 ret = do_ulp_so(sock, "tls"); 248 if (ret == 0) 249 X("setsockopt"); 250 } else if (proto == IPPROTO_MPTCP) { 251 ret = do_ulp_so(sock, "tls"); 252 if (ret != -1) 253 X("setsockopt"); 254 } 255 256 ret = do_ulp_so(sock, "mptcp"); 257 if (ret != -1) 258 X("setsockopt"); 259 260 #undef X 261 } 262 263 #define SOCK_TEST_TCPULP(s, p) sock_test_tcpulp((s), (p), __LINE__) 264 265 static int sock_listen_mptcp(const char * const listenaddr, 266 const char * const port) 267 { 268 int sock = -1; 269 struct addrinfo hints = { 270 .ai_protocol = IPPROTO_TCP, 271 .ai_socktype = SOCK_STREAM, 272 .ai_flags = AI_PASSIVE | AI_NUMERICHOST 273 }; 274 275 hints.ai_family = pf; 276 277 struct addrinfo *a, *addr; 278 int one = 1; 279 280 xgetaddrinfo(listenaddr, port, &hints, &addr); 281 hints.ai_family = pf; 282 283 for (a = addr; a; a = a->ai_next) { 284 sock = socket(a->ai_family, a->ai_socktype, cfg_sock_proto); 285 if (sock < 0) 286 continue; 287 288 SOCK_TEST_TCPULP(sock, cfg_sock_proto); 289 290 if (-1 == setsockopt(sock, SOL_SOCKET, SO_REUSEADDR, &one, 291 sizeof(one))) 292 perror("setsockopt"); 293 294 if (cfg_sockopt_types.transparent) 295 set_transparent(sock, pf); 296 297 if (bind(sock, a->ai_addr, a->ai_addrlen) == 0) 298 break; /* success */ 299 300 perror("bind"); 301 close(sock); 302 sock = -1; 303 } 304 305 freeaddrinfo(addr); 306 307 if (sock < 0) { 308 fprintf(stderr, "Could not create listen socket\n"); 309 return sock; 310 } 311 312 SOCK_TEST_TCPULP(sock, cfg_sock_proto); 313 314 if (listen(sock, 20)) { 315 perror("listen"); 316 close(sock); 317 return -1; 318 } 319 320 SOCK_TEST_TCPULP(sock, cfg_sock_proto); 321 322 return sock; 323 } 324 325 static int sock_connect_mptcp(const char * const remoteaddr, 326 const char * const port, int proto, 327 struct addrinfo **peer) 328 { 329 struct addrinfo hints = { 330 .ai_protocol = IPPROTO_TCP, 331 .ai_socktype = SOCK_STREAM, 332 }; 333 struct addrinfo *a, *addr; 334 int sock = -1; 335 336 hints.ai_family = pf; 337 338 xgetaddrinfo(remoteaddr, port, &hints, &addr); 339 for (a = addr; a; a = a->ai_next) { 340 sock = socket(a->ai_family, a->ai_socktype, proto); 341 if (sock < 0) { 342 perror("socket"); 343 continue; 344 } 345 346 SOCK_TEST_TCPULP(sock, proto); 347 348 if (cfg_mark) 349 set_mark(sock, cfg_mark); 350 351 if (connect(sock, a->ai_addr, a->ai_addrlen) == 0) { 352 *peer = a; 353 break; /* success */ 354 } 355 356 perror("connect()"); 357 close(sock); 358 sock = -1; 359 } 360 361 freeaddrinfo(addr); 362 if (sock != -1) 363 SOCK_TEST_TCPULP(sock, proto); 364 return sock; 365 } 366 367 static size_t do_rnd_write(const int fd, char *buf, const size_t len) 368 { 369 static bool first = true; 370 unsigned int do_w; 371 ssize_t bw; 372 373 do_w = rand() & 0xffff; 374 if (do_w == 0 || do_w > len) 375 do_w = len; 376 377 if (cfg_join && first && do_w > 100) 378 do_w = 100; 379 380 if (cfg_remove && do_w > cfg_do_w) 381 do_w = cfg_do_w; 382 383 bw = write(fd, buf, do_w); 384 if (bw < 0) 385 perror("write"); 386 387 /* let the join handshake complete, before going on */ 388 if (cfg_join && first) { 389 usleep(200000); 390 first = false; 391 } 392 393 if (cfg_remove) 394 usleep(200000); 395 396 return bw; 397 } 398 399 static size_t do_write(const int fd, char *buf, const size_t len) 400 { 401 size_t offset = 0; 402 403 while (offset < len) { 404 size_t written; 405 ssize_t bw; 406 407 bw = write(fd, buf + offset, len - offset); 408 if (bw < 0) { 409 perror("write"); 410 return 0; 411 } 412 413 written = (size_t)bw; 414 offset += written; 415 } 416 417 return offset; 418 } 419 420 static void process_cmsg(struct msghdr *msgh) 421 { 422 struct __kernel_timespec ts; 423 bool inq_found = false; 424 bool ts_found = false; 425 unsigned int inq = 0; 426 struct cmsghdr *cmsg; 427 428 for (cmsg = CMSG_FIRSTHDR(msgh); cmsg ; cmsg = CMSG_NXTHDR(msgh, cmsg)) { 429 if (cmsg->cmsg_level == SOL_SOCKET && cmsg->cmsg_type == SO_TIMESTAMPNS_NEW) { 430 memcpy(&ts, CMSG_DATA(cmsg), sizeof(ts)); 431 ts_found = true; 432 continue; 433 } 434 if (cmsg->cmsg_level == IPPROTO_TCP && cmsg->cmsg_type == TCP_CM_INQ) { 435 memcpy(&inq, CMSG_DATA(cmsg), sizeof(inq)); 436 inq_found = true; 437 continue; 438 } 439 440 } 441 442 if (cfg_cmsg_types.timestampns) { 443 if (!ts_found) 444 xerror("TIMESTAMPNS not present\n"); 445 } 446 447 if (cfg_cmsg_types.tcp_inq) { 448 if (!inq_found) 449 xerror("TCP_INQ not present\n"); 450 451 if (inq > 1024) 452 xerror("tcp_inq %u is larger than one kbyte\n", inq); 453 tcp_inq.last = inq; 454 } 455 } 456 457 static ssize_t do_recvmsg_cmsg(const int fd, char *buf, const size_t len) 458 { 459 char msg_buf[8192]; 460 struct iovec iov = { 461 .iov_base = buf, 462 .iov_len = len, 463 }; 464 struct msghdr msg = { 465 .msg_iov = &iov, 466 .msg_iovlen = 1, 467 .msg_control = msg_buf, 468 .msg_controllen = sizeof(msg_buf), 469 }; 470 int flags = 0; 471 unsigned int last_hint = tcp_inq.last; 472 int ret = recvmsg(fd, &msg, flags); 473 474 if (ret <= 0) { 475 if (ret == 0 && tcp_inq.expect_eof) 476 return ret; 477 478 if (ret == 0 && cfg_cmsg_types.tcp_inq) 479 if (last_hint != 1 && last_hint != 0) 480 xerror("EOF but last tcp_inq hint was %u\n", last_hint); 481 482 return ret; 483 } 484 485 if (tcp_inq.expect_eof) 486 xerror("expected EOF, last_hint %u, now %u\n", 487 last_hint, tcp_inq.last); 488 489 if (msg.msg_controllen && !cfg_cmsg_types.cmsg_enabled) 490 xerror("got %lu bytes of cmsg data, expected 0\n", 491 (unsigned long)msg.msg_controllen); 492 493 if (msg.msg_controllen == 0 && cfg_cmsg_types.cmsg_enabled) 494 xerror("%s\n", "got no cmsg data"); 495 496 if (msg.msg_controllen) 497 process_cmsg(&msg); 498 499 if (cfg_cmsg_types.tcp_inq) { 500 if ((size_t)ret < len && last_hint > (unsigned int)ret) { 501 if (ret + 1 != (int)last_hint) { 502 int next = read(fd, msg_buf, sizeof(msg_buf)); 503 504 xerror("read %u of %u, last_hint was %u tcp_inq hint now %u next_read returned %d/%m\n", 505 ret, (unsigned int)len, last_hint, tcp_inq.last, next); 506 } else { 507 tcp_inq.expect_eof = true; 508 } 509 } 510 } 511 512 return ret; 513 } 514 515 static ssize_t do_rnd_read(const int fd, char *buf, const size_t len) 516 { 517 int ret = 0; 518 char tmp[16384]; 519 size_t cap = rand(); 520 521 cap &= 0xffff; 522 523 if (cap == 0) 524 cap = 1; 525 else if (cap > len) 526 cap = len; 527 528 if (cfg_peek == CFG_WITH_PEEK) { 529 ret = recv(fd, buf, cap, MSG_PEEK); 530 ret = (ret < 0) ? ret : read(fd, tmp, ret); 531 } else if (cfg_peek == CFG_AFTER_PEEK) { 532 ret = recv(fd, buf, cap, MSG_PEEK); 533 ret = (ret < 0) ? ret : read(fd, buf, cap); 534 } else if (cfg_cmsg_types.cmsg_enabled) { 535 ret = do_recvmsg_cmsg(fd, buf, cap); 536 } else { 537 ret = read(fd, buf, cap); 538 } 539 540 return ret; 541 } 542 543 static void set_nonblock(int fd, bool nonblock) 544 { 545 int flags = fcntl(fd, F_GETFL); 546 547 if (flags == -1) 548 return; 549 550 if (nonblock) 551 fcntl(fd, F_SETFL, flags | O_NONBLOCK); 552 else 553 fcntl(fd, F_SETFL, flags & ~O_NONBLOCK); 554 } 555 556 static int copyfd_io_poll(int infd, int peerfd, int outfd, bool *in_closed_after_out) 557 { 558 struct pollfd fds = { 559 .fd = peerfd, 560 .events = POLLIN | POLLOUT, 561 }; 562 unsigned int woff = 0, wlen = 0; 563 char wbuf[8192]; 564 565 set_nonblock(peerfd, true); 566 567 for (;;) { 568 char rbuf[8192]; 569 ssize_t len; 570 571 if (fds.events == 0) 572 break; 573 574 switch (poll(&fds, 1, poll_timeout)) { 575 case -1: 576 if (errno == EINTR) 577 continue; 578 perror("poll"); 579 return 1; 580 case 0: 581 fprintf(stderr, "%s: poll timed out (events: " 582 "POLLIN %u, POLLOUT %u)\n", __func__, 583 fds.events & POLLIN, fds.events & POLLOUT); 584 return 2; 585 } 586 587 if (fds.revents & POLLIN) { 588 len = do_rnd_read(peerfd, rbuf, sizeof(rbuf)); 589 if (len == 0) { 590 /* no more data to receive: 591 * peer has closed its write side 592 */ 593 fds.events &= ~POLLIN; 594 595 if ((fds.events & POLLOUT) == 0) { 596 *in_closed_after_out = true; 597 /* and nothing more to send */ 598 break; 599 } 600 601 /* Else, still have data to transmit */ 602 } else if (len < 0) { 603 perror("read"); 604 return 3; 605 } 606 607 do_write(outfd, rbuf, len); 608 } 609 610 if (fds.revents & POLLOUT) { 611 if (wlen == 0) { 612 woff = 0; 613 wlen = read(infd, wbuf, sizeof(wbuf)); 614 } 615 616 if (wlen > 0) { 617 ssize_t bw; 618 619 bw = do_rnd_write(peerfd, wbuf + woff, wlen); 620 if (bw < 0) 621 return 111; 622 623 woff += bw; 624 wlen -= bw; 625 } else if (wlen == 0) { 626 /* We have no more data to send. */ 627 fds.events &= ~POLLOUT; 628 629 if ((fds.events & POLLIN) == 0) 630 /* ... and peer also closed already */ 631 break; 632 633 /* ... but we still receive. 634 * Close our write side, ev. give some time 635 * for address notification and/or checking 636 * the current status 637 */ 638 if (cfg_wait) 639 usleep(cfg_wait); 640 shutdown(peerfd, SHUT_WR); 641 } else { 642 if (errno == EINTR) 643 continue; 644 perror("read"); 645 return 4; 646 } 647 } 648 649 if (fds.revents & (POLLERR | POLLNVAL)) { 650 fprintf(stderr, "Unexpected revents: " 651 "POLLERR/POLLNVAL(%x)\n", fds.revents); 652 return 5; 653 } 654 } 655 656 /* leave some time for late join/announce */ 657 if (cfg_remove) 658 usleep(cfg_wait); 659 660 return 0; 661 } 662 663 static int do_recvfile(int infd, int outfd) 664 { 665 ssize_t r; 666 667 do { 668 char buf[16384]; 669 670 r = do_rnd_read(infd, buf, sizeof(buf)); 671 if (r > 0) { 672 if (write(outfd, buf, r) != r) 673 break; 674 } else if (r < 0) { 675 perror("read"); 676 } 677 } while (r > 0); 678 679 return (int)r; 680 } 681 682 static int do_mmap(int infd, int outfd, unsigned int size) 683 { 684 char *inbuf = mmap(NULL, size, PROT_READ, MAP_SHARED, infd, 0); 685 ssize_t ret = 0, off = 0; 686 size_t rem; 687 688 if (inbuf == MAP_FAILED) { 689 perror("mmap"); 690 return 1; 691 } 692 693 rem = size; 694 695 while (rem > 0) { 696 ret = write(outfd, inbuf + off, rem); 697 698 if (ret < 0) { 699 perror("write"); 700 break; 701 } 702 703 off += ret; 704 rem -= ret; 705 } 706 707 munmap(inbuf, size); 708 return rem; 709 } 710 711 static int get_infd_size(int fd) 712 { 713 struct stat sb; 714 ssize_t count; 715 int err; 716 717 err = fstat(fd, &sb); 718 if (err < 0) { 719 perror("fstat"); 720 return -1; 721 } 722 723 if ((sb.st_mode & S_IFMT) != S_IFREG) { 724 fprintf(stderr, "%s: stdin is not a regular file\n", __func__); 725 return -2; 726 } 727 728 count = sb.st_size; 729 if (count > INT_MAX) { 730 fprintf(stderr, "File too large: %zu\n", count); 731 return -3; 732 } 733 734 return (int)count; 735 } 736 737 static int do_sendfile(int infd, int outfd, unsigned int count) 738 { 739 while (count > 0) { 740 ssize_t r; 741 742 r = sendfile(outfd, infd, NULL, count); 743 if (r < 0) { 744 perror("sendfile"); 745 return 3; 746 } 747 748 count -= r; 749 } 750 751 return 0; 752 } 753 754 static int copyfd_io_mmap(int infd, int peerfd, int outfd, 755 unsigned int size, bool *in_closed_after_out) 756 { 757 int err; 758 759 if (listen_mode) { 760 err = do_recvfile(peerfd, outfd); 761 if (err) 762 return err; 763 764 err = do_mmap(infd, peerfd, size); 765 } else { 766 err = do_mmap(infd, peerfd, size); 767 if (err) 768 return err; 769 770 shutdown(peerfd, SHUT_WR); 771 772 err = do_recvfile(peerfd, outfd); 773 *in_closed_after_out = true; 774 } 775 776 return err; 777 } 778 779 static int copyfd_io_sendfile(int infd, int peerfd, int outfd, 780 unsigned int size, bool *in_closed_after_out) 781 { 782 int err; 783 784 if (listen_mode) { 785 err = do_recvfile(peerfd, outfd); 786 if (err) 787 return err; 788 789 err = do_sendfile(infd, peerfd, size); 790 } else { 791 err = do_sendfile(infd, peerfd, size); 792 if (err) 793 return err; 794 err = do_recvfile(peerfd, outfd); 795 *in_closed_after_out = true; 796 } 797 798 return err; 799 } 800 801 static int copyfd_io(int infd, int peerfd, int outfd, bool close_peerfd) 802 { 803 bool in_closed_after_out = false; 804 struct timespec start, end; 805 int file_size; 806 int ret; 807 808 if (cfg_time && (clock_gettime(CLOCK_MONOTONIC, &start) < 0)) 809 xerror("can not fetch start time %d", errno); 810 811 switch (cfg_mode) { 812 case CFG_MODE_POLL: 813 ret = copyfd_io_poll(infd, peerfd, outfd, &in_closed_after_out); 814 break; 815 816 case CFG_MODE_MMAP: 817 file_size = get_infd_size(infd); 818 if (file_size < 0) 819 return file_size; 820 ret = copyfd_io_mmap(infd, peerfd, outfd, file_size, &in_closed_after_out); 821 break; 822 823 case CFG_MODE_SENDFILE: 824 file_size = get_infd_size(infd); 825 if (file_size < 0) 826 return file_size; 827 ret = copyfd_io_sendfile(infd, peerfd, outfd, file_size, &in_closed_after_out); 828 break; 829 830 default: 831 fprintf(stderr, "Invalid mode %d\n", cfg_mode); 832 833 die_usage(); 834 return 1; 835 } 836 837 if (ret) 838 return ret; 839 840 if (close_peerfd) 841 close(peerfd); 842 843 if (cfg_time) { 844 unsigned int delta_ms; 845 846 if (clock_gettime(CLOCK_MONOTONIC, &end) < 0) 847 xerror("can not fetch end time %d", errno); 848 delta_ms = (end.tv_sec - start.tv_sec) * 1000 + (end.tv_nsec - start.tv_nsec) / 1000000; 849 if (delta_ms > cfg_time) { 850 xerror("transfer slower than expected! runtime %d ms, expected %d ms", 851 delta_ms, cfg_time); 852 } 853 854 /* show the runtime only if this end shutdown(wr) before receiving the EOF, 855 * (that is, if this end got the longer runtime) 856 */ 857 if (in_closed_after_out) 858 fprintf(stderr, "%d", delta_ms); 859 } 860 861 return 0; 862 } 863 864 static void check_sockaddr(int pf, struct sockaddr_storage *ss, 865 socklen_t salen) 866 { 867 struct sockaddr_in6 *sin6; 868 struct sockaddr_in *sin; 869 socklen_t wanted_size = 0; 870 871 switch (pf) { 872 case AF_INET: 873 wanted_size = sizeof(*sin); 874 sin = (void *)ss; 875 if (!sin->sin_port) 876 fprintf(stderr, "accept: something wrong: ip connection from port 0"); 877 break; 878 case AF_INET6: 879 wanted_size = sizeof(*sin6); 880 sin6 = (void *)ss; 881 if (!sin6->sin6_port) 882 fprintf(stderr, "accept: something wrong: ipv6 connection from port 0"); 883 break; 884 default: 885 fprintf(stderr, "accept: Unknown pf %d, salen %u\n", pf, salen); 886 return; 887 } 888 889 if (salen != wanted_size) 890 fprintf(stderr, "accept: size mismatch, got %d expected %d\n", 891 (int)salen, wanted_size); 892 893 if (ss->ss_family != pf) 894 fprintf(stderr, "accept: pf mismatch, expect %d, ss_family is %d\n", 895 (int)ss->ss_family, pf); 896 } 897 898 static void check_getpeername(int fd, struct sockaddr_storage *ss, socklen_t salen) 899 { 900 struct sockaddr_storage peerss; 901 socklen_t peersalen = sizeof(peerss); 902 903 if (getpeername(fd, (struct sockaddr *)&peerss, &peersalen) < 0) { 904 perror("getpeername"); 905 return; 906 } 907 908 if (peersalen != salen) { 909 fprintf(stderr, "%s: %d vs %d\n", __func__, peersalen, salen); 910 return; 911 } 912 913 if (memcmp(ss, &peerss, peersalen)) { 914 char a[INET6_ADDRSTRLEN]; 915 char b[INET6_ADDRSTRLEN]; 916 char c[INET6_ADDRSTRLEN]; 917 char d[INET6_ADDRSTRLEN]; 918 919 xgetnameinfo((struct sockaddr *)ss, salen, 920 a, sizeof(a), b, sizeof(b)); 921 922 xgetnameinfo((struct sockaddr *)&peerss, peersalen, 923 c, sizeof(c), d, sizeof(d)); 924 925 fprintf(stderr, "%s: memcmp failure: accept %s vs peername %s, %s vs %s salen %d vs %d\n", 926 __func__, a, c, b, d, peersalen, salen); 927 } 928 } 929 930 static void check_getpeername_connect(int fd) 931 { 932 struct sockaddr_storage ss; 933 socklen_t salen = sizeof(ss); 934 char a[INET6_ADDRSTRLEN]; 935 char b[INET6_ADDRSTRLEN]; 936 937 if (getpeername(fd, (struct sockaddr *)&ss, &salen) < 0) { 938 perror("getpeername"); 939 return; 940 } 941 942 xgetnameinfo((struct sockaddr *)&ss, salen, 943 a, sizeof(a), b, sizeof(b)); 944 945 if (strcmp(cfg_host, a) || strcmp(cfg_port, b)) 946 fprintf(stderr, "%s: %s vs %s, %s vs %s\n", __func__, 947 cfg_host, a, cfg_port, b); 948 } 949 950 static void maybe_close(int fd) 951 { 952 unsigned int r = rand(); 953 954 if (!(cfg_join || cfg_remove || cfg_repeat > 1) && (r & 1)) 955 close(fd); 956 } 957 958 int main_loop_s(int listensock) 959 { 960 struct sockaddr_storage ss; 961 struct pollfd polls; 962 socklen_t salen; 963 int remotesock; 964 int fd = 0; 965 966 again: 967 polls.fd = listensock; 968 polls.events = POLLIN; 969 970 switch (poll(&polls, 1, poll_timeout)) { 971 case -1: 972 perror("poll"); 973 return 1; 974 case 0: 975 fprintf(stderr, "%s: timed out\n", __func__); 976 close(listensock); 977 return 2; 978 } 979 980 salen = sizeof(ss); 981 remotesock = accept(listensock, (struct sockaddr *)&ss, &salen); 982 if (remotesock >= 0) { 983 maybe_close(listensock); 984 check_sockaddr(pf, &ss, salen); 985 check_getpeername(remotesock, &ss, salen); 986 987 if (cfg_input) { 988 fd = open(cfg_input, O_RDONLY); 989 if (fd < 0) 990 xerror("can't open %s: %d", cfg_input, errno); 991 } 992 993 SOCK_TEST_TCPULP(remotesock, 0); 994 995 copyfd_io(fd, remotesock, 1, true); 996 } else { 997 perror("accept"); 998 return 1; 999 } 1000 1001 if (--cfg_repeat > 0) { 1002 if (cfg_input) 1003 close(fd); 1004 goto again; 1005 } 1006 1007 return 0; 1008 } 1009 1010 static void init_rng(void) 1011 { 1012 int fd = open("/dev/urandom", O_RDONLY); 1013 unsigned int foo; 1014 1015 if (fd > 0) { 1016 int ret = read(fd, &foo, sizeof(foo)); 1017 1018 if (ret < 0) 1019 srand(fd + foo); 1020 close(fd); 1021 } 1022 1023 srand(foo); 1024 } 1025 1026 static void xsetsockopt(int fd, int level, int optname, const void *optval, socklen_t optlen) 1027 { 1028 int err; 1029 1030 err = setsockopt(fd, level, optname, optval, optlen); 1031 if (err) { 1032 perror("setsockopt"); 1033 exit(1); 1034 } 1035 } 1036 1037 static void apply_cmsg_types(int fd, const struct cfg_cmsg_types *cmsg) 1038 { 1039 static const unsigned int on = 1; 1040 1041 if (cmsg->timestampns) 1042 xsetsockopt(fd, SOL_SOCKET, SO_TIMESTAMPNS_NEW, &on, sizeof(on)); 1043 if (cmsg->tcp_inq) 1044 xsetsockopt(fd, IPPROTO_TCP, TCP_INQ, &on, sizeof(on)); 1045 } 1046 1047 static void parse_cmsg_types(const char *type) 1048 { 1049 char *next = strchr(type, ','); 1050 unsigned int len = 0; 1051 1052 cfg_cmsg_types.cmsg_enabled = 1; 1053 1054 if (next) { 1055 parse_cmsg_types(next + 1); 1056 len = next - type; 1057 } else { 1058 len = strlen(type); 1059 } 1060 1061 if (strncmp(type, "TIMESTAMPNS", len) == 0) { 1062 cfg_cmsg_types.timestampns = 1; 1063 return; 1064 } 1065 1066 if (strncmp(type, "TCPINQ", len) == 0) { 1067 cfg_cmsg_types.tcp_inq = 1; 1068 return; 1069 } 1070 1071 fprintf(stderr, "Unrecognized cmsg option %s\n", type); 1072 exit(1); 1073 } 1074 1075 static void parse_setsock_options(const char *name) 1076 { 1077 char *next = strchr(name, ','); 1078 unsigned int len = 0; 1079 1080 if (next) { 1081 parse_setsock_options(next + 1); 1082 len = next - name; 1083 } else { 1084 len = strlen(name); 1085 } 1086 1087 if (strncmp(name, "TRANSPARENT", len) == 0) { 1088 cfg_sockopt_types.transparent = 1; 1089 return; 1090 } 1091 1092 fprintf(stderr, "Unrecognized setsockopt option %s\n", name); 1093 exit(1); 1094 } 1095 1096 void xdisconnect(int fd, int addrlen) 1097 { 1098 struct sockaddr_storage empty; 1099 int msec_sleep = 10; 1100 int queued = 1; 1101 int i; 1102 1103 shutdown(fd, SHUT_WR); 1104 1105 /* while until the pending data is completely flushed, the later 1106 * disconnect will bypass/ignore/drop any pending data. 1107 */ 1108 for (i = 0; ; i += msec_sleep) { 1109 if (ioctl(fd, SIOCOUTQ, &queued) < 0) 1110 xerror("can't query out socket queue: %d", errno); 1111 1112 if (!queued) 1113 break; 1114 1115 if (i > poll_timeout) 1116 xerror("timeout while waiting for spool to complete"); 1117 usleep(msec_sleep * 1000); 1118 } 1119 1120 memset(&empty, 0, sizeof(empty)); 1121 empty.ss_family = AF_UNSPEC; 1122 if (connect(fd, (struct sockaddr *)&empty, addrlen) < 0) 1123 xerror("can't disconnect: %d", errno); 1124 } 1125 1126 int main_loop(void) 1127 { 1128 int fd, ret, fd_in = 0; 1129 struct addrinfo *peer; 1130 1131 /* listener is ready. */ 1132 fd = sock_connect_mptcp(cfg_host, cfg_port, cfg_sock_proto, &peer); 1133 if (fd < 0) 1134 return 2; 1135 1136 again: 1137 check_getpeername_connect(fd); 1138 1139 SOCK_TEST_TCPULP(fd, cfg_sock_proto); 1140 1141 if (cfg_rcvbuf) 1142 set_rcvbuf(fd, cfg_rcvbuf); 1143 if (cfg_sndbuf) 1144 set_sndbuf(fd, cfg_sndbuf); 1145 if (cfg_cmsg_types.cmsg_enabled) 1146 apply_cmsg_types(fd, &cfg_cmsg_types); 1147 1148 if (cfg_input) { 1149 fd_in = open(cfg_input, O_RDONLY); 1150 if (fd < 0) 1151 xerror("can't open %s:%d", cfg_input, errno); 1152 } 1153 1154 /* close the client socket open only if we are not going to reconnect */ 1155 ret = copyfd_io(fd_in, fd, 1, cfg_repeat == 1); 1156 if (ret) 1157 return ret; 1158 1159 if (--cfg_repeat > 0) { 1160 xdisconnect(fd, peer->ai_addrlen); 1161 1162 /* the socket could be unblocking at this point, we need the 1163 * connect to be blocking 1164 */ 1165 set_nonblock(fd, false); 1166 if (connect(fd, peer->ai_addr, peer->ai_addrlen)) 1167 xerror("can't reconnect: %d", errno); 1168 if (cfg_input) 1169 close(fd_in); 1170 goto again; 1171 } 1172 return 0; 1173 } 1174 1175 int parse_proto(const char *proto) 1176 { 1177 if (!strcasecmp(proto, "MPTCP")) 1178 return IPPROTO_MPTCP; 1179 if (!strcasecmp(proto, "TCP")) 1180 return IPPROTO_TCP; 1181 1182 fprintf(stderr, "Unknown protocol: %s\n.", proto); 1183 die_usage(); 1184 1185 /* silence compiler warning */ 1186 return 0; 1187 } 1188 1189 int parse_mode(const char *mode) 1190 { 1191 if (!strcasecmp(mode, "poll")) 1192 return CFG_MODE_POLL; 1193 if (!strcasecmp(mode, "mmap")) 1194 return CFG_MODE_MMAP; 1195 if (!strcasecmp(mode, "sendfile")) 1196 return CFG_MODE_SENDFILE; 1197 1198 fprintf(stderr, "Unknown test mode: %s\n", mode); 1199 fprintf(stderr, "Supported modes are:\n"); 1200 fprintf(stderr, "\t\t\"poll\" - interleaved read/write using poll()\n"); 1201 fprintf(stderr, "\t\t\"mmap\" - send entire input file (mmap+write), then read response (-l will read input first)\n"); 1202 fprintf(stderr, "\t\t\"sendfile\" - send entire input file (sendfile), then read response (-l will read input first)\n"); 1203 1204 die_usage(); 1205 1206 /* silence compiler warning */ 1207 return 0; 1208 } 1209 1210 int parse_peek(const char *mode) 1211 { 1212 if (!strcasecmp(mode, "saveWithPeek")) 1213 return CFG_WITH_PEEK; 1214 if (!strcasecmp(mode, "saveAfterPeek")) 1215 return CFG_AFTER_PEEK; 1216 1217 fprintf(stderr, "Unknown: %s\n", mode); 1218 fprintf(stderr, "Supported MSG_PEEK mode are:\n"); 1219 fprintf(stderr, 1220 "\t\t\"saveWithPeek\" - recv data with flags 'MSG_PEEK' and save the peek data into file\n"); 1221 fprintf(stderr, 1222 "\t\t\"saveAfterPeek\" - read and save data into file after recv with flags 'MSG_PEEK'\n"); 1223 1224 die_usage(); 1225 1226 /* silence compiler warning */ 1227 return 0; 1228 } 1229 1230 static int parse_int(const char *size) 1231 { 1232 unsigned long s; 1233 1234 errno = 0; 1235 1236 s = strtoul(size, NULL, 0); 1237 1238 if (errno) { 1239 fprintf(stderr, "Invalid sndbuf size %s (%s)\n", 1240 size, strerror(errno)); 1241 die_usage(); 1242 } 1243 1244 if (s > INT_MAX) { 1245 fprintf(stderr, "Invalid sndbuf size %s (%s)\n", 1246 size, strerror(ERANGE)); 1247 die_usage(); 1248 } 1249 1250 return (int)s; 1251 } 1252 1253 static void parse_opts(int argc, char **argv) 1254 { 1255 int c; 1256 1257 while ((c = getopt(argc, argv, "6c:hi:I:jlm:M:o:p:P:r:R:s:S:t:T:w:")) != -1) { 1258 switch (c) { 1259 case 'j': 1260 cfg_join = true; 1261 cfg_mode = CFG_MODE_POLL; 1262 break; 1263 case 'r': 1264 cfg_remove = true; 1265 cfg_mode = CFG_MODE_POLL; 1266 cfg_wait = 400000; 1267 cfg_do_w = atoi(optarg); 1268 if (cfg_do_w <= 0) 1269 cfg_do_w = 50; 1270 break; 1271 case 'i': 1272 cfg_input = optarg; 1273 break; 1274 case 'I': 1275 cfg_repeat = atoi(optarg); 1276 break; 1277 case 'l': 1278 listen_mode = true; 1279 break; 1280 case 'p': 1281 cfg_port = optarg; 1282 break; 1283 case 's': 1284 cfg_sock_proto = parse_proto(optarg); 1285 break; 1286 case 'h': 1287 die_usage(); 1288 break; 1289 case '6': 1290 pf = AF_INET6; 1291 break; 1292 case 't': 1293 poll_timeout = atoi(optarg) * 1000; 1294 if (poll_timeout <= 0) 1295 poll_timeout = -1; 1296 break; 1297 case 'T': 1298 cfg_time = atoi(optarg); 1299 break; 1300 case 'm': 1301 cfg_mode = parse_mode(optarg); 1302 break; 1303 case 'S': 1304 cfg_sndbuf = parse_int(optarg); 1305 break; 1306 case 'R': 1307 cfg_rcvbuf = parse_int(optarg); 1308 break; 1309 case 'w': 1310 cfg_wait = atoi(optarg)*1000000; 1311 break; 1312 case 'M': 1313 cfg_mark = strtol(optarg, NULL, 0); 1314 break; 1315 case 'P': 1316 cfg_peek = parse_peek(optarg); 1317 break; 1318 case 'c': 1319 parse_cmsg_types(optarg); 1320 break; 1321 case 'o': 1322 parse_setsock_options(optarg); 1323 break; 1324 } 1325 } 1326 1327 if (optind + 1 != argc) 1328 die_usage(); 1329 cfg_host = argv[optind]; 1330 1331 if (strchr(cfg_host, ':')) 1332 pf = AF_INET6; 1333 } 1334 1335 int main(int argc, char *argv[]) 1336 { 1337 init_rng(); 1338 1339 signal(SIGUSR1, handle_signal); 1340 parse_opts(argc, argv); 1341 1342 if (listen_mode) { 1343 int fd = sock_listen_mptcp(cfg_host, cfg_port); 1344 1345 if (fd < 0) 1346 return 1; 1347 1348 if (cfg_rcvbuf) 1349 set_rcvbuf(fd, cfg_rcvbuf); 1350 if (cfg_sndbuf) 1351 set_sndbuf(fd, cfg_sndbuf); 1352 if (cfg_mark) 1353 set_mark(fd, cfg_mark); 1354 if (cfg_cmsg_types.cmsg_enabled) 1355 apply_cmsg_types(fd, &cfg_cmsg_types); 1356 1357 return main_loop_s(fd); 1358 } 1359 1360 return main_loop(); 1361 } 1362