tcp_input.c revision 1.423
1 /* $NetBSD: tcp_input.c,v 1.423 2020/09/13 11:47:12 roy Exp $ */ 2 3 /* 4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 5 * All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 3. Neither the name of the project nor the names of its contributors 16 * may be used to endorse or promote products derived from this software 17 * without specific prior written permission. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * SUCH DAMAGE. 30 */ 31 32 /* 33 * @(#)COPYRIGHT 1.1 (NRL) 17 January 1995 34 * 35 * NRL grants permission for redistribution and use in source and binary 36 * forms, with or without modification, of the software and documentation 37 * created at NRL provided that the following conditions are met: 38 * 39 * 1. Redistributions of source code must retain the above copyright 40 * notice, this list of conditions and the following disclaimer. 41 * 2. Redistributions in binary form must reproduce the above copyright 42 * notice, this list of conditions and the following disclaimer in the 43 * documentation and/or other materials provided with the distribution. 44 * 3. All advertising materials mentioning features or use of this software 45 * must display the following acknowledgements: 46 * This product includes software developed by the University of 47 * California, Berkeley and its contributors. 48 * This product includes software developed at the Information 49 * Technology Division, US Naval Research Laboratory. 50 * 4. Neither the name of the NRL nor the names of its contributors 51 * may be used to endorse or promote products derived from this software 52 * without specific prior written permission. 53 * 54 * THE SOFTWARE PROVIDED BY NRL IS PROVIDED BY NRL AND CONTRIBUTORS ``AS 55 * IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 56 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A 57 * PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL NRL OR 58 * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, 59 * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, 60 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR 61 * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF 62 * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING 63 * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS 64 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 65 * 66 * The views and conclusions contained in the software and documentation 67 * are those of the authors and should not be interpreted as representing 68 * official policies, either expressed or implied, of the US Naval 69 * Research Laboratory (NRL). 70 */ 71 72 /*- 73 * Copyright (c) 1997, 1998, 1999, 2001, 2005, 2006, 74 * 2011 The NetBSD Foundation, Inc. 75 * All rights reserved. 76 * 77 * This code is derived from software contributed to The NetBSD Foundation 78 * by Coyote Point Systems, Inc. 79 * This code is derived from software contributed to The NetBSD Foundation 80 * by Jason R. Thorpe and Kevin M. Lahey of the Numerical Aerospace Simulation 81 * Facility, NASA Ames Research Center. 82 * This code is derived from software contributed to The NetBSD Foundation 83 * by Charles M. Hannum. 84 * This code is derived from software contributed to The NetBSD Foundation 85 * by Rui Paulo. 86 * 87 * Redistribution and use in source and binary forms, with or without 88 * modification, are permitted provided that the following conditions 89 * are met: 90 * 1. Redistributions of source code must retain the above copyright 91 * notice, this list of conditions and the following disclaimer. 92 * 2. Redistributions in binary form must reproduce the above copyright 93 * notice, this list of conditions and the following disclaimer in the 94 * documentation and/or other materials provided with the distribution. 95 * 96 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 97 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 98 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 99 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 100 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 101 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 102 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 103 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 104 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 105 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 106 * POSSIBILITY OF SUCH DAMAGE. 107 */ 108 109 /* 110 * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1994, 1995 111 * The Regents of the University of California. All rights reserved. 112 * 113 * Redistribution and use in source and binary forms, with or without 114 * modification, are permitted provided that the following conditions 115 * are met: 116 * 1. Redistributions of source code must retain the above copyright 117 * notice, this list of conditions and the following disclaimer. 118 * 2. Redistributions in binary form must reproduce the above copyright 119 * notice, this list of conditions and the following disclaimer in the 120 * documentation and/or other materials provided with the distribution. 121 * 3. Neither the name of the University nor the names of its contributors 122 * may be used to endorse or promote products derived from this software 123 * without specific prior written permission. 124 * 125 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 126 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 127 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 128 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 129 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 130 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 131 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 132 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 133 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 134 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 135 * SUCH DAMAGE. 136 * 137 * @(#)tcp_input.c 8.12 (Berkeley) 5/24/95 138 */ 139 140 /* 141 * TODO list for SYN cache stuff: 142 * 143 * Find room for a "state" field, which is needed to keep a 144 * compressed state for TIME_WAIT TCBs. It's been noted already 145 * that this is fairly important for very high-volume web and 146 * mail servers, which use a large number of short-lived 147 * connections. 148 */ 149 150 #include <sys/cdefs.h> 151 __KERNEL_RCSID(0, "$NetBSD: tcp_input.c,v 1.423 2020/09/13 11:47:12 roy Exp $"); 152 153 #ifdef _KERNEL_OPT 154 #include "opt_inet.h" 155 #include "opt_ipsec.h" 156 #include "opt_inet_csum.h" 157 #include "opt_tcp_debug.h" 158 #endif 159 160 #include <sys/param.h> 161 #include <sys/systm.h> 162 #include <sys/malloc.h> 163 #include <sys/mbuf.h> 164 #include <sys/protosw.h> 165 #include <sys/socket.h> 166 #include <sys/socketvar.h> 167 #include <sys/errno.h> 168 #include <sys/syslog.h> 169 #include <sys/pool.h> 170 #include <sys/domain.h> 171 #include <sys/kernel.h> 172 #ifdef TCP_SIGNATURE 173 #include <sys/md5.h> 174 #endif 175 #include <sys/lwp.h> /* for lwp0 */ 176 #include <sys/cprng.h> 177 178 #include <net/if.h> 179 #include <net/if_types.h> 180 181 #include <netinet/in.h> 182 #include <netinet/in_systm.h> 183 #include <netinet/ip.h> 184 #include <netinet/in_pcb.h> 185 #include <netinet/in_var.h> 186 #include <netinet/ip_var.h> 187 #include <netinet/in_offload.h> 188 189 #if NARP > 0 190 #include <netinet/if_inarp.h> 191 #endif 192 #ifdef INET6 193 #include <netinet/ip6.h> 194 #include <netinet6/ip6_var.h> 195 #include <netinet6/in6_pcb.h> 196 #include <netinet6/ip6_var.h> 197 #include <netinet6/in6_var.h> 198 #include <netinet/icmp6.h> 199 #include <netinet6/nd6.h> 200 #ifdef TCP_SIGNATURE 201 #include <netinet6/scope6_var.h> 202 #endif 203 #endif 204 205 #ifndef INET6 206 #include <netinet/ip6.h> 207 #endif 208 209 #include <netinet/tcp.h> 210 #include <netinet/tcp_fsm.h> 211 #include <netinet/tcp_seq.h> 212 #include <netinet/tcp_timer.h> 213 #include <netinet/tcp_var.h> 214 #include <netinet/tcp_private.h> 215 #include <netinet/tcp_congctl.h> 216 #include <netinet/tcp_debug.h> 217 218 #ifdef INET6 219 #include "faith.h" 220 #if defined(NFAITH) && NFAITH > 0 221 #include <net/if_faith.h> 222 #endif 223 #endif 224 225 #ifdef IPSEC 226 #include <netipsec/ipsec.h> 227 #include <netipsec/key.h> 228 #ifdef INET6 229 #include <netipsec/ipsec6.h> 230 #endif 231 #endif /* IPSEC*/ 232 233 #include <netinet/tcp_vtw.h> 234 235 int tcprexmtthresh = 3; 236 int tcp_log_refused; 237 238 int tcp_do_autorcvbuf = 1; 239 int tcp_autorcvbuf_inc = 16 * 1024; 240 int tcp_autorcvbuf_max = 256 * 1024; 241 int tcp_msl = (TCPTV_MSL / PR_SLOWHZ); 242 243 static int tcp_rst_ppslim_count = 0; 244 static struct timeval tcp_rst_ppslim_last; 245 static int tcp_ackdrop_ppslim_count = 0; 246 static struct timeval tcp_ackdrop_ppslim_last; 247 248 static void syn_cache_timer(void *); 249 250 #define TCP_PAWS_IDLE (24U * 24 * 60 * 60 * PR_SLOWHZ) 251 252 /* for modulo comparisons of timestamps */ 253 #define TSTMP_LT(a,b) ((int)((a)-(b)) < 0) 254 #define TSTMP_GEQ(a,b) ((int)((a)-(b)) >= 0) 255 256 /* 257 * Neighbor Discovery, Neighbor Unreachability Detection Upper layer hint. 258 */ 259 static void 260 nd_hint(struct tcpcb *tp) 261 { 262 struct route *ro = NULL; 263 struct rtentry *rt; 264 265 if (tp == NULL) 266 return; 267 268 switch (tp->t_family) { 269 #if NARP > 0 270 case AF_INET: 271 if (tp->t_inpcb != NULL) 272 ro = &tp->t_inpcb->inp_route; 273 break; 274 #endif 275 #ifdef INET6 276 case AF_INET6: 277 if (tp->t_in6pcb != NULL) 278 ro = &tp->t_in6pcb->in6p_route; 279 break; 280 #endif 281 } 282 283 if (ro == NULL) 284 return; 285 286 rt = rtcache_validate(ro); 287 if (rt == NULL) 288 return; 289 290 switch (tp->t_family) { 291 #if NARP > 0 292 case AF_INET: 293 arp_nud_hint(rt); 294 break; 295 #endif 296 #ifdef INET6 297 case AF_INET6: 298 nd6_nud_hint(rt); 299 break; 300 #endif 301 } 302 303 rtcache_unref(rt, ro); 304 } 305 306 /* 307 * Compute ACK transmission behavior. Delay the ACK unless 308 * we have already delayed an ACK (must send an ACK every two segments). 309 * We also ACK immediately if we received a PUSH and the ACK-on-PUSH 310 * option is enabled. 311 */ 312 static void 313 tcp_setup_ack(struct tcpcb *tp, const struct tcphdr *th) 314 { 315 316 if (tp->t_flags & TF_DELACK || 317 (tcp_ack_on_push && th->th_flags & TH_PUSH)) 318 tp->t_flags |= TF_ACKNOW; 319 else 320 TCP_SET_DELACK(tp); 321 } 322 323 static void 324 icmp_check(struct tcpcb *tp, const struct tcphdr *th, int acked) 325 { 326 327 /* 328 * If we had a pending ICMP message that refers to data that have 329 * just been acknowledged, disregard the recorded ICMP message. 330 */ 331 if ((tp->t_flags & TF_PMTUD_PEND) && 332 SEQ_GT(th->th_ack, tp->t_pmtud_th_seq)) 333 tp->t_flags &= ~TF_PMTUD_PEND; 334 335 /* 336 * Keep track of the largest chunk of data 337 * acknowledged since last PMTU update 338 */ 339 if (tp->t_pmtud_mss_acked < acked) 340 tp->t_pmtud_mss_acked = acked; 341 } 342 343 /* 344 * Convert TCP protocol fields to host order for easier processing. 345 */ 346 static void 347 tcp_fields_to_host(struct tcphdr *th) 348 { 349 350 NTOHL(th->th_seq); 351 NTOHL(th->th_ack); 352 NTOHS(th->th_win); 353 NTOHS(th->th_urp); 354 } 355 356 /* 357 * ... and reverse the above. 358 */ 359 static void 360 tcp_fields_to_net(struct tcphdr *th) 361 { 362 363 HTONL(th->th_seq); 364 HTONL(th->th_ack); 365 HTONS(th->th_win); 366 HTONS(th->th_urp); 367 } 368 369 static void 370 tcp_urp_drop(struct tcphdr *th, int todrop, int *tiflags) 371 { 372 if (th->th_urp > todrop) { 373 th->th_urp -= todrop; 374 } else { 375 *tiflags &= ~TH_URG; 376 th->th_urp = 0; 377 } 378 } 379 380 #ifdef TCP_CSUM_COUNTERS 381 #include <sys/device.h> 382 383 extern struct evcnt tcp_hwcsum_ok; 384 extern struct evcnt tcp_hwcsum_bad; 385 extern struct evcnt tcp_hwcsum_data; 386 extern struct evcnt tcp_swcsum; 387 #if defined(INET6) 388 extern struct evcnt tcp6_hwcsum_ok; 389 extern struct evcnt tcp6_hwcsum_bad; 390 extern struct evcnt tcp6_hwcsum_data; 391 extern struct evcnt tcp6_swcsum; 392 #endif /* defined(INET6) */ 393 394 #define TCP_CSUM_COUNTER_INCR(ev) (ev)->ev_count++ 395 396 #else 397 398 #define TCP_CSUM_COUNTER_INCR(ev) /* nothing */ 399 400 #endif /* TCP_CSUM_COUNTERS */ 401 402 #ifdef TCP_REASS_COUNTERS 403 #include <sys/device.h> 404 405 extern struct evcnt tcp_reass_; 406 extern struct evcnt tcp_reass_empty; 407 extern struct evcnt tcp_reass_iteration[8]; 408 extern struct evcnt tcp_reass_prependfirst; 409 extern struct evcnt tcp_reass_prepend; 410 extern struct evcnt tcp_reass_insert; 411 extern struct evcnt tcp_reass_inserttail; 412 extern struct evcnt tcp_reass_append; 413 extern struct evcnt tcp_reass_appendtail; 414 extern struct evcnt tcp_reass_overlaptail; 415 extern struct evcnt tcp_reass_overlapfront; 416 extern struct evcnt tcp_reass_segdup; 417 extern struct evcnt tcp_reass_fragdup; 418 419 #define TCP_REASS_COUNTER_INCR(ev) (ev)->ev_count++ 420 421 #else 422 423 #define TCP_REASS_COUNTER_INCR(ev) /* nothing */ 424 425 #endif /* TCP_REASS_COUNTERS */ 426 427 static int tcp_reass(struct tcpcb *, const struct tcphdr *, struct mbuf *, 428 int); 429 static int tcp_dooptions(struct tcpcb *, const u_char *, int, 430 struct tcphdr *, struct mbuf *, int, struct tcp_opt_info *); 431 432 static void tcp4_log_refused(const struct ip *, const struct tcphdr *); 433 #ifdef INET6 434 static void tcp6_log_refused(const struct ip6_hdr *, const struct tcphdr *); 435 #endif 436 437 #if defined(MBUFTRACE) 438 struct mowner tcp_reass_mowner = MOWNER_INIT("tcp", "reass"); 439 #endif /* defined(MBUFTRACE) */ 440 441 static struct pool tcpipqent_pool; 442 443 void 444 tcpipqent_init(void) 445 { 446 447 pool_init(&tcpipqent_pool, sizeof(struct ipqent), 0, 0, 0, "tcpipqepl", 448 NULL, IPL_VM); 449 } 450 451 struct ipqent * 452 tcpipqent_alloc(void) 453 { 454 struct ipqent *ipqe; 455 int s; 456 457 s = splvm(); 458 ipqe = pool_get(&tcpipqent_pool, PR_NOWAIT); 459 splx(s); 460 461 return ipqe; 462 } 463 464 void 465 tcpipqent_free(struct ipqent *ipqe) 466 { 467 int s; 468 469 s = splvm(); 470 pool_put(&tcpipqent_pool, ipqe); 471 splx(s); 472 } 473 474 /* 475 * Insert segment ti into reassembly queue of tcp with 476 * control block tp. Return TH_FIN if reassembly now includes 477 * a segment with FIN. 478 */ 479 static int 480 tcp_reass(struct tcpcb *tp, const struct tcphdr *th, struct mbuf *m, int tlen) 481 { 482 struct ipqent *p, *q, *nq, *tiqe = NULL; 483 struct socket *so = NULL; 484 int pkt_flags; 485 tcp_seq pkt_seq; 486 unsigned pkt_len; 487 u_long rcvpartdupbyte = 0; 488 u_long rcvoobyte; 489 #ifdef TCP_REASS_COUNTERS 490 u_int count = 0; 491 #endif 492 uint64_t *tcps; 493 494 if (tp->t_inpcb) 495 so = tp->t_inpcb->inp_socket; 496 #ifdef INET6 497 else if (tp->t_in6pcb) 498 so = tp->t_in6pcb->in6p_socket; 499 #endif 500 501 TCP_REASS_LOCK_CHECK(tp); 502 503 /* 504 * Call with th==NULL after become established to 505 * force pre-ESTABLISHED data up to user socket. 506 */ 507 if (th == NULL) 508 goto present; 509 510 m_claimm(m, &tcp_reass_mowner); 511 512 rcvoobyte = tlen; 513 /* 514 * Copy these to local variables because the TCP header gets munged 515 * while we are collapsing mbufs. 516 */ 517 pkt_seq = th->th_seq; 518 pkt_len = tlen; 519 pkt_flags = th->th_flags; 520 521 TCP_REASS_COUNTER_INCR(&tcp_reass_); 522 523 if ((p = TAILQ_LAST(&tp->segq, ipqehead)) != NULL) { 524 /* 525 * When we miss a packet, the vast majority of time we get 526 * packets that follow it in order. So optimize for that. 527 */ 528 if (pkt_seq == p->ipqe_seq + p->ipqe_len) { 529 p->ipqe_len += pkt_len; 530 p->ipqe_flags |= pkt_flags; 531 m_cat(p->ipqe_m, m); 532 m = NULL; 533 tiqe = p; 534 TAILQ_REMOVE(&tp->timeq, p, ipqe_timeq); 535 TCP_REASS_COUNTER_INCR(&tcp_reass_appendtail); 536 goto skip_replacement; 537 } 538 /* 539 * While we're here, if the pkt is completely beyond 540 * anything we have, just insert it at the tail. 541 */ 542 if (SEQ_GT(pkt_seq, p->ipqe_seq + p->ipqe_len)) { 543 TCP_REASS_COUNTER_INCR(&tcp_reass_inserttail); 544 goto insert_it; 545 } 546 } 547 548 q = TAILQ_FIRST(&tp->segq); 549 550 if (q != NULL) { 551 /* 552 * If this segment immediately precedes the first out-of-order 553 * block, simply slap the segment in front of it and (mostly) 554 * skip the complicated logic. 555 */ 556 if (pkt_seq + pkt_len == q->ipqe_seq) { 557 q->ipqe_seq = pkt_seq; 558 q->ipqe_len += pkt_len; 559 q->ipqe_flags |= pkt_flags; 560 m_cat(m, q->ipqe_m); 561 q->ipqe_m = m; 562 tiqe = q; 563 TAILQ_REMOVE(&tp->timeq, q, ipqe_timeq); 564 TCP_REASS_COUNTER_INCR(&tcp_reass_prependfirst); 565 goto skip_replacement; 566 } 567 } else { 568 TCP_REASS_COUNTER_INCR(&tcp_reass_empty); 569 } 570 571 /* 572 * Find a segment which begins after this one does. 573 */ 574 for (p = NULL; q != NULL; q = nq) { 575 nq = TAILQ_NEXT(q, ipqe_q); 576 #ifdef TCP_REASS_COUNTERS 577 count++; 578 #endif 579 580 /* 581 * If the received segment is just right after this 582 * fragment, merge the two together and then check 583 * for further overlaps. 584 */ 585 if (q->ipqe_seq + q->ipqe_len == pkt_seq) { 586 pkt_len += q->ipqe_len; 587 pkt_flags |= q->ipqe_flags; 588 pkt_seq = q->ipqe_seq; 589 m_cat(q->ipqe_m, m); 590 m = q->ipqe_m; 591 TCP_REASS_COUNTER_INCR(&tcp_reass_append); 592 goto free_ipqe; 593 } 594 595 /* 596 * If the received segment is completely past this 597 * fragment, we need to go to the next fragment. 598 */ 599 if (SEQ_LT(q->ipqe_seq + q->ipqe_len, pkt_seq)) { 600 p = q; 601 continue; 602 } 603 604 /* 605 * If the fragment is past the received segment, 606 * it (or any following) can't be concatenated. 607 */ 608 if (SEQ_GT(q->ipqe_seq, pkt_seq + pkt_len)) { 609 TCP_REASS_COUNTER_INCR(&tcp_reass_insert); 610 break; 611 } 612 613 /* 614 * We've received all the data in this segment before. 615 * Mark it as a duplicate and return. 616 */ 617 if (SEQ_LEQ(q->ipqe_seq, pkt_seq) && 618 SEQ_GEQ(q->ipqe_seq + q->ipqe_len, pkt_seq + pkt_len)) { 619 tcps = TCP_STAT_GETREF(); 620 tcps[TCP_STAT_RCVDUPPACK]++; 621 tcps[TCP_STAT_RCVDUPBYTE] += pkt_len; 622 TCP_STAT_PUTREF(); 623 tcp_new_dsack(tp, pkt_seq, pkt_len); 624 m_freem(m); 625 if (tiqe != NULL) { 626 tcpipqent_free(tiqe); 627 } 628 TCP_REASS_COUNTER_INCR(&tcp_reass_segdup); 629 goto out; 630 } 631 632 /* 633 * Received segment completely overlaps this fragment 634 * so we drop the fragment (this keeps the temporal 635 * ordering of segments correct). 636 */ 637 if (SEQ_GEQ(q->ipqe_seq, pkt_seq) && 638 SEQ_LEQ(q->ipqe_seq + q->ipqe_len, pkt_seq + pkt_len)) { 639 rcvpartdupbyte += q->ipqe_len; 640 m_freem(q->ipqe_m); 641 TCP_REASS_COUNTER_INCR(&tcp_reass_fragdup); 642 goto free_ipqe; 643 } 644 645 /* 646 * Received segment extends past the end of the fragment. 647 * Drop the overlapping bytes, merge the fragment and 648 * segment, and treat as a longer received packet. 649 */ 650 if (SEQ_LT(q->ipqe_seq, pkt_seq) && 651 SEQ_GT(q->ipqe_seq + q->ipqe_len, pkt_seq)) { 652 int overlap = q->ipqe_seq + q->ipqe_len - pkt_seq; 653 m_adj(m, overlap); 654 rcvpartdupbyte += overlap; 655 m_cat(q->ipqe_m, m); 656 m = q->ipqe_m; 657 pkt_seq = q->ipqe_seq; 658 pkt_len += q->ipqe_len - overlap; 659 rcvoobyte -= overlap; 660 TCP_REASS_COUNTER_INCR(&tcp_reass_overlaptail); 661 goto free_ipqe; 662 } 663 664 /* 665 * Received segment extends past the front of the fragment. 666 * Drop the overlapping bytes on the received packet. The 667 * packet will then be concatenated with this fragment a 668 * bit later. 669 */ 670 if (SEQ_GT(q->ipqe_seq, pkt_seq) && 671 SEQ_LT(q->ipqe_seq, pkt_seq + pkt_len)) { 672 int overlap = pkt_seq + pkt_len - q->ipqe_seq; 673 m_adj(m, -overlap); 674 pkt_len -= overlap; 675 rcvpartdupbyte += overlap; 676 TCP_REASS_COUNTER_INCR(&tcp_reass_overlapfront); 677 rcvoobyte -= overlap; 678 } 679 680 /* 681 * If the received segment immediately precedes this 682 * fragment then tack the fragment onto this segment 683 * and reinsert the data. 684 */ 685 if (q->ipqe_seq == pkt_seq + pkt_len) { 686 pkt_len += q->ipqe_len; 687 pkt_flags |= q->ipqe_flags; 688 m_cat(m, q->ipqe_m); 689 TAILQ_REMOVE(&tp->segq, q, ipqe_q); 690 TAILQ_REMOVE(&tp->timeq, q, ipqe_timeq); 691 tp->t_segqlen--; 692 KASSERT(tp->t_segqlen >= 0); 693 KASSERT(tp->t_segqlen != 0 || 694 (TAILQ_EMPTY(&tp->segq) && 695 TAILQ_EMPTY(&tp->timeq))); 696 if (tiqe == NULL) { 697 tiqe = q; 698 } else { 699 tcpipqent_free(q); 700 } 701 TCP_REASS_COUNTER_INCR(&tcp_reass_prepend); 702 break; 703 } 704 705 /* 706 * If the fragment is before the segment, remember it. 707 * When this loop is terminated, p will contain the 708 * pointer to the fragment that is right before the 709 * received segment. 710 */ 711 if (SEQ_LEQ(q->ipqe_seq, pkt_seq)) 712 p = q; 713 714 continue; 715 716 /* 717 * This is a common operation. It also will allow 718 * to save doing a malloc/free in most instances. 719 */ 720 free_ipqe: 721 TAILQ_REMOVE(&tp->segq, q, ipqe_q); 722 TAILQ_REMOVE(&tp->timeq, q, ipqe_timeq); 723 tp->t_segqlen--; 724 KASSERT(tp->t_segqlen >= 0); 725 KASSERT(tp->t_segqlen != 0 || 726 (TAILQ_EMPTY(&tp->segq) && TAILQ_EMPTY(&tp->timeq))); 727 if (tiqe == NULL) { 728 tiqe = q; 729 } else { 730 tcpipqent_free(q); 731 } 732 } 733 734 #ifdef TCP_REASS_COUNTERS 735 if (count > 7) 736 TCP_REASS_COUNTER_INCR(&tcp_reass_iteration[0]); 737 else if (count > 0) 738 TCP_REASS_COUNTER_INCR(&tcp_reass_iteration[count]); 739 #endif 740 741 insert_it: 742 /* 743 * Allocate a new queue entry (block) since the received segment 744 * did not collapse onto any other out-of-order block. If it had 745 * collapsed, tiqe would not be NULL and we would be reusing it. 746 * 747 * If the allocation fails, drop the packet. 748 */ 749 if (tiqe == NULL) { 750 tiqe = tcpipqent_alloc(); 751 if (tiqe == NULL) { 752 TCP_STATINC(TCP_STAT_RCVMEMDROP); 753 m_freem(m); 754 goto out; 755 } 756 } 757 758 /* 759 * Update the counters. 760 */ 761 tp->t_rcvoopack++; 762 tcps = TCP_STAT_GETREF(); 763 tcps[TCP_STAT_RCVOOPACK]++; 764 tcps[TCP_STAT_RCVOOBYTE] += rcvoobyte; 765 if (rcvpartdupbyte) { 766 tcps[TCP_STAT_RCVPARTDUPPACK]++; 767 tcps[TCP_STAT_RCVPARTDUPBYTE] += rcvpartdupbyte; 768 } 769 TCP_STAT_PUTREF(); 770 771 /* 772 * Insert the new fragment queue entry into both queues. 773 */ 774 tiqe->ipqe_m = m; 775 tiqe->ipqe_seq = pkt_seq; 776 tiqe->ipqe_len = pkt_len; 777 tiqe->ipqe_flags = pkt_flags; 778 if (p == NULL) { 779 TAILQ_INSERT_HEAD(&tp->segq, tiqe, ipqe_q); 780 } else { 781 TAILQ_INSERT_AFTER(&tp->segq, p, tiqe, ipqe_q); 782 } 783 tp->t_segqlen++; 784 785 skip_replacement: 786 TAILQ_INSERT_HEAD(&tp->timeq, tiqe, ipqe_timeq); 787 788 present: 789 /* 790 * Present data to user, advancing rcv_nxt through 791 * completed sequence space. 792 */ 793 if (TCPS_HAVEESTABLISHED(tp->t_state) == 0) 794 goto out; 795 q = TAILQ_FIRST(&tp->segq); 796 if (q == NULL || q->ipqe_seq != tp->rcv_nxt) 797 goto out; 798 if (tp->t_state == TCPS_SYN_RECEIVED && q->ipqe_len) 799 goto out; 800 801 tp->rcv_nxt += q->ipqe_len; 802 pkt_flags = q->ipqe_flags & TH_FIN; 803 nd_hint(tp); 804 805 TAILQ_REMOVE(&tp->segq, q, ipqe_q); 806 TAILQ_REMOVE(&tp->timeq, q, ipqe_timeq); 807 tp->t_segqlen--; 808 KASSERT(tp->t_segqlen >= 0); 809 KASSERT(tp->t_segqlen != 0 || 810 (TAILQ_EMPTY(&tp->segq) && TAILQ_EMPTY(&tp->timeq))); 811 if (so->so_state & SS_CANTRCVMORE) 812 m_freem(q->ipqe_m); 813 else 814 sbappendstream(&so->so_rcv, q->ipqe_m); 815 tcpipqent_free(q); 816 TCP_REASS_UNLOCK(tp); 817 sorwakeup(so); 818 return pkt_flags; 819 820 out: 821 TCP_REASS_UNLOCK(tp); 822 return 0; 823 } 824 825 #ifdef INET6 826 int 827 tcp6_input(struct mbuf **mp, int *offp, int proto) 828 { 829 struct mbuf *m = *mp; 830 831 /* 832 * draft-itojun-ipv6-tcp-to-anycast 833 * better place to put this in? 834 */ 835 if (m->m_flags & M_ANYCAST6) { 836 struct ip6_hdr *ip6; 837 if (m->m_len < sizeof(struct ip6_hdr)) { 838 if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) { 839 TCP_STATINC(TCP_STAT_RCVSHORT); 840 return IPPROTO_DONE; 841 } 842 } 843 ip6 = mtod(m, struct ip6_hdr *); 844 icmp6_error(m, ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_ADDR, 845 (char *)&ip6->ip6_dst - (char *)ip6); 846 return IPPROTO_DONE; 847 } 848 849 tcp_input(m, *offp, proto); 850 return IPPROTO_DONE; 851 } 852 #endif 853 854 static void 855 tcp4_log_refused(const struct ip *ip, const struct tcphdr *th) 856 { 857 char src[INET_ADDRSTRLEN]; 858 char dst[INET_ADDRSTRLEN]; 859 860 if (ip) { 861 in_print(src, sizeof(src), &ip->ip_src); 862 in_print(dst, sizeof(dst), &ip->ip_dst); 863 } else { 864 strlcpy(src, "(unknown)", sizeof(src)); 865 strlcpy(dst, "(unknown)", sizeof(dst)); 866 } 867 log(LOG_INFO, 868 "Connection attempt to TCP %s:%d from %s:%d\n", 869 dst, ntohs(th->th_dport), 870 src, ntohs(th->th_sport)); 871 } 872 873 #ifdef INET6 874 static void 875 tcp6_log_refused(const struct ip6_hdr *ip6, const struct tcphdr *th) 876 { 877 char src[INET6_ADDRSTRLEN]; 878 char dst[INET6_ADDRSTRLEN]; 879 880 if (ip6) { 881 in6_print(src, sizeof(src), &ip6->ip6_src); 882 in6_print(dst, sizeof(dst), &ip6->ip6_dst); 883 } else { 884 strlcpy(src, "(unknown v6)", sizeof(src)); 885 strlcpy(dst, "(unknown v6)", sizeof(dst)); 886 } 887 log(LOG_INFO, 888 "Connection attempt to TCP [%s]:%d from [%s]:%d\n", 889 dst, ntohs(th->th_dport), 890 src, ntohs(th->th_sport)); 891 } 892 #endif 893 894 /* 895 * Checksum extended TCP header and data. 896 */ 897 int 898 tcp_input_checksum(int af, struct mbuf *m, const struct tcphdr *th, 899 int toff, int off, int tlen) 900 { 901 struct ifnet *rcvif; 902 int s; 903 904 /* 905 * XXX it's better to record and check if this mbuf is 906 * already checked. 907 */ 908 909 rcvif = m_get_rcvif(m, &s); 910 if (__predict_false(rcvif == NULL)) 911 goto badcsum; /* XXX */ 912 913 switch (af) { 914 case AF_INET: 915 switch (m->m_pkthdr.csum_flags & 916 ((rcvif->if_csum_flags_rx & M_CSUM_TCPv4) | 917 M_CSUM_TCP_UDP_BAD | M_CSUM_DATA)) { 918 case M_CSUM_TCPv4|M_CSUM_TCP_UDP_BAD: 919 TCP_CSUM_COUNTER_INCR(&tcp_hwcsum_bad); 920 goto badcsum; 921 922 case M_CSUM_TCPv4|M_CSUM_DATA: { 923 u_int32_t hw_csum = m->m_pkthdr.csum_data; 924 925 TCP_CSUM_COUNTER_INCR(&tcp_hwcsum_data); 926 if (m->m_pkthdr.csum_flags & M_CSUM_NO_PSEUDOHDR) { 927 const struct ip *ip = 928 mtod(m, const struct ip *); 929 930 hw_csum = in_cksum_phdr(ip->ip_src.s_addr, 931 ip->ip_dst.s_addr, 932 htons(hw_csum + tlen + off + IPPROTO_TCP)); 933 } 934 if ((hw_csum ^ 0xffff) != 0) 935 goto badcsum; 936 break; 937 } 938 939 case M_CSUM_TCPv4: 940 /* Checksum was okay. */ 941 TCP_CSUM_COUNTER_INCR(&tcp_hwcsum_ok); 942 break; 943 944 default: 945 /* 946 * Must compute it ourselves. Maybe skip checksum 947 * on loopback interfaces. 948 */ 949 if (__predict_true(!(rcvif->if_flags & IFF_LOOPBACK) || 950 tcp_do_loopback_cksum)) { 951 TCP_CSUM_COUNTER_INCR(&tcp_swcsum); 952 if (in4_cksum(m, IPPROTO_TCP, toff, 953 tlen + off) != 0) 954 goto badcsum; 955 } 956 break; 957 } 958 break; 959 960 #ifdef INET6 961 case AF_INET6: 962 switch (m->m_pkthdr.csum_flags & 963 ((rcvif->if_csum_flags_rx & M_CSUM_TCPv6) | 964 M_CSUM_TCP_UDP_BAD | M_CSUM_DATA)) { 965 case M_CSUM_TCPv6|M_CSUM_TCP_UDP_BAD: 966 TCP_CSUM_COUNTER_INCR(&tcp6_hwcsum_bad); 967 goto badcsum; 968 969 #if 0 /* notyet */ 970 case M_CSUM_TCPv6|M_CSUM_DATA: 971 #endif 972 973 case M_CSUM_TCPv6: 974 /* Checksum was okay. */ 975 TCP_CSUM_COUNTER_INCR(&tcp6_hwcsum_ok); 976 break; 977 978 default: 979 /* 980 * Must compute it ourselves. Maybe skip checksum 981 * on loopback interfaces. 982 */ 983 if (__predict_true((m->m_flags & M_LOOP) == 0 || 984 tcp_do_loopback_cksum)) { 985 TCP_CSUM_COUNTER_INCR(&tcp6_swcsum); 986 if (in6_cksum(m, IPPROTO_TCP, toff, 987 tlen + off) != 0) 988 goto badcsum; 989 } 990 } 991 break; 992 #endif /* INET6 */ 993 } 994 m_put_rcvif(rcvif, &s); 995 996 return 0; 997 998 badcsum: 999 m_put_rcvif(rcvif, &s); 1000 TCP_STATINC(TCP_STAT_RCVBADSUM); 1001 return -1; 1002 } 1003 1004 /* 1005 * When a packet arrives addressed to a vestigial tcpbp, we 1006 * nevertheless have to respond to it per the spec. 1007 * 1008 * This code is duplicated from the one in tcp_input(). 1009 */ 1010 static void tcp_vtw_input(struct tcphdr *th, vestigial_inpcb_t *vp, 1011 struct mbuf *m, int tlen) 1012 { 1013 int tiflags; 1014 int todrop; 1015 uint32_t t_flags = 0; 1016 uint64_t *tcps; 1017 1018 tiflags = th->th_flags; 1019 todrop = vp->rcv_nxt - th->th_seq; 1020 1021 if (todrop > 0) { 1022 if (tiflags & TH_SYN) { 1023 tiflags &= ~TH_SYN; 1024 th->th_seq++; 1025 tcp_urp_drop(th, 1, &tiflags); 1026 todrop--; 1027 } 1028 if (todrop > tlen || 1029 (todrop == tlen && (tiflags & TH_FIN) == 0)) { 1030 /* 1031 * Any valid FIN or RST must be to the left of the 1032 * window. At this point the FIN or RST must be a 1033 * duplicate or out of sequence; drop it. 1034 */ 1035 if (tiflags & TH_RST) 1036 goto drop; 1037 tiflags &= ~(TH_FIN|TH_RST); 1038 1039 /* 1040 * Send an ACK to resynchronize and drop any data. 1041 * But keep on processing for RST or ACK. 1042 */ 1043 t_flags |= TF_ACKNOW; 1044 todrop = tlen; 1045 tcps = TCP_STAT_GETREF(); 1046 tcps[TCP_STAT_RCVDUPPACK] += 1; 1047 tcps[TCP_STAT_RCVDUPBYTE] += todrop; 1048 TCP_STAT_PUTREF(); 1049 } else if ((tiflags & TH_RST) && 1050 th->th_seq != vp->rcv_nxt) { 1051 /* 1052 * Test for reset before adjusting the sequence 1053 * number for overlapping data. 1054 */ 1055 goto dropafterack_ratelim; 1056 } else { 1057 tcps = TCP_STAT_GETREF(); 1058 tcps[TCP_STAT_RCVPARTDUPPACK] += 1; 1059 tcps[TCP_STAT_RCVPARTDUPBYTE] += todrop; 1060 TCP_STAT_PUTREF(); 1061 } 1062 1063 // tcp_new_dsack(tp, th->th_seq, todrop); 1064 // hdroptlen += todrop; /*drop from head afterwards*/ 1065 1066 th->th_seq += todrop; 1067 tlen -= todrop; 1068 tcp_urp_drop(th, todrop, &tiflags); 1069 } 1070 1071 /* 1072 * If new data are received on a connection after the 1073 * user processes are gone, then RST the other end. 1074 */ 1075 if (tlen) { 1076 TCP_STATINC(TCP_STAT_RCVAFTERCLOSE); 1077 goto dropwithreset; 1078 } 1079 1080 /* 1081 * If segment ends after window, drop trailing data 1082 * (and PUSH and FIN); if nothing left, just ACK. 1083 */ 1084 todrop = (th->th_seq + tlen) - (vp->rcv_nxt + vp->rcv_wnd); 1085 1086 if (todrop > 0) { 1087 TCP_STATINC(TCP_STAT_RCVPACKAFTERWIN); 1088 if (todrop >= tlen) { 1089 /* 1090 * The segment actually starts after the window. 1091 * th->th_seq + tlen - vp->rcv_nxt - vp->rcv_wnd >= tlen 1092 * th->th_seq - vp->rcv_nxt - vp->rcv_wnd >= 0 1093 * th->th_seq >= vp->rcv_nxt + vp->rcv_wnd 1094 */ 1095 TCP_STATADD(TCP_STAT_RCVBYTEAFTERWIN, tlen); 1096 1097 /* 1098 * If a new connection request is received 1099 * while in TIME_WAIT, drop the old connection 1100 * and start over if the sequence numbers 1101 * are above the previous ones. 1102 */ 1103 if ((tiflags & TH_SYN) && 1104 SEQ_GT(th->th_seq, vp->rcv_nxt)) { 1105 /* 1106 * We only support this in the !NOFDREF case, which 1107 * is to say: not here. 1108 */ 1109 goto dropwithreset; 1110 } 1111 1112 /* 1113 * If window is closed can only take segments at 1114 * window edge, and have to drop data and PUSH from 1115 * incoming segments. Continue processing, but 1116 * remember to ack. Otherwise, drop segment 1117 * and (if not RST) ack. 1118 */ 1119 if (vp->rcv_wnd == 0 && th->th_seq == vp->rcv_nxt) { 1120 t_flags |= TF_ACKNOW; 1121 TCP_STATINC(TCP_STAT_RCVWINPROBE); 1122 } else { 1123 goto dropafterack; 1124 } 1125 } else { 1126 TCP_STATADD(TCP_STAT_RCVBYTEAFTERWIN, todrop); 1127 } 1128 m_adj(m, -todrop); 1129 tlen -= todrop; 1130 tiflags &= ~(TH_PUSH|TH_FIN); 1131 } 1132 1133 if (tiflags & TH_RST) { 1134 if (th->th_seq != vp->rcv_nxt) 1135 goto dropafterack_ratelim; 1136 1137 vtw_del(vp->ctl, vp->vtw); 1138 goto drop; 1139 } 1140 1141 /* 1142 * If the ACK bit is off we drop the segment and return. 1143 */ 1144 if ((tiflags & TH_ACK) == 0) { 1145 if (t_flags & TF_ACKNOW) 1146 goto dropafterack; 1147 goto drop; 1148 } 1149 1150 /* 1151 * In TIME_WAIT state the only thing that should arrive 1152 * is a retransmission of the remote FIN. Acknowledge 1153 * it and restart the finack timer. 1154 */ 1155 vtw_restart(vp); 1156 goto dropafterack; 1157 1158 dropafterack: 1159 /* 1160 * Generate an ACK dropping incoming segment if it occupies 1161 * sequence space, where the ACK reflects our state. 1162 */ 1163 if (tiflags & TH_RST) 1164 goto drop; 1165 goto dropafterack2; 1166 1167 dropafterack_ratelim: 1168 /* 1169 * We may want to rate-limit ACKs against SYN/RST attack. 1170 */ 1171 if (ppsratecheck(&tcp_ackdrop_ppslim_last, &tcp_ackdrop_ppslim_count, 1172 tcp_ackdrop_ppslim) == 0) { 1173 /* XXX stat */ 1174 goto drop; 1175 } 1176 /* ...fall into dropafterack2... */ 1177 1178 dropafterack2: 1179 (void)tcp_respond(0, m, m, th, th->th_seq + tlen, th->th_ack, TH_ACK); 1180 return; 1181 1182 dropwithreset: 1183 /* 1184 * Generate a RST, dropping incoming segment. 1185 * Make ACK acceptable to originator of segment. 1186 */ 1187 if (tiflags & TH_RST) 1188 goto drop; 1189 1190 if (tiflags & TH_ACK) { 1191 tcp_respond(0, m, m, th, (tcp_seq)0, th->th_ack, TH_RST); 1192 } else { 1193 if (tiflags & TH_SYN) 1194 ++tlen; 1195 (void)tcp_respond(0, m, m, th, th->th_seq + tlen, (tcp_seq)0, 1196 TH_RST|TH_ACK); 1197 } 1198 return; 1199 drop: 1200 m_freem(m); 1201 } 1202 1203 /* 1204 * TCP input routine, follows pages 65-76 of RFC 793 very closely. 1205 */ 1206 void 1207 tcp_input(struct mbuf *m, int off, int proto) 1208 { 1209 struct tcphdr *th; 1210 struct ip *ip; 1211 struct inpcb *inp; 1212 #ifdef INET6 1213 struct ip6_hdr *ip6; 1214 struct in6pcb *in6p; 1215 #endif 1216 u_int8_t *optp = NULL; 1217 int optlen = 0; 1218 int len, tlen, hdroptlen = 0; 1219 struct tcpcb *tp = NULL; 1220 int tiflags; 1221 struct socket *so = NULL; 1222 int todrop, acked, ourfinisacked, needoutput = 0; 1223 bool dupseg; 1224 #ifdef TCP_DEBUG 1225 short ostate = 0; 1226 #endif 1227 u_long tiwin; 1228 struct tcp_opt_info opti; 1229 int thlen, iphlen; 1230 int af; /* af on the wire */ 1231 struct mbuf *tcp_saveti = NULL; 1232 uint32_t ts_rtt; 1233 uint8_t iptos; 1234 uint64_t *tcps; 1235 vestigial_inpcb_t vestige; 1236 1237 vestige.valid = 0; 1238 1239 MCLAIM(m, &tcp_rx_mowner); 1240 1241 TCP_STATINC(TCP_STAT_RCVTOTAL); 1242 1243 memset(&opti, 0, sizeof(opti)); 1244 opti.ts_present = 0; 1245 opti.maxseg = 0; 1246 1247 /* 1248 * RFC1122 4.2.3.10, p. 104: discard bcast/mcast SYN. 1249 * 1250 * TCP is, by definition, unicast, so we reject all 1251 * multicast outright. 1252 * 1253 * Note, there are additional src/dst address checks in 1254 * the AF-specific code below. 1255 */ 1256 if (m->m_flags & (M_BCAST|M_MCAST)) { 1257 /* XXX stat */ 1258 goto drop; 1259 } 1260 #ifdef INET6 1261 if (m->m_flags & M_ANYCAST6) { 1262 /* XXX stat */ 1263 goto drop; 1264 } 1265 #endif 1266 1267 M_REGION_GET(th, struct tcphdr *, m, off, sizeof(struct tcphdr)); 1268 if (th == NULL) { 1269 TCP_STATINC(TCP_STAT_RCVSHORT); 1270 return; 1271 } 1272 1273 /* 1274 * Enforce alignment requirements that are violated in 1275 * some cases, see kern/50766 for details. 1276 */ 1277 if (TCP_HDR_ALIGNED_P(th) == 0) { 1278 m = m_copyup(m, off + sizeof(struct tcphdr), 0); 1279 if (m == NULL) { 1280 TCP_STATINC(TCP_STAT_RCVSHORT); 1281 return; 1282 } 1283 th = (struct tcphdr *)(mtod(m, char *) + off); 1284 } 1285 KASSERT(TCP_HDR_ALIGNED_P(th)); 1286 1287 /* 1288 * Get IP and TCP header. 1289 * Note: IP leaves IP header in first mbuf. 1290 */ 1291 ip = mtod(m, struct ip *); 1292 #ifdef INET6 1293 ip6 = mtod(m, struct ip6_hdr *); 1294 #endif 1295 switch (ip->ip_v) { 1296 case 4: 1297 af = AF_INET; 1298 iphlen = sizeof(struct ip); 1299 1300 if (IN_MULTICAST(ip->ip_dst.s_addr) || 1301 in_broadcast(ip->ip_dst, m_get_rcvif_NOMPSAFE(m))) 1302 goto drop; 1303 1304 /* We do the checksum after PCB lookup... */ 1305 len = ntohs(ip->ip_len); 1306 tlen = len - off; 1307 iptos = ip->ip_tos; 1308 break; 1309 #ifdef INET6 1310 case 6: 1311 iphlen = sizeof(struct ip6_hdr); 1312 af = AF_INET6; 1313 1314 /* 1315 * Be proactive about unspecified IPv6 address in source. 1316 * As we use all-zero to indicate unbounded/unconnected pcb, 1317 * unspecified IPv6 address can be used to confuse us. 1318 * 1319 * Note that packets with unspecified IPv6 destination is 1320 * already dropped in ip6_input. 1321 */ 1322 if (IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src)) { 1323 /* XXX stat */ 1324 goto drop; 1325 } 1326 1327 /* 1328 * Make sure destination address is not multicast. 1329 * Source address checked in ip6_input(). 1330 */ 1331 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) { 1332 /* XXX stat */ 1333 goto drop; 1334 } 1335 1336 /* We do the checksum after PCB lookup... */ 1337 len = m->m_pkthdr.len; 1338 tlen = len - off; 1339 iptos = (ntohl(ip6->ip6_flow) >> 20) & 0xff; 1340 break; 1341 #endif 1342 default: 1343 m_freem(m); 1344 return; 1345 } 1346 1347 1348 /* 1349 * Check that TCP offset makes sense, pull out TCP options and 1350 * adjust length. 1351 */ 1352 thlen = th->th_off << 2; 1353 if (thlen < sizeof(struct tcphdr) || thlen > tlen) { 1354 TCP_STATINC(TCP_STAT_RCVBADOFF); 1355 goto drop; 1356 } 1357 tlen -= thlen; 1358 1359 if (thlen > sizeof(struct tcphdr)) { 1360 M_REGION_GET(th, struct tcphdr *, m, off, thlen); 1361 if (th == NULL) { 1362 TCP_STATINC(TCP_STAT_RCVSHORT); 1363 return; 1364 } 1365 KASSERT(TCP_HDR_ALIGNED_P(th)); 1366 optlen = thlen - sizeof(struct tcphdr); 1367 optp = ((u_int8_t *)th) + sizeof(struct tcphdr); 1368 1369 /* 1370 * Do quick retrieval of timestamp options. 1371 * 1372 * If timestamp is the only option and it's formatted as 1373 * recommended in RFC 1323 appendix A, we quickly get the 1374 * values now and don't bother calling tcp_dooptions(), 1375 * etc. 1376 */ 1377 if ((optlen == TCPOLEN_TSTAMP_APPA || 1378 (optlen > TCPOLEN_TSTAMP_APPA && 1379 optp[TCPOLEN_TSTAMP_APPA] == TCPOPT_EOL)) && 1380 be32dec(optp) == TCPOPT_TSTAMP_HDR && 1381 (th->th_flags & TH_SYN) == 0) { 1382 opti.ts_present = 1; 1383 opti.ts_val = be32dec(optp + 4); 1384 opti.ts_ecr = be32dec(optp + 8); 1385 optp = NULL; /* we've parsed the options */ 1386 } 1387 } 1388 tiflags = th->th_flags; 1389 1390 /* 1391 * Checksum extended TCP header and data 1392 */ 1393 if (tcp_input_checksum(af, m, th, off, thlen, tlen)) 1394 goto badcsum; 1395 1396 /* 1397 * Locate pcb for segment. 1398 */ 1399 findpcb: 1400 inp = NULL; 1401 #ifdef INET6 1402 in6p = NULL; 1403 #endif 1404 switch (af) { 1405 case AF_INET: 1406 inp = in_pcblookup_connect(&tcbtable, ip->ip_src, th->th_sport, 1407 ip->ip_dst, th->th_dport, &vestige); 1408 if (inp == NULL && !vestige.valid) { 1409 TCP_STATINC(TCP_STAT_PCBHASHMISS); 1410 inp = in_pcblookup_bind(&tcbtable, ip->ip_dst, 1411 th->th_dport); 1412 } 1413 #ifdef INET6 1414 if (inp == NULL && !vestige.valid) { 1415 struct in6_addr s, d; 1416 1417 /* mapped addr case */ 1418 in6_in_2_v4mapin6(&ip->ip_src, &s); 1419 in6_in_2_v4mapin6(&ip->ip_dst, &d); 1420 in6p = in6_pcblookup_connect(&tcbtable, &s, 1421 th->th_sport, &d, th->th_dport, 0, &vestige); 1422 if (in6p == 0 && !vestige.valid) { 1423 TCP_STATINC(TCP_STAT_PCBHASHMISS); 1424 in6p = in6_pcblookup_bind(&tcbtable, &d, 1425 th->th_dport, 0); 1426 } 1427 } 1428 #endif 1429 #ifndef INET6 1430 if (inp == NULL && !vestige.valid) 1431 #else 1432 if (inp == NULL && in6p == NULL && !vestige.valid) 1433 #endif 1434 { 1435 TCP_STATINC(TCP_STAT_NOPORT); 1436 if (tcp_log_refused && 1437 (tiflags & (TH_RST|TH_ACK|TH_SYN)) == TH_SYN) { 1438 tcp4_log_refused(ip, th); 1439 } 1440 tcp_fields_to_host(th); 1441 goto dropwithreset_ratelim; 1442 } 1443 #if defined(IPSEC) 1444 if (ipsec_used) { 1445 if (inp && ipsec_in_reject(m, inp)) { 1446 goto drop; 1447 } 1448 #ifdef INET6 1449 else if (in6p && ipsec_in_reject(m, in6p)) { 1450 goto drop; 1451 } 1452 #endif 1453 } 1454 #endif /*IPSEC*/ 1455 break; 1456 #ifdef INET6 1457 case AF_INET6: 1458 { 1459 int faith; 1460 1461 #if defined(NFAITH) && NFAITH > 0 1462 faith = faithprefix(&ip6->ip6_dst); 1463 #else 1464 faith = 0; 1465 #endif 1466 in6p = in6_pcblookup_connect(&tcbtable, &ip6->ip6_src, 1467 th->th_sport, &ip6->ip6_dst, th->th_dport, faith, &vestige); 1468 if (!in6p && !vestige.valid) { 1469 TCP_STATINC(TCP_STAT_PCBHASHMISS); 1470 in6p = in6_pcblookup_bind(&tcbtable, &ip6->ip6_dst, 1471 th->th_dport, faith); 1472 } 1473 if (!in6p && !vestige.valid) { 1474 TCP_STATINC(TCP_STAT_NOPORT); 1475 if (tcp_log_refused && 1476 (tiflags & (TH_RST|TH_ACK|TH_SYN)) == TH_SYN) { 1477 tcp6_log_refused(ip6, th); 1478 } 1479 tcp_fields_to_host(th); 1480 goto dropwithreset_ratelim; 1481 } 1482 #if defined(IPSEC) 1483 if (ipsec_used && in6p && ipsec_in_reject(m, in6p)) { 1484 goto drop; 1485 } 1486 #endif 1487 break; 1488 } 1489 #endif 1490 } 1491 1492 tcp_fields_to_host(th); 1493 1494 /* 1495 * If the state is CLOSED (i.e., TCB does not exist) then 1496 * all data in the incoming segment is discarded. 1497 * If the TCB exists but is in CLOSED state, it is embryonic, 1498 * but should either do a listen or a connect soon. 1499 */ 1500 tp = NULL; 1501 so = NULL; 1502 if (inp) { 1503 /* Check the minimum TTL for socket. */ 1504 if (ip->ip_ttl < inp->inp_ip_minttl) 1505 goto drop; 1506 1507 tp = intotcpcb(inp); 1508 so = inp->inp_socket; 1509 } 1510 #ifdef INET6 1511 else if (in6p) { 1512 tp = in6totcpcb(in6p); 1513 so = in6p->in6p_socket; 1514 } 1515 #endif 1516 else if (vestige.valid) { 1517 /* We do not support the resurrection of vtw tcpcps. */ 1518 tcp_vtw_input(th, &vestige, m, tlen); 1519 m = NULL; 1520 goto drop; 1521 } 1522 1523 if (tp == NULL) 1524 goto dropwithreset_ratelim; 1525 if (tp->t_state == TCPS_CLOSED) 1526 goto drop; 1527 1528 KASSERT(so->so_lock == softnet_lock); 1529 KASSERT(solocked(so)); 1530 1531 /* Unscale the window into a 32-bit value. */ 1532 if ((tiflags & TH_SYN) == 0) 1533 tiwin = th->th_win << tp->snd_scale; 1534 else 1535 tiwin = th->th_win; 1536 1537 #ifdef INET6 1538 /* save packet options if user wanted */ 1539 if (in6p && (in6p->in6p_flags & IN6P_CONTROLOPTS)) { 1540 if (in6p->in6p_options) { 1541 m_freem(in6p->in6p_options); 1542 in6p->in6p_options = NULL; 1543 } 1544 ip6_savecontrol(in6p, &in6p->in6p_options, ip6, m); 1545 } 1546 #endif 1547 1548 if (so->so_options & SO_DEBUG) { 1549 #ifdef TCP_DEBUG 1550 ostate = tp->t_state; 1551 #endif 1552 1553 tcp_saveti = NULL; 1554 if (iphlen + sizeof(struct tcphdr) > MHLEN) 1555 goto nosave; 1556 1557 if (m->m_len > iphlen && (m->m_flags & M_EXT) == 0) { 1558 tcp_saveti = m_copym(m, 0, iphlen, M_DONTWAIT); 1559 if (tcp_saveti == NULL) 1560 goto nosave; 1561 } else { 1562 MGETHDR(tcp_saveti, M_DONTWAIT, MT_HEADER); 1563 if (tcp_saveti == NULL) 1564 goto nosave; 1565 MCLAIM(m, &tcp_mowner); 1566 tcp_saveti->m_len = iphlen; 1567 m_copydata(m, 0, iphlen, 1568 mtod(tcp_saveti, void *)); 1569 } 1570 1571 if (M_TRAILINGSPACE(tcp_saveti) < sizeof(struct tcphdr)) { 1572 m_freem(tcp_saveti); 1573 tcp_saveti = NULL; 1574 } else { 1575 tcp_saveti->m_len += sizeof(struct tcphdr); 1576 memcpy(mtod(tcp_saveti, char *) + iphlen, th, 1577 sizeof(struct tcphdr)); 1578 } 1579 nosave:; 1580 } 1581 1582 if (so->so_options & SO_ACCEPTCONN) { 1583 union syn_cache_sa src; 1584 union syn_cache_sa dst; 1585 1586 KASSERT(tp->t_state == TCPS_LISTEN); 1587 1588 memset(&src, 0, sizeof(src)); 1589 memset(&dst, 0, sizeof(dst)); 1590 switch (af) { 1591 case AF_INET: 1592 src.sin.sin_len = sizeof(struct sockaddr_in); 1593 src.sin.sin_family = AF_INET; 1594 src.sin.sin_addr = ip->ip_src; 1595 src.sin.sin_port = th->th_sport; 1596 1597 dst.sin.sin_len = sizeof(struct sockaddr_in); 1598 dst.sin.sin_family = AF_INET; 1599 dst.sin.sin_addr = ip->ip_dst; 1600 dst.sin.sin_port = th->th_dport; 1601 break; 1602 #ifdef INET6 1603 case AF_INET6: 1604 src.sin6.sin6_len = sizeof(struct sockaddr_in6); 1605 src.sin6.sin6_family = AF_INET6; 1606 src.sin6.sin6_addr = ip6->ip6_src; 1607 src.sin6.sin6_port = th->th_sport; 1608 1609 dst.sin6.sin6_len = sizeof(struct sockaddr_in6); 1610 dst.sin6.sin6_family = AF_INET6; 1611 dst.sin6.sin6_addr = ip6->ip6_dst; 1612 dst.sin6.sin6_port = th->th_dport; 1613 break; 1614 #endif 1615 } 1616 1617 if ((tiflags & (TH_RST|TH_ACK|TH_SYN)) != TH_SYN) { 1618 if (tiflags & TH_RST) { 1619 syn_cache_reset(&src.sa, &dst.sa, th); 1620 } else if ((tiflags & (TH_ACK|TH_SYN)) == 1621 (TH_ACK|TH_SYN)) { 1622 /* 1623 * Received a SYN,ACK. This should never 1624 * happen while we are in LISTEN. Send an RST. 1625 */ 1626 goto badsyn; 1627 } else if (tiflags & TH_ACK) { 1628 so = syn_cache_get(&src.sa, &dst.sa, th, so, m); 1629 if (so == NULL) { 1630 /* 1631 * We don't have a SYN for this ACK; 1632 * send an RST. 1633 */ 1634 goto badsyn; 1635 } else if (so == (struct socket *)(-1)) { 1636 /* 1637 * We were unable to create the 1638 * connection. If the 3-way handshake 1639 * was completed, and RST has been 1640 * sent to the peer. Since the mbuf 1641 * might be in use for the reply, do 1642 * not free it. 1643 */ 1644 m = NULL; 1645 } else { 1646 /* 1647 * We have created a full-blown 1648 * connection. 1649 */ 1650 tp = NULL; 1651 inp = NULL; 1652 #ifdef INET6 1653 in6p = NULL; 1654 #endif 1655 switch (so->so_proto->pr_domain->dom_family) { 1656 case AF_INET: 1657 inp = sotoinpcb(so); 1658 tp = intotcpcb(inp); 1659 break; 1660 #ifdef INET6 1661 case AF_INET6: 1662 in6p = sotoin6pcb(so); 1663 tp = in6totcpcb(in6p); 1664 break; 1665 #endif 1666 } 1667 if (tp == NULL) 1668 goto badsyn; /*XXX*/ 1669 tiwin <<= tp->snd_scale; 1670 goto after_listen; 1671 } 1672 } else { 1673 /* 1674 * None of RST, SYN or ACK was set. 1675 * This is an invalid packet for a 1676 * TCB in LISTEN state. Send a RST. 1677 */ 1678 goto badsyn; 1679 } 1680 } else { 1681 /* 1682 * Received a SYN. 1683 */ 1684 1685 #ifdef INET6 1686 /* 1687 * If deprecated address is forbidden, we do 1688 * not accept SYN to deprecated interface 1689 * address to prevent any new inbound 1690 * connection from getting established. 1691 * When we do not accept SYN, we send a TCP 1692 * RST, with deprecated source address (instead 1693 * of dropping it). We compromise it as it is 1694 * much better for peer to send a RST, and 1695 * RST will be the final packet for the 1696 * exchange. 1697 * 1698 * If we do not forbid deprecated addresses, we 1699 * accept the SYN packet. RFC2462 does not 1700 * suggest dropping SYN in this case. 1701 * If we decipher RFC2462 5.5.4, it says like 1702 * this: 1703 * 1. use of deprecated addr with existing 1704 * communication is okay - "SHOULD continue 1705 * to be used" 1706 * 2. use of it with new communication: 1707 * (2a) "SHOULD NOT be used if alternate 1708 * address with sufficient scope is 1709 * available" 1710 * (2b) nothing mentioned otherwise. 1711 * Here we fall into (2b) case as we have no 1712 * choice in our source address selection - we 1713 * must obey the peer. 1714 * 1715 * The wording in RFC2462 is confusing, and 1716 * there are multiple description text for 1717 * deprecated address handling - worse, they 1718 * are not exactly the same. I believe 5.5.4 1719 * is the best one, so we follow 5.5.4. 1720 */ 1721 if (af == AF_INET6 && !ip6_use_deprecated) { 1722 struct in6_ifaddr *ia6; 1723 int s; 1724 struct ifnet *rcvif = m_get_rcvif(m, &s); 1725 if (rcvif == NULL) 1726 goto dropwithreset; /* XXX */ 1727 if ((ia6 = in6ifa_ifpwithaddr(rcvif, 1728 &ip6->ip6_dst)) && 1729 (ia6->ia6_flags & IN6_IFF_DEPRECATED)) { 1730 tp = NULL; 1731 m_put_rcvif(rcvif, &s); 1732 goto dropwithreset; 1733 } 1734 m_put_rcvif(rcvif, &s); 1735 } 1736 #endif 1737 1738 /* 1739 * LISTEN socket received a SYN from itself? This 1740 * can't possibly be valid; drop the packet. 1741 */ 1742 if (th->th_sport == th->th_dport) { 1743 int eq = 0; 1744 1745 switch (af) { 1746 case AF_INET: 1747 eq = in_hosteq(ip->ip_src, ip->ip_dst); 1748 break; 1749 #ifdef INET6 1750 case AF_INET6: 1751 eq = IN6_ARE_ADDR_EQUAL(&ip6->ip6_src, 1752 &ip6->ip6_dst); 1753 break; 1754 #endif 1755 } 1756 if (eq) { 1757 TCP_STATINC(TCP_STAT_BADSYN); 1758 goto drop; 1759 } 1760 } 1761 1762 /* 1763 * SYN looks ok; create compressed TCP 1764 * state for it. 1765 */ 1766 if (so->so_qlen <= so->so_qlimit && 1767 syn_cache_add(&src.sa, &dst.sa, th, off, 1768 so, m, optp, optlen, &opti)) 1769 m = NULL; 1770 } 1771 1772 goto drop; 1773 } 1774 1775 after_listen: 1776 /* 1777 * From here on, we're dealing with !LISTEN. 1778 */ 1779 KASSERT(tp->t_state != TCPS_LISTEN); 1780 1781 /* 1782 * Segment received on connection. 1783 * Reset idle time and keep-alive timer. 1784 */ 1785 tp->t_rcvtime = tcp_now; 1786 if (TCPS_HAVEESTABLISHED(tp->t_state)) 1787 TCP_TIMER_ARM(tp, TCPT_KEEP, tp->t_keepidle); 1788 1789 /* 1790 * Process options. 1791 */ 1792 #ifdef TCP_SIGNATURE 1793 if (optp || (tp->t_flags & TF_SIGNATURE)) 1794 #else 1795 if (optp) 1796 #endif 1797 if (tcp_dooptions(tp, optp, optlen, th, m, off, &opti) < 0) 1798 goto drop; 1799 1800 if (TCP_SACK_ENABLED(tp)) { 1801 tcp_del_sackholes(tp, th); 1802 } 1803 1804 if (TCP_ECN_ALLOWED(tp)) { 1805 if (tiflags & TH_CWR) { 1806 tp->t_flags &= ~TF_ECN_SND_ECE; 1807 } 1808 switch (iptos & IPTOS_ECN_MASK) { 1809 case IPTOS_ECN_CE: 1810 tp->t_flags |= TF_ECN_SND_ECE; 1811 TCP_STATINC(TCP_STAT_ECN_CE); 1812 break; 1813 case IPTOS_ECN_ECT0: 1814 TCP_STATINC(TCP_STAT_ECN_ECT); 1815 break; 1816 case IPTOS_ECN_ECT1: 1817 /* XXX: ignore for now -- rpaulo */ 1818 break; 1819 } 1820 /* 1821 * Congestion experienced. 1822 * Ignore if we are already trying to recover. 1823 */ 1824 if ((tiflags & TH_ECE) && SEQ_GEQ(tp->snd_una, tp->snd_recover)) 1825 tp->t_congctl->cong_exp(tp); 1826 } 1827 1828 if (opti.ts_present && opti.ts_ecr) { 1829 /* 1830 * Calculate the RTT from the returned time stamp and the 1831 * connection's time base. If the time stamp is later than 1832 * the current time, or is extremely old, fall back to non-1323 1833 * RTT calculation. Since ts_rtt is unsigned, we can test both 1834 * at the same time. 1835 * 1836 * Note that ts_rtt is in units of slow ticks (500 1837 * ms). Since most earthbound RTTs are < 500 ms, 1838 * observed values will have large quantization noise. 1839 * Our smoothed RTT is then the fraction of observed 1840 * samples that are 1 tick instead of 0 (times 500 1841 * ms). 1842 * 1843 * ts_rtt is increased by 1 to denote a valid sample, 1844 * with 0 indicating an invalid measurement. This 1845 * extra 1 must be removed when ts_rtt is used, or 1846 * else an erroneous extra 500 ms will result. 1847 */ 1848 ts_rtt = TCP_TIMESTAMP(tp) - opti.ts_ecr + 1; 1849 if (ts_rtt > TCP_PAWS_IDLE) 1850 ts_rtt = 0; 1851 } else { 1852 ts_rtt = 0; 1853 } 1854 1855 /* 1856 * Fast path: check for the two common cases of a uni-directional 1857 * data transfer. If: 1858 * o We are in the ESTABLISHED state, and 1859 * o The packet has no control flags, and 1860 * o The packet is in-sequence, and 1861 * o The window didn't change, and 1862 * o We are not retransmitting 1863 * It's a candidate. 1864 * 1865 * If the length (tlen) is zero and the ack moved forward, we're 1866 * the sender side of the transfer. Just free the data acked and 1867 * wake any higher level process that was blocked waiting for 1868 * space. 1869 * 1870 * If the length is non-zero and the ack didn't move, we're the 1871 * receiver side. If we're getting packets in-order (the reassembly 1872 * queue is empty), add the data to the socket buffer and note 1873 * that we need a delayed ack. 1874 */ 1875 if (tp->t_state == TCPS_ESTABLISHED && 1876 (tiflags & (TH_SYN|TH_FIN|TH_RST|TH_URG|TH_ECE|TH_CWR|TH_ACK)) 1877 == TH_ACK && 1878 (!opti.ts_present || TSTMP_GEQ(opti.ts_val, tp->ts_recent)) && 1879 th->th_seq == tp->rcv_nxt && 1880 tiwin && tiwin == tp->snd_wnd && 1881 tp->snd_nxt == tp->snd_max) { 1882 1883 /* 1884 * If last ACK falls within this segment's sequence numbers, 1885 * record the timestamp. 1886 * NOTE that the test is modified according to the latest 1887 * proposal of the tcplw (at) cray.com list (Braden 1993/04/26). 1888 * 1889 * note that we already know 1890 * TSTMP_GEQ(opti.ts_val, tp->ts_recent) 1891 */ 1892 if (opti.ts_present && SEQ_LEQ(th->th_seq, tp->last_ack_sent)) { 1893 tp->ts_recent_age = tcp_now; 1894 tp->ts_recent = opti.ts_val; 1895 } 1896 1897 if (tlen == 0) { 1898 /* Ack prediction. */ 1899 if (SEQ_GT(th->th_ack, tp->snd_una) && 1900 SEQ_LEQ(th->th_ack, tp->snd_max) && 1901 tp->snd_cwnd >= tp->snd_wnd && 1902 tp->t_partialacks < 0) { 1903 /* 1904 * this is a pure ack for outstanding data. 1905 */ 1906 if (ts_rtt) 1907 tcp_xmit_timer(tp, ts_rtt - 1); 1908 else if (tp->t_rtttime && 1909 SEQ_GT(th->th_ack, tp->t_rtseq)) 1910 tcp_xmit_timer(tp, 1911 tcp_now - tp->t_rtttime); 1912 acked = th->th_ack - tp->snd_una; 1913 tcps = TCP_STAT_GETREF(); 1914 tcps[TCP_STAT_PREDACK]++; 1915 tcps[TCP_STAT_RCVACKPACK]++; 1916 tcps[TCP_STAT_RCVACKBYTE] += acked; 1917 TCP_STAT_PUTREF(); 1918 nd_hint(tp); 1919 1920 if (acked > (tp->t_lastoff - tp->t_inoff)) 1921 tp->t_lastm = NULL; 1922 sbdrop(&so->so_snd, acked); 1923 tp->t_lastoff -= acked; 1924 1925 icmp_check(tp, th, acked); 1926 1927 tp->snd_una = th->th_ack; 1928 tp->snd_fack = tp->snd_una; 1929 if (SEQ_LT(tp->snd_high, tp->snd_una)) 1930 tp->snd_high = tp->snd_una; 1931 /* 1932 * drag snd_wl2 along so only newer 1933 * ACKs can update the window size. 1934 * also avoids the state where snd_wl2 1935 * is eventually larger than th_ack and thus 1936 * blocking the window update mechanism and 1937 * the connection gets stuck for a loooong 1938 * time in the zero sized send window state. 1939 * 1940 * see PR/kern 55567 1941 */ 1942 tp->snd_wl2 = tp->snd_una; 1943 1944 m_freem(m); 1945 1946 /* 1947 * If all outstanding data are acked, stop 1948 * retransmit timer, otherwise restart timer 1949 * using current (possibly backed-off) value. 1950 * If process is waiting for space, 1951 * wakeup/selnotify/signal. If data 1952 * are ready to send, let tcp_output 1953 * decide between more output or persist. 1954 */ 1955 if (tp->snd_una == tp->snd_max) 1956 TCP_TIMER_DISARM(tp, TCPT_REXMT); 1957 else if (TCP_TIMER_ISARMED(tp, 1958 TCPT_PERSIST) == 0) 1959 TCP_TIMER_ARM(tp, TCPT_REXMT, 1960 tp->t_rxtcur); 1961 1962 sowwakeup(so); 1963 if (so->so_snd.sb_cc) { 1964 KERNEL_LOCK(1, NULL); 1965 (void)tcp_output(tp); 1966 KERNEL_UNLOCK_ONE(NULL); 1967 } 1968 if (tcp_saveti) 1969 m_freem(tcp_saveti); 1970 return; 1971 } 1972 } else if (th->th_ack == tp->snd_una && 1973 TAILQ_FIRST(&tp->segq) == NULL && 1974 tlen <= sbspace(&so->so_rcv)) { 1975 int newsize = 0; 1976 1977 /* 1978 * this is a pure, in-sequence data packet 1979 * with nothing on the reassembly queue and 1980 * we have enough buffer space to take it. 1981 */ 1982 tp->rcv_nxt += tlen; 1983 1984 /* 1985 * Pull rcv_up up to prevent seq wrap relative to 1986 * rcv_nxt. 1987 */ 1988 tp->rcv_up = tp->rcv_nxt; 1989 1990 /* 1991 * Pull snd_wl1 up to prevent seq wrap relative to 1992 * th_seq. 1993 */ 1994 tp->snd_wl1 = th->th_seq; 1995 1996 tcps = TCP_STAT_GETREF(); 1997 tcps[TCP_STAT_PREDDAT]++; 1998 tcps[TCP_STAT_RCVPACK]++; 1999 tcps[TCP_STAT_RCVBYTE] += tlen; 2000 TCP_STAT_PUTREF(); 2001 nd_hint(tp); 2002 /* 2003 * Automatic sizing enables the performance of large buffers 2004 * and most of the efficiency of small ones by only allocating 2005 * space when it is needed. 2006 * 2007 * On the receive side the socket buffer memory is only rarely 2008 * used to any significant extent. This allows us to be much 2009 * more aggressive in scaling the receive socket buffer. For 2010 * the case that the buffer space is actually used to a large 2011 * extent and we run out of kernel memory we can simply drop 2012 * the new segments; TCP on the sender will just retransmit it 2013 * later. Setting the buffer size too big may only consume too 2014 * much kernel memory if the application doesn't read() from 2015 * the socket or packet loss or reordering makes use of the 2016 * reassembly queue. 2017 * 2018 * The criteria to step up the receive buffer one notch are: 2019 * 1. the number of bytes received during the time it takes 2020 * one timestamp to be reflected back to us (the RTT); 2021 * 2. received bytes per RTT is within seven eighth of the 2022 * current socket buffer size; 2023 * 3. receive buffer size has not hit maximal automatic size; 2024 * 2025 * This algorithm does one step per RTT at most and only if 2026 * we receive a bulk stream w/o packet losses or reorderings. 2027 * Shrinking the buffer during idle times is not necessary as 2028 * it doesn't consume any memory when idle. 2029 * 2030 * TODO: Only step up if the application is actually serving 2031 * the buffer to better manage the socket buffer resources. 2032 */ 2033 if (tcp_do_autorcvbuf && 2034 opti.ts_ecr && 2035 (so->so_rcv.sb_flags & SB_AUTOSIZE)) { 2036 if (opti.ts_ecr > tp->rfbuf_ts && 2037 opti.ts_ecr - tp->rfbuf_ts < PR_SLOWHZ) { 2038 if (tp->rfbuf_cnt > 2039 (so->so_rcv.sb_hiwat / 8 * 7) && 2040 so->so_rcv.sb_hiwat < 2041 tcp_autorcvbuf_max) { 2042 newsize = 2043 uimin(so->so_rcv.sb_hiwat + 2044 tcp_autorcvbuf_inc, 2045 tcp_autorcvbuf_max); 2046 } 2047 /* Start over with next RTT. */ 2048 tp->rfbuf_ts = 0; 2049 tp->rfbuf_cnt = 0; 2050 } else 2051 tp->rfbuf_cnt += tlen; /* add up */ 2052 } 2053 2054 /* 2055 * Drop TCP, IP headers and TCP options then add data 2056 * to socket buffer. 2057 */ 2058 if (so->so_state & SS_CANTRCVMORE) { 2059 m_freem(m); 2060 } else { 2061 /* 2062 * Set new socket buffer size. 2063 * Give up when limit is reached. 2064 */ 2065 if (newsize) 2066 if (!sbreserve(&so->so_rcv, 2067 newsize, so)) 2068 so->so_rcv.sb_flags &= ~SB_AUTOSIZE; 2069 m_adj(m, off + thlen); 2070 sbappendstream(&so->so_rcv, m); 2071 } 2072 sorwakeup(so); 2073 tcp_setup_ack(tp, th); 2074 if (tp->t_flags & TF_ACKNOW) { 2075 KERNEL_LOCK(1, NULL); 2076 (void)tcp_output(tp); 2077 KERNEL_UNLOCK_ONE(NULL); 2078 } 2079 if (tcp_saveti) 2080 m_freem(tcp_saveti); 2081 return; 2082 } 2083 } 2084 2085 /* 2086 * Compute mbuf offset to TCP data segment. 2087 */ 2088 hdroptlen = off + thlen; 2089 2090 /* 2091 * Calculate amount of space in receive window. Receive window is 2092 * amount of space in rcv queue, but not less than advertised 2093 * window. 2094 */ 2095 { 2096 int win; 2097 win = sbspace(&so->so_rcv); 2098 if (win < 0) 2099 win = 0; 2100 tp->rcv_wnd = imax(win, (int)(tp->rcv_adv - tp->rcv_nxt)); 2101 } 2102 2103 /* Reset receive buffer auto scaling when not in bulk receive mode. */ 2104 tp->rfbuf_ts = 0; 2105 tp->rfbuf_cnt = 0; 2106 2107 switch (tp->t_state) { 2108 /* 2109 * If the state is SYN_SENT: 2110 * if seg contains an ACK, but not for our SYN, drop the input. 2111 * if seg contains a RST, then drop the connection. 2112 * if seg does not contain SYN, then drop it. 2113 * Otherwise this is an acceptable SYN segment 2114 * initialize tp->rcv_nxt and tp->irs 2115 * if seg contains ack then advance tp->snd_una 2116 * if seg contains a ECE and ECN support is enabled, the stream 2117 * is ECN capable. 2118 * if SYN has been acked change to ESTABLISHED else SYN_RCVD state 2119 * arrange for segment to be acked (eventually) 2120 * continue processing rest of data/controls, beginning with URG 2121 */ 2122 case TCPS_SYN_SENT: 2123 if ((tiflags & TH_ACK) && 2124 (SEQ_LEQ(th->th_ack, tp->iss) || 2125 SEQ_GT(th->th_ack, tp->snd_max))) 2126 goto dropwithreset; 2127 if (tiflags & TH_RST) { 2128 if (tiflags & TH_ACK) 2129 tp = tcp_drop(tp, ECONNREFUSED); 2130 goto drop; 2131 } 2132 if ((tiflags & TH_SYN) == 0) 2133 goto drop; 2134 if (tiflags & TH_ACK) { 2135 tp->snd_una = th->th_ack; 2136 if (SEQ_LT(tp->snd_nxt, tp->snd_una)) 2137 tp->snd_nxt = tp->snd_una; 2138 if (SEQ_LT(tp->snd_high, tp->snd_una)) 2139 tp->snd_high = tp->snd_una; 2140 TCP_TIMER_DISARM(tp, TCPT_REXMT); 2141 2142 if ((tiflags & TH_ECE) && tcp_do_ecn) { 2143 tp->t_flags |= TF_ECN_PERMIT; 2144 TCP_STATINC(TCP_STAT_ECN_SHS); 2145 } 2146 } 2147 tp->irs = th->th_seq; 2148 tcp_rcvseqinit(tp); 2149 tp->t_flags |= TF_ACKNOW; 2150 tcp_mss_from_peer(tp, opti.maxseg); 2151 2152 /* 2153 * Initialize the initial congestion window. If we 2154 * had to retransmit the SYN, we must initialize cwnd 2155 * to 1 segment (i.e. the Loss Window). 2156 */ 2157 if (tp->t_flags & TF_SYN_REXMT) 2158 tp->snd_cwnd = tp->t_peermss; 2159 else { 2160 int ss = tcp_init_win; 2161 if (inp != NULL && in_localaddr(inp->inp_faddr)) 2162 ss = tcp_init_win_local; 2163 #ifdef INET6 2164 if (in6p != NULL && in6_localaddr(&in6p->in6p_faddr)) 2165 ss = tcp_init_win_local; 2166 #endif 2167 tp->snd_cwnd = TCP_INITIAL_WINDOW(ss, tp->t_peermss); 2168 } 2169 2170 tcp_rmx_rtt(tp); 2171 if (tiflags & TH_ACK) { 2172 TCP_STATINC(TCP_STAT_CONNECTS); 2173 /* 2174 * move tcp_established before soisconnected 2175 * because upcall handler can drive tcp_output 2176 * functionality. 2177 * XXX we might call soisconnected at the end of 2178 * all processing 2179 */ 2180 tcp_established(tp); 2181 soisconnected(so); 2182 /* Do window scaling on this connection? */ 2183 if ((tp->t_flags & (TF_RCVD_SCALE|TF_REQ_SCALE)) == 2184 (TF_RCVD_SCALE|TF_REQ_SCALE)) { 2185 tp->snd_scale = tp->requested_s_scale; 2186 tp->rcv_scale = tp->request_r_scale; 2187 } 2188 TCP_REASS_LOCK(tp); 2189 (void)tcp_reass(tp, NULL, NULL, tlen); 2190 /* 2191 * if we didn't have to retransmit the SYN, 2192 * use its rtt as our initial srtt & rtt var. 2193 */ 2194 if (tp->t_rtttime) 2195 tcp_xmit_timer(tp, tcp_now - tp->t_rtttime); 2196 } else { 2197 tp->t_state = TCPS_SYN_RECEIVED; 2198 } 2199 2200 /* 2201 * Advance th->th_seq to correspond to first data byte. 2202 * If data, trim to stay within window, 2203 * dropping FIN if necessary. 2204 */ 2205 th->th_seq++; 2206 if (tlen > tp->rcv_wnd) { 2207 todrop = tlen - tp->rcv_wnd; 2208 m_adj(m, -todrop); 2209 tlen = tp->rcv_wnd; 2210 tiflags &= ~TH_FIN; 2211 tcps = TCP_STAT_GETREF(); 2212 tcps[TCP_STAT_RCVPACKAFTERWIN]++; 2213 tcps[TCP_STAT_RCVBYTEAFTERWIN] += todrop; 2214 TCP_STAT_PUTREF(); 2215 } 2216 tp->snd_wl1 = th->th_seq - 1; 2217 tp->rcv_up = th->th_seq; 2218 goto step6; 2219 2220 /* 2221 * If the state is SYN_RECEIVED: 2222 * If seg contains an ACK, but not for our SYN, drop the input 2223 * and generate an RST. See page 36, rfc793 2224 */ 2225 case TCPS_SYN_RECEIVED: 2226 if ((tiflags & TH_ACK) && 2227 (SEQ_LEQ(th->th_ack, tp->iss) || 2228 SEQ_GT(th->th_ack, tp->snd_max))) 2229 goto dropwithreset; 2230 break; 2231 } 2232 2233 /* 2234 * From here on, we're dealing with !LISTEN and !SYN_SENT. 2235 */ 2236 KASSERT(tp->t_state != TCPS_LISTEN && 2237 tp->t_state != TCPS_SYN_SENT); 2238 2239 /* 2240 * RFC1323 PAWS: if we have a timestamp reply on this segment and 2241 * it's less than ts_recent, drop it. 2242 */ 2243 if (opti.ts_present && (tiflags & TH_RST) == 0 && tp->ts_recent && 2244 TSTMP_LT(opti.ts_val, tp->ts_recent)) { 2245 /* Check to see if ts_recent is over 24 days old. */ 2246 if (tcp_now - tp->ts_recent_age > TCP_PAWS_IDLE) { 2247 /* 2248 * Invalidate ts_recent. If this segment updates 2249 * ts_recent, the age will be reset later and ts_recent 2250 * will get a valid value. If it does not, setting 2251 * ts_recent to zero will at least satisfy the 2252 * requirement that zero be placed in the timestamp 2253 * echo reply when ts_recent isn't valid. The 2254 * age isn't reset until we get a valid ts_recent 2255 * because we don't want out-of-order segments to be 2256 * dropped when ts_recent is old. 2257 */ 2258 tp->ts_recent = 0; 2259 } else { 2260 tcps = TCP_STAT_GETREF(); 2261 tcps[TCP_STAT_RCVDUPPACK]++; 2262 tcps[TCP_STAT_RCVDUPBYTE] += tlen; 2263 tcps[TCP_STAT_PAWSDROP]++; 2264 TCP_STAT_PUTREF(); 2265 tcp_new_dsack(tp, th->th_seq, tlen); 2266 goto dropafterack; 2267 } 2268 } 2269 2270 /* 2271 * Check that at least some bytes of the segment are within the 2272 * receive window. If segment begins before rcv_nxt, drop leading 2273 * data (and SYN); if nothing left, just ack. 2274 */ 2275 todrop = tp->rcv_nxt - th->th_seq; 2276 dupseg = false; 2277 if (todrop > 0) { 2278 if (tiflags & TH_SYN) { 2279 tiflags &= ~TH_SYN; 2280 th->th_seq++; 2281 tcp_urp_drop(th, 1, &tiflags); 2282 todrop--; 2283 } 2284 if (todrop > tlen || 2285 (todrop == tlen && (tiflags & TH_FIN) == 0)) { 2286 /* 2287 * Any valid FIN or RST must be to the left of the 2288 * window. At this point the FIN or RST must be a 2289 * duplicate or out of sequence; drop it. 2290 */ 2291 if (tiflags & TH_RST) 2292 goto drop; 2293 tiflags &= ~(TH_FIN|TH_RST); 2294 2295 /* 2296 * Send an ACK to resynchronize and drop any data. 2297 * But keep on processing for RST or ACK. 2298 */ 2299 tp->t_flags |= TF_ACKNOW; 2300 todrop = tlen; 2301 dupseg = true; 2302 tcps = TCP_STAT_GETREF(); 2303 tcps[TCP_STAT_RCVDUPPACK]++; 2304 tcps[TCP_STAT_RCVDUPBYTE] += todrop; 2305 TCP_STAT_PUTREF(); 2306 } else if ((tiflags & TH_RST) && th->th_seq != tp->rcv_nxt) { 2307 /* 2308 * Test for reset before adjusting the sequence 2309 * number for overlapping data. 2310 */ 2311 goto dropafterack_ratelim; 2312 } else { 2313 tcps = TCP_STAT_GETREF(); 2314 tcps[TCP_STAT_RCVPARTDUPPACK]++; 2315 tcps[TCP_STAT_RCVPARTDUPBYTE] += todrop; 2316 TCP_STAT_PUTREF(); 2317 } 2318 tcp_new_dsack(tp, th->th_seq, todrop); 2319 hdroptlen += todrop; /* drop from head afterwards (m_adj) */ 2320 th->th_seq += todrop; 2321 tlen -= todrop; 2322 tcp_urp_drop(th, todrop, &tiflags); 2323 } 2324 2325 /* 2326 * If new data is received on a connection after the user processes 2327 * are gone, then RST the other end. 2328 */ 2329 if ((so->so_state & SS_NOFDREF) && 2330 tp->t_state > TCPS_CLOSE_WAIT && tlen) { 2331 tp = tcp_close(tp); 2332 TCP_STATINC(TCP_STAT_RCVAFTERCLOSE); 2333 goto dropwithreset; 2334 } 2335 2336 /* 2337 * If the segment ends after the window, drop trailing data (and 2338 * PUSH and FIN); if nothing left, just ACK. 2339 */ 2340 todrop = (th->th_seq + tlen) - (tp->rcv_nxt + tp->rcv_wnd); 2341 if (todrop > 0) { 2342 TCP_STATINC(TCP_STAT_RCVPACKAFTERWIN); 2343 if (todrop >= tlen) { 2344 /* 2345 * The segment actually starts after the window. 2346 * th->th_seq + tlen - tp->rcv_nxt - tp->rcv_wnd >= tlen 2347 * th->th_seq - tp->rcv_nxt - tp->rcv_wnd >= 0 2348 * th->th_seq >= tp->rcv_nxt + tp->rcv_wnd 2349 */ 2350 TCP_STATADD(TCP_STAT_RCVBYTEAFTERWIN, tlen); 2351 2352 /* 2353 * If a new connection request is received while in 2354 * TIME_WAIT, drop the old connection and start over 2355 * if the sequence numbers are above the previous 2356 * ones. 2357 * 2358 * NOTE: We need to put the header fields back into 2359 * network order. 2360 */ 2361 if ((tiflags & TH_SYN) && 2362 tp->t_state == TCPS_TIME_WAIT && 2363 SEQ_GT(th->th_seq, tp->rcv_nxt)) { 2364 tp = tcp_close(tp); 2365 tcp_fields_to_net(th); 2366 m_freem(tcp_saveti); 2367 tcp_saveti = NULL; 2368 goto findpcb; 2369 } 2370 2371 /* 2372 * If window is closed can only take segments at 2373 * window edge, and have to drop data and PUSH from 2374 * incoming segments. Continue processing, but 2375 * remember to ack. Otherwise, drop segment 2376 * and (if not RST) ack. 2377 */ 2378 if (tp->rcv_wnd == 0 && th->th_seq == tp->rcv_nxt) { 2379 KASSERT(todrop == tlen); 2380 tp->t_flags |= TF_ACKNOW; 2381 TCP_STATINC(TCP_STAT_RCVWINPROBE); 2382 } else { 2383 goto dropafterack; 2384 } 2385 } else { 2386 TCP_STATADD(TCP_STAT_RCVBYTEAFTERWIN, todrop); 2387 } 2388 m_adj(m, -todrop); 2389 tlen -= todrop; 2390 tiflags &= ~(TH_PUSH|TH_FIN); 2391 } 2392 2393 /* 2394 * If last ACK falls within this segment's sequence numbers, 2395 * record the timestamp. 2396 * NOTE: 2397 * 1) That the test incorporates suggestions from the latest 2398 * proposal of the tcplw (at) cray.com list (Braden 1993/04/26). 2399 * 2) That updating only on newer timestamps interferes with 2400 * our earlier PAWS tests, so this check should be solely 2401 * predicated on the sequence space of this segment. 2402 * 3) That we modify the segment boundary check to be 2403 * Last.ACK.Sent <= SEG.SEQ + SEG.Len 2404 * instead of RFC1323's 2405 * Last.ACK.Sent < SEG.SEQ + SEG.Len, 2406 * This modified check allows us to overcome RFC1323's 2407 * limitations as described in Stevens TCP/IP Illustrated 2408 * Vol. 2 p.869. In such cases, we can still calculate the 2409 * RTT correctly when RCV.NXT == Last.ACK.Sent. 2410 */ 2411 if (opti.ts_present && 2412 SEQ_LEQ(th->th_seq, tp->last_ack_sent) && 2413 SEQ_LEQ(tp->last_ack_sent, th->th_seq + tlen + 2414 ((tiflags & (TH_SYN|TH_FIN)) != 0))) { 2415 tp->ts_recent_age = tcp_now; 2416 tp->ts_recent = opti.ts_val; 2417 } 2418 2419 /* 2420 * If the RST bit is set examine the state: 2421 * RECEIVED state: 2422 * If passive open, return to LISTEN state. 2423 * If active open, inform user that connection was refused. 2424 * ESTABLISHED, FIN_WAIT_1, FIN_WAIT2, CLOSE_WAIT states: 2425 * Inform user that connection was reset, and close tcb. 2426 * CLOSING, LAST_ACK, TIME_WAIT states: 2427 * Close the tcb. 2428 */ 2429 if (tiflags & TH_RST) { 2430 if (th->th_seq != tp->rcv_nxt) 2431 goto dropafterack_ratelim; 2432 2433 switch (tp->t_state) { 2434 case TCPS_SYN_RECEIVED: 2435 so->so_error = ECONNREFUSED; 2436 goto close; 2437 2438 case TCPS_ESTABLISHED: 2439 case TCPS_FIN_WAIT_1: 2440 case TCPS_FIN_WAIT_2: 2441 case TCPS_CLOSE_WAIT: 2442 so->so_error = ECONNRESET; 2443 close: 2444 tp->t_state = TCPS_CLOSED; 2445 TCP_STATINC(TCP_STAT_DROPS); 2446 tp = tcp_close(tp); 2447 goto drop; 2448 2449 case TCPS_CLOSING: 2450 case TCPS_LAST_ACK: 2451 case TCPS_TIME_WAIT: 2452 tp = tcp_close(tp); 2453 goto drop; 2454 } 2455 } 2456 2457 /* 2458 * Since we've covered the SYN-SENT and SYN-RECEIVED states above 2459 * we must be in a synchronized state. RFC793 states (under Reset 2460 * Generation) that any unacceptable segment (an out-of-order SYN 2461 * qualifies) received in a synchronized state must elicit only an 2462 * empty acknowledgment segment ... and the connection remains in 2463 * the same state. 2464 */ 2465 if (tiflags & TH_SYN) { 2466 if (tp->rcv_nxt == th->th_seq) { 2467 tcp_respond(tp, m, m, th, (tcp_seq)0, th->th_ack - 1, 2468 TH_ACK); 2469 if (tcp_saveti) 2470 m_freem(tcp_saveti); 2471 return; 2472 } 2473 2474 goto dropafterack_ratelim; 2475 } 2476 2477 /* 2478 * If the ACK bit is off we drop the segment and return. 2479 */ 2480 if ((tiflags & TH_ACK) == 0) { 2481 if (tp->t_flags & TF_ACKNOW) 2482 goto dropafterack; 2483 goto drop; 2484 } 2485 2486 /* 2487 * From here on, we're doing ACK processing. 2488 */ 2489 2490 switch (tp->t_state) { 2491 /* 2492 * In SYN_RECEIVED state if the ack ACKs our SYN then enter 2493 * ESTABLISHED state and continue processing, otherwise 2494 * send an RST. 2495 */ 2496 case TCPS_SYN_RECEIVED: 2497 if (SEQ_GT(tp->snd_una, th->th_ack) || 2498 SEQ_GT(th->th_ack, tp->snd_max)) 2499 goto dropwithreset; 2500 TCP_STATINC(TCP_STAT_CONNECTS); 2501 soisconnected(so); 2502 tcp_established(tp); 2503 /* Do window scaling? */ 2504 if ((tp->t_flags & (TF_RCVD_SCALE|TF_REQ_SCALE)) == 2505 (TF_RCVD_SCALE|TF_REQ_SCALE)) { 2506 tp->snd_scale = tp->requested_s_scale; 2507 tp->rcv_scale = tp->request_r_scale; 2508 } 2509 TCP_REASS_LOCK(tp); 2510 (void)tcp_reass(tp, NULL, NULL, tlen); 2511 tp->snd_wl1 = th->th_seq - 1; 2512 /* FALLTHROUGH */ 2513 2514 /* 2515 * In ESTABLISHED state: drop duplicate ACKs; ACK out of range 2516 * ACKs. If the ack is in the range 2517 * tp->snd_una < th->th_ack <= tp->snd_max 2518 * then advance tp->snd_una to th->th_ack and drop 2519 * data from the retransmission queue. If this ACK reflects 2520 * more up to date window information we update our window information. 2521 */ 2522 case TCPS_ESTABLISHED: 2523 case TCPS_FIN_WAIT_1: 2524 case TCPS_FIN_WAIT_2: 2525 case TCPS_CLOSE_WAIT: 2526 case TCPS_CLOSING: 2527 case TCPS_LAST_ACK: 2528 case TCPS_TIME_WAIT: 2529 if (SEQ_LEQ(th->th_ack, tp->snd_una)) { 2530 if (tlen == 0 && !dupseg && tiwin == tp->snd_wnd) { 2531 TCP_STATINC(TCP_STAT_RCVDUPACK); 2532 /* 2533 * If we have outstanding data (other than 2534 * a window probe), this is a completely 2535 * duplicate ack (ie, window info didn't 2536 * change), the ack is the biggest we've 2537 * seen and we've seen exactly our rexmt 2538 * threshhold of them, assume a packet 2539 * has been dropped and retransmit it. 2540 * Kludge snd_nxt & the congestion 2541 * window so we send only this one 2542 * packet. 2543 */ 2544 if (TCP_TIMER_ISARMED(tp, TCPT_REXMT) == 0 || 2545 th->th_ack != tp->snd_una) 2546 tp->t_dupacks = 0; 2547 else if (tp->t_partialacks < 0 && 2548 (++tp->t_dupacks == tcprexmtthresh || 2549 TCP_FACK_FASTRECOV(tp))) { 2550 /* 2551 * Do the fast retransmit, and adjust 2552 * congestion control paramenters. 2553 */ 2554 if (tp->t_congctl->fast_retransmit(tp, th)) { 2555 /* False fast retransmit */ 2556 break; 2557 } 2558 goto drop; 2559 } else if (tp->t_dupacks > tcprexmtthresh) { 2560 tp->snd_cwnd += tp->t_segsz; 2561 KERNEL_LOCK(1, NULL); 2562 (void)tcp_output(tp); 2563 KERNEL_UNLOCK_ONE(NULL); 2564 goto drop; 2565 } 2566 } else { 2567 /* 2568 * If the ack appears to be very old, only 2569 * allow data that is in-sequence. This 2570 * makes it somewhat more difficult to insert 2571 * forged data by guessing sequence numbers. 2572 * Sent an ack to try to update the send 2573 * sequence number on the other side. 2574 */ 2575 if (tlen && th->th_seq != tp->rcv_nxt && 2576 SEQ_LT(th->th_ack, 2577 tp->snd_una - tp->max_sndwnd)) 2578 goto dropafterack; 2579 } 2580 break; 2581 } 2582 /* 2583 * If the congestion window was inflated to account 2584 * for the other side's cached packets, retract it. 2585 */ 2586 tp->t_congctl->fast_retransmit_newack(tp, th); 2587 2588 if (SEQ_GT(th->th_ack, tp->snd_max)) { 2589 TCP_STATINC(TCP_STAT_RCVACKTOOMUCH); 2590 goto dropafterack; 2591 } 2592 acked = th->th_ack - tp->snd_una; 2593 tcps = TCP_STAT_GETREF(); 2594 tcps[TCP_STAT_RCVACKPACK]++; 2595 tcps[TCP_STAT_RCVACKBYTE] += acked; 2596 TCP_STAT_PUTREF(); 2597 2598 /* 2599 * If we have a timestamp reply, update smoothed 2600 * round trip time. If no timestamp is present but 2601 * transmit timer is running and timed sequence 2602 * number was acked, update smoothed round trip time. 2603 * Since we now have an rtt measurement, cancel the 2604 * timer backoff (cf., Phil Karn's retransmit alg.). 2605 * Recompute the initial retransmit timer. 2606 */ 2607 if (ts_rtt) 2608 tcp_xmit_timer(tp, ts_rtt - 1); 2609 else if (tp->t_rtttime && SEQ_GT(th->th_ack, tp->t_rtseq)) 2610 tcp_xmit_timer(tp, tcp_now - tp->t_rtttime); 2611 2612 /* 2613 * If all outstanding data is acked, stop retransmit 2614 * timer and remember to restart (more output or persist). 2615 * If there is more data to be acked, restart retransmit 2616 * timer, using current (possibly backed-off) value. 2617 */ 2618 if (th->th_ack == tp->snd_max) { 2619 TCP_TIMER_DISARM(tp, TCPT_REXMT); 2620 needoutput = 1; 2621 } else if (TCP_TIMER_ISARMED(tp, TCPT_PERSIST) == 0) 2622 TCP_TIMER_ARM(tp, TCPT_REXMT, tp->t_rxtcur); 2623 2624 /* 2625 * New data has been acked, adjust the congestion window. 2626 */ 2627 tp->t_congctl->newack(tp, th); 2628 2629 nd_hint(tp); 2630 if (acked > so->so_snd.sb_cc) { 2631 tp->snd_wnd -= so->so_snd.sb_cc; 2632 sbdrop(&so->so_snd, (int)so->so_snd.sb_cc); 2633 ourfinisacked = 1; 2634 } else { 2635 if (acked > (tp->t_lastoff - tp->t_inoff)) 2636 tp->t_lastm = NULL; 2637 sbdrop(&so->so_snd, acked); 2638 tp->t_lastoff -= acked; 2639 if (tp->snd_wnd > acked) 2640 tp->snd_wnd -= acked; 2641 else 2642 tp->snd_wnd = 0; 2643 ourfinisacked = 0; 2644 } 2645 sowwakeup(so); 2646 2647 icmp_check(tp, th, acked); 2648 2649 tp->snd_una = th->th_ack; 2650 if (SEQ_GT(tp->snd_una, tp->snd_fack)) 2651 tp->snd_fack = tp->snd_una; 2652 if (SEQ_LT(tp->snd_nxt, tp->snd_una)) 2653 tp->snd_nxt = tp->snd_una; 2654 if (SEQ_LT(tp->snd_high, tp->snd_una)) 2655 tp->snd_high = tp->snd_una; 2656 2657 switch (tp->t_state) { 2658 2659 /* 2660 * In FIN_WAIT_1 STATE in addition to the processing 2661 * for the ESTABLISHED state if our FIN is now acknowledged 2662 * then enter FIN_WAIT_2. 2663 */ 2664 case TCPS_FIN_WAIT_1: 2665 if (ourfinisacked) { 2666 /* 2667 * If we can't receive any more 2668 * data, then closing user can proceed. 2669 * Starting the timer is contrary to the 2670 * specification, but if we don't get a FIN 2671 * we'll hang forever. 2672 */ 2673 if (so->so_state & SS_CANTRCVMORE) { 2674 soisdisconnected(so); 2675 if (tp->t_maxidle > 0) 2676 TCP_TIMER_ARM(tp, TCPT_2MSL, 2677 tp->t_maxidle); 2678 } 2679 tp->t_state = TCPS_FIN_WAIT_2; 2680 } 2681 break; 2682 2683 /* 2684 * In CLOSING STATE in addition to the processing for 2685 * the ESTABLISHED state if the ACK acknowledges our FIN 2686 * then enter the TIME-WAIT state, otherwise ignore 2687 * the segment. 2688 */ 2689 case TCPS_CLOSING: 2690 if (ourfinisacked) { 2691 tp->t_state = TCPS_TIME_WAIT; 2692 tcp_canceltimers(tp); 2693 TCP_TIMER_ARM(tp, TCPT_2MSL, 2 * tp->t_msl); 2694 soisdisconnected(so); 2695 } 2696 break; 2697 2698 /* 2699 * In LAST_ACK, we may still be waiting for data to drain 2700 * and/or to be acked, as well as for the ack of our FIN. 2701 * If our FIN is now acknowledged, delete the TCB, 2702 * enter the closed state and return. 2703 */ 2704 case TCPS_LAST_ACK: 2705 if (ourfinisacked) { 2706 tp = tcp_close(tp); 2707 goto drop; 2708 } 2709 break; 2710 2711 /* 2712 * In TIME_WAIT state the only thing that should arrive 2713 * is a retransmission of the remote FIN. Acknowledge 2714 * it and restart the finack timer. 2715 */ 2716 case TCPS_TIME_WAIT: 2717 TCP_TIMER_ARM(tp, TCPT_2MSL, 2 * tp->t_msl); 2718 goto dropafterack; 2719 } 2720 } 2721 2722 step6: 2723 /* 2724 * Update window information. 2725 * Don't look at window if no ACK: TAC's send garbage on first SYN. 2726 */ 2727 if ((tiflags & TH_ACK) && (SEQ_LT(tp->snd_wl1, th->th_seq) || 2728 (tp->snd_wl1 == th->th_seq && (SEQ_LT(tp->snd_wl2, th->th_ack) || 2729 (tp->snd_wl2 == th->th_ack && tiwin > tp->snd_wnd))))) { 2730 /* keep track of pure window updates */ 2731 if (tlen == 0 && 2732 tp->snd_wl2 == th->th_ack && tiwin > tp->snd_wnd) 2733 TCP_STATINC(TCP_STAT_RCVWINUPD); 2734 tp->snd_wnd = tiwin; 2735 tp->snd_wl1 = th->th_seq; 2736 tp->snd_wl2 = th->th_ack; 2737 if (tp->snd_wnd > tp->max_sndwnd) 2738 tp->max_sndwnd = tp->snd_wnd; 2739 needoutput = 1; 2740 } 2741 2742 /* 2743 * Process segments with URG. 2744 */ 2745 if ((tiflags & TH_URG) && th->th_urp && 2746 TCPS_HAVERCVDFIN(tp->t_state) == 0) { 2747 /* 2748 * This is a kludge, but if we receive and accept 2749 * random urgent pointers, we'll crash in 2750 * soreceive. It's hard to imagine someone 2751 * actually wanting to send this much urgent data. 2752 */ 2753 if (th->th_urp + so->so_rcv.sb_cc > sb_max) { 2754 th->th_urp = 0; /* XXX */ 2755 tiflags &= ~TH_URG; /* XXX */ 2756 goto dodata; /* XXX */ 2757 } 2758 2759 /* 2760 * If this segment advances the known urgent pointer, 2761 * then mark the data stream. This should not happen 2762 * in CLOSE_WAIT, CLOSING, LAST_ACK or TIME_WAIT STATES since 2763 * a FIN has been received from the remote side. 2764 * In these states we ignore the URG. 2765 * 2766 * According to RFC961 (Assigned Protocols), 2767 * the urgent pointer points to the last octet 2768 * of urgent data. We continue, however, 2769 * to consider it to indicate the first octet 2770 * of data past the urgent section as the original 2771 * spec states (in one of two places). 2772 */ 2773 if (SEQ_GT(th->th_seq+th->th_urp, tp->rcv_up)) { 2774 tp->rcv_up = th->th_seq + th->th_urp; 2775 so->so_oobmark = so->so_rcv.sb_cc + 2776 (tp->rcv_up - tp->rcv_nxt) - 1; 2777 if (so->so_oobmark == 0) 2778 so->so_state |= SS_RCVATMARK; 2779 sohasoutofband(so); 2780 tp->t_oobflags &= ~(TCPOOB_HAVEDATA | TCPOOB_HADDATA); 2781 } 2782 2783 /* 2784 * Remove out of band data so doesn't get presented to user. 2785 * This can happen independent of advancing the URG pointer, 2786 * but if two URG's are pending at once, some out-of-band 2787 * data may creep in... ick. 2788 */ 2789 if (th->th_urp <= (u_int16_t)tlen && 2790 (so->so_options & SO_OOBINLINE) == 0) 2791 tcp_pulloutofband(so, th, m, hdroptlen); 2792 } else { 2793 /* 2794 * If no out of band data is expected, 2795 * pull receive urgent pointer along 2796 * with the receive window. 2797 */ 2798 if (SEQ_GT(tp->rcv_nxt, tp->rcv_up)) 2799 tp->rcv_up = tp->rcv_nxt; 2800 } 2801 dodata: 2802 2803 /* 2804 * Process the segment text, merging it into the TCP sequencing queue, 2805 * and arranging for acknowledgement of receipt if necessary. 2806 * This process logically involves adjusting tp->rcv_wnd as data 2807 * is presented to the user (this happens in tcp_usrreq.c, 2808 * tcp_rcvd()). If a FIN has already been received on this 2809 * connection then we just ignore the text. 2810 */ 2811 if ((tlen || (tiflags & TH_FIN)) && 2812 TCPS_HAVERCVDFIN(tp->t_state) == 0) { 2813 /* 2814 * Handle the common case: 2815 * o Segment is the next to be received, and 2816 * o The queue is empty, and 2817 * o The connection is established 2818 * In this case, we avoid calling tcp_reass. 2819 * 2820 * tcp_setup_ack: set DELACK for segments received in order, 2821 * but ack immediately when segments are out of order (so that 2822 * fast retransmit can work). 2823 */ 2824 TCP_REASS_LOCK(tp); 2825 if (th->th_seq == tp->rcv_nxt && 2826 TAILQ_FIRST(&tp->segq) == NULL && 2827 tp->t_state == TCPS_ESTABLISHED) { 2828 tcp_setup_ack(tp, th); 2829 tp->rcv_nxt += tlen; 2830 tiflags = th->th_flags & TH_FIN; 2831 tcps = TCP_STAT_GETREF(); 2832 tcps[TCP_STAT_RCVPACK]++; 2833 tcps[TCP_STAT_RCVBYTE] += tlen; 2834 TCP_STAT_PUTREF(); 2835 nd_hint(tp); 2836 if (so->so_state & SS_CANTRCVMORE) { 2837 m_freem(m); 2838 } else { 2839 m_adj(m, hdroptlen); 2840 sbappendstream(&(so)->so_rcv, m); 2841 } 2842 TCP_REASS_UNLOCK(tp); 2843 sorwakeup(so); 2844 } else { 2845 m_adj(m, hdroptlen); 2846 tiflags = tcp_reass(tp, th, m, tlen); 2847 tp->t_flags |= TF_ACKNOW; 2848 } 2849 2850 /* 2851 * Note the amount of data that peer has sent into 2852 * our window, in order to estimate the sender's 2853 * buffer size. 2854 */ 2855 len = so->so_rcv.sb_hiwat - (tp->rcv_adv - tp->rcv_nxt); 2856 } else { 2857 m_freem(m); 2858 m = NULL; 2859 tiflags &= ~TH_FIN; 2860 } 2861 2862 /* 2863 * If FIN is received ACK the FIN and let the user know 2864 * that the connection is closing. Ignore a FIN received before 2865 * the connection is fully established. 2866 */ 2867 if ((tiflags & TH_FIN) && TCPS_HAVEESTABLISHED(tp->t_state)) { 2868 if (TCPS_HAVERCVDFIN(tp->t_state) == 0) { 2869 socantrcvmore(so); 2870 tp->t_flags |= TF_ACKNOW; 2871 tp->rcv_nxt++; 2872 } 2873 switch (tp->t_state) { 2874 2875 /* 2876 * In ESTABLISHED STATE enter the CLOSE_WAIT state. 2877 */ 2878 case TCPS_ESTABLISHED: 2879 tp->t_state = TCPS_CLOSE_WAIT; 2880 break; 2881 2882 /* 2883 * If still in FIN_WAIT_1 STATE FIN has not been acked so 2884 * enter the CLOSING state. 2885 */ 2886 case TCPS_FIN_WAIT_1: 2887 tp->t_state = TCPS_CLOSING; 2888 break; 2889 2890 /* 2891 * In FIN_WAIT_2 state enter the TIME_WAIT state, 2892 * starting the time-wait timer, turning off the other 2893 * standard timers. 2894 */ 2895 case TCPS_FIN_WAIT_2: 2896 tp->t_state = TCPS_TIME_WAIT; 2897 tcp_canceltimers(tp); 2898 TCP_TIMER_ARM(tp, TCPT_2MSL, 2 * tp->t_msl); 2899 soisdisconnected(so); 2900 break; 2901 2902 /* 2903 * In TIME_WAIT state restart the 2 MSL time_wait timer. 2904 */ 2905 case TCPS_TIME_WAIT: 2906 TCP_TIMER_ARM(tp, TCPT_2MSL, 2 * tp->t_msl); 2907 break; 2908 } 2909 } 2910 #ifdef TCP_DEBUG 2911 if (so->so_options & SO_DEBUG) 2912 tcp_trace(TA_INPUT, ostate, tp, tcp_saveti, 0); 2913 #endif 2914 2915 /* 2916 * Return any desired output. 2917 */ 2918 if (needoutput || (tp->t_flags & TF_ACKNOW)) { 2919 KERNEL_LOCK(1, NULL); 2920 (void)tcp_output(tp); 2921 KERNEL_UNLOCK_ONE(NULL); 2922 } 2923 if (tcp_saveti) 2924 m_freem(tcp_saveti); 2925 2926 if (tp->t_state == TCPS_TIME_WAIT 2927 && (so->so_state & SS_NOFDREF) 2928 && (tp->t_inpcb || af != AF_INET) 2929 && (tp->t_in6pcb || af != AF_INET6) 2930 && ((af == AF_INET ? tcp4_vtw_enable : tcp6_vtw_enable) & 1) != 0 2931 && TAILQ_EMPTY(&tp->segq) 2932 && vtw_add(af, tp)) { 2933 ; 2934 } 2935 return; 2936 2937 badsyn: 2938 /* 2939 * Received a bad SYN. Increment counters and dropwithreset. 2940 */ 2941 TCP_STATINC(TCP_STAT_BADSYN); 2942 tp = NULL; 2943 goto dropwithreset; 2944 2945 dropafterack: 2946 /* 2947 * Generate an ACK dropping incoming segment if it occupies 2948 * sequence space, where the ACK reflects our state. 2949 */ 2950 if (tiflags & TH_RST) 2951 goto drop; 2952 goto dropafterack2; 2953 2954 dropafterack_ratelim: 2955 /* 2956 * We may want to rate-limit ACKs against SYN/RST attack. 2957 */ 2958 if (ppsratecheck(&tcp_ackdrop_ppslim_last, &tcp_ackdrop_ppslim_count, 2959 tcp_ackdrop_ppslim) == 0) { 2960 /* XXX stat */ 2961 goto drop; 2962 } 2963 2964 dropafterack2: 2965 m_freem(m); 2966 tp->t_flags |= TF_ACKNOW; 2967 KERNEL_LOCK(1, NULL); 2968 (void)tcp_output(tp); 2969 KERNEL_UNLOCK_ONE(NULL); 2970 if (tcp_saveti) 2971 m_freem(tcp_saveti); 2972 return; 2973 2974 dropwithreset_ratelim: 2975 /* 2976 * We may want to rate-limit RSTs in certain situations, 2977 * particularly if we are sending an RST in response to 2978 * an attempt to connect to or otherwise communicate with 2979 * a port for which we have no socket. 2980 */ 2981 if (ppsratecheck(&tcp_rst_ppslim_last, &tcp_rst_ppslim_count, 2982 tcp_rst_ppslim) == 0) { 2983 /* XXX stat */ 2984 goto drop; 2985 } 2986 2987 dropwithreset: 2988 /* 2989 * Generate a RST, dropping incoming segment. 2990 * Make ACK acceptable to originator of segment. 2991 */ 2992 if (tiflags & TH_RST) 2993 goto drop; 2994 if (tiflags & TH_ACK) { 2995 (void)tcp_respond(tp, m, m, th, (tcp_seq)0, th->th_ack, TH_RST); 2996 } else { 2997 if (tiflags & TH_SYN) 2998 tlen++; 2999 (void)tcp_respond(tp, m, m, th, th->th_seq + tlen, (tcp_seq)0, 3000 TH_RST|TH_ACK); 3001 } 3002 if (tcp_saveti) 3003 m_freem(tcp_saveti); 3004 return; 3005 3006 badcsum: 3007 drop: 3008 /* 3009 * Drop space held by incoming segment and return. 3010 */ 3011 if (tp) { 3012 if (tp->t_inpcb) 3013 so = tp->t_inpcb->inp_socket; 3014 #ifdef INET6 3015 else if (tp->t_in6pcb) 3016 so = tp->t_in6pcb->in6p_socket; 3017 #endif 3018 else 3019 so = NULL; 3020 #ifdef TCP_DEBUG 3021 if (so && (so->so_options & SO_DEBUG) != 0) 3022 tcp_trace(TA_DROP, ostate, tp, tcp_saveti, 0); 3023 #endif 3024 } 3025 if (tcp_saveti) 3026 m_freem(tcp_saveti); 3027 m_freem(m); 3028 return; 3029 } 3030 3031 #ifdef TCP_SIGNATURE 3032 int 3033 tcp_signature_apply(void *fstate, void *data, u_int len) 3034 { 3035 3036 MD5Update(fstate, (u_char *)data, len); 3037 return (0); 3038 } 3039 3040 struct secasvar * 3041 tcp_signature_getsav(struct mbuf *m) 3042 { 3043 struct ip *ip; 3044 struct ip6_hdr *ip6; 3045 3046 ip = mtod(m, struct ip *); 3047 switch (ip->ip_v) { 3048 case 4: 3049 ip = mtod(m, struct ip *); 3050 ip6 = NULL; 3051 break; 3052 case 6: 3053 ip = NULL; 3054 ip6 = mtod(m, struct ip6_hdr *); 3055 break; 3056 default: 3057 return (NULL); 3058 } 3059 3060 #ifdef IPSEC 3061 union sockaddr_union dst; 3062 3063 /* Extract the destination from the IP header in the mbuf. */ 3064 memset(&dst, 0, sizeof(union sockaddr_union)); 3065 if (ip != NULL) { 3066 dst.sa.sa_len = sizeof(struct sockaddr_in); 3067 dst.sa.sa_family = AF_INET; 3068 dst.sin.sin_addr = ip->ip_dst; 3069 } else { 3070 dst.sa.sa_len = sizeof(struct sockaddr_in6); 3071 dst.sa.sa_family = AF_INET6; 3072 dst.sin6.sin6_addr = ip6->ip6_dst; 3073 } 3074 3075 /* 3076 * Look up an SADB entry which matches the address of the peer. 3077 */ 3078 return KEY_LOOKUP_SA(&dst, IPPROTO_TCP, htonl(TCP_SIG_SPI), 0, 0); 3079 #else 3080 return NULL; 3081 #endif 3082 } 3083 3084 int 3085 tcp_signature(struct mbuf *m, struct tcphdr *th, int thoff, 3086 struct secasvar *sav, char *sig) 3087 { 3088 MD5_CTX ctx; 3089 struct ip *ip; 3090 struct ipovly *ipovly; 3091 #ifdef INET6 3092 struct ip6_hdr *ip6; 3093 struct ip6_hdr_pseudo ip6pseudo; 3094 #endif 3095 struct ippseudo ippseudo; 3096 struct tcphdr th0; 3097 int l, tcphdrlen; 3098 3099 if (sav == NULL) 3100 return (-1); 3101 3102 tcphdrlen = th->th_off * 4; 3103 3104 switch (mtod(m, struct ip *)->ip_v) { 3105 case 4: 3106 MD5Init(&ctx); 3107 ip = mtod(m, struct ip *); 3108 memset(&ippseudo, 0, sizeof(ippseudo)); 3109 ipovly = (struct ipovly *)ip; 3110 ippseudo.ippseudo_src = ipovly->ih_src; 3111 ippseudo.ippseudo_dst = ipovly->ih_dst; 3112 ippseudo.ippseudo_pad = 0; 3113 ippseudo.ippseudo_p = IPPROTO_TCP; 3114 ippseudo.ippseudo_len = htons(m->m_pkthdr.len - thoff); 3115 MD5Update(&ctx, (char *)&ippseudo, sizeof(ippseudo)); 3116 break; 3117 #if INET6 3118 case 6: 3119 MD5Init(&ctx); 3120 ip6 = mtod(m, struct ip6_hdr *); 3121 memset(&ip6pseudo, 0, sizeof(ip6pseudo)); 3122 ip6pseudo.ip6ph_src = ip6->ip6_src; 3123 in6_clearscope(&ip6pseudo.ip6ph_src); 3124 ip6pseudo.ip6ph_dst = ip6->ip6_dst; 3125 in6_clearscope(&ip6pseudo.ip6ph_dst); 3126 ip6pseudo.ip6ph_len = htons(m->m_pkthdr.len - thoff); 3127 ip6pseudo.ip6ph_nxt = IPPROTO_TCP; 3128 MD5Update(&ctx, (char *)&ip6pseudo, sizeof(ip6pseudo)); 3129 break; 3130 #endif 3131 default: 3132 return (-1); 3133 } 3134 3135 th0 = *th; 3136 th0.th_sum = 0; 3137 MD5Update(&ctx, (char *)&th0, sizeof(th0)); 3138 3139 l = m->m_pkthdr.len - thoff - tcphdrlen; 3140 if (l > 0) 3141 m_apply(m, thoff + tcphdrlen, 3142 m->m_pkthdr.len - thoff - tcphdrlen, 3143 tcp_signature_apply, &ctx); 3144 3145 MD5Update(&ctx, _KEYBUF(sav->key_auth), _KEYLEN(sav->key_auth)); 3146 MD5Final(sig, &ctx); 3147 3148 return (0); 3149 } 3150 #endif 3151 3152 /* 3153 * Parse and process tcp options. 3154 * 3155 * Returns -1 if this segment should be dropped. (eg. wrong signature) 3156 * Otherwise returns 0. 3157 */ 3158 static int 3159 tcp_dooptions(struct tcpcb *tp, const u_char *cp, int cnt, struct tcphdr *th, 3160 struct mbuf *m, int toff, struct tcp_opt_info *oi) 3161 { 3162 u_int16_t mss; 3163 int opt, optlen = 0; 3164 #ifdef TCP_SIGNATURE 3165 void *sigp = NULL; 3166 char sigbuf[TCP_SIGLEN]; 3167 struct secasvar *sav = NULL; 3168 #endif 3169 3170 for (; cp && cnt > 0; cnt -= optlen, cp += optlen) { 3171 opt = cp[0]; 3172 if (opt == TCPOPT_EOL) 3173 break; 3174 if (opt == TCPOPT_NOP) 3175 optlen = 1; 3176 else { 3177 if (cnt < 2) 3178 break; 3179 optlen = cp[1]; 3180 if (optlen < 2 || optlen > cnt) 3181 break; 3182 } 3183 switch (opt) { 3184 3185 default: 3186 continue; 3187 3188 case TCPOPT_MAXSEG: 3189 if (optlen != TCPOLEN_MAXSEG) 3190 continue; 3191 if (!(th->th_flags & TH_SYN)) 3192 continue; 3193 if (TCPS_HAVERCVDSYN(tp->t_state)) 3194 continue; 3195 memcpy(&mss, cp + 2, sizeof(mss)); 3196 oi->maxseg = ntohs(mss); 3197 break; 3198 3199 case TCPOPT_WINDOW: 3200 if (optlen != TCPOLEN_WINDOW) 3201 continue; 3202 if (!(th->th_flags & TH_SYN)) 3203 continue; 3204 if (TCPS_HAVERCVDSYN(tp->t_state)) 3205 continue; 3206 tp->t_flags |= TF_RCVD_SCALE; 3207 tp->requested_s_scale = cp[2]; 3208 if (tp->requested_s_scale > TCP_MAX_WINSHIFT) { 3209 char buf[INET6_ADDRSTRLEN]; 3210 struct ip *ip = mtod(m, struct ip *); 3211 #ifdef INET6 3212 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); 3213 #endif 3214 3215 switch (ip->ip_v) { 3216 case 4: 3217 in_print(buf, sizeof(buf), 3218 &ip->ip_src); 3219 break; 3220 #ifdef INET6 3221 case 6: 3222 in6_print(buf, sizeof(buf), 3223 &ip6->ip6_src); 3224 break; 3225 #endif 3226 default: 3227 strlcpy(buf, "(unknown)", sizeof(buf)); 3228 break; 3229 } 3230 3231 log(LOG_ERR, "TCP: invalid wscale %d from %s, " 3232 "assuming %d\n", 3233 tp->requested_s_scale, buf, 3234 TCP_MAX_WINSHIFT); 3235 tp->requested_s_scale = TCP_MAX_WINSHIFT; 3236 } 3237 break; 3238 3239 case TCPOPT_TIMESTAMP: 3240 if (optlen != TCPOLEN_TIMESTAMP) 3241 continue; 3242 oi->ts_present = 1; 3243 memcpy(&oi->ts_val, cp + 2, sizeof(oi->ts_val)); 3244 NTOHL(oi->ts_val); 3245 memcpy(&oi->ts_ecr, cp + 6, sizeof(oi->ts_ecr)); 3246 NTOHL(oi->ts_ecr); 3247 3248 if (!(th->th_flags & TH_SYN)) 3249 continue; 3250 if (TCPS_HAVERCVDSYN(tp->t_state)) 3251 continue; 3252 /* 3253 * A timestamp received in a SYN makes 3254 * it ok to send timestamp requests and replies. 3255 */ 3256 tp->t_flags |= TF_RCVD_TSTMP; 3257 tp->ts_recent = oi->ts_val; 3258 tp->ts_recent_age = tcp_now; 3259 break; 3260 3261 case TCPOPT_SACK_PERMITTED: 3262 if (optlen != TCPOLEN_SACK_PERMITTED) 3263 continue; 3264 if (!(th->th_flags & TH_SYN)) 3265 continue; 3266 if (TCPS_HAVERCVDSYN(tp->t_state)) 3267 continue; 3268 if (tcp_do_sack) { 3269 tp->t_flags |= TF_SACK_PERMIT; 3270 tp->t_flags |= TF_WILL_SACK; 3271 } 3272 break; 3273 3274 case TCPOPT_SACK: 3275 tcp_sack_option(tp, th, cp, optlen); 3276 break; 3277 #ifdef TCP_SIGNATURE 3278 case TCPOPT_SIGNATURE: 3279 if (optlen != TCPOLEN_SIGNATURE) 3280 continue; 3281 if (sigp && 3282 !consttime_memequal(sigp, cp + 2, TCP_SIGLEN)) 3283 return (-1); 3284 3285 sigp = sigbuf; 3286 memcpy(sigbuf, cp + 2, TCP_SIGLEN); 3287 tp->t_flags |= TF_SIGNATURE; 3288 break; 3289 #endif 3290 } 3291 } 3292 3293 #ifndef TCP_SIGNATURE 3294 return 0; 3295 #else 3296 if (tp->t_flags & TF_SIGNATURE) { 3297 sav = tcp_signature_getsav(m); 3298 if (sav == NULL && tp->t_state == TCPS_LISTEN) 3299 return (-1); 3300 } 3301 3302 if ((sigp ? TF_SIGNATURE : 0) ^ (tp->t_flags & TF_SIGNATURE)) 3303 goto out; 3304 3305 if (sigp) { 3306 char sig[TCP_SIGLEN]; 3307 3308 tcp_fields_to_net(th); 3309 if (tcp_signature(m, th, toff, sav, sig) < 0) { 3310 tcp_fields_to_host(th); 3311 goto out; 3312 } 3313 tcp_fields_to_host(th); 3314 3315 if (!consttime_memequal(sig, sigp, TCP_SIGLEN)) { 3316 TCP_STATINC(TCP_STAT_BADSIG); 3317 goto out; 3318 } else 3319 TCP_STATINC(TCP_STAT_GOODSIG); 3320 3321 key_sa_recordxfer(sav, m); 3322 KEY_SA_UNREF(&sav); 3323 } 3324 return 0; 3325 out: 3326 if (sav != NULL) 3327 KEY_SA_UNREF(&sav); 3328 return -1; 3329 #endif 3330 } 3331 3332 /* 3333 * Pull out of band byte out of a segment so 3334 * it doesn't appear in the user's data queue. 3335 * It is still reflected in the segment length for 3336 * sequencing purposes. 3337 */ 3338 void 3339 tcp_pulloutofband(struct socket *so, struct tcphdr *th, 3340 struct mbuf *m, int off) 3341 { 3342 int cnt = off + th->th_urp - 1; 3343 3344 while (cnt >= 0) { 3345 if (m->m_len > cnt) { 3346 char *cp = mtod(m, char *) + cnt; 3347 struct tcpcb *tp = sototcpcb(so); 3348 3349 tp->t_iobc = *cp; 3350 tp->t_oobflags |= TCPOOB_HAVEDATA; 3351 memmove(cp, cp + 1, (unsigned)(m->m_len - cnt - 1)); 3352 m->m_len--; 3353 return; 3354 } 3355 cnt -= m->m_len; 3356 m = m->m_next; 3357 if (m == NULL) 3358 break; 3359 } 3360 panic("tcp_pulloutofband"); 3361 } 3362 3363 /* 3364 * Collect new round-trip time estimate 3365 * and update averages and current timeout. 3366 * 3367 * rtt is in units of slow ticks (typically 500 ms) -- essentially the 3368 * difference of two timestamps. 3369 */ 3370 void 3371 tcp_xmit_timer(struct tcpcb *tp, uint32_t rtt) 3372 { 3373 int32_t delta; 3374 3375 TCP_STATINC(TCP_STAT_RTTUPDATED); 3376 if (tp->t_srtt != 0) { 3377 /* 3378 * Compute the amount to add to srtt for smoothing, 3379 * *alpha, or 2^(-TCP_RTT_SHIFT). Because 3380 * srtt is stored in 1/32 slow ticks, we conceptually 3381 * shift left 5 bits, subtract srtt to get the 3382 * diference, and then shift right by TCP_RTT_SHIFT 3383 * (3) to obtain 1/8 of the difference. 3384 */ 3385 delta = (rtt << 2) - (tp->t_srtt >> TCP_RTT_SHIFT); 3386 /* 3387 * This can never happen, because delta's lowest 3388 * possible value is 1/8 of t_srtt. But if it does, 3389 * set srtt to some reasonable value, here chosen 3390 * as 1/8 tick. 3391 */ 3392 if ((tp->t_srtt += delta) <= 0) 3393 tp->t_srtt = 1 << 2; 3394 /* 3395 * RFC2988 requires that rttvar be updated first. 3396 * This code is compliant because "delta" is the old 3397 * srtt minus the new observation (scaled). 3398 * 3399 * RFC2988 says: 3400 * rttvar = (1-beta) * rttvar + beta * |srtt-observed| 3401 * 3402 * delta is in units of 1/32 ticks, and has then been 3403 * divided by 8. This is equivalent to being in 1/16s 3404 * units and divided by 4. Subtract from it 1/4 of 3405 * the existing rttvar to form the (signed) amount to 3406 * adjust. 3407 */ 3408 if (delta < 0) 3409 delta = -delta; 3410 delta -= (tp->t_rttvar >> TCP_RTTVAR_SHIFT); 3411 /* 3412 * As with srtt, this should never happen. There is 3413 * no support in RFC2988 for this operation. But 1/4s 3414 * as rttvar when faced with something arguably wrong 3415 * is ok. 3416 */ 3417 if ((tp->t_rttvar += delta) <= 0) 3418 tp->t_rttvar = 1 << 2; 3419 3420 /* 3421 * If srtt exceeds .01 second, ensure we use the 'remote' MSL 3422 * Problem is: it doesn't work. Disabled by defaulting 3423 * tcp_rttlocal to 0; see corresponding code in 3424 * tcp_subr that selects local vs remote in a different way. 3425 * 3426 * The static branch prediction hint here should be removed 3427 * when the rtt estimator is fixed and the rtt_enable code 3428 * is turned back on. 3429 */ 3430 if (__predict_false(tcp_rttlocal) && tcp_msl_enable 3431 && tp->t_srtt > tcp_msl_remote_threshold 3432 && tp->t_msl < tcp_msl_remote) { 3433 tp->t_msl = MIN(tcp_msl_remote, TCP_MAXMSL); 3434 } 3435 } else { 3436 /* 3437 * This is the first measurement. Per RFC2988, 2.2, 3438 * set rtt=R and srtt=R/2. 3439 * For srtt, storage representation is 1/32 ticks, 3440 * so shift left by 5. 3441 * For rttvar, storage representation is 1/16 ticks, 3442 * So shift left by 4, but then right by 1 to halve. 3443 */ 3444 tp->t_srtt = rtt << (TCP_RTT_SHIFT + 2); 3445 tp->t_rttvar = rtt << (TCP_RTTVAR_SHIFT + 2 - 1); 3446 } 3447 tp->t_rtttime = 0; 3448 tp->t_rxtshift = 0; 3449 3450 /* 3451 * the retransmit should happen at rtt + 4 * rttvar. 3452 * Because of the way we do the smoothing, srtt and rttvar 3453 * will each average +1/2 tick of bias. When we compute 3454 * the retransmit timer, we want 1/2 tick of rounding and 3455 * 1 extra tick because of +-1/2 tick uncertainty in the 3456 * firing of the timer. The bias will give us exactly the 3457 * 1.5 tick we need. But, because the bias is 3458 * statistical, we have to test that we don't drop below 3459 * the minimum feasible timer (which is 2 ticks). 3460 */ 3461 TCPT_RANGESET(tp->t_rxtcur, TCP_REXMTVAL(tp), 3462 uimax(tp->t_rttmin, rtt + 2), TCPTV_REXMTMAX); 3463 3464 /* 3465 * We received an ack for a packet that wasn't retransmitted; 3466 * it is probably safe to discard any error indications we've 3467 * received recently. This isn't quite right, but close enough 3468 * for now (a route might have failed after we sent a segment, 3469 * and the return path might not be symmetrical). 3470 */ 3471 tp->t_softerror = 0; 3472 } 3473 3474 3475 /* 3476 * TCP compressed state engine. Currently used to hold compressed 3477 * state for SYN_RECEIVED. 3478 */ 3479 3480 u_long syn_cache_count; 3481 u_int32_t syn_hash1, syn_hash2; 3482 3483 #define SYN_HASH(sa, sp, dp) \ 3484 ((((sa)->s_addr^syn_hash1)*(((((u_int32_t)(dp))<<16) + \ 3485 ((u_int32_t)(sp)))^syn_hash2))) 3486 #ifndef INET6 3487 #define SYN_HASHALL(hash, src, dst) \ 3488 do { \ 3489 hash = SYN_HASH(&((const struct sockaddr_in *)(src))->sin_addr, \ 3490 ((const struct sockaddr_in *)(src))->sin_port, \ 3491 ((const struct sockaddr_in *)(dst))->sin_port); \ 3492 } while (/*CONSTCOND*/ 0) 3493 #else 3494 #define SYN_HASH6(sa, sp, dp) \ 3495 ((((sa)->s6_addr32[0] ^ (sa)->s6_addr32[3] ^ syn_hash1) * \ 3496 (((((u_int32_t)(dp))<<16) + ((u_int32_t)(sp)))^syn_hash2)) \ 3497 & 0x7fffffff) 3498 3499 #define SYN_HASHALL(hash, src, dst) \ 3500 do { \ 3501 switch ((src)->sa_family) { \ 3502 case AF_INET: \ 3503 hash = SYN_HASH(&((const struct sockaddr_in *)(src))->sin_addr, \ 3504 ((const struct sockaddr_in *)(src))->sin_port, \ 3505 ((const struct sockaddr_in *)(dst))->sin_port); \ 3506 break; \ 3507 case AF_INET6: \ 3508 hash = SYN_HASH6(&((const struct sockaddr_in6 *)(src))->sin6_addr, \ 3509 ((const struct sockaddr_in6 *)(src))->sin6_port, \ 3510 ((const struct sockaddr_in6 *)(dst))->sin6_port); \ 3511 break; \ 3512 default: \ 3513 hash = 0; \ 3514 } \ 3515 } while (/*CONSTCOND*/0) 3516 #endif /* INET6 */ 3517 3518 static struct pool syn_cache_pool; 3519 3520 /* 3521 * We don't estimate RTT with SYNs, so each packet starts with the default 3522 * RTT and each timer step has a fixed timeout value. 3523 */ 3524 static inline void 3525 syn_cache_timer_arm(struct syn_cache *sc) 3526 { 3527 3528 TCPT_RANGESET(sc->sc_rxtcur, 3529 TCPTV_SRTTDFLT * tcp_backoff[sc->sc_rxtshift], TCPTV_MIN, 3530 TCPTV_REXMTMAX); 3531 callout_reset(&sc->sc_timer, 3532 sc->sc_rxtcur * (hz / PR_SLOWHZ), syn_cache_timer, sc); 3533 } 3534 3535 #define SYN_CACHE_TIMESTAMP(sc) (tcp_now - (sc)->sc_timebase) 3536 3537 static inline void 3538 syn_cache_rm(struct syn_cache *sc) 3539 { 3540 TAILQ_REMOVE(&tcp_syn_cache[sc->sc_bucketidx].sch_bucket, 3541 sc, sc_bucketq); 3542 sc->sc_tp = NULL; 3543 LIST_REMOVE(sc, sc_tpq); 3544 tcp_syn_cache[sc->sc_bucketidx].sch_length--; 3545 callout_stop(&sc->sc_timer); 3546 syn_cache_count--; 3547 } 3548 3549 static inline void 3550 syn_cache_put(struct syn_cache *sc) 3551 { 3552 if (sc->sc_ipopts) 3553 (void) m_free(sc->sc_ipopts); 3554 rtcache_free(&sc->sc_route); 3555 sc->sc_flags |= SCF_DEAD; 3556 if (!callout_invoking(&sc->sc_timer)) 3557 callout_schedule(&(sc)->sc_timer, 1); 3558 } 3559 3560 void 3561 syn_cache_init(void) 3562 { 3563 int i; 3564 3565 pool_init(&syn_cache_pool, sizeof(struct syn_cache), 0, 0, 0, 3566 "synpl", NULL, IPL_SOFTNET); 3567 3568 /* Initialize the hash buckets. */ 3569 for (i = 0; i < tcp_syn_cache_size; i++) 3570 TAILQ_INIT(&tcp_syn_cache[i].sch_bucket); 3571 } 3572 3573 void 3574 syn_cache_insert(struct syn_cache *sc, struct tcpcb *tp) 3575 { 3576 struct syn_cache_head *scp; 3577 struct syn_cache *sc2; 3578 int s; 3579 3580 /* 3581 * If there are no entries in the hash table, reinitialize 3582 * the hash secrets. 3583 */ 3584 if (syn_cache_count == 0) { 3585 syn_hash1 = cprng_fast32(); 3586 syn_hash2 = cprng_fast32(); 3587 } 3588 3589 SYN_HASHALL(sc->sc_hash, &sc->sc_src.sa, &sc->sc_dst.sa); 3590 sc->sc_bucketidx = sc->sc_hash % tcp_syn_cache_size; 3591 scp = &tcp_syn_cache[sc->sc_bucketidx]; 3592 3593 /* 3594 * Make sure that we don't overflow the per-bucket 3595 * limit or the total cache size limit. 3596 */ 3597 s = splsoftnet(); 3598 if (scp->sch_length >= tcp_syn_bucket_limit) { 3599 TCP_STATINC(TCP_STAT_SC_BUCKETOVERFLOW); 3600 /* 3601 * The bucket is full. Toss the oldest element in the 3602 * bucket. This will be the first entry in the bucket. 3603 */ 3604 sc2 = TAILQ_FIRST(&scp->sch_bucket); 3605 #ifdef DIAGNOSTIC 3606 /* 3607 * This should never happen; we should always find an 3608 * entry in our bucket. 3609 */ 3610 if (sc2 == NULL) 3611 panic("syn_cache_insert: bucketoverflow: impossible"); 3612 #endif 3613 syn_cache_rm(sc2); 3614 syn_cache_put(sc2); /* calls pool_put but see spl above */ 3615 } else if (syn_cache_count >= tcp_syn_cache_limit) { 3616 struct syn_cache_head *scp2, *sce; 3617 3618 TCP_STATINC(TCP_STAT_SC_OVERFLOWED); 3619 /* 3620 * The cache is full. Toss the oldest entry in the 3621 * first non-empty bucket we can find. 3622 * 3623 * XXX We would really like to toss the oldest 3624 * entry in the cache, but we hope that this 3625 * condition doesn't happen very often. 3626 */ 3627 scp2 = scp; 3628 if (TAILQ_EMPTY(&scp2->sch_bucket)) { 3629 sce = &tcp_syn_cache[tcp_syn_cache_size]; 3630 for (++scp2; scp2 != scp; scp2++) { 3631 if (scp2 >= sce) 3632 scp2 = &tcp_syn_cache[0]; 3633 if (! TAILQ_EMPTY(&scp2->sch_bucket)) 3634 break; 3635 } 3636 #ifdef DIAGNOSTIC 3637 /* 3638 * This should never happen; we should always find a 3639 * non-empty bucket. 3640 */ 3641 if (scp2 == scp) 3642 panic("syn_cache_insert: cacheoverflow: " 3643 "impossible"); 3644 #endif 3645 } 3646 sc2 = TAILQ_FIRST(&scp2->sch_bucket); 3647 syn_cache_rm(sc2); 3648 syn_cache_put(sc2); /* calls pool_put but see spl above */ 3649 } 3650 3651 /* 3652 * Initialize the entry's timer. 3653 */ 3654 sc->sc_rxttot = 0; 3655 sc->sc_rxtshift = 0; 3656 syn_cache_timer_arm(sc); 3657 3658 /* Link it from tcpcb entry */ 3659 LIST_INSERT_HEAD(&tp->t_sc, sc, sc_tpq); 3660 3661 /* Put it into the bucket. */ 3662 TAILQ_INSERT_TAIL(&scp->sch_bucket, sc, sc_bucketq); 3663 scp->sch_length++; 3664 syn_cache_count++; 3665 3666 TCP_STATINC(TCP_STAT_SC_ADDED); 3667 splx(s); 3668 } 3669 3670 /* 3671 * Walk the timer queues, looking for SYN,ACKs that need to be retransmitted. 3672 * If we have retransmitted an entry the maximum number of times, expire 3673 * that entry. 3674 */ 3675 static void 3676 syn_cache_timer(void *arg) 3677 { 3678 struct syn_cache *sc = arg; 3679 3680 mutex_enter(softnet_lock); 3681 KERNEL_LOCK(1, NULL); 3682 3683 callout_ack(&sc->sc_timer); 3684 3685 if (__predict_false(sc->sc_flags & SCF_DEAD)) { 3686 TCP_STATINC(TCP_STAT_SC_DELAYED_FREE); 3687 goto free; 3688 } 3689 3690 if (__predict_false(sc->sc_rxtshift == TCP_MAXRXTSHIFT)) { 3691 /* Drop it -- too many retransmissions. */ 3692 goto dropit; 3693 } 3694 3695 /* 3696 * Compute the total amount of time this entry has 3697 * been on a queue. If this entry has been on longer 3698 * than the keep alive timer would allow, expire it. 3699 */ 3700 sc->sc_rxttot += sc->sc_rxtcur; 3701 if (sc->sc_rxttot >= MIN(tcp_keepinit, TCP_TIMER_MAXTICKS)) 3702 goto dropit; 3703 3704 TCP_STATINC(TCP_STAT_SC_RETRANSMITTED); 3705 (void)syn_cache_respond(sc); 3706 3707 /* Advance the timer back-off. */ 3708 sc->sc_rxtshift++; 3709 syn_cache_timer_arm(sc); 3710 3711 goto out; 3712 3713 dropit: 3714 TCP_STATINC(TCP_STAT_SC_TIMED_OUT); 3715 syn_cache_rm(sc); 3716 if (sc->sc_ipopts) 3717 (void) m_free(sc->sc_ipopts); 3718 rtcache_free(&sc->sc_route); 3719 3720 free: 3721 callout_destroy(&sc->sc_timer); 3722 pool_put(&syn_cache_pool, sc); 3723 3724 out: 3725 KERNEL_UNLOCK_ONE(NULL); 3726 mutex_exit(softnet_lock); 3727 } 3728 3729 /* 3730 * Remove syn cache created by the specified tcb entry, 3731 * because this does not make sense to keep them 3732 * (if there's no tcb entry, syn cache entry will never be used) 3733 */ 3734 void 3735 syn_cache_cleanup(struct tcpcb *tp) 3736 { 3737 struct syn_cache *sc, *nsc; 3738 int s; 3739 3740 s = splsoftnet(); 3741 3742 for (sc = LIST_FIRST(&tp->t_sc); sc != NULL; sc = nsc) { 3743 nsc = LIST_NEXT(sc, sc_tpq); 3744 3745 #ifdef DIAGNOSTIC 3746 if (sc->sc_tp != tp) 3747 panic("invalid sc_tp in syn_cache_cleanup"); 3748 #endif 3749 syn_cache_rm(sc); 3750 syn_cache_put(sc); /* calls pool_put but see spl above */ 3751 } 3752 /* just for safety */ 3753 LIST_INIT(&tp->t_sc); 3754 3755 splx(s); 3756 } 3757 3758 /* 3759 * Find an entry in the syn cache. 3760 */ 3761 struct syn_cache * 3762 syn_cache_lookup(const struct sockaddr *src, const struct sockaddr *dst, 3763 struct syn_cache_head **headp) 3764 { 3765 struct syn_cache *sc; 3766 struct syn_cache_head *scp; 3767 u_int32_t hash; 3768 int s; 3769 3770 SYN_HASHALL(hash, src, dst); 3771 3772 scp = &tcp_syn_cache[hash % tcp_syn_cache_size]; 3773 *headp = scp; 3774 s = splsoftnet(); 3775 for (sc = TAILQ_FIRST(&scp->sch_bucket); sc != NULL; 3776 sc = TAILQ_NEXT(sc, sc_bucketq)) { 3777 if (sc->sc_hash != hash) 3778 continue; 3779 if (!memcmp(&sc->sc_src, src, src->sa_len) && 3780 !memcmp(&sc->sc_dst, dst, dst->sa_len)) { 3781 splx(s); 3782 return (sc); 3783 } 3784 } 3785 splx(s); 3786 return (NULL); 3787 } 3788 3789 /* 3790 * This function gets called when we receive an ACK for a socket in the 3791 * LISTEN state. We look up the connection in the syn cache, and if it's 3792 * there, we pull it out of the cache and turn it into a full-blown 3793 * connection in the SYN-RECEIVED state. 3794 * 3795 * The return values may not be immediately obvious, and their effects 3796 * can be subtle, so here they are: 3797 * 3798 * NULL SYN was not found in cache; caller should drop the 3799 * packet and send an RST. 3800 * 3801 * -1 We were unable to create the new connection, and are 3802 * aborting it. An ACK,RST is being sent to the peer 3803 * (unless we got screwey sequence numbers; see below), 3804 * because the 3-way handshake has been completed. Caller 3805 * should not free the mbuf, since we may be using it. If 3806 * we are not, we will free it. 3807 * 3808 * Otherwise, the return value is a pointer to the new socket 3809 * associated with the connection. 3810 */ 3811 struct socket * 3812 syn_cache_get(struct sockaddr *src, struct sockaddr *dst, 3813 struct tcphdr *th, struct socket *so, struct mbuf *m) 3814 { 3815 struct syn_cache *sc; 3816 struct syn_cache_head *scp; 3817 struct inpcb *inp = NULL; 3818 #ifdef INET6 3819 struct in6pcb *in6p = NULL; 3820 #endif 3821 struct tcpcb *tp; 3822 int s; 3823 struct socket *oso; 3824 3825 s = splsoftnet(); 3826 if ((sc = syn_cache_lookup(src, dst, &scp)) == NULL) { 3827 splx(s); 3828 return NULL; 3829 } 3830 3831 /* 3832 * Verify the sequence and ack numbers. Try getting the correct 3833 * response again. 3834 */ 3835 if ((th->th_ack != sc->sc_iss + 1) || 3836 SEQ_LEQ(th->th_seq, sc->sc_irs) || 3837 SEQ_GT(th->th_seq, sc->sc_irs + 1 + sc->sc_win)) { 3838 m_freem(m); 3839 (void)syn_cache_respond(sc); 3840 splx(s); 3841 return ((struct socket *)(-1)); 3842 } 3843 3844 /* Remove this cache entry */ 3845 syn_cache_rm(sc); 3846 splx(s); 3847 3848 /* 3849 * Ok, create the full blown connection, and set things up 3850 * as they would have been set up if we had created the 3851 * connection when the SYN arrived. If we can't create 3852 * the connection, abort it. 3853 */ 3854 /* 3855 * inp still has the OLD in_pcb stuff, set the 3856 * v6-related flags on the new guy, too. This is 3857 * done particularly for the case where an AF_INET6 3858 * socket is bound only to a port, and a v4 connection 3859 * comes in on that port. 3860 * we also copy the flowinfo from the original pcb 3861 * to the new one. 3862 */ 3863 oso = so; 3864 so = sonewconn(so, true); 3865 if (so == NULL) 3866 goto resetandabort; 3867 3868 switch (so->so_proto->pr_domain->dom_family) { 3869 case AF_INET: 3870 inp = sotoinpcb(so); 3871 break; 3872 #ifdef INET6 3873 case AF_INET6: 3874 in6p = sotoin6pcb(so); 3875 break; 3876 #endif 3877 } 3878 3879 switch (src->sa_family) { 3880 case AF_INET: 3881 if (inp) { 3882 inp->inp_laddr = ((struct sockaddr_in *)dst)->sin_addr; 3883 inp->inp_lport = ((struct sockaddr_in *)dst)->sin_port; 3884 inp->inp_options = ip_srcroute(m); 3885 in_pcbstate(inp, INP_BOUND); 3886 if (inp->inp_options == NULL) { 3887 inp->inp_options = sc->sc_ipopts; 3888 sc->sc_ipopts = NULL; 3889 } 3890 } 3891 #ifdef INET6 3892 else if (in6p) { 3893 /* IPv4 packet to AF_INET6 socket */ 3894 memset(&in6p->in6p_laddr, 0, sizeof(in6p->in6p_laddr)); 3895 in6p->in6p_laddr.s6_addr16[5] = htons(0xffff); 3896 bcopy(&((struct sockaddr_in *)dst)->sin_addr, 3897 &in6p->in6p_laddr.s6_addr32[3], 3898 sizeof(((struct sockaddr_in *)dst)->sin_addr)); 3899 in6p->in6p_lport = ((struct sockaddr_in *)dst)->sin_port; 3900 in6totcpcb(in6p)->t_family = AF_INET; 3901 if (sotoin6pcb(oso)->in6p_flags & IN6P_IPV6_V6ONLY) 3902 in6p->in6p_flags |= IN6P_IPV6_V6ONLY; 3903 else 3904 in6p->in6p_flags &= ~IN6P_IPV6_V6ONLY; 3905 in6_pcbstate(in6p, IN6P_BOUND); 3906 } 3907 #endif 3908 break; 3909 #ifdef INET6 3910 case AF_INET6: 3911 if (in6p) { 3912 in6p->in6p_laddr = ((struct sockaddr_in6 *)dst)->sin6_addr; 3913 in6p->in6p_lport = ((struct sockaddr_in6 *)dst)->sin6_port; 3914 in6_pcbstate(in6p, IN6P_BOUND); 3915 } 3916 break; 3917 #endif 3918 } 3919 3920 #ifdef INET6 3921 if (in6p && in6totcpcb(in6p)->t_family == AF_INET6 && sotoinpcb(oso)) { 3922 struct in6pcb *oin6p = sotoin6pcb(oso); 3923 /* inherit socket options from the listening socket */ 3924 in6p->in6p_flags |= (oin6p->in6p_flags & IN6P_CONTROLOPTS); 3925 if (in6p->in6p_flags & IN6P_CONTROLOPTS) { 3926 m_freem(in6p->in6p_options); 3927 in6p->in6p_options = NULL; 3928 } 3929 ip6_savecontrol(in6p, &in6p->in6p_options, 3930 mtod(m, struct ip6_hdr *), m); 3931 } 3932 #endif 3933 3934 /* 3935 * Give the new socket our cached route reference. 3936 */ 3937 if (inp) { 3938 rtcache_copy(&inp->inp_route, &sc->sc_route); 3939 rtcache_free(&sc->sc_route); 3940 } 3941 #ifdef INET6 3942 else { 3943 rtcache_copy(&in6p->in6p_route, &sc->sc_route); 3944 rtcache_free(&sc->sc_route); 3945 } 3946 #endif 3947 3948 if (inp) { 3949 struct sockaddr_in sin; 3950 memcpy(&sin, src, src->sa_len); 3951 if (in_pcbconnect(inp, &sin, &lwp0)) { 3952 goto resetandabort; 3953 } 3954 } 3955 #ifdef INET6 3956 else if (in6p) { 3957 struct sockaddr_in6 sin6; 3958 memcpy(&sin6, src, src->sa_len); 3959 if (src->sa_family == AF_INET) { 3960 /* IPv4 packet to AF_INET6 socket */ 3961 in6_sin_2_v4mapsin6((struct sockaddr_in *)src, &sin6); 3962 } 3963 if (in6_pcbconnect(in6p, &sin6, NULL)) { 3964 goto resetandabort; 3965 } 3966 } 3967 #endif 3968 else { 3969 goto resetandabort; 3970 } 3971 3972 if (inp) 3973 tp = intotcpcb(inp); 3974 #ifdef INET6 3975 else if (in6p) 3976 tp = in6totcpcb(in6p); 3977 #endif 3978 else 3979 tp = NULL; 3980 3981 tp->t_flags = sototcpcb(oso)->t_flags & TF_NODELAY; 3982 if (sc->sc_request_r_scale != 15) { 3983 tp->requested_s_scale = sc->sc_requested_s_scale; 3984 tp->request_r_scale = sc->sc_request_r_scale; 3985 tp->snd_scale = sc->sc_requested_s_scale; 3986 tp->rcv_scale = sc->sc_request_r_scale; 3987 tp->t_flags |= TF_REQ_SCALE|TF_RCVD_SCALE; 3988 } 3989 if (sc->sc_flags & SCF_TIMESTAMP) 3990 tp->t_flags |= TF_REQ_TSTMP|TF_RCVD_TSTMP; 3991 tp->ts_timebase = sc->sc_timebase; 3992 3993 tp->t_template = tcp_template(tp); 3994 if (tp->t_template == 0) { 3995 tp = tcp_drop(tp, ENOBUFS); /* destroys socket */ 3996 so = NULL; 3997 m_freem(m); 3998 goto abort; 3999 } 4000 4001 tp->iss = sc->sc_iss; 4002 tp->irs = sc->sc_irs; 4003 tcp_sendseqinit(tp); 4004 tcp_rcvseqinit(tp); 4005 tp->t_state = TCPS_SYN_RECEIVED; 4006 TCP_TIMER_ARM(tp, TCPT_KEEP, tp->t_keepinit); 4007 TCP_STATINC(TCP_STAT_ACCEPTS); 4008 4009 if ((sc->sc_flags & SCF_SACK_PERMIT) && tcp_do_sack) 4010 tp->t_flags |= TF_WILL_SACK; 4011 4012 if ((sc->sc_flags & SCF_ECN_PERMIT) && tcp_do_ecn) 4013 tp->t_flags |= TF_ECN_PERMIT; 4014 4015 #ifdef TCP_SIGNATURE 4016 if (sc->sc_flags & SCF_SIGNATURE) 4017 tp->t_flags |= TF_SIGNATURE; 4018 #endif 4019 4020 /* Initialize tp->t_ourmss before we deal with the peer's! */ 4021 tp->t_ourmss = sc->sc_ourmaxseg; 4022 tcp_mss_from_peer(tp, sc->sc_peermaxseg); 4023 4024 /* 4025 * Initialize the initial congestion window. If we 4026 * had to retransmit the SYN,ACK, we must initialize cwnd 4027 * to 1 segment (i.e. the Loss Window). 4028 */ 4029 if (sc->sc_rxtshift) 4030 tp->snd_cwnd = tp->t_peermss; 4031 else { 4032 int ss = tcp_init_win; 4033 if (inp != NULL && in_localaddr(inp->inp_faddr)) 4034 ss = tcp_init_win_local; 4035 #ifdef INET6 4036 if (in6p != NULL && in6_localaddr(&in6p->in6p_faddr)) 4037 ss = tcp_init_win_local; 4038 #endif 4039 tp->snd_cwnd = TCP_INITIAL_WINDOW(ss, tp->t_peermss); 4040 } 4041 4042 tcp_rmx_rtt(tp); 4043 tp->snd_wl1 = sc->sc_irs; 4044 tp->rcv_up = sc->sc_irs + 1; 4045 4046 /* 4047 * This is what whould have happened in tcp_output() when 4048 * the SYN,ACK was sent. 4049 */ 4050 tp->snd_up = tp->snd_una; 4051 tp->snd_max = tp->snd_nxt = tp->iss+1; 4052 TCP_TIMER_ARM(tp, TCPT_REXMT, tp->t_rxtcur); 4053 if (sc->sc_win > 0 && SEQ_GT(tp->rcv_nxt + sc->sc_win, tp->rcv_adv)) 4054 tp->rcv_adv = tp->rcv_nxt + sc->sc_win; 4055 tp->last_ack_sent = tp->rcv_nxt; 4056 tp->t_partialacks = -1; 4057 tp->t_dupacks = 0; 4058 4059 TCP_STATINC(TCP_STAT_SC_COMPLETED); 4060 s = splsoftnet(); 4061 syn_cache_put(sc); 4062 splx(s); 4063 return so; 4064 4065 resetandabort: 4066 (void)tcp_respond(NULL, m, m, th, (tcp_seq)0, th->th_ack, TH_RST); 4067 abort: 4068 if (so != NULL) { 4069 (void) soqremque(so, 1); 4070 (void) soabort(so); 4071 mutex_enter(softnet_lock); 4072 } 4073 s = splsoftnet(); 4074 syn_cache_put(sc); 4075 splx(s); 4076 TCP_STATINC(TCP_STAT_SC_ABORTED); 4077 return ((struct socket *)(-1)); 4078 } 4079 4080 /* 4081 * This function is called when we get a RST for a 4082 * non-existent connection, so that we can see if the 4083 * connection is in the syn cache. If it is, zap it. 4084 */ 4085 4086 void 4087 syn_cache_reset(struct sockaddr *src, struct sockaddr *dst, struct tcphdr *th) 4088 { 4089 struct syn_cache *sc; 4090 struct syn_cache_head *scp; 4091 int s = splsoftnet(); 4092 4093 if ((sc = syn_cache_lookup(src, dst, &scp)) == NULL) { 4094 splx(s); 4095 return; 4096 } 4097 if (SEQ_LT(th->th_seq, sc->sc_irs) || 4098 SEQ_GT(th->th_seq, sc->sc_irs+1)) { 4099 splx(s); 4100 return; 4101 } 4102 syn_cache_rm(sc); 4103 TCP_STATINC(TCP_STAT_SC_RESET); 4104 syn_cache_put(sc); /* calls pool_put but see spl above */ 4105 splx(s); 4106 } 4107 4108 void 4109 syn_cache_unreach(const struct sockaddr *src, const struct sockaddr *dst, 4110 struct tcphdr *th) 4111 { 4112 struct syn_cache *sc; 4113 struct syn_cache_head *scp; 4114 int s; 4115 4116 s = splsoftnet(); 4117 if ((sc = syn_cache_lookup(src, dst, &scp)) == NULL) { 4118 splx(s); 4119 return; 4120 } 4121 /* If the sequence number != sc_iss, then it's a bogus ICMP msg */ 4122 if (ntohl(th->th_seq) != sc->sc_iss) { 4123 splx(s); 4124 return; 4125 } 4126 4127 /* 4128 * If we've retransmitted 3 times and this is our second error, 4129 * we remove the entry. Otherwise, we allow it to continue on. 4130 * This prevents us from incorrectly nuking an entry during a 4131 * spurious network outage. 4132 * 4133 * See tcp_notify(). 4134 */ 4135 if ((sc->sc_flags & SCF_UNREACH) == 0 || sc->sc_rxtshift < 3) { 4136 sc->sc_flags |= SCF_UNREACH; 4137 splx(s); 4138 return; 4139 } 4140 4141 syn_cache_rm(sc); 4142 TCP_STATINC(TCP_STAT_SC_UNREACH); 4143 syn_cache_put(sc); /* calls pool_put but see spl above */ 4144 splx(s); 4145 } 4146 4147 /* 4148 * Given a LISTEN socket and an inbound SYN request, add this to the syn 4149 * cache, and send back a segment: 4150 * <SEQ=ISS><ACK=RCV_NXT><CTL=SYN,ACK> 4151 * to the source. 4152 * 4153 * IMPORTANT NOTE: We do _NOT_ ACK data that might accompany the SYN. 4154 * Doing so would require that we hold onto the data and deliver it 4155 * to the application. However, if we are the target of a SYN-flood 4156 * DoS attack, an attacker could send data which would eventually 4157 * consume all available buffer space if it were ACKed. By not ACKing 4158 * the data, we avoid this DoS scenario. 4159 */ 4160 int 4161 syn_cache_add(struct sockaddr *src, struct sockaddr *dst, struct tcphdr *th, 4162 unsigned int toff, struct socket *so, struct mbuf *m, u_char *optp, 4163 int optlen, struct tcp_opt_info *oi) 4164 { 4165 struct tcpcb tb, *tp; 4166 long win; 4167 struct syn_cache *sc; 4168 struct syn_cache_head *scp; 4169 struct mbuf *ipopts; 4170 int s; 4171 4172 tp = sototcpcb(so); 4173 4174 /* 4175 * Initialize some local state. 4176 */ 4177 win = sbspace(&so->so_rcv); 4178 if (win > TCP_MAXWIN) 4179 win = TCP_MAXWIN; 4180 4181 #ifdef TCP_SIGNATURE 4182 if (optp || (tp->t_flags & TF_SIGNATURE)) 4183 #else 4184 if (optp) 4185 #endif 4186 { 4187 tb.t_flags = tcp_do_rfc1323 ? (TF_REQ_SCALE|TF_REQ_TSTMP) : 0; 4188 #ifdef TCP_SIGNATURE 4189 tb.t_flags |= (tp->t_flags & TF_SIGNATURE); 4190 #endif 4191 tb.t_state = TCPS_LISTEN; 4192 if (tcp_dooptions(&tb, optp, optlen, th, m, toff, oi) < 0) 4193 return 0; 4194 } else 4195 tb.t_flags = 0; 4196 4197 switch (src->sa_family) { 4198 case AF_INET: 4199 /* Remember the IP options, if any. */ 4200 ipopts = ip_srcroute(m); 4201 break; 4202 default: 4203 ipopts = NULL; 4204 } 4205 4206 /* 4207 * See if we already have an entry for this connection. 4208 * If we do, resend the SYN,ACK. We do not count this 4209 * as a retransmission (XXX though maybe we should). 4210 */ 4211 if ((sc = syn_cache_lookup(src, dst, &scp)) != NULL) { 4212 TCP_STATINC(TCP_STAT_SC_DUPESYN); 4213 if (ipopts) { 4214 /* 4215 * If we were remembering a previous source route, 4216 * forget it and use the new one we've been given. 4217 */ 4218 if (sc->sc_ipopts) 4219 (void)m_free(sc->sc_ipopts); 4220 sc->sc_ipopts = ipopts; 4221 } 4222 sc->sc_timestamp = tb.ts_recent; 4223 m_freem(m); 4224 if (syn_cache_respond(sc) == 0) { 4225 uint64_t *tcps = TCP_STAT_GETREF(); 4226 tcps[TCP_STAT_SNDACKS]++; 4227 tcps[TCP_STAT_SNDTOTAL]++; 4228 TCP_STAT_PUTREF(); 4229 } 4230 return 1; 4231 } 4232 4233 s = splsoftnet(); 4234 sc = pool_get(&syn_cache_pool, PR_NOWAIT); 4235 splx(s); 4236 if (sc == NULL) { 4237 if (ipopts) 4238 (void)m_free(ipopts); 4239 return 0; 4240 } 4241 4242 /* 4243 * Fill in the cache, and put the necessary IP and TCP 4244 * options into the reply. 4245 */ 4246 memset(sc, 0, sizeof(struct syn_cache)); 4247 callout_init(&sc->sc_timer, CALLOUT_MPSAFE); 4248 memcpy(&sc->sc_src, src, src->sa_len); 4249 memcpy(&sc->sc_dst, dst, dst->sa_len); 4250 sc->sc_flags = 0; 4251 sc->sc_ipopts = ipopts; 4252 sc->sc_irs = th->th_seq; 4253 switch (src->sa_family) { 4254 case AF_INET: 4255 { 4256 struct sockaddr_in *srcin = (void *)src; 4257 struct sockaddr_in *dstin = (void *)dst; 4258 4259 sc->sc_iss = tcp_new_iss1(&dstin->sin_addr, 4260 &srcin->sin_addr, dstin->sin_port, 4261 srcin->sin_port, sizeof(dstin->sin_addr), 0); 4262 break; 4263 } 4264 #ifdef INET6 4265 case AF_INET6: 4266 { 4267 struct sockaddr_in6 *srcin6 = (void *)src; 4268 struct sockaddr_in6 *dstin6 = (void *)dst; 4269 4270 sc->sc_iss = tcp_new_iss1(&dstin6->sin6_addr, 4271 &srcin6->sin6_addr, dstin6->sin6_port, 4272 srcin6->sin6_port, sizeof(dstin6->sin6_addr), 0); 4273 break; 4274 } 4275 #endif 4276 } 4277 sc->sc_peermaxseg = oi->maxseg; 4278 sc->sc_ourmaxseg = tcp_mss_to_advertise(m->m_flags & M_PKTHDR ? 4279 m_get_rcvif_NOMPSAFE(m) : NULL, sc->sc_src.sa.sa_family); 4280 sc->sc_win = win; 4281 sc->sc_timebase = tcp_now - 1; /* see tcp_newtcpcb() */ 4282 sc->sc_timestamp = tb.ts_recent; 4283 if ((tb.t_flags & (TF_REQ_TSTMP|TF_RCVD_TSTMP)) == 4284 (TF_REQ_TSTMP|TF_RCVD_TSTMP)) 4285 sc->sc_flags |= SCF_TIMESTAMP; 4286 if ((tb.t_flags & (TF_RCVD_SCALE|TF_REQ_SCALE)) == 4287 (TF_RCVD_SCALE|TF_REQ_SCALE)) { 4288 sc->sc_requested_s_scale = tb.requested_s_scale; 4289 sc->sc_request_r_scale = 0; 4290 /* 4291 * Pick the smallest possible scaling factor that 4292 * will still allow us to scale up to sb_max. 4293 * 4294 * We do this because there are broken firewalls that 4295 * will corrupt the window scale option, leading to 4296 * the other endpoint believing that our advertised 4297 * window is unscaled. At scale factors larger than 4298 * 5 the unscaled window will drop below 1500 bytes, 4299 * leading to serious problems when traversing these 4300 * broken firewalls. 4301 * 4302 * With the default sbmax of 256K, a scale factor 4303 * of 3 will be chosen by this algorithm. Those who 4304 * choose a larger sbmax should watch out 4305 * for the compatiblity problems mentioned above. 4306 * 4307 * RFC1323: The Window field in a SYN (i.e., a <SYN> 4308 * or <SYN,ACK>) segment itself is never scaled. 4309 */ 4310 while (sc->sc_request_r_scale < TCP_MAX_WINSHIFT && 4311 (TCP_MAXWIN << sc->sc_request_r_scale) < sb_max) 4312 sc->sc_request_r_scale++; 4313 } else { 4314 sc->sc_requested_s_scale = 15; 4315 sc->sc_request_r_scale = 15; 4316 } 4317 if ((tb.t_flags & TF_SACK_PERMIT) && tcp_do_sack) 4318 sc->sc_flags |= SCF_SACK_PERMIT; 4319 4320 /* 4321 * ECN setup packet received. 4322 */ 4323 if ((th->th_flags & (TH_ECE|TH_CWR)) && tcp_do_ecn) 4324 sc->sc_flags |= SCF_ECN_PERMIT; 4325 4326 #ifdef TCP_SIGNATURE 4327 if (tb.t_flags & TF_SIGNATURE) 4328 sc->sc_flags |= SCF_SIGNATURE; 4329 #endif 4330 sc->sc_tp = tp; 4331 m_freem(m); 4332 if (syn_cache_respond(sc) == 0) { 4333 uint64_t *tcps = TCP_STAT_GETREF(); 4334 tcps[TCP_STAT_SNDACKS]++; 4335 tcps[TCP_STAT_SNDTOTAL]++; 4336 TCP_STAT_PUTREF(); 4337 syn_cache_insert(sc, tp); 4338 } else { 4339 s = splsoftnet(); 4340 /* 4341 * syn_cache_put() will try to schedule the timer, so 4342 * we need to initialize it 4343 */ 4344 syn_cache_timer_arm(sc); 4345 syn_cache_put(sc); 4346 splx(s); 4347 TCP_STATINC(TCP_STAT_SC_DROPPED); 4348 } 4349 return 1; 4350 } 4351 4352 /* 4353 * syn_cache_respond: (re)send SYN+ACK. 4354 * 4355 * Returns 0 on success. 4356 */ 4357 4358 int 4359 syn_cache_respond(struct syn_cache *sc) 4360 { 4361 #ifdef INET6 4362 struct rtentry *rt = NULL; 4363 #endif 4364 struct route *ro; 4365 u_int8_t *optp; 4366 int optlen, error; 4367 u_int16_t tlen; 4368 struct ip *ip = NULL; 4369 #ifdef INET6 4370 struct ip6_hdr *ip6 = NULL; 4371 #endif 4372 struct tcpcb *tp; 4373 struct tcphdr *th; 4374 struct mbuf *m; 4375 u_int hlen; 4376 #ifdef TCP_SIGNATURE 4377 struct secasvar *sav = NULL; 4378 u_int8_t *sigp = NULL; 4379 #endif 4380 4381 ro = &sc->sc_route; 4382 switch (sc->sc_src.sa.sa_family) { 4383 case AF_INET: 4384 hlen = sizeof(struct ip); 4385 break; 4386 #ifdef INET6 4387 case AF_INET6: 4388 hlen = sizeof(struct ip6_hdr); 4389 break; 4390 #endif 4391 default: 4392 return EAFNOSUPPORT; 4393 } 4394 4395 /* Worst case scanario, since we don't know the option size yet. */ 4396 tlen = hlen + sizeof(struct tcphdr) + MAX_TCPOPTLEN; 4397 KASSERT(max_linkhdr + tlen <= MCLBYTES); 4398 4399 /* 4400 * Create the IP+TCP header from scratch. 4401 */ 4402 MGETHDR(m, M_DONTWAIT, MT_DATA); 4403 if (m && (max_linkhdr + tlen) > MHLEN) { 4404 MCLGET(m, M_DONTWAIT); 4405 if ((m->m_flags & M_EXT) == 0) { 4406 m_freem(m); 4407 m = NULL; 4408 } 4409 } 4410 if (m == NULL) 4411 return ENOBUFS; 4412 MCLAIM(m, &tcp_tx_mowner); 4413 4414 tp = sc->sc_tp; 4415 4416 /* Fixup the mbuf. */ 4417 m->m_data += max_linkhdr; 4418 m_reset_rcvif(m); 4419 memset(mtod(m, void *), 0, tlen); 4420 4421 switch (sc->sc_src.sa.sa_family) { 4422 case AF_INET: 4423 ip = mtod(m, struct ip *); 4424 ip->ip_v = 4; 4425 ip->ip_dst = sc->sc_src.sin.sin_addr; 4426 ip->ip_src = sc->sc_dst.sin.sin_addr; 4427 ip->ip_p = IPPROTO_TCP; 4428 th = (struct tcphdr *)(ip + 1); 4429 th->th_dport = sc->sc_src.sin.sin_port; 4430 th->th_sport = sc->sc_dst.sin.sin_port; 4431 break; 4432 #ifdef INET6 4433 case AF_INET6: 4434 ip6 = mtod(m, struct ip6_hdr *); 4435 ip6->ip6_vfc = IPV6_VERSION; 4436 ip6->ip6_dst = sc->sc_src.sin6.sin6_addr; 4437 ip6->ip6_src = sc->sc_dst.sin6.sin6_addr; 4438 ip6->ip6_nxt = IPPROTO_TCP; 4439 /* ip6_plen will be updated in ip6_output() */ 4440 th = (struct tcphdr *)(ip6 + 1); 4441 th->th_dport = sc->sc_src.sin6.sin6_port; 4442 th->th_sport = sc->sc_dst.sin6.sin6_port; 4443 break; 4444 #endif 4445 default: 4446 panic("%s: impossible (1)", __func__); 4447 } 4448 4449 th->th_seq = htonl(sc->sc_iss); 4450 th->th_ack = htonl(sc->sc_irs + 1); 4451 th->th_flags = TH_SYN|TH_ACK; 4452 th->th_win = htons(sc->sc_win); 4453 /* th_x2, th_sum, th_urp already 0 from memset */ 4454 4455 /* Tack on the TCP options. */ 4456 optp = (u_int8_t *)(th + 1); 4457 optlen = 0; 4458 *optp++ = TCPOPT_MAXSEG; 4459 *optp++ = TCPOLEN_MAXSEG; 4460 *optp++ = (sc->sc_ourmaxseg >> 8) & 0xff; 4461 *optp++ = sc->sc_ourmaxseg & 0xff; 4462 optlen += TCPOLEN_MAXSEG; 4463 4464 if (sc->sc_request_r_scale != 15) { 4465 *((u_int32_t *)optp) = htonl(TCPOPT_NOP << 24 | 4466 TCPOPT_WINDOW << 16 | TCPOLEN_WINDOW << 8 | 4467 sc->sc_request_r_scale); 4468 optp += TCPOLEN_WINDOW + TCPOLEN_NOP; 4469 optlen += TCPOLEN_WINDOW + TCPOLEN_NOP; 4470 } 4471 4472 if (sc->sc_flags & SCF_SACK_PERMIT) { 4473 /* Let the peer know that we will SACK. */ 4474 *optp++ = TCPOPT_SACK_PERMITTED; 4475 *optp++ = TCPOLEN_SACK_PERMITTED; 4476 optlen += TCPOLEN_SACK_PERMITTED; 4477 } 4478 4479 if (sc->sc_flags & SCF_TIMESTAMP) { 4480 while (optlen % 4 != 2) { 4481 optlen += TCPOLEN_NOP; 4482 *optp++ = TCPOPT_NOP; 4483 } 4484 *optp++ = TCPOPT_TIMESTAMP; 4485 *optp++ = TCPOLEN_TIMESTAMP; 4486 u_int32_t *lp = (u_int32_t *)(optp); 4487 /* Form timestamp option as shown in appendix A of RFC 1323. */ 4488 *lp++ = htonl(SYN_CACHE_TIMESTAMP(sc)); 4489 *lp = htonl(sc->sc_timestamp); 4490 optp += TCPOLEN_TIMESTAMP - 2; 4491 optlen += TCPOLEN_TIMESTAMP; 4492 } 4493 4494 #ifdef TCP_SIGNATURE 4495 if (sc->sc_flags & SCF_SIGNATURE) { 4496 sav = tcp_signature_getsav(m); 4497 if (sav == NULL) { 4498 m_freem(m); 4499 return EPERM; 4500 } 4501 4502 *optp++ = TCPOPT_SIGNATURE; 4503 *optp++ = TCPOLEN_SIGNATURE; 4504 sigp = optp; 4505 memset(optp, 0, TCP_SIGLEN); 4506 optp += TCP_SIGLEN; 4507 optlen += TCPOLEN_SIGNATURE; 4508 } 4509 #endif 4510 4511 /* 4512 * Terminate and pad TCP options to a 4 byte boundary. 4513 * 4514 * According to RFC793: "The content of the header beyond the 4515 * End-of-Option option must be header padding (i.e., zero)." 4516 * And later: "The padding is composed of zeros." 4517 */ 4518 if (optlen % 4) { 4519 optlen += TCPOLEN_EOL; 4520 *optp++ = TCPOPT_EOL; 4521 } 4522 while (optlen % 4) { 4523 optlen += TCPOLEN_PAD; 4524 *optp++ = TCPOPT_PAD; 4525 } 4526 4527 /* Compute the actual values now that we've added the options. */ 4528 tlen = hlen + sizeof(struct tcphdr) + optlen; 4529 m->m_len = m->m_pkthdr.len = tlen; 4530 th->th_off = (sizeof(struct tcphdr) + optlen) >> 2; 4531 4532 #ifdef TCP_SIGNATURE 4533 if (sav) { 4534 (void)tcp_signature(m, th, hlen, sav, sigp); 4535 key_sa_recordxfer(sav, m); 4536 KEY_SA_UNREF(&sav); 4537 } 4538 #endif 4539 4540 /* 4541 * Send ECN SYN-ACK setup packet. 4542 * Routes can be asymetric, so, even if we receive a packet 4543 * with ECE and CWR set, we must not assume no one will block 4544 * the ECE packet we are about to send. 4545 */ 4546 if ((sc->sc_flags & SCF_ECN_PERMIT) && tp && 4547 SEQ_GEQ(tp->snd_nxt, tp->snd_max)) { 4548 th->th_flags |= TH_ECE; 4549 TCP_STATINC(TCP_STAT_ECN_SHS); 4550 4551 /* 4552 * draft-ietf-tcpm-ecnsyn-00.txt 4553 * 4554 * "[...] a TCP node MAY respond to an ECN-setup 4555 * SYN packet by setting ECT in the responding 4556 * ECN-setup SYN/ACK packet, indicating to routers 4557 * that the SYN/ACK packet is ECN-Capable. 4558 * This allows a congested router along the path 4559 * to mark the packet instead of dropping the 4560 * packet as an indication of congestion." 4561 * 4562 * "[...] There can be a great benefit in setting 4563 * an ECN-capable codepoint in SYN/ACK packets [...] 4564 * Congestion is most likely to occur in 4565 * the server-to-client direction. As a result, 4566 * setting an ECN-capable codepoint in SYN/ACK 4567 * packets can reduce the occurence of three-second 4568 * retransmit timeouts resulting from the drop 4569 * of SYN/ACK packets." 4570 * 4571 * Page 4 and 6, January 2006. 4572 */ 4573 4574 switch (sc->sc_src.sa.sa_family) { 4575 case AF_INET: 4576 ip->ip_tos |= IPTOS_ECN_ECT0; 4577 break; 4578 #ifdef INET6 4579 case AF_INET6: 4580 ip6->ip6_flow |= htonl(IPTOS_ECN_ECT0 << 20); 4581 break; 4582 #endif 4583 } 4584 TCP_STATINC(TCP_STAT_ECN_ECT); 4585 } 4586 4587 4588 /* 4589 * Compute the packet's checksum. 4590 * 4591 * Fill in some straggling IP bits. Note the stack expects 4592 * ip_len to be in host order, for convenience. 4593 */ 4594 switch (sc->sc_src.sa.sa_family) { 4595 case AF_INET: 4596 ip->ip_len = htons(tlen - hlen); 4597 th->th_sum = 0; 4598 th->th_sum = in4_cksum(m, IPPROTO_TCP, hlen, tlen - hlen); 4599 ip->ip_len = htons(tlen); 4600 ip->ip_ttl = ip_defttl; 4601 /* XXX tos? */ 4602 break; 4603 #ifdef INET6 4604 case AF_INET6: 4605 ip6->ip6_plen = htons(tlen - hlen); 4606 th->th_sum = 0; 4607 th->th_sum = in6_cksum(m, IPPROTO_TCP, hlen, tlen - hlen); 4608 ip6->ip6_vfc &= ~IPV6_VERSION_MASK; 4609 ip6->ip6_vfc |= IPV6_VERSION; 4610 ip6->ip6_plen = htons(tlen - hlen); 4611 /* ip6_hlim will be initialized afterwards */ 4612 /* XXX flowlabel? */ 4613 break; 4614 #endif 4615 } 4616 4617 /* XXX use IPsec policy on listening socket, on SYN ACK */ 4618 tp = sc->sc_tp; 4619 4620 switch (sc->sc_src.sa.sa_family) { 4621 case AF_INET: 4622 error = ip_output(m, sc->sc_ipopts, ro, 4623 (ip_mtudisc ? IP_MTUDISC : 0), 4624 NULL, tp ? tp->t_inpcb : NULL); 4625 break; 4626 #ifdef INET6 4627 case AF_INET6: 4628 ip6->ip6_hlim = in6_selecthlim(NULL, 4629 (rt = rtcache_validate(ro)) != NULL ? rt->rt_ifp : NULL); 4630 rtcache_unref(rt, ro); 4631 4632 error = ip6_output(m, NULL /*XXX*/, ro, 0, NULL, 4633 tp ? tp->t_in6pcb : NULL, NULL); 4634 break; 4635 #endif 4636 default: 4637 panic("%s: impossible (2)", __func__); 4638 } 4639 4640 return error; 4641 } 4642
Indexes created Sun Sep 21 19:09:51 GMT 2025