1 /* $NetBSD: tcp_input.c,v 1.441 2024/10/08 06:17:14 rin 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 #include <sys/cdefs.h> 141 __KERNEL_RCSID(0, "$NetBSD: tcp_input.c,v 1.441 2024/10/08 06:17:14 rin Exp $"); 142 143 #ifdef _KERNEL_OPT 144 #include "opt_inet.h" 145 #include "opt_ipsec.h" 146 #include "opt_inet_csum.h" 147 #include "opt_tcp_debug.h" 148 #endif 149 150 #include <sys/param.h> 151 #include <sys/systm.h> 152 #include <sys/malloc.h> 153 #include <sys/mbuf.h> 154 #include <sys/protosw.h> 155 #include <sys/socket.h> 156 #include <sys/socketvar.h> 157 #include <sys/errno.h> 158 #include <sys/syslog.h> 159 #include <sys/pool.h> 160 #include <sys/domain.h> 161 #include <sys/kernel.h> 162 #ifdef TCP_SIGNATURE 163 #include <sys/md5.h> 164 #endif 165 #include <sys/lwp.h> /* for lwp0 */ 166 #include <sys/cprng.h> 167 168 #include <net/if.h> 169 #include <net/if_types.h> 170 171 #include <netinet/in.h> 172 #include <netinet/in_systm.h> 173 #include <netinet/ip.h> 174 #include <netinet/in_pcb.h> 175 #include <netinet/in_var.h> 176 #include <netinet/ip_var.h> 177 #include <netinet/in_offload.h> 178 179 #if NARP > 0 180 #include <netinet/if_inarp.h> 181 #endif 182 #ifdef INET6 183 #include <netinet/ip6.h> 184 #include <netinet6/ip6_var.h> 185 #include <netinet6/in6_pcb.h> 186 #include <netinet6/ip6_var.h> 187 #include <netinet6/in6_var.h> 188 #include <netinet/icmp6.h> 189 #include <netinet6/nd6.h> 190 #ifdef TCP_SIGNATURE 191 #include <netinet6/scope6_var.h> 192 #endif 193 #endif 194 195 #ifndef INET6 196 #include <netinet/ip6.h> 197 #endif 198 199 #include <netinet/tcp.h> 200 #include <netinet/tcp_fsm.h> 201 #include <netinet/tcp_seq.h> 202 #include <netinet/tcp_timer.h> 203 #include <netinet/tcp_var.h> 204 #include <netinet/tcp_private.h> 205 #include <netinet/tcp_congctl.h> 206 #include <netinet/tcp_debug.h> 207 #include <netinet/tcp_syncache.h> 208 209 #ifdef INET6 210 #include "faith.h" 211 #if defined(NFAITH) && NFAITH > 0 212 #include <net/if_faith.h> 213 #endif 214 #endif 215 216 #ifdef IPSEC 217 #include <netipsec/ipsec.h> 218 #include <netipsec/key.h> 219 #ifdef INET6 220 #include <netipsec/ipsec6.h> 221 #endif 222 #endif /* IPSEC*/ 223 224 #include <netinet/tcp_vtw.h> 225 226 int tcprexmtthresh = 3; 227 int tcp_log_refused; 228 229 int tcp_do_autorcvbuf = 1; 230 int tcp_autorcvbuf_inc = 16 * 1024; 231 int tcp_autorcvbuf_max = 256 * 1024; 232 int tcp_msl = (TCPTV_MSL / PR_SLOWHZ); 233 234 int tcp_reass_maxqueuelen = 100; 235 236 static int tcp_rst_ppslim_count = 0; 237 static struct timeval tcp_rst_ppslim_last; 238 static int tcp_ackdrop_ppslim_count = 0; 239 static struct timeval tcp_ackdrop_ppslim_last; 240 241 #define TCP_PAWS_IDLE (24U * 24 * 60 * 60 * PR_SLOWHZ) 242 243 /* for modulo comparisons of timestamps */ 244 #define TSTMP_LT(a,b) ((int)((a)-(b)) < 0) 245 #define TSTMP_GEQ(a,b) ((int)((a)-(b)) >= 0) 246 247 /* 248 * Neighbor Discovery, Neighbor Unreachability Detection Upper layer hint. 249 */ 250 static void 251 nd_hint(struct tcpcb *tp) 252 { 253 struct route *ro = NULL; 254 struct rtentry *rt; 255 256 if (tp == NULL) 257 return; 258 259 ro = &tp->t_inpcb->inp_route; 260 if (ro == NULL) 261 return; 262 263 rt = rtcache_validate(ro); 264 if (rt == NULL) 265 return; 266 267 switch (tp->t_family) { 268 #if NARP > 0 269 case AF_INET: 270 arp_nud_hint(rt); 271 break; 272 #endif 273 #ifdef INET6 274 case AF_INET6: 275 nd6_nud_hint(rt); 276 break; 277 #endif 278 } 279 280 rtcache_unref(rt, ro); 281 } 282 283 /* 284 * Compute ACK transmission behavior. Delay the ACK unless 285 * we have already delayed an ACK (must send an ACK every two segments). 286 * We also ACK immediately if we received a PUSH and the ACK-on-PUSH 287 * option is enabled. 288 */ 289 static void 290 tcp_setup_ack(struct tcpcb *tp, const struct tcphdr *th) 291 { 292 293 if (tp->t_flags & TF_DELACK || 294 (tcp_ack_on_push && th->th_flags & TH_PUSH)) 295 tp->t_flags |= TF_ACKNOW; 296 else 297 TCP_SET_DELACK(tp); 298 } 299 300 static void 301 icmp_check(struct tcpcb *tp, const struct tcphdr *th, int acked) 302 { 303 304 /* 305 * If we had a pending ICMP message that refers to data that have 306 * just been acknowledged, disregard the recorded ICMP message. 307 */ 308 if ((tp->t_flags & TF_PMTUD_PEND) && 309 SEQ_GT(th->th_ack, tp->t_pmtud_th_seq)) 310 tp->t_flags &= ~TF_PMTUD_PEND; 311 312 /* 313 * Keep track of the largest chunk of data 314 * acknowledged since last PMTU update 315 */ 316 if (tp->t_pmtud_mss_acked < acked) 317 tp->t_pmtud_mss_acked = acked; 318 } 319 320 /* 321 * Convert TCP protocol fields to host order for easier processing. 322 */ 323 static void 324 tcp_fields_to_host(struct tcphdr *th) 325 { 326 327 NTOHL(th->th_seq); 328 NTOHL(th->th_ack); 329 NTOHS(th->th_win); 330 NTOHS(th->th_urp); 331 } 332 333 /* 334 * ... and reverse the above. 335 */ 336 static void 337 tcp_fields_to_net(struct tcphdr *th) 338 { 339 340 HTONL(th->th_seq); 341 HTONL(th->th_ack); 342 HTONS(th->th_win); 343 HTONS(th->th_urp); 344 } 345 346 static void 347 tcp_urp_drop(struct tcphdr *th, int todrop, int *tiflags) 348 { 349 if (th->th_urp > todrop) { 350 th->th_urp -= todrop; 351 } else { 352 *tiflags &= ~TH_URG; 353 th->th_urp = 0; 354 } 355 } 356 357 #ifdef TCP_CSUM_COUNTERS 358 #include <sys/device.h> 359 360 extern struct evcnt tcp_hwcsum_ok; 361 extern struct evcnt tcp_hwcsum_bad; 362 extern struct evcnt tcp_hwcsum_data; 363 extern struct evcnt tcp_swcsum; 364 #if defined(INET6) 365 extern struct evcnt tcp6_hwcsum_ok; 366 extern struct evcnt tcp6_hwcsum_bad; 367 extern struct evcnt tcp6_hwcsum_data; 368 extern struct evcnt tcp6_swcsum; 369 #endif /* defined(INET6) */ 370 371 #define TCP_CSUM_COUNTER_INCR(ev) (ev)->ev_count++ 372 373 #else 374 375 #define TCP_CSUM_COUNTER_INCR(ev) /* nothing */ 376 377 #endif /* TCP_CSUM_COUNTERS */ 378 379 #ifdef TCP_REASS_COUNTERS 380 #include <sys/device.h> 381 382 extern struct evcnt tcp_reass_; 383 extern struct evcnt tcp_reass_empty; 384 extern struct evcnt tcp_reass_iteration[8]; 385 extern struct evcnt tcp_reass_prependfirst; 386 extern struct evcnt tcp_reass_prepend; 387 extern struct evcnt tcp_reass_insert; 388 extern struct evcnt tcp_reass_inserttail; 389 extern struct evcnt tcp_reass_append; 390 extern struct evcnt tcp_reass_appendtail; 391 extern struct evcnt tcp_reass_overlaptail; 392 extern struct evcnt tcp_reass_overlapfront; 393 extern struct evcnt tcp_reass_segdup; 394 extern struct evcnt tcp_reass_fragdup; 395 396 #define TCP_REASS_COUNTER_INCR(ev) (ev)->ev_count++ 397 398 #else 399 400 #define TCP_REASS_COUNTER_INCR(ev) /* nothing */ 401 402 #endif /* TCP_REASS_COUNTERS */ 403 404 static int tcp_reass(struct tcpcb *, const struct tcphdr *, struct mbuf *, 405 int); 406 407 static void tcp4_log_refused(const struct ip *, const struct tcphdr *); 408 #ifdef INET6 409 static void tcp6_log_refused(const struct ip6_hdr *, const struct tcphdr *); 410 #endif 411 412 #if defined(MBUFTRACE) 413 struct mowner tcp_reass_mowner = MOWNER_INIT("tcp", "reass"); 414 #endif /* defined(MBUFTRACE) */ 415 416 static struct pool tcpipqent_pool; 417 418 void 419 tcpipqent_init(void) 420 { 421 422 pool_init(&tcpipqent_pool, sizeof(struct ipqent), 0, 0, 0, "tcpipqepl", 423 NULL, IPL_VM); 424 } 425 426 struct ipqent * 427 tcpipqent_alloc(void) 428 { 429 struct ipqent *ipqe; 430 int s; 431 432 s = splvm(); 433 ipqe = pool_get(&tcpipqent_pool, PR_NOWAIT); 434 splx(s); 435 436 return ipqe; 437 } 438 439 void 440 tcpipqent_free(struct ipqent *ipqe) 441 { 442 int s; 443 444 s = splvm(); 445 pool_put(&tcpipqent_pool, ipqe); 446 splx(s); 447 } 448 449 /* 450 * Insert segment ti into reassembly queue of tcp with 451 * control block tp. Return TH_FIN if reassembly now includes 452 * a segment with FIN. 453 */ 454 static int 455 tcp_reass(struct tcpcb *tp, const struct tcphdr *th, struct mbuf *m, int tlen) 456 { 457 struct ipqent *p, *q, *nq, *tiqe = NULL; 458 struct socket *so = NULL; 459 int pkt_flags; 460 tcp_seq pkt_seq; 461 unsigned pkt_len; 462 u_long rcvpartdupbyte = 0; 463 u_long rcvoobyte; 464 #ifdef TCP_REASS_COUNTERS 465 u_int count = 0; 466 #endif 467 net_stat_ref_t tcps; 468 469 so = tp->t_inpcb->inp_socket; 470 471 TCP_REASS_LOCK_CHECK(tp); 472 473 /* 474 * Call with th==NULL after become established to 475 * force pre-ESTABLISHED data up to user socket. 476 */ 477 if (th == NULL) 478 goto present; 479 480 m_claimm(m, &tcp_reass_mowner); 481 482 rcvoobyte = tlen; 483 /* 484 * Copy these to local variables because the TCP header gets munged 485 * while we are collapsing mbufs. 486 */ 487 pkt_seq = th->th_seq; 488 pkt_len = tlen; 489 pkt_flags = th->th_flags; 490 491 TCP_REASS_COUNTER_INCR(&tcp_reass_); 492 493 if ((p = TAILQ_LAST(&tp->segq, ipqehead)) != NULL) { 494 /* 495 * When we miss a packet, the vast majority of time we get 496 * packets that follow it in order. So optimize for that. 497 */ 498 if (pkt_seq == p->ipqe_seq + p->ipqe_len) { 499 p->ipqe_len += pkt_len; 500 p->ipqe_flags |= pkt_flags; 501 m_cat(p->ipqe_m, m); 502 m = NULL; 503 tiqe = p; 504 TAILQ_REMOVE(&tp->timeq, p, ipqe_timeq); 505 TCP_REASS_COUNTER_INCR(&tcp_reass_appendtail); 506 goto skip_replacement; 507 } 508 /* 509 * While we're here, if the pkt is completely beyond 510 * anything we have, just insert it at the tail. 511 */ 512 if (SEQ_GT(pkt_seq, p->ipqe_seq + p->ipqe_len)) { 513 TCP_REASS_COUNTER_INCR(&tcp_reass_inserttail); 514 goto insert_it; 515 } 516 } 517 518 q = TAILQ_FIRST(&tp->segq); 519 520 if (q != NULL) { 521 /* 522 * If this segment immediately precedes the first out-of-order 523 * block, simply slap the segment in front of it and (mostly) 524 * skip the complicated logic. 525 */ 526 if (pkt_seq + pkt_len == q->ipqe_seq) { 527 q->ipqe_seq = pkt_seq; 528 q->ipqe_len += pkt_len; 529 q->ipqe_flags |= pkt_flags; 530 m_cat(m, q->ipqe_m); 531 q->ipqe_m = m; 532 tiqe = q; 533 TAILQ_REMOVE(&tp->timeq, q, ipqe_timeq); 534 TCP_REASS_COUNTER_INCR(&tcp_reass_prependfirst); 535 goto skip_replacement; 536 } 537 } else { 538 TCP_REASS_COUNTER_INCR(&tcp_reass_empty); 539 } 540 541 /* 542 * Find a segment which begins after this one does. 543 */ 544 for (p = NULL; q != NULL; q = nq) { 545 nq = TAILQ_NEXT(q, ipqe_q); 546 #ifdef TCP_REASS_COUNTERS 547 count++; 548 #endif 549 550 /* 551 * If the received segment is just right after this 552 * fragment, merge the two together and then check 553 * for further overlaps. 554 */ 555 if (q->ipqe_seq + q->ipqe_len == pkt_seq) { 556 pkt_len += q->ipqe_len; 557 pkt_flags |= q->ipqe_flags; 558 pkt_seq = q->ipqe_seq; 559 m_cat(q->ipqe_m, m); 560 m = q->ipqe_m; 561 TCP_REASS_COUNTER_INCR(&tcp_reass_append); 562 goto free_ipqe; 563 } 564 565 /* 566 * If the received segment is completely past this 567 * fragment, we need to go to the next fragment. 568 */ 569 if (SEQ_LT(q->ipqe_seq + q->ipqe_len, pkt_seq)) { 570 p = q; 571 continue; 572 } 573 574 /* 575 * If the fragment is past the received segment, 576 * it (or any following) can't be concatenated. 577 */ 578 if (SEQ_GT(q->ipqe_seq, pkt_seq + pkt_len)) { 579 TCP_REASS_COUNTER_INCR(&tcp_reass_insert); 580 break; 581 } 582 583 /* 584 * We've received all the data in this segment before. 585 * Mark it as a duplicate and return. 586 */ 587 if (SEQ_LEQ(q->ipqe_seq, pkt_seq) && 588 SEQ_GEQ(q->ipqe_seq + q->ipqe_len, pkt_seq + pkt_len)) { 589 tcps = TCP_STAT_GETREF(); 590 _NET_STATINC_REF(tcps, TCP_STAT_RCVDUPPACK); 591 _NET_STATADD_REF(tcps, TCP_STAT_RCVDUPBYTE, pkt_len); 592 TCP_STAT_PUTREF(); 593 tcp_new_dsack(tp, pkt_seq, pkt_len); 594 m_freem(m); 595 if (tiqe != NULL) { 596 tcpipqent_free(tiqe); 597 } 598 TCP_REASS_COUNTER_INCR(&tcp_reass_segdup); 599 goto out; 600 } 601 602 /* 603 * Received segment completely overlaps this fragment 604 * so we drop the fragment (this keeps the temporal 605 * ordering of segments correct). 606 */ 607 if (SEQ_GEQ(q->ipqe_seq, pkt_seq) && 608 SEQ_LEQ(q->ipqe_seq + q->ipqe_len, pkt_seq + pkt_len)) { 609 rcvpartdupbyte += q->ipqe_len; 610 m_freem(q->ipqe_m); 611 TCP_REASS_COUNTER_INCR(&tcp_reass_fragdup); 612 goto free_ipqe; 613 } 614 615 /* 616 * Received segment extends past the end of the fragment. 617 * Drop the overlapping bytes, merge the fragment and 618 * segment, and treat as a longer received packet. 619 */ 620 if (SEQ_LT(q->ipqe_seq, pkt_seq) && 621 SEQ_GT(q->ipqe_seq + q->ipqe_len, pkt_seq)) { 622 int overlap = q->ipqe_seq + q->ipqe_len - pkt_seq; 623 m_adj(m, overlap); 624 rcvpartdupbyte += overlap; 625 m_cat(q->ipqe_m, m); 626 m = q->ipqe_m; 627 pkt_seq = q->ipqe_seq; 628 pkt_len += q->ipqe_len - overlap; 629 rcvoobyte -= overlap; 630 TCP_REASS_COUNTER_INCR(&tcp_reass_overlaptail); 631 goto free_ipqe; 632 } 633 634 /* 635 * Received segment extends past the front of the fragment. 636 * Drop the overlapping bytes on the received packet. The 637 * packet will then be concatenated with this fragment a 638 * bit later. 639 */ 640 if (SEQ_GT(q->ipqe_seq, pkt_seq) && 641 SEQ_LT(q->ipqe_seq, pkt_seq + pkt_len)) { 642 int overlap = pkt_seq + pkt_len - q->ipqe_seq; 643 m_adj(m, -overlap); 644 pkt_len -= overlap; 645 rcvpartdupbyte += overlap; 646 TCP_REASS_COUNTER_INCR(&tcp_reass_overlapfront); 647 rcvoobyte -= overlap; 648 } 649 650 /* 651 * If the received segment immediately precedes this 652 * fragment then tack the fragment onto this segment 653 * and reinsert the data. 654 */ 655 if (q->ipqe_seq == pkt_seq + pkt_len) { 656 pkt_len += q->ipqe_len; 657 pkt_flags |= q->ipqe_flags; 658 m_cat(m, q->ipqe_m); 659 TAILQ_REMOVE(&tp->segq, q, ipqe_q); 660 TAILQ_REMOVE(&tp->timeq, q, ipqe_timeq); 661 tp->t_segqlen--; 662 KASSERT(tp->t_segqlen >= 0); 663 KASSERT(tp->t_segqlen != 0 || 664 (TAILQ_EMPTY(&tp->segq) && 665 TAILQ_EMPTY(&tp->timeq))); 666 if (tiqe == NULL) { 667 tiqe = q; 668 } else { 669 tcpipqent_free(q); 670 } 671 TCP_REASS_COUNTER_INCR(&tcp_reass_prepend); 672 break; 673 } 674 675 /* 676 * If the fragment is before the segment, remember it. 677 * When this loop is terminated, p will contain the 678 * pointer to the fragment that is right before the 679 * received segment. 680 */ 681 if (SEQ_LEQ(q->ipqe_seq, pkt_seq)) 682 p = q; 683 684 continue; 685 686 /* 687 * This is a common operation. It also will allow 688 * to save doing a malloc/free in most instances. 689 */ 690 free_ipqe: 691 TAILQ_REMOVE(&tp->segq, q, ipqe_q); 692 TAILQ_REMOVE(&tp->timeq, q, ipqe_timeq); 693 tp->t_segqlen--; 694 KASSERT(tp->t_segqlen >= 0); 695 KASSERT(tp->t_segqlen != 0 || 696 (TAILQ_EMPTY(&tp->segq) && TAILQ_EMPTY(&tp->timeq))); 697 if (tiqe == NULL) { 698 tiqe = q; 699 } else { 700 tcpipqent_free(q); 701 } 702 } 703 704 #ifdef TCP_REASS_COUNTERS 705 if (count > 7) 706 TCP_REASS_COUNTER_INCR(&tcp_reass_iteration[0]); 707 else if (count > 0) 708 TCP_REASS_COUNTER_INCR(&tcp_reass_iteration[count]); 709 #endif 710 711 insert_it: 712 /* limit tcp segments per reassembly queue */ 713 if (tp->t_segqlen > tcp_reass_maxqueuelen) { 714 TCP_STATINC(TCP_STAT_RCVMEMDROP); 715 m_freem(m); 716 goto out; 717 } 718 719 /* 720 * Allocate a new queue entry (block) since the received segment 721 * did not collapse onto any other out-of-order block. If it had 722 * collapsed, tiqe would not be NULL and we would be reusing it. 723 * 724 * If the allocation fails, drop the packet. 725 */ 726 if (tiqe == NULL) { 727 tiqe = tcpipqent_alloc(); 728 if (tiqe == NULL) { 729 TCP_STATINC(TCP_STAT_RCVMEMDROP); 730 m_freem(m); 731 goto out; 732 } 733 } 734 735 /* 736 * Update the counters. 737 */ 738 tp->t_rcvoopack++; 739 tcps = TCP_STAT_GETREF(); 740 _NET_STATINC_REF(tcps, TCP_STAT_RCVOOPACK); 741 _NET_STATADD_REF(tcps, TCP_STAT_RCVOOBYTE, rcvoobyte); 742 if (rcvpartdupbyte) { 743 _NET_STATINC_REF(tcps, TCP_STAT_RCVPARTDUPPACK); 744 _NET_STATADD_REF(tcps, TCP_STAT_RCVPARTDUPBYTE, 745 rcvpartdupbyte); 746 } 747 TCP_STAT_PUTREF(); 748 749 /* 750 * Insert the new fragment queue entry into both queues. 751 */ 752 tiqe->ipqe_m = m; 753 tiqe->ipqe_seq = pkt_seq; 754 tiqe->ipqe_len = pkt_len; 755 tiqe->ipqe_flags = pkt_flags; 756 if (p == NULL) { 757 TAILQ_INSERT_HEAD(&tp->segq, tiqe, ipqe_q); 758 } else { 759 TAILQ_INSERT_AFTER(&tp->segq, p, tiqe, ipqe_q); 760 } 761 tp->t_segqlen++; 762 763 skip_replacement: 764 TAILQ_INSERT_HEAD(&tp->timeq, tiqe, ipqe_timeq); 765 766 present: 767 /* 768 * Present data to user, advancing rcv_nxt through 769 * completed sequence space. 770 */ 771 if (TCPS_HAVEESTABLISHED(tp->t_state) == 0) 772 goto out; 773 q = TAILQ_FIRST(&tp->segq); 774 if (q == NULL || q->ipqe_seq != tp->rcv_nxt) 775 goto out; 776 if (tp->t_state == TCPS_SYN_RECEIVED && q->ipqe_len) 777 goto out; 778 779 tp->rcv_nxt += q->ipqe_len; 780 pkt_flags = q->ipqe_flags & TH_FIN; 781 nd_hint(tp); 782 783 TAILQ_REMOVE(&tp->segq, q, ipqe_q); 784 TAILQ_REMOVE(&tp->timeq, q, ipqe_timeq); 785 tp->t_segqlen--; 786 KASSERT(tp->t_segqlen >= 0); 787 KASSERT(tp->t_segqlen != 0 || 788 (TAILQ_EMPTY(&tp->segq) && TAILQ_EMPTY(&tp->timeq))); 789 if (so->so_state & SS_CANTRCVMORE) 790 m_freem(q->ipqe_m); 791 else 792 sbappendstream(&so->so_rcv, q->ipqe_m); 793 tcpipqent_free(q); 794 TCP_REASS_UNLOCK(tp); 795 sorwakeup(so); 796 return pkt_flags; 797 798 out: 799 TCP_REASS_UNLOCK(tp); 800 return 0; 801 } 802 803 #ifdef INET6 804 int 805 tcp6_input(struct mbuf **mp, int *offp, int proto) 806 { 807 struct mbuf *m = *mp; 808 809 /* 810 * draft-itojun-ipv6-tcp-to-anycast 811 * better place to put this in? 812 */ 813 if (m->m_flags & M_ANYCAST6) { 814 struct ip6_hdr *ip6; 815 if (m->m_len < sizeof(struct ip6_hdr)) { 816 if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) { 817 TCP_STATINC(TCP_STAT_RCVSHORT); 818 return IPPROTO_DONE; 819 } 820 } 821 ip6 = mtod(m, struct ip6_hdr *); 822 icmp6_error(m, ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_ADDR, 823 (char *)&ip6->ip6_dst - (char *)ip6); 824 return IPPROTO_DONE; 825 } 826 827 tcp_input(m, *offp, proto); 828 return IPPROTO_DONE; 829 } 830 #endif 831 832 static void 833 tcp4_log_refused(const struct ip *ip, const struct tcphdr *th) 834 { 835 char src[INET_ADDRSTRLEN]; 836 char dst[INET_ADDRSTRLEN]; 837 838 if (ip) { 839 in_print(src, sizeof(src), &ip->ip_src); 840 in_print(dst, sizeof(dst), &ip->ip_dst); 841 } else { 842 strlcpy(src, "(unknown)", sizeof(src)); 843 strlcpy(dst, "(unknown)", sizeof(dst)); 844 } 845 log(LOG_INFO, 846 "Connection attempt to TCP %s:%d from %s:%d\n", 847 dst, ntohs(th->th_dport), 848 src, ntohs(th->th_sport)); 849 } 850 851 #ifdef INET6 852 static void 853 tcp6_log_refused(const struct ip6_hdr *ip6, const struct tcphdr *th) 854 { 855 char src[INET6_ADDRSTRLEN]; 856 char dst[INET6_ADDRSTRLEN]; 857 858 if (ip6) { 859 in6_print(src, sizeof(src), &ip6->ip6_src); 860 in6_print(dst, sizeof(dst), &ip6->ip6_dst); 861 } else { 862 strlcpy(src, "(unknown v6)", sizeof(src)); 863 strlcpy(dst, "(unknown v6)", sizeof(dst)); 864 } 865 log(LOG_INFO, 866 "Connection attempt to TCP [%s]:%d from [%s]:%d\n", 867 dst, ntohs(th->th_dport), 868 src, ntohs(th->th_sport)); 869 } 870 #endif 871 872 /* 873 * Checksum extended TCP header and data. 874 */ 875 int 876 tcp_input_checksum(int af, struct mbuf *m, const struct tcphdr *th, 877 int toff, int off, int tlen) 878 { 879 struct ifnet *rcvif; 880 int s; 881 882 /* 883 * XXX it's better to record and check if this mbuf is 884 * already checked. 885 */ 886 887 rcvif = m_get_rcvif(m, &s); 888 if (__predict_false(rcvif == NULL)) 889 goto badcsum; /* XXX */ 890 891 switch (af) { 892 case AF_INET: 893 switch (m->m_pkthdr.csum_flags & 894 ((rcvif->if_csum_flags_rx & M_CSUM_TCPv4) | 895 M_CSUM_TCP_UDP_BAD | M_CSUM_DATA)) { 896 case M_CSUM_TCPv4|M_CSUM_TCP_UDP_BAD: 897 TCP_CSUM_COUNTER_INCR(&tcp_hwcsum_bad); 898 goto badcsum; 899 900 case M_CSUM_TCPv4|M_CSUM_DATA: { 901 u_int32_t hw_csum = m->m_pkthdr.csum_data; 902 903 TCP_CSUM_COUNTER_INCR(&tcp_hwcsum_data); 904 if (m->m_pkthdr.csum_flags & M_CSUM_NO_PSEUDOHDR) { 905 const struct ip *ip = 906 mtod(m, const struct ip *); 907 908 hw_csum = in_cksum_phdr(ip->ip_src.s_addr, 909 ip->ip_dst.s_addr, 910 htons(hw_csum + tlen + off + IPPROTO_TCP)); 911 } 912 if ((hw_csum ^ 0xffff) != 0) 913 goto badcsum; 914 break; 915 } 916 917 case M_CSUM_TCPv4: 918 /* Checksum was okay. */ 919 TCP_CSUM_COUNTER_INCR(&tcp_hwcsum_ok); 920 break; 921 922 default: 923 /* 924 * Must compute it ourselves. Maybe skip checksum 925 * on loopback interfaces. 926 */ 927 if (__predict_true(!(rcvif->if_flags & IFF_LOOPBACK) || 928 tcp_do_loopback_cksum)) { 929 TCP_CSUM_COUNTER_INCR(&tcp_swcsum); 930 if (in4_cksum(m, IPPROTO_TCP, toff, 931 tlen + off) != 0) 932 goto badcsum; 933 } 934 break; 935 } 936 break; 937 938 #ifdef INET6 939 case AF_INET6: 940 switch (m->m_pkthdr.csum_flags & 941 ((rcvif->if_csum_flags_rx & M_CSUM_TCPv6) | 942 M_CSUM_TCP_UDP_BAD | M_CSUM_DATA)) { 943 case M_CSUM_TCPv6|M_CSUM_TCP_UDP_BAD: 944 TCP_CSUM_COUNTER_INCR(&tcp6_hwcsum_bad); 945 goto badcsum; 946 947 #if 0 /* notyet */ 948 case M_CSUM_TCPv6|M_CSUM_DATA: 949 #endif 950 951 case M_CSUM_TCPv6: 952 /* Checksum was okay. */ 953 TCP_CSUM_COUNTER_INCR(&tcp6_hwcsum_ok); 954 break; 955 956 default: 957 /* 958 * Must compute it ourselves. Maybe skip checksum 959 * on loopback interfaces. 960 */ 961 if (__predict_true((m->m_flags & M_LOOP) == 0 || 962 tcp_do_loopback_cksum)) { 963 TCP_CSUM_COUNTER_INCR(&tcp6_swcsum); 964 if (in6_cksum(m, IPPROTO_TCP, toff, 965 tlen + off) != 0) 966 goto badcsum; 967 } 968 } 969 break; 970 #endif /* INET6 */ 971 } 972 m_put_rcvif(rcvif, &s); 973 974 return 0; 975 976 badcsum: 977 m_put_rcvif(rcvif, &s); 978 TCP_STATINC(TCP_STAT_RCVBADSUM); 979 return -1; 980 } 981 982 /* 983 * When a packet arrives addressed to a vestigial tcpbp, we 984 * nevertheless have to respond to it per the spec. 985 * 986 * This code is duplicated from the one in tcp_input(). 987 */ 988 static void tcp_vtw_input(struct tcphdr *th, vestigial_inpcb_t *vp, 989 struct mbuf *m, int tlen) 990 { 991 int tiflags; 992 int todrop; 993 uint32_t t_flags = 0; 994 net_stat_ref_t tcps; 995 996 tiflags = th->th_flags; 997 todrop = vp->rcv_nxt - th->th_seq; 998 999 if (todrop > 0) { 1000 if (tiflags & TH_SYN) { 1001 tiflags &= ~TH_SYN; 1002 th->th_seq++; 1003 tcp_urp_drop(th, 1, &tiflags); 1004 todrop--; 1005 } 1006 if (todrop > tlen || 1007 (todrop == tlen && (tiflags & TH_FIN) == 0)) { 1008 /* 1009 * Any valid FIN or RST must be to the left of the 1010 * window. At this point the FIN or RST must be a 1011 * duplicate or out of sequence; drop it. 1012 */ 1013 if (tiflags & TH_RST) 1014 goto drop; 1015 tiflags &= ~(TH_FIN|TH_RST); 1016 1017 /* 1018 * Send an ACK to resynchronize and drop any data. 1019 * But keep on processing for RST or ACK. 1020 */ 1021 t_flags |= TF_ACKNOW; 1022 todrop = tlen; 1023 tcps = TCP_STAT_GETREF(); 1024 _NET_STATINC_REF(tcps, TCP_STAT_RCVDUPPACK); 1025 _NET_STATADD_REF(tcps, TCP_STAT_RCVDUPBYTE, todrop); 1026 TCP_STAT_PUTREF(); 1027 } else if ((tiflags & TH_RST) && 1028 th->th_seq != vp->rcv_nxt) { 1029 /* 1030 * Test for reset before adjusting the sequence 1031 * number for overlapping data. 1032 */ 1033 goto dropafterack_ratelim; 1034 } else { 1035 tcps = TCP_STAT_GETREF(); 1036 _NET_STATINC_REF(tcps, TCP_STAT_RCVPARTDUPPACK); 1037 _NET_STATADD_REF(tcps, TCP_STAT_RCVPARTDUPBYTE, 1038 todrop); 1039 TCP_STAT_PUTREF(); 1040 } 1041 1042 // tcp_new_dsack(tp, th->th_seq, todrop); 1043 // hdroptlen += todrop; /*drop from head afterwards*/ 1044 1045 th->th_seq += todrop; 1046 tlen -= todrop; 1047 tcp_urp_drop(th, todrop, &tiflags); 1048 } 1049 1050 /* 1051 * If new data are received on a connection after the 1052 * user processes are gone, then RST the other end. 1053 */ 1054 if (tlen) { 1055 TCP_STATINC(TCP_STAT_RCVAFTERCLOSE); 1056 goto dropwithreset; 1057 } 1058 1059 /* 1060 * If segment ends after window, drop trailing data 1061 * (and PUSH and FIN); if nothing left, just ACK. 1062 */ 1063 todrop = (th->th_seq + tlen) - (vp->rcv_nxt + vp->rcv_wnd); 1064 1065 if (todrop > 0) { 1066 TCP_STATINC(TCP_STAT_RCVPACKAFTERWIN); 1067 if (todrop >= tlen) { 1068 /* 1069 * The segment actually starts after the window. 1070 * th->th_seq + tlen - vp->rcv_nxt - vp->rcv_wnd >= tlen 1071 * th->th_seq - vp->rcv_nxt - vp->rcv_wnd >= 0 1072 * th->th_seq >= vp->rcv_nxt + vp->rcv_wnd 1073 */ 1074 TCP_STATADD(TCP_STAT_RCVBYTEAFTERWIN, tlen); 1075 1076 /* 1077 * If a new connection request is received 1078 * while in TIME_WAIT, drop the old connection 1079 * and start over if the sequence numbers 1080 * are above the previous ones. 1081 */ 1082 if ((tiflags & TH_SYN) && 1083 SEQ_GT(th->th_seq, vp->rcv_nxt)) { 1084 /* 1085 * We only support this in the !NOFDREF case, which 1086 * is to say: not here. 1087 */ 1088 goto dropwithreset; 1089 } 1090 1091 /* 1092 * If window is closed can only take segments at 1093 * window edge, and have to drop data and PUSH from 1094 * incoming segments. Continue processing, but 1095 * remember to ack. Otherwise, drop segment 1096 * and (if not RST) ack. 1097 */ 1098 if (vp->rcv_wnd == 0 && th->th_seq == vp->rcv_nxt) { 1099 t_flags |= TF_ACKNOW; 1100 TCP_STATINC(TCP_STAT_RCVWINPROBE); 1101 } else { 1102 goto dropafterack; 1103 } 1104 } else { 1105 TCP_STATADD(TCP_STAT_RCVBYTEAFTERWIN, todrop); 1106 } 1107 m_adj(m, -todrop); 1108 tlen -= todrop; 1109 tiflags &= ~(TH_PUSH|TH_FIN); 1110 } 1111 1112 if (tiflags & TH_RST) { 1113 if (th->th_seq != vp->rcv_nxt) 1114 goto dropafterack_ratelim; 1115 1116 vtw_del(vp->ctl, vp->vtw); 1117 goto drop; 1118 } 1119 1120 /* 1121 * If the ACK bit is off we drop the segment and return. 1122 */ 1123 if ((tiflags & TH_ACK) == 0) { 1124 if (t_flags & TF_ACKNOW) 1125 goto dropafterack; 1126 goto drop; 1127 } 1128 1129 /* 1130 * In TIME_WAIT state the only thing that should arrive 1131 * is a retransmission of the remote FIN. Acknowledge 1132 * it and restart the finack timer. 1133 */ 1134 vtw_restart(vp); 1135 goto dropafterack; 1136 1137 dropafterack: 1138 /* 1139 * Generate an ACK dropping incoming segment if it occupies 1140 * sequence space, where the ACK reflects our state. 1141 */ 1142 if (tiflags & TH_RST) 1143 goto drop; 1144 goto dropafterack2; 1145 1146 dropafterack_ratelim: 1147 /* 1148 * We may want to rate-limit ACKs against SYN/RST attack. 1149 */ 1150 if (ppsratecheck(&tcp_ackdrop_ppslim_last, &tcp_ackdrop_ppslim_count, 1151 tcp_ackdrop_ppslim) == 0) { 1152 /* XXX stat */ 1153 goto drop; 1154 } 1155 /* ...fall into dropafterack2... */ 1156 1157 dropafterack2: 1158 (void)tcp_respond(0, m, m, th, th->th_seq + tlen, th->th_ack, TH_ACK); 1159 return; 1160 1161 dropwithreset: 1162 /* 1163 * Generate a RST, dropping incoming segment. 1164 * Make ACK acceptable to originator of segment. 1165 */ 1166 if (tiflags & TH_RST) 1167 goto drop; 1168 1169 if (tiflags & TH_ACK) { 1170 tcp_respond(0, m, m, th, (tcp_seq)0, th->th_ack, TH_RST); 1171 } else { 1172 if (tiflags & TH_SYN) 1173 ++tlen; 1174 (void)tcp_respond(0, m, m, th, th->th_seq + tlen, (tcp_seq)0, 1175 TH_RST|TH_ACK); 1176 } 1177 return; 1178 drop: 1179 m_freem(m); 1180 } 1181 1182 /* 1183 * TCP input routine, follows pages 65-76 of RFC 793 very closely. 1184 */ 1185 void 1186 tcp_input(struct mbuf *m, int off, int proto) 1187 { 1188 struct tcphdr *th; 1189 struct ip *ip; 1190 struct inpcb *inp; 1191 #ifdef INET6 1192 struct ip6_hdr *ip6; 1193 #endif 1194 u_int8_t *optp = NULL; 1195 int optlen = 0; 1196 int len, tlen, hdroptlen = 0; 1197 struct tcpcb *tp = NULL; 1198 int tiflags; 1199 struct socket *so = NULL; 1200 int todrop, acked, ourfinisacked, needoutput = 0; 1201 bool dupseg; 1202 #ifdef TCP_DEBUG 1203 short ostate = 0; 1204 #endif 1205 u_long tiwin; 1206 struct tcp_opt_info opti; 1207 int thlen, iphlen; 1208 int af; /* af on the wire */ 1209 struct mbuf *tcp_saveti = NULL; 1210 uint32_t ts_rtt; 1211 uint8_t iptos; 1212 net_stat_ref_t tcps; 1213 vestigial_inpcb_t vestige; 1214 1215 vestige.valid = 0; 1216 1217 MCLAIM(m, &tcp_rx_mowner); 1218 1219 TCP_STATINC(TCP_STAT_RCVTOTAL); 1220 1221 memset(&opti, 0, sizeof(opti)); 1222 opti.ts_present = 0; 1223 opti.maxseg = 0; 1224 1225 /* 1226 * RFC1122 4.2.3.10, p. 104: discard bcast/mcast SYN. 1227 * 1228 * TCP is, by definition, unicast, so we reject all 1229 * multicast outright. 1230 * 1231 * Note, there are additional src/dst address checks in 1232 * the AF-specific code below. 1233 */ 1234 if (m->m_flags & (M_BCAST|M_MCAST)) { 1235 /* XXX stat */ 1236 goto drop; 1237 } 1238 #ifdef INET6 1239 if (m->m_flags & M_ANYCAST6) { 1240 /* XXX stat */ 1241 goto drop; 1242 } 1243 #endif 1244 1245 M_REGION_GET(th, struct tcphdr *, m, off, sizeof(struct tcphdr)); 1246 if (th == NULL) { 1247 TCP_STATINC(TCP_STAT_RCVSHORT); 1248 return; 1249 } 1250 1251 /* 1252 * Enforce alignment requirements that are violated in 1253 * some cases, see kern/50766 for details. 1254 */ 1255 if (ACCESSIBLE_POINTER(th, struct tcphdr) == 0) { 1256 m = m_copyup(m, off + sizeof(struct tcphdr), 0); 1257 if (m == NULL) { 1258 TCP_STATINC(TCP_STAT_RCVSHORT); 1259 return; 1260 } 1261 th = (struct tcphdr *)(mtod(m, char *) + off); 1262 } 1263 KASSERT(ACCESSIBLE_POINTER(th, struct tcphdr)); 1264 1265 /* 1266 * Get IP and TCP header. 1267 * Note: IP leaves IP header in first mbuf. 1268 */ 1269 ip = mtod(m, struct ip *); 1270 #ifdef INET6 1271 ip6 = mtod(m, struct ip6_hdr *); 1272 #endif 1273 switch (ip->ip_v) { 1274 case 4: 1275 af = AF_INET; 1276 iphlen = sizeof(struct ip); 1277 1278 if (IN_MULTICAST(ip->ip_dst.s_addr) || 1279 in_broadcast(ip->ip_dst, m_get_rcvif_NOMPSAFE(m))) 1280 goto drop; 1281 1282 /* We do the checksum after PCB lookup... */ 1283 len = ntohs(ip->ip_len); 1284 tlen = len - off; 1285 iptos = ip->ip_tos; 1286 break; 1287 #ifdef INET6 1288 case 6: 1289 iphlen = sizeof(struct ip6_hdr); 1290 af = AF_INET6; 1291 1292 /* 1293 * Be proactive about unspecified IPv6 address in source. 1294 * As we use all-zero to indicate unbounded/unconnected pcb, 1295 * unspecified IPv6 address can be used to confuse us. 1296 * 1297 * Note that packets with unspecified IPv6 destination is 1298 * already dropped in ip6_input. 1299 */ 1300 if (IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src)) { 1301 /* XXX stat */ 1302 goto drop; 1303 } 1304 1305 /* 1306 * Make sure destination address is not multicast. 1307 * Source address checked in ip6_input(). 1308 */ 1309 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) { 1310 /* XXX stat */ 1311 goto drop; 1312 } 1313 1314 /* We do the checksum after PCB lookup... */ 1315 len = m->m_pkthdr.len; 1316 tlen = len - off; 1317 iptos = (ntohl(ip6->ip6_flow) >> 20) & 0xff; 1318 break; 1319 #endif 1320 default: 1321 m_freem(m); 1322 return; 1323 } 1324 1325 1326 /* 1327 * Check that TCP offset makes sense, pull out TCP options and 1328 * adjust length. 1329 */ 1330 thlen = th->th_off << 2; 1331 if (thlen < sizeof(struct tcphdr) || thlen > tlen) { 1332 TCP_STATINC(TCP_STAT_RCVBADOFF); 1333 goto drop; 1334 } 1335 tlen -= thlen; 1336 1337 if (thlen > sizeof(struct tcphdr)) { 1338 M_REGION_GET(th, struct tcphdr *, m, off, thlen); 1339 if (th == NULL) { 1340 TCP_STATINC(TCP_STAT_RCVSHORT); 1341 return; 1342 } 1343 KASSERT(ACCESSIBLE_POINTER(th, struct tcphdr)); 1344 optlen = thlen - sizeof(struct tcphdr); 1345 optp = ((u_int8_t *)th) + sizeof(struct tcphdr); 1346 1347 /* 1348 * Do quick retrieval of timestamp options. 1349 * 1350 * If timestamp is the only option and it's formatted as 1351 * recommended in RFC 1323 appendix A, we quickly get the 1352 * values now and don't bother calling tcp_dooptions(), 1353 * etc. 1354 */ 1355 if ((optlen == TCPOLEN_TSTAMP_APPA || 1356 (optlen > TCPOLEN_TSTAMP_APPA && 1357 optp[TCPOLEN_TSTAMP_APPA] == TCPOPT_EOL)) && 1358 be32dec(optp) == TCPOPT_TSTAMP_HDR && 1359 (th->th_flags & TH_SYN) == 0) { 1360 opti.ts_present = 1; 1361 opti.ts_val = be32dec(optp + 4); 1362 opti.ts_ecr = be32dec(optp + 8); 1363 optp = NULL; /* we've parsed the options */ 1364 } 1365 } 1366 tiflags = th->th_flags; 1367 1368 /* 1369 * Checksum extended TCP header and data 1370 */ 1371 if (tcp_input_checksum(af, m, th, off, thlen, tlen)) 1372 goto badcsum; 1373 1374 /* 1375 * Locate pcb for segment. 1376 */ 1377 findpcb: 1378 inp = NULL; 1379 switch (af) { 1380 case AF_INET: 1381 inp = inpcb_lookup(&tcbtable, ip->ip_src, th->th_sport, 1382 ip->ip_dst, th->th_dport, &vestige); 1383 if (inp == NULL && !vestige.valid) { 1384 TCP_STATINC(TCP_STAT_PCBHASHMISS); 1385 inp = inpcb_lookup_bound(&tcbtable, ip->ip_dst, 1386 th->th_dport); 1387 } 1388 #ifdef INET6 1389 if (inp == NULL && !vestige.valid) { 1390 struct in6_addr s, d; 1391 1392 /* mapped addr case */ 1393 in6_in_2_v4mapin6(&ip->ip_src, &s); 1394 in6_in_2_v4mapin6(&ip->ip_dst, &d); 1395 inp = in6pcb_lookup(&tcbtable, &s, 1396 th->th_sport, &d, th->th_dport, 0, &vestige); 1397 if (inp == NULL && !vestige.valid) { 1398 TCP_STATINC(TCP_STAT_PCBHASHMISS); 1399 inp = in6pcb_lookup_bound(&tcbtable, &d, 1400 th->th_dport, 0); 1401 } 1402 } 1403 #endif 1404 if (inp == NULL && !vestige.valid) { 1405 TCP_STATINC(TCP_STAT_NOPORT); 1406 if (tcp_log_refused && 1407 (tiflags & (TH_RST|TH_ACK|TH_SYN)) == TH_SYN) { 1408 tcp4_log_refused(ip, th); 1409 } 1410 tcp_fields_to_host(th); 1411 goto dropwithreset_ratelim; 1412 } 1413 #if defined(IPSEC) 1414 if (ipsec_used) { 1415 if (inp && ipsec_in_reject(m, inp)) 1416 goto drop; 1417 } 1418 #endif /*IPSEC*/ 1419 break; 1420 #ifdef INET6 1421 case AF_INET6: 1422 { 1423 int faith; 1424 1425 #if defined(NFAITH) && NFAITH > 0 1426 faith = faithprefix(&ip6->ip6_dst); 1427 #else 1428 faith = 0; 1429 #endif 1430 inp = in6pcb_lookup(&tcbtable, &ip6->ip6_src, 1431 th->th_sport, &ip6->ip6_dst, th->th_dport, faith, &vestige); 1432 if (inp == NULL && !vestige.valid) { 1433 TCP_STATINC(TCP_STAT_PCBHASHMISS); 1434 inp = in6pcb_lookup_bound(&tcbtable, &ip6->ip6_dst, 1435 th->th_dport, faith); 1436 } 1437 if (inp == NULL && !vestige.valid) { 1438 TCP_STATINC(TCP_STAT_NOPORT); 1439 if (tcp_log_refused && 1440 (tiflags & (TH_RST|TH_ACK|TH_SYN)) == TH_SYN) { 1441 tcp6_log_refused(ip6, th); 1442 } 1443 tcp_fields_to_host(th); 1444 goto dropwithreset_ratelim; 1445 } 1446 #if defined(IPSEC) 1447 if (ipsec_used && inp && ipsec_in_reject(m, inp)) 1448 goto drop; 1449 #endif 1450 break; 1451 } 1452 #endif 1453 } 1454 1455 tcp_fields_to_host(th); 1456 1457 /* 1458 * If the state is CLOSED (i.e., TCB does not exist) then 1459 * all data in the incoming segment is discarded. 1460 * If the TCB exists but is in CLOSED state, it is embryonic, 1461 * but should either do a listen or a connect soon. 1462 */ 1463 tp = NULL; 1464 so = NULL; 1465 if (inp) { 1466 /* Check the minimum TTL for socket. */ 1467 if (inp->inp_af == AF_INET && ip->ip_ttl < in4p_ip_minttl(inp)) 1468 goto drop; 1469 1470 tp = intotcpcb(inp); 1471 so = inp->inp_socket; 1472 } else if (vestige.valid) { 1473 /* We do not support the resurrection of vtw tcpcps. */ 1474 tcp_vtw_input(th, &vestige, m, tlen); 1475 m = NULL; 1476 goto drop; 1477 } 1478 1479 if (tp == NULL) 1480 goto dropwithreset_ratelim; 1481 if (tp->t_state == TCPS_CLOSED) 1482 goto drop; 1483 1484 KASSERT(so->so_lock == softnet_lock); 1485 KASSERT(solocked(so)); 1486 1487 /* Unscale the window into a 32-bit value. */ 1488 if ((tiflags & TH_SYN) == 0) 1489 tiwin = th->th_win << tp->snd_scale; 1490 else 1491 tiwin = th->th_win; 1492 1493 #ifdef INET6 1494 /* save packet options if user wanted */ 1495 if (inp->inp_af == AF_INET6 && (inp->inp_flags & IN6P_CONTROLOPTS)) { 1496 m_freem(inp->inp_options); 1497 inp->inp_options = NULL; 1498 ip6_savecontrol(inp, &inp->inp_options, ip6, m); 1499 } 1500 #endif 1501 1502 if (so->so_options & SO_DEBUG) { 1503 #ifdef TCP_DEBUG 1504 ostate = tp->t_state; 1505 #endif 1506 1507 tcp_saveti = NULL; 1508 if (iphlen + sizeof(struct tcphdr) > MHLEN) 1509 goto nosave; 1510 1511 if (m->m_len > iphlen && (m->m_flags & M_EXT) == 0) { 1512 tcp_saveti = m_copym(m, 0, iphlen, M_DONTWAIT); 1513 if (tcp_saveti == NULL) 1514 goto nosave; 1515 } else { 1516 MGETHDR(tcp_saveti, M_DONTWAIT, MT_HEADER); 1517 if (tcp_saveti == NULL) 1518 goto nosave; 1519 MCLAIM(m, &tcp_mowner); 1520 tcp_saveti->m_len = iphlen; 1521 m_copydata(m, 0, iphlen, 1522 mtod(tcp_saveti, void *)); 1523 } 1524 1525 if (M_TRAILINGSPACE(tcp_saveti) < sizeof(struct tcphdr)) { 1526 m_freem(tcp_saveti); 1527 tcp_saveti = NULL; 1528 } else { 1529 tcp_saveti->m_len += sizeof(struct tcphdr); 1530 memcpy(mtod(tcp_saveti, char *) + iphlen, th, 1531 sizeof(struct tcphdr)); 1532 } 1533 nosave:; 1534 } 1535 1536 if (so->so_options & SO_ACCEPTCONN) { 1537 union syn_cache_sa src; 1538 union syn_cache_sa dst; 1539 1540 KASSERT(tp->t_state == TCPS_LISTEN); 1541 1542 memset(&src, 0, sizeof(src)); 1543 memset(&dst, 0, sizeof(dst)); 1544 switch (af) { 1545 case AF_INET: 1546 src.sin.sin_len = sizeof(struct sockaddr_in); 1547 src.sin.sin_family = AF_INET; 1548 src.sin.sin_addr = ip->ip_src; 1549 src.sin.sin_port = th->th_sport; 1550 1551 dst.sin.sin_len = sizeof(struct sockaddr_in); 1552 dst.sin.sin_family = AF_INET; 1553 dst.sin.sin_addr = ip->ip_dst; 1554 dst.sin.sin_port = th->th_dport; 1555 break; 1556 #ifdef INET6 1557 case AF_INET6: 1558 src.sin6.sin6_len = sizeof(struct sockaddr_in6); 1559 src.sin6.sin6_family = AF_INET6; 1560 src.sin6.sin6_addr = ip6->ip6_src; 1561 src.sin6.sin6_port = th->th_sport; 1562 1563 dst.sin6.sin6_len = sizeof(struct sockaddr_in6); 1564 dst.sin6.sin6_family = AF_INET6; 1565 dst.sin6.sin6_addr = ip6->ip6_dst; 1566 dst.sin6.sin6_port = th->th_dport; 1567 break; 1568 #endif 1569 } 1570 1571 if ((tiflags & (TH_RST|TH_ACK|TH_SYN)) != TH_SYN) { 1572 if (tiflags & TH_RST) { 1573 syn_cache_reset(&src.sa, &dst.sa, th); 1574 } else if ((tiflags & (TH_ACK|TH_SYN)) == 1575 (TH_ACK|TH_SYN)) { 1576 /* 1577 * Received a SYN,ACK. This should never 1578 * happen while we are in LISTEN. Send an RST. 1579 */ 1580 goto badsyn; 1581 } else if (tiflags & TH_ACK) { 1582 so = syn_cache_get(&src.sa, &dst.sa, th, so, m); 1583 if (so == NULL) { 1584 /* 1585 * We don't have a SYN for this ACK; 1586 * send an RST. 1587 */ 1588 goto badsyn; 1589 } else if (so == (struct socket *)(-1)) { 1590 /* 1591 * We were unable to create the 1592 * connection. If the 3-way handshake 1593 * was completed, and RST has been 1594 * sent to the peer. Since the mbuf 1595 * might be in use for the reply, do 1596 * not free it. 1597 */ 1598 m = NULL; 1599 } else { 1600 /* 1601 * We have created a full-blown 1602 * connection. 1603 */ 1604 inp = sotoinpcb(so); 1605 tp = intotcpcb(inp); 1606 if (tp == NULL) 1607 goto badsyn; /*XXX*/ 1608 tiwin <<= tp->snd_scale; 1609 goto after_listen; 1610 } 1611 } else { 1612 /* 1613 * None of RST, SYN or ACK was set. 1614 * This is an invalid packet for a 1615 * TCB in LISTEN state. Send a RST. 1616 */ 1617 goto badsyn; 1618 } 1619 } else { 1620 /* 1621 * Received a SYN. 1622 */ 1623 1624 #ifdef INET6 1625 /* 1626 * If deprecated address is forbidden, we do 1627 * not accept SYN to deprecated interface 1628 * address to prevent any new inbound 1629 * connection from getting established. 1630 * When we do not accept SYN, we send a TCP 1631 * RST, with deprecated source address (instead 1632 * of dropping it). We compromise it as it is 1633 * much better for peer to send a RST, and 1634 * RST will be the final packet for the 1635 * exchange. 1636 * 1637 * If we do not forbid deprecated addresses, we 1638 * accept the SYN packet. RFC2462 does not 1639 * suggest dropping SYN in this case. 1640 * If we decipher RFC2462 5.5.4, it says like 1641 * this: 1642 * 1. use of deprecated addr with existing 1643 * communication is okay - "SHOULD continue 1644 * to be used" 1645 * 2. use of it with new communication: 1646 * (2a) "SHOULD NOT be used if alternate 1647 * address with sufficient scope is 1648 * available" 1649 * (2b) nothing mentioned otherwise. 1650 * Here we fall into (2b) case as we have no 1651 * choice in our source address selection - we 1652 * must obey the peer. 1653 * 1654 * The wording in RFC2462 is confusing, and 1655 * there are multiple description text for 1656 * deprecated address handling - worse, they 1657 * are not exactly the same. I believe 5.5.4 1658 * is the best one, so we follow 5.5.4. 1659 */ 1660 if (af == AF_INET6 && !ip6_use_deprecated) { 1661 struct in6_ifaddr *ia6; 1662 int s; 1663 struct ifnet *rcvif = m_get_rcvif(m, &s); 1664 if (rcvif == NULL) 1665 goto dropwithreset; /* XXX */ 1666 if ((ia6 = in6ifa_ifpwithaddr(rcvif, 1667 &ip6->ip6_dst)) && 1668 (ia6->ia6_flags & IN6_IFF_DEPRECATED)) { 1669 tp = NULL; 1670 m_put_rcvif(rcvif, &s); 1671 goto dropwithreset; 1672 } 1673 m_put_rcvif(rcvif, &s); 1674 } 1675 #endif 1676 1677 /* 1678 * LISTEN socket received a SYN from itself? This 1679 * can't possibly be valid; drop the packet. 1680 */ 1681 if (th->th_sport == th->th_dport) { 1682 int eq = 0; 1683 1684 switch (af) { 1685 case AF_INET: 1686 eq = in_hosteq(ip->ip_src, ip->ip_dst); 1687 break; 1688 #ifdef INET6 1689 case AF_INET6: 1690 eq = IN6_ARE_ADDR_EQUAL(&ip6->ip6_src, 1691 &ip6->ip6_dst); 1692 break; 1693 #endif 1694 } 1695 if (eq) { 1696 TCP_STATINC(TCP_STAT_BADSYN); 1697 goto drop; 1698 } 1699 } 1700 1701 /* 1702 * SYN looks ok; create compressed TCP 1703 * state for it. 1704 */ 1705 if (so->so_qlen <= so->so_qlimit && 1706 syn_cache_add(&src.sa, &dst.sa, th, off, 1707 so, m, optp, optlen, &opti)) 1708 m = NULL; 1709 } 1710 1711 goto drop; 1712 } 1713 1714 after_listen: 1715 /* 1716 * From here on, we're dealing with !LISTEN. 1717 */ 1718 KASSERT(tp->t_state != TCPS_LISTEN); 1719 1720 /* 1721 * Segment received on connection. 1722 * Reset idle time and keep-alive timer. 1723 */ 1724 tp->t_rcvtime = tcp_now; 1725 if (TCPS_HAVEESTABLISHED(tp->t_state)) 1726 TCP_TIMER_ARM(tp, TCPT_KEEP, tp->t_keepidle); 1727 1728 /* 1729 * Process options. 1730 */ 1731 #ifdef TCP_SIGNATURE 1732 if (optp || (tp->t_flags & TF_SIGNATURE)) 1733 #else 1734 if (optp) 1735 #endif 1736 if (tcp_dooptions(tp, optp, optlen, th, m, off, &opti) < 0) 1737 goto drop; 1738 1739 if (TCP_SACK_ENABLED(tp)) { 1740 tcp_del_sackholes(tp, th); 1741 } 1742 1743 if (TCP_ECN_ALLOWED(tp)) { 1744 if (tiflags & TH_CWR) { 1745 tp->t_flags &= ~TF_ECN_SND_ECE; 1746 } 1747 switch (iptos & IPTOS_ECN_MASK) { 1748 case IPTOS_ECN_CE: 1749 tp->t_flags |= TF_ECN_SND_ECE; 1750 TCP_STATINC(TCP_STAT_ECN_CE); 1751 break; 1752 case IPTOS_ECN_ECT0: 1753 TCP_STATINC(TCP_STAT_ECN_ECT); 1754 break; 1755 case IPTOS_ECN_ECT1: 1756 /* XXX: ignore for now -- rpaulo */ 1757 break; 1758 } 1759 /* 1760 * Congestion experienced. 1761 * Ignore if we are already trying to recover. 1762 */ 1763 if ((tiflags & TH_ECE) && SEQ_GEQ(tp->snd_una, tp->snd_recover)) 1764 tp->t_congctl->cong_exp(tp); 1765 } 1766 1767 if (opti.ts_present && opti.ts_ecr) { 1768 /* 1769 * Calculate the RTT from the returned time stamp and the 1770 * connection's time base. If the time stamp is later than 1771 * the current time, or is extremely old, fall back to non-1323 1772 * RTT calculation. Since ts_rtt is unsigned, we can test both 1773 * at the same time. 1774 * 1775 * Note that ts_rtt is in units of slow ticks (500 1776 * ms). Since most earthbound RTTs are < 500 ms, 1777 * observed values will have large quantization noise. 1778 * Our smoothed RTT is then the fraction of observed 1779 * samples that are 1 tick instead of 0 (times 500 1780 * ms). 1781 * 1782 * ts_rtt is increased by 1 to denote a valid sample, 1783 * with 0 indicating an invalid measurement. This 1784 * extra 1 must be removed when ts_rtt is used, or 1785 * else an erroneous extra 500 ms will result. 1786 */ 1787 ts_rtt = TCP_TIMESTAMP(tp) - opti.ts_ecr + 1; 1788 if (ts_rtt > TCP_PAWS_IDLE) 1789 ts_rtt = 0; 1790 } else { 1791 ts_rtt = 0; 1792 } 1793 1794 /* 1795 * Fast path: check for the two common cases of a uni-directional 1796 * data transfer. If: 1797 * o We are in the ESTABLISHED state, and 1798 * o The packet has no control flags, and 1799 * o The packet is in-sequence, and 1800 * o The window didn't change, and 1801 * o We are not retransmitting 1802 * It's a candidate. 1803 * 1804 * If the length (tlen) is zero and the ack moved forward, we're 1805 * the sender side of the transfer. Just free the data acked and 1806 * wake any higher level process that was blocked waiting for 1807 * space. 1808 * 1809 * If the length is non-zero and the ack didn't move, we're the 1810 * receiver side. If we're getting packets in-order (the reassembly 1811 * queue is empty), add the data to the socket buffer and note 1812 * that we need a delayed ack. 1813 */ 1814 if (tp->t_state == TCPS_ESTABLISHED && 1815 (tiflags & (TH_SYN|TH_FIN|TH_RST|TH_URG|TH_ECE|TH_CWR|TH_ACK)) 1816 == TH_ACK && 1817 (!opti.ts_present || TSTMP_GEQ(opti.ts_val, tp->ts_recent)) && 1818 th->th_seq == tp->rcv_nxt && 1819 tiwin && tiwin == tp->snd_wnd && 1820 tp->snd_nxt == tp->snd_max) { 1821 1822 /* 1823 * If last ACK falls within this segment's sequence numbers, 1824 * record the timestamp. 1825 * NOTE that the test is modified according to the latest 1826 * proposal of the tcplw@cray.com list (Braden 1993/04/26). 1827 * 1828 * note that we already know 1829 * TSTMP_GEQ(opti.ts_val, tp->ts_recent) 1830 */ 1831 if (opti.ts_present && SEQ_LEQ(th->th_seq, tp->last_ack_sent)) { 1832 tp->ts_recent_age = tcp_now; 1833 tp->ts_recent = opti.ts_val; 1834 } 1835 1836 if (tlen == 0) { 1837 /* Ack prediction. */ 1838 if (SEQ_GT(th->th_ack, tp->snd_una) && 1839 SEQ_LEQ(th->th_ack, tp->snd_max) && 1840 tp->snd_cwnd >= tp->snd_wnd && 1841 tp->t_partialacks < 0) { 1842 /* 1843 * this is a pure ack for outstanding data. 1844 */ 1845 if (ts_rtt) 1846 tcp_xmit_timer(tp, ts_rtt - 1); 1847 else if (tp->t_rtttime && 1848 SEQ_GT(th->th_ack, tp->t_rtseq)) 1849 tcp_xmit_timer(tp, 1850 tcp_now - tp->t_rtttime); 1851 acked = th->th_ack - tp->snd_una; 1852 tcps = TCP_STAT_GETREF(); 1853 _NET_STATINC_REF(tcps, TCP_STAT_PREDACK); 1854 _NET_STATINC_REF(tcps, TCP_STAT_RCVACKPACK); 1855 _NET_STATADD_REF(tcps, TCP_STAT_RCVACKBYTE, 1856 acked); 1857 TCP_STAT_PUTREF(); 1858 nd_hint(tp); 1859 1860 if (acked > (tp->t_lastoff - tp->t_inoff)) 1861 tp->t_lastm = NULL; 1862 sbdrop(&so->so_snd, acked); 1863 tp->t_lastoff -= acked; 1864 1865 icmp_check(tp, th, acked); 1866 1867 tp->snd_una = th->th_ack; 1868 tp->snd_fack = tp->snd_una; 1869 if (SEQ_LT(tp->snd_high, tp->snd_una)) 1870 tp->snd_high = tp->snd_una; 1871 /* 1872 * drag snd_wl2 along so only newer 1873 * ACKs can update the window size. 1874 * also avoids the state where snd_wl2 1875 * is eventually larger than th_ack and thus 1876 * blocking the window update mechanism and 1877 * the connection gets stuck for a loooong 1878 * time in the zero sized send window state. 1879 * 1880 * see PR/kern 55567 1881 */ 1882 tp->snd_wl2 = tp->snd_una; 1883 1884 m_freem(m); 1885 1886 /* 1887 * If all outstanding data are acked, stop 1888 * retransmit timer, otherwise restart timer 1889 * using current (possibly backed-off) value. 1890 * If process is waiting for space, 1891 * wakeup/selnotify/signal. If data 1892 * are ready to send, let tcp_output 1893 * decide between more output or persist. 1894 */ 1895 if (tp->snd_una == tp->snd_max) 1896 TCP_TIMER_DISARM(tp, TCPT_REXMT); 1897 else if (TCP_TIMER_ISARMED(tp, 1898 TCPT_PERSIST) == 0) 1899 TCP_TIMER_ARM(tp, TCPT_REXMT, 1900 tp->t_rxtcur); 1901 1902 sowwakeup(so); 1903 if (so->so_snd.sb_cc) { 1904 KERNEL_LOCK(1, NULL); 1905 (void)tcp_output(tp); 1906 KERNEL_UNLOCK_ONE(NULL); 1907 } 1908 m_freem(tcp_saveti); 1909 return; 1910 } 1911 } else if (th->th_ack == tp->snd_una && 1912 TAILQ_FIRST(&tp->segq) == NULL && 1913 tlen <= sbspace(&so->so_rcv)) { 1914 int newsize = 0; 1915 1916 /* 1917 * this is a pure, in-sequence data packet 1918 * with nothing on the reassembly queue and 1919 * we have enough buffer space to take it. 1920 */ 1921 tp->rcv_nxt += tlen; 1922 1923 /* 1924 * Pull rcv_up up to prevent seq wrap relative to 1925 * rcv_nxt. 1926 */ 1927 tp->rcv_up = tp->rcv_nxt; 1928 1929 /* 1930 * Pull snd_wl1 up to prevent seq wrap relative to 1931 * th_seq. 1932 */ 1933 tp->snd_wl1 = th->th_seq; 1934 1935 tcps = TCP_STAT_GETREF(); 1936 _NET_STATINC_REF(tcps, TCP_STAT_PREDDAT); 1937 _NET_STATINC_REF(tcps, TCP_STAT_RCVPACK); 1938 _NET_STATADD_REF(tcps, TCP_STAT_RCVBYTE, tlen); 1939 TCP_STAT_PUTREF(); 1940 nd_hint(tp); 1941 /* 1942 * Automatic sizing enables the performance of large buffers 1943 * and most of the efficiency of small ones by only allocating 1944 * space when it is needed. 1945 * 1946 * On the receive side the socket buffer memory is only rarely 1947 * used to any significant extent. This allows us to be much 1948 * more aggressive in scaling the receive socket buffer. For 1949 * the case that the buffer space is actually used to a large 1950 * extent and we run out of kernel memory we can simply drop 1951 * the new segments; TCP on the sender will just retransmit it 1952 * later. Setting the buffer size too big may only consume too 1953 * much kernel memory if the application doesn't read() from 1954 * the socket or packet loss or reordering makes use of the 1955 * reassembly queue. 1956 * 1957 * The criteria to step up the receive buffer one notch are: 1958 * 1. the number of bytes received during the time it takes 1959 * one timestamp to be reflected back to us (the RTT); 1960 * 2. received bytes per RTT is within seven eighth of the 1961 * current socket buffer size; 1962 * 3. receive buffer size has not hit maximal automatic size; 1963 * 1964 * This algorithm does one step per RTT at most and only if 1965 * we receive a bulk stream w/o packet losses or reorderings. 1966 * Shrinking the buffer during idle times is not necessary as 1967 * it doesn't consume any memory when idle. 1968 * 1969 * TODO: Only step up if the application is actually serving 1970 * the buffer to better manage the socket buffer resources. 1971 */ 1972 if (tcp_do_autorcvbuf && 1973 opti.ts_ecr && 1974 (so->so_rcv.sb_flags & SB_AUTOSIZE)) { 1975 if (opti.ts_ecr > tp->rfbuf_ts && 1976 opti.ts_ecr - tp->rfbuf_ts < PR_SLOWHZ) { 1977 if (tp->rfbuf_cnt > 1978 (so->so_rcv.sb_hiwat / 8 * 7) && 1979 so->so_rcv.sb_hiwat < 1980 tcp_autorcvbuf_max) { 1981 newsize = 1982 uimin(so->so_rcv.sb_hiwat + 1983 tcp_autorcvbuf_inc, 1984 tcp_autorcvbuf_max); 1985 } 1986 /* Start over with next RTT. */ 1987 tp->rfbuf_ts = 0; 1988 tp->rfbuf_cnt = 0; 1989 } else 1990 tp->rfbuf_cnt += tlen; /* add up */ 1991 } 1992 1993 /* 1994 * Drop TCP, IP headers and TCP options then add data 1995 * to socket buffer. 1996 */ 1997 if (so->so_state & SS_CANTRCVMORE) { 1998 m_freem(m); 1999 } else { 2000 /* 2001 * Set new socket buffer size. 2002 * Give up when limit is reached. 2003 */ 2004 if (newsize) 2005 if (!sbreserve(&so->so_rcv, 2006 newsize, so)) 2007 so->so_rcv.sb_flags &= ~SB_AUTOSIZE; 2008 m_adj(m, off + thlen); 2009 sbappendstream(&so->so_rcv, m); 2010 } 2011 sorwakeup(so); 2012 tcp_setup_ack(tp, th); 2013 if (tp->t_flags & TF_ACKNOW) { 2014 KERNEL_LOCK(1, NULL); 2015 (void)tcp_output(tp); 2016 KERNEL_UNLOCK_ONE(NULL); 2017 } 2018 m_freem(tcp_saveti); 2019 return; 2020 } 2021 } 2022 2023 /* 2024 * Compute mbuf offset to TCP data segment. 2025 */ 2026 hdroptlen = off + thlen; 2027 2028 /* 2029 * Calculate amount of space in receive window. Receive window is 2030 * amount of space in rcv queue, but not less than advertised 2031 * window. 2032 */ 2033 { 2034 int win; 2035 win = sbspace(&so->so_rcv); 2036 if (win < 0) 2037 win = 0; 2038 tp->rcv_wnd = imax(win, (int)(tp->rcv_adv - tp->rcv_nxt)); 2039 } 2040 2041 /* Reset receive buffer auto scaling when not in bulk receive mode. */ 2042 tp->rfbuf_ts = 0; 2043 tp->rfbuf_cnt = 0; 2044 2045 switch (tp->t_state) { 2046 /* 2047 * If the state is SYN_SENT: 2048 * if seg contains an ACK, but not for our SYN, drop the input. 2049 * if seg contains a RST, then drop the connection. 2050 * if seg does not contain SYN, then drop it. 2051 * Otherwise this is an acceptable SYN segment 2052 * initialize tp->rcv_nxt and tp->irs 2053 * if seg contains ack then advance tp->snd_una 2054 * if seg contains a ECE and ECN support is enabled, the stream 2055 * is ECN capable. 2056 * if SYN has been acked change to ESTABLISHED else SYN_RCVD state 2057 * arrange for segment to be acked (eventually) 2058 * continue processing rest of data/controls, beginning with URG 2059 */ 2060 case TCPS_SYN_SENT: 2061 if ((tiflags & TH_ACK) && 2062 (SEQ_LEQ(th->th_ack, tp->iss) || 2063 SEQ_GT(th->th_ack, tp->snd_max))) 2064 goto dropwithreset; 2065 if (tiflags & TH_RST) { 2066 if (tiflags & TH_ACK) 2067 tp = tcp_drop(tp, ECONNREFUSED); 2068 goto drop; 2069 } 2070 if ((tiflags & TH_SYN) == 0) 2071 goto drop; 2072 if (tiflags & TH_ACK) { 2073 tp->snd_una = th->th_ack; 2074 if (SEQ_LT(tp->snd_nxt, tp->snd_una)) 2075 tp->snd_nxt = tp->snd_una; 2076 if (SEQ_LT(tp->snd_high, tp->snd_una)) 2077 tp->snd_high = tp->snd_una; 2078 TCP_TIMER_DISARM(tp, TCPT_REXMT); 2079 2080 if ((tiflags & TH_ECE) && tcp_do_ecn) { 2081 tp->t_flags |= TF_ECN_PERMIT; 2082 TCP_STATINC(TCP_STAT_ECN_SHS); 2083 } 2084 } 2085 tp->irs = th->th_seq; 2086 tcp_rcvseqinit(tp); 2087 tp->t_flags |= TF_ACKNOW; 2088 tcp_mss_from_peer(tp, opti.maxseg); 2089 2090 /* 2091 * Initialize the initial congestion window. If we 2092 * had to retransmit the SYN, we must initialize cwnd 2093 * to 1 segment (i.e. the Loss Window). 2094 */ 2095 if (tp->t_flags & TF_SYN_REXMT) 2096 tp->snd_cwnd = tp->t_peermss; 2097 else { 2098 int ss = tcp_init_win; 2099 if (inp->inp_af == AF_INET && in_localaddr(in4p_faddr(inp))) 2100 ss = tcp_init_win_local; 2101 #ifdef INET6 2102 else if (inp->inp_af == AF_INET6 && in6_localaddr(&in6p_faddr(inp))) 2103 ss = tcp_init_win_local; 2104 #endif 2105 tp->snd_cwnd = TCP_INITIAL_WINDOW(ss, tp->t_peermss); 2106 } 2107 2108 tcp_rmx_rtt(tp); 2109 if (tiflags & TH_ACK) { 2110 TCP_STATINC(TCP_STAT_CONNECTS); 2111 /* 2112 * move tcp_established before soisconnected 2113 * because upcall handler can drive tcp_output 2114 * functionality. 2115 * XXX we might call soisconnected at the end of 2116 * all processing 2117 */ 2118 tcp_established(tp); 2119 soisconnected(so); 2120 /* Do window scaling on this connection? */ 2121 if ((tp->t_flags & (TF_RCVD_SCALE|TF_REQ_SCALE)) == 2122 (TF_RCVD_SCALE|TF_REQ_SCALE)) { 2123 tp->snd_scale = tp->requested_s_scale; 2124 tp->rcv_scale = tp->request_r_scale; 2125 } 2126 TCP_REASS_LOCK(tp); 2127 (void)tcp_reass(tp, NULL, NULL, tlen); 2128 /* 2129 * if we didn't have to retransmit the SYN, 2130 * use its rtt as our initial srtt & rtt var. 2131 */ 2132 if (tp->t_rtttime) 2133 tcp_xmit_timer(tp, tcp_now - tp->t_rtttime); 2134 } else { 2135 tp->t_state = TCPS_SYN_RECEIVED; 2136 } 2137 2138 /* 2139 * Advance th->th_seq to correspond to first data byte. 2140 * If data, trim to stay within window, 2141 * dropping FIN if necessary. 2142 */ 2143 th->th_seq++; 2144 if (tlen > tp->rcv_wnd) { 2145 todrop = tlen - tp->rcv_wnd; 2146 m_adj(m, -todrop); 2147 tlen = tp->rcv_wnd; 2148 tiflags &= ~TH_FIN; 2149 tcps = TCP_STAT_GETREF(); 2150 _NET_STATINC_REF(tcps, TCP_STAT_RCVPACKAFTERWIN); 2151 _NET_STATADD_REF(tcps, TCP_STAT_RCVBYTEAFTERWIN, 2152 todrop); 2153 TCP_STAT_PUTREF(); 2154 } 2155 tp->snd_wl1 = th->th_seq - 1; 2156 tp->rcv_up = th->th_seq; 2157 goto step6; 2158 2159 /* 2160 * If the state is SYN_RECEIVED: 2161 * If seg contains an ACK, but not for our SYN, drop the input 2162 * and generate an RST. See page 36, rfc793 2163 */ 2164 case TCPS_SYN_RECEIVED: 2165 if ((tiflags & TH_ACK) && 2166 (SEQ_LEQ(th->th_ack, tp->iss) || 2167 SEQ_GT(th->th_ack, tp->snd_max))) 2168 goto dropwithreset; 2169 break; 2170 } 2171 2172 /* 2173 * From here on, we're dealing with !LISTEN and !SYN_SENT. 2174 */ 2175 KASSERT(tp->t_state != TCPS_LISTEN && 2176 tp->t_state != TCPS_SYN_SENT); 2177 2178 /* 2179 * RFC1323 PAWS: if we have a timestamp reply on this segment and 2180 * it's less than ts_recent, drop it. 2181 */ 2182 if (opti.ts_present && (tiflags & TH_RST) == 0 && tp->ts_recent && 2183 TSTMP_LT(opti.ts_val, tp->ts_recent)) { 2184 /* Check to see if ts_recent is over 24 days old. */ 2185 if (tcp_now - tp->ts_recent_age > TCP_PAWS_IDLE) { 2186 /* 2187 * Invalidate ts_recent. If this segment updates 2188 * ts_recent, the age will be reset later and ts_recent 2189 * will get a valid value. If it does not, setting 2190 * ts_recent to zero will at least satisfy the 2191 * requirement that zero be placed in the timestamp 2192 * echo reply when ts_recent isn't valid. The 2193 * age isn't reset until we get a valid ts_recent 2194 * because we don't want out-of-order segments to be 2195 * dropped when ts_recent is old. 2196 */ 2197 tp->ts_recent = 0; 2198 } else { 2199 tcps = TCP_STAT_GETREF(); 2200 _NET_STATINC_REF(tcps, TCP_STAT_RCVDUPPACK); 2201 _NET_STATADD_REF(tcps, TCP_STAT_RCVDUPBYTE, tlen); 2202 _NET_STATINC_REF(tcps, TCP_STAT_PAWSDROP); 2203 TCP_STAT_PUTREF(); 2204 tcp_new_dsack(tp, th->th_seq, tlen); 2205 goto dropafterack; 2206 } 2207 } 2208 2209 /* 2210 * Check that at least some bytes of the segment are within the 2211 * receive window. If segment begins before rcv_nxt, drop leading 2212 * data (and SYN); if nothing left, just ack. 2213 */ 2214 todrop = tp->rcv_nxt - th->th_seq; 2215 dupseg = false; 2216 if (todrop > 0) { 2217 if (tiflags & TH_SYN) { 2218 tiflags &= ~TH_SYN; 2219 th->th_seq++; 2220 tcp_urp_drop(th, 1, &tiflags); 2221 todrop--; 2222 } 2223 if (todrop > tlen || 2224 (todrop == tlen && (tiflags & TH_FIN) == 0)) { 2225 /* 2226 * Any valid FIN or RST must be to the left of the 2227 * window. At this point the FIN or RST must be a 2228 * duplicate or out of sequence; drop it. 2229 */ 2230 if (tiflags & TH_RST) 2231 goto drop; 2232 tiflags &= ~(TH_FIN|TH_RST); 2233 2234 /* 2235 * Send an ACK to resynchronize and drop any data. 2236 * But keep on processing for RST or ACK. 2237 */ 2238 tp->t_flags |= TF_ACKNOW; 2239 todrop = tlen; 2240 dupseg = true; 2241 tcps = TCP_STAT_GETREF(); 2242 _NET_STATINC_REF(tcps, TCP_STAT_RCVDUPPACK); 2243 _NET_STATADD_REF(tcps, TCP_STAT_RCVDUPBYTE, todrop); 2244 TCP_STAT_PUTREF(); 2245 } else if ((tiflags & TH_RST) && th->th_seq != tp->rcv_nxt) { 2246 /* 2247 * Test for reset before adjusting the sequence 2248 * number for overlapping data. 2249 */ 2250 goto dropafterack_ratelim; 2251 } else { 2252 tcps = TCP_STAT_GETREF(); 2253 _NET_STATINC_REF(tcps, TCP_STAT_RCVPARTDUPPACK); 2254 _NET_STATADD_REF(tcps, TCP_STAT_RCVPARTDUPBYTE, 2255 todrop); 2256 TCP_STAT_PUTREF(); 2257 } 2258 tcp_new_dsack(tp, th->th_seq, todrop); 2259 hdroptlen += todrop; /* drop from head afterwards (m_adj) */ 2260 th->th_seq += todrop; 2261 tlen -= todrop; 2262 tcp_urp_drop(th, todrop, &tiflags); 2263 } 2264 2265 /* 2266 * If new data is received on a connection after the user processes 2267 * are gone, then RST the other end. 2268 */ 2269 if ((so->so_state & SS_NOFDREF) && 2270 tp->t_state > TCPS_CLOSE_WAIT && tlen) { 2271 tp = tcp_close(tp); 2272 TCP_STATINC(TCP_STAT_RCVAFTERCLOSE); 2273 goto dropwithreset; 2274 } 2275 2276 /* 2277 * If the segment ends after the window, drop trailing data (and 2278 * PUSH and FIN); if nothing left, just ACK. 2279 */ 2280 todrop = (th->th_seq + tlen) - (tp->rcv_nxt + tp->rcv_wnd); 2281 if (todrop > 0) { 2282 TCP_STATINC(TCP_STAT_RCVPACKAFTERWIN); 2283 if (todrop >= tlen) { 2284 /* 2285 * The segment actually starts after the window. 2286 * th->th_seq + tlen - tp->rcv_nxt - tp->rcv_wnd >= tlen 2287 * th->th_seq - tp->rcv_nxt - tp->rcv_wnd >= 0 2288 * th->th_seq >= tp->rcv_nxt + tp->rcv_wnd 2289 */ 2290 TCP_STATADD(TCP_STAT_RCVBYTEAFTERWIN, tlen); 2291 2292 /* 2293 * If a new connection request is received while in 2294 * TIME_WAIT, drop the old connection and start over 2295 * if the sequence numbers are above the previous 2296 * ones. 2297 * 2298 * NOTE: We need to put the header fields back into 2299 * network order. 2300 */ 2301 if ((tiflags & TH_SYN) && 2302 tp->t_state == TCPS_TIME_WAIT && 2303 SEQ_GT(th->th_seq, tp->rcv_nxt)) { 2304 tp = tcp_close(tp); 2305 tcp_fields_to_net(th); 2306 m_freem(tcp_saveti); 2307 tcp_saveti = NULL; 2308 goto findpcb; 2309 } 2310 2311 /* 2312 * If window is closed can only take segments at 2313 * window edge, and have to drop data and PUSH from 2314 * incoming segments. Continue processing, but 2315 * remember to ack. Otherwise, drop segment 2316 * and (if not RST) ack. 2317 */ 2318 if (tp->rcv_wnd == 0 && th->th_seq == tp->rcv_nxt) { 2319 KASSERT(todrop == tlen); 2320 tp->t_flags |= TF_ACKNOW; 2321 TCP_STATINC(TCP_STAT_RCVWINPROBE); 2322 } else { 2323 goto dropafterack; 2324 } 2325 } else { 2326 TCP_STATADD(TCP_STAT_RCVBYTEAFTERWIN, todrop); 2327 } 2328 m_adj(m, -todrop); 2329 tlen -= todrop; 2330 tiflags &= ~(TH_PUSH|TH_FIN); 2331 } 2332 2333 /* 2334 * If last ACK falls within this segment's sequence numbers, 2335 * record the timestamp. 2336 * NOTE: 2337 * 1) That the test incorporates suggestions from the latest 2338 * proposal of the tcplw@cray.com list (Braden 1993/04/26). 2339 * 2) That updating only on newer timestamps interferes with 2340 * our earlier PAWS tests, so this check should be solely 2341 * predicated on the sequence space of this segment. 2342 * 3) That we modify the segment boundary check to be 2343 * Last.ACK.Sent <= SEG.SEQ + SEG.Len 2344 * instead of RFC1323's 2345 * Last.ACK.Sent < SEG.SEQ + SEG.Len, 2346 * This modified check allows us to overcome RFC1323's 2347 * limitations as described in Stevens TCP/IP Illustrated 2348 * Vol. 2 p.869. In such cases, we can still calculate the 2349 * RTT correctly when RCV.NXT == Last.ACK.Sent. 2350 */ 2351 if (opti.ts_present && 2352 SEQ_LEQ(th->th_seq, tp->last_ack_sent) && 2353 SEQ_LEQ(tp->last_ack_sent, th->th_seq + tlen + 2354 ((tiflags & (TH_SYN|TH_FIN)) != 0))) { 2355 tp->ts_recent_age = tcp_now; 2356 tp->ts_recent = opti.ts_val; 2357 } 2358 2359 /* 2360 * If the RST bit is set examine the state: 2361 * RECEIVED state: 2362 * If passive open, return to LISTEN state. 2363 * If active open, inform user that connection was refused. 2364 * ESTABLISHED, FIN_WAIT_1, FIN_WAIT2, CLOSE_WAIT states: 2365 * Inform user that connection was reset, and close tcb. 2366 * CLOSING, LAST_ACK, TIME_WAIT states: 2367 * Close the tcb. 2368 */ 2369 if (tiflags & TH_RST) { 2370 if (th->th_seq != tp->rcv_nxt) 2371 goto dropafterack_ratelim; 2372 2373 switch (tp->t_state) { 2374 case TCPS_SYN_RECEIVED: 2375 so->so_error = ECONNREFUSED; 2376 goto close; 2377 2378 case TCPS_ESTABLISHED: 2379 case TCPS_FIN_WAIT_1: 2380 case TCPS_FIN_WAIT_2: 2381 case TCPS_CLOSE_WAIT: 2382 so->so_error = ECONNRESET; 2383 close: 2384 tp->t_state = TCPS_CLOSED; 2385 TCP_STATINC(TCP_STAT_DROPS); 2386 tp = tcp_close(tp); 2387 goto drop; 2388 2389 case TCPS_CLOSING: 2390 case TCPS_LAST_ACK: 2391 case TCPS_TIME_WAIT: 2392 tp = tcp_close(tp); 2393 goto drop; 2394 } 2395 } 2396 2397 /* 2398 * Since we've covered the SYN-SENT and SYN-RECEIVED states above 2399 * we must be in a synchronized state. RFC793 states (under Reset 2400 * Generation) that any unacceptable segment (an out-of-order SYN 2401 * qualifies) received in a synchronized state must elicit only an 2402 * empty acknowledgment segment ... and the connection remains in 2403 * the same state. 2404 */ 2405 if (tiflags & TH_SYN) { 2406 if (tp->rcv_nxt == th->th_seq) { 2407 tcp_respond(tp, m, m, th, (tcp_seq)0, th->th_ack - 1, 2408 TH_ACK); 2409 m_freem(tcp_saveti); 2410 return; 2411 } 2412 2413 goto dropafterack_ratelim; 2414 } 2415 2416 /* 2417 * If the ACK bit is off we drop the segment and return. 2418 */ 2419 if ((tiflags & TH_ACK) == 0) { 2420 if (tp->t_flags & TF_ACKNOW) 2421 goto dropafterack; 2422 goto drop; 2423 } 2424 2425 /* 2426 * From here on, we're doing ACK processing. 2427 */ 2428 2429 switch (tp->t_state) { 2430 /* 2431 * In SYN_RECEIVED state if the ack ACKs our SYN then enter 2432 * ESTABLISHED state and continue processing, otherwise 2433 * send an RST. 2434 */ 2435 case TCPS_SYN_RECEIVED: 2436 if (SEQ_GT(tp->snd_una, th->th_ack) || 2437 SEQ_GT(th->th_ack, tp->snd_max)) 2438 goto dropwithreset; 2439 TCP_STATINC(TCP_STAT_CONNECTS); 2440 soisconnected(so); 2441 tcp_established(tp); 2442 /* Do window scaling? */ 2443 if ((tp->t_flags & (TF_RCVD_SCALE|TF_REQ_SCALE)) == 2444 (TF_RCVD_SCALE|TF_REQ_SCALE)) { 2445 tp->snd_scale = tp->requested_s_scale; 2446 tp->rcv_scale = tp->request_r_scale; 2447 } 2448 TCP_REASS_LOCK(tp); 2449 (void)tcp_reass(tp, NULL, NULL, tlen); 2450 tp->snd_wl1 = th->th_seq - 1; 2451 /* FALLTHROUGH */ 2452 2453 /* 2454 * In ESTABLISHED state: drop duplicate ACKs; ACK out of range 2455 * ACKs. If the ack is in the range 2456 * tp->snd_una < th->th_ack <= tp->snd_max 2457 * then advance tp->snd_una to th->th_ack and drop 2458 * data from the retransmission queue. If this ACK reflects 2459 * more up to date window information we update our window information. 2460 */ 2461 case TCPS_ESTABLISHED: 2462 case TCPS_FIN_WAIT_1: 2463 case TCPS_FIN_WAIT_2: 2464 case TCPS_CLOSE_WAIT: 2465 case TCPS_CLOSING: 2466 case TCPS_LAST_ACK: 2467 case TCPS_TIME_WAIT: 2468 if (SEQ_LEQ(th->th_ack, tp->snd_una)) { 2469 if (tlen == 0 && !dupseg && tiwin == tp->snd_wnd) { 2470 TCP_STATINC(TCP_STAT_RCVDUPACK); 2471 /* 2472 * If we have outstanding data (other than 2473 * a window probe), this is a completely 2474 * duplicate ack (ie, window info didn't 2475 * change), the ack is the biggest we've 2476 * seen and we've seen exactly our rexmt 2477 * threshold of them, assume a packet 2478 * has been dropped and retransmit it. 2479 * Kludge snd_nxt & the congestion 2480 * window so we send only this one 2481 * packet. 2482 */ 2483 if (TCP_TIMER_ISARMED(tp, TCPT_REXMT) == 0 || 2484 th->th_ack != tp->snd_una) 2485 tp->t_dupacks = 0; 2486 else if (tp->t_partialacks < 0 && 2487 (++tp->t_dupacks == tcprexmtthresh || 2488 TCP_FACK_FASTRECOV(tp))) { 2489 /* 2490 * Do the fast retransmit, and adjust 2491 * congestion control parameters. 2492 */ 2493 if (tp->t_congctl->fast_retransmit(tp, th)) { 2494 /* False fast retransmit */ 2495 break; 2496 } 2497 goto drop; 2498 } else if (tp->t_dupacks > tcprexmtthresh) { 2499 tp->snd_cwnd += tp->t_segsz; 2500 KERNEL_LOCK(1, NULL); 2501 (void)tcp_output(tp); 2502 KERNEL_UNLOCK_ONE(NULL); 2503 goto drop; 2504 } 2505 } else { 2506 /* 2507 * If the ack appears to be very old, only 2508 * allow data that is in-sequence. This 2509 * makes it somewhat more difficult to insert 2510 * forged data by guessing sequence numbers. 2511 * Sent an ack to try to update the send 2512 * sequence number on the other side. 2513 */ 2514 if (tlen && th->th_seq != tp->rcv_nxt && 2515 SEQ_LT(th->th_ack, 2516 tp->snd_una - tp->max_sndwnd)) 2517 goto dropafterack; 2518 } 2519 break; 2520 } 2521 /* 2522 * If the congestion window was inflated to account 2523 * for the other side's cached packets, retract it. 2524 */ 2525 tp->t_congctl->fast_retransmit_newack(tp, th); 2526 2527 if (SEQ_GT(th->th_ack, tp->snd_max)) { 2528 TCP_STATINC(TCP_STAT_RCVACKTOOMUCH); 2529 goto dropafterack; 2530 } 2531 acked = th->th_ack - tp->snd_una; 2532 tcps = TCP_STAT_GETREF(); 2533 _NET_STATINC_REF(tcps, TCP_STAT_RCVACKPACK); 2534 _NET_STATADD_REF(tcps, TCP_STAT_RCVACKBYTE, acked); 2535 TCP_STAT_PUTREF(); 2536 2537 /* 2538 * If we have a timestamp reply, update smoothed 2539 * round trip time. If no timestamp is present but 2540 * transmit timer is running and timed sequence 2541 * number was acked, update smoothed round trip time. 2542 * Since we now have an rtt measurement, cancel the 2543 * timer backoff (cf., Phil Karn's retransmit alg.). 2544 * Recompute the initial retransmit timer. 2545 */ 2546 if (ts_rtt) 2547 tcp_xmit_timer(tp, ts_rtt - 1); 2548 else if (tp->t_rtttime && SEQ_GT(th->th_ack, tp->t_rtseq)) 2549 tcp_xmit_timer(tp, tcp_now - tp->t_rtttime); 2550 2551 /* 2552 * If all outstanding data is acked, stop retransmit 2553 * timer and remember to restart (more output or persist). 2554 * If there is more data to be acked, restart retransmit 2555 * timer, using current (possibly backed-off) value. 2556 */ 2557 if (th->th_ack == tp->snd_max) { 2558 TCP_TIMER_DISARM(tp, TCPT_REXMT); 2559 needoutput = 1; 2560 } else if (TCP_TIMER_ISARMED(tp, TCPT_PERSIST) == 0) 2561 TCP_TIMER_ARM(tp, TCPT_REXMT, tp->t_rxtcur); 2562 2563 /* 2564 * New data has been acked, adjust the congestion window. 2565 */ 2566 tp->t_congctl->newack(tp, th); 2567 2568 nd_hint(tp); 2569 if (acked > so->so_snd.sb_cc) { 2570 tp->snd_wnd -= so->so_snd.sb_cc; 2571 sbdrop(&so->so_snd, (int)so->so_snd.sb_cc); 2572 ourfinisacked = 1; 2573 } else { 2574 if (acked > (tp->t_lastoff - tp->t_inoff)) 2575 tp->t_lastm = NULL; 2576 sbdrop(&so->so_snd, acked); 2577 tp->t_lastoff -= acked; 2578 if (tp->snd_wnd > acked) 2579 tp->snd_wnd -= acked; 2580 else 2581 tp->snd_wnd = 0; 2582 ourfinisacked = 0; 2583 } 2584 sowwakeup(so); 2585 2586 icmp_check(tp, th, acked); 2587 2588 tp->snd_una = th->th_ack; 2589 if (SEQ_GT(tp->snd_una, tp->snd_fack)) 2590 tp->snd_fack = tp->snd_una; 2591 if (SEQ_LT(tp->snd_nxt, tp->snd_una)) 2592 tp->snd_nxt = tp->snd_una; 2593 if (SEQ_LT(tp->snd_high, tp->snd_una)) 2594 tp->snd_high = tp->snd_una; 2595 2596 switch (tp->t_state) { 2597 2598 /* 2599 * In FIN_WAIT_1 STATE in addition to the processing 2600 * for the ESTABLISHED state if our FIN is now acknowledged 2601 * then enter FIN_WAIT_2. 2602 */ 2603 case TCPS_FIN_WAIT_1: 2604 if (ourfinisacked) { 2605 /* 2606 * If we can't receive any more 2607 * data, then closing user can proceed. 2608 * Starting the timer is contrary to the 2609 * specification, but if we don't get a FIN 2610 * we'll hang forever. 2611 */ 2612 if (so->so_state & SS_CANTRCVMORE) { 2613 soisdisconnected(so); 2614 if (tp->t_maxidle > 0) 2615 TCP_TIMER_ARM(tp, TCPT_2MSL, 2616 tp->t_maxidle); 2617 } 2618 tp->t_state = TCPS_FIN_WAIT_2; 2619 } 2620 break; 2621 2622 /* 2623 * In CLOSING STATE in addition to the processing for 2624 * the ESTABLISHED state if the ACK acknowledges our FIN 2625 * then enter the TIME-WAIT state, otherwise ignore 2626 * the segment. 2627 */ 2628 case TCPS_CLOSING: 2629 if (ourfinisacked) { 2630 tp->t_state = TCPS_TIME_WAIT; 2631 tcp_canceltimers(tp); 2632 TCP_TIMER_ARM(tp, TCPT_2MSL, 2 * tp->t_msl); 2633 soisdisconnected(so); 2634 } 2635 break; 2636 2637 /* 2638 * In LAST_ACK, we may still be waiting for data to drain 2639 * and/or to be acked, as well as for the ack of our FIN. 2640 * If our FIN is now acknowledged, delete the TCB, 2641 * enter the closed state and return. 2642 */ 2643 case TCPS_LAST_ACK: 2644 if (ourfinisacked) { 2645 tp = tcp_close(tp); 2646 goto drop; 2647 } 2648 break; 2649 2650 /* 2651 * In TIME_WAIT state the only thing that should arrive 2652 * is a retransmission of the remote FIN. Acknowledge 2653 * it and restart the finack timer. 2654 */ 2655 case TCPS_TIME_WAIT: 2656 TCP_TIMER_ARM(tp, TCPT_2MSL, 2 * tp->t_msl); 2657 goto dropafterack; 2658 } 2659 } 2660 2661 step6: 2662 /* 2663 * Update window information. 2664 * Don't look at window if no ACK: TAC's send garbage on first SYN. 2665 */ 2666 if ((tiflags & TH_ACK) && (SEQ_LT(tp->snd_wl1, th->th_seq) || 2667 (tp->snd_wl1 == th->th_seq && (SEQ_LT(tp->snd_wl2, th->th_ack) || 2668 (tp->snd_wl2 == th->th_ack && tiwin > tp->snd_wnd))))) { 2669 /* keep track of pure window updates */ 2670 if (tlen == 0 && 2671 tp->snd_wl2 == th->th_ack && tiwin > tp->snd_wnd) 2672 TCP_STATINC(TCP_STAT_RCVWINUPD); 2673 tp->snd_wnd = tiwin; 2674 tp->snd_wl1 = th->th_seq; 2675 tp->snd_wl2 = th->th_ack; 2676 if (tp->snd_wnd > tp->max_sndwnd) 2677 tp->max_sndwnd = tp->snd_wnd; 2678 needoutput = 1; 2679 } 2680 2681 /* 2682 * Process segments with URG. 2683 */ 2684 if ((tiflags & TH_URG) && th->th_urp && 2685 TCPS_HAVERCVDFIN(tp->t_state) == 0) { 2686 /* 2687 * This is a kludge, but if we receive and accept 2688 * random urgent pointers, we'll crash in 2689 * soreceive. It's hard to imagine someone 2690 * actually wanting to send this much urgent data. 2691 */ 2692 if (th->th_urp + so->so_rcv.sb_cc > sb_max) { 2693 th->th_urp = 0; /* XXX */ 2694 tiflags &= ~TH_URG; /* XXX */ 2695 goto dodata; /* XXX */ 2696 } 2697 2698 /* 2699 * If this segment advances the known urgent pointer, 2700 * then mark the data stream. This should not happen 2701 * in CLOSE_WAIT, CLOSING, LAST_ACK or TIME_WAIT STATES since 2702 * a FIN has been received from the remote side. 2703 * In these states we ignore the URG. 2704 * 2705 * According to RFC961 (Assigned Protocols), 2706 * the urgent pointer points to the last octet 2707 * of urgent data. We continue, however, 2708 * to consider it to indicate the first octet 2709 * of data past the urgent section as the original 2710 * spec states (in one of two places). 2711 */ 2712 if (SEQ_GT(th->th_seq+th->th_urp, tp->rcv_up)) { 2713 tp->rcv_up = th->th_seq + th->th_urp; 2714 so->so_oobmark = so->so_rcv.sb_cc + 2715 (tp->rcv_up - tp->rcv_nxt) - 1; 2716 if (so->so_oobmark == 0) 2717 so->so_state |= SS_RCVATMARK; 2718 sohasoutofband(so); 2719 tp->t_oobflags &= ~(TCPOOB_HAVEDATA | TCPOOB_HADDATA); 2720 } 2721 2722 /* 2723 * Remove out of band data so doesn't get presented to user. 2724 * This can happen independent of advancing the URG pointer, 2725 * but if two URG's are pending at once, some out-of-band 2726 * data may creep in... ick. 2727 */ 2728 if (th->th_urp <= (u_int16_t)tlen && 2729 (so->so_options & SO_OOBINLINE) == 0) 2730 tcp_pulloutofband(so, th, m, hdroptlen); 2731 } else { 2732 /* 2733 * If no out of band data is expected, 2734 * pull receive urgent pointer along 2735 * with the receive window. 2736 */ 2737 if (SEQ_GT(tp->rcv_nxt, tp->rcv_up)) 2738 tp->rcv_up = tp->rcv_nxt; 2739 } 2740 dodata: 2741 2742 /* 2743 * Process the segment text, merging it into the TCP sequencing queue, 2744 * and arranging for acknowledgement of receipt if necessary. 2745 * This process logically involves adjusting tp->rcv_wnd as data 2746 * is presented to the user (this happens in tcp_usrreq.c, 2747 * tcp_rcvd()). If a FIN has already been received on this 2748 * connection then we just ignore the text. 2749 */ 2750 if ((tlen || (tiflags & TH_FIN)) && 2751 TCPS_HAVERCVDFIN(tp->t_state) == 0) { 2752 /* 2753 * Handle the common case: 2754 * o Segment is the next to be received, and 2755 * o The queue is empty, and 2756 * o The connection is established 2757 * In this case, we avoid calling tcp_reass. 2758 * 2759 * tcp_setup_ack: set DELACK for segments received in order, 2760 * but ack immediately when segments are out of order (so that 2761 * fast retransmit can work). 2762 */ 2763 TCP_REASS_LOCK(tp); 2764 if (th->th_seq == tp->rcv_nxt && 2765 TAILQ_FIRST(&tp->segq) == NULL && 2766 tp->t_state == TCPS_ESTABLISHED) { 2767 tcp_setup_ack(tp, th); 2768 tp->rcv_nxt += tlen; 2769 tiflags = th->th_flags & TH_FIN; 2770 tcps = TCP_STAT_GETREF(); 2771 _NET_STATINC_REF(tcps, TCP_STAT_RCVPACK); 2772 _NET_STATADD_REF(tcps, TCP_STAT_RCVBYTE, tlen); 2773 TCP_STAT_PUTREF(); 2774 nd_hint(tp); 2775 if (so->so_state & SS_CANTRCVMORE) { 2776 m_freem(m); 2777 } else { 2778 m_adj(m, hdroptlen); 2779 sbappendstream(&(so)->so_rcv, m); 2780 } 2781 TCP_REASS_UNLOCK(tp); 2782 sorwakeup(so); 2783 } else { 2784 m_adj(m, hdroptlen); 2785 tiflags = tcp_reass(tp, th, m, tlen); 2786 tp->t_flags |= TF_ACKNOW; 2787 } 2788 2789 /* 2790 * Note the amount of data that peer has sent into 2791 * our window, in order to estimate the sender's 2792 * buffer size. 2793 */ 2794 len = so->so_rcv.sb_hiwat - (tp->rcv_adv - tp->rcv_nxt); 2795 } else { 2796 m_freem(m); 2797 m = NULL; 2798 tiflags &= ~TH_FIN; 2799 } 2800 2801 /* 2802 * If FIN is received ACK the FIN and let the user know 2803 * that the connection is closing. Ignore a FIN received before 2804 * the connection is fully established. 2805 */ 2806 if ((tiflags & TH_FIN) && TCPS_HAVEESTABLISHED(tp->t_state)) { 2807 if (TCPS_HAVERCVDFIN(tp->t_state) == 0) { 2808 socantrcvmore(so); 2809 tp->t_flags |= TF_ACKNOW; 2810 tp->rcv_nxt++; 2811 } 2812 switch (tp->t_state) { 2813 2814 /* 2815 * In ESTABLISHED STATE enter the CLOSE_WAIT state. 2816 */ 2817 case TCPS_ESTABLISHED: 2818 tp->t_state = TCPS_CLOSE_WAIT; 2819 break; 2820 2821 /* 2822 * If still in FIN_WAIT_1 STATE FIN has not been acked so 2823 * enter the CLOSING state. 2824 */ 2825 case TCPS_FIN_WAIT_1: 2826 tp->t_state = TCPS_CLOSING; 2827 break; 2828 2829 /* 2830 * In FIN_WAIT_2 state enter the TIME_WAIT state, 2831 * starting the time-wait timer, turning off the other 2832 * standard timers. 2833 */ 2834 case TCPS_FIN_WAIT_2: 2835 tp->t_state = TCPS_TIME_WAIT; 2836 tcp_canceltimers(tp); 2837 TCP_TIMER_ARM(tp, TCPT_2MSL, 2 * tp->t_msl); 2838 soisdisconnected(so); 2839 break; 2840 2841 /* 2842 * In TIME_WAIT state restart the 2 MSL time_wait timer. 2843 */ 2844 case TCPS_TIME_WAIT: 2845 TCP_TIMER_ARM(tp, TCPT_2MSL, 2 * tp->t_msl); 2846 break; 2847 } 2848 } 2849 #ifdef TCP_DEBUG 2850 if (so->so_options & SO_DEBUG) 2851 tcp_trace(TA_INPUT, ostate, tp, tcp_saveti, 0); 2852 #endif 2853 2854 /* 2855 * Return any desired output. 2856 */ 2857 if (needoutput || (tp->t_flags & TF_ACKNOW)) { 2858 KERNEL_LOCK(1, NULL); 2859 (void)tcp_output(tp); 2860 KERNEL_UNLOCK_ONE(NULL); 2861 } 2862 m_freem(tcp_saveti); 2863 2864 if (tp->t_state == TCPS_TIME_WAIT 2865 && (so->so_state & SS_NOFDREF) 2866 && (tp->t_inpcb || af != AF_INET || af != AF_INET6) 2867 && ((af == AF_INET ? tcp4_vtw_enable : tcp6_vtw_enable) & 1) != 0 2868 && TAILQ_EMPTY(&tp->segq) 2869 && vtw_add(af, tp)) { 2870 ; 2871 } 2872 return; 2873 2874 badsyn: 2875 /* 2876 * Received a bad SYN. Increment counters and dropwithreset. 2877 */ 2878 TCP_STATINC(TCP_STAT_BADSYN); 2879 tp = NULL; 2880 goto dropwithreset; 2881 2882 dropafterack: 2883 /* 2884 * Generate an ACK dropping incoming segment if it occupies 2885 * sequence space, where the ACK reflects our state. 2886 */ 2887 if (tiflags & TH_RST) 2888 goto drop; 2889 goto dropafterack2; 2890 2891 dropafterack_ratelim: 2892 /* 2893 * We may want to rate-limit ACKs against SYN/RST attack. 2894 */ 2895 if (ppsratecheck(&tcp_ackdrop_ppslim_last, &tcp_ackdrop_ppslim_count, 2896 tcp_ackdrop_ppslim) == 0) { 2897 /* XXX stat */ 2898 goto drop; 2899 } 2900 2901 dropafterack2: 2902 m_freem(m); 2903 tp->t_flags |= TF_ACKNOW; 2904 KERNEL_LOCK(1, NULL); 2905 (void)tcp_output(tp); 2906 KERNEL_UNLOCK_ONE(NULL); 2907 m_freem(tcp_saveti); 2908 return; 2909 2910 dropwithreset_ratelim: 2911 /* 2912 * We may want to rate-limit RSTs in certain situations, 2913 * particularly if we are sending an RST in response to 2914 * an attempt to connect to or otherwise communicate with 2915 * a port for which we have no socket. 2916 */ 2917 if (ppsratecheck(&tcp_rst_ppslim_last, &tcp_rst_ppslim_count, 2918 tcp_rst_ppslim) == 0) { 2919 /* XXX stat */ 2920 goto drop; 2921 } 2922 2923 dropwithreset: 2924 /* 2925 * Generate a RST, dropping incoming segment. 2926 * Make ACK acceptable to originator of segment. 2927 */ 2928 if (tiflags & TH_RST) 2929 goto drop; 2930 if (tiflags & TH_ACK) { 2931 (void)tcp_respond(tp, m, m, th, (tcp_seq)0, th->th_ack, TH_RST); 2932 } else { 2933 if (tiflags & TH_SYN) 2934 tlen++; 2935 (void)tcp_respond(tp, m, m, th, th->th_seq + tlen, (tcp_seq)0, 2936 TH_RST|TH_ACK); 2937 } 2938 m_freem(tcp_saveti); 2939 return; 2940 2941 badcsum: 2942 drop: 2943 /* 2944 * Drop space held by incoming segment and return. 2945 */ 2946 if (tp) { 2947 so = tp->t_inpcb->inp_socket; 2948 #ifdef TCP_DEBUG 2949 if (so && (so->so_options & SO_DEBUG) != 0) 2950 tcp_trace(TA_DROP, ostate, tp, tcp_saveti, 0); 2951 #endif 2952 } 2953 m_freem(tcp_saveti); 2954 m_freem(m); 2955 return; 2956 } 2957 2958 #ifdef TCP_SIGNATURE 2959 int 2960 tcp_signature_apply(void *fstate, void *data, u_int len) 2961 { 2962 2963 MD5Update(fstate, (u_char *)data, len); 2964 return (0); 2965 } 2966 2967 struct secasvar * 2968 tcp_signature_getsav(struct mbuf *m) 2969 { 2970 struct ip *ip; 2971 struct ip6_hdr *ip6; 2972 2973 ip = mtod(m, struct ip *); 2974 switch (ip->ip_v) { 2975 case 4: 2976 ip = mtod(m, struct ip *); 2977 ip6 = NULL; 2978 break; 2979 case 6: 2980 ip = NULL; 2981 ip6 = mtod(m, struct ip6_hdr *); 2982 break; 2983 default: 2984 return (NULL); 2985 } 2986 2987 #ifdef IPSEC 2988 union sockaddr_union dst; 2989 2990 /* Extract the destination from the IP header in the mbuf. */ 2991 memset(&dst, 0, sizeof(union sockaddr_union)); 2992 if (ip != NULL) { 2993 dst.sa.sa_len = sizeof(struct sockaddr_in); 2994 dst.sa.sa_family = AF_INET; 2995 dst.sin.sin_addr = ip->ip_dst; 2996 } else { 2997 dst.sa.sa_len = sizeof(struct sockaddr_in6); 2998 dst.sa.sa_family = AF_INET6; 2999 dst.sin6.sin6_addr = ip6->ip6_dst; 3000 } 3001 3002 /* 3003 * Look up an SADB entry which matches the address of the peer. 3004 */ 3005 return KEY_LOOKUP_SA(&dst, IPPROTO_TCP, htonl(TCP_SIG_SPI), 0, 0); 3006 #else 3007 return NULL; 3008 #endif 3009 } 3010 3011 int 3012 tcp_signature(struct mbuf *m, struct tcphdr *th, int thoff, 3013 struct secasvar *sav, char *sig) 3014 { 3015 MD5_CTX ctx; 3016 struct ip *ip; 3017 struct ipovly *ipovly; 3018 #ifdef INET6 3019 struct ip6_hdr *ip6; 3020 struct ip6_hdr_pseudo ip6pseudo; 3021 #endif 3022 struct ippseudo ippseudo; 3023 struct tcphdr th0; 3024 int l, tcphdrlen; 3025 3026 if (sav == NULL) 3027 return (-1); 3028 3029 tcphdrlen = th->th_off * 4; 3030 3031 switch (mtod(m, struct ip *)->ip_v) { 3032 case 4: 3033 MD5Init(&ctx); 3034 ip = mtod(m, struct ip *); 3035 memset(&ippseudo, 0, sizeof(ippseudo)); 3036 ipovly = (struct ipovly *)ip; 3037 ippseudo.ippseudo_src = ipovly->ih_src; 3038 ippseudo.ippseudo_dst = ipovly->ih_dst; 3039 ippseudo.ippseudo_pad = 0; 3040 ippseudo.ippseudo_p = IPPROTO_TCP; 3041 ippseudo.ippseudo_len = htons(m->m_pkthdr.len - thoff); 3042 MD5Update(&ctx, (char *)&ippseudo, sizeof(ippseudo)); 3043 break; 3044 #if INET6 3045 case 6: 3046 MD5Init(&ctx); 3047 ip6 = mtod(m, struct ip6_hdr *); 3048 memset(&ip6pseudo, 0, sizeof(ip6pseudo)); 3049 ip6pseudo.ip6ph_src = ip6->ip6_src; 3050 in6_clearscope(&ip6pseudo.ip6ph_src); 3051 ip6pseudo.ip6ph_dst = ip6->ip6_dst; 3052 in6_clearscope(&ip6pseudo.ip6ph_dst); 3053 ip6pseudo.ip6ph_len = htons(m->m_pkthdr.len - thoff); 3054 ip6pseudo.ip6ph_nxt = IPPROTO_TCP; 3055 MD5Update(&ctx, (char *)&ip6pseudo, sizeof(ip6pseudo)); 3056 break; 3057 #endif 3058 default: 3059 return (-1); 3060 } 3061 3062 th0 = *th; 3063 th0.th_sum = 0; 3064 MD5Update(&ctx, (char *)&th0, sizeof(th0)); 3065 3066 l = m->m_pkthdr.len - thoff - tcphdrlen; 3067 if (l > 0) 3068 m_apply(m, thoff + tcphdrlen, 3069 m->m_pkthdr.len - thoff - tcphdrlen, 3070 tcp_signature_apply, &ctx); 3071 3072 MD5Update(&ctx, _KEYBUF(sav->key_auth), _KEYLEN(sav->key_auth)); 3073 MD5Final(sig, &ctx); 3074 3075 return (0); 3076 } 3077 #endif 3078 3079 /* 3080 * Parse and process tcp options. 3081 * 3082 * Returns -1 if this segment should be dropped. (eg. wrong signature) 3083 * Otherwise returns 0. 3084 */ 3085 int 3086 tcp_dooptions(struct tcpcb *tp, const u_char *cp, int cnt, struct tcphdr *th, 3087 struct mbuf *m, int toff, struct tcp_opt_info *oi) 3088 { 3089 u_int16_t mss; 3090 int opt, optlen = 0; 3091 #ifdef TCP_SIGNATURE 3092 void *sigp = NULL; 3093 char sigbuf[TCP_SIGLEN]; 3094 struct secasvar *sav = NULL; 3095 #endif 3096 3097 for (; cp && cnt > 0; cnt -= optlen, cp += optlen) { 3098 opt = cp[0]; 3099 if (opt == TCPOPT_EOL) 3100 break; 3101 if (opt == TCPOPT_NOP) 3102 optlen = 1; 3103 else { 3104 if (cnt < 2) 3105 break; 3106 optlen = cp[1]; 3107 if (optlen < 2 || optlen > cnt) 3108 break; 3109 } 3110 switch (opt) { 3111 3112 default: 3113 continue; 3114 3115 case TCPOPT_MAXSEG: 3116 if (optlen != TCPOLEN_MAXSEG) 3117 continue; 3118 if (!(th->th_flags & TH_SYN)) 3119 continue; 3120 if (TCPS_HAVERCVDSYN(tp->t_state)) 3121 continue; 3122 memcpy(&mss, cp + 2, sizeof(mss)); 3123 oi->maxseg = ntohs(mss); 3124 break; 3125 3126 case TCPOPT_WINDOW: 3127 if (optlen != TCPOLEN_WINDOW) 3128 continue; 3129 if (!(th->th_flags & TH_SYN)) 3130 continue; 3131 if (TCPS_HAVERCVDSYN(tp->t_state)) 3132 continue; 3133 tp->t_flags |= TF_RCVD_SCALE; 3134 tp->requested_s_scale = cp[2]; 3135 if (tp->requested_s_scale > TCP_MAX_WINSHIFT) { 3136 char buf[INET6_ADDRSTRLEN]; 3137 struct ip *ip = mtod(m, struct ip *); 3138 #ifdef INET6 3139 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); 3140 #endif 3141 3142 switch (ip->ip_v) { 3143 case 4: 3144 in_print(buf, sizeof(buf), 3145 &ip->ip_src); 3146 break; 3147 #ifdef INET6 3148 case 6: 3149 in6_print(buf, sizeof(buf), 3150 &ip6->ip6_src); 3151 break; 3152 #endif 3153 default: 3154 strlcpy(buf, "(unknown)", sizeof(buf)); 3155 break; 3156 } 3157 3158 log(LOG_ERR, "TCP: invalid wscale %d from %s, " 3159 "assuming %d\n", 3160 tp->requested_s_scale, buf, 3161 TCP_MAX_WINSHIFT); 3162 tp->requested_s_scale = TCP_MAX_WINSHIFT; 3163 } 3164 break; 3165 3166 case TCPOPT_TIMESTAMP: 3167 if (optlen != TCPOLEN_TIMESTAMP) 3168 continue; 3169 oi->ts_present = 1; 3170 memcpy(&oi->ts_val, cp + 2, sizeof(oi->ts_val)); 3171 NTOHL(oi->ts_val); 3172 memcpy(&oi->ts_ecr, cp + 6, sizeof(oi->ts_ecr)); 3173 NTOHL(oi->ts_ecr); 3174 3175 if (!(th->th_flags & TH_SYN)) 3176 continue; 3177 if (TCPS_HAVERCVDSYN(tp->t_state)) 3178 continue; 3179 /* 3180 * A timestamp received in a SYN makes 3181 * it ok to send timestamp requests and replies. 3182 */ 3183 tp->t_flags |= TF_RCVD_TSTMP; 3184 tp->ts_recent = oi->ts_val; 3185 tp->ts_recent_age = tcp_now; 3186 break; 3187 3188 case TCPOPT_SACK_PERMITTED: 3189 if (optlen != TCPOLEN_SACK_PERMITTED) 3190 continue; 3191 if (!(th->th_flags & TH_SYN)) 3192 continue; 3193 if (TCPS_HAVERCVDSYN(tp->t_state)) 3194 continue; 3195 if (tcp_do_sack) { 3196 tp->t_flags |= TF_SACK_PERMIT; 3197 tp->t_flags |= TF_WILL_SACK; 3198 } 3199 break; 3200 3201 case TCPOPT_SACK: 3202 tcp_sack_option(tp, th, cp, optlen); 3203 break; 3204 #ifdef TCP_SIGNATURE 3205 case TCPOPT_SIGNATURE: 3206 if (optlen != TCPOLEN_SIGNATURE) 3207 continue; 3208 if (sigp && 3209 !consttime_memequal(sigp, cp + 2, TCP_SIGLEN)) 3210 return (-1); 3211 3212 sigp = sigbuf; 3213 memcpy(sigbuf, cp + 2, TCP_SIGLEN); 3214 tp->t_flags |= TF_SIGNATURE; 3215 break; 3216 #endif 3217 } 3218 } 3219 3220 #ifndef TCP_SIGNATURE 3221 return 0; 3222 #else 3223 if (tp->t_flags & TF_SIGNATURE) { 3224 sav = tcp_signature_getsav(m); 3225 if (sav == NULL && tp->t_state == TCPS_LISTEN) 3226 return (-1); 3227 } 3228 3229 if ((sigp ? TF_SIGNATURE : 0) ^ (tp->t_flags & TF_SIGNATURE)) 3230 goto out; 3231 3232 if (sigp) { 3233 char sig[TCP_SIGLEN]; 3234 3235 tcp_fields_to_net(th); 3236 if (tcp_signature(m, th, toff, sav, sig) < 0) { 3237 tcp_fields_to_host(th); 3238 goto out; 3239 } 3240 tcp_fields_to_host(th); 3241 3242 if (!consttime_memequal(sig, sigp, TCP_SIGLEN)) { 3243 TCP_STATINC(TCP_STAT_BADSIG); 3244 goto out; 3245 } else 3246 TCP_STATINC(TCP_STAT_GOODSIG); 3247 3248 key_sa_recordxfer(sav, m); 3249 KEY_SA_UNREF(&sav); 3250 } 3251 return 0; 3252 out: 3253 if (sav != NULL) 3254 KEY_SA_UNREF(&sav); 3255 return -1; 3256 #endif 3257 } 3258 3259 /* 3260 * Pull out of band byte out of a segment so 3261 * it doesn't appear in the user's data queue. 3262 * It is still reflected in the segment length for 3263 * sequencing purposes. 3264 */ 3265 void 3266 tcp_pulloutofband(struct socket *so, struct tcphdr *th, 3267 struct mbuf *m, int off) 3268 { 3269 int cnt = off + th->th_urp - 1; 3270 3271 while (cnt >= 0) { 3272 if (m->m_len > cnt) { 3273 char *cp = mtod(m, char *) + cnt; 3274 struct tcpcb *tp = sototcpcb(so); 3275 3276 tp->t_iobc = *cp; 3277 tp->t_oobflags |= TCPOOB_HAVEDATA; 3278 memmove(cp, cp + 1, (unsigned)(m->m_len - cnt - 1)); 3279 m->m_len--; 3280 return; 3281 } 3282 cnt -= m->m_len; 3283 m = m->m_next; 3284 if (m == NULL) 3285 break; 3286 } 3287 panic("tcp_pulloutofband"); 3288 } 3289 3290 /* 3291 * Collect new round-trip time estimate 3292 * and update averages and current timeout. 3293 * 3294 * rtt is in units of slow ticks (typically 500 ms) -- essentially the 3295 * difference of two timestamps. 3296 */ 3297 void 3298 tcp_xmit_timer(struct tcpcb *tp, uint32_t rtt) 3299 { 3300 int32_t delta; 3301 3302 TCP_STATINC(TCP_STAT_RTTUPDATED); 3303 if (tp->t_srtt != 0) { 3304 /* 3305 * Compute the amount to add to srtt for smoothing, 3306 * *alpha, or 2^(-TCP_RTT_SHIFT). Because 3307 * srtt is stored in 1/32 slow ticks, we conceptually 3308 * shift left 5 bits, subtract srtt to get the 3309 * difference, and then shift right by TCP_RTT_SHIFT 3310 * (3) to obtain 1/8 of the difference. 3311 */ 3312 delta = (rtt << 2) - (tp->t_srtt >> TCP_RTT_SHIFT); 3313 /* 3314 * This can never happen, because delta's lowest 3315 * possible value is 1/8 of t_srtt. But if it does, 3316 * set srtt to some reasonable value, here chosen 3317 * as 1/8 tick. 3318 */ 3319 if ((tp->t_srtt += delta) <= 0) 3320 tp->t_srtt = 1 << 2; 3321 /* 3322 * RFC2988 requires that rttvar be updated first. 3323 * This code is compliant because "delta" is the old 3324 * srtt minus the new observation (scaled). 3325 * 3326 * RFC2988 says: 3327 * rttvar = (1-beta) * rttvar + beta * |srtt-observed| 3328 * 3329 * delta is in units of 1/32 ticks, and has then been 3330 * divided by 8. This is equivalent to being in 1/16s 3331 * units and divided by 4. Subtract from it 1/4 of 3332 * the existing rttvar to form the (signed) amount to 3333 * adjust. 3334 */ 3335 if (delta < 0) 3336 delta = -delta; 3337 delta -= (tp->t_rttvar >> TCP_RTTVAR_SHIFT); 3338 /* 3339 * As with srtt, this should never happen. There is 3340 * no support in RFC2988 for this operation. But 1/4s 3341 * as rttvar when faced with something arguably wrong 3342 * is ok. 3343 */ 3344 if ((tp->t_rttvar += delta) <= 0) 3345 tp->t_rttvar = 1 << 2; 3346 3347 /* 3348 * If srtt exceeds .01 second, ensure we use the 'remote' MSL 3349 * Problem is: it doesn't work. Disabled by defaulting 3350 * tcp_rttlocal to 0; see corresponding code in 3351 * tcp_subr that selects local vs remote in a different way. 3352 * 3353 * The static branch prediction hint here should be removed 3354 * when the rtt estimator is fixed and the rtt_enable code 3355 * is turned back on. 3356 */ 3357 if (__predict_false(tcp_rttlocal) && tcp_msl_enable 3358 && tp->t_srtt > tcp_msl_remote_threshold 3359 && tp->t_msl < tcp_msl_remote) { 3360 tp->t_msl = MIN(tcp_msl_remote, TCP_MAXMSL); 3361 } 3362 } else { 3363 /* 3364 * This is the first measurement. Per RFC2988, 2.2, 3365 * set rtt=R and srtt=R/2. 3366 * For srtt, storage representation is 1/32 ticks, 3367 * so shift left by 5. 3368 * For rttvar, storage representation is 1/16 ticks, 3369 * So shift left by 4, but then right by 1 to halve. 3370 */ 3371 tp->t_srtt = rtt << (TCP_RTT_SHIFT + 2); 3372 tp->t_rttvar = rtt << (TCP_RTTVAR_SHIFT + 2 - 1); 3373 } 3374 tp->t_rtttime = 0; 3375 tp->t_rxtshift = 0; 3376 3377 /* 3378 * the retransmit should happen at rtt + 4 * rttvar. 3379 * Because of the way we do the smoothing, srtt and rttvar 3380 * will each average +1/2 tick of bias. When we compute 3381 * the retransmit timer, we want 1/2 tick of rounding and 3382 * 1 extra tick because of +-1/2 tick uncertainty in the 3383 * firing of the timer. The bias will give us exactly the 3384 * 1.5 tick we need. But, because the bias is 3385 * statistical, we have to test that we don't drop below 3386 * the minimum feasible timer (which is 2 ticks). 3387 */ 3388 TCPT_RANGESET(tp->t_rxtcur, TCP_REXMTVAL(tp), 3389 uimax(tp->t_rttmin, rtt + 2), TCPTV_REXMTMAX); 3390 3391 /* 3392 * We received an ack for a packet that wasn't retransmitted; 3393 * it is probably safe to discard any error indications we've 3394 * received recently. This isn't quite right, but close enough 3395 * for now (a route might have failed after we sent a segment, 3396 * and the return path might not be symmetrical). 3397 */ 3398 tp->t_softerror = 0; 3399 } 3400