xref: /netbsd-src/sys/netinet/tcp_subr.c (revision 946379e7b37692fc43f68eb0d1c10daa0a7f3b6c)
1 /*	$NetBSD: tcp_subr.c,v 1.264 2015/09/07 01:56:50 ozaki-r 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 (c) 1997, 1998, 2000, 2001, 2008 The NetBSD Foundation, Inc.
34  * All rights reserved.
35  *
36  * This code is derived from software contributed to The NetBSD Foundation
37  * by Jason R. Thorpe and Kevin M. Lahey of the Numerical Aerospace Simulation
38  * Facility, NASA Ames Research Center.
39  *
40  * Redistribution and use in source and binary forms, with or without
41  * modification, are permitted provided that the following conditions
42  * are met:
43  * 1. Redistributions of source code must retain the above copyright
44  *    notice, this list of conditions and the following disclaimer.
45  * 2. Redistributions in binary form must reproduce the above copyright
46  *    notice, this list of conditions and the following disclaimer in the
47  *    documentation and/or other materials provided with the distribution.
48  *
49  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
50  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
51  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
52  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
53  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
54  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
55  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
56  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
57  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
58  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
59  * POSSIBILITY OF SUCH DAMAGE.
60  */
61 
62 /*
63  * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995
64  *	The Regents of the University of California.  All rights reserved.
65  *
66  * Redistribution and use in source and binary forms, with or without
67  * modification, are permitted provided that the following conditions
68  * are met:
69  * 1. Redistributions of source code must retain the above copyright
70  *    notice, this list of conditions and the following disclaimer.
71  * 2. Redistributions in binary form must reproduce the above copyright
72  *    notice, this list of conditions and the following disclaimer in the
73  *    documentation and/or other materials provided with the distribution.
74  * 3. Neither the name of the University nor the names of its contributors
75  *    may be used to endorse or promote products derived from this software
76  *    without specific prior written permission.
77  *
78  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
79  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
80  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
81  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
82  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
83  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
84  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
85  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
86  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
87  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
88  * SUCH DAMAGE.
89  *
90  *	@(#)tcp_subr.c	8.2 (Berkeley) 5/24/95
91  */
92 
93 #include <sys/cdefs.h>
94 __KERNEL_RCSID(0, "$NetBSD: tcp_subr.c,v 1.264 2015/09/07 01:56:50 ozaki-r Exp $");
95 
96 #ifdef _KERNEL_OPT
97 #include "opt_inet.h"
98 #include "opt_ipsec.h"
99 #include "opt_tcp_compat_42.h"
100 #include "opt_inet_csum.h"
101 #include "opt_mbuftrace.h"
102 #endif
103 
104 #include <sys/param.h>
105 #include <sys/atomic.h>
106 #include <sys/proc.h>
107 #include <sys/systm.h>
108 #include <sys/mbuf.h>
109 #include <sys/once.h>
110 #include <sys/socket.h>
111 #include <sys/socketvar.h>
112 #include <sys/protosw.h>
113 #include <sys/errno.h>
114 #include <sys/kernel.h>
115 #include <sys/pool.h>
116 #include <sys/md5.h>
117 #include <sys/cprng.h>
118 
119 #include <net/route.h>
120 #include <net/if.h>
121 
122 #include <netinet/in.h>
123 #include <netinet/in_systm.h>
124 #include <netinet/ip.h>
125 #include <netinet/in_pcb.h>
126 #include <netinet/ip_var.h>
127 #include <netinet/ip_icmp.h>
128 
129 #ifdef INET6
130 #ifndef INET
131 #include <netinet/in.h>
132 #endif
133 #include <netinet/ip6.h>
134 #include <netinet6/in6_pcb.h>
135 #include <netinet6/ip6_var.h>
136 #include <netinet6/in6_var.h>
137 #include <netinet6/ip6protosw.h>
138 #include <netinet/icmp6.h>
139 #include <netinet6/nd6.h>
140 #endif
141 
142 #include <netinet/tcp.h>
143 #include <netinet/tcp_fsm.h>
144 #include <netinet/tcp_seq.h>
145 #include <netinet/tcp_timer.h>
146 #include <netinet/tcp_var.h>
147 #include <netinet/tcp_vtw.h>
148 #include <netinet/tcp_private.h>
149 #include <netinet/tcp_congctl.h>
150 #include <netinet/tcpip.h>
151 
152 #ifdef IPSEC
153 #include <netipsec/ipsec.h>
154 #include <netipsec/xform.h>
155 #ifdef INET6
156 #include <netipsec/ipsec6.h>
157 #endif
158  #include <netipsec/key.h>
159 #endif	/* IPSEC*/
160 
161 
162 struct	inpcbtable tcbtable;	/* head of queue of active tcpcb's */
163 u_int32_t tcp_now;		/* slow ticks, for RFC 1323 timestamps */
164 
165 percpu_t *tcpstat_percpu;
166 
167 /* patchable/settable parameters for tcp */
168 int 	tcp_mssdflt = TCP_MSS;
169 int	tcp_minmss = TCP_MINMSS;
170 int 	tcp_rttdflt = TCPTV_SRTTDFLT / PR_SLOWHZ;
171 int	tcp_do_rfc1323 = 1;	/* window scaling / timestamps (obsolete) */
172 int	tcp_do_rfc1948 = 0;	/* ISS by cryptographic hash */
173 int	tcp_do_sack = 1;	/* selective acknowledgement */
174 int	tcp_do_win_scale = 1;	/* RFC1323 window scaling */
175 int	tcp_do_timestamps = 1;	/* RFC1323 timestamps */
176 int	tcp_ack_on_push = 0;	/* set to enable immediate ACK-on-PUSH */
177 int	tcp_do_ecn = 0;		/* Explicit Congestion Notification */
178 #ifndef TCP_INIT_WIN
179 #define	TCP_INIT_WIN	4	/* initial slow start window */
180 #endif
181 #ifndef TCP_INIT_WIN_LOCAL
182 #define	TCP_INIT_WIN_LOCAL 4	/* initial slow start window for local nets */
183 #endif
184 /*
185  * Up to 5 we scale linearly, to reach 3 * 1460; then (iw) * 1460.
186  * This is to simulate current behavior for iw == 4
187  */
188 int tcp_init_win_max[] = {
189 	 1 * 1460,
190 	 1 * 1460,
191 	 2 * 1460,
192 	 2 * 1460,
193 	 3 * 1460,
194 	 5 * 1460,
195 	 6 * 1460,
196 	 7 * 1460,
197 	 8 * 1460,
198 	 9 * 1460,
199 	10 * 1460
200 };
201 int	tcp_init_win = TCP_INIT_WIN;
202 int	tcp_init_win_local = TCP_INIT_WIN_LOCAL;
203 int	tcp_mss_ifmtu = 0;
204 #ifdef TCP_COMPAT_42
205 int	tcp_compat_42 = 1;
206 #else
207 int	tcp_compat_42 = 0;
208 #endif
209 int	tcp_rst_ppslim = 100;	/* 100pps */
210 int	tcp_ackdrop_ppslim = 100;	/* 100pps */
211 int	tcp_do_loopback_cksum = 0;
212 int	tcp_do_abc = 1;		/* RFC3465 Appropriate byte counting. */
213 int	tcp_abc_aggressive = 1;	/* 1: L=2*SMSS  0: L=1*SMSS */
214 int	tcp_sack_tp_maxholes = 32;
215 int	tcp_sack_globalmaxholes = 1024;
216 int	tcp_sack_globalholes = 0;
217 int	tcp_ecn_maxretries = 1;
218 int	tcp_msl_enable = 1;		/* enable TIME_WAIT truncation	*/
219 int	tcp_msl_loop   = PR_SLOWHZ;	/* MSL for loopback		*/
220 int	tcp_msl_local  = 5 * PR_SLOWHZ;	/* MSL for 'local'		*/
221 int	tcp_msl_remote = TCPTV_MSL;	/* MSL otherwise		*/
222 int	tcp_msl_remote_threshold = TCPTV_SRTTDFLT;	/* RTT threshold */
223 int	tcp_rttlocal = 0;		/* Use RTT to decide who's 'local' */
224 
225 int	tcp4_vtw_enable = 0;		/* 1 to enable */
226 int	tcp6_vtw_enable = 0;		/* 1 to enable */
227 int	tcp_vtw_was_enabled = 0;
228 int	tcp_vtw_entries = 1 << 4;	/* 16 vestigial TIME_WAIT entries */
229 
230 /* tcb hash */
231 #ifndef TCBHASHSIZE
232 #define	TCBHASHSIZE	128
233 #endif
234 int	tcbhashsize = TCBHASHSIZE;
235 
236 /* syn hash parameters */
237 #define	TCP_SYN_HASH_SIZE	293
238 #define	TCP_SYN_BUCKET_SIZE	35
239 int	tcp_syn_cache_size = TCP_SYN_HASH_SIZE;
240 int	tcp_syn_cache_limit = TCP_SYN_HASH_SIZE*TCP_SYN_BUCKET_SIZE;
241 int	tcp_syn_bucket_limit = 3*TCP_SYN_BUCKET_SIZE;
242 struct	syn_cache_head tcp_syn_cache[TCP_SYN_HASH_SIZE];
243 
244 int	tcp_freeq(struct tcpcb *);
245 static int	tcp_iss_secret_init(void);
246 
247 #ifdef INET
248 static void	tcp_mtudisc_callback(struct in_addr);
249 #endif
250 
251 #ifdef INET6
252 void	tcp6_mtudisc(struct in6pcb *, int);
253 #endif
254 
255 static struct pool tcpcb_pool;
256 
257 static int tcp_drainwanted;
258 
259 #ifdef TCP_CSUM_COUNTERS
260 #include <sys/device.h>
261 
262 #if defined(INET)
263 struct evcnt tcp_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
264     NULL, "tcp", "hwcsum bad");
265 struct evcnt tcp_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
266     NULL, "tcp", "hwcsum ok");
267 struct evcnt tcp_hwcsum_data = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
268     NULL, "tcp", "hwcsum data");
269 struct evcnt tcp_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
270     NULL, "tcp", "swcsum");
271 
272 EVCNT_ATTACH_STATIC(tcp_hwcsum_bad);
273 EVCNT_ATTACH_STATIC(tcp_hwcsum_ok);
274 EVCNT_ATTACH_STATIC(tcp_hwcsum_data);
275 EVCNT_ATTACH_STATIC(tcp_swcsum);
276 #endif /* defined(INET) */
277 
278 #if defined(INET6)
279 struct evcnt tcp6_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
280     NULL, "tcp6", "hwcsum bad");
281 struct evcnt tcp6_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
282     NULL, "tcp6", "hwcsum ok");
283 struct evcnt tcp6_hwcsum_data = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
284     NULL, "tcp6", "hwcsum data");
285 struct evcnt tcp6_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
286     NULL, "tcp6", "swcsum");
287 
288 EVCNT_ATTACH_STATIC(tcp6_hwcsum_bad);
289 EVCNT_ATTACH_STATIC(tcp6_hwcsum_ok);
290 EVCNT_ATTACH_STATIC(tcp6_hwcsum_data);
291 EVCNT_ATTACH_STATIC(tcp6_swcsum);
292 #endif /* defined(INET6) */
293 #endif /* TCP_CSUM_COUNTERS */
294 
295 
296 #ifdef TCP_OUTPUT_COUNTERS
297 #include <sys/device.h>
298 
299 struct evcnt tcp_output_bigheader = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
300     NULL, "tcp", "output big header");
301 struct evcnt tcp_output_predict_hit = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
302     NULL, "tcp", "output predict hit");
303 struct evcnt tcp_output_predict_miss = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
304     NULL, "tcp", "output predict miss");
305 struct evcnt tcp_output_copysmall = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
306     NULL, "tcp", "output copy small");
307 struct evcnt tcp_output_copybig = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
308     NULL, "tcp", "output copy big");
309 struct evcnt tcp_output_refbig = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
310     NULL, "tcp", "output reference big");
311 
312 EVCNT_ATTACH_STATIC(tcp_output_bigheader);
313 EVCNT_ATTACH_STATIC(tcp_output_predict_hit);
314 EVCNT_ATTACH_STATIC(tcp_output_predict_miss);
315 EVCNT_ATTACH_STATIC(tcp_output_copysmall);
316 EVCNT_ATTACH_STATIC(tcp_output_copybig);
317 EVCNT_ATTACH_STATIC(tcp_output_refbig);
318 
319 #endif /* TCP_OUTPUT_COUNTERS */
320 
321 #ifdef TCP_REASS_COUNTERS
322 #include <sys/device.h>
323 
324 struct evcnt tcp_reass_ = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
325     NULL, "tcp_reass", "calls");
326 struct evcnt tcp_reass_empty = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
327     &tcp_reass_, "tcp_reass", "insert into empty queue");
328 struct evcnt tcp_reass_iteration[8] = {
329     EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", ">7 iterations"),
330     EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "1 iteration"),
331     EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "2 iterations"),
332     EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "3 iterations"),
333     EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "4 iterations"),
334     EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "5 iterations"),
335     EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "6 iterations"),
336     EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &tcp_reass_, "tcp_reass", "7 iterations"),
337 };
338 struct evcnt tcp_reass_prependfirst = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
339     &tcp_reass_, "tcp_reass", "prepend to first");
340 struct evcnt tcp_reass_prepend = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
341     &tcp_reass_, "tcp_reass", "prepend");
342 struct evcnt tcp_reass_insert = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
343     &tcp_reass_, "tcp_reass", "insert");
344 struct evcnt tcp_reass_inserttail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
345     &tcp_reass_, "tcp_reass", "insert at tail");
346 struct evcnt tcp_reass_append = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
347     &tcp_reass_, "tcp_reass", "append");
348 struct evcnt tcp_reass_appendtail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
349     &tcp_reass_, "tcp_reass", "append to tail fragment");
350 struct evcnt tcp_reass_overlaptail = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
351     &tcp_reass_, "tcp_reass", "overlap at end");
352 struct evcnt tcp_reass_overlapfront = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
353     &tcp_reass_, "tcp_reass", "overlap at start");
354 struct evcnt tcp_reass_segdup = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
355     &tcp_reass_, "tcp_reass", "duplicate segment");
356 struct evcnt tcp_reass_fragdup = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
357     &tcp_reass_, "tcp_reass", "duplicate fragment");
358 
359 EVCNT_ATTACH_STATIC(tcp_reass_);
360 EVCNT_ATTACH_STATIC(tcp_reass_empty);
361 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 0);
362 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 1);
363 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 2);
364 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 3);
365 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 4);
366 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 5);
367 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 6);
368 EVCNT_ATTACH_STATIC2(tcp_reass_iteration, 7);
369 EVCNT_ATTACH_STATIC(tcp_reass_prependfirst);
370 EVCNT_ATTACH_STATIC(tcp_reass_prepend);
371 EVCNT_ATTACH_STATIC(tcp_reass_insert);
372 EVCNT_ATTACH_STATIC(tcp_reass_inserttail);
373 EVCNT_ATTACH_STATIC(tcp_reass_append);
374 EVCNT_ATTACH_STATIC(tcp_reass_appendtail);
375 EVCNT_ATTACH_STATIC(tcp_reass_overlaptail);
376 EVCNT_ATTACH_STATIC(tcp_reass_overlapfront);
377 EVCNT_ATTACH_STATIC(tcp_reass_segdup);
378 EVCNT_ATTACH_STATIC(tcp_reass_fragdup);
379 
380 #endif /* TCP_REASS_COUNTERS */
381 
382 #ifdef MBUFTRACE
383 struct mowner tcp_mowner = MOWNER_INIT("tcp", "");
384 struct mowner tcp_rx_mowner = MOWNER_INIT("tcp", "rx");
385 struct mowner tcp_tx_mowner = MOWNER_INIT("tcp", "tx");
386 struct mowner tcp_sock_mowner = MOWNER_INIT("tcp", "sock");
387 struct mowner tcp_sock_rx_mowner = MOWNER_INIT("tcp", "sock rx");
388 struct mowner tcp_sock_tx_mowner = MOWNER_INIT("tcp", "sock tx");
389 #endif
390 
391 callout_t tcp_slowtimo_ch;
392 
393 static int
394 do_tcpinit(void)
395 {
396 
397 	in_pcbinit(&tcbtable, tcbhashsize, tcbhashsize);
398 	pool_init(&tcpcb_pool, sizeof(struct tcpcb), 0, 0, 0, "tcpcbpl",
399 	    NULL, IPL_SOFTNET);
400 
401 	tcp_usrreq_init();
402 
403 	/* Initialize timer state. */
404 	tcp_timer_init();
405 
406 	/* Initialize the compressed state engine. */
407 	syn_cache_init();
408 
409 	/* Initialize the congestion control algorithms. */
410 	tcp_congctl_init();
411 
412 	/* Initialize the TCPCB template. */
413 	tcp_tcpcb_template();
414 
415 	/* Initialize reassembly queue */
416 	tcpipqent_init();
417 
418 	/* SACK */
419 	tcp_sack_init();
420 
421 	MOWNER_ATTACH(&tcp_tx_mowner);
422 	MOWNER_ATTACH(&tcp_rx_mowner);
423 	MOWNER_ATTACH(&tcp_reass_mowner);
424 	MOWNER_ATTACH(&tcp_sock_mowner);
425 	MOWNER_ATTACH(&tcp_sock_tx_mowner);
426 	MOWNER_ATTACH(&tcp_sock_rx_mowner);
427 	MOWNER_ATTACH(&tcp_mowner);
428 
429 	tcpstat_percpu = percpu_alloc(sizeof(uint64_t) * TCP_NSTATS);
430 
431 	vtw_earlyinit();
432 
433 	callout_init(&tcp_slowtimo_ch, CALLOUT_MPSAFE);
434 	callout_reset(&tcp_slowtimo_ch, 1, tcp_slowtimo, NULL);
435 
436 	return 0;
437 }
438 
439 void
440 tcp_init_common(unsigned basehlen)
441 {
442 	static ONCE_DECL(dotcpinit);
443 	unsigned hlen = basehlen + sizeof(struct tcphdr);
444 	unsigned oldhlen;
445 
446 	if (max_linkhdr + hlen > MHLEN)
447 		panic("tcp_init");
448 	while ((oldhlen = max_protohdr) < hlen)
449 		atomic_cas_uint(&max_protohdr, oldhlen, hlen);
450 
451 	RUN_ONCE(&dotcpinit, do_tcpinit);
452 }
453 
454 /*
455  * Tcp initialization
456  */
457 void
458 tcp_init(void)
459 {
460 
461 	icmp_mtudisc_callback_register(tcp_mtudisc_callback);
462 
463 	tcp_init_common(sizeof(struct ip));
464 }
465 
466 /*
467  * Create template to be used to send tcp packets on a connection.
468  * Call after host entry created, allocates an mbuf and fills
469  * in a skeletal tcp/ip header, minimizing the amount of work
470  * necessary when the connection is used.
471  */
472 struct mbuf *
473 tcp_template(struct tcpcb *tp)
474 {
475 	struct inpcb *inp = tp->t_inpcb;
476 #ifdef INET6
477 	struct in6pcb *in6p = tp->t_in6pcb;
478 #endif
479 	struct tcphdr *n;
480 	struct mbuf *m;
481 	int hlen;
482 
483 	switch (tp->t_family) {
484 	case AF_INET:
485 		hlen = sizeof(struct ip);
486 		if (inp)
487 			break;
488 #ifdef INET6
489 		if (in6p) {
490 			/* mapped addr case */
491 			if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_laddr)
492 			 && IN6_IS_ADDR_V4MAPPED(&in6p->in6p_faddr))
493 				break;
494 		}
495 #endif
496 		return NULL;	/*EINVAL*/
497 #ifdef INET6
498 	case AF_INET6:
499 		hlen = sizeof(struct ip6_hdr);
500 		if (in6p) {
501 			/* more sainty check? */
502 			break;
503 		}
504 		return NULL;	/*EINVAL*/
505 #endif
506 	default:
507 		hlen = 0;	/*pacify gcc*/
508 		return NULL;	/*EAFNOSUPPORT*/
509 	}
510 #ifdef DIAGNOSTIC
511 	if (hlen + sizeof(struct tcphdr) > MCLBYTES)
512 		panic("mclbytes too small for t_template");
513 #endif
514 	m = tp->t_template;
515 	if (m && m->m_len == hlen + sizeof(struct tcphdr))
516 		;
517 	else {
518 		if (m)
519 			m_freem(m);
520 		m = tp->t_template = NULL;
521 		MGETHDR(m, M_DONTWAIT, MT_HEADER);
522 		if (m && hlen + sizeof(struct tcphdr) > MHLEN) {
523 			MCLGET(m, M_DONTWAIT);
524 			if ((m->m_flags & M_EXT) == 0) {
525 				m_free(m);
526 				m = NULL;
527 			}
528 		}
529 		if (m == NULL)
530 			return NULL;
531 		MCLAIM(m, &tcp_mowner);
532 		m->m_pkthdr.len = m->m_len = hlen + sizeof(struct tcphdr);
533 	}
534 
535 	memset(mtod(m, void *), 0, m->m_len);
536 
537 	n = (struct tcphdr *)(mtod(m, char *) + hlen);
538 
539 	switch (tp->t_family) {
540 	case AF_INET:
541 	    {
542 		struct ipovly *ipov;
543 		mtod(m, struct ip *)->ip_v = 4;
544 		mtod(m, struct ip *)->ip_hl = hlen >> 2;
545 		ipov = mtod(m, struct ipovly *);
546 		ipov->ih_pr = IPPROTO_TCP;
547 		ipov->ih_len = htons(sizeof(struct tcphdr));
548 		if (inp) {
549 			ipov->ih_src = inp->inp_laddr;
550 			ipov->ih_dst = inp->inp_faddr;
551 		}
552 #ifdef INET6
553 		else if (in6p) {
554 			/* mapped addr case */
555 			bcopy(&in6p->in6p_laddr.s6_addr32[3], &ipov->ih_src,
556 				sizeof(ipov->ih_src));
557 			bcopy(&in6p->in6p_faddr.s6_addr32[3], &ipov->ih_dst,
558 				sizeof(ipov->ih_dst));
559 		}
560 #endif
561 		/*
562 		 * Compute the pseudo-header portion of the checksum
563 		 * now.  We incrementally add in the TCP option and
564 		 * payload lengths later, and then compute the TCP
565 		 * checksum right before the packet is sent off onto
566 		 * the wire.
567 		 */
568 		n->th_sum = in_cksum_phdr(ipov->ih_src.s_addr,
569 		    ipov->ih_dst.s_addr,
570 		    htons(sizeof(struct tcphdr) + IPPROTO_TCP));
571 		break;
572 	    }
573 #ifdef INET6
574 	case AF_INET6:
575 	    {
576 		struct ip6_hdr *ip6;
577 		mtod(m, struct ip *)->ip_v = 6;
578 		ip6 = mtod(m, struct ip6_hdr *);
579 		ip6->ip6_nxt = IPPROTO_TCP;
580 		ip6->ip6_plen = htons(sizeof(struct tcphdr));
581 		ip6->ip6_src = in6p->in6p_laddr;
582 		ip6->ip6_dst = in6p->in6p_faddr;
583 		ip6->ip6_flow = in6p->in6p_flowinfo & IPV6_FLOWINFO_MASK;
584 		if (ip6_auto_flowlabel) {
585 			ip6->ip6_flow &= ~IPV6_FLOWLABEL_MASK;
586 			ip6->ip6_flow |=
587 			    (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK);
588 		}
589 		ip6->ip6_vfc &= ~IPV6_VERSION_MASK;
590 		ip6->ip6_vfc |= IPV6_VERSION;
591 
592 		/*
593 		 * Compute the pseudo-header portion of the checksum
594 		 * now.  We incrementally add in the TCP option and
595 		 * payload lengths later, and then compute the TCP
596 		 * checksum right before the packet is sent off onto
597 		 * the wire.
598 		 */
599 		n->th_sum = in6_cksum_phdr(&in6p->in6p_laddr,
600 		    &in6p->in6p_faddr, htonl(sizeof(struct tcphdr)),
601 		    htonl(IPPROTO_TCP));
602 		break;
603 	    }
604 #endif
605 	}
606 	if (inp) {
607 		n->th_sport = inp->inp_lport;
608 		n->th_dport = inp->inp_fport;
609 	}
610 #ifdef INET6
611 	else if (in6p) {
612 		n->th_sport = in6p->in6p_lport;
613 		n->th_dport = in6p->in6p_fport;
614 	}
615 #endif
616 	n->th_seq = 0;
617 	n->th_ack = 0;
618 	n->th_x2 = 0;
619 	n->th_off = 5;
620 	n->th_flags = 0;
621 	n->th_win = 0;
622 	n->th_urp = 0;
623 	return (m);
624 }
625 
626 /*
627  * Send a single message to the TCP at address specified by
628  * the given TCP/IP header.  If m == 0, then we make a copy
629  * of the tcpiphdr at ti and send directly to the addressed host.
630  * This is used to force keep alive messages out using the TCP
631  * template for a connection tp->t_template.  If flags are given
632  * then we send a message back to the TCP which originated the
633  * segment ti, and discard the mbuf containing it and any other
634  * attached mbufs.
635  *
636  * In any case the ack and sequence number of the transmitted
637  * segment are as specified by the parameters.
638  */
639 int
640 tcp_respond(struct tcpcb *tp, struct mbuf *mtemplate, struct mbuf *m,
641     struct tcphdr *th0, tcp_seq ack, tcp_seq seq, int flags)
642 {
643 	struct route *ro;
644 	int error, tlen, win = 0;
645 	int hlen;
646 	struct ip *ip;
647 #ifdef INET6
648 	struct ip6_hdr *ip6;
649 #endif
650 	int family;	/* family on packet, not inpcb/in6pcb! */
651 	struct tcphdr *th;
652 	struct socket *so;
653 
654 	if (tp != NULL && (flags & TH_RST) == 0) {
655 #ifdef DIAGNOSTIC
656 		if (tp->t_inpcb && tp->t_in6pcb)
657 			panic("tcp_respond: both t_inpcb and t_in6pcb are set");
658 #endif
659 #ifdef INET
660 		if (tp->t_inpcb)
661 			win = sbspace(&tp->t_inpcb->inp_socket->so_rcv);
662 #endif
663 #ifdef INET6
664 		if (tp->t_in6pcb)
665 			win = sbspace(&tp->t_in6pcb->in6p_socket->so_rcv);
666 #endif
667 	}
668 
669 	th = NULL;	/* Quell uninitialized warning */
670 	ip = NULL;
671 #ifdef INET6
672 	ip6 = NULL;
673 #endif
674 	if (m == 0) {
675 		if (!mtemplate)
676 			return EINVAL;
677 
678 		/* get family information from template */
679 		switch (mtod(mtemplate, struct ip *)->ip_v) {
680 		case 4:
681 			family = AF_INET;
682 			hlen = sizeof(struct ip);
683 			break;
684 #ifdef INET6
685 		case 6:
686 			family = AF_INET6;
687 			hlen = sizeof(struct ip6_hdr);
688 			break;
689 #endif
690 		default:
691 			return EAFNOSUPPORT;
692 		}
693 
694 		MGETHDR(m, M_DONTWAIT, MT_HEADER);
695 		if (m) {
696 			MCLAIM(m, &tcp_tx_mowner);
697 			MCLGET(m, M_DONTWAIT);
698 			if ((m->m_flags & M_EXT) == 0) {
699 				m_free(m);
700 				m = NULL;
701 			}
702 		}
703 		if (m == NULL)
704 			return (ENOBUFS);
705 
706 		if (tcp_compat_42)
707 			tlen = 1;
708 		else
709 			tlen = 0;
710 
711 		m->m_data += max_linkhdr;
712 		bcopy(mtod(mtemplate, void *), mtod(m, void *),
713 			mtemplate->m_len);
714 		switch (family) {
715 		case AF_INET:
716 			ip = mtod(m, struct ip *);
717 			th = (struct tcphdr *)(ip + 1);
718 			break;
719 #ifdef INET6
720 		case AF_INET6:
721 			ip6 = mtod(m, struct ip6_hdr *);
722 			th = (struct tcphdr *)(ip6 + 1);
723 			break;
724 #endif
725 #if 0
726 		default:
727 			/* noone will visit here */
728 			m_freem(m);
729 			return EAFNOSUPPORT;
730 #endif
731 		}
732 		flags = TH_ACK;
733 	} else {
734 
735 		if ((m->m_flags & M_PKTHDR) == 0) {
736 #if 0
737 			printf("non PKTHDR to tcp_respond\n");
738 #endif
739 			m_freem(m);
740 			return EINVAL;
741 		}
742 #ifdef DIAGNOSTIC
743 		if (!th0)
744 			panic("th0 == NULL in tcp_respond");
745 #endif
746 
747 		/* get family information from m */
748 		switch (mtod(m, struct ip *)->ip_v) {
749 		case 4:
750 			family = AF_INET;
751 			hlen = sizeof(struct ip);
752 			ip = mtod(m, struct ip *);
753 			break;
754 #ifdef INET6
755 		case 6:
756 			family = AF_INET6;
757 			hlen = sizeof(struct ip6_hdr);
758 			ip6 = mtod(m, struct ip6_hdr *);
759 			break;
760 #endif
761 		default:
762 			m_freem(m);
763 			return EAFNOSUPPORT;
764 		}
765 		/* clear h/w csum flags inherited from rx packet */
766 		m->m_pkthdr.csum_flags = 0;
767 
768 		if ((flags & TH_SYN) == 0 || sizeof(*th0) > (th0->th_off << 2))
769 			tlen = sizeof(*th0);
770 		else
771 			tlen = th0->th_off << 2;
772 
773 		if (m->m_len > hlen + tlen && (m->m_flags & M_EXT) == 0 &&
774 		    mtod(m, char *) + hlen == (char *)th0) {
775 			m->m_len = hlen + tlen;
776 			m_freem(m->m_next);
777 			m->m_next = NULL;
778 		} else {
779 			struct mbuf *n;
780 
781 #ifdef DIAGNOSTIC
782 			if (max_linkhdr + hlen + tlen > MCLBYTES) {
783 				m_freem(m);
784 				return EMSGSIZE;
785 			}
786 #endif
787 			MGETHDR(n, M_DONTWAIT, MT_HEADER);
788 			if (n && max_linkhdr + hlen + tlen > MHLEN) {
789 				MCLGET(n, M_DONTWAIT);
790 				if ((n->m_flags & M_EXT) == 0) {
791 					m_freem(n);
792 					n = NULL;
793 				}
794 			}
795 			if (!n) {
796 				m_freem(m);
797 				return ENOBUFS;
798 			}
799 
800 			MCLAIM(n, &tcp_tx_mowner);
801 			n->m_data += max_linkhdr;
802 			n->m_len = hlen + tlen;
803 			m_copyback(n, 0, hlen, mtod(m, void *));
804 			m_copyback(n, hlen, tlen, (void *)th0);
805 
806 			m_freem(m);
807 			m = n;
808 			n = NULL;
809 		}
810 
811 #define xchg(a,b,type) { type t; t=a; a=b; b=t; }
812 		switch (family) {
813 		case AF_INET:
814 			ip = mtod(m, struct ip *);
815 			th = (struct tcphdr *)(ip + 1);
816 			ip->ip_p = IPPROTO_TCP;
817 			xchg(ip->ip_dst, ip->ip_src, struct in_addr);
818 			ip->ip_p = IPPROTO_TCP;
819 			break;
820 #ifdef INET6
821 		case AF_INET6:
822 			ip6 = mtod(m, struct ip6_hdr *);
823 			th = (struct tcphdr *)(ip6 + 1);
824 			ip6->ip6_nxt = IPPROTO_TCP;
825 			xchg(ip6->ip6_dst, ip6->ip6_src, struct in6_addr);
826 			ip6->ip6_nxt = IPPROTO_TCP;
827 			break;
828 #endif
829 #if 0
830 		default:
831 			/* noone will visit here */
832 			m_freem(m);
833 			return EAFNOSUPPORT;
834 #endif
835 		}
836 		xchg(th->th_dport, th->th_sport, u_int16_t);
837 #undef xchg
838 		tlen = 0;	/*be friendly with the following code*/
839 	}
840 	th->th_seq = htonl(seq);
841 	th->th_ack = htonl(ack);
842 	th->th_x2 = 0;
843 	if ((flags & TH_SYN) == 0) {
844 		if (tp)
845 			win >>= tp->rcv_scale;
846 		if (win > TCP_MAXWIN)
847 			win = TCP_MAXWIN;
848 		th->th_win = htons((u_int16_t)win);
849 		th->th_off = sizeof (struct tcphdr) >> 2;
850 		tlen += sizeof(*th);
851 	} else
852 		tlen += th->th_off << 2;
853 	m->m_len = hlen + tlen;
854 	m->m_pkthdr.len = hlen + tlen;
855 	m->m_pkthdr.rcvif = NULL;
856 	th->th_flags = flags;
857 	th->th_urp = 0;
858 
859 	switch (family) {
860 #ifdef INET
861 	case AF_INET:
862 	    {
863 		struct ipovly *ipov = (struct ipovly *)ip;
864 		memset(ipov->ih_x1, 0, sizeof ipov->ih_x1);
865 		ipov->ih_len = htons((u_int16_t)tlen);
866 
867 		th->th_sum = 0;
868 		th->th_sum = in_cksum(m, hlen + tlen);
869 		ip->ip_len = htons(hlen + tlen);
870 		ip->ip_ttl = ip_defttl;
871 		break;
872 	    }
873 #endif
874 #ifdef INET6
875 	case AF_INET6:
876 	    {
877 		th->th_sum = 0;
878 		th->th_sum = in6_cksum(m, IPPROTO_TCP, sizeof(struct ip6_hdr),
879 				tlen);
880 		ip6->ip6_plen = htons(tlen);
881 		if (tp && tp->t_in6pcb)
882 			ip6->ip6_hlim = in6_selecthlim_rt(tp->t_in6pcb);
883 		else
884 			ip6->ip6_hlim = ip6_defhlim;
885 		ip6->ip6_flow &= ~IPV6_FLOWINFO_MASK;
886 		if (ip6_auto_flowlabel) {
887 			ip6->ip6_flow |=
888 			    (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK);
889 		}
890 		break;
891 	    }
892 #endif
893 	}
894 
895 	if (tp && tp->t_inpcb)
896 		so = tp->t_inpcb->inp_socket;
897 #ifdef INET6
898 	else if (tp && tp->t_in6pcb)
899 		so = tp->t_in6pcb->in6p_socket;
900 #endif
901 	else
902 		so = NULL;
903 
904 	if (tp != NULL && tp->t_inpcb != NULL) {
905 		ro = &tp->t_inpcb->inp_route;
906 #ifdef DIAGNOSTIC
907 		if (family != AF_INET)
908 			panic("tcp_respond: address family mismatch");
909 		if (!in_hosteq(ip->ip_dst, tp->t_inpcb->inp_faddr)) {
910 			panic("tcp_respond: ip_dst %x != inp_faddr %x",
911 			    ntohl(ip->ip_dst.s_addr),
912 			    ntohl(tp->t_inpcb->inp_faddr.s_addr));
913 		}
914 #endif
915 	}
916 #ifdef INET6
917 	else if (tp != NULL && tp->t_in6pcb != NULL) {
918 		ro = (struct route *)&tp->t_in6pcb->in6p_route;
919 #ifdef DIAGNOSTIC
920 		if (family == AF_INET) {
921 			if (!IN6_IS_ADDR_V4MAPPED(&tp->t_in6pcb->in6p_faddr))
922 				panic("tcp_respond: not mapped addr");
923 			if (memcmp(&ip->ip_dst,
924 			    &tp->t_in6pcb->in6p_faddr.s6_addr32[3],
925 			    sizeof(ip->ip_dst)) != 0) {
926 				panic("tcp_respond: ip_dst != in6p_faddr");
927 			}
928 		} else if (family == AF_INET6) {
929 			if (!IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst,
930 			    &tp->t_in6pcb->in6p_faddr))
931 				panic("tcp_respond: ip6_dst != in6p_faddr");
932 		} else
933 			panic("tcp_respond: address family mismatch");
934 #endif
935 	}
936 #endif
937 	else
938 		ro = NULL;
939 
940 	switch (family) {
941 #ifdef INET
942 	case AF_INET:
943 		error = ip_output(m, NULL, ro,
944 		    (tp && tp->t_mtudisc ? IP_MTUDISC : 0), NULL, so);
945 		break;
946 #endif
947 #ifdef INET6
948 	case AF_INET6:
949 		error = ip6_output(m, NULL, ro, 0, NULL, so, NULL);
950 		break;
951 #endif
952 	default:
953 		error = EAFNOSUPPORT;
954 		break;
955 	}
956 
957 	return (error);
958 }
959 
960 /*
961  * Template TCPCB.  Rather than zeroing a new TCPCB and initializing
962  * a bunch of members individually, we maintain this template for the
963  * static and mostly-static components of the TCPCB, and copy it into
964  * the new TCPCB instead.
965  */
966 static struct tcpcb tcpcb_template = {
967 	.t_srtt = TCPTV_SRTTBASE,
968 	.t_rttmin = TCPTV_MIN,
969 
970 	.snd_cwnd = TCP_MAXWIN << TCP_MAX_WINSHIFT,
971 	.snd_ssthresh = TCP_MAXWIN << TCP_MAX_WINSHIFT,
972 	.snd_numholes = 0,
973 	.snd_cubic_wmax = 0,
974 	.snd_cubic_wmax_last = 0,
975 	.snd_cubic_ctime = 0,
976 
977 	.t_partialacks = -1,
978 	.t_bytes_acked = 0,
979 	.t_sndrexmitpack = 0,
980 	.t_rcvoopack = 0,
981 	.t_sndzerowin = 0,
982 };
983 
984 /*
985  * Updates the TCPCB template whenever a parameter that would affect
986  * the template is changed.
987  */
988 void
989 tcp_tcpcb_template(void)
990 {
991 	struct tcpcb *tp = &tcpcb_template;
992 	int flags;
993 
994 	tp->t_peermss = tcp_mssdflt;
995 	tp->t_ourmss = tcp_mssdflt;
996 	tp->t_segsz = tcp_mssdflt;
997 
998 	flags = 0;
999 	if (tcp_do_rfc1323 && tcp_do_win_scale)
1000 		flags |= TF_REQ_SCALE;
1001 	if (tcp_do_rfc1323 && tcp_do_timestamps)
1002 		flags |= TF_REQ_TSTMP;
1003 	tp->t_flags = flags;
1004 
1005 	/*
1006 	 * Init srtt to TCPTV_SRTTBASE (0), so we can tell that we have no
1007 	 * rtt estimate.  Set rttvar so that srtt + 2 * rttvar gives
1008 	 * reasonable initial retransmit time.
1009 	 */
1010 	tp->t_rttvar = tcp_rttdflt * PR_SLOWHZ << (TCP_RTTVAR_SHIFT + 2 - 1);
1011 	TCPT_RANGESET(tp->t_rxtcur, TCP_REXMTVAL(tp),
1012 	    TCPTV_MIN, TCPTV_REXMTMAX);
1013 
1014 	/* Keep Alive */
1015 	tp->t_keepinit = tcp_keepinit;
1016 	tp->t_keepidle = tcp_keepidle;
1017 	tp->t_keepintvl = tcp_keepintvl;
1018 	tp->t_keepcnt = tcp_keepcnt;
1019 	tp->t_maxidle = tp->t_keepcnt * tp->t_keepintvl;
1020 
1021 	/* MSL */
1022 	tp->t_msl = TCPTV_MSL;
1023 }
1024 
1025 /*
1026  * Create a new TCP control block, making an
1027  * empty reassembly queue and hooking it to the argument
1028  * protocol control block.
1029  */
1030 /* family selects inpcb, or in6pcb */
1031 struct tcpcb *
1032 tcp_newtcpcb(int family, void *aux)
1033 {
1034 	struct tcpcb *tp;
1035 	int i;
1036 
1037 	/* XXX Consider using a pool_cache for speed. */
1038 	tp = pool_get(&tcpcb_pool, PR_NOWAIT);	/* splsoftnet via tcp_usrreq */
1039 	if (tp == NULL)
1040 		return (NULL);
1041 	memcpy(tp, &tcpcb_template, sizeof(*tp));
1042 	TAILQ_INIT(&tp->segq);
1043 	TAILQ_INIT(&tp->timeq);
1044 	tp->t_family = family;		/* may be overridden later on */
1045 	TAILQ_INIT(&tp->snd_holes);
1046 	LIST_INIT(&tp->t_sc);		/* XXX can template this */
1047 
1048 	/* Don't sweat this loop; hopefully the compiler will unroll it. */
1049 	for (i = 0; i < TCPT_NTIMERS; i++) {
1050 		callout_init(&tp->t_timer[i], CALLOUT_MPSAFE);
1051 		TCP_TIMER_INIT(tp, i);
1052 	}
1053 	callout_init(&tp->t_delack_ch, CALLOUT_MPSAFE);
1054 
1055 	switch (family) {
1056 	case AF_INET:
1057 	    {
1058 		struct inpcb *inp = (struct inpcb *)aux;
1059 
1060 		inp->inp_ip.ip_ttl = ip_defttl;
1061 		inp->inp_ppcb = (void *)tp;
1062 
1063 		tp->t_inpcb = inp;
1064 		tp->t_mtudisc = ip_mtudisc;
1065 		break;
1066 	    }
1067 #ifdef INET6
1068 	case AF_INET6:
1069 	    {
1070 		struct in6pcb *in6p = (struct in6pcb *)aux;
1071 
1072 		in6p->in6p_ip6.ip6_hlim = in6_selecthlim_rt(in6p);
1073 		in6p->in6p_ppcb = (void *)tp;
1074 
1075 		tp->t_in6pcb = in6p;
1076 		/* for IPv6, always try to run path MTU discovery */
1077 		tp->t_mtudisc = 1;
1078 		break;
1079 	    }
1080 #endif /* INET6 */
1081 	default:
1082 		for (i = 0; i < TCPT_NTIMERS; i++)
1083 			callout_destroy(&tp->t_timer[i]);
1084 		callout_destroy(&tp->t_delack_ch);
1085 		pool_put(&tcpcb_pool, tp);	/* splsoftnet via tcp_usrreq */
1086 		return (NULL);
1087 	}
1088 
1089 	/*
1090 	 * Initialize our timebase.  When we send timestamps, we take
1091 	 * the delta from tcp_now -- this means each connection always
1092 	 * gets a timebase of 1, which makes it, among other things,
1093 	 * more difficult to determine how long a system has been up,
1094 	 * and thus how many TCP sequence increments have occurred.
1095 	 *
1096 	 * We start with 1, because 0 doesn't work with linux, which
1097 	 * considers timestamp 0 in a SYN packet as a bug and disables
1098 	 * timestamps.
1099 	 */
1100 	tp->ts_timebase = tcp_now - 1;
1101 
1102 	tcp_congctl_select(tp, tcp_congctl_global_name);
1103 
1104 	return (tp);
1105 }
1106 
1107 /*
1108  * Drop a TCP connection, reporting
1109  * the specified error.  If connection is synchronized,
1110  * then send a RST to peer.
1111  */
1112 struct tcpcb *
1113 tcp_drop(struct tcpcb *tp, int errno)
1114 {
1115 	struct socket *so = NULL;
1116 
1117 #ifdef DIAGNOSTIC
1118 	if (tp->t_inpcb && tp->t_in6pcb)
1119 		panic("tcp_drop: both t_inpcb and t_in6pcb are set");
1120 #endif
1121 #ifdef INET
1122 	if (tp->t_inpcb)
1123 		so = tp->t_inpcb->inp_socket;
1124 #endif
1125 #ifdef INET6
1126 	if (tp->t_in6pcb)
1127 		so = tp->t_in6pcb->in6p_socket;
1128 #endif
1129 	if (!so)
1130 		return NULL;
1131 
1132 	if (TCPS_HAVERCVDSYN(tp->t_state)) {
1133 		tp->t_state = TCPS_CLOSED;
1134 		(void) tcp_output(tp);
1135 		TCP_STATINC(TCP_STAT_DROPS);
1136 	} else
1137 		TCP_STATINC(TCP_STAT_CONNDROPS);
1138 	if (errno == ETIMEDOUT && tp->t_softerror)
1139 		errno = tp->t_softerror;
1140 	so->so_error = errno;
1141 	return (tcp_close(tp));
1142 }
1143 
1144 /*
1145  * Close a TCP control block:
1146  *	discard all space held by the tcp
1147  *	discard internet protocol block
1148  *	wake up any sleepers
1149  */
1150 struct tcpcb *
1151 tcp_close(struct tcpcb *tp)
1152 {
1153 	struct inpcb *inp;
1154 #ifdef INET6
1155 	struct in6pcb *in6p;
1156 #endif
1157 	struct socket *so;
1158 #ifdef RTV_RTT
1159 	struct rtentry *rt;
1160 #endif
1161 	struct route *ro;
1162 	int j;
1163 
1164 	inp = tp->t_inpcb;
1165 #ifdef INET6
1166 	in6p = tp->t_in6pcb;
1167 #endif
1168 	so = NULL;
1169 	ro = NULL;
1170 	if (inp) {
1171 		so = inp->inp_socket;
1172 		ro = &inp->inp_route;
1173 	}
1174 #ifdef INET6
1175 	else if (in6p) {
1176 		so = in6p->in6p_socket;
1177 		ro = (struct route *)&in6p->in6p_route;
1178 	}
1179 #endif
1180 
1181 #ifdef RTV_RTT
1182 	/*
1183 	 * If we sent enough data to get some meaningful characteristics,
1184 	 * save them in the routing entry.  'Enough' is arbitrarily
1185 	 * defined as the sendpipesize (default 4K) * 16.  This would
1186 	 * give us 16 rtt samples assuming we only get one sample per
1187 	 * window (the usual case on a long haul net).  16 samples is
1188 	 * enough for the srtt filter to converge to within 5% of the correct
1189 	 * value; fewer samples and we could save a very bogus rtt.
1190 	 *
1191 	 * Don't update the default route's characteristics and don't
1192 	 * update anything that the user "locked".
1193 	 */
1194 	if (SEQ_LT(tp->iss + so->so_snd.sb_hiwat * 16, tp->snd_max) &&
1195 	    ro && (rt = rtcache_validate(ro)) != NULL &&
1196 	    !in_nullhost(satocsin(rt_getkey(rt))->sin_addr)) {
1197 		u_long i = 0;
1198 
1199 		if ((rt->rt_rmx.rmx_locks & RTV_RTT) == 0) {
1200 			i = tp->t_srtt *
1201 			    ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTT_SHIFT + 2));
1202 			if (rt->rt_rmx.rmx_rtt && i)
1203 				/*
1204 				 * filter this update to half the old & half
1205 				 * the new values, converting scale.
1206 				 * See route.h and tcp_var.h for a
1207 				 * description of the scaling constants.
1208 				 */
1209 				rt->rt_rmx.rmx_rtt =
1210 				    (rt->rt_rmx.rmx_rtt + i) / 2;
1211 			else
1212 				rt->rt_rmx.rmx_rtt = i;
1213 		}
1214 		if ((rt->rt_rmx.rmx_locks & RTV_RTTVAR) == 0) {
1215 			i = tp->t_rttvar *
1216 			    ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTTVAR_SHIFT + 2));
1217 			if (rt->rt_rmx.rmx_rttvar && i)
1218 				rt->rt_rmx.rmx_rttvar =
1219 				    (rt->rt_rmx.rmx_rttvar + i) / 2;
1220 			else
1221 				rt->rt_rmx.rmx_rttvar = i;
1222 		}
1223 		/*
1224 		 * update the pipelimit (ssthresh) if it has been updated
1225 		 * already or if a pipesize was specified & the threshhold
1226 		 * got below half the pipesize.  I.e., wait for bad news
1227 		 * before we start updating, then update on both good
1228 		 * and bad news.
1229 		 */
1230 		if (((rt->rt_rmx.rmx_locks & RTV_SSTHRESH) == 0 &&
1231 		    (i = tp->snd_ssthresh) && rt->rt_rmx.rmx_ssthresh) ||
1232 		    i < (rt->rt_rmx.rmx_sendpipe / 2)) {
1233 			/*
1234 			 * convert the limit from user data bytes to
1235 			 * packets then to packet data bytes.
1236 			 */
1237 			i = (i + tp->t_segsz / 2) / tp->t_segsz;
1238 			if (i < 2)
1239 				i = 2;
1240 			i *= (u_long)(tp->t_segsz + sizeof (struct tcpiphdr));
1241 			if (rt->rt_rmx.rmx_ssthresh)
1242 				rt->rt_rmx.rmx_ssthresh =
1243 				    (rt->rt_rmx.rmx_ssthresh + i) / 2;
1244 			else
1245 				rt->rt_rmx.rmx_ssthresh = i;
1246 		}
1247 	}
1248 #endif /* RTV_RTT */
1249 	/* free the reassembly queue, if any */
1250 	TCP_REASS_LOCK(tp);
1251 	(void) tcp_freeq(tp);
1252 	TCP_REASS_UNLOCK(tp);
1253 
1254 	/* free the SACK holes list. */
1255 	tcp_free_sackholes(tp);
1256 	tcp_congctl_release(tp);
1257 	syn_cache_cleanup(tp);
1258 
1259 	if (tp->t_template) {
1260 		m_free(tp->t_template);
1261 		tp->t_template = NULL;
1262 	}
1263 
1264 	/*
1265 	 * Detaching the pcb will unlock the socket/tcpcb, and stopping
1266 	 * the timers can also drop the lock.  We need to prevent access
1267 	 * to the tcpcb as it's half torn down.  Flag the pcb as dead
1268 	 * (prevents access by timers) and only then detach it.
1269 	 */
1270 	tp->t_flags |= TF_DEAD;
1271 	if (inp) {
1272 		inp->inp_ppcb = 0;
1273 		soisdisconnected(so);
1274 		in_pcbdetach(inp);
1275 	}
1276 #ifdef INET6
1277 	else if (in6p) {
1278 		in6p->in6p_ppcb = 0;
1279 		soisdisconnected(so);
1280 		in6_pcbdetach(in6p);
1281 	}
1282 #endif
1283 	/*
1284 	 * pcb is no longer visble elsewhere, so we can safely release
1285 	 * the lock in callout_halt() if needed.
1286 	 */
1287 	TCP_STATINC(TCP_STAT_CLOSED);
1288 	for (j = 0; j < TCPT_NTIMERS; j++) {
1289 		callout_halt(&tp->t_timer[j], softnet_lock);
1290 		callout_destroy(&tp->t_timer[j]);
1291 	}
1292 	callout_halt(&tp->t_delack_ch, softnet_lock);
1293 	callout_destroy(&tp->t_delack_ch);
1294 	pool_put(&tcpcb_pool, tp);
1295 
1296 	return NULL;
1297 }
1298 
1299 int
1300 tcp_freeq(struct tcpcb *tp)
1301 {
1302 	struct ipqent *qe;
1303 	int rv = 0;
1304 #ifdef TCPREASS_DEBUG
1305 	int i = 0;
1306 #endif
1307 
1308 	TCP_REASS_LOCK_CHECK(tp);
1309 
1310 	while ((qe = TAILQ_FIRST(&tp->segq)) != NULL) {
1311 #ifdef TCPREASS_DEBUG
1312 		printf("tcp_freeq[%p,%d]: %u:%u(%u) 0x%02x\n",
1313 			tp, i++, qe->ipqe_seq, qe->ipqe_seq + qe->ipqe_len,
1314 			qe->ipqe_len, qe->ipqe_flags & (TH_SYN|TH_FIN|TH_RST));
1315 #endif
1316 		TAILQ_REMOVE(&tp->segq, qe, ipqe_q);
1317 		TAILQ_REMOVE(&tp->timeq, qe, ipqe_timeq);
1318 		m_freem(qe->ipqe_m);
1319 		tcpipqent_free(qe);
1320 		rv = 1;
1321 	}
1322 	tp->t_segqlen = 0;
1323 	KASSERT(TAILQ_EMPTY(&tp->timeq));
1324 	return (rv);
1325 }
1326 
1327 void
1328 tcp_fasttimo(void)
1329 {
1330 	if (tcp_drainwanted) {
1331 		tcp_drain();
1332 		tcp_drainwanted = 0;
1333 	}
1334 }
1335 
1336 void
1337 tcp_drainstub(void)
1338 {
1339 	tcp_drainwanted = 1;
1340 }
1341 
1342 /*
1343  * Protocol drain routine.  Called when memory is in short supply.
1344  * Called from pr_fasttimo thus a callout context.
1345  */
1346 void
1347 tcp_drain(void)
1348 {
1349 	struct inpcb_hdr *inph;
1350 	struct tcpcb *tp;
1351 
1352 	mutex_enter(softnet_lock);
1353 	KERNEL_LOCK(1, NULL);
1354 
1355 	/*
1356 	 * Free the sequence queue of all TCP connections.
1357 	 */
1358 	TAILQ_FOREACH(inph, &tcbtable.inpt_queue, inph_queue) {
1359 		switch (inph->inph_af) {
1360 		case AF_INET:
1361 			tp = intotcpcb((struct inpcb *)inph);
1362 			break;
1363 #ifdef INET6
1364 		case AF_INET6:
1365 			tp = in6totcpcb((struct in6pcb *)inph);
1366 			break;
1367 #endif
1368 		default:
1369 			tp = NULL;
1370 			break;
1371 		}
1372 		if (tp != NULL) {
1373 			/*
1374 			 * We may be called from a device's interrupt
1375 			 * context.  If the tcpcb is already busy,
1376 			 * just bail out now.
1377 			 */
1378 			if (tcp_reass_lock_try(tp) == 0)
1379 				continue;
1380 			if (tcp_freeq(tp))
1381 				TCP_STATINC(TCP_STAT_CONNSDRAINED);
1382 			TCP_REASS_UNLOCK(tp);
1383 		}
1384 	}
1385 
1386 	KERNEL_UNLOCK_ONE(NULL);
1387 	mutex_exit(softnet_lock);
1388 }
1389 
1390 /*
1391  * Notify a tcp user of an asynchronous error;
1392  * store error as soft error, but wake up user
1393  * (for now, won't do anything until can select for soft error).
1394  */
1395 void
1396 tcp_notify(struct inpcb *inp, int error)
1397 {
1398 	struct tcpcb *tp = (struct tcpcb *)inp->inp_ppcb;
1399 	struct socket *so = inp->inp_socket;
1400 
1401 	/*
1402 	 * Ignore some errors if we are hooked up.
1403 	 * If connection hasn't completed, has retransmitted several times,
1404 	 * and receives a second error, give up now.  This is better
1405 	 * than waiting a long time to establish a connection that
1406 	 * can never complete.
1407 	 */
1408 	if (tp->t_state == TCPS_ESTABLISHED &&
1409 	     (error == EHOSTUNREACH || error == ENETUNREACH ||
1410 	      error == EHOSTDOWN)) {
1411 		return;
1412 	} else if (TCPS_HAVEESTABLISHED(tp->t_state) == 0 &&
1413 	    tp->t_rxtshift > 3 && tp->t_softerror)
1414 		so->so_error = error;
1415 	else
1416 		tp->t_softerror = error;
1417 	cv_broadcast(&so->so_cv);
1418 	sorwakeup(so);
1419 	sowwakeup(so);
1420 }
1421 
1422 #ifdef INET6
1423 void
1424 tcp6_notify(struct in6pcb *in6p, int error)
1425 {
1426 	struct tcpcb *tp = (struct tcpcb *)in6p->in6p_ppcb;
1427 	struct socket *so = in6p->in6p_socket;
1428 
1429 	/*
1430 	 * Ignore some errors if we are hooked up.
1431 	 * If connection hasn't completed, has retransmitted several times,
1432 	 * and receives a second error, give up now.  This is better
1433 	 * than waiting a long time to establish a connection that
1434 	 * can never complete.
1435 	 */
1436 	if (tp->t_state == TCPS_ESTABLISHED &&
1437 	     (error == EHOSTUNREACH || error == ENETUNREACH ||
1438 	      error == EHOSTDOWN)) {
1439 		return;
1440 	} else if (TCPS_HAVEESTABLISHED(tp->t_state) == 0 &&
1441 	    tp->t_rxtshift > 3 && tp->t_softerror)
1442 		so->so_error = error;
1443 	else
1444 		tp->t_softerror = error;
1445 	cv_broadcast(&so->so_cv);
1446 	sorwakeup(so);
1447 	sowwakeup(so);
1448 }
1449 #endif
1450 
1451 #ifdef INET6
1452 void *
1453 tcp6_ctlinput(int cmd, const struct sockaddr *sa, void *d)
1454 {
1455 	struct tcphdr th;
1456 	void (*notify)(struct in6pcb *, int) = tcp6_notify;
1457 	int nmatch;
1458 	struct ip6_hdr *ip6;
1459 	const struct sockaddr_in6 *sa6_src = NULL;
1460 	const struct sockaddr_in6 *sa6 = (const struct sockaddr_in6 *)sa;
1461 	struct mbuf *m;
1462 	int off;
1463 
1464 	if (sa->sa_family != AF_INET6 ||
1465 	    sa->sa_len != sizeof(struct sockaddr_in6))
1466 		return NULL;
1467 	if ((unsigned)cmd >= PRC_NCMDS)
1468 		return NULL;
1469 	else if (cmd == PRC_QUENCH) {
1470 		/*
1471 		 * Don't honor ICMP Source Quench messages meant for
1472 		 * TCP connections.
1473 		 */
1474 		return NULL;
1475 	} else if (PRC_IS_REDIRECT(cmd))
1476 		notify = in6_rtchange, d = NULL;
1477 	else if (cmd == PRC_MSGSIZE)
1478 		; /* special code is present, see below */
1479 	else if (cmd == PRC_HOSTDEAD)
1480 		d = NULL;
1481 	else if (inet6ctlerrmap[cmd] == 0)
1482 		return NULL;
1483 
1484 	/* if the parameter is from icmp6, decode it. */
1485 	if (d != NULL) {
1486 		struct ip6ctlparam *ip6cp = (struct ip6ctlparam *)d;
1487 		m = ip6cp->ip6c_m;
1488 		ip6 = ip6cp->ip6c_ip6;
1489 		off = ip6cp->ip6c_off;
1490 		sa6_src = ip6cp->ip6c_src;
1491 	} else {
1492 		m = NULL;
1493 		ip6 = NULL;
1494 		sa6_src = &sa6_any;
1495 		off = 0;
1496 	}
1497 
1498 	if (ip6) {
1499 		/*
1500 		 * XXX: We assume that when ip6 is non NULL,
1501 		 * M and OFF are valid.
1502 		 */
1503 
1504 		/* check if we can safely examine src and dst ports */
1505 		if (m->m_pkthdr.len < off + sizeof(th)) {
1506 			if (cmd == PRC_MSGSIZE)
1507 				icmp6_mtudisc_update((struct ip6ctlparam *)d, 0);
1508 			return NULL;
1509 		}
1510 
1511 		memset(&th, 0, sizeof(th));
1512 		m_copydata(m, off, sizeof(th), (void *)&th);
1513 
1514 		if (cmd == PRC_MSGSIZE) {
1515 			int valid = 0;
1516 
1517 			/*
1518 			 * Check to see if we have a valid TCP connection
1519 			 * corresponding to the address in the ICMPv6 message
1520 			 * payload.
1521 			 */
1522 			if (in6_pcblookup_connect(&tcbtable, &sa6->sin6_addr,
1523 			    th.th_dport,
1524 			    (const struct in6_addr *)&sa6_src->sin6_addr,
1525 						  th.th_sport, 0, 0))
1526 				valid++;
1527 
1528 			/*
1529 			 * Depending on the value of "valid" and routing table
1530 			 * size (mtudisc_{hi,lo}wat), we will:
1531 			 * - recalcurate the new MTU and create the
1532 			 *   corresponding routing entry, or
1533 			 * - ignore the MTU change notification.
1534 			 */
1535 			icmp6_mtudisc_update((struct ip6ctlparam *)d, valid);
1536 
1537 			/*
1538 			 * no need to call in6_pcbnotify, it should have been
1539 			 * called via callback if necessary
1540 			 */
1541 			return NULL;
1542 		}
1543 
1544 		nmatch = in6_pcbnotify(&tcbtable, sa, th.th_dport,
1545 		    (const struct sockaddr *)sa6_src, th.th_sport, cmd, NULL, notify);
1546 		if (nmatch == 0 && syn_cache_count &&
1547 		    (inet6ctlerrmap[cmd] == EHOSTUNREACH ||
1548 		     inet6ctlerrmap[cmd] == ENETUNREACH ||
1549 		     inet6ctlerrmap[cmd] == EHOSTDOWN))
1550 			syn_cache_unreach((const struct sockaddr *)sa6_src,
1551 					  sa, &th);
1552 	} else {
1553 		(void) in6_pcbnotify(&tcbtable, sa, 0,
1554 		    (const struct sockaddr *)sa6_src, 0, cmd, NULL, notify);
1555 	}
1556 
1557 	return NULL;
1558 }
1559 #endif
1560 
1561 #ifdef INET
1562 /* assumes that ip header and tcp header are contiguous on mbuf */
1563 void *
1564 tcp_ctlinput(int cmd, const struct sockaddr *sa, void *v)
1565 {
1566 	struct ip *ip = v;
1567 	struct tcphdr *th;
1568 	struct icmp *icp;
1569 	extern const int inetctlerrmap[];
1570 	void (*notify)(struct inpcb *, int) = tcp_notify;
1571 	int errno;
1572 	int nmatch;
1573 	struct tcpcb *tp;
1574 	u_int mtu;
1575 	tcp_seq seq;
1576 	struct inpcb *inp;
1577 #ifdef INET6
1578 	struct in6pcb *in6p;
1579 	struct in6_addr src6, dst6;
1580 #endif
1581 
1582 	if (sa->sa_family != AF_INET ||
1583 	    sa->sa_len != sizeof(struct sockaddr_in))
1584 		return NULL;
1585 	if ((unsigned)cmd >= PRC_NCMDS)
1586 		return NULL;
1587 	errno = inetctlerrmap[cmd];
1588 	if (cmd == PRC_QUENCH)
1589 		/*
1590 		 * Don't honor ICMP Source Quench messages meant for
1591 		 * TCP connections.
1592 		 */
1593 		return NULL;
1594 	else if (PRC_IS_REDIRECT(cmd))
1595 		notify = in_rtchange, ip = 0;
1596 	else if (cmd == PRC_MSGSIZE && ip && ip->ip_v == 4) {
1597 		/*
1598 		 * Check to see if we have a valid TCP connection
1599 		 * corresponding to the address in the ICMP message
1600 		 * payload.
1601 		 *
1602 		 * Boundary check is made in icmp_input(), with ICMP_ADVLENMIN.
1603 		 */
1604 		th = (struct tcphdr *)((char *)ip + (ip->ip_hl << 2));
1605 #ifdef INET6
1606 		memset(&src6, 0, sizeof(src6));
1607 		memset(&dst6, 0, sizeof(dst6));
1608 		src6.s6_addr16[5] = dst6.s6_addr16[5] = 0xffff;
1609 		memcpy(&src6.s6_addr32[3], &ip->ip_src, sizeof(struct in_addr));
1610 		memcpy(&dst6.s6_addr32[3], &ip->ip_dst, sizeof(struct in_addr));
1611 #endif
1612 		if ((inp = in_pcblookup_connect(&tcbtable, ip->ip_dst,
1613 						th->th_dport, ip->ip_src, th->th_sport, 0)) != NULL)
1614 #ifdef INET6
1615 			in6p = NULL;
1616 #else
1617 			;
1618 #endif
1619 #ifdef INET6
1620 		else if ((in6p = in6_pcblookup_connect(&tcbtable, &dst6,
1621 						       th->th_dport, &src6, th->th_sport, 0, 0)) != NULL)
1622 			;
1623 #endif
1624 		else
1625 			return NULL;
1626 
1627 		/*
1628 		 * Now that we've validated that we are actually communicating
1629 		 * with the host indicated in the ICMP message, locate the
1630 		 * ICMP header, recalculate the new MTU, and create the
1631 		 * corresponding routing entry.
1632 		 */
1633 		icp = (struct icmp *)((char *)ip -
1634 		    offsetof(struct icmp, icmp_ip));
1635 		if (inp) {
1636 			if ((tp = intotcpcb(inp)) == NULL)
1637 				return NULL;
1638 		}
1639 #ifdef INET6
1640 		else if (in6p) {
1641 			if ((tp = in6totcpcb(in6p)) == NULL)
1642 				return NULL;
1643 		}
1644 #endif
1645 		else
1646 			return NULL;
1647 		seq = ntohl(th->th_seq);
1648 		if (SEQ_LT(seq, tp->snd_una) || SEQ_GT(seq, tp->snd_max))
1649 			return NULL;
1650 		/*
1651 		 * If the ICMP message advertises a Next-Hop MTU
1652 		 * equal or larger than the maximum packet size we have
1653 		 * ever sent, drop the message.
1654 		 */
1655 		mtu = (u_int)ntohs(icp->icmp_nextmtu);
1656 		if (mtu >= tp->t_pmtud_mtu_sent)
1657 			return NULL;
1658 		if (mtu >= tcp_hdrsz(tp) + tp->t_pmtud_mss_acked) {
1659 			/*
1660 			 * Calculate new MTU, and create corresponding
1661 			 * route (traditional PMTUD).
1662 			 */
1663 			tp->t_flags &= ~TF_PMTUD_PEND;
1664 			icmp_mtudisc(icp, ip->ip_dst);
1665 		} else {
1666 			/*
1667 			 * Record the information got in the ICMP
1668 			 * message; act on it later.
1669 			 * If we had already recorded an ICMP message,
1670 			 * replace the old one only if the new message
1671 			 * refers to an older TCP segment
1672 			 */
1673 			if (tp->t_flags & TF_PMTUD_PEND) {
1674 				if (SEQ_LT(tp->t_pmtud_th_seq, seq))
1675 					return NULL;
1676 			} else
1677 				tp->t_flags |= TF_PMTUD_PEND;
1678 			tp->t_pmtud_th_seq = seq;
1679 			tp->t_pmtud_nextmtu = icp->icmp_nextmtu;
1680 			tp->t_pmtud_ip_len = icp->icmp_ip.ip_len;
1681 			tp->t_pmtud_ip_hl = icp->icmp_ip.ip_hl;
1682 		}
1683 		return NULL;
1684 	} else if (cmd == PRC_HOSTDEAD)
1685 		ip = 0;
1686 	else if (errno == 0)
1687 		return NULL;
1688 	if (ip && ip->ip_v == 4 && sa->sa_family == AF_INET) {
1689 		th = (struct tcphdr *)((char *)ip + (ip->ip_hl << 2));
1690 		nmatch = in_pcbnotify(&tcbtable, satocsin(sa)->sin_addr,
1691 		    th->th_dport, ip->ip_src, th->th_sport, errno, notify);
1692 		if (nmatch == 0 && syn_cache_count &&
1693 		    (inetctlerrmap[cmd] == EHOSTUNREACH ||
1694 		    inetctlerrmap[cmd] == ENETUNREACH ||
1695 		    inetctlerrmap[cmd] == EHOSTDOWN)) {
1696 			struct sockaddr_in sin;
1697 			memset(&sin, 0, sizeof(sin));
1698 			sin.sin_len = sizeof(sin);
1699 			sin.sin_family = AF_INET;
1700 			sin.sin_port = th->th_sport;
1701 			sin.sin_addr = ip->ip_src;
1702 			syn_cache_unreach((struct sockaddr *)&sin, sa, th);
1703 		}
1704 
1705 		/* XXX mapped address case */
1706 	} else
1707 		in_pcbnotifyall(&tcbtable, satocsin(sa)->sin_addr, errno,
1708 		    notify);
1709 	return NULL;
1710 }
1711 
1712 /*
1713  * When a source quench is received, we are being notified of congestion.
1714  * Close the congestion window down to the Loss Window (one segment).
1715  * We will gradually open it again as we proceed.
1716  */
1717 void
1718 tcp_quench(struct inpcb *inp, int errno)
1719 {
1720 	struct tcpcb *tp = intotcpcb(inp);
1721 
1722 	if (tp) {
1723 		tp->snd_cwnd = tp->t_segsz;
1724 		tp->t_bytes_acked = 0;
1725 	}
1726 }
1727 #endif
1728 
1729 #ifdef INET6
1730 void
1731 tcp6_quench(struct in6pcb *in6p, int errno)
1732 {
1733 	struct tcpcb *tp = in6totcpcb(in6p);
1734 
1735 	if (tp) {
1736 		tp->snd_cwnd = tp->t_segsz;
1737 		tp->t_bytes_acked = 0;
1738 	}
1739 }
1740 #endif
1741 
1742 #ifdef INET
1743 /*
1744  * Path MTU Discovery handlers.
1745  */
1746 void
1747 tcp_mtudisc_callback(struct in_addr faddr)
1748 {
1749 #ifdef INET6
1750 	struct in6_addr in6;
1751 #endif
1752 
1753 	in_pcbnotifyall(&tcbtable, faddr, EMSGSIZE, tcp_mtudisc);
1754 #ifdef INET6
1755 	memset(&in6, 0, sizeof(in6));
1756 	in6.s6_addr16[5] = 0xffff;
1757 	memcpy(&in6.s6_addr32[3], &faddr, sizeof(struct in_addr));
1758 	tcp6_mtudisc_callback(&in6);
1759 #endif
1760 }
1761 
1762 /*
1763  * On receipt of path MTU corrections, flush old route and replace it
1764  * with the new one.  Retransmit all unacknowledged packets, to ensure
1765  * that all packets will be received.
1766  */
1767 void
1768 tcp_mtudisc(struct inpcb *inp, int errno)
1769 {
1770 	struct tcpcb *tp = intotcpcb(inp);
1771 	struct rtentry *rt;
1772 
1773 	if (tp == NULL)
1774 		return;
1775 
1776 	rt = in_pcbrtentry(inp);
1777 	if (rt != NULL) {
1778 		/*
1779 		 * If this was not a host route, remove and realloc.
1780 		 */
1781 		if ((rt->rt_flags & RTF_HOST) == 0) {
1782 			in_rtchange(inp, errno);
1783 			if ((rt = in_pcbrtentry(inp)) == NULL)
1784 				return;
1785 		}
1786 
1787 		/*
1788 		 * Slow start out of the error condition.  We
1789 		 * use the MTU because we know it's smaller
1790 		 * than the previously transmitted segment.
1791 		 *
1792 		 * Note: This is more conservative than the
1793 		 * suggestion in draft-floyd-incr-init-win-03.
1794 		 */
1795 		if (rt->rt_rmx.rmx_mtu != 0)
1796 			tp->snd_cwnd =
1797 			    TCP_INITIAL_WINDOW(tcp_init_win,
1798 			    rt->rt_rmx.rmx_mtu);
1799 	}
1800 
1801 	/*
1802 	 * Resend unacknowledged packets.
1803 	 */
1804 	tp->snd_nxt = tp->sack_newdata = tp->snd_una;
1805 	tcp_output(tp);
1806 }
1807 #endif /* INET */
1808 
1809 #ifdef INET6
1810 /*
1811  * Path MTU Discovery handlers.
1812  */
1813 void
1814 tcp6_mtudisc_callback(struct in6_addr *faddr)
1815 {
1816 	struct sockaddr_in6 sin6;
1817 
1818 	memset(&sin6, 0, sizeof(sin6));
1819 	sin6.sin6_family = AF_INET6;
1820 	sin6.sin6_len = sizeof(struct sockaddr_in6);
1821 	sin6.sin6_addr = *faddr;
1822 	(void) in6_pcbnotify(&tcbtable, (struct sockaddr *)&sin6, 0,
1823 	    (const struct sockaddr *)&sa6_any, 0, PRC_MSGSIZE, NULL, tcp6_mtudisc);
1824 }
1825 
1826 void
1827 tcp6_mtudisc(struct in6pcb *in6p, int errno)
1828 {
1829 	struct tcpcb *tp = in6totcpcb(in6p);
1830 	struct rtentry *rt = in6_pcbrtentry(in6p);
1831 
1832 	if (tp != 0) {
1833 		if (rt != 0) {
1834 			/*
1835 			 * If this was not a host route, remove and realloc.
1836 			 */
1837 			if ((rt->rt_flags & RTF_HOST) == 0) {
1838 				in6_rtchange(in6p, errno);
1839 				if ((rt = in6_pcbrtentry(in6p)) == 0)
1840 					return;
1841 			}
1842 
1843 			/*
1844 			 * Slow start out of the error condition.  We
1845 			 * use the MTU because we know it's smaller
1846 			 * than the previously transmitted segment.
1847 			 *
1848 			 * Note: This is more conservative than the
1849 			 * suggestion in draft-floyd-incr-init-win-03.
1850 			 */
1851 			if (rt->rt_rmx.rmx_mtu != 0)
1852 				tp->snd_cwnd =
1853 				    TCP_INITIAL_WINDOW(tcp_init_win,
1854 				    rt->rt_rmx.rmx_mtu);
1855 		}
1856 
1857 		/*
1858 		 * Resend unacknowledged packets.
1859 		 */
1860 		tp->snd_nxt = tp->sack_newdata = tp->snd_una;
1861 		tcp_output(tp);
1862 	}
1863 }
1864 #endif /* INET6 */
1865 
1866 /*
1867  * Compute the MSS to advertise to the peer.  Called only during
1868  * the 3-way handshake.  If we are the server (peer initiated
1869  * connection), we are called with a pointer to the interface
1870  * on which the SYN packet arrived.  If we are the client (we
1871  * initiated connection), we are called with a pointer to the
1872  * interface out which this connection should go.
1873  *
1874  * NOTE: Do not subtract IP option/extension header size nor IPsec
1875  * header size from MSS advertisement.  MSS option must hold the maximum
1876  * segment size we can accept, so it must always be:
1877  *	 max(if mtu) - ip header - tcp header
1878  */
1879 u_long
1880 tcp_mss_to_advertise(const struct ifnet *ifp, int af)
1881 {
1882 	extern u_long in_maxmtu;
1883 	u_long mss = 0;
1884 	u_long hdrsiz;
1885 
1886 	/*
1887 	 * In order to avoid defeating path MTU discovery on the peer,
1888 	 * we advertise the max MTU of all attached networks as our MSS,
1889 	 * per RFC 1191, section 3.1.
1890 	 *
1891 	 * We provide the option to advertise just the MTU of
1892 	 * the interface on which we hope this connection will
1893 	 * be receiving.  If we are responding to a SYN, we
1894 	 * will have a pretty good idea about this, but when
1895 	 * initiating a connection there is a bit more doubt.
1896 	 *
1897 	 * We also need to ensure that loopback has a large enough
1898 	 * MSS, as the loopback MTU is never included in in_maxmtu.
1899 	 */
1900 
1901 	if (ifp != NULL)
1902 		switch (af) {
1903 		case AF_INET:
1904 			mss = ifp->if_mtu;
1905 			break;
1906 #ifdef INET6
1907 		case AF_INET6:
1908 			mss = IN6_LINKMTU(ifp);
1909 			break;
1910 #endif
1911 		}
1912 
1913 	if (tcp_mss_ifmtu == 0)
1914 		switch (af) {
1915 		case AF_INET:
1916 			mss = max(in_maxmtu, mss);
1917 			break;
1918 #ifdef INET6
1919 		case AF_INET6:
1920 			mss = max(in6_maxmtu, mss);
1921 			break;
1922 #endif
1923 		}
1924 
1925 	switch (af) {
1926 	case AF_INET:
1927 		hdrsiz = sizeof(struct ip);
1928 		break;
1929 #ifdef INET6
1930 	case AF_INET6:
1931 		hdrsiz = sizeof(struct ip6_hdr);
1932 		break;
1933 #endif
1934 	default:
1935 		hdrsiz = 0;
1936 		break;
1937 	}
1938 	hdrsiz += sizeof(struct tcphdr);
1939 	if (mss > hdrsiz)
1940 		mss -= hdrsiz;
1941 
1942 	mss = max(tcp_mssdflt, mss);
1943 	return (mss);
1944 }
1945 
1946 /*
1947  * Set connection variables based on the peer's advertised MSS.
1948  * We are passed the TCPCB for the actual connection.  If we
1949  * are the server, we are called by the compressed state engine
1950  * when the 3-way handshake is complete.  If we are the client,
1951  * we are called when we receive the SYN,ACK from the server.
1952  *
1953  * NOTE: Our advertised MSS value must be initialized in the TCPCB
1954  * before this routine is called!
1955  */
1956 void
1957 tcp_mss_from_peer(struct tcpcb *tp, int offer)
1958 {
1959 	struct socket *so;
1960 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH)
1961 	struct rtentry *rt;
1962 #endif
1963 	u_long bufsize;
1964 	int mss;
1965 
1966 #ifdef DIAGNOSTIC
1967 	if (tp->t_inpcb && tp->t_in6pcb)
1968 		panic("tcp_mss_from_peer: both t_inpcb and t_in6pcb are set");
1969 #endif
1970 	so = NULL;
1971 	rt = NULL;
1972 #ifdef INET
1973 	if (tp->t_inpcb) {
1974 		so = tp->t_inpcb->inp_socket;
1975 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH)
1976 		rt = in_pcbrtentry(tp->t_inpcb);
1977 #endif
1978 	}
1979 #endif
1980 #ifdef INET6
1981 	if (tp->t_in6pcb) {
1982 		so = tp->t_in6pcb->in6p_socket;
1983 #if defined(RTV_SPIPE) || defined(RTV_SSTHRESH)
1984 		rt = in6_pcbrtentry(tp->t_in6pcb);
1985 #endif
1986 	}
1987 #endif
1988 
1989 	/*
1990 	 * As per RFC1122, use the default MSS value, unless they
1991 	 * sent us an offer.  Do not accept offers less than 256 bytes.
1992 	 */
1993 	mss = tcp_mssdflt;
1994 	if (offer)
1995 		mss = offer;
1996 	mss = max(mss, 256);		/* sanity */
1997 	tp->t_peermss = mss;
1998 	mss -= tcp_optlen(tp);
1999 #ifdef INET
2000 	if (tp->t_inpcb)
2001 		mss -= ip_optlen(tp->t_inpcb);
2002 #endif
2003 #ifdef INET6
2004 	if (tp->t_in6pcb)
2005 		mss -= ip6_optlen(tp->t_in6pcb);
2006 #endif
2007 
2008 	/*
2009 	 * If there's a pipesize, change the socket buffer to that size.
2010 	 * Make the socket buffer an integral number of MSS units.  If
2011 	 * the MSS is larger than the socket buffer, artificially decrease
2012 	 * the MSS.
2013 	 */
2014 #ifdef RTV_SPIPE
2015 	if (rt != NULL && rt->rt_rmx.rmx_sendpipe != 0)
2016 		bufsize = rt->rt_rmx.rmx_sendpipe;
2017 	else
2018 #endif
2019 	{
2020 		KASSERT(so != NULL);
2021 		bufsize = so->so_snd.sb_hiwat;
2022 	}
2023 	if (bufsize < mss)
2024 		mss = bufsize;
2025 	else {
2026 		bufsize = roundup(bufsize, mss);
2027 		if (bufsize > sb_max)
2028 			bufsize = sb_max;
2029 		(void) sbreserve(&so->so_snd, bufsize, so);
2030 	}
2031 	tp->t_segsz = mss;
2032 
2033 #ifdef RTV_SSTHRESH
2034 	if (rt != NULL && rt->rt_rmx.rmx_ssthresh) {
2035 		/*
2036 		 * There's some sort of gateway or interface buffer
2037 		 * limit on the path.  Use this to set the slow
2038 		 * start threshold, but set the threshold to no less
2039 		 * than 2 * MSS.
2040 		 */
2041 		tp->snd_ssthresh = max(2 * mss, rt->rt_rmx.rmx_ssthresh);
2042 	}
2043 #endif
2044 }
2045 
2046 /*
2047  * Processing necessary when a TCP connection is established.
2048  */
2049 void
2050 tcp_established(struct tcpcb *tp)
2051 {
2052 	struct socket *so;
2053 #ifdef RTV_RPIPE
2054 	struct rtentry *rt;
2055 #endif
2056 	u_long bufsize;
2057 
2058 #ifdef DIAGNOSTIC
2059 	if (tp->t_inpcb && tp->t_in6pcb)
2060 		panic("tcp_established: both t_inpcb and t_in6pcb are set");
2061 #endif
2062 	so = NULL;
2063 	rt = NULL;
2064 #ifdef INET
2065 	/* This is a while() to reduce the dreadful stairstepping below */
2066 	while (tp->t_inpcb) {
2067 		so = tp->t_inpcb->inp_socket;
2068 #if defined(RTV_RPIPE)
2069 		rt = in_pcbrtentry(tp->t_inpcb);
2070 #endif
2071 		if (__predict_true(tcp_msl_enable)) {
2072 			if (tp->t_inpcb->inp_laddr.s_addr == INADDR_LOOPBACK) {
2073 				tp->t_msl = tcp_msl_loop ? tcp_msl_loop : (TCPTV_MSL >> 2);
2074 				break;
2075 			}
2076 
2077 			if (__predict_false(tcp_rttlocal)) {
2078 				/* This may be adjusted by tcp_input */
2079 				tp->t_msl = tcp_msl_local ? tcp_msl_local : (TCPTV_MSL >> 1);
2080 				break;
2081 			}
2082 			if (in_localaddr(tp->t_inpcb->inp_faddr)) {
2083 				tp->t_msl = tcp_msl_local ? tcp_msl_local : (TCPTV_MSL >> 1);
2084 				break;
2085 			}
2086 		}
2087 		tp->t_msl = tcp_msl_remote ? tcp_msl_remote : TCPTV_MSL;
2088 		break;
2089 	}
2090 #endif
2091 #ifdef INET6
2092 	/* The !tp->t_inpcb lets the compiler know it can't be v4 *and* v6 */
2093 	while (!tp->t_inpcb && tp->t_in6pcb) {
2094 		so = tp->t_in6pcb->in6p_socket;
2095 #if defined(RTV_RPIPE)
2096 		rt = in6_pcbrtentry(tp->t_in6pcb);
2097 #endif
2098 		if (__predict_true(tcp_msl_enable)) {
2099 			extern const struct in6_addr in6addr_loopback;
2100 
2101 			if (IN6_ARE_ADDR_EQUAL(&tp->t_in6pcb->in6p_laddr,
2102 					       &in6addr_loopback)) {
2103 				tp->t_msl = tcp_msl_loop ? tcp_msl_loop : (TCPTV_MSL >> 2);
2104 				break;
2105 			}
2106 
2107 			if (__predict_false(tcp_rttlocal)) {
2108 				/* This may be adjusted by tcp_input */
2109 				tp->t_msl = tcp_msl_local ? tcp_msl_local : (TCPTV_MSL >> 1);
2110 				break;
2111 			}
2112 			if (in6_localaddr(&tp->t_in6pcb->in6p_faddr)) {
2113 				tp->t_msl = tcp_msl_local ? tcp_msl_local : (TCPTV_MSL >> 1);
2114 				break;
2115 			}
2116 		}
2117 		tp->t_msl = tcp_msl_remote ? tcp_msl_remote : TCPTV_MSL;
2118 		break;
2119 	}
2120 #endif
2121 
2122 	tp->t_state = TCPS_ESTABLISHED;
2123 	TCP_TIMER_ARM(tp, TCPT_KEEP, tp->t_keepidle);
2124 
2125 #ifdef RTV_RPIPE
2126 	if (rt != NULL && rt->rt_rmx.rmx_recvpipe != 0)
2127 		bufsize = rt->rt_rmx.rmx_recvpipe;
2128 	else
2129 #endif
2130 	{
2131 		KASSERT(so != NULL);
2132 		bufsize = so->so_rcv.sb_hiwat;
2133 	}
2134 	if (bufsize > tp->t_ourmss) {
2135 		bufsize = roundup(bufsize, tp->t_ourmss);
2136 		if (bufsize > sb_max)
2137 			bufsize = sb_max;
2138 		(void) sbreserve(&so->so_rcv, bufsize, so);
2139 	}
2140 }
2141 
2142 /*
2143  * Check if there's an initial rtt or rttvar.  Convert from the
2144  * route-table units to scaled multiples of the slow timeout timer.
2145  * Called only during the 3-way handshake.
2146  */
2147 void
2148 tcp_rmx_rtt(struct tcpcb *tp)
2149 {
2150 #ifdef RTV_RTT
2151 	struct rtentry *rt = NULL;
2152 	int rtt;
2153 
2154 #ifdef DIAGNOSTIC
2155 	if (tp->t_inpcb && tp->t_in6pcb)
2156 		panic("tcp_rmx_rtt: both t_inpcb and t_in6pcb are set");
2157 #endif
2158 #ifdef INET
2159 	if (tp->t_inpcb)
2160 		rt = in_pcbrtentry(tp->t_inpcb);
2161 #endif
2162 #ifdef INET6
2163 	if (tp->t_in6pcb)
2164 		rt = in6_pcbrtentry(tp->t_in6pcb);
2165 #endif
2166 	if (rt == NULL)
2167 		return;
2168 
2169 	if (tp->t_srtt == 0 && (rtt = rt->rt_rmx.rmx_rtt)) {
2170 		/*
2171 		 * XXX The lock bit for MTU indicates that the value
2172 		 * is also a minimum value; this is subject to time.
2173 		 */
2174 		if (rt->rt_rmx.rmx_locks & RTV_RTT)
2175 			TCPT_RANGESET(tp->t_rttmin,
2176 			    rtt / (RTM_RTTUNIT / PR_SLOWHZ),
2177 			    TCPTV_MIN, TCPTV_REXMTMAX);
2178 		tp->t_srtt = rtt /
2179 		    ((RTM_RTTUNIT / PR_SLOWHZ) >> (TCP_RTT_SHIFT + 2));
2180 		if (rt->rt_rmx.rmx_rttvar) {
2181 			tp->t_rttvar = rt->rt_rmx.rmx_rttvar /
2182 			    ((RTM_RTTUNIT / PR_SLOWHZ) >>
2183 				(TCP_RTTVAR_SHIFT + 2));
2184 		} else {
2185 			/* Default variation is +- 1 rtt */
2186 			tp->t_rttvar =
2187 			    tp->t_srtt >> (TCP_RTT_SHIFT - TCP_RTTVAR_SHIFT);
2188 		}
2189 		TCPT_RANGESET(tp->t_rxtcur,
2190 		    ((tp->t_srtt >> 2) + tp->t_rttvar) >> (1 + 2),
2191 		    tp->t_rttmin, TCPTV_REXMTMAX);
2192 	}
2193 #endif
2194 }
2195 
2196 tcp_seq	 tcp_iss_seq = 0;	/* tcp initial seq # */
2197 
2198 /*
2199  * Get a new sequence value given a tcp control block
2200  */
2201 tcp_seq
2202 tcp_new_iss(struct tcpcb *tp, tcp_seq addin)
2203 {
2204 
2205 #ifdef INET
2206 	if (tp->t_inpcb != NULL) {
2207 		return (tcp_new_iss1(&tp->t_inpcb->inp_laddr,
2208 		    &tp->t_inpcb->inp_faddr, tp->t_inpcb->inp_lport,
2209 		    tp->t_inpcb->inp_fport, sizeof(tp->t_inpcb->inp_laddr),
2210 		    addin));
2211 	}
2212 #endif
2213 #ifdef INET6
2214 	if (tp->t_in6pcb != NULL) {
2215 		return (tcp_new_iss1(&tp->t_in6pcb->in6p_laddr,
2216 		    &tp->t_in6pcb->in6p_faddr, tp->t_in6pcb->in6p_lport,
2217 		    tp->t_in6pcb->in6p_fport, sizeof(tp->t_in6pcb->in6p_laddr),
2218 		    addin));
2219 	}
2220 #endif
2221 	/* Not possible. */
2222 	panic("tcp_new_iss");
2223 }
2224 
2225 static u_int8_t tcp_iss_secret[16];	/* 128 bits; should be plenty */
2226 
2227 /*
2228  * Initialize RFC 1948 ISS Secret
2229  */
2230 static int
2231 tcp_iss_secret_init(void)
2232 {
2233 	cprng_strong(kern_cprng,
2234 	    tcp_iss_secret, sizeof(tcp_iss_secret), 0);
2235 
2236 	return 0;
2237 }
2238 
2239 /*
2240  * This routine actually generates a new TCP initial sequence number.
2241  */
2242 tcp_seq
2243 tcp_new_iss1(void *laddr, void *faddr, u_int16_t lport, u_int16_t fport,
2244     size_t addrsz, tcp_seq addin)
2245 {
2246 	tcp_seq tcp_iss;
2247 
2248 	if (tcp_do_rfc1948) {
2249 		MD5_CTX ctx;
2250 		u_int8_t hash[16];	/* XXX MD5 knowledge */
2251 		static ONCE_DECL(tcp_iss_secret_control);
2252 
2253 		/*
2254 		 * If we haven't been here before, initialize our cryptographic
2255 		 * hash secret.
2256 		 */
2257 		RUN_ONCE(&tcp_iss_secret_control, tcp_iss_secret_init);
2258 
2259 		/*
2260 		 * Compute the base value of the ISS.  It is a hash
2261 		 * of (saddr, sport, daddr, dport, secret).
2262 		 */
2263 		MD5Init(&ctx);
2264 
2265 		MD5Update(&ctx, (u_char *) laddr, addrsz);
2266 		MD5Update(&ctx, (u_char *) &lport, sizeof(lport));
2267 
2268 		MD5Update(&ctx, (u_char *) faddr, addrsz);
2269 		MD5Update(&ctx, (u_char *) &fport, sizeof(fport));
2270 
2271 		MD5Update(&ctx, tcp_iss_secret, sizeof(tcp_iss_secret));
2272 
2273 		MD5Final(hash, &ctx);
2274 
2275 		memcpy(&tcp_iss, hash, sizeof(tcp_iss));
2276 
2277 		/*
2278 		 * Now increment our "timer", and add it in to
2279 		 * the computed value.
2280 		 *
2281 		 * XXX Use `addin'?
2282 		 * XXX TCP_ISSINCR too large to use?
2283 		 */
2284 		tcp_iss_seq += TCP_ISSINCR;
2285 #ifdef TCPISS_DEBUG
2286 		printf("ISS hash 0x%08x, ", tcp_iss);
2287 #endif
2288 		tcp_iss += tcp_iss_seq + addin;
2289 #ifdef TCPISS_DEBUG
2290 		printf("new ISS 0x%08x\n", tcp_iss);
2291 #endif
2292 	} else {
2293 		/*
2294 		 * Randomize.
2295 		 */
2296 		tcp_iss = cprng_fast32();
2297 
2298 		/*
2299 		 * If we were asked to add some amount to a known value,
2300 		 * we will take a random value obtained above, mask off
2301 		 * the upper bits, and add in the known value.  We also
2302 		 * add in a constant to ensure that we are at least a
2303 		 * certain distance from the original value.
2304 		 *
2305 		 * This is used when an old connection is in timed wait
2306 		 * and we have a new one coming in, for instance.
2307 		 */
2308 		if (addin != 0) {
2309 #ifdef TCPISS_DEBUG
2310 			printf("Random %08x, ", tcp_iss);
2311 #endif
2312 			tcp_iss &= TCP_ISS_RANDOM_MASK;
2313 			tcp_iss += addin + TCP_ISSINCR;
2314 #ifdef TCPISS_DEBUG
2315 			printf("Old ISS %08x, ISS %08x\n", addin, tcp_iss);
2316 #endif
2317 		} else {
2318 			tcp_iss &= TCP_ISS_RANDOM_MASK;
2319 			tcp_iss += tcp_iss_seq;
2320 			tcp_iss_seq += TCP_ISSINCR;
2321 #ifdef TCPISS_DEBUG
2322 			printf("ISS %08x\n", tcp_iss);
2323 #endif
2324 		}
2325 	}
2326 
2327 	if (tcp_compat_42) {
2328 		/*
2329 		 * Limit it to the positive range for really old TCP
2330 		 * implementations.
2331 		 * Just AND off the top bit instead of checking if
2332 		 * is set first - saves a branch 50% of the time.
2333 		 */
2334 		tcp_iss &= 0x7fffffff;		/* XXX */
2335 	}
2336 
2337 	return (tcp_iss);
2338 }
2339 
2340 #if defined(IPSEC)
2341 /* compute ESP/AH header size for TCP, including outer IP header. */
2342 size_t
2343 ipsec4_hdrsiz_tcp(struct tcpcb *tp)
2344 {
2345 	struct inpcb *inp;
2346 	size_t hdrsiz;
2347 
2348 	/* XXX mapped addr case (tp->t_in6pcb) */
2349 	if (!tp || !tp->t_template || !(inp = tp->t_inpcb))
2350 		return 0;
2351 	switch (tp->t_family) {
2352 	case AF_INET:
2353 		/* XXX: should use currect direction. */
2354 		hdrsiz = ipsec4_hdrsiz(tp->t_template, IPSEC_DIR_OUTBOUND, inp);
2355 		break;
2356 	default:
2357 		hdrsiz = 0;
2358 		break;
2359 	}
2360 
2361 	return hdrsiz;
2362 }
2363 
2364 #ifdef INET6
2365 size_t
2366 ipsec6_hdrsiz_tcp(struct tcpcb *tp)
2367 {
2368 	struct in6pcb *in6p;
2369 	size_t hdrsiz;
2370 
2371 	if (!tp || !tp->t_template || !(in6p = tp->t_in6pcb))
2372 		return 0;
2373 	switch (tp->t_family) {
2374 	case AF_INET6:
2375 		/* XXX: should use currect direction. */
2376 		hdrsiz = ipsec6_hdrsiz(tp->t_template, IPSEC_DIR_OUTBOUND, in6p);
2377 		break;
2378 	case AF_INET:
2379 		/* mapped address case - tricky */
2380 	default:
2381 		hdrsiz = 0;
2382 		break;
2383 	}
2384 
2385 	return hdrsiz;
2386 }
2387 #endif
2388 #endif /*IPSEC*/
2389 
2390 /*
2391  * Determine the length of the TCP options for this connection.
2392  *
2393  * XXX:  What do we do for SACK, when we add that?  Just reserve
2394  *       all of the space?  Otherwise we can't exactly be incrementing
2395  *       cwnd by an amount that varies depending on the amount we last
2396  *       had to SACK!
2397  */
2398 
2399 u_int
2400 tcp_optlen(struct tcpcb *tp)
2401 {
2402 	u_int optlen;
2403 
2404 	optlen = 0;
2405 	if ((tp->t_flags & (TF_REQ_TSTMP|TF_RCVD_TSTMP|TF_NOOPT)) ==
2406 	    (TF_REQ_TSTMP | TF_RCVD_TSTMP))
2407 		optlen += TCPOLEN_TSTAMP_APPA;
2408 
2409 #ifdef TCP_SIGNATURE
2410 	if (tp->t_flags & TF_SIGNATURE)
2411 		optlen += TCPOLEN_SIGNATURE + 2;
2412 #endif /* TCP_SIGNATURE */
2413 
2414 	return optlen;
2415 }
2416 
2417 u_int
2418 tcp_hdrsz(struct tcpcb *tp)
2419 {
2420 	u_int hlen;
2421 
2422 	switch (tp->t_family) {
2423 #ifdef INET6
2424 	case AF_INET6:
2425 		hlen = sizeof(struct ip6_hdr);
2426 		break;
2427 #endif
2428 	case AF_INET:
2429 		hlen = sizeof(struct ip);
2430 		break;
2431 	default:
2432 		hlen = 0;
2433 		break;
2434 	}
2435 	hlen += sizeof(struct tcphdr);
2436 
2437 	if ((tp->t_flags & (TF_REQ_TSTMP|TF_NOOPT)) == TF_REQ_TSTMP &&
2438 	    (tp->t_flags & TF_RCVD_TSTMP) == TF_RCVD_TSTMP)
2439 		hlen += TCPOLEN_TSTAMP_APPA;
2440 #ifdef TCP_SIGNATURE
2441 	if (tp->t_flags & TF_SIGNATURE)
2442 		hlen += TCPOLEN_SIGLEN;
2443 #endif
2444 	return hlen;
2445 }
2446 
2447 void
2448 tcp_statinc(u_int stat)
2449 {
2450 
2451 	KASSERT(stat < TCP_NSTATS);
2452 	TCP_STATINC(stat);
2453 }
2454 
2455 void
2456 tcp_statadd(u_int stat, uint64_t val)
2457 {
2458 
2459 	KASSERT(stat < TCP_NSTATS);
2460 	TCP_STATADD(stat, val);
2461 }
2462