2 * Copyright (c) 2003, 2004 Jeffrey M. Hsu. All rights reserved.
3 * Copyright (c) 2003, 2004 The DragonFly Project. All rights reserved.
5 * This code is derived from software contributed to The DragonFly Project
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in the
15 * documentation and/or other materials provided with the distribution.
16 * 3. Neither the name of The DragonFly Project nor the names of its
17 * contributors may be used to endorse or promote products derived
18 * from this software without specific, prior written permission.
20 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
21 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
22 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
23 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
24 * COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
25 * INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING,
26 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
27 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
28 * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
29 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
30 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35 * Copyright (c) 2003, 2004 Jeffrey M. Hsu. All rights reserved.
37 * License terms: all terms for the DragonFly license above plus the following:
39 * 4. All advertising materials mentioning features or use of this software
40 * must display the following acknowledgement:
42 * This product includes software developed by Jeffrey M. Hsu
43 * for the DragonFly Project.
45 * This requirement may be waived with permission from Jeffrey Hsu.
46 * This requirement will sunset and may be removed on July 8 2005,
47 * after which the standard DragonFly license (as shown above) will
52 * Copyright (c) 1982, 1986, 1993, 1994, 1995
53 * The Regents of the University of California. All rights reserved.
55 * Redistribution and use in source and binary forms, with or without
56 * modification, are permitted provided that the following conditions
58 * 1. Redistributions of source code must retain the above copyright
59 * notice, this list of conditions and the following disclaimer.
60 * 2. Redistributions in binary form must reproduce the above copyright
61 * notice, this list of conditions and the following disclaimer in the
62 * documentation and/or other materials provided with the distribution.
63 * 3. All advertising materials mentioning features or use of this software
64 * must display the following acknowledgement:
65 * This product includes software developed by the University of
66 * California, Berkeley and its contributors.
67 * 4. Neither the name of the University nor the names of its contributors
68 * may be used to endorse or promote products derived from this software
69 * without specific prior written permission.
71 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
72 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
73 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
74 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
75 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
76 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
77 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
78 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
79 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
80 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
83 * @(#)tcp_var.h 8.4 (Berkeley) 5/24/95
84 * $FreeBSD: src/sys/netinet/tcp_var.h,v 1.56.2.13 2003/02/03 02:34:07 hsu Exp $
85 * $DragonFly: src/sys/netinet/tcp_var.h,v 1.29 2005/03/04 05:57:50 hsu Exp $
88 #ifndef _NETINET_TCP_VAR_H_
89 #define _NETINET_TCP_VAR_H_
91 #include <netinet/in_pcb.h> /* needed for in_conninfo, inp_gen_t */
94 * Kernel variables for tcp.
96 extern int tcp_do_rfc1323;
97 extern int tcp_do_rfc1644;
98 extern int tcp_do_sack;
99 extern int tcp_do_smartsack;
101 /* TCP segment queue entry */
103 LIST_ENTRY(tseg_qent) tqe_q;
104 int tqe_len; /* TCP segment data length */
105 struct tcphdr *tqe_th; /* a pointer to tcp header */
106 struct mbuf *tqe_m; /* mbuf contains packet */
108 LIST_HEAD(tsegqe_head, tseg_qent);
109 extern int tcp_reass_maxseg;
110 extern int tcp_reass_qsize;
111 #ifdef MALLOC_DECLARE
112 MALLOC_DECLARE(M_TSEGQ);
116 u_char tt_ipgen[40]; /* the size must be of max ip header, now IPv6 */
120 #define tcp6cb tcpcb /* for KAME src sync over BSD*'s */
122 struct raw_sackblock { /* covers [start, end) */
127 /* maximum number of SACK blocks that will fit in the TCP option space */
128 #define MAX_SACK_REPORT_BLOCKS 4
130 TAILQ_HEAD(sackblock_list, sackblock);
134 struct sackblock_list sackblocks;
135 tcp_seq lostseq; /* passed SACK lost test */
136 struct sackblock *lastfound; /* search hint */
140 * Tcp control block, one per tcp; fields:
141 * Organized for 16 byte cacheline efficiency.
144 struct tsegqe_head t_segq;
145 int t_dupacks; /* consecutive dup acks recd */
146 int tt_cpu; /* sanity check the cpu */
147 struct callout *tt_rexmt; /* retransmit timer */
149 struct callout *tt_persist; /* retransmit persistence */
150 struct callout *tt_keep; /* keepalive */
151 struct callout *tt_2msl; /* 2*msl TIME_WAIT timer */
152 struct callout *tt_delack; /* delayed ACK timer */
154 struct inpcb *t_inpcb; /* back pointer to internet pcb */
155 int t_state; /* state of this connection */
157 #define TF_ACKNOW 0x00000001 /* ack peer immediately */
158 #define TF_DELACK 0x00000002 /* ack, but try to delay it */
159 #define TF_NODELAY 0x00000004 /* don't delay packets to coalesce */
160 #define TF_NOOPT 0x00000008 /* don't use tcp options */
161 #define TF_SENTFIN 0x00000010 /* have sent FIN */
162 #define TF_REQ_SCALE 0x00000020 /* have/will request window scaling */
163 #define TF_RCVD_SCALE 0x00000040 /* other side has requested scaling */
164 #define TF_REQ_TSTMP 0x00000080 /* have/will request timestamps */
165 #define TF_RCVD_TSTMP 0x00000100 /* a timestamp was received in SYN */
166 #define TF_SACK_PERMITTED 0x00000200 /* other side said I could SACK */
167 #define TF_NEEDSYN 0x00000400 /* send SYN (implicit state) */
168 #define TF_NEEDFIN 0x00000800 /* send FIN (implicit state) */
169 #define TF_NOPUSH 0x00001000 /* don't push */
170 #define TF_REQ_CC 0x00002000 /* have/will request CC */
171 #define TF_RCVD_CC 0x00004000 /* a CC was received in SYN */
172 #define TF_SENDCCNEW 0x00008000 /* send CCnew instead of CC in SYN */
173 #define TF_MORETOCOME 0x00010000 /* More data to be appended to sock */
174 #define TF_LQ_OVERFLOW 0x00020000 /* listen queue overflow */
175 #define TF_LASTIDLE 0x00040000 /* connection was previously idle */
176 #define TF_RXWIN0SENT 0x00080000 /* sent a receiver win 0 in response */
177 #define TF_FASTRECOVERY 0x00100000 /* in NewReno Fast Recovery */
178 #define TF_WASFRECOVERY 0x00200000 /* was in NewReno Fast Recovery */
179 #define TF_FIRSTACCACK 0x00400000 /* Look for 1st acceptable ACK. */
180 #define TF_FASTREXMT 0x00800000 /* Did Fast Retransmit. */
181 #define TF_EARLYREXMT 0x01000000 /* Did Early (Fast) Retransmit. */
182 #define TF_FORCE 0x02000000 /* Set if forcing out a byte */
183 #define TF_ONOUTPUTQ 0x04000000 /* on t_outputq list */
184 #define TF_DUPSEG 0x08000000 /* last seg a duplicate */
185 #define TF_ENCLOSESEG 0x10000000 /* enclosing SACK block */
186 #define TF_SACKLEFT 0x20000000 /* send SACK blocks from left side */
187 tcp_seq snd_up; /* send urgent pointer */
189 tcp_seq snd_una; /* send unacknowledged */
190 tcp_seq snd_recover; /* for use with NewReno Fast Recovery */
191 tcp_seq snd_max; /* highest sequence number sent;
192 * used to recognize retransmits
194 tcp_seq snd_nxt; /* send next */
196 tcp_seq snd_wl1; /* window update seg seq number */
197 tcp_seq snd_wl2; /* window update seg ack number */
198 tcp_seq iss; /* initial send sequence number */
199 tcp_seq irs; /* initial receive sequence number */
201 tcp_seq rcv_nxt; /* receive next */
202 tcp_seq rcv_adv; /* advertised window */
203 u_long rcv_wnd; /* receive window */
204 tcp_seq rcv_up; /* receive urgent pointer */
206 u_long snd_wnd; /* send window */
207 u_long snd_cwnd; /* congestion-controlled window */
208 u_long snd_bwnd; /* bandwidth-controlled window */
209 u_long snd_ssthresh; /* snd_cwnd size threshold for
210 * for slow start exponential to
213 u_long snd_bandwidth; /* calculated bandwidth or 0 */
215 u_int t_maxopd; /* mss plus options */
217 u_long t_rcvtime; /* inactivity time */
218 u_long t_starttime; /* time connection was established */
219 int t_rtttime; /* round trip time */
220 tcp_seq t_rtseq; /* sequence number being timed */
222 int t_bw_rtttime; /* used for bandwidth calculation */
223 tcp_seq t_bw_rtseq; /* used for bandwidth calculation */
225 int t_rxtcur; /* current retransmit value (ticks) */
226 u_int t_maxseg; /* maximum segment size */
227 int t_srtt; /* smoothed round-trip time */
228 int t_rttvar; /* variance in round-trip time */
230 int t_rxtshift; /* log(2) of rexmt exp. backoff */
231 u_int t_rttmin; /* minimum rtt allowed */
232 u_int t_rttbest; /* best rtt we've seen */
233 u_long t_rttupdated; /* number of times rtt sampled */
234 u_long max_sndwnd; /* largest window peer has offered */
236 int t_softerror; /* possible error not yet reported */
237 /* out-of-band data */
238 char t_oobflags; /* have some */
239 char t_iobc; /* input character */
240 #define TCPOOB_HAVEDATA 0x01
241 #define TCPOOB_HADDATA 0x02
242 /* RFC 1323 variables */
243 u_char snd_scale; /* window scaling for send window */
244 u_char rcv_scale; /* window scaling for recv window */
245 u_char request_r_scale; /* pending window scaling */
246 u_char requested_s_scale;
247 u_long ts_recent; /* timestamp echo data */
249 u_long ts_recent_age; /* when last updated */
250 tcp_seq last_ack_sent;
251 /* RFC 1644 variables */
252 tcp_cc cc_send; /* send connection count */
253 tcp_cc cc_recv; /* receive connection count */
255 u_long snd_cwnd_prev; /* cwnd prior to retransmit */
256 u_long snd_ssthresh_prev; /* ssthresh prior to retransmit */
257 tcp_seq snd_recover_prev; /* snd_recover prior to retransmit */
258 u_long t_badrxtwin; /* window for retransmit recovery */
259 u_long t_rexmtTS; /* timestamp of last retransmit */
260 u_char snd_limited; /* segments limited transmitted */
262 tcp_seq rexmt_high; /* highest seq # retransmitted + 1 */
263 tcp_seq snd_max_rexmt; /* snd_max when rexmting snd_una */
264 struct scoreboard scb; /* sack scoreboard */
265 struct raw_sackblock reportblk; /* incoming segment or D-SACK block */
266 struct raw_sackblock encloseblk;
268 struct raw_sackblock sackhistory[MAX_SACK_REPORT_BLOCKS]; /* reported */
269 TAILQ_ENTRY(tcpcb) t_outputq; /* tcp_output needed list */
272 #define IN_FASTRECOVERY(tp) (tp->t_flags & TF_FASTRECOVERY)
273 #define ENTER_FASTRECOVERY(tp) tp->t_flags |= TF_FASTRECOVERY
274 #define EXIT_FASTRECOVERY(tp) tp->t_flags &= ~TF_FASTRECOVERY
280 #define tcpstat tcpstats_ary[_GD->gd_cpuid]
282 #define tcpstat tcpstats_ary[0]
286 extern struct tcp_stats tcpstats_ary[MAXCPU];
288 static const int tcprexmtthresh = 3;
295 u_long tcps_connattempt; /* connections initiated */
296 u_long tcps_accepts; /* connections accepted */
297 u_long tcps_connects; /* connections established */
298 u_long tcps_drops; /* connections dropped */
299 u_long tcps_conndrops; /* embryonic connections dropped */
300 u_long tcps_closed; /* conn. closed (includes drops) */
301 u_long tcps_segstimed; /* segs where we tried to get rtt */
302 u_long tcps_rttupdated; /* times we succeeded */
303 u_long tcps_delack; /* delayed acks sent */
304 u_long tcps_timeoutdrop; /* conn. dropped in rxmt timeout */
305 u_long tcps_rexmttimeo; /* retransmit timeouts */
306 u_long tcps_persisttimeo; /* persist timeouts */
307 u_long tcps_keeptimeo; /* keepalive timeouts */
308 u_long tcps_keepprobe; /* keepalive probes sent */
309 u_long tcps_keepdrops; /* connections dropped in keepalive */
311 u_long tcps_sndtotal; /* total packets sent */
312 u_long tcps_sndpack; /* data packets sent */
313 u_long tcps_sndbyte; /* data bytes sent */
314 u_long tcps_sndrexmitpack; /* data packets retransmitted */
315 u_long tcps_sndrexmitbyte; /* data bytes retransmitted */
316 u_long tcps_sndfastrexmit; /* Fast Retransmissions */
317 u_long tcps_sndearlyrexmit; /* early Fast Retransmissions */
318 u_long tcps_sndlimited; /* Limited Transmit packets */
319 u_long tcps_sndrtobad; /* spurious RTO retransmissions */
320 u_long tcps_sndfastrexmitbad; /* spurious Fast Retransmissions */
321 u_long tcps_sndearlyrexmitbad; /* spurious early Fast Retransmissions,
322 a subset of tcps_sndfastrexmitbad */
323 u_long tcps_eifeldetected; /* Eifel-detected spurious rexmits */
324 u_long tcps_rttcantdetect; /* Eifel but not 1/2 RTT-detectable */
325 u_long tcps_rttdetected; /* RTT-detected spurious RTO rexmits */
326 u_long tcps_sndacks; /* ack-only packets sent */
327 u_long tcps_sndprobe; /* window probes sent */
328 u_long tcps_sndurg; /* packets sent with URG only */
329 u_long tcps_sndwinup; /* window update-only packets sent */
330 u_long tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */
331 u_long tcps_sndsackpack; /* packets sent by SACK recovery alg */
332 u_long tcps_sndsackbyte; /* bytes sent by SACK recovery */
333 u_long tcps_snduna3; /* re-retransmit snd_una on 3 new seg */
334 u_long tcps_snduna1; /* re-retransmit snd_una on 1 new seg */
336 u_long tcps_rcvtotal; /* total packets received */
337 u_long tcps_rcvpack; /* packets received in sequence */
338 u_long tcps_rcvbyte; /* bytes received in sequence */
339 u_long tcps_rcvbadsum; /* packets received with ccksum errs */
340 u_long tcps_rcvbadoff; /* packets received with bad offset */
341 u_long tcps_rcvmemdrop; /* packets dropped for lack of memory */
342 u_long tcps_rcvshort; /* packets received too short */
343 u_long tcps_rcvduppack; /* duplicate-only packets received */
344 u_long tcps_rcvdupbyte; /* duplicate-only bytes received */
345 u_long tcps_rcvpartduppack; /* packets with some duplicate data */
346 u_long tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */
347 u_long tcps_rcvoopack; /* out-of-order packets received */
348 u_long tcps_rcvoobyte; /* out-of-order bytes received */
349 u_long tcps_rcvpackafterwin; /* packets with data after window */
350 u_long tcps_rcvbyteafterwin; /* bytes rcvd after window */
351 u_long tcps_rcvafterclose; /* packets rcvd after "close" */
352 u_long tcps_rcvwinprobe; /* rcvd window probe packets */
353 u_long tcps_rcvdupack; /* rcvd duplicate acks */
354 u_long tcps_rcvacktoomuch; /* rcvd acks for unsent data */
355 u_long tcps_rcvackpack; /* rcvd ack packets */
356 u_long tcps_rcvackbyte; /* bytes acked by rcvd acks */
357 u_long tcps_rcvwinupd; /* rcvd window update packets */
358 u_long tcps_pawsdrop; /* segments dropped due to PAWS */
359 u_long tcps_predack; /* times hdr predict ok for acks */
360 u_long tcps_preddat; /* times hdr predict ok for data pkts */
361 u_long tcps_pcbcachemiss;
362 u_long tcps_cachedrtt; /* times cached RTT in route updated */
363 u_long tcps_cachedrttvar; /* times cached rttvar updated */
364 u_long tcps_cachedssthresh; /* times cached ssthresh updated */
365 u_long tcps_usedrtt; /* times RTT initialized from route */
366 u_long tcps_usedrttvar; /* times RTTVAR initialized from rt */
367 u_long tcps_usedssthresh; /* times ssthresh initialized from rt*/
368 u_long tcps_persistdrop; /* timeout in persist state */
369 u_long tcps_badsyn; /* bogus SYN, e.g. premature ACK */
370 u_long tcps_mturesent; /* resends due to MTU discovery */
371 u_long tcps_listendrop; /* listen queue overflows */
373 u_long tcps_sc_added; /* entry added to syncache */
374 u_long tcps_sc_retransmitted; /* syncache entry was retransmitted */
375 u_long tcps_sc_dupsyn; /* duplicate SYN packet */
376 u_long tcps_sc_dropped; /* could not reply to packet */
377 u_long tcps_sc_completed; /* successful extraction of entry */
378 u_long tcps_sc_bucketoverflow; /* syncache per-bucket limit hit */
379 u_long tcps_sc_cacheoverflow; /* syncache cache limit hit */
380 u_long tcps_sc_reset; /* RST removed entry from syncache */
381 u_long tcps_sc_stale; /* timed out or listen socket gone */
382 u_long tcps_sc_aborted; /* syncache entry aborted */
383 u_long tcps_sc_badack; /* removed due to bad ACK */
384 u_long tcps_sc_unreach; /* ICMP unreachable received */
385 u_long tcps_sc_zonefail; /* zalloc() failed */
386 u_long tcps_sc_sendcookie; /* SYN cookie sent */
387 u_long tcps_sc_recvcookie; /* SYN cookie received */
391 * Structure to hold TCP options that are only used during segment
392 * processing (in tcp_input), but not held in the tcpcb.
393 * It's basically used to reduce the number of parameters
397 u_long to_flags; /* which options are present */
398 #define TOF_TS 0x0001 /* timestamp */
399 #define TOF_CC 0x0002 /* CC and CCnew are exclusive */
400 #define TOF_CCNEW 0x0004
401 #define TOF_CCECHO 0x0008
402 #define TOF_MSS 0x0010
403 #define TOF_SCALE 0x0020
404 #define TOF_SACK_PERMITTED 0x0040
405 #define TOF_SACK 0x0080
408 tcp_cc to_cc; /* holds CC or CCnew */
411 u_int8_t to_requested_s_scale;
414 struct raw_sackblock *to_sackblocks;
418 inp_gen_t sc_inp_gencnt; /* pointer check */
419 struct tcpcb *sc_tp; /* tcb for listening socket */
420 struct mbuf *sc_ipopts; /* source route */
421 struct in_conninfo sc_inc; /* addresses */
422 #define sc_route sc_inc.inc_route
423 #define sc_route6 sc_inc.inc6_route
424 u_int32_t sc_tsrecent;
425 tcp_cc sc_cc_send; /* holds CC or CCnew */
427 tcp_seq sc_irs; /* seq from peer */
428 tcp_seq sc_iss; /* our ISS */
429 u_long sc_rxttime; /* retransmit time */
430 u_int16_t sc_rxtslot; /* retransmit counter */
431 u_int16_t sc_peer_mss; /* peer's MSS */
432 u_int16_t sc_wnd; /* advertised window */
433 u_int8_t sc_requested_s_scale:4,
434 sc_request_r_scale:4;
436 #define SCF_NOOPT 0x01 /* no TCP options */
437 #define SCF_WINSCALE 0x02 /* negotiated window scaling */
438 #define SCF_TIMESTAMP 0x04 /* negotiated timestamps */
439 #define SCF_CC 0x08 /* negotiated CC */
440 #define SCF_UNREACH 0x10 /* icmp unreachable received */
441 #define SCF_SACK_PERMITTED 0x20 /* saw SACK permitted option */
442 TAILQ_ENTRY(syncache) sc_hash;
443 TAILQ_ENTRY(syncache) sc_timerq;
446 struct syncache_head {
447 TAILQ_HEAD(, syncache) sch_bucket;
452 * The TAO cache entry which is stored in the protocol family specific
453 * portion of the route metrics.
456 tcp_cc tao_cc; /* latest CC in valid SYN */
457 tcp_cc tao_ccsent; /* latest CC sent to peer */
458 u_short tao_mssopt; /* peer's cached MSS */
461 u_short tao_flags; /* cache status flags */
462 #define TAOF_DONT 0x0001 /* peer doesn't understand rfc1644 */
463 #define TAOF_OK 0x0002 /* peer does understand rfc1644 */
464 #define TAOF_UNDEF 0 /* we don't know yet */
468 #define rmx_taop(rt) ((struct rmxp_tao *)(rt).rmx_filler)
470 #define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb)
471 #define sototcpcb(so) (intotcpcb(sotoinpcb(so)))
474 * The smoothed round-trip time and estimated variance
475 * are stored as fixed point numbers scaled by the values below.
476 * For convenience, these scales are also used in smoothing the average
477 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed).
478 * With these scales, srtt has 3 bits to the right of the binary point,
479 * and thus an "ALPHA" of 0.875. rttvar has 2 bits to the right of the
480 * binary point, and is smoothed with an ALPHA of 0.75.
482 #define TCP_RTT_SCALE 32 /* multiplier for srtt; 3 bits frac. */
483 #define TCP_RTT_SHIFT 5 /* shift for srtt; 3 bits frac. */
484 #define TCP_RTTVAR_SCALE 16 /* multiplier for rttvar; 2 bits */
485 #define TCP_RTTVAR_SHIFT 4 /* shift for rttvar; 2 bits */
486 #define TCP_DELTA_SHIFT 2 /* see tcp_input.c */
489 * The initial retransmission should happen at rtt + 4 * rttvar.
490 * Because of the way we do the smoothing, srtt and rttvar
491 * will each average +1/2 tick of bias. When we compute
492 * the retransmit timer, we want 1/2 tick of rounding and
493 * 1 extra tick because of +-1/2 tick uncertainty in the
494 * firing of the timer. The bias will give us exactly the
495 * 1.5 tick we need. But, because the bias is
496 * statistical, we have to test that we don't drop below
497 * the minimum feasible timer (which is 2 ticks).
498 * This version of the macro adapted from a paper by Lawrence
499 * Brakmo and Larry Peterson which outlines a problem caused
500 * by insufficient precision in the original implementation,
501 * which results in inappropriately large RTO values for very
504 #define TCP_REXMTVAL(tp) \
505 max((tp)->t_rttmin, (((tp)->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT)) \
506 + (tp)->t_rttvar) >> TCP_DELTA_SHIFT)
509 * TCB structure exported to user-land via sysctl(3).
510 * Evil hack: declare only if in_pcb.h and sys/socketvar.h have been
511 * included. Not all of our clients do.
513 #if defined(_NETINET_IN_PCB_H_) && defined(_SYS_SOCKETVAR_H_)
518 struct xsocket xt_socket;
519 u_quad_t xt_alignment_hack;
524 * Names for TCP sysctl objects
526 #define TCPCTL_DO_RFC1323 1 /* use RFC-1323 extensions */
527 #define TCPCTL_DO_RFC1644 2 /* use RFC-1644 extensions */
528 #define TCPCTL_MSSDFLT 3 /* MSS default */
529 #define TCPCTL_STATS 4 /* statistics (read-only) */
530 #define TCPCTL_RTTDFLT 5 /* default RTT estimate */
531 #define TCPCTL_KEEPIDLE 6 /* keepalive idle timer */
532 #define TCPCTL_KEEPINTVL 7 /* interval to send keepalives */
533 #define TCPCTL_SENDSPACE 8 /* send buffer space */
534 #define TCPCTL_RECVSPACE 9 /* receive buffer space */
535 #define TCPCTL_KEEPINIT 10 /* timeout for establishing syn */
536 #define TCPCTL_PCBLIST 11 /* list of all outstanding PCBs */
537 #define TCPCTL_DELACKTIME 12 /* time before sending delayed ACK */
538 #define TCPCTL_V6MSSDFLT 13 /* MSS default for IPv6 */
539 #define TCPCTL_MAXID 14
541 #define TCPCTL_NAMES { \
543 { "rfc1323", CTLTYPE_INT }, \
544 { "rfc1644", CTLTYPE_INT }, \
545 { "mssdflt", CTLTYPE_INT }, \
546 { "stats", CTLTYPE_STRUCT }, \
547 { "rttdflt", CTLTYPE_INT }, \
548 { "keepidle", CTLTYPE_INT }, \
549 { "keepintvl", CTLTYPE_INT }, \
550 { "sendspace", CTLTYPE_INT }, \
551 { "recvspace", CTLTYPE_INT }, \
552 { "keepinit", CTLTYPE_INT }, \
553 { "pcblist", CTLTYPE_STRUCT }, \
554 { "delacktime", CTLTYPE_INT }, \
555 { "v6mssdflt", CTLTYPE_INT }, \
560 SYSCTL_DECL(_net_inet_tcp);
563 #define TCP_DO_SACK(tp) ((tp)->t_flags & TF_SACK_PERMITTED)
565 TAILQ_HEAD(tcpcbackqhead,tcpcb);
567 extern struct inpcbinfo tcbinfo[];
568 extern struct tcpcbackqhead tcpcbackq[];
570 extern int tcp_mssdflt; /* XXX */
571 extern int tcp_delack_enabled;
572 extern int path_mtu_discovery;
574 int tcp_addrcpu(in_addr_t faddr, in_port_t fport,
575 in_addr_t laddr, in_port_t lport);
577 tcp_addrport(in_addr_t faddr, in_port_t fport,
578 in_addr_t laddr, in_port_t lport);
579 void tcp_canceltimers (struct tcpcb *);
581 tcp_close (struct tcpcb *);
582 void tcpmsg_service_loop (void *);
583 void tcp_ctlinput (int, struct sockaddr *, void *);
584 int tcp_ctloutput (struct socket *, struct sockopt *);
588 tcp_drop (struct tcpcb *, int);
589 void tcp_drain (void);
590 void tcp_fasttimo (void);
592 tcp_gettaocache (struct in_conninfo *);
593 void tcp_init (void);
594 void tcp_thread_init (void);
595 void tcp_input (struct mbuf *, ...);
596 void tcp_mss (struct tcpcb *, int);
597 int tcp_mssopt (struct tcpcb *);
598 void tcp_drop_syn_sent (struct inpcb *, int);
599 void tcp_mtudisc (struct inpcb *, int);
601 tcp_newtcpcb (struct inpcb *);
602 int tcp_output (struct tcpcb *);
603 void tcp_quench (struct inpcb *, int);
604 void tcp_respond (struct tcpcb *, void *,
605 struct tcphdr *, struct mbuf *, tcp_seq, tcp_seq, int);
607 tcp_rtlookup (struct in_conninfo *);
608 void tcp_sack_cleanup(struct scoreboard *scb);
609 int tcp_sack_ndsack_blocks(struct raw_sackblock *blocks,
610 const int numblocks, tcp_seq snd_una);
611 void tcp_sack_fill_report(struct tcpcb *tp, u_char *opt, u_int *plen);
613 tcp_sack_has_sacked(struct scoreboard *scb, u_int amount);
614 void tcp_sack_init(void);
615 void tcp_sack_tcpcb_init(struct tcpcb *tp);
616 uint32_t tcp_sack_compute_pipe(struct tcpcb *tp);
618 tcp_sack_nextseg(struct tcpcb *tp, tcp_seq *nextrexmt, uint32_t *len,
621 void tcp_sack_revert_scoreboard(struct scoreboard *scb, tcp_seq snd_una,
623 void tcp_sack_save_scoreboard(struct scoreboard *scb);
625 void tcp_sack_skip_sacked(struct scoreboard *scb, tcp_seq *prexmt);
626 void tcp_sack_update_scoreboard(struct tcpcb *tp, struct tcpopt *to);
627 void tcp_save_congestion_state(struct tcpcb *tp);
628 void tcp_revert_congestion_state(struct tcpcb *tp);
629 void tcp_setpersist (struct tcpcb *);
630 void tcp_slowtimo (void);
631 struct tcptemp *tcp_maketemplate (struct tcpcb *);
632 void tcp_freetemplate (struct tcptemp *);
633 void tcp_fillheaders (struct tcpcb *, void *, void *);
635 tcp_soport(struct socket *, struct sockaddr *nam, int req);
637 tcp_timers (struct tcpcb *, int);
638 void tcp_trace (int, int, struct tcpcb *, void *, struct tcphdr *,
640 void tcp_xmit_bandwidth_limit(struct tcpcb *tp, tcp_seq ack_seq);
641 void syncache_init(void);
642 void syncache_unreach(struct in_conninfo *, struct tcphdr *);
643 int syncache_expand(struct in_conninfo *, struct tcphdr *,
644 struct socket **, struct mbuf *);
645 int syncache_add(struct in_conninfo *, struct tcpopt *,
646 struct tcphdr *, struct socket **, struct mbuf *);
647 void syncache_chkrst(struct in_conninfo *, struct tcphdr *);
648 void syncache_badack(struct in_conninfo *);
650 extern struct pr_usrreqs tcp_usrreqs;
651 extern u_long tcp_sendspace;
652 extern u_long tcp_recvspace;
653 tcp_seq tcp_new_isn (struct tcpcb *);
657 #endif /* _NETINET_TCP_VAR_H_ */