2 * Copyright (c) 1982, 1986, 1993, 1994, 1995
3 * The Regents of the University of California. All rights reserved.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
13 * 4. Neither the name of the University nor the names of its contributors
14 * may be used to endorse or promote products derived from this software
15 * without specific prior written permission.
17 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * @(#)tcp_var.h 8.4 (Berkeley) 5/24/95
33 #ifndef _NETINET_TCP_VAR_H_
34 #define _NETINET_TCP_VAR_H_
36 #include <netinet/tcp.h>
41 * Kernel variables for tcp.
44 extern int tcp_do_rfc1323;
47 /* TCP segment queue entry */
49 LIST_ENTRY(tseg_qent) tqe_q;
50 int tqe_len; /* TCP segment data length */
51 struct tcphdr *tqe_th; /* a pointer to tcp header */
52 struct mbuf *tqe_m; /* mbuf contains packet */
54 LIST_HEAD(tsegqe_head, tseg_qent);
56 extern int tcp_reass_qsize;
58 extern struct uma_zone *tcp_reass_zone;
61 tcp_seq start; /* start seq no. of sack block */
62 tcp_seq end; /* end seq no. */
66 tcp_seq start; /* start seq no. of hole */
67 tcp_seq end; /* end seq no. */
68 tcp_seq rxmit; /* next seq. no in hole to be retransmitted */
69 TAILQ_ENTRY(sackhole) scblink; /* scoreboard linkage */
73 struct sackhole *nexthole;
74 int sack_bytes_rexmit;
76 int ispare; /* explicit pad for 64bit alignment */
77 uint64_t _pad[2]; /* 1 sacked_bytes, 1 TBD */
81 u_char tt_ipgen[40]; /* the size must be of max ip header, now IPv6 */
85 #define tcp6cb tcpcb /* for KAME src sync over BSD*'s */
87 /* Neighbor Discovery, Neighbor Unreachability Detection Upper layer hint. */
89 #define ND6_HINT(tp) \
91 if ((tp) && (tp)->t_inpcb && \
92 ((tp)->t_inpcb->inp_vflag & INP_IPV6) != 0) \
93 nd6_nud_hint(NULL, NULL, 0); \
100 * Tcp control block, one per tcp; fields:
101 * Organized for 16 byte cacheline efficiency.
104 struct tsegqe_head t_segq; /* segment reassembly queue */
105 void *t_pspare[2]; /* new reassembly queue */
106 int t_segqlen; /* segment reassembly queue length */
107 int t_dupacks; /* consecutive dup acks recd */
109 struct tcp_timer *t_timers; /* All the TCP timers in one struct */
111 struct inpcb *t_inpcb; /* back pointer to internet pcb */
112 int t_state; /* state of this connection */
115 struct vnet *t_vnet; /* back pointer to parent vnet */
117 tcp_seq snd_una; /* send unacknowledged */
118 tcp_seq snd_max; /* highest sequence number sent;
119 * used to recognize retransmits
121 tcp_seq snd_nxt; /* send next */
122 tcp_seq snd_up; /* send urgent pointer */
124 tcp_seq snd_wl1; /* window update seg seq number */
125 tcp_seq snd_wl2; /* window update seg ack number */
126 tcp_seq iss; /* initial send sequence number */
127 tcp_seq irs; /* initial receive sequence number */
129 tcp_seq rcv_nxt; /* receive next */
130 tcp_seq rcv_adv; /* advertised window */
131 u_long rcv_wnd; /* receive window */
132 tcp_seq rcv_up; /* receive urgent pointer */
134 u_long snd_wnd; /* send window */
135 u_long snd_cwnd; /* congestion-controlled window */
136 u_long snd_bwnd; /* bandwidth-controlled window */
137 u_long snd_ssthresh; /* snd_cwnd size threshold for
138 * for slow start exponential to
141 u_long snd_bandwidth; /* calculated bandwidth or 0 */
142 tcp_seq snd_recover; /* for use in NewReno Fast Recovery */
144 u_int t_maxopd; /* mss plus options */
146 u_int t_rcvtime; /* inactivity time */
147 u_int t_starttime; /* time connection was established */
148 u_int t_rtttime; /* RTT measurement start time */
149 tcp_seq t_rtseq; /* sequence number being timed */
151 u_int t_bw_rtttime; /* used for bandwidth calculation */
152 tcp_seq t_bw_rtseq; /* used for bandwidth calculation */
154 int t_rxtcur; /* current retransmit value (ticks) */
155 u_int t_maxseg; /* maximum segment size */
156 int t_srtt; /* smoothed round-trip time */
157 int t_rttvar; /* variance in round-trip time */
159 int t_rxtshift; /* log(2) of rexmt exp. backoff */
160 u_int t_rttmin; /* minimum rtt allowed */
161 u_int t_rttbest; /* best rtt we've seen */
162 u_long t_rttupdated; /* number of times rtt sampled */
163 u_long max_sndwnd; /* largest window peer has offered */
165 int t_softerror; /* possible error not yet reported */
166 /* out-of-band data */
167 char t_oobflags; /* have some */
168 char t_iobc; /* input character */
169 /* RFC 1323 variables */
170 u_char snd_scale; /* window scaling for send window */
171 u_char rcv_scale; /* window scaling for recv window */
172 u_char request_r_scale; /* pending window scaling */
173 u_int32_t ts_recent; /* timestamp echo data */
174 u_int ts_recent_age; /* when last updated */
175 u_int32_t ts_offset; /* our timestamp offset */
177 tcp_seq last_ack_sent;
179 u_long snd_cwnd_prev; /* cwnd prior to retransmit */
180 u_long snd_ssthresh_prev; /* ssthresh prior to retransmit */
181 tcp_seq snd_recover_prev; /* snd_recover prior to retransmit */
182 u_int t_badrxtwin; /* window for retransmit recovery */
183 u_char snd_limited; /* segments limited transmitted */
184 /* SACK related state */
185 int snd_numholes; /* number of holes seen by sender */
186 TAILQ_HEAD(sackhole_head, sackhole) snd_holes;
187 /* SACK scoreboard (sorted) */
188 tcp_seq snd_fack; /* last seq number(+1) sack'd by rcv'r*/
189 int rcv_numsacks; /* # distinct sack blks present */
190 struct sackblk sackblks[MAX_SACK_BLKS]; /* seq nos. of sack blocks */
191 tcp_seq sack_newdata; /* New data xmitted in this recovery
192 episode starts at this seq number */
193 struct sackhint sackhint; /* SACK scoreboard hint */
194 int t_rttlow; /* smallest observerved RTT */
195 u_int32_t rfbuf_ts; /* recv buffer autoscaling timestamp */
196 int rfbuf_cnt; /* recv buffer autoscaling byte count */
197 struct toe_usrreqs *t_tu; /* offload operations vector */
198 void *t_toe; /* TOE pcb pointer */
199 int t_bytes_acked; /* # bytes acked during current RTT */
201 int t_ispare; /* explicit pad for 64bit alignment */
202 void *t_pspare2[6]; /* 2 CC / 4 TBD */
203 uint64_t _pad[12]; /* 7 UTO, 5 TBD (1-2 CC/RTT?) */
207 * Flags and utility macros for the t_flags field.
209 #define TF_ACKNOW 0x000001 /* ack peer immediately */
210 #define TF_DELACK 0x000002 /* ack, but try to delay it */
211 #define TF_NODELAY 0x000004 /* don't delay packets to coalesce */
212 #define TF_NOOPT 0x000008 /* don't use tcp options */
213 #define TF_SENTFIN 0x000010 /* have sent FIN */
214 #define TF_REQ_SCALE 0x000020 /* have/will request window scaling */
215 #define TF_RCVD_SCALE 0x000040 /* other side has requested scaling */
216 #define TF_REQ_TSTMP 0x000080 /* have/will request timestamps */
217 #define TF_RCVD_TSTMP 0x000100 /* a timestamp was received in SYN */
218 #define TF_SACK_PERMIT 0x000200 /* other side said I could SACK */
219 #define TF_NEEDSYN 0x000400 /* send SYN (implicit state) */
220 #define TF_NEEDFIN 0x000800 /* send FIN (implicit state) */
221 #define TF_NOPUSH 0x001000 /* don't push */
222 #define TF_MORETOCOME 0x010000 /* More data to be appended to sock */
223 #define TF_LQ_OVERFLOW 0x020000 /* listen queue overflow */
224 #define TF_LASTIDLE 0x040000 /* connection was previously idle */
225 #define TF_RXWIN0SENT 0x080000 /* sent a receiver win 0 in response */
226 #define TF_FASTRECOVERY 0x100000 /* in NewReno Fast Recovery */
227 #define TF_WASFRECOVERY 0x200000 /* was in NewReno Fast Recovery */
228 #define TF_SIGNATURE 0x400000 /* require MD5 digests (RFC2385) */
229 #define TF_FORCEDATA 0x800000 /* force out a byte */
230 #define TF_TSO 0x1000000 /* TSO enabled on this connection */
231 #define TF_TOE 0x2000000 /* this connection is offloaded */
232 #define TF_ECN_PERMIT 0x4000000 /* connection ECN-ready */
233 #define TF_ECN_SND_CWR 0x8000000 /* ECN CWR in queue */
234 #define TF_ECN_SND_ECE 0x10000000 /* ECN ECE in queue */
236 #define IN_FASTRECOVERY(tp) (tp->t_flags & TF_FASTRECOVERY)
237 #define ENTER_FASTRECOVERY(tp) tp->t_flags |= TF_FASTRECOVERY
238 #define EXIT_FASTRECOVERY(tp) tp->t_flags &= ~TF_FASTRECOVERY
241 * Flags for the t_oobflags field.
243 #define TCPOOB_HAVEDATA 0x01
244 #define TCPOOB_HADDATA 0x02
248 * Defines which are needed by the xform_tcp module and tcp_[in|out]put
249 * for SADB verification and lookup.
251 #define TCP_SIGLEN 16 /* length of computed digest in bytes */
252 #define TCP_KEYLEN_MIN 1 /* minimum length of TCP-MD5 key */
253 #define TCP_KEYLEN_MAX 80 /* maximum length of TCP-MD5 key */
255 * Only a single SA per host may be specified at this time. An SPI is
256 * needed in order for the KEY_ALLOCSA() lookup to work.
258 #define TCP_SIG_SPI 0x1000
259 #endif /* TCP_SIGNATURE */
262 * Structure to hold TCP options that are only used during segment
263 * processing (in tcp_input), but not held in the tcpcb.
264 * It's basically used to reduce the number of parameters
265 * to tcp_dooptions and tcp_addoptions.
266 * The binary order of the to_flags is relevant for packing of the
267 * options in tcp_addoptions.
270 u_int64_t to_flags; /* which options are present */
271 #define TOF_MSS 0x0001 /* maximum segment size */
272 #define TOF_SCALE 0x0002 /* window scaling */
273 #define TOF_SACKPERM 0x0004 /* SACK permitted */
274 #define TOF_TS 0x0010 /* timestamp */
275 #define TOF_SIGNATURE 0x0040 /* TCP-MD5 signature option (RFC2385) */
276 #define TOF_SACK 0x0080 /* Peer sent SACK option */
277 #define TOF_MAXOPT 0x0100
278 u_int32_t to_tsval; /* new timestamp */
279 u_int32_t to_tsecr; /* reflected timestamp */
280 u_char *to_sacks; /* pointer to the first SACK blocks */
281 u_char *to_signature; /* pointer to the TCP-MD5 signature */
282 u_int16_t to_mss; /* maximum segment size */
283 u_int8_t to_wscale; /* window scaling */
284 u_int8_t to_nsacks; /* number of SACK blocks */
288 * Flags for tcp_dooptions.
290 #define TO_SYN 0x01 /* parse SYN-only options */
292 struct hc_metrics_lite { /* must stay in sync with hc_metrics */
293 u_long rmx_mtu; /* MTU for this path */
294 u_long rmx_ssthresh; /* outbound gateway buffer limit */
295 u_long rmx_rtt; /* estimated round trip time */
296 u_long rmx_rttvar; /* estimated rtt variance */
297 u_long rmx_bandwidth; /* estimated bandwidth */
298 u_long rmx_cwnd; /* congestion window */
299 u_long rmx_sendpipe; /* outbound delay-bandwidth product */
300 u_long rmx_recvpipe; /* inbound delay-bandwidth product */
303 #ifndef _NETINET_IN_PCB_H_
305 #endif /* _NETINET_IN_PCB_H_ */
308 struct inpcb *tw_inpcb; /* XXX back pointer to internet pcb */
313 u_short last_win; /* cached window value */
314 u_short tw_so_options; /* copy of so_options */
315 struct ucred *tw_cred; /* user credentials */
317 u_int32_t ts_offset; /* our timestamp offset */
320 TAILQ_ENTRY(tcptw) tw_2msl;
323 #define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb)
324 #define intotw(ip) ((struct tcptw *)(ip)->inp_ppcb)
325 #define sototcpcb(so) (intotcpcb(sotoinpcb(so)))
328 * The smoothed round-trip time and estimated variance
329 * are stored as fixed point numbers scaled by the values below.
330 * For convenience, these scales are also used in smoothing the average
331 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed).
332 * With these scales, srtt has 3 bits to the right of the binary point,
333 * and thus an "ALPHA" of 0.875. rttvar has 2 bits to the right of the
334 * binary point, and is smoothed with an ALPHA of 0.75.
336 #define TCP_RTT_SCALE 32 /* multiplier for srtt; 3 bits frac. */
337 #define TCP_RTT_SHIFT 5 /* shift for srtt; 3 bits frac. */
338 #define TCP_RTTVAR_SCALE 16 /* multiplier for rttvar; 2 bits */
339 #define TCP_RTTVAR_SHIFT 4 /* shift for rttvar; 2 bits */
340 #define TCP_DELTA_SHIFT 2 /* see tcp_input.c */
343 * The initial retransmission should happen at rtt + 4 * rttvar.
344 * Because of the way we do the smoothing, srtt and rttvar
345 * will each average +1/2 tick of bias. When we compute
346 * the retransmit timer, we want 1/2 tick of rounding and
347 * 1 extra tick because of +-1/2 tick uncertainty in the
348 * firing of the timer. The bias will give us exactly the
349 * 1.5 tick we need. But, because the bias is
350 * statistical, we have to test that we don't drop below
351 * the minimum feasible timer (which is 2 ticks).
352 * This version of the macro adapted from a paper by Lawrence
353 * Brakmo and Larry Peterson which outlines a problem caused
354 * by insufficient precision in the original implementation,
355 * which results in inappropriately large RTO values for very
358 #define TCP_REXMTVAL(tp) \
359 max((tp)->t_rttmin, (((tp)->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT)) \
360 + (tp)->t_rttvar) >> TCP_DELTA_SHIFT)
364 * Many of these should be kept per connection,
365 * but that's inconvenient at the moment.
368 u_long tcps_connattempt; /* connections initiated */
369 u_long tcps_accepts; /* connections accepted */
370 u_long tcps_connects; /* connections established */
371 u_long tcps_drops; /* connections dropped */
372 u_long tcps_conndrops; /* embryonic connections dropped */
373 u_long tcps_minmssdrops; /* average minmss too low drops */
374 u_long tcps_closed; /* conn. closed (includes drops) */
375 u_long tcps_segstimed; /* segs where we tried to get rtt */
376 u_long tcps_rttupdated; /* times we succeeded */
377 u_long tcps_delack; /* delayed acks sent */
378 u_long tcps_timeoutdrop; /* conn. dropped in rxmt timeout */
379 u_long tcps_rexmttimeo; /* retransmit timeouts */
380 u_long tcps_persisttimeo; /* persist timeouts */
381 u_long tcps_keeptimeo; /* keepalive timeouts */
382 u_long tcps_keepprobe; /* keepalive probes sent */
383 u_long tcps_keepdrops; /* connections dropped in keepalive */
385 u_long tcps_sndtotal; /* total packets sent */
386 u_long tcps_sndpack; /* data packets sent */
387 u_long tcps_sndbyte; /* data bytes sent */
388 u_long tcps_sndrexmitpack; /* data packets retransmitted */
389 u_long tcps_sndrexmitbyte; /* data bytes retransmitted */
390 u_long tcps_sndrexmitbad; /* unnecessary packet retransmissions */
391 u_long tcps_sndacks; /* ack-only packets sent */
392 u_long tcps_sndprobe; /* window probes sent */
393 u_long tcps_sndurg; /* packets sent with URG only */
394 u_long tcps_sndwinup; /* window update-only packets sent */
395 u_long tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */
397 u_long tcps_rcvtotal; /* total packets received */
398 u_long tcps_rcvpack; /* packets received in sequence */
399 u_long tcps_rcvbyte; /* bytes received in sequence */
400 u_long tcps_rcvbadsum; /* packets received with ccksum errs */
401 u_long tcps_rcvbadoff; /* packets received with bad offset */
402 u_long tcps_rcvmemdrop; /* packets dropped for lack of memory */
403 u_long tcps_rcvshort; /* packets received too short */
404 u_long tcps_rcvduppack; /* duplicate-only packets received */
405 u_long tcps_rcvdupbyte; /* duplicate-only bytes received */
406 u_long tcps_rcvpartduppack; /* packets with some duplicate data */
407 u_long tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */
408 u_long tcps_rcvoopack; /* out-of-order packets received */
409 u_long tcps_rcvoobyte; /* out-of-order bytes received */
410 u_long tcps_rcvpackafterwin; /* packets with data after window */
411 u_long tcps_rcvbyteafterwin; /* bytes rcvd after window */
412 u_long tcps_rcvafterclose; /* packets rcvd after "close" */
413 u_long tcps_rcvwinprobe; /* rcvd window probe packets */
414 u_long tcps_rcvdupack; /* rcvd duplicate acks */
415 u_long tcps_rcvacktoomuch; /* rcvd acks for unsent data */
416 u_long tcps_rcvackpack; /* rcvd ack packets */
417 u_long tcps_rcvackbyte; /* bytes acked by rcvd acks */
418 u_long tcps_rcvwinupd; /* rcvd window update packets */
419 u_long tcps_pawsdrop; /* segments dropped due to PAWS */
420 u_long tcps_predack; /* times hdr predict ok for acks */
421 u_long tcps_preddat; /* times hdr predict ok for data pkts */
422 u_long tcps_pcbcachemiss;
423 u_long tcps_cachedrtt; /* times cached RTT in route updated */
424 u_long tcps_cachedrttvar; /* times cached rttvar updated */
425 u_long tcps_cachedssthresh; /* times cached ssthresh updated */
426 u_long tcps_usedrtt; /* times RTT initialized from route */
427 u_long tcps_usedrttvar; /* times RTTVAR initialized from rt */
428 u_long tcps_usedssthresh; /* times ssthresh initialized from rt*/
429 u_long tcps_persistdrop; /* timeout in persist state */
430 u_long tcps_badsyn; /* bogus SYN, e.g. premature ACK */
431 u_long tcps_mturesent; /* resends due to MTU discovery */
432 u_long tcps_listendrop; /* listen queue overflows */
433 u_long tcps_badrst; /* ignored RSTs in the window */
435 u_long tcps_sc_added; /* entry added to syncache */
436 u_long tcps_sc_retransmitted; /* syncache entry was retransmitted */
437 u_long tcps_sc_dupsyn; /* duplicate SYN packet */
438 u_long tcps_sc_dropped; /* could not reply to packet */
439 u_long tcps_sc_completed; /* successful extraction of entry */
440 u_long tcps_sc_bucketoverflow; /* syncache per-bucket limit hit */
441 u_long tcps_sc_cacheoverflow; /* syncache cache limit hit */
442 u_long tcps_sc_reset; /* RST removed entry from syncache */
443 u_long tcps_sc_stale; /* timed out or listen socket gone */
444 u_long tcps_sc_aborted; /* syncache entry aborted */
445 u_long tcps_sc_badack; /* removed due to bad ACK */
446 u_long tcps_sc_unreach; /* ICMP unreachable received */
447 u_long tcps_sc_zonefail; /* zalloc() failed */
448 u_long tcps_sc_sendcookie; /* SYN cookie sent */
449 u_long tcps_sc_recvcookie; /* SYN cookie received */
451 u_long tcps_hc_added; /* entry added to hostcache */
452 u_long tcps_hc_bucketoverflow; /* hostcache per bucket limit hit */
454 u_long tcps_finwait2_drops; /* Drop FIN_WAIT_2 connection after time limit */
456 /* SACK related stats */
457 u_long tcps_sack_recovery_episode; /* SACK recovery episodes */
458 u_long tcps_sack_rexmits; /* SACK rexmit segments */
459 u_long tcps_sack_rexmit_bytes; /* SACK rexmit bytes */
460 u_long tcps_sack_rcv_blocks; /* SACK blocks (options) received */
461 u_long tcps_sack_send_blocks; /* SACK blocks (options) sent */
462 u_long tcps_sack_sboverflow; /* times scoreboard overflowed */
464 /* ECN related stats */
465 u_long tcps_ecn_ce; /* ECN Congestion Experienced */
466 u_long tcps_ecn_ect0; /* ECN Capable Transport */
467 u_long tcps_ecn_ect1; /* ECN Capable Transport */
468 u_long tcps_ecn_shs; /* ECN successful handshakes */
469 u_long tcps_ecn_rcwnd; /* # times ECN reduced the cwnd */
471 u_long _pad[12]; /* 6 UTO, 6 TBD */
475 #define TCPSTAT_ADD(name, val) V_tcpstat.name += (val)
476 #define TCPSTAT_INC(name) TCPSTAT_ADD(name, 1)
480 * TCB structure exported to user-land via sysctl(3).
481 * Evil hack: declare only if in_pcb.h and sys/socketvar.h have been
482 * included. Not all of our clients do.
484 #if defined(_NETINET_IN_PCB_H_) && defined(_SYS_SOCKETVAR_H_)
489 struct xsocket xt_socket;
490 u_quad_t xt_alignment_hack;
495 * Names for TCP sysctl objects
497 #define TCPCTL_DO_RFC1323 1 /* use RFC-1323 extensions */
498 #define TCPCTL_MSSDFLT 3 /* MSS default */
499 #define TCPCTL_STATS 4 /* statistics (read-only) */
500 #define TCPCTL_RTTDFLT 5 /* default RTT estimate */
501 #define TCPCTL_KEEPIDLE 6 /* keepalive idle timer */
502 #define TCPCTL_KEEPINTVL 7 /* interval to send keepalives */
503 #define TCPCTL_SENDSPACE 8 /* send buffer space */
504 #define TCPCTL_RECVSPACE 9 /* receive buffer space */
505 #define TCPCTL_KEEPINIT 10 /* timeout for establishing syn */
506 #define TCPCTL_PCBLIST 11 /* list of all outstanding PCBs */
507 #define TCPCTL_DELACKTIME 12 /* time before sending delayed ACK */
508 #define TCPCTL_V6MSSDFLT 13 /* MSS default for IPv6 */
509 #define TCPCTL_SACK 14 /* Selective Acknowledgement,rfc 2018 */
510 #define TCPCTL_DROP 15 /* drop tcp connection */
511 #define TCPCTL_MAXID 16
512 #define TCPCTL_FINWAIT2_TIMEOUT 17
514 #define TCPCTL_NAMES { \
516 { "rfc1323", CTLTYPE_INT }, \
517 { "mssdflt", CTLTYPE_INT }, \
518 { "stats", CTLTYPE_STRUCT }, \
519 { "rttdflt", CTLTYPE_INT }, \
520 { "keepidle", CTLTYPE_INT }, \
521 { "keepintvl", CTLTYPE_INT }, \
522 { "sendspace", CTLTYPE_INT }, \
523 { "recvspace", CTLTYPE_INT }, \
524 { "keepinit", CTLTYPE_INT }, \
525 { "pcblist", CTLTYPE_STRUCT }, \
526 { "delacktime", CTLTYPE_INT }, \
527 { "v6mssdflt", CTLTYPE_INT }, \
528 { "maxid", CTLTYPE_INT }, \
534 SYSCTL_DECL(_net_inet_tcp);
535 SYSCTL_DECL(_net_inet_tcp_sack);
536 MALLOC_DECLARE(M_TCPLOG);
539 extern int tcp_log_in_vain;
541 #ifdef VIMAGE_GLOBALS
542 extern struct inpcbhead tcb; /* head of queue of active tcpcb's */
543 extern struct inpcbinfo tcbinfo;
544 extern struct tcpstat tcpstat; /* tcp statistics */
545 extern int tcp_mssdflt; /* XXX */
546 extern int tcp_minmss;
547 extern int tcp_delack_enabled;
548 extern int tcp_do_newreno;
549 extern int path_mtu_discovery;
551 extern int ss_fltsz_local;
553 extern int blackhole;
554 extern int drop_synfin;
555 extern int tcp_do_rfc3042;
556 extern int tcp_do_rfc3390;
557 extern int tcp_insecure_rst;
558 extern int tcp_do_autorcvbuf;
559 extern int tcp_autorcvbuf_inc;
560 extern int tcp_autorcvbuf_max;
561 extern int tcp_do_rfc3465;
562 extern int tcp_abc_l_var;
564 extern int tcp_do_tso;
565 extern int tcp_do_autosndbuf;
566 extern int tcp_autosndbuf_inc;
567 extern int tcp_autosndbuf_max;
569 extern int nolocaltimewait;
571 extern int tcp_do_sack; /* SACK enabled/disabled */
572 extern int tcp_sack_maxholes;
573 extern int tcp_sack_globalmaxholes;
574 extern int tcp_sack_globalholes;
575 extern int tcp_sc_rst_sock_fail; /* RST on sock alloc failure */
576 extern int tcp_do_ecn; /* TCP ECN enabled/disabled */
577 extern int tcp_ecn_maxretries;
578 #endif /* VIMAGE_GLOBALS */
580 int tcp_addoptions(struct tcpopt *, u_char *);
582 tcp_close(struct tcpcb *);
583 void tcp_discardcb(struct tcpcb *);
584 void tcp_twstart(struct tcpcb *);
586 int tcp_twrecycleable(struct tcptw *tw);
588 void tcp_twclose(struct tcptw *_tw, int _reuse);
589 void tcp_ctlinput(int, struct sockaddr *, void *);
590 int tcp_ctloutput(struct socket *, struct sockopt *);
592 tcp_drop(struct tcpcb *, int);
593 void tcp_drain(void);
594 void tcp_fasttimo(void);
597 void tcp_destroy(void);
599 void tcp_fini(void *);
600 char *tcp_log_addrs(struct in_conninfo *, struct tcphdr *, void *,
602 int tcp_reass(struct tcpcb *, struct tcphdr *, int *, struct mbuf *);
603 void tcp_reass_init(void);
604 void tcp_input(struct mbuf *, int);
605 u_long tcp_maxmtu(struct in_conninfo *, int *);
606 u_long tcp_maxmtu6(struct in_conninfo *, int *);
607 void tcp_mss_update(struct tcpcb *, int, struct hc_metrics_lite *, int *);
608 void tcp_mss(struct tcpcb *, int);
609 int tcp_mssopt(struct in_conninfo *);
611 tcp_drop_syn_sent(struct inpcb *, int);
613 tcp_mtudisc(struct inpcb *, int);
615 tcp_newtcpcb(struct inpcb *);
616 int tcp_output(struct tcpcb *);
617 void tcp_respond(struct tcpcb *, void *,
618 struct tcphdr *, struct mbuf *, tcp_seq, tcp_seq, int);
619 void tcp_tw_init(void);
621 void tcp_tw_destroy(void);
623 void tcp_tw_zone_change(void);
624 int tcp_twcheck(struct inpcb *, struct tcpopt *, struct tcphdr *,
626 int tcp_twrespond(struct tcptw *, int);
627 void tcp_setpersist(struct tcpcb *);
629 int tcp_signature_compute(struct mbuf *, int, int, int, u_char *, u_int);
631 void tcp_slowtimo(void);
633 tcpip_maketemplate(struct inpcb *);
634 void tcpip_fillheaders(struct inpcb *, void *, void *);
635 void tcp_timer_activate(struct tcpcb *, int, u_int);
636 int tcp_timer_active(struct tcpcb *, int);
637 void tcp_trace(short, short, struct tcpcb *, void *, struct tcphdr *, int);
638 void tcp_xmit_bandwidth_limit(struct tcpcb *tp, tcp_seq ack_seq);
640 * All tcp_hc_* functions are IPv4 and IPv6 (via in_conninfo)
642 void tcp_hc_init(void);
644 void tcp_hc_destroy(void);
646 void tcp_hc_get(struct in_conninfo *, struct hc_metrics_lite *);
647 u_long tcp_hc_getmtu(struct in_conninfo *);
648 void tcp_hc_updatemtu(struct in_conninfo *, u_long);
649 void tcp_hc_update(struct in_conninfo *, struct hc_metrics_lite *);
651 extern struct pr_usrreqs tcp_usrreqs;
652 extern u_long tcp_sendspace;
653 extern u_long tcp_recvspace;
654 tcp_seq tcp_new_isn(struct tcpcb *);
656 void tcp_sack_doack(struct tcpcb *, struct tcpopt *, tcp_seq);
657 void tcp_update_sack_list(struct tcpcb *tp, tcp_seq rcv_laststart, tcp_seq rcv_lastend);
658 void tcp_clean_sackreport(struct tcpcb *tp);
659 void tcp_sack_adjust(struct tcpcb *tp);
660 struct sackhole *tcp_sack_output(struct tcpcb *tp, int *sack_bytes_rexmt);
661 void tcp_sack_partialack(struct tcpcb *, struct tcphdr *);
662 void tcp_free_sackholes(struct tcpcb *tp);
663 int tcp_newreno(struct tcpcb *, struct tcphdr *);
664 u_long tcp_seq_subtract(u_long, u_long );
668 #endif /* _NETINET_TCP_VAR_H_ */