tcp_var.h revision 159725
1/*-
2 * Copyright (c) 1982, 1986, 1993, 1994, 1995
3 *	The Regents of the University of California.  All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
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.
16 *
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
27 * SUCH DAMAGE.
28 *
29 *	@(#)tcp_var.h	8.4 (Berkeley) 5/24/95
30 * $FreeBSD: head/sys/netinet/tcp_var.h 159725 2006-06-18 12:26:11Z andre $
31 */
32
33#ifndef _NETINET_TCP_VAR_H_
34#define _NETINET_TCP_VAR_H_
35
36#include <netinet/tcp.h>
37
38/*
39 * Kernel variables for tcp.
40 */
41extern int	tcp_do_rfc1323;
42
43/* TCP segment queue entry */
44struct tseg_qent {
45	LIST_ENTRY(tseg_qent) tqe_q;
46	int	tqe_len;		/* TCP segment data length */
47	struct	tcphdr *tqe_th;		/* a pointer to tcp header */
48	struct	mbuf	*tqe_m;		/* mbuf contains packet */
49};
50LIST_HEAD(tsegqe_head, tseg_qent);
51extern int	tcp_reass_qsize;
52extern struct uma_zone *tcp_reass_zone;
53
54struct sackblk {
55	tcp_seq start;		/* start seq no. of sack block */
56	tcp_seq end;		/* end seq no. */
57};
58
59struct sackhole {
60	tcp_seq start;		/* start seq no. of hole */
61	tcp_seq end;		/* end seq no. */
62	tcp_seq rxmit;		/* next seq. no in hole to be retransmitted */
63	TAILQ_ENTRY(sackhole) scblink;	/* scoreboard linkage */
64};
65
66struct sackhint {
67	struct sackhole	*nexthole;
68	int		sack_bytes_rexmit;
69};
70
71struct tcptemp {
72	u_char	tt_ipgen[40]; /* the size must be of max ip header, now IPv6 */
73	struct	tcphdr tt_t;
74};
75
76#define tcp6cb		tcpcb  /* for KAME src sync over BSD*'s */
77
78/*
79 * Tcp control block, one per tcp; fields:
80 * Organized for 16 byte cacheline efficiency.
81 */
82struct tcpcb {
83	struct	tsegqe_head t_segq;	/* segment reassembly queue */
84	int	t_segqlen;		/* segment reassembly queue length */
85	int	t_dupacks;		/* consecutive dup acks recd */
86
87	struct	callout *tt_rexmt;	/* retransmit timer */
88	struct	callout *tt_persist;	/* retransmit persistence */
89	struct	callout *tt_keep;	/* keepalive */
90	struct	callout *tt_2msl;	/* 2*msl TIME_WAIT timer */
91	struct	callout *tt_delack;	/* delayed ACK timer */
92
93	struct	inpcb *t_inpcb;		/* back pointer to internet pcb */
94	int	t_state;		/* state of this connection */
95	u_int	t_flags;
96#define	TF_ACKNOW	0x000001	/* ack peer immediately */
97#define	TF_DELACK	0x000002	/* ack, but try to delay it */
98#define	TF_NODELAY	0x000004	/* don't delay packets to coalesce */
99#define	TF_NOOPT	0x000008	/* don't use tcp options */
100#define	TF_SENTFIN	0x000010	/* have sent FIN */
101#define	TF_REQ_SCALE	0x000020	/* have/will request window scaling */
102#define	TF_RCVD_SCALE	0x000040	/* other side has requested scaling */
103#define	TF_REQ_TSTMP	0x000080	/* have/will request timestamps */
104#define	TF_RCVD_TSTMP	0x000100	/* a timestamp was received in SYN */
105#define	TF_SACK_PERMIT	0x000200	/* other side said I could SACK */
106#define	TF_NEEDSYN	0x000400	/* send SYN (implicit state) */
107#define	TF_NEEDFIN	0x000800	/* send FIN (implicit state) */
108#define	TF_NOPUSH	0x001000	/* don't push */
109#define	TF_MORETOCOME	0x010000	/* More data to be appended to sock */
110#define	TF_LQ_OVERFLOW	0x020000	/* listen queue overflow */
111#define	TF_LASTIDLE	0x040000	/* connection was previously idle */
112#define	TF_RXWIN0SENT	0x080000	/* sent a receiver win 0 in response */
113#define	TF_FASTRECOVERY	0x100000	/* in NewReno Fast Recovery */
114#define	TF_WASFRECOVERY	0x200000	/* was in NewReno Fast Recovery */
115#define	TF_SIGNATURE	0x400000	/* require MD5 digests (RFC2385) */
116#define	TF_FORCEDATA	0x800000	/* force out a byte */
117
118	tcp_seq	snd_una;		/* send unacknowledged */
119	tcp_seq	snd_max;		/* highest sequence number sent;
120					 * used to recognize retransmits
121					 */
122	tcp_seq	snd_nxt;		/* send next */
123	tcp_seq	snd_up;			/* send urgent pointer */
124
125	tcp_seq	snd_wl1;		/* window update seg seq number */
126	tcp_seq	snd_wl2;		/* window update seg ack number */
127	tcp_seq	iss;			/* initial send sequence number */
128	tcp_seq	irs;			/* initial receive sequence number */
129
130	tcp_seq	rcv_nxt;		/* receive next */
131	tcp_seq	rcv_adv;		/* advertised window */
132	u_long	rcv_wnd;		/* receive window */
133	tcp_seq	rcv_up;			/* receive urgent pointer */
134
135	u_long	snd_wnd;		/* send window */
136	u_long	snd_cwnd;		/* congestion-controlled window */
137	u_long	snd_bwnd;		/* bandwidth-controlled window */
138	u_long	snd_ssthresh;		/* snd_cwnd size threshold for
139					 * for slow start exponential to
140					 * linear switch
141					 */
142	u_long	snd_bandwidth;		/* calculated bandwidth or 0 */
143	tcp_seq	snd_recover;		/* for use in NewReno Fast Recovery */
144
145	u_int	t_maxopd;		/* mss plus options */
146
147	u_long	t_rcvtime;		/* inactivity time */
148	u_long	t_starttime;		/* time connection was established */
149	int	t_rtttime;		/* round trip time */
150	tcp_seq	t_rtseq;		/* sequence number being timed */
151
152	int	t_bw_rtttime;		/* used for bandwidth calculation */
153	tcp_seq	t_bw_rtseq;		/* used for bandwidth calculation */
154
155	int	t_rxtcur;		/* current retransmit value (ticks) */
156	u_int	t_maxseg;		/* maximum segment size */
157	int	t_srtt;			/* smoothed round-trip time */
158	int	t_rttvar;		/* variance in round-trip time */
159
160	int	t_rxtshift;		/* log(2) of rexmt exp. backoff */
161	u_int	t_rttmin;		/* minimum rtt allowed */
162	u_int	t_rttbest;		/* best rtt we've seen */
163	u_long	t_rttupdated;		/* number of times rtt sampled */
164	u_long	max_sndwnd;		/* largest window peer has offered */
165
166	int	t_softerror;		/* possible error not yet reported */
167/* out-of-band data */
168	char	t_oobflags;		/* have some */
169	char	t_iobc;			/* input character */
170#define	TCPOOB_HAVEDATA	0x01
171#define	TCPOOB_HADDATA	0x02
172/* RFC 1323 variables */
173	u_char	snd_scale;		/* window scaling for send window */
174	u_char	rcv_scale;		/* window scaling for recv window */
175	u_char	request_r_scale;	/* pending window scaling */
176	u_char	requested_s_scale;	/* unused, to be reused later */
177	u_long	ts_recent;		/* timestamp echo data */
178
179	u_long	ts_recent_age;		/* when last updated */
180	tcp_seq	last_ack_sent;
181/* experimental */
182	u_long	snd_cwnd_prev;		/* cwnd prior to retransmit */
183	u_long	snd_ssthresh_prev;	/* ssthresh prior to retransmit */
184	tcp_seq	snd_recover_prev;	/* snd_recover prior to retransmit */
185	u_long	t_badrxtwin;		/* window for retransmit recovery */
186	u_char	snd_limited;		/* segments limited transmitted */
187/* anti DoS counters */
188	u_long	rcv_second;		/* start of interval second */
189	u_long	rcv_pps;		/* received packets per second */
190	u_long	rcv_byps;		/* received bytes per second */
191/* SACK related state */
192	int	sack_enable;		/* enable SACK for this connection */
193	int	snd_numholes;		/* number of holes seen by sender */
194	TAILQ_HEAD(sackhole_head, sackhole) snd_holes;
195					/* SACK scoreboard (sorted) */
196	tcp_seq	snd_fack;		/* last seq number(+1) sack'd by rcv'r*/
197	int	rcv_numsacks;		/* # distinct sack blks present */
198	struct sackblk sackblks[MAX_SACK_BLKS]; /* seq nos. of sack blocks */
199	tcp_seq sack_newdata;		/* New data xmitted in this recovery
200					   episode starts at this seq number */
201	struct sackhint	sackhint;	/* SACK scoreboard hint */
202	int	t_rttlow;		/* smallest observerved RTT */
203};
204
205#define IN_FASTRECOVERY(tp)	(tp->t_flags & TF_FASTRECOVERY)
206#define ENTER_FASTRECOVERY(tp)	tp->t_flags |= TF_FASTRECOVERY
207#define EXIT_FASTRECOVERY(tp)	tp->t_flags &= ~TF_FASTRECOVERY
208
209#ifdef TCP_SIGNATURE
210/*
211 * Defines which are needed by the xform_tcp module and tcp_[in|out]put
212 * for SADB verification and lookup.
213 */
214#define	TCP_SIGLEN	16	/* length of computed digest in bytes */
215#define	TCP_KEYLEN_MIN	1	/* minimum length of TCP-MD5 key */
216#define	TCP_KEYLEN_MAX	80	/* maximum length of TCP-MD5 key */
217/*
218 * Only a single SA per host may be specified at this time. An SPI is
219 * needed in order for the KEY_ALLOCSA() lookup to work.
220 */
221#define	TCP_SIG_SPI	0x1000
222#endif /* TCP_SIGNATURE */
223
224/*
225 * Structure to hold TCP options that are only used during segment
226 * processing (in tcp_input), but not held in the tcpcb.
227 * It's basically used to reduce the number of parameters
228 * to tcp_dooptions.
229 */
230struct tcpopt {
231	u_long		to_flags;	/* which options are present */
232#define TOF_TS		0x0001		/* timestamp */
233#define	TOF_MSS		0x0010
234#define	TOF_SCALE	0x0020
235#define	TOF_SIGNATURE	0x0040		/* signature option present */
236#define	TOF_SIGLEN	0x0080		/* signature length valid (RFC2385) */
237#define	TOF_SACK	0x0100		/* Peer sent SACK option */
238	u_int32_t	to_tsval;
239	u_int32_t	to_tsecr;
240	u_int16_t	to_mss;
241	u_int8_t	to_requested_s_scale;
242	u_int8_t	to_nsacks;	/* number of SACK blocks */
243	u_char		*to_sacks;	/* pointer to the first SACK blocks */
244};
245
246struct hc_metrics_lite {	/* must stay in sync with hc_metrics */
247	u_long	rmx_mtu;	/* MTU for this path */
248	u_long	rmx_ssthresh;	/* outbound gateway buffer limit */
249	u_long	rmx_rtt;	/* estimated round trip time */
250	u_long	rmx_rttvar;	/* estimated rtt variance */
251	u_long	rmx_bandwidth;	/* estimated bandwidth */
252	u_long	rmx_cwnd;	/* congestion window */
253	u_long	rmx_sendpipe;   /* outbound delay-bandwidth product */
254	u_long	rmx_recvpipe;   /* inbound delay-bandwidth product */
255};
256
257#ifndef _NETINET_IN_PCB_H_
258struct in_conninfo;
259#endif /* _NETINET_IN_PCB_H_ */
260
261struct tcptw {
262	struct inpcb	*tw_inpcb;	/* XXX back pointer to internet pcb */
263	tcp_seq		snd_nxt;
264	tcp_seq		rcv_nxt;
265	tcp_seq		iss;
266	tcp_seq		irs;
267	u_short		last_win;	/* cached window value */
268	u_short		tw_so_options;	/* copy of so_options */
269	struct ucred	*tw_cred;	/* user credentials */
270	u_long		t_recent;
271	u_long		t_starttime;
272	int		tw_time;
273	LIST_ENTRY(tcptw) tw_2msl;
274};
275
276#define	intotcpcb(ip)	((struct tcpcb *)(ip)->inp_ppcb)
277#define	intotw(ip)	((struct tcptw *)(ip)->inp_ppcb)
278#define	sototcpcb(so)	(intotcpcb(sotoinpcb(so)))
279
280/*
281 * The smoothed round-trip time and estimated variance
282 * are stored as fixed point numbers scaled by the values below.
283 * For convenience, these scales are also used in smoothing the average
284 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed).
285 * With these scales, srtt has 3 bits to the right of the binary point,
286 * and thus an "ALPHA" of 0.875.  rttvar has 2 bits to the right of the
287 * binary point, and is smoothed with an ALPHA of 0.75.
288 */
289#define	TCP_RTT_SCALE		32	/* multiplier for srtt; 3 bits frac. */
290#define	TCP_RTT_SHIFT		5	/* shift for srtt; 3 bits frac. */
291#define	TCP_RTTVAR_SCALE	16	/* multiplier for rttvar; 2 bits */
292#define	TCP_RTTVAR_SHIFT	4	/* shift for rttvar; 2 bits */
293#define	TCP_DELTA_SHIFT		2	/* see tcp_input.c */
294
295/*
296 * The initial retransmission should happen at rtt + 4 * rttvar.
297 * Because of the way we do the smoothing, srtt and rttvar
298 * will each average +1/2 tick of bias.  When we compute
299 * the retransmit timer, we want 1/2 tick of rounding and
300 * 1 extra tick because of +-1/2 tick uncertainty in the
301 * firing of the timer.  The bias will give us exactly the
302 * 1.5 tick we need.  But, because the bias is
303 * statistical, we have to test that we don't drop below
304 * the minimum feasible timer (which is 2 ticks).
305 * This version of the macro adapted from a paper by Lawrence
306 * Brakmo and Larry Peterson which outlines a problem caused
307 * by insufficient precision in the original implementation,
308 * which results in inappropriately large RTO values for very
309 * fast networks.
310 */
311#define	TCP_REXMTVAL(tp) \
312	max((tp)->t_rttmin, (((tp)->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT))  \
313	  + (tp)->t_rttvar) >> TCP_DELTA_SHIFT)
314
315/*
316 * TCP statistics.
317 * Many of these should be kept per connection,
318 * but that's inconvenient at the moment.
319 */
320struct	tcpstat {
321	u_long	tcps_connattempt;	/* connections initiated */
322	u_long	tcps_accepts;		/* connections accepted */
323	u_long	tcps_connects;		/* connections established */
324	u_long	tcps_drops;		/* connections dropped */
325	u_long	tcps_conndrops;		/* embryonic connections dropped */
326	u_long	tcps_minmssdrops;	/* average minmss too low drops */
327	u_long	tcps_closed;		/* conn. closed (includes drops) */
328	u_long	tcps_segstimed;		/* segs where we tried to get rtt */
329	u_long	tcps_rttupdated;	/* times we succeeded */
330	u_long	tcps_delack;		/* delayed acks sent */
331	u_long	tcps_timeoutdrop;	/* conn. dropped in rxmt timeout */
332	u_long	tcps_rexmttimeo;	/* retransmit timeouts */
333	u_long	tcps_persisttimeo;	/* persist timeouts */
334	u_long	tcps_keeptimeo;		/* keepalive timeouts */
335	u_long	tcps_keepprobe;		/* keepalive probes sent */
336	u_long	tcps_keepdrops;		/* connections dropped in keepalive */
337
338	u_long	tcps_sndtotal;		/* total packets sent */
339	u_long	tcps_sndpack;		/* data packets sent */
340	u_long	tcps_sndbyte;		/* data bytes sent */
341	u_long	tcps_sndrexmitpack;	/* data packets retransmitted */
342	u_long	tcps_sndrexmitbyte;	/* data bytes retransmitted */
343	u_long	tcps_sndrexmitbad;	/* unnecessary packet retransmissions */
344	u_long	tcps_sndacks;		/* ack-only packets sent */
345	u_long	tcps_sndprobe;		/* window probes sent */
346	u_long	tcps_sndurg;		/* packets sent with URG only */
347	u_long	tcps_sndwinup;		/* window update-only packets sent */
348	u_long	tcps_sndctrl;		/* control (SYN|FIN|RST) packets sent */
349
350	u_long	tcps_rcvtotal;		/* total packets received */
351	u_long	tcps_rcvpack;		/* packets received in sequence */
352	u_long	tcps_rcvbyte;		/* bytes received in sequence */
353	u_long	tcps_rcvbadsum;		/* packets received with ccksum errs */
354	u_long	tcps_rcvbadoff;		/* packets received with bad offset */
355	u_long	tcps_rcvmemdrop;	/* packets dropped for lack of memory */
356	u_long	tcps_rcvshort;		/* packets received too short */
357	u_long	tcps_rcvduppack;	/* duplicate-only packets received */
358	u_long	tcps_rcvdupbyte;	/* duplicate-only bytes received */
359	u_long	tcps_rcvpartduppack;	/* packets with some duplicate data */
360	u_long	tcps_rcvpartdupbyte;	/* dup. bytes in part-dup. packets */
361	u_long	tcps_rcvoopack;		/* out-of-order packets received */
362	u_long	tcps_rcvoobyte;		/* out-of-order bytes received */
363	u_long	tcps_rcvpackafterwin;	/* packets with data after window */
364	u_long	tcps_rcvbyteafterwin;	/* bytes rcvd after window */
365	u_long	tcps_rcvafterclose;	/* packets rcvd after "close" */
366	u_long	tcps_rcvwinprobe;	/* rcvd window probe packets */
367	u_long	tcps_rcvdupack;		/* rcvd duplicate acks */
368	u_long	tcps_rcvacktoomuch;	/* rcvd acks for unsent data */
369	u_long	tcps_rcvackpack;	/* rcvd ack packets */
370	u_long	tcps_rcvackbyte;	/* bytes acked by rcvd acks */
371	u_long	tcps_rcvwinupd;		/* rcvd window update packets */
372	u_long	tcps_pawsdrop;		/* segments dropped due to PAWS */
373	u_long	tcps_predack;		/* times hdr predict ok for acks */
374	u_long	tcps_preddat;		/* times hdr predict ok for data pkts */
375	u_long	tcps_pcbcachemiss;
376	u_long	tcps_cachedrtt;		/* times cached RTT in route updated */
377	u_long	tcps_cachedrttvar;	/* times cached rttvar updated */
378	u_long	tcps_cachedssthresh;	/* times cached ssthresh updated */
379	u_long	tcps_usedrtt;		/* times RTT initialized from route */
380	u_long	tcps_usedrttvar;	/* times RTTVAR initialized from rt */
381	u_long	tcps_usedssthresh;	/* times ssthresh initialized from rt*/
382	u_long	tcps_persistdrop;	/* timeout in persist state */
383	u_long	tcps_badsyn;		/* bogus SYN, e.g. premature ACK */
384	u_long	tcps_mturesent;		/* resends due to MTU discovery */
385	u_long	tcps_listendrop;	/* listen queue overflows */
386	u_long	tcps_badrst;		/* ignored RSTs in the window */
387
388	u_long	tcps_sc_added;		/* entry added to syncache */
389	u_long	tcps_sc_retransmitted;	/* syncache entry was retransmitted */
390	u_long	tcps_sc_dupsyn;		/* duplicate SYN packet */
391	u_long	tcps_sc_dropped;	/* could not reply to packet */
392	u_long	tcps_sc_completed;	/* successful extraction of entry */
393	u_long	tcps_sc_bucketoverflow;	/* syncache per-bucket limit hit */
394	u_long	tcps_sc_cacheoverflow;	/* syncache cache limit hit */
395	u_long	tcps_sc_reset;		/* RST removed entry from syncache */
396	u_long	tcps_sc_stale;		/* timed out or listen socket gone */
397	u_long	tcps_sc_aborted;	/* syncache entry aborted */
398	u_long	tcps_sc_badack;		/* removed due to bad ACK */
399	u_long	tcps_sc_unreach;	/* ICMP unreachable received */
400	u_long	tcps_sc_zonefail;	/* zalloc() failed */
401	u_long	tcps_sc_sendcookie;	/* SYN cookie sent */
402	u_long	tcps_sc_recvcookie;	/* SYN cookie received */
403
404	u_long	tcps_hc_added;		/* entry added to hostcache */
405	u_long	tcps_hc_bucketoverflow;	/* hostcache per bucket limit hit */
406
407	/* SACK related stats */
408	u_long	tcps_sack_recovery_episode; /* SACK recovery episodes */
409	u_long  tcps_sack_rexmits;	    /* SACK rexmit segments   */
410	u_long  tcps_sack_rexmit_bytes;	    /* SACK rexmit bytes      */
411	u_long  tcps_sack_rcv_blocks;	    /* SACK blocks (options) received */
412	u_long  tcps_sack_send_blocks;	    /* SACK blocks (options) sent     */
413	u_long  tcps_sack_sboverflow; 	    /* times scoreboard overflowed */
414};
415
416/*
417 * TCB structure exported to user-land via sysctl(3).
418 * Evil hack: declare only if in_pcb.h and sys/socketvar.h have been
419 * included.  Not all of our clients do.
420 */
421#if defined(_NETINET_IN_PCB_H_) && defined(_SYS_SOCKETVAR_H_)
422struct	xtcpcb {
423	size_t	xt_len;
424	struct	inpcb	xt_inp;
425	struct	tcpcb	xt_tp;
426	struct	xsocket	xt_socket;
427	u_quad_t	xt_alignment_hack;
428};
429#endif
430
431/*
432 * Names for TCP sysctl objects
433 */
434#define	TCPCTL_DO_RFC1323	1	/* use RFC-1323 extensions */
435#define	TCPCTL_MSSDFLT		3	/* MSS default */
436#define TCPCTL_STATS		4	/* statistics (read-only) */
437#define	TCPCTL_RTTDFLT		5	/* default RTT estimate */
438#define	TCPCTL_KEEPIDLE		6	/* keepalive idle timer */
439#define	TCPCTL_KEEPINTVL	7	/* interval to send keepalives */
440#define	TCPCTL_SENDSPACE	8	/* send buffer space */
441#define	TCPCTL_RECVSPACE	9	/* receive buffer space */
442#define	TCPCTL_KEEPINIT		10	/* timeout for establishing syn */
443#define	TCPCTL_PCBLIST		11	/* list of all outstanding PCBs */
444#define	TCPCTL_DELACKTIME	12	/* time before sending delayed ACK */
445#define	TCPCTL_V6MSSDFLT	13	/* MSS default for IPv6 */
446#define	TCPCTL_SACK		14	/* Selective Acknowledgement,rfc 2018 */
447#define	TCPCTL_DROP		15	/* drop tcp connection */
448#define	TCPCTL_MAXID		16
449
450#define TCPCTL_NAMES { \
451	{ 0, 0 }, \
452	{ "rfc1323", CTLTYPE_INT }, \
453	{ "mssdflt", CTLTYPE_INT }, \
454	{ "stats", CTLTYPE_STRUCT }, \
455	{ "rttdflt", CTLTYPE_INT }, \
456	{ "keepidle", CTLTYPE_INT }, \
457	{ "keepintvl", CTLTYPE_INT }, \
458	{ "sendspace", CTLTYPE_INT }, \
459	{ "recvspace", CTLTYPE_INT }, \
460	{ "keepinit", CTLTYPE_INT }, \
461	{ "pcblist", CTLTYPE_STRUCT }, \
462	{ "delacktime", CTLTYPE_INT }, \
463	{ "v6mssdflt", CTLTYPE_INT }, \
464	{ "maxid", CTLTYPE_INT }, \
465}
466
467
468#ifdef _KERNEL
469#ifdef SYSCTL_DECL
470SYSCTL_DECL(_net_inet_tcp);
471SYSCTL_DECL(_net_inet_tcp_sack);
472#endif
473
474extern	struct inpcbhead tcb;		/* head of queue of active tcpcb's */
475extern	struct inpcbinfo tcbinfo;
476extern	struct tcpstat tcpstat;	/* tcp statistics */
477extern	int tcp_mssdflt;	/* XXX */
478extern	int tcp_minmss;
479extern	int tcp_minmssoverload;
480extern	int tcp_delack_enabled;
481extern	int tcp_do_newreno;
482extern	int path_mtu_discovery;
483extern	int ss_fltsz;
484extern	int ss_fltsz_local;
485
486extern	int tcp_do_sack;	/* SACK enabled/disabled */
487
488struct tcpcb *
489	 tcp_close(struct tcpcb *);
490void	 tcp_discardcb(struct tcpcb *);
491void	 tcp_twstart(struct tcpcb *);
492int	 tcp_twrecycleable(struct tcptw *tw);
493void	 tcp_twclose(struct tcptw *_tw, int _reuse);
494void	 tcp_ctlinput(int, struct sockaddr *, void *);
495int	 tcp_ctloutput(struct socket *, struct sockopt *);
496struct tcpcb *
497	 tcp_drop(struct tcpcb *, int);
498void	 tcp_drain(void);
499void	 tcp_fasttimo(void);
500void	 tcp_init(void);
501void	 tcp_fini(void *);
502void	 tcp_reass_init(void);
503void	 tcp_input(struct mbuf *, int);
504u_long	 tcp_maxmtu(struct in_conninfo *);
505u_long	 tcp_maxmtu6(struct in_conninfo *);
506void	 tcp_mss(struct tcpcb *, int);
507int	 tcp_mssopt(struct in_conninfo *);
508struct inpcb *
509	 tcp_drop_syn_sent(struct inpcb *, int);
510struct inpcb *
511	 tcp_mtudisc(struct inpcb *, int);
512struct tcpcb *
513	 tcp_newtcpcb(struct inpcb *);
514int	 tcp_output(struct tcpcb *);
515void	 tcp_respond(struct tcpcb *, void *,
516	    struct tcphdr *, struct mbuf *, tcp_seq, tcp_seq, int);
517int	 tcp_twrespond(struct tcptw *, int);
518void	 tcp_setpersist(struct tcpcb *);
519#ifdef TCP_SIGNATURE
520int	 tcp_signature_compute(struct mbuf *, int, int, int, u_char *, u_int);
521#endif
522void	 tcp_slowtimo(void);
523struct tcptemp *
524	 tcpip_maketemplate(struct inpcb *);
525void	 tcpip_fillheaders(struct inpcb *, void *, void *);
526struct tcpcb *
527	 tcp_timers(struct tcpcb *, int);
528void	 tcp_trace(int, int, struct tcpcb *, void *, struct tcphdr *, int);
529void	 tcp_xmit_bandwidth_limit(struct tcpcb *tp, tcp_seq ack_seq);
530void	 syncache_init(void);
531void	 syncache_unreach(struct in_conninfo *, struct tcphdr *);
532int	 syncache_expand(struct in_conninfo *, struct tcphdr *,
533	     struct socket **, struct mbuf *);
534int	 syncache_add(struct in_conninfo *, struct tcpopt *,
535	     struct tcphdr *, struct inpcb *, struct socket **, struct mbuf *);
536void	 syncache_chkrst(struct in_conninfo *, struct tcphdr *);
537void	 syncache_badack(struct in_conninfo *);
538/*
539 * All tcp_hc_* functions are IPv4 and IPv6 (via in_conninfo)
540 */
541void	 tcp_hc_init(void);
542void	 tcp_hc_get(struct in_conninfo *, struct hc_metrics_lite *);
543u_long	 tcp_hc_getmtu(struct in_conninfo *);
544void	 tcp_hc_updatemtu(struct in_conninfo *, u_long);
545void	 tcp_hc_update(struct in_conninfo *, struct hc_metrics_lite *);
546
547extern	struct pr_usrreqs tcp_usrreqs;
548extern	u_long tcp_sendspace;
549extern	u_long tcp_recvspace;
550tcp_seq tcp_new_isn(struct tcpcb *);
551
552void	 tcp_sack_doack(struct tcpcb *, struct tcpopt *, tcp_seq);
553void	 tcp_update_sack_list(struct tcpcb *tp, tcp_seq rcv_laststart, tcp_seq rcv_lastend);
554void	 tcp_clean_sackreport(struct tcpcb *tp);
555void	 tcp_sack_adjust(struct tcpcb *tp);
556struct sackhole *tcp_sack_output(struct tcpcb *tp, int *sack_bytes_rexmt);
557void	 tcp_sack_partialack(struct tcpcb *, struct tcphdr *);
558void	 tcp_free_sackholes(struct tcpcb *tp);
559int	 tcp_newreno(struct tcpcb *, struct tcphdr *);
560u_long	 tcp_seq_subtract(u_long, u_long );
561
562#endif /* _KERNEL */
563
564#endif /* _NETINET_TCP_VAR_H_ */
565