tcp_var.h revision 169541
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 169541 2007-05-13 22:16:13Z 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/* Neighbor Discovery, Neighbor Unreachability Detection Upper layer hint. */
79#ifdef INET6
80#define ND6_HINT(tp)						\
81do {								\
82	if ((tp) && (tp)->t_inpcb &&				\
83	    ((tp)->t_inpcb->inp_vflag & INP_IPV6) != 0)		\
84		nd6_nud_hint(NULL, NULL, 0);			\
85} while (0)
86#else
87#define ND6_HINT(tp)
88#endif
89
90/*
91 * Tcp control block, one per tcp; fields:
92 * Organized for 16 byte cacheline efficiency.
93 */
94struct tcpcb {
95	struct	tsegqe_head t_segq;	/* segment reassembly queue */
96	int	t_segqlen;		/* segment reassembly queue length */
97	int	t_dupacks;		/* consecutive dup acks recd */
98
99	struct	tcp_timer *t_timers;	/* retransmit timer */
100
101	struct	inpcb *t_inpcb;		/* back pointer to internet pcb */
102	int	t_state;		/* state of this connection */
103	u_int	t_flags;
104#define	TF_ACKNOW	0x000001	/* ack peer immediately */
105#define	TF_DELACK	0x000002	/* ack, but try to delay it */
106#define	TF_NODELAY	0x000004	/* don't delay packets to coalesce */
107#define	TF_NOOPT	0x000008	/* don't use tcp options */
108#define	TF_SENTFIN	0x000010	/* have sent FIN */
109#define	TF_REQ_SCALE	0x000020	/* have/will request window scaling */
110#define	TF_RCVD_SCALE	0x000040	/* other side has requested scaling */
111#define	TF_REQ_TSTMP	0x000080	/* have/will request timestamps */
112#define	TF_RCVD_TSTMP	0x000100	/* a timestamp was received in SYN */
113#define	TF_SACK_PERMIT	0x000200	/* other side said I could SACK */
114#define	TF_NEEDSYN	0x000400	/* send SYN (implicit state) */
115#define	TF_NEEDFIN	0x000800	/* send FIN (implicit state) */
116#define	TF_NOPUSH	0x001000	/* don't push */
117#define	TF_MORETOCOME	0x010000	/* More data to be appended to sock */
118#define	TF_LQ_OVERFLOW	0x020000	/* listen queue overflow */
119#define	TF_LASTIDLE	0x040000	/* connection was previously idle */
120#define	TF_RXWIN0SENT	0x080000	/* sent a receiver win 0 in response */
121#define	TF_FASTRECOVERY	0x100000	/* in NewReno Fast Recovery */
122#define	TF_WASFRECOVERY	0x200000	/* was in NewReno Fast Recovery */
123#define	TF_SIGNATURE	0x400000	/* require MD5 digests (RFC2385) */
124#define	TF_FORCEDATA	0x800000	/* force out a byte */
125#define	TF_TSO		0x1000000	/* TSO enabled on this connection */
126
127	tcp_seq	snd_una;		/* send unacknowledged */
128	tcp_seq	snd_max;		/* highest sequence number sent;
129					 * used to recognize retransmits
130					 */
131	tcp_seq	snd_nxt;		/* send next */
132	tcp_seq	snd_up;			/* send urgent pointer */
133
134	tcp_seq	snd_wl1;		/* window update seg seq number */
135	tcp_seq	snd_wl2;		/* window update seg ack number */
136	tcp_seq	iss;			/* initial send sequence number */
137	tcp_seq	irs;			/* initial receive sequence number */
138
139	tcp_seq	rcv_nxt;		/* receive next */
140	tcp_seq	rcv_adv;		/* advertised window */
141	u_long	rcv_wnd;		/* receive window */
142	tcp_seq	rcv_up;			/* receive urgent pointer */
143
144	u_long	snd_wnd;		/* send window */
145	u_long	snd_cwnd;		/* congestion-controlled window */
146	u_long	snd_bwnd;		/* bandwidth-controlled window */
147	u_long	snd_ssthresh;		/* snd_cwnd size threshold for
148					 * for slow start exponential to
149					 * linear switch
150					 */
151	u_long	snd_bandwidth;		/* calculated bandwidth or 0 */
152	tcp_seq	snd_recover;		/* for use in NewReno Fast Recovery */
153
154	u_int	t_maxopd;		/* mss plus options */
155
156	u_long	t_rcvtime;		/* inactivity time */
157	u_long	t_starttime;		/* time connection was established */
158	int	t_rtttime;		/* round trip time */
159	tcp_seq	t_rtseq;		/* sequence number being timed */
160
161	int	t_bw_rtttime;		/* used for bandwidth calculation */
162	tcp_seq	t_bw_rtseq;		/* used for bandwidth calculation */
163
164	int	t_rxtcur;		/* current retransmit value (ticks) */
165	u_int	t_maxseg;		/* maximum segment size */
166	int	t_srtt;			/* smoothed round-trip time */
167	int	t_rttvar;		/* variance in round-trip time */
168
169	int	t_rxtshift;		/* log(2) of rexmt exp. backoff */
170	u_int	t_rttmin;		/* minimum rtt allowed */
171	u_int	t_rttbest;		/* best rtt we've seen */
172	u_long	t_rttupdated;		/* number of times rtt sampled */
173	u_long	max_sndwnd;		/* largest window peer has offered */
174
175	int	t_softerror;		/* possible error not yet reported */
176/* out-of-band data */
177	char	t_oobflags;		/* have some */
178	char	t_iobc;			/* input character */
179#define	TCPOOB_HAVEDATA	0x01
180#define	TCPOOB_HADDATA	0x02
181/* RFC 1323 variables */
182	u_char	snd_scale;		/* window scaling for send window */
183	u_char	rcv_scale;		/* window scaling for recv window */
184	u_char	request_r_scale;	/* pending window scaling */
185	u_int32_t  ts_recent;		/* timestamp echo data */
186	u_long	ts_recent_age;		/* when last updated */
187	u_int32_t  ts_offset;		/* our timestamp offset */
188
189	tcp_seq	last_ack_sent;
190/* experimental */
191	u_long	snd_cwnd_prev;		/* cwnd prior to retransmit */
192	u_long	snd_ssthresh_prev;	/* ssthresh prior to retransmit */
193	tcp_seq	snd_recover_prev;	/* snd_recover prior to retransmit */
194	u_long	t_badrxtwin;		/* window for retransmit recovery */
195	u_char	snd_limited;		/* segments limited transmitted */
196/* SACK related state */
197	int	snd_numholes;		/* number of holes seen by sender */
198	TAILQ_HEAD(sackhole_head, sackhole) snd_holes;
199					/* SACK scoreboard (sorted) */
200	tcp_seq	snd_fack;		/* last seq number(+1) sack'd by rcv'r*/
201	int	rcv_numsacks;		/* # distinct sack blks present */
202	struct sackblk sackblks[MAX_SACK_BLKS]; /* seq nos. of sack blocks */
203	tcp_seq sack_newdata;		/* New data xmitted in this recovery
204					   episode starts at this seq number */
205	struct sackhint	sackhint;	/* SACK scoreboard hint */
206	int	t_rttlow;		/* smallest observerved RTT */
207	u_int32_t	rfbuf_ts;	/* recv buffer autoscaling timestamp */
208	int	rfbuf_cnt;		/* recv buffer autoscaling byte count */
209};
210
211#define IN_FASTRECOVERY(tp)	(tp->t_flags & TF_FASTRECOVERY)
212#define ENTER_FASTRECOVERY(tp)	tp->t_flags |= TF_FASTRECOVERY
213#define EXIT_FASTRECOVERY(tp)	tp->t_flags &= ~TF_FASTRECOVERY
214
215#ifdef TCP_SIGNATURE
216/*
217 * Defines which are needed by the xform_tcp module and tcp_[in|out]put
218 * for SADB verification and lookup.
219 */
220#define	TCP_SIGLEN	16	/* length of computed digest in bytes */
221#define	TCP_KEYLEN_MIN	1	/* minimum length of TCP-MD5 key */
222#define	TCP_KEYLEN_MAX	80	/* maximum length of TCP-MD5 key */
223/*
224 * Only a single SA per host may be specified at this time. An SPI is
225 * needed in order for the KEY_ALLOCSA() lookup to work.
226 */
227#define	TCP_SIG_SPI	0x1000
228#endif /* TCP_SIGNATURE */
229
230/*
231 * Structure to hold TCP options that are only used during segment
232 * processing (in tcp_input), but not held in the tcpcb.
233 * It's basically used to reduce the number of parameters
234 * to tcp_dooptions and tcp_addoptions.
235 * The binary order of the to_flags is relevant for packing of the
236 * options in tcp_addoptions.
237 */
238struct tcpopt {
239	u_long		to_flags;	/* which options are present */
240#define	TOF_MSS		0x0001		/* maximum segment size */
241#define	TOF_SCALE	0x0002		/* window scaling */
242#define	TOF_SACKPERM	0x0004		/* SACK permitted */
243#define	TOF_TS		0x0010		/* timestamp */
244#define	TOF_SIGNATURE	0x0040		/* TCP-MD5 signature option (RFC2385) */
245#define	TOF_SACK	0x0080		/* Peer sent SACK option */
246#define	TOF_MAXOPT	0x0100
247	u_int32_t	to_tsval;	/* new timestamp */
248	u_int32_t	to_tsecr;	/* reflected timestamp */
249	u_int16_t	to_mss;		/* maximum segment size */
250	u_int8_t	to_wscale;	/* window scaling */
251	u_int8_t	to_nsacks;	/* number of SACK blocks */
252	u_char		*to_sacks;	/* pointer to the first SACK blocks */
253	u_char		*to_signature;	/* pointer to the TCP-MD5 signature */
254};
255
256/*
257 * Flags for tcp_dooptions.
258 */
259#define	TO_SYN		0x01		/* parse SYN-only options */
260
261struct hc_metrics_lite {	/* must stay in sync with hc_metrics */
262	u_long	rmx_mtu;	/* MTU for this path */
263	u_long	rmx_ssthresh;	/* outbound gateway buffer limit */
264	u_long	rmx_rtt;	/* estimated round trip time */
265	u_long	rmx_rttvar;	/* estimated rtt variance */
266	u_long	rmx_bandwidth;	/* estimated bandwidth */
267	u_long	rmx_cwnd;	/* congestion window */
268	u_long	rmx_sendpipe;   /* outbound delay-bandwidth product */
269	u_long	rmx_recvpipe;   /* inbound delay-bandwidth product */
270};
271
272#ifndef _NETINET_IN_PCB_H_
273struct in_conninfo;
274#endif /* _NETINET_IN_PCB_H_ */
275
276struct tcptw {
277	struct inpcb	*tw_inpcb;	/* XXX back pointer to internet pcb */
278	tcp_seq		snd_nxt;
279	tcp_seq		rcv_nxt;
280	tcp_seq		iss;
281	tcp_seq		irs;
282	u_short		last_win;	/* cached window value */
283	u_short		tw_so_options;	/* copy of so_options */
284	struct ucred	*tw_cred;	/* user credentials */
285	u_long		t_recent;
286	u_int32_t	ts_offset;	/* our timestamp offset */
287	u_long		t_starttime;
288	int		tw_time;
289	TAILQ_ENTRY(tcptw) tw_2msl;
290};
291
292#define	intotcpcb(ip)	((struct tcpcb *)(ip)->inp_ppcb)
293#define	intotw(ip)	((struct tcptw *)(ip)->inp_ppcb)
294#define	sototcpcb(so)	(intotcpcb(sotoinpcb(so)))
295
296/*
297 * The smoothed round-trip time and estimated variance
298 * are stored as fixed point numbers scaled by the values below.
299 * For convenience, these scales are also used in smoothing the average
300 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed).
301 * With these scales, srtt has 3 bits to the right of the binary point,
302 * and thus an "ALPHA" of 0.875.  rttvar has 2 bits to the right of the
303 * binary point, and is smoothed with an ALPHA of 0.75.
304 */
305#define	TCP_RTT_SCALE		32	/* multiplier for srtt; 3 bits frac. */
306#define	TCP_RTT_SHIFT		5	/* shift for srtt; 3 bits frac. */
307#define	TCP_RTTVAR_SCALE	16	/* multiplier for rttvar; 2 bits */
308#define	TCP_RTTVAR_SHIFT	4	/* shift for rttvar; 2 bits */
309#define	TCP_DELTA_SHIFT		2	/* see tcp_input.c */
310
311/*
312 * The initial retransmission should happen at rtt + 4 * rttvar.
313 * Because of the way we do the smoothing, srtt and rttvar
314 * will each average +1/2 tick of bias.  When we compute
315 * the retransmit timer, we want 1/2 tick of rounding and
316 * 1 extra tick because of +-1/2 tick uncertainty in the
317 * firing of the timer.  The bias will give us exactly the
318 * 1.5 tick we need.  But, because the bias is
319 * statistical, we have to test that we don't drop below
320 * the minimum feasible timer (which is 2 ticks).
321 * This version of the macro adapted from a paper by Lawrence
322 * Brakmo and Larry Peterson which outlines a problem caused
323 * by insufficient precision in the original implementation,
324 * which results in inappropriately large RTO values for very
325 * fast networks.
326 */
327#define	TCP_REXMTVAL(tp) \
328	max((tp)->t_rttmin, (((tp)->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT))  \
329	  + (tp)->t_rttvar) >> TCP_DELTA_SHIFT)
330
331/*
332 * TCP statistics.
333 * Many of these should be kept per connection,
334 * but that's inconvenient at the moment.
335 */
336struct	tcpstat {
337	u_long	tcps_connattempt;	/* connections initiated */
338	u_long	tcps_accepts;		/* connections accepted */
339	u_long	tcps_connects;		/* connections established */
340	u_long	tcps_drops;		/* connections dropped */
341	u_long	tcps_conndrops;		/* embryonic connections dropped */
342	u_long	tcps_minmssdrops;	/* average minmss too low drops */
343	u_long	tcps_closed;		/* conn. closed (includes drops) */
344	u_long	tcps_segstimed;		/* segs where we tried to get rtt */
345	u_long	tcps_rttupdated;	/* times we succeeded */
346	u_long	tcps_delack;		/* delayed acks sent */
347	u_long	tcps_timeoutdrop;	/* conn. dropped in rxmt timeout */
348	u_long	tcps_rexmttimeo;	/* retransmit timeouts */
349	u_long	tcps_persisttimeo;	/* persist timeouts */
350	u_long	tcps_keeptimeo;		/* keepalive timeouts */
351	u_long	tcps_keepprobe;		/* keepalive probes sent */
352	u_long	tcps_keepdrops;		/* connections dropped in keepalive */
353
354	u_long	tcps_sndtotal;		/* total packets sent */
355	u_long	tcps_sndpack;		/* data packets sent */
356	u_long	tcps_sndbyte;		/* data bytes sent */
357	u_long	tcps_sndrexmitpack;	/* data packets retransmitted */
358	u_long	tcps_sndrexmitbyte;	/* data bytes retransmitted */
359	u_long	tcps_sndrexmitbad;	/* unnecessary packet retransmissions */
360	u_long	tcps_sndacks;		/* ack-only packets sent */
361	u_long	tcps_sndprobe;		/* window probes sent */
362	u_long	tcps_sndurg;		/* packets sent with URG only */
363	u_long	tcps_sndwinup;		/* window update-only packets sent */
364	u_long	tcps_sndctrl;		/* control (SYN|FIN|RST) packets sent */
365
366	u_long	tcps_rcvtotal;		/* total packets received */
367	u_long	tcps_rcvpack;		/* packets received in sequence */
368	u_long	tcps_rcvbyte;		/* bytes received in sequence */
369	u_long	tcps_rcvbadsum;		/* packets received with ccksum errs */
370	u_long	tcps_rcvbadoff;		/* packets received with bad offset */
371	u_long	tcps_rcvmemdrop;	/* packets dropped for lack of memory */
372	u_long	tcps_rcvshort;		/* packets received too short */
373	u_long	tcps_rcvduppack;	/* duplicate-only packets received */
374	u_long	tcps_rcvdupbyte;	/* duplicate-only bytes received */
375	u_long	tcps_rcvpartduppack;	/* packets with some duplicate data */
376	u_long	tcps_rcvpartdupbyte;	/* dup. bytes in part-dup. packets */
377	u_long	tcps_rcvoopack;		/* out-of-order packets received */
378	u_long	tcps_rcvoobyte;		/* out-of-order bytes received */
379	u_long	tcps_rcvpackafterwin;	/* packets with data after window */
380	u_long	tcps_rcvbyteafterwin;	/* bytes rcvd after window */
381	u_long	tcps_rcvafterclose;	/* packets rcvd after "close" */
382	u_long	tcps_rcvwinprobe;	/* rcvd window probe packets */
383	u_long	tcps_rcvdupack;		/* rcvd duplicate acks */
384	u_long	tcps_rcvacktoomuch;	/* rcvd acks for unsent data */
385	u_long	tcps_rcvackpack;	/* rcvd ack packets */
386	u_long	tcps_rcvackbyte;	/* bytes acked by rcvd acks */
387	u_long	tcps_rcvwinupd;		/* rcvd window update packets */
388	u_long	tcps_pawsdrop;		/* segments dropped due to PAWS */
389	u_long	tcps_predack;		/* times hdr predict ok for acks */
390	u_long	tcps_preddat;		/* times hdr predict ok for data pkts */
391	u_long	tcps_pcbcachemiss;
392	u_long	tcps_cachedrtt;		/* times cached RTT in route updated */
393	u_long	tcps_cachedrttvar;	/* times cached rttvar updated */
394	u_long	tcps_cachedssthresh;	/* times cached ssthresh updated */
395	u_long	tcps_usedrtt;		/* times RTT initialized from route */
396	u_long	tcps_usedrttvar;	/* times RTTVAR initialized from rt */
397	u_long	tcps_usedssthresh;	/* times ssthresh initialized from rt*/
398	u_long	tcps_persistdrop;	/* timeout in persist state */
399	u_long	tcps_badsyn;		/* bogus SYN, e.g. premature ACK */
400	u_long	tcps_mturesent;		/* resends due to MTU discovery */
401	u_long	tcps_listendrop;	/* listen queue overflows */
402	u_long	tcps_badrst;		/* ignored RSTs in the window */
403
404	u_long	tcps_sc_added;		/* entry added to syncache */
405	u_long	tcps_sc_retransmitted;	/* syncache entry was retransmitted */
406	u_long	tcps_sc_dupsyn;		/* duplicate SYN packet */
407	u_long	tcps_sc_dropped;	/* could not reply to packet */
408	u_long	tcps_sc_completed;	/* successful extraction of entry */
409	u_long	tcps_sc_bucketoverflow;	/* syncache per-bucket limit hit */
410	u_long	tcps_sc_cacheoverflow;	/* syncache cache limit hit */
411	u_long	tcps_sc_reset;		/* RST removed entry from syncache */
412	u_long	tcps_sc_stale;		/* timed out or listen socket gone */
413	u_long	tcps_sc_aborted;	/* syncache entry aborted */
414	u_long	tcps_sc_badack;		/* removed due to bad ACK */
415	u_long	tcps_sc_unreach;	/* ICMP unreachable received */
416	u_long	tcps_sc_zonefail;	/* zalloc() failed */
417	u_long	tcps_sc_sendcookie;	/* SYN cookie sent */
418	u_long	tcps_sc_recvcookie;	/* SYN cookie received */
419
420	u_long	tcps_hc_added;		/* entry added to hostcache */
421	u_long	tcps_hc_bucketoverflow;	/* hostcache per bucket limit hit */
422
423	u_long  tcps_finwait2_drops;    /* Drop FIN_WAIT_2 connection after time limit */
424
425	/* SACK related stats */
426	u_long	tcps_sack_recovery_episode; /* SACK recovery episodes */
427	u_long  tcps_sack_rexmits;	    /* SACK rexmit segments   */
428	u_long  tcps_sack_rexmit_bytes;	    /* SACK rexmit bytes      */
429	u_long  tcps_sack_rcv_blocks;	    /* SACK blocks (options) received */
430	u_long  tcps_sack_send_blocks;	    /* SACK blocks (options) sent     */
431	u_long  tcps_sack_sboverflow; 	    /* times scoreboard overflowed */
432};
433
434/*
435 * TCB structure exported to user-land via sysctl(3).
436 * Evil hack: declare only if in_pcb.h and sys/socketvar.h have been
437 * included.  Not all of our clients do.
438 */
439#if defined(_NETINET_IN_PCB_H_) && defined(_SYS_SOCKETVAR_H_)
440struct	xtcpcb {
441	size_t	xt_len;
442	struct	inpcb	xt_inp;
443	struct	tcpcb	xt_tp;
444	struct	xsocket	xt_socket;
445	u_quad_t	xt_alignment_hack;
446};
447#endif
448
449/*
450 * Names for TCP sysctl objects
451 */
452#define	TCPCTL_DO_RFC1323	1	/* use RFC-1323 extensions */
453#define	TCPCTL_MSSDFLT		3	/* MSS default */
454#define TCPCTL_STATS		4	/* statistics (read-only) */
455#define	TCPCTL_RTTDFLT		5	/* default RTT estimate */
456#define	TCPCTL_KEEPIDLE		6	/* keepalive idle timer */
457#define	TCPCTL_KEEPINTVL	7	/* interval to send keepalives */
458#define	TCPCTL_SENDSPACE	8	/* send buffer space */
459#define	TCPCTL_RECVSPACE	9	/* receive buffer space */
460#define	TCPCTL_KEEPINIT		10	/* timeout for establishing syn */
461#define	TCPCTL_PCBLIST		11	/* list of all outstanding PCBs */
462#define	TCPCTL_DELACKTIME	12	/* time before sending delayed ACK */
463#define	TCPCTL_V6MSSDFLT	13	/* MSS default for IPv6 */
464#define	TCPCTL_SACK		14	/* Selective Acknowledgement,rfc 2018 */
465#define	TCPCTL_DROP		15	/* drop tcp connection */
466#define	TCPCTL_MAXID		16
467#define TCPCTL_FINWAIT2_TIMEOUT        17
468
469#define TCPCTL_NAMES { \
470	{ 0, 0 }, \
471	{ "rfc1323", CTLTYPE_INT }, \
472	{ "mssdflt", CTLTYPE_INT }, \
473	{ "stats", CTLTYPE_STRUCT }, \
474	{ "rttdflt", CTLTYPE_INT }, \
475	{ "keepidle", CTLTYPE_INT }, \
476	{ "keepintvl", CTLTYPE_INT }, \
477	{ "sendspace", CTLTYPE_INT }, \
478	{ "recvspace", CTLTYPE_INT }, \
479	{ "keepinit", CTLTYPE_INT }, \
480	{ "pcblist", CTLTYPE_STRUCT }, \
481	{ "delacktime", CTLTYPE_INT }, \
482	{ "v6mssdflt", CTLTYPE_INT }, \
483	{ "maxid", CTLTYPE_INT }, \
484}
485
486
487#ifdef _KERNEL
488#ifdef SYSCTL_DECL
489SYSCTL_DECL(_net_inet_tcp);
490SYSCTL_DECL(_net_inet_tcp_sack);
491#endif
492
493extern	struct inpcbhead tcb;		/* head of queue of active tcpcb's */
494extern	struct inpcbinfo tcbinfo;
495extern	struct tcpstat tcpstat;	/* tcp statistics */
496extern	int tcp_mssdflt;	/* XXX */
497extern	int tcp_minmss;
498extern	int tcp_delack_enabled;
499extern	int tcp_do_newreno;
500extern	int path_mtu_discovery;
501extern	int ss_fltsz;
502extern	int ss_fltsz_local;
503
504extern	int tcp_do_sack;	/* SACK enabled/disabled */
505
506int	 tcp_addoptions(struct tcpopt *, u_char *);
507struct tcpcb *
508	 tcp_close(struct tcpcb *);
509void	 tcp_discardcb(struct tcpcb *);
510void	 tcp_twstart(struct tcpcb *);
511#if 0
512int	 tcp_twrecycleable(struct tcptw *tw);
513#endif
514void	 tcp_twclose(struct tcptw *_tw, int _reuse);
515void	 tcp_ctlinput(int, struct sockaddr *, void *);
516int	 tcp_ctloutput(struct socket *, struct sockopt *);
517struct tcpcb *
518	 tcp_drop(struct tcpcb *, int);
519void	 tcp_drain(void);
520void	 tcp_fasttimo(void);
521void	 tcp_init(void);
522void	 tcp_fini(void *);
523int	 tcp_reass(struct tcpcb *, struct tcphdr *, int *, struct mbuf *);
524void	 tcp_reass_init(void);
525void	 tcp_input(struct mbuf *, int);
526u_long	 tcp_maxmtu(struct in_conninfo *, int *);
527u_long	 tcp_maxmtu6(struct in_conninfo *, int *);
528void	 tcp_mss(struct tcpcb *, int);
529int	 tcp_mssopt(struct in_conninfo *);
530struct inpcb *
531	 tcp_drop_syn_sent(struct inpcb *, int);
532struct inpcb *
533	 tcp_mtudisc(struct inpcb *, int);
534struct tcpcb *
535	 tcp_newtcpcb(struct inpcb *);
536int	 tcp_output(struct tcpcb *);
537void	 tcp_respond(struct tcpcb *, void *,
538	    struct tcphdr *, struct mbuf *, tcp_seq, tcp_seq, int);
539void	 tcp_tw_init(void);
540void	 tcp_tw_zone_change(void);
541int	 tcp_twrespond(struct tcptw *, int);
542void	 tcp_setpersist(struct tcpcb *);
543#ifdef TCP_SIGNATURE
544int	 tcp_signature_compute(struct mbuf *, int, int, int, u_char *, u_int);
545#endif
546void	 tcp_slowtimo(void);
547struct tcptemp *
548	 tcpip_maketemplate(struct inpcb *);
549void	 tcpip_fillheaders(struct inpcb *, void *, void *);
550void	 tcp_timer_activate(struct tcpcb *, int, u_int);
551int	 tcp_timer_active(struct tcpcb *, int);
552void	 tcp_trace(short, short, struct tcpcb *, void *, struct tcphdr *, int);
553void	 tcp_xmit_bandwidth_limit(struct tcpcb *tp, tcp_seq ack_seq);
554void	 syncache_init(void);
555void	 syncache_unreach(struct in_conninfo *, struct tcphdr *);
556int	 syncache_expand(struct in_conninfo *, struct tcpopt *,
557	     struct tcphdr *, struct socket **, struct mbuf *);
558void	 syncache_add(struct in_conninfo *, struct tcpopt *,
559	     struct tcphdr *, struct inpcb *, struct socket **, struct mbuf *);
560void	 syncache_chkrst(struct in_conninfo *, struct tcphdr *);
561void	 syncache_badack(struct in_conninfo *);
562/*
563 * All tcp_hc_* functions are IPv4 and IPv6 (via in_conninfo)
564 */
565void	 tcp_hc_init(void);
566void	 tcp_hc_get(struct in_conninfo *, struct hc_metrics_lite *);
567u_long	 tcp_hc_getmtu(struct in_conninfo *);
568void	 tcp_hc_updatemtu(struct in_conninfo *, u_long);
569void	 tcp_hc_update(struct in_conninfo *, struct hc_metrics_lite *);
570
571extern	struct pr_usrreqs tcp_usrreqs;
572extern	u_long tcp_sendspace;
573extern	u_long tcp_recvspace;
574tcp_seq tcp_new_isn(struct tcpcb *);
575
576void	 tcp_sack_doack(struct tcpcb *, struct tcpopt *, tcp_seq);
577void	 tcp_update_sack_list(struct tcpcb *tp, tcp_seq rcv_laststart, tcp_seq rcv_lastend);
578void	 tcp_clean_sackreport(struct tcpcb *tp);
579void	 tcp_sack_adjust(struct tcpcb *tp);
580struct sackhole *tcp_sack_output(struct tcpcb *tp, int *sack_bytes_rexmt);
581void	 tcp_sack_partialack(struct tcpcb *, struct tcphdr *);
582void	 tcp_free_sackholes(struct tcpcb *tp);
583int	 tcp_newreno(struct tcpcb *, struct tcphdr *);
584u_long	 tcp_seq_subtract(u_long, u_long );
585
586#endif /* _KERNEL */
587
588#endif /* _NETINET_TCP_VAR_H_ */
589