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 * 3. All advertising materials mentioning features or use of this software
14 * must display the following acknowledgement:
15 * This product includes software developed by the University of
16 * California, Berkeley and its contributors.
17 * 4. Neither the name of the University nor the names of its contributors
18 * may be used to endorse or promote products derived from this software
19 * without specific prior written permission.
20 *
21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31 * SUCH DAMAGE.
32 *
33 * @(#)tcp_var.h 8.4 (Berkeley) 5/24/95
| 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 * 3. All advertising materials mentioning features or use of this software
14 * must display the following acknowledgement:
15 * This product includes software developed by the University of
16 * California, Berkeley and its contributors.
17 * 4. Neither the name of the University nor the names of its contributors
18 * may be used to endorse or promote products derived from this software
19 * without specific prior written permission.
20 *
21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31 * SUCH DAMAGE.
32 *
33 * @(#)tcp_var.h 8.4 (Berkeley) 5/24/95
|
34 * $FreeBSD: head/sys/netinet/tcp_var.h 52904 1999-11-05 14:41:39Z shin $
| 34 * $FreeBSD: head/sys/netinet/tcp_var.h 55009 1999-12-22 19:13:38Z shin $
|
35 */
36
37#ifndef _NETINET_TCP_VAR_H_
38#define _NETINET_TCP_VAR_H_
39/*
40 * Kernel variables for tcp.
41 */
42
43/*
44 * Tcp control block, one per tcp; fields:
45 * Organized for 16 byte cacheline efficiency.
46 */
47struct tcpcb {
48 struct mbuf *t_segq;
49 int t_dupacks; /* consecutive dup acks recd */
50 struct tcpiphdr *t_template; /* skeletal packet for transmit */
51
52 struct callout *tt_rexmt; /* retransmit timer */
53 struct callout *tt_persist; /* retransmit persistence */
54 struct callout *tt_keep; /* keepalive */
55 struct callout *tt_2msl; /* 2*msl TIME_WAIT timer */
56 struct callout *tt_delack; /* delayed ACK timer */
57
58 struct inpcb *t_inpcb; /* back pointer to internet pcb */
59 int t_state; /* state of this connection */
60 u_int t_flags;
61#define TF_ACKNOW 0x00001 /* ack peer immediately */
62#define TF_DELACK 0x00002 /* ack, but try to delay it */
63#define TF_NODELAY 0x00004 /* don't delay packets to coalesce */
64#define TF_NOOPT 0x00008 /* don't use tcp options */
65#define TF_SENTFIN 0x00010 /* have sent FIN */
66#define TF_REQ_SCALE 0x00020 /* have/will request window scaling */
67#define TF_RCVD_SCALE 0x00040 /* other side has requested scaling */
68#define TF_REQ_TSTMP 0x00080 /* have/will request timestamps */
69#define TF_RCVD_TSTMP 0x00100 /* a timestamp was received in SYN */
70#define TF_SACK_PERMIT 0x00200 /* other side said I could SACK */
71#define TF_NEEDSYN 0x00400 /* send SYN (implicit state) */
72#define TF_NEEDFIN 0x00800 /* send FIN (implicit state) */
73#define TF_NOPUSH 0x01000 /* don't push */
74#define TF_REQ_CC 0x02000 /* have/will request CC */
75#define TF_RCVD_CC 0x04000 /* a CC was received in SYN */
76#define TF_SENDCCNEW 0x08000 /* send CCnew instead of CC in SYN */
77#define TF_MORETOCOME 0x10000 /* More data to be appended to sock */
78 int t_force; /* 1 if forcing out a byte */
79
80 tcp_seq snd_una; /* send unacknowledged */
81 tcp_seq snd_max; /* highest sequence number sent;
82 * used to recognize retransmits
83 */
84 tcp_seq snd_nxt; /* send next */
85 tcp_seq snd_up; /* send urgent pointer */
86
87 tcp_seq snd_wl1; /* window update seg seq number */
88 tcp_seq snd_wl2; /* window update seg ack number */
89 tcp_seq iss; /* initial send sequence number */
90 tcp_seq irs; /* initial receive sequence number */
91
92 tcp_seq rcv_nxt; /* receive next */
93 tcp_seq rcv_adv; /* advertised window */
94 u_long rcv_wnd; /* receive window */
95 tcp_seq rcv_up; /* receive urgent pointer */
96
97 u_long snd_wnd; /* send window */
98 u_long snd_cwnd; /* congestion-controlled window */
99 u_long snd_ssthresh; /* snd_cwnd size threshold for
100 * for slow start exponential to
101 * linear switch
102 */
103 u_int t_maxopd; /* mss plus options */
104
105 u_long t_rcvtime; /* inactivity time */
106 u_long t_starttime; /* time connection was established */
107 int t_rtttime; /* round trip time */
108 tcp_seq t_rtseq; /* sequence number being timed */
109
110 int t_rxtcur; /* current retransmit value (ticks) */
111 u_int t_maxseg; /* maximum segment size */
112 int t_srtt; /* smoothed round-trip time */
113 int t_rttvar; /* variance in round-trip time */
114
115 int t_rxtshift; /* log(2) of rexmt exp. backoff */
116 u_int t_rttmin; /* minimum rtt allowed */
117 u_long t_rttupdated; /* number of times rtt sampled */
118 u_long max_sndwnd; /* largest window peer has offered */
119
120 int t_softerror; /* possible error not yet reported */
121/* out-of-band data */
122 char t_oobflags; /* have some */
123 char t_iobc; /* input character */
124#define TCPOOB_HAVEDATA 0x01
125#define TCPOOB_HADDATA 0x02
126/* RFC 1323 variables */
127 u_char snd_scale; /* window scaling for send window */
128 u_char rcv_scale; /* window scaling for recv window */
129 u_char request_r_scale; /* pending window scaling */
130 u_char requested_s_scale;
131 u_long ts_recent; /* timestamp echo data */
132
133 u_long ts_recent_age; /* when last updated */
134 tcp_seq last_ack_sent;
135/* RFC 1644 variables */
136 tcp_cc cc_send; /* send connection count */
137 tcp_cc cc_recv; /* receive connection count */
138/* experimental */
139 u_long snd_cwnd_prev; /* cwnd prior to retransmit */
140 u_long snd_ssthresh_prev; /* ssthresh prior to retransmit */
141 u_long t_badrxtwin; /* window for retransmit recovery */
142};
143
144/*
145 * Structure to hold TCP options that are only used during segment
146 * processing (in tcp_input), but not held in the tcpcb.
147 * It's basically used to reduce the number of parameters
148 * to tcp_dooptions.
149 */
150struct tcpopt {
151 u_long to_flag; /* which options are present */
152#define TOF_TS 0x0001 /* timestamp */
153#define TOF_CC 0x0002 /* CC and CCnew are exclusive */
154#define TOF_CCNEW 0x0004
155#define TOF_CCECHO 0x0008
156 u_long to_tsval;
157 u_long to_tsecr;
158 tcp_cc to_cc; /* holds CC or CCnew */
159 tcp_cc to_ccecho;
160};
161
162/*
163 * The TAO cache entry which is stored in the protocol family specific
164 * portion of the route metrics.
165 */
166struct rmxp_tao {
167 tcp_cc tao_cc; /* latest CC in valid SYN */
168 tcp_cc tao_ccsent; /* latest CC sent to peer */
169 u_short tao_mssopt; /* peer's cached MSS */
170#ifdef notyet
171 u_short tao_flags; /* cache status flags */
172#define TAOF_DONT 0x0001 /* peer doesn't understand rfc1644 */
173#define TAOF_OK 0x0002 /* peer does understand rfc1644 */
174#define TAOF_UNDEF 0 /* we don't know yet */
175#endif /* notyet */
176};
177#define rmx_taop(r) ((struct rmxp_tao *)(r).rmx_filler)
178
179#define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb)
180#define sototcpcb(so) (intotcpcb(sotoinpcb(so)))
181
182/*
183 * The smoothed round-trip time and estimated variance
184 * are stored as fixed point numbers scaled by the values below.
185 * For convenience, these scales are also used in smoothing the average
186 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed).
187 * With these scales, srtt has 3 bits to the right of the binary point,
188 * and thus an "ALPHA" of 0.875. rttvar has 2 bits to the right of the
189 * binary point, and is smoothed with an ALPHA of 0.75.
190 */
191#define TCP_RTT_SCALE 32 /* multiplier for srtt; 3 bits frac. */
192#define TCP_RTT_SHIFT 5 /* shift for srtt; 3 bits frac. */
193#define TCP_RTTVAR_SCALE 16 /* multiplier for rttvar; 2 bits */
194#define TCP_RTTVAR_SHIFT 4 /* shift for rttvar; 2 bits */
195#define TCP_DELTA_SHIFT 2 /* see tcp_input.c */
196
197/*
198 * The initial retransmission should happen at rtt + 4 * rttvar.
199 * Because of the way we do the smoothing, srtt and rttvar
200 * will each average +1/2 tick of bias. When we compute
201 * the retransmit timer, we want 1/2 tick of rounding and
202 * 1 extra tick because of +-1/2 tick uncertainty in the
203 * firing of the timer. The bias will give us exactly the
204 * 1.5 tick we need. But, because the bias is
205 * statistical, we have to test that we don't drop below
206 * the minimum feasible timer (which is 2 ticks).
207 * This version of the macro adapted from a paper by Lawrence
208 * Brakmo and Larry Peterson which outlines a problem caused
209 * by insufficient precision in the original implementation,
210 * which results in inappropriately large RTO values for very
211 * fast networks.
212 */
213#define TCP_REXMTVAL(tp) \
214 max((tp)->t_rttmin, (((tp)->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT)) \
215 + (tp)->t_rttvar) >> TCP_DELTA_SHIFT)
216
217/*
218 * TCP statistics.
219 * Many of these should be kept per connection,
220 * but that's inconvenient at the moment.
221 */
222struct tcpstat {
223 u_long tcps_connattempt; /* connections initiated */
224 u_long tcps_accepts; /* connections accepted */
225 u_long tcps_connects; /* connections established */
226 u_long tcps_drops; /* connections dropped */
227 u_long tcps_conndrops; /* embryonic connections dropped */
228 u_long tcps_closed; /* conn. closed (includes drops) */
229 u_long tcps_segstimed; /* segs where we tried to get rtt */
230 u_long tcps_rttupdated; /* times we succeeded */
231 u_long tcps_delack; /* delayed acks sent */
232 u_long tcps_timeoutdrop; /* conn. dropped in rxmt timeout */
233 u_long tcps_rexmttimeo; /* retransmit timeouts */
234 u_long tcps_persisttimeo; /* persist timeouts */
235 u_long tcps_keeptimeo; /* keepalive timeouts */
236 u_long tcps_keepprobe; /* keepalive probes sent */
237 u_long tcps_keepdrops; /* connections dropped in keepalive */
238
239 u_long tcps_sndtotal; /* total packets sent */
240 u_long tcps_sndpack; /* data packets sent */
241 u_long tcps_sndbyte; /* data bytes sent */
242 u_long tcps_sndrexmitpack; /* data packets retransmitted */
243 u_long tcps_sndrexmitbyte; /* data bytes retransmitted */
244 u_long tcps_sndacks; /* ack-only packets sent */
245 u_long tcps_sndprobe; /* window probes sent */
246 u_long tcps_sndurg; /* packets sent with URG only */
247 u_long tcps_sndwinup; /* window update-only packets sent */
248 u_long tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */
249
250 u_long tcps_rcvtotal; /* total packets received */
251 u_long tcps_rcvpack; /* packets received in sequence */
252 u_long tcps_rcvbyte; /* bytes received in sequence */
253 u_long tcps_rcvbadsum; /* packets received with ccksum errs */
254 u_long tcps_rcvbadoff; /* packets received with bad offset */
255 u_long tcps_rcvmemdrop; /* packets dropped for lack of memory */
256 u_long tcps_rcvshort; /* packets received too short */
257 u_long tcps_rcvduppack; /* duplicate-only packets received */
258 u_long tcps_rcvdupbyte; /* duplicate-only bytes received */
259 u_long tcps_rcvpartduppack; /* packets with some duplicate data */
260 u_long tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */
261 u_long tcps_rcvoopack; /* out-of-order packets received */
262 u_long tcps_rcvoobyte; /* out-of-order bytes received */
263 u_long tcps_rcvpackafterwin; /* packets with data after window */
264 u_long tcps_rcvbyteafterwin; /* bytes rcvd after window */
265 u_long tcps_rcvafterclose; /* packets rcvd after "close" */
266 u_long tcps_rcvwinprobe; /* rcvd window probe packets */
267 u_long tcps_rcvdupack; /* rcvd duplicate acks */
268 u_long tcps_rcvacktoomuch; /* rcvd acks for unsent data */
269 u_long tcps_rcvackpack; /* rcvd ack packets */
270 u_long tcps_rcvackbyte; /* bytes acked by rcvd acks */
271 u_long tcps_rcvwinupd; /* rcvd window update packets */
272 u_long tcps_pawsdrop; /* segments dropped due to PAWS */
273 u_long tcps_predack; /* times hdr predict ok for acks */
274 u_long tcps_preddat; /* times hdr predict ok for data pkts */
275 u_long tcps_pcbcachemiss;
276 u_long tcps_cachedrtt; /* times cached RTT in route updated */
277 u_long tcps_cachedrttvar; /* times cached rttvar updated */
278 u_long tcps_cachedssthresh; /* times cached ssthresh updated */
279 u_long tcps_usedrtt; /* times RTT initialized from route */
280 u_long tcps_usedrttvar; /* times RTTVAR initialized from rt */
281 u_long tcps_usedssthresh; /* times ssthresh initialized from rt*/
282 u_long tcps_persistdrop; /* timeout in persist state */
283 u_long tcps_badsyn; /* bogus SYN, e.g. premature ACK */
284 u_long tcps_mturesent; /* resends due to MTU discovery */
285 u_long tcps_listendrop; /* listen queue overflows */
286};
287
288/*
289 * TCB structure exported to user-land via sysctl(3).
290 * Evil hack: declare only if in_pcb.h and sys/socketvar.h have been
291 * included. Not all of our clients do.
292 */
293#if defined(_NETINET_IN_PCB_H_) && defined(_SYS_SOCKETVAR_H_)
294struct xtcpcb {
295 size_t xt_len;
296 struct inpcb xt_inp;
297 struct tcpcb xt_tp;
298 struct xsocket xt_socket;
299 u_quad_t xt_alignment_hack;
300};
301#endif
302
303/*
304 * Names for TCP sysctl objects
305 */
306#define TCPCTL_DO_RFC1323 1 /* use RFC-1323 extensions */
307#define TCPCTL_DO_RFC1644 2 /* use RFC-1644 extensions */
308#define TCPCTL_MSSDFLT 3 /* MSS default */
309#define TCPCTL_STATS 4 /* statistics (read-only) */
310#define TCPCTL_RTTDFLT 5 /* default RTT estimate */
311#define TCPCTL_KEEPIDLE 6 /* keepalive idle timer */
312#define TCPCTL_KEEPINTVL 7 /* interval to send keepalives */
313#define TCPCTL_SENDSPACE 8 /* send buffer space */
314#define TCPCTL_RECVSPACE 9 /* receive buffer space */
315#define TCPCTL_KEEPINIT 10 /* receive buffer space */
316#define TCPCTL_PCBLIST 11 /* list of all outstanding PCBs */
317#define TCPCTL_DELACKTIME 12 /* time before sending delayed ACK */
318#define TCPCTL_V6MSSDFLT 13 /* MSS default for IPv6 */
319#define TCPCTL_MAXID 14
320
321#define TCPCTL_NAMES { \
322 { 0, 0 }, \
323 { "rfc1323", CTLTYPE_INT }, \
324 { "rfc1644", CTLTYPE_INT }, \
325 { "mssdflt", CTLTYPE_INT }, \
326 { "stats", CTLTYPE_STRUCT }, \
327 { "rttdflt", CTLTYPE_INT }, \
328 { "keepidle", CTLTYPE_INT }, \
329 { "keepintvl", CTLTYPE_INT }, \
330 { "sendspace", CTLTYPE_INT }, \
331 { "recvspace", CTLTYPE_INT }, \
332 { "keepinit", CTLTYPE_INT }, \
333 { "pcblist", CTLTYPE_STRUCT }, \
334 { "delacktime", CTLTYPE_INT }, \
335 { "v6mssdflt", CTLTYPE_INT }, \
336}
337
338
339#ifdef KERNEL
340#ifdef SYSCTL_DECL
341SYSCTL_DECL(_net_inet_tcp);
342#endif
343
344extern struct inpcbhead tcb; /* head of queue of active tcpcb's */
345extern struct inpcbinfo tcbinfo;
346extern struct tcpstat tcpstat; /* tcp statistics */
347extern int tcp_mssdflt; /* XXX */
348extern int tcp_delack_enabled;
349extern int ss_fltsz;
350extern int ss_fltsz_local;
351
352void tcp_canceltimers __P((struct tcpcb *));
353struct tcpcb *
354 tcp_close __P((struct tcpcb *));
355void tcp_ctlinput __P((int, struct sockaddr *, void *));
356int tcp_ctloutput __P((struct socket *, struct sockopt *));
357struct tcpcb *
358 tcp_drop __P((struct tcpcb *, int));
359void tcp_drain __P((void));
360void tcp_fasttimo __P((void));
361struct rmxp_tao *
362 tcp_gettaocache __P((struct inpcb *));
363void tcp_init __P((void));
| 35 */
36
37#ifndef _NETINET_TCP_VAR_H_
38#define _NETINET_TCP_VAR_H_
39/*
40 * Kernel variables for tcp.
41 */
42
43/*
44 * Tcp control block, one per tcp; fields:
45 * Organized for 16 byte cacheline efficiency.
46 */
47struct tcpcb {
48 struct mbuf *t_segq;
49 int t_dupacks; /* consecutive dup acks recd */
50 struct tcpiphdr *t_template; /* skeletal packet for transmit */
51
52 struct callout *tt_rexmt; /* retransmit timer */
53 struct callout *tt_persist; /* retransmit persistence */
54 struct callout *tt_keep; /* keepalive */
55 struct callout *tt_2msl; /* 2*msl TIME_WAIT timer */
56 struct callout *tt_delack; /* delayed ACK timer */
57
58 struct inpcb *t_inpcb; /* back pointer to internet pcb */
59 int t_state; /* state of this connection */
60 u_int t_flags;
61#define TF_ACKNOW 0x00001 /* ack peer immediately */
62#define TF_DELACK 0x00002 /* ack, but try to delay it */
63#define TF_NODELAY 0x00004 /* don't delay packets to coalesce */
64#define TF_NOOPT 0x00008 /* don't use tcp options */
65#define TF_SENTFIN 0x00010 /* have sent FIN */
66#define TF_REQ_SCALE 0x00020 /* have/will request window scaling */
67#define TF_RCVD_SCALE 0x00040 /* other side has requested scaling */
68#define TF_REQ_TSTMP 0x00080 /* have/will request timestamps */
69#define TF_RCVD_TSTMP 0x00100 /* a timestamp was received in SYN */
70#define TF_SACK_PERMIT 0x00200 /* other side said I could SACK */
71#define TF_NEEDSYN 0x00400 /* send SYN (implicit state) */
72#define TF_NEEDFIN 0x00800 /* send FIN (implicit state) */
73#define TF_NOPUSH 0x01000 /* don't push */
74#define TF_REQ_CC 0x02000 /* have/will request CC */
75#define TF_RCVD_CC 0x04000 /* a CC was received in SYN */
76#define TF_SENDCCNEW 0x08000 /* send CCnew instead of CC in SYN */
77#define TF_MORETOCOME 0x10000 /* More data to be appended to sock */
78 int t_force; /* 1 if forcing out a byte */
79
80 tcp_seq snd_una; /* send unacknowledged */
81 tcp_seq snd_max; /* highest sequence number sent;
82 * used to recognize retransmits
83 */
84 tcp_seq snd_nxt; /* send next */
85 tcp_seq snd_up; /* send urgent pointer */
86
87 tcp_seq snd_wl1; /* window update seg seq number */
88 tcp_seq snd_wl2; /* window update seg ack number */
89 tcp_seq iss; /* initial send sequence number */
90 tcp_seq irs; /* initial receive sequence number */
91
92 tcp_seq rcv_nxt; /* receive next */
93 tcp_seq rcv_adv; /* advertised window */
94 u_long rcv_wnd; /* receive window */
95 tcp_seq rcv_up; /* receive urgent pointer */
96
97 u_long snd_wnd; /* send window */
98 u_long snd_cwnd; /* congestion-controlled window */
99 u_long snd_ssthresh; /* snd_cwnd size threshold for
100 * for slow start exponential to
101 * linear switch
102 */
103 u_int t_maxopd; /* mss plus options */
104
105 u_long t_rcvtime; /* inactivity time */
106 u_long t_starttime; /* time connection was established */
107 int t_rtttime; /* round trip time */
108 tcp_seq t_rtseq; /* sequence number being timed */
109
110 int t_rxtcur; /* current retransmit value (ticks) */
111 u_int t_maxseg; /* maximum segment size */
112 int t_srtt; /* smoothed round-trip time */
113 int t_rttvar; /* variance in round-trip time */
114
115 int t_rxtshift; /* log(2) of rexmt exp. backoff */
116 u_int t_rttmin; /* minimum rtt allowed */
117 u_long t_rttupdated; /* number of times rtt sampled */
118 u_long max_sndwnd; /* largest window peer has offered */
119
120 int t_softerror; /* possible error not yet reported */
121/* out-of-band data */
122 char t_oobflags; /* have some */
123 char t_iobc; /* input character */
124#define TCPOOB_HAVEDATA 0x01
125#define TCPOOB_HADDATA 0x02
126/* RFC 1323 variables */
127 u_char snd_scale; /* window scaling for send window */
128 u_char rcv_scale; /* window scaling for recv window */
129 u_char request_r_scale; /* pending window scaling */
130 u_char requested_s_scale;
131 u_long ts_recent; /* timestamp echo data */
132
133 u_long ts_recent_age; /* when last updated */
134 tcp_seq last_ack_sent;
135/* RFC 1644 variables */
136 tcp_cc cc_send; /* send connection count */
137 tcp_cc cc_recv; /* receive connection count */
138/* experimental */
139 u_long snd_cwnd_prev; /* cwnd prior to retransmit */
140 u_long snd_ssthresh_prev; /* ssthresh prior to retransmit */
141 u_long t_badrxtwin; /* window for retransmit recovery */
142};
143
144/*
145 * Structure to hold TCP options that are only used during segment
146 * processing (in tcp_input), but not held in the tcpcb.
147 * It's basically used to reduce the number of parameters
148 * to tcp_dooptions.
149 */
150struct tcpopt {
151 u_long to_flag; /* which options are present */
152#define TOF_TS 0x0001 /* timestamp */
153#define TOF_CC 0x0002 /* CC and CCnew are exclusive */
154#define TOF_CCNEW 0x0004
155#define TOF_CCECHO 0x0008
156 u_long to_tsval;
157 u_long to_tsecr;
158 tcp_cc to_cc; /* holds CC or CCnew */
159 tcp_cc to_ccecho;
160};
161
162/*
163 * The TAO cache entry which is stored in the protocol family specific
164 * portion of the route metrics.
165 */
166struct rmxp_tao {
167 tcp_cc tao_cc; /* latest CC in valid SYN */
168 tcp_cc tao_ccsent; /* latest CC sent to peer */
169 u_short tao_mssopt; /* peer's cached MSS */
170#ifdef notyet
171 u_short tao_flags; /* cache status flags */
172#define TAOF_DONT 0x0001 /* peer doesn't understand rfc1644 */
173#define TAOF_OK 0x0002 /* peer does understand rfc1644 */
174#define TAOF_UNDEF 0 /* we don't know yet */
175#endif /* notyet */
176};
177#define rmx_taop(r) ((struct rmxp_tao *)(r).rmx_filler)
178
179#define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb)
180#define sototcpcb(so) (intotcpcb(sotoinpcb(so)))
181
182/*
183 * The smoothed round-trip time and estimated variance
184 * are stored as fixed point numbers scaled by the values below.
185 * For convenience, these scales are also used in smoothing the average
186 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed).
187 * With these scales, srtt has 3 bits to the right of the binary point,
188 * and thus an "ALPHA" of 0.875. rttvar has 2 bits to the right of the
189 * binary point, and is smoothed with an ALPHA of 0.75.
190 */
191#define TCP_RTT_SCALE 32 /* multiplier for srtt; 3 bits frac. */
192#define TCP_RTT_SHIFT 5 /* shift for srtt; 3 bits frac. */
193#define TCP_RTTVAR_SCALE 16 /* multiplier for rttvar; 2 bits */
194#define TCP_RTTVAR_SHIFT 4 /* shift for rttvar; 2 bits */
195#define TCP_DELTA_SHIFT 2 /* see tcp_input.c */
196
197/*
198 * The initial retransmission should happen at rtt + 4 * rttvar.
199 * Because of the way we do the smoothing, srtt and rttvar
200 * will each average +1/2 tick of bias. When we compute
201 * the retransmit timer, we want 1/2 tick of rounding and
202 * 1 extra tick because of +-1/2 tick uncertainty in the
203 * firing of the timer. The bias will give us exactly the
204 * 1.5 tick we need. But, because the bias is
205 * statistical, we have to test that we don't drop below
206 * the minimum feasible timer (which is 2 ticks).
207 * This version of the macro adapted from a paper by Lawrence
208 * Brakmo and Larry Peterson which outlines a problem caused
209 * by insufficient precision in the original implementation,
210 * which results in inappropriately large RTO values for very
211 * fast networks.
212 */
213#define TCP_REXMTVAL(tp) \
214 max((tp)->t_rttmin, (((tp)->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT)) \
215 + (tp)->t_rttvar) >> TCP_DELTA_SHIFT)
216
217/*
218 * TCP statistics.
219 * Many of these should be kept per connection,
220 * but that's inconvenient at the moment.
221 */
222struct tcpstat {
223 u_long tcps_connattempt; /* connections initiated */
224 u_long tcps_accepts; /* connections accepted */
225 u_long tcps_connects; /* connections established */
226 u_long tcps_drops; /* connections dropped */
227 u_long tcps_conndrops; /* embryonic connections dropped */
228 u_long tcps_closed; /* conn. closed (includes drops) */
229 u_long tcps_segstimed; /* segs where we tried to get rtt */
230 u_long tcps_rttupdated; /* times we succeeded */
231 u_long tcps_delack; /* delayed acks sent */
232 u_long tcps_timeoutdrop; /* conn. dropped in rxmt timeout */
233 u_long tcps_rexmttimeo; /* retransmit timeouts */
234 u_long tcps_persisttimeo; /* persist timeouts */
235 u_long tcps_keeptimeo; /* keepalive timeouts */
236 u_long tcps_keepprobe; /* keepalive probes sent */
237 u_long tcps_keepdrops; /* connections dropped in keepalive */
238
239 u_long tcps_sndtotal; /* total packets sent */
240 u_long tcps_sndpack; /* data packets sent */
241 u_long tcps_sndbyte; /* data bytes sent */
242 u_long tcps_sndrexmitpack; /* data packets retransmitted */
243 u_long tcps_sndrexmitbyte; /* data bytes retransmitted */
244 u_long tcps_sndacks; /* ack-only packets sent */
245 u_long tcps_sndprobe; /* window probes sent */
246 u_long tcps_sndurg; /* packets sent with URG only */
247 u_long tcps_sndwinup; /* window update-only packets sent */
248 u_long tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */
249
250 u_long tcps_rcvtotal; /* total packets received */
251 u_long tcps_rcvpack; /* packets received in sequence */
252 u_long tcps_rcvbyte; /* bytes received in sequence */
253 u_long tcps_rcvbadsum; /* packets received with ccksum errs */
254 u_long tcps_rcvbadoff; /* packets received with bad offset */
255 u_long tcps_rcvmemdrop; /* packets dropped for lack of memory */
256 u_long tcps_rcvshort; /* packets received too short */
257 u_long tcps_rcvduppack; /* duplicate-only packets received */
258 u_long tcps_rcvdupbyte; /* duplicate-only bytes received */
259 u_long tcps_rcvpartduppack; /* packets with some duplicate data */
260 u_long tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */
261 u_long tcps_rcvoopack; /* out-of-order packets received */
262 u_long tcps_rcvoobyte; /* out-of-order bytes received */
263 u_long tcps_rcvpackafterwin; /* packets with data after window */
264 u_long tcps_rcvbyteafterwin; /* bytes rcvd after window */
265 u_long tcps_rcvafterclose; /* packets rcvd after "close" */
266 u_long tcps_rcvwinprobe; /* rcvd window probe packets */
267 u_long tcps_rcvdupack; /* rcvd duplicate acks */
268 u_long tcps_rcvacktoomuch; /* rcvd acks for unsent data */
269 u_long tcps_rcvackpack; /* rcvd ack packets */
270 u_long tcps_rcvackbyte; /* bytes acked by rcvd acks */
271 u_long tcps_rcvwinupd; /* rcvd window update packets */
272 u_long tcps_pawsdrop; /* segments dropped due to PAWS */
273 u_long tcps_predack; /* times hdr predict ok for acks */
274 u_long tcps_preddat; /* times hdr predict ok for data pkts */
275 u_long tcps_pcbcachemiss;
276 u_long tcps_cachedrtt; /* times cached RTT in route updated */
277 u_long tcps_cachedrttvar; /* times cached rttvar updated */
278 u_long tcps_cachedssthresh; /* times cached ssthresh updated */
279 u_long tcps_usedrtt; /* times RTT initialized from route */
280 u_long tcps_usedrttvar; /* times RTTVAR initialized from rt */
281 u_long tcps_usedssthresh; /* times ssthresh initialized from rt*/
282 u_long tcps_persistdrop; /* timeout in persist state */
283 u_long tcps_badsyn; /* bogus SYN, e.g. premature ACK */
284 u_long tcps_mturesent; /* resends due to MTU discovery */
285 u_long tcps_listendrop; /* listen queue overflows */
286};
287
288/*
289 * TCB structure exported to user-land via sysctl(3).
290 * Evil hack: declare only if in_pcb.h and sys/socketvar.h have been
291 * included. Not all of our clients do.
292 */
293#if defined(_NETINET_IN_PCB_H_) && defined(_SYS_SOCKETVAR_H_)
294struct xtcpcb {
295 size_t xt_len;
296 struct inpcb xt_inp;
297 struct tcpcb xt_tp;
298 struct xsocket xt_socket;
299 u_quad_t xt_alignment_hack;
300};
301#endif
302
303/*
304 * Names for TCP sysctl objects
305 */
306#define TCPCTL_DO_RFC1323 1 /* use RFC-1323 extensions */
307#define TCPCTL_DO_RFC1644 2 /* use RFC-1644 extensions */
308#define TCPCTL_MSSDFLT 3 /* MSS default */
309#define TCPCTL_STATS 4 /* statistics (read-only) */
310#define TCPCTL_RTTDFLT 5 /* default RTT estimate */
311#define TCPCTL_KEEPIDLE 6 /* keepalive idle timer */
312#define TCPCTL_KEEPINTVL 7 /* interval to send keepalives */
313#define TCPCTL_SENDSPACE 8 /* send buffer space */
314#define TCPCTL_RECVSPACE 9 /* receive buffer space */
315#define TCPCTL_KEEPINIT 10 /* receive buffer space */
316#define TCPCTL_PCBLIST 11 /* list of all outstanding PCBs */
317#define TCPCTL_DELACKTIME 12 /* time before sending delayed ACK */
318#define TCPCTL_V6MSSDFLT 13 /* MSS default for IPv6 */
319#define TCPCTL_MAXID 14
320
321#define TCPCTL_NAMES { \
322 { 0, 0 }, \
323 { "rfc1323", CTLTYPE_INT }, \
324 { "rfc1644", CTLTYPE_INT }, \
325 { "mssdflt", CTLTYPE_INT }, \
326 { "stats", CTLTYPE_STRUCT }, \
327 { "rttdflt", CTLTYPE_INT }, \
328 { "keepidle", CTLTYPE_INT }, \
329 { "keepintvl", CTLTYPE_INT }, \
330 { "sendspace", CTLTYPE_INT }, \
331 { "recvspace", CTLTYPE_INT }, \
332 { "keepinit", CTLTYPE_INT }, \
333 { "pcblist", CTLTYPE_STRUCT }, \
334 { "delacktime", CTLTYPE_INT }, \
335 { "v6mssdflt", CTLTYPE_INT }, \
336}
337
338
339#ifdef KERNEL
340#ifdef SYSCTL_DECL
341SYSCTL_DECL(_net_inet_tcp);
342#endif
343
344extern struct inpcbhead tcb; /* head of queue of active tcpcb's */
345extern struct inpcbinfo tcbinfo;
346extern struct tcpstat tcpstat; /* tcp statistics */
347extern int tcp_mssdflt; /* XXX */
348extern int tcp_delack_enabled;
349extern int ss_fltsz;
350extern int ss_fltsz_local;
351
352void tcp_canceltimers __P((struct tcpcb *));
353struct tcpcb *
354 tcp_close __P((struct tcpcb *));
355void tcp_ctlinput __P((int, struct sockaddr *, void *));
356int tcp_ctloutput __P((struct socket *, struct sockopt *));
357struct tcpcb *
358 tcp_drop __P((struct tcpcb *, int));
359void tcp_drain __P((void));
360void tcp_fasttimo __P((void));
361struct rmxp_tao *
362 tcp_gettaocache __P((struct inpcb *));
363void tcp_init __P((void));
|
364void tcp_input __P((struct mbuf *, int));
| 364void tcp_input __P((struct mbuf *, int, int));
|
365void tcp_mss __P((struct tcpcb *, int));
366int tcp_mssopt __P((struct tcpcb *));
367void tcp_mtudisc __P((struct inpcb *, int));
368struct tcpcb *
369 tcp_newtcpcb __P((struct inpcb *));
370int tcp_output __P((struct tcpcb *));
371void tcp_quench __P((struct inpcb *, int));
372void tcp_respond __P((struct tcpcb *,
373 struct tcpiphdr *, struct mbuf *, tcp_seq, tcp_seq, int));
374struct rtentry *
375 tcp_rtlookup __P((struct inpcb *));
376void tcp_setpersist __P((struct tcpcb *));
377void tcp_slowtimo __P((void));
378struct tcpiphdr *
379 tcp_template __P((struct tcpcb *));
380struct tcpcb *
381 tcp_timers __P((struct tcpcb *, int));
382void tcp_trace __P((int, int, struct tcpcb *, struct tcpiphdr *, int));
383
384extern struct pr_usrreqs tcp_usrreqs;
385extern u_long tcp_sendspace;
386extern u_long tcp_recvspace;
387
388#endif /* KERNEL */
389
390#endif /* _NETINET_TCP_VAR_H_ */
| 365void tcp_mss __P((struct tcpcb *, int));
366int tcp_mssopt __P((struct tcpcb *));
367void tcp_mtudisc __P((struct inpcb *, int));
368struct tcpcb *
369 tcp_newtcpcb __P((struct inpcb *));
370int tcp_output __P((struct tcpcb *));
371void tcp_quench __P((struct inpcb *, int));
372void tcp_respond __P((struct tcpcb *,
373 struct tcpiphdr *, struct mbuf *, tcp_seq, tcp_seq, int));
374struct rtentry *
375 tcp_rtlookup __P((struct inpcb *));
376void tcp_setpersist __P((struct tcpcb *));
377void tcp_slowtimo __P((void));
378struct tcpiphdr *
379 tcp_template __P((struct tcpcb *));
380struct tcpcb *
381 tcp_timers __P((struct tcpcb *, int));
382void tcp_trace __P((int, int, struct tcpcb *, struct tcpiphdr *, int));
383
384extern struct pr_usrreqs tcp_usrreqs;
385extern u_long tcp_sendspace;
386extern u_long tcp_recvspace;
387
388#endif /* KERNEL */
389
390#endif /* _NETINET_TCP_VAR_H_ */
|