tcp_var.h revision 1.20 1 /* $NetBSD: tcp_var.h,v 1.20 1997/07/23 21:26:54 thorpej Exp $ */
2
3 /*
4 * Copyright (c) 1982, 1986, 1993, 1994
5 * The Regents of the University of California. All rights reserved.
6 *
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
9 * are met:
10 * 1. Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 * 3. All advertising materials mentioning features or use of this software
16 * must display the following acknowledgement:
17 * This product includes software developed by the University of
18 * California, Berkeley and its contributors.
19 * 4. Neither the name of the University nor the names of its contributors
20 * may be used to endorse or promote products derived from this software
21 * without specific prior written permission.
22 *
23 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
24 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
25 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
26 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
27 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
28 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
29 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
30 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
31 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
32 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33 * SUCH DAMAGE.
34 *
35 * @(#)tcp_var.h 8.3 (Berkeley) 4/10/94
36 */
37
38 /*
39 * Kernel variables for tcp.
40 */
41
42 /*
43 * Tcp control block, one per tcp; fields:
44 */
45 struct tcpcb {
46 struct ipqehead segq; /* sequencing queue */
47 short t_state; /* state of this connection */
48 short t_timer[TCPT_NTIMERS]; /* tcp timers */
49 short t_rxtshift; /* log(2) of rexmt exp. backoff */
50 short t_rxtcur; /* current retransmit value */
51 short t_dupacks; /* consecutive dup acks recd */
52 u_short t_maxseg; /* maximum segment size */
53 char t_force; /* 1 if forcing out a byte */
54 u_short t_flags;
55 #define TF_ACKNOW 0x0001 /* ack peer immediately */
56 #define TF_DELACK 0x0002 /* ack, but try to delay it */
57 #define TF_NODELAY 0x0004 /* don't delay packets to coalesce */
58 #define TF_NOOPT 0x0008 /* don't use tcp options */
59 #define TF_SENTFIN 0x0010 /* have sent FIN */
60 #define TF_REQ_SCALE 0x0020 /* have/will request window scaling */
61 #define TF_RCVD_SCALE 0x0040 /* other side has requested scaling */
62 #define TF_REQ_TSTMP 0x0080 /* have/will request timestamps */
63 #define TF_RCVD_TSTMP 0x0100 /* a timestamp was received in SYN */
64 #define TF_SACK_PERMIT 0x0200 /* other side said I could SACK */
65
66 struct tcpiphdr *t_template; /* skeletal packet for transmit */
67 struct inpcb *t_inpcb; /* back pointer to internet pcb */
68 /*
69 * The following fields are used as in the protocol specification.
70 * See RFC783, Dec. 1981, page 21.
71 */
72 /* send sequence variables */
73 tcp_seq snd_una; /* send unacknowledged */
74 tcp_seq snd_nxt; /* send next */
75 tcp_seq snd_up; /* send urgent pointer */
76 tcp_seq snd_wl1; /* window update seg seq number */
77 tcp_seq snd_wl2; /* window update seg ack number */
78 tcp_seq iss; /* initial send sequence number */
79 u_long snd_wnd; /* send window */
80 /* receive sequence variables */
81 u_long rcv_wnd; /* receive window */
82 tcp_seq rcv_nxt; /* receive next */
83 tcp_seq rcv_up; /* receive urgent pointer */
84 tcp_seq irs; /* initial receive sequence number */
85 /*
86 * Additional variables for this implementation.
87 */
88 /* receive variables */
89 tcp_seq rcv_adv; /* advertised window */
90 /* retransmit variables */
91 tcp_seq snd_max; /* highest sequence number sent;
92 * used to recognize retransmits
93 */
94 /* congestion control (for slow start, source quench, retransmit after loss) */
95 u_long snd_cwnd; /* congestion-controlled window */
96 u_long snd_ssthresh; /* snd_cwnd size threshhold for
97 * for slow start exponential to
98 * linear switch
99 */
100 /*
101 * transmit timing stuff. See below for scale of srtt and rttvar.
102 * "Variance" is actually smoothed difference.
103 */
104 short t_idle; /* inactivity time */
105 short t_rtt; /* round trip time */
106 tcp_seq t_rtseq; /* sequence number being timed */
107 short t_srtt; /* smoothed round-trip time */
108 short t_rttvar; /* variance in round-trip time */
109 u_short t_rttmin; /* minimum rtt allowed */
110 u_long max_sndwnd; /* largest window peer has offered */
111
112 /* out-of-band data */
113 char t_oobflags; /* have some */
114 char t_iobc; /* input character */
115 #define TCPOOB_HAVEDATA 0x01
116 #define TCPOOB_HADDATA 0x02
117 short t_softerror; /* possible error not yet reported */
118
119 /* RFC 1323 variables */
120 u_char snd_scale; /* window scaling for send window */
121 u_char rcv_scale; /* window scaling for recv window */
122 u_char request_r_scale; /* pending window scaling */
123 u_char requested_s_scale;
124 u_int32_t ts_recent; /* timestamp echo data */
125 u_int32_t ts_recent_age; /* when last updated */
126 tcp_seq last_ack_sent;
127
128 /* TUBA stuff */
129 caddr_t t_tuba_pcb; /* next level down pcb for TCP over z */
130 };
131
132 /*
133 * Handy way of passing around TCP option info.
134 */
135 struct tcp_opt_info {
136 int ts_present;
137 u_int32_t ts_val;
138 u_int32_t ts_ecr;
139 u_int16_t maxseg;
140 };
141
142 /*
143 * This structure should not exceed 32 bytes.
144 * XXX On the Alpha, it's already 36-bytes, which rounds to 40.
145 * XXX Need to eliminate the pointer.
146 */
147 struct syn_cache {
148 struct syn_cache *sc_next;
149 u_int32_t sc_tstmp : 1,
150 sc_hash : 31;
151 struct in_addr sc_src;
152 struct in_addr sc_dst;
153 tcp_seq sc_irs;
154 tcp_seq sc_iss;
155 u_int16_t sc_sport;
156 u_int16_t sc_dport;
157 u_int16_t sc_peermaxseg;
158 u_int8_t sc_timer;
159 u_int8_t sc_request_r_scale : 4,
160 sc_requested_s_scale : 4;
161 };
162
163 struct syn_cache_head {
164 struct syn_cache *sch_first; /* First entry in the bucket */
165 struct syn_cache *sch_last; /* Last entry in the bucket */
166 struct syn_cache_head *sch_headq; /* The next non-empty bucket */
167 int16_t sch_timer_sum; /* Total time in this bucket */
168 u_int16_t sch_length; /* @ # elements in bucket */
169 };
170
171 #define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb)
172 #define sototcpcb(so) (intotcpcb(sotoinpcb(so)))
173
174 /*
175 * The smoothed round-trip time and estimated variance
176 * are stored as fixed point numbers scaled by the values below.
177 * For convenience, these scales are also used in smoothing the average
178 * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed).
179 * With these scales, srtt has 3 bits to the right of the binary point,
180 * and thus an "ALPHA" of 0.875. rttvar has 2 bits to the right of the
181 * binary point, and is smoothed with an ALPHA of 0.75.
182 */
183 #define TCP_RTT_SHIFT 3 /* shift for srtt; 3 bits frac. */
184 #define TCP_RTTVAR_SHIFT 2 /* multiplier for rttvar; 2 bits */
185
186 /*
187 * The initial retransmission should happen at rtt + 4 * rttvar.
188 * Because of the way we do the smoothing, srtt and rttvar
189 * will each average +1/2 tick of bias. When we compute
190 * the retransmit timer, we want 1/2 tick of rounding and
191 * 1 extra tick because of +-1/2 tick uncertainty in the
192 * firing of the timer. The bias will give us exactly the
193 * 1.5 tick we need. But, because the bias is
194 * statistical, we have to test that we don't drop below
195 * the minimum feasible timer (which is 2 ticks).
196 * This macro assumes that the value of 1<<TCP_RTTVAR_SHIFT
197 * is the same as the multiplier for rttvar.
198 */
199 #define TCP_REXMTVAL(tp) \
200 ((((tp)->t_srtt >> TCP_RTT_SHIFT) + (tp)->t_rttvar) >> 2)
201
202 /*
203 * TCP statistics.
204 * Many of these should be kept per connection,
205 * but that's inconvenient at the moment.
206 */
207 struct tcpstat {
208 u_long tcps_connattempt; /* connections initiated */
209 u_long tcps_accepts; /* connections accepted */
210 u_long tcps_connects; /* connections established */
211 u_long tcps_drops; /* connections dropped */
212 u_long tcps_conndrops; /* embryonic connections dropped */
213 u_long tcps_closed; /* conn. closed (includes drops) */
214 u_long tcps_segstimed; /* segs where we tried to get rtt */
215 u_long tcps_rttupdated; /* times we succeeded */
216 u_long tcps_delack; /* delayed acks sent */
217 u_long tcps_timeoutdrop; /* conn. dropped in rxmt timeout */
218 u_long tcps_rexmttimeo; /* retransmit timeouts */
219 u_long tcps_persisttimeo; /* persist timeouts */
220 u_long tcps_keeptimeo; /* keepalive timeouts */
221 u_long tcps_keepprobe; /* keepalive probes sent */
222 u_long tcps_keepdrops; /* connections dropped in keepalive */
223
224 u_long tcps_sndtotal; /* total packets sent */
225 u_long tcps_sndpack; /* data packets sent */
226 u_long tcps_sndbyte; /* data bytes sent */
227 u_long tcps_sndrexmitpack; /* data packets retransmitted */
228 u_long tcps_sndrexmitbyte; /* data bytes retransmitted */
229 u_long tcps_sndacks; /* ack-only packets sent */
230 u_long tcps_sndprobe; /* window probes sent */
231 u_long tcps_sndurg; /* packets sent with URG only */
232 u_long tcps_sndwinup; /* window update-only packets sent */
233 u_long tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */
234
235 u_long tcps_rcvtotal; /* total packets received */
236 u_long tcps_rcvpack; /* packets received in sequence */
237 u_long tcps_rcvbyte; /* bytes received in sequence */
238 u_long tcps_rcvbadsum; /* packets received with ccksum errs */
239 u_long tcps_rcvbadoff; /* packets received with bad offset */
240 u_long tcps_rcvmemdrop; /* packets dropped for lack of memory */
241 u_long tcps_rcvshort; /* packets received too short */
242 u_long tcps_rcvduppack; /* duplicate-only packets received */
243 u_long tcps_rcvdupbyte; /* duplicate-only bytes received */
244 u_long tcps_rcvpartduppack; /* packets with some duplicate data */
245 u_long tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */
246 u_long tcps_rcvoopack; /* out-of-order packets received */
247 u_long tcps_rcvoobyte; /* out-of-order bytes received */
248 u_long tcps_rcvpackafterwin; /* packets with data after window */
249 u_long tcps_rcvbyteafterwin; /* bytes rcvd after window */
250 u_long tcps_rcvafterclose; /* packets rcvd after "close" */
251 u_long tcps_rcvwinprobe; /* rcvd window probe packets */
252 u_long tcps_rcvdupack; /* rcvd duplicate acks */
253 u_long tcps_rcvacktoomuch; /* rcvd acks for unsent data */
254 u_long tcps_rcvackpack; /* rcvd ack packets */
255 u_long tcps_rcvackbyte; /* bytes acked by rcvd acks */
256 u_long tcps_rcvwinupd; /* rcvd window update packets */
257 u_long tcps_pawsdrop; /* segments dropped due to PAWS */
258 u_long tcps_predack; /* times hdr predict ok for acks */
259 u_long tcps_preddat; /* times hdr predict ok for data pkts */
260
261 u_long tcps_pcbhashmiss; /* input packets missing pcb hash */
262 u_long tcps_noport; /* no socket on port */
263 u_long tcps_badsyn; /* received ack for which we have
264 no SYN in compressed state */
265
266 /* These statistics deal with the SYN cache. */
267 u_long tcps_sc_added; /* # of entries added */
268 u_long tcps_sc_completed; /* # of connections completed */
269 u_long tcps_sc_timed_out; /* # of entries timed out */
270 u_long tcps_sc_overflowed; /* # dropped due to overflow */
271 u_long tcps_sc_reset; /* # dropped due to RST */
272 u_long tcps_sc_unreach; /* # dropped due to ICMP unreach */
273 u_long tcps_sc_bucketoverflow; /* # dropped due to bucket overflow */
274 u_long tcps_sc_aborted; /* # of entries aborted (no mem) */
275 u_long tcps_sc_dupesyn; /* # of duplicate SYNs received */
276 u_long tcps_sc_dropped; /* # of SYNs dropped (no route/mem) */
277 u_long tcps_sc_collisions; /* # of hash collisions */
278 };
279
280 /*
281 * Names for TCP sysctl objects.
282 */
283 /* enable/disable RFC1323 timestamps/scaling */
284 #define TCPCTL_RFC1323 1
285 #define TCPCTL_MAXID 2
286
287 #define TCPCTL_NAMES { \
288 { 0, 0 }, \
289 { "rfc1323", CTLTYPE_INT }, \
290 }
291
292 #ifdef _KERNEL
293 struct inpcbtable tcbtable; /* head of queue of active tcpcb's */
294 struct tcpstat tcpstat; /* tcp statistics */
295 u_int32_t tcp_now; /* for RFC 1323 timestamps */
296 extern int tcp_do_rfc1323; /* enabled/disabled? */
297
298 int tcp_syn_cache_size;
299 int tcp_syn_cache_timeo;
300 struct syn_cache_head tcp_syn_cache[], *tcp_syn_cache_first;
301 u_long syn_cache_count;
302
303 int tcp_attach __P((struct socket *));
304 void tcp_canceltimers __P((struct tcpcb *));
305 struct tcpcb *
306 tcp_close __P((struct tcpcb *));
307 void *tcp_ctlinput __P((int, struct sockaddr *, void *));
308 int tcp_ctloutput __P((int, struct socket *, int, int, struct mbuf **));
309 struct tcpcb *
310 tcp_disconnect __P((struct tcpcb *));
311 struct tcpcb *
312 tcp_drop __P((struct tcpcb *, int));
313 void tcp_dooptions __P((struct tcpcb *,
314 u_char *, int, struct tcpiphdr *, struct tcp_opt_info *));
315 void tcp_drain __P((void));
316 void tcp_fasttimo __P((void));
317 void tcp_init __P((void));
318 void tcp_input __P((struct mbuf *, ...));
319 int tcp_mss __P((struct tcpcb *, u_int));
320 struct tcpcb *
321 tcp_newtcpcb __P((struct inpcb *));
322 void tcp_notify __P((struct inpcb *, int));
323 int tcp_output __P((struct tcpcb *));
324 void tcp_pulloutofband __P((struct socket *,
325 struct tcpiphdr *, struct mbuf *));
326 void tcp_quench __P((struct inpcb *, int));
327 int tcp_reass __P((struct tcpcb *, struct tcpiphdr *, struct mbuf *));
328 int tcp_respond __P((struct tcpcb *,
329 struct tcpiphdr *, struct mbuf *, tcp_seq, tcp_seq, int));
330 void tcp_setpersist __P((struct tcpcb *));
331 void tcp_slowtimo __P((void));
332 struct tcpiphdr *
333 tcp_template __P((struct tcpcb *));
334 struct tcpcb *
335 tcp_timers __P((struct tcpcb *, int));
336 void tcp_trace __P((int, int, struct tcpcb *, struct tcpiphdr *, int));
337 struct tcpcb *
338 tcp_usrclosed __P((struct tcpcb *));
339 int tcp_sysctl __P((int *, u_int, void *, size_t *, void *, size_t));
340 int tcp_usrreq __P((struct socket *,
341 int, struct mbuf *, struct mbuf *, struct mbuf *, struct proc *));
342 void tcp_xmit_timer __P((struct tcpcb *, int));
343
344 int syn_cache_add __P((struct socket *, struct mbuf *, u_char *,
345 int, struct tcp_opt_info *));
346 void syn_cache_unreach __P((struct ip *, struct tcphdr *));
347 struct socket *
348 syn_cache_get __P((struct socket *so, struct mbuf *));
349 void syn_cache_insert __P((struct syn_cache *, struct syn_cache ***,
350 struct syn_cache_head **));
351 struct syn_cache *
352 syn_cache_lookup __P((struct tcpiphdr *, struct syn_cache ***,
353 struct syn_cache_head **));
354 void syn_cache_reset __P((struct tcpiphdr *));
355 int syn_cache_respond __P((struct syn_cache *, struct mbuf *,
356 register struct tcpiphdr *, long, u_long));
357 void syn_cache_timer __P((int));
358
359 #endif
360