|
|
1.1 root 1: /*
2: * Copyright (c) 2000 Apple Computer, Inc. All rights reserved.
3: *
4: * @APPLE_LICENSE_HEADER_START@
5: *
6: * The contents of this file constitute Original Code as defined in and
7: * are subject to the Apple Public Source License Version 1.1 (the
8: * "License"). You may not use this file except in compliance with the
9: * License. Please obtain a copy of the License at
10: * http://www.apple.com/publicsource and read it before using this file.
11: *
12: * This Original Code and all software distributed under the License are
13: * distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY KIND, EITHER
14: * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
15: * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
16: * FITNESS FOR A PARTICULAR PURPOSE OR NON-INFRINGEMENT. Please see the
17: * License for the specific language governing rights and limitations
18: * under the License.
19: *
20: * @APPLE_LICENSE_HEADER_END@
21: */
22: /*
23: * Copyright (c) 1982, 1986, 1993, 1994, 1995
24: * The Regents of the University of California. All rights reserved.
25: *
26: * Redistribution and use in source and binary forms, with or without
27: * modification, are permitted provided that the following conditions
28: * are met:
29: * 1. Redistributions of source code must retain the above copyright
30: * notice, this list of conditions and the following disclaimer.
31: * 2. Redistributions in binary form must reproduce the above copyright
32: * notice, this list of conditions and the following disclaimer in the
33: * documentation and/or other materials provided with the distribution.
34: * 3. All advertising materials mentioning features or use of this software
35: * must display the following acknowledgement:
36: * This product includes software developed by the University of
37: * California, Berkeley and its contributors.
38: * 4. Neither the name of the University nor the names of its contributors
39: * may be used to endorse or promote products derived from this software
40: * without specific prior written permission.
41: *
42: * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
43: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
44: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
45: * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
46: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
47: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
48: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
49: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
50: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
51: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
52: * SUCH DAMAGE.
53: *
54: * @(#)tcp_var.h 8.4 (Berkeley) 5/24/95
55: */
56:
57: #ifndef _NETINET_TCP_VAR_H_
58: #define _NETINET_TCP_VAR_H_
59: /*
60: * Kernel variables for tcp.
61: */
62:
63: /*
64: * Tcp control block, one per tcp; fields:
65: * Organized for 16 byte cacheline efficiency.
66: */
67: struct tcpcb {
68: struct mbuf *t_segq;
69: int t_dupacks; /* consecutive dup acks recd */
70: struct tcpiphdr *t_template; /* skeletal packet for transmit */
71:
72: int t_timer[TCPT_NTIMERS]; /* tcp timers */
73:
74: struct inpcb *t_inpcb; /* back pointer to internet pcb */
75: int t_state; /* state of this connection */
76: u_int t_flags;
77: #define TF_ACKNOW 0x00001 /* ack peer immediately */
78: #define TF_DELACK 0x00002 /* ack, but try to delay it */
79: #define TF_NODELAY 0x00004 /* don't delay packets to coalesce */
80: #define TF_NOOPT 0x00008 /* don't use tcp options */
81: #define TF_SENTFIN 0x00010 /* have sent FIN */
82: #define TF_REQ_SCALE 0x00020 /* have/will request window scaling */
83: #define TF_RCVD_SCALE 0x00040 /* other side has requested scaling */
84: #define TF_REQ_TSTMP 0x00080 /* have/will request timestamps */
85: #define TF_RCVD_TSTMP 0x00100 /* a timestamp was received in SYN */
86: #define TF_SACK_PERMIT 0x00200 /* other side said I could SACK */
87: #define TF_NEEDSYN 0x00400 /* send SYN (implicit state) */
88: #define TF_NEEDFIN 0x00800 /* send FIN (implicit state) */
89: #define TF_NOPUSH 0x01000 /* don't push */
90: #define TF_REQ_CC 0x02000 /* have/will request CC */
91: #define TF_RCVD_CC 0x04000 /* a CC was received in SYN */
92: #define TF_SENDCCNEW 0x08000 /* send CCnew instead of CC in SYN */
93: #define TF_MORETOCOME 0x10000 /* More data to be appended to sock */
94: int t_force; /* 1 if forcing out a byte */
95:
96: tcp_seq snd_una; /* send unacknowledged */
97: tcp_seq snd_max; /* highest sequence number sent;
98: * used to recognize retransmits
99: */
100: tcp_seq snd_nxt; /* send next */
101: tcp_seq snd_up; /* send urgent pointer */
102:
103: tcp_seq snd_wl1; /* window update seg seq number */
104: tcp_seq snd_wl2; /* window update seg ack number */
105: tcp_seq iss; /* initial send sequence number */
106: tcp_seq irs; /* initial receive sequence number */
107:
108: tcp_seq rcv_nxt; /* receive next */
109: tcp_seq rcv_adv; /* advertised window */
110: u_long rcv_wnd; /* receive window */
111: tcp_seq rcv_up; /* receive urgent pointer */
112:
113: u_long snd_wnd; /* send window */
114: u_long snd_cwnd; /* congestion-controlled window */
115: u_long snd_ssthresh; /* snd_cwnd size threshold for
116: * for slow start exponential to
117: * linear switch
118: */
119: u_int t_maxopd; /* mss plus options */
120:
121: u_int t_idle; /* inactivity time */
122: u_long t_duration; /* connection duration */
123: int t_rtt; /* round trip time */
124: tcp_seq t_rtseq; /* sequence number being timed */
125:
126: int t_rxtcur; /* current retransmit value */
127: u_int t_maxseg; /* maximum segment size */
128: int t_srtt; /* smoothed round-trip time */
129: int t_rttvar; /* variance in round-trip time */
130:
131: int t_rxtshift; /* log(2) of rexmt exp. backoff */
132: u_int t_rttmin; /* minimum rtt allowed */
133: u_long t_rttupdated; /* number of times rtt sampled */
134: u_long max_sndwnd; /* largest window peer has offered */
135:
136: int t_softerror; /* possible error not yet reported */
137: /* out-of-band data */
138: char t_oobflags; /* have some */
139: char t_iobc; /* input character */
140: #define TCPOOB_HAVEDATA 0x01
141: #define TCPOOB_HADDATA 0x02
142: /* RFC 1323 variables */
143: u_char snd_scale; /* window scaling for send window */
144: u_char rcv_scale; /* window scaling for recv window */
145: u_char request_r_scale; /* pending window scaling */
146: u_char requested_s_scale;
147: u_long ts_recent; /* timestamp echo data */
148:
149: u_long ts_recent_age; /* when last updated */
150: tcp_seq last_ack_sent;
151: /* RFC 1644 variables */
152: tcp_cc cc_send; /* send connection count */
153: tcp_cc cc_recv; /* receive connection count */
154: };
155:
156: /*
157: * Structure to hold TCP options that are only used during segment
158: * processing (in tcp_input), but not held in the tcpcb.
159: * It's basically used to reduce the number of parameters
160: * to tcp_dooptions.
161: */
162: struct tcpopt {
163: u_long to_flag; /* which options are present */
164: #define TOF_TS 0x0001 /* timestamp */
165: #define TOF_CC 0x0002 /* CC and CCnew are exclusive */
166: #define TOF_CCNEW 0x0004
167: #define TOF_CCECHO 0x0008
168: u_long to_tsval;
169: u_long to_tsecr;
170: tcp_cc to_cc; /* holds CC or CCnew */
171: tcp_cc to_ccecho;
172: };
173:
174: /*
175: * The TAO cache entry which is stored in the protocol family specific
176: * portion of the route metrics.
177: */
178: struct rmxp_tao {
179: tcp_cc tao_cc; /* latest CC in valid SYN */
180: tcp_cc tao_ccsent; /* latest CC sent to peer */
181: u_short tao_mssopt; /* peer's cached MSS */
182: #ifdef notyet
183: u_short tao_flags; /* cache status flags */
184: #define TAOF_DONT 0x0001 /* peer doesn't understand rfc1644 */
185: #define TAOF_OK 0x0002 /* peer does understand rfc1644 */
186: #define TAOF_UNDEF 0 /* we don't know yet */
187: #endif /* notyet */
188: };
189: #define rmx_taop(r) ((struct rmxp_tao *)(r).rmx_filler)
190:
191: #define intotcpcb(ip) ((struct tcpcb *)(ip)->inp_ppcb)
192: #define sototcpcb(so) (intotcpcb(sotoinpcb(so)))
193:
194: /*
195: * The smoothed round-trip time and estimated variance
196: * are stored as fixed point numbers scaled by the values below.
197: * For convenience, these scales are also used in smoothing the average
198: * (smoothed = (1/scale)sample + ((scale-1)/scale)smoothed).
199: * With these scales, srtt has 3 bits to the right of the binary point,
200: * and thus an "ALPHA" of 0.875. rttvar has 2 bits to the right of the
201: * binary point, and is smoothed with an ALPHA of 0.75.
202: */
203: #define TCP_RTT_SCALE 32 /* multiplier for srtt; 3 bits frac. */
204: #define TCP_RTT_SHIFT 5 /* shift for srtt; 3 bits frac. */
205: #define TCP_RTTVAR_SCALE 16 /* multiplier for rttvar; 2 bits */
206: #define TCP_RTTVAR_SHIFT 4 /* shift for rttvar; 2 bits */
207: #define TCP_DELTA_SHIFT 2 /* see tcp_input.c */
208:
209: /*
210: * The initial retransmission should happen at rtt + 4 * rttvar.
211: * Because of the way we do the smoothing, srtt and rttvar
212: * will each average +1/2 tick of bias. When we compute
213: * the retransmit timer, we want 1/2 tick of rounding and
214: * 1 extra tick because of +-1/2 tick uncertainty in the
215: * firing of the timer. The bias will give us exactly the
216: * 1.5 tick we need. But, because the bias is
217: * statistical, we have to test that we don't drop below
218: * the minimum feasible timer (which is 2 ticks).
219: * This version of the macro adapted from a paper by Lawrence
220: * Brakmo and Larry Peterson which outlines a problem caused
221: * by insufficient precision in the original implementation,
222: * which results in inappropriately large RTO values for very
223: * fast networks.
224: */
225: #define TCP_REXMTVAL(tp) \
226: max((tp)->t_rttmin, (((tp)->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT)) \
227: + (tp)->t_rttvar) >> TCP_DELTA_SHIFT)
228:
229: /*
230: * TCP statistics.
231: * Many of these should be kept per connection,
232: * but that's inconvenient at the moment.
233: */
234: struct tcpstat {
235: u_long tcps_connattempt; /* connections initiated */
236: u_long tcps_accepts; /* connections accepted */
237: u_long tcps_connects; /* connections established */
238: u_long tcps_drops; /* connections dropped */
239: u_long tcps_conndrops; /* embryonic connections dropped */
240: u_long tcps_closed; /* conn. closed (includes drops) */
241: u_long tcps_segstimed; /* segs where we tried to get rtt */
242: u_long tcps_rttupdated; /* times we succeeded */
243: u_long tcps_delack; /* delayed acks sent */
244: u_long tcps_timeoutdrop; /* conn. dropped in rxmt timeout */
245: u_long tcps_rexmttimeo; /* retransmit timeouts */
246: u_long tcps_persisttimeo; /* persist timeouts */
247: u_long tcps_keeptimeo; /* keepalive timeouts */
248: u_long tcps_keepprobe; /* keepalive probes sent */
249: u_long tcps_keepdrops; /* connections dropped in keepalive */
250:
251: u_long tcps_sndtotal; /* total packets sent */
252: u_long tcps_sndpack; /* data packets sent */
253: u_long tcps_sndbyte; /* data bytes sent */
254: u_long tcps_sndrexmitpack; /* data packets retransmitted */
255: u_long tcps_sndrexmitbyte; /* data bytes retransmitted */
256: u_long tcps_sndacks; /* ack-only packets sent */
257: u_long tcps_sndprobe; /* window probes sent */
258: u_long tcps_sndurg; /* packets sent with URG only */
259: u_long tcps_sndwinup; /* window update-only packets sent */
260: u_long tcps_sndctrl; /* control (SYN|FIN|RST) packets sent */
261:
262: u_long tcps_rcvtotal; /* total packets received */
263: u_long tcps_rcvpack; /* packets received in sequence */
264: u_long tcps_rcvbyte; /* bytes received in sequence */
265: u_long tcps_rcvbadsum; /* packets received with ccksum errs */
266: u_long tcps_rcvbadoff; /* packets received with bad offset */
267: u_long tcps_rcvshort; /* packets received too short */
268: u_long tcps_rcvduppack; /* duplicate-only packets received */
269: u_long tcps_rcvdupbyte; /* duplicate-only bytes received */
270: u_long tcps_rcvpartduppack; /* packets with some duplicate data */
271: u_long tcps_rcvpartdupbyte; /* dup. bytes in part-dup. packets */
272: u_long tcps_rcvoopack; /* out-of-order packets received */
273: u_long tcps_rcvoobyte; /* out-of-order bytes received */
274: u_long tcps_rcvpackafterwin; /* packets with data after window */
275: u_long tcps_rcvbyteafterwin; /* bytes rcvd after window */
276: u_long tcps_rcvafterclose; /* packets rcvd after "close" */
277: u_long tcps_rcvwinprobe; /* rcvd window probe packets */
278: u_long tcps_rcvdupack; /* rcvd duplicate acks */
279: u_long tcps_rcvacktoomuch; /* rcvd acks for unsent data */
280: u_long tcps_rcvackpack; /* rcvd ack packets */
281: u_long tcps_rcvackbyte; /* bytes acked by rcvd acks */
282: u_long tcps_rcvwinupd; /* rcvd window update packets */
283: u_long tcps_pawsdrop; /* segments dropped due to PAWS */
284: u_long tcps_predack; /* times hdr predict ok for acks */
285: u_long tcps_preddat; /* times hdr predict ok for data pkts */
286: u_long tcps_pcbcachemiss;
287: u_long tcps_cachedrtt; /* times cached RTT in route updated */
288: u_long tcps_cachedrttvar; /* times cached rttvar updated */
289: u_long tcps_cachedssthresh; /* times cached ssthresh updated */
290: u_long tcps_usedrtt; /* times RTT initialized from route */
291: u_long tcps_usedrttvar; /* times RTTVAR initialized from rt */
292: u_long tcps_usedssthresh; /* times ssthresh initialized from rt*/
293: u_long tcps_persistdrop; /* timeout in persist state */
294: u_long tcps_badsyn; /* bogus SYN, e.g. premature ACK */
295: u_long tcps_mturesent; /* resends due to MTU discovery */
296: u_long tcps_listendrop; /* listen queue overflows */
297: };
298:
299: /*
300: * TCB structure exported to user-land via sysctl(3).
301: * Evil hack: declare only if in_pcb.h and sys/socketvar.h have been
302: * included. Not all of our clients do.
303: */
304: #if defined(_NETINET_IN_PCB_H_) && defined(_SYS_SOCKETVAR_H_)
305: struct xtcpcb {
306: size_t xt_len;
307: struct inpcb xt_inp;
308: struct tcpcb xt_tp;
309: struct xsocket xt_socket;
310: u_quad_t xt_alignment_hack;
311: };
312: #endif
313:
314: /*
315: * Names for TCP sysctl objects
316: */
317: #define TCPCTL_DO_RFC1323 1 /* use RFC-1323 extensions */
318: #define TCPCTL_DO_RFC1644 2 /* use RFC-1644 extensions */
319: #define TCPCTL_MSSDFLT 3 /* MSS default */
320: #define TCPCTL_STATS 4 /* statistics (read-only) */
321: #define TCPCTL_RTTDFLT 5 /* default RTT estimate */
322: #define TCPCTL_KEEPIDLE 6 /* keepalive idle timer */
323: #define TCPCTL_KEEPINTVL 7 /* interval to send keepalives */
324: #define TCPCTL_SENDSPACE 8 /* send buffer space */
325: #define TCPCTL_RECVSPACE 9 /* receive buffer space */
326: #define TCPCTL_KEEPINIT 10 /* receive buffer space */
327: #define TCPCTL_PCBLIST 11 /* list of all outstanding PCBs */
328: #define TCPCTL_MAXID 12
329:
330: #define TCPCTL_NAMES { \
331: { 0, 0 }, \
332: { "rfc1323", CTLTYPE_INT }, \
333: { "rfc1644", CTLTYPE_INT }, \
334: { "mssdflt", CTLTYPE_INT }, \
335: { "stats", CTLTYPE_STRUCT }, \
336: { "rttdflt", CTLTYPE_INT }, \
337: { "keepidle", CTLTYPE_INT }, \
338: { "keepintvl", CTLTYPE_INT }, \
339: { "sendspace", CTLTYPE_INT }, \
340: { "recvspace", CTLTYPE_INT }, \
341: { "keepinit", CTLTYPE_INT }, \
342: { "pcblist", CTLTYPE_STRUCT }, \
343: }
344:
345: #ifdef KERNEL
346: #ifdef SYSCTL_DECL
347: SYSCTL_DECL(_net_inet_tcp);
348: #endif
349:
350: extern struct inpcbhead tcb; /* head of queue of active tcpcb's */
351: extern struct inpcbinfo tcbinfo;
352: extern struct tcpstat tcpstat; /* tcp statistics */
353: extern int tcp_mssdflt; /* XXX */
354: extern u_long tcp_now; /* for RFC 1323 timestamps */
355: extern int tcp_delack_enabled;
356:
357: void tcp_canceltimers __P((struct tcpcb *));
358: struct tcpcb *
359: tcp_close __P((struct tcpcb *));
360: void tcp_ctlinput __P((int, struct sockaddr *, void *));
361: int tcp_ctloutput __P((struct socket *, struct sockopt *));
362: struct tcpcb *
363: tcp_drop __P((struct tcpcb *, int));
364: void tcp_drain __P((void));
365: void tcp_fasttimo __P((void));
366: struct rmxp_tao *
367: tcp_gettaocache __P((struct inpcb *));
368: void tcp_init __P((void));
369: void tcp_input __P((struct mbuf *, int));
370: void tcp_mss __P((struct tcpcb *, int));
371: int tcp_mssopt __P((struct tcpcb *));
372: void tcp_mtudisc __P((struct inpcb *, int));
373: struct tcpcb *
374: tcp_newtcpcb __P((struct inpcb *));
375: int tcp_output __P((struct tcpcb *));
376: void tcp_quench __P((struct inpcb *, int));
377: void tcp_respond __P((struct tcpcb *,
378: struct tcpiphdr *, struct mbuf *, tcp_seq, tcp_seq, int));
379: struct rtentry *
380: tcp_rtlookup __P((struct inpcb *));
381: void tcp_setpersist __P((struct tcpcb *));
382: void tcp_slowtimo __P((void));
383: struct tcpiphdr *
384: tcp_template __P((struct tcpcb *));
385: struct tcpcb *
386: tcp_timers __P((struct tcpcb *, int));
387: void tcp_trace __P((int, int, struct tcpcb *, struct tcpiphdr *, int));
388:
389: extern struct pr_usrreqs tcp_usrreqs;
390: extern u_long tcp_sendspace;
391: extern u_long tcp_recvspace;
392:
393: #endif /* KERNEL */
394:
395: #endif /* _NETINET_TCP_VAR_H_ */
This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.