Annotation of qemu/slirp/tcp_input.c, revision 1.1.1.5

1.1       root        1: /*
                      2:  * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1994
                      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.
1.1.1.4   root       13:  * 3. Neither the name of the University nor the names of its contributors
1.1       root       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_input.c 8.5 (Berkeley) 4/10/94
                     30:  * tcp_input.c,v 1.10 1994/10/13 18:36:32 wollman Exp
                     31:  */
                     32: 
                     33: /*
                     34:  * Changes and additions relating to SLiRP
                     35:  * Copyright (c) 1995 Danny Gasparovski.
1.1.1.3   root       36:  *
                     37:  * Please read the file COPYRIGHT for the
1.1       root       38:  * terms and conditions of the copyright.
                     39:  */
                     40: 
                     41: #include <slirp.h>
                     42: #include "ip_icmp.h"
                     43: 
1.1.1.3   root       44: #define        TCPREXMTTHRESH 3
1.1       root       45: 
                     46: #define TCP_PAWS_IDLE  (24 * 24 * 60 * 60 * PR_SLOWHZ)
                     47: 
                     48: /* for modulo comparisons of timestamps */
                     49: #define TSTMP_LT(a,b)  ((int)((a)-(b)) < 0)
                     50: #define TSTMP_GEQ(a,b) ((int)((a)-(b)) >= 0)
                     51: 
                     52: /*
                     53:  * Insert segment ti into reassembly queue of tcp with
                     54:  * control block tp.  Return TH_FIN if reassembly now includes
                     55:  * a segment with FIN.  The macro form does the common case inline
                     56:  * (segment is the next to be received on an established connection,
                     57:  * and the queue is empty), avoiding linkage into and removal
                     58:  * from the queue and repetition of various conversions.
                     59:  * Set DELACK for segments received in order, but ack immediately
                     60:  * when segments are out of order (so fast retransmit can work).
                     61:  */
                     62: #ifdef TCP_ACK_HACK
                     63: #define TCP_REASS(tp, ti, m, so, flags) {\
                     64:        if ((ti)->ti_seq == (tp)->rcv_nxt && \
1.1.1.4   root       65:            tcpfrag_list_empty(tp) && \
1.1       root       66:            (tp)->t_state == TCPS_ESTABLISHED) {\
                     67:                if (ti->ti_flags & TH_PUSH) \
                     68:                        tp->t_flags |= TF_ACKNOW; \
                     69:                else \
                     70:                        tp->t_flags |= TF_DELACK; \
                     71:                (tp)->rcv_nxt += (ti)->ti_len; \
                     72:                flags = (ti)->ti_flags & TH_FIN; \
                     73:                if (so->so_emu) { \
                     74:                       if (tcp_emu((so),(m))) sbappend((so), (m)); \
                     75:               } else \
                     76:                       sbappend((so), (m)); \
                     77:        } else {\
                     78:                (flags) = tcp_reass((tp), (ti), (m)); \
                     79:                tp->t_flags |= TF_ACKNOW; \
                     80:        } \
                     81: }
                     82: #else
                     83: #define        TCP_REASS(tp, ti, m, so, flags) { \
                     84:        if ((ti)->ti_seq == (tp)->rcv_nxt && \
1.1.1.4   root       85:         tcpfrag_list_empty(tp) && \
1.1       root       86:            (tp)->t_state == TCPS_ESTABLISHED) { \
                     87:                tp->t_flags |= TF_DELACK; \
                     88:                (tp)->rcv_nxt += (ti)->ti_len; \
                     89:                flags = (ti)->ti_flags & TH_FIN; \
                     90:                if (so->so_emu) { \
                     91:                        if (tcp_emu((so),(m))) sbappend(so, (m)); \
                     92:                } else \
                     93:                        sbappend((so), (m)); \
                     94:        } else { \
                     95:                (flags) = tcp_reass((tp), (ti), (m)); \
                     96:                tp->t_flags |= TF_ACKNOW; \
                     97:        } \
                     98: }
                     99: #endif
1.1.1.3   root      100: static void tcp_dooptions(struct tcpcb *tp, u_char *cp, int cnt,
                    101:                           struct tcpiphdr *ti);
                    102: static void tcp_xmit_timer(register struct tcpcb *tp, int rtt);
                    103: 
                    104: static int
                    105: tcp_reass(register struct tcpcb *tp, register struct tcpiphdr *ti,
                    106:           struct mbuf *m)
1.1       root      107: {
                    108:        register struct tcpiphdr *q;
                    109:        struct socket *so = tp->t_socket;
                    110:        int flags;
1.1.1.3   root      111: 
1.1       root      112:        /*
1.1.1.5 ! root      113:         * Call with ti==NULL after become established to
1.1       root      114:         * force pre-ESTABLISHED data up to user socket.
                    115:         */
1.1.1.5 ! root      116:         if (ti == NULL)
1.1       root      117:                goto present;
                    118: 
                    119:        /*
                    120:         * Find a segment which begins after this one does.
                    121:         */
1.1.1.4   root      122:        for (q = tcpfrag_list_first(tp); !tcpfrag_list_end(q, tp);
                    123:             q = tcpiphdr_next(q))
1.1       root      124:                if (SEQ_GT(q->ti_seq, ti->ti_seq))
                    125:                        break;
                    126: 
                    127:        /*
                    128:         * If there is a preceding segment, it may provide some of
                    129:         * our data already.  If so, drop the data from the incoming
                    130:         * segment.  If it provides all of our data, drop us.
                    131:         */
1.1.1.4   root      132:        if (!tcpfrag_list_end(tcpiphdr_prev(q), tp)) {
1.1       root      133:                register int i;
1.1.1.4   root      134:                q = tcpiphdr_prev(q);
1.1       root      135:                /* conversion to int (in i) handles seq wraparound */
                    136:                i = q->ti_seq + q->ti_len - ti->ti_seq;
                    137:                if (i > 0) {
                    138:                        if (i >= ti->ti_len) {
                    139:                                m_freem(m);
                    140:                                /*
                    141:                                 * Try to present any queued data
                    142:                                 * at the left window edge to the user.
                    143:                                 * This is needed after the 3-WHS
                    144:                                 * completes.
                    145:                                 */
                    146:                                goto present;   /* ??? */
                    147:                        }
                    148:                        m_adj(m, i);
                    149:                        ti->ti_len -= i;
                    150:                        ti->ti_seq += i;
                    151:                }
1.1.1.4   root      152:                q = tcpiphdr_next(q);
1.1       root      153:        }
1.1.1.4   root      154:        ti->ti_mbuf = m;
1.1       root      155: 
                    156:        /*
                    157:         * While we overlap succeeding segments trim them or,
                    158:         * if they are completely covered, dequeue them.
                    159:         */
1.1.1.4   root      160:        while (!tcpfrag_list_end(q, tp)) {
1.1       root      161:                register int i = (ti->ti_seq + ti->ti_len) - q->ti_seq;
                    162:                if (i <= 0)
                    163:                        break;
                    164:                if (i < q->ti_len) {
                    165:                        q->ti_seq += i;
                    166:                        q->ti_len -= i;
1.1.1.4   root      167:                        m_adj(q->ti_mbuf, i);
1.1       root      168:                        break;
                    169:                }
1.1.1.4   root      170:                q = tcpiphdr_next(q);
                    171:                m = tcpiphdr_prev(q)->ti_mbuf;
                    172:                remque(tcpiphdr2qlink(tcpiphdr_prev(q)));
1.1       root      173:                m_freem(m);
                    174:        }
                    175: 
                    176:        /*
                    177:         * Stick new segment in its place.
                    178:         */
1.1.1.4   root      179:        insque(tcpiphdr2qlink(ti), tcpiphdr2qlink(tcpiphdr_prev(q)));
1.1       root      180: 
                    181: present:
                    182:        /*
                    183:         * Present data to user, advancing rcv_nxt through
                    184:         * completed sequence space.
                    185:         */
                    186:        if (!TCPS_HAVEESTABLISHED(tp->t_state))
                    187:                return (0);
1.1.1.4   root      188:        ti = tcpfrag_list_first(tp);
                    189:        if (tcpfrag_list_end(ti, tp) || ti->ti_seq != tp->rcv_nxt)
1.1       root      190:                return (0);
                    191:        if (tp->t_state == TCPS_SYN_RECEIVED && ti->ti_len)
                    192:                return (0);
                    193:        do {
                    194:                tp->rcv_nxt += ti->ti_len;
                    195:                flags = ti->ti_flags & TH_FIN;
1.1.1.4   root      196:                remque(tcpiphdr2qlink(ti));
                    197:                m = ti->ti_mbuf;
                    198:                ti = tcpiphdr_next(ti);
1.1       root      199:                if (so->so_state & SS_FCANTSENDMORE)
                    200:                        m_freem(m);
                    201:                else {
                    202:                        if (so->so_emu) {
                    203:                                if (tcp_emu(so,m)) sbappend(so, m);
                    204:                        } else
                    205:                                sbappend(so, m);
                    206:                }
                    207:        } while (ti != (struct tcpiphdr *)tp && ti->ti_seq == tp->rcv_nxt);
                    208:        return (flags);
                    209: }
                    210: 
                    211: /*
                    212:  * TCP input routine, follows pages 65-76 of the
                    213:  * protocol specification dated September, 1981 very closely.
                    214:  */
                    215: void
1.1.1.5 ! root      216: tcp_input(struct mbuf *m, int iphlen, struct socket *inso)
1.1       root      217: {
                    218:        struct ip save_ip, *ip;
                    219:        register struct tcpiphdr *ti;
                    220:        caddr_t optp = NULL;
                    221:        int optlen = 0;
                    222:        int len, tlen, off;
1.1.1.5 ! root      223:         register struct tcpcb *tp = NULL;
1.1       root      224:        register int tiflags;
1.1.1.5 ! root      225:         struct socket *so = NULL;
1.1       root      226:        int todrop, acked, ourfinisacked, needoutput = 0;
                    227:        int iss = 0;
                    228:        u_long tiwin;
                    229:        int ret;
1.1.1.4   root      230:     struct ex_list *ex_ptr;
1.1.1.5 ! root      231:     Slirp *slirp;
1.1       root      232: 
                    233:        DEBUG_CALL("tcp_input");
1.1.1.3   root      234:        DEBUG_ARGS((dfd," m = %8lx  iphlen = %2d  inso = %lx\n",
1.1       root      235:                    (long )m, iphlen, (long )inso ));
1.1.1.3   root      236: 
1.1       root      237:        /*
                    238:         * If called with m == 0, then we're continuing the connect
                    239:         */
                    240:        if (m == NULL) {
                    241:                so = inso;
1.1.1.5 ! root      242:                slirp = so->slirp;
1.1.1.3   root      243: 
1.1       root      244:                /* Re-set a few variables */
                    245:                tp = sototcpcb(so);
                    246:                m = so->so_m;
1.1.1.5 ! root      247:                 so->so_m = NULL;
1.1       root      248:                ti = so->so_ti;
                    249:                tiwin = ti->ti_win;
                    250:                tiflags = ti->ti_flags;
1.1.1.3   root      251: 
1.1       root      252:                goto cont_conn;
                    253:        }
1.1.1.5 ! root      254:        slirp = m->slirp;
1.1.1.3   root      255: 
1.1       root      256:        /*
                    257:         * Get IP and TCP header together in first mbuf.
                    258:         * Note: IP leaves IP header in first mbuf.
                    259:         */
                    260:        ti = mtod(m, struct tcpiphdr *);
                    261:        if (iphlen > sizeof(struct ip )) {
                    262:          ip_stripoptions(m, (struct mbuf *)0);
                    263:          iphlen=sizeof(struct ip );
                    264:        }
                    265:        /* XXX Check if too short */
1.1.1.3   root      266: 
1.1       root      267: 
                    268:        /*
                    269:         * Save a copy of the IP header in case we want restore it
                    270:         * for sending an ICMP error message in response.
                    271:         */
                    272:        ip=mtod(m, struct ip *);
1.1.1.3   root      273:        save_ip = *ip;
1.1       root      274:        save_ip.ip_len+= iphlen;
                    275: 
                    276:        /*
                    277:         * Checksum extended TCP header and data.
                    278:         */
                    279:        tlen = ((struct ip *)ti)->ip_len;
1.1.1.5 ! root      280:         tcpiphdr2qlink(ti)->next = tcpiphdr2qlink(ti)->prev = NULL;
        !           281:         memset(&ti->ti_i.ih_mbuf, 0 , sizeof(struct mbuf_ptr));
1.1       root      282:        ti->ti_x1 = 0;
                    283:        ti->ti_len = htons((u_int16_t)tlen);
                    284:        len = sizeof(struct ip ) + tlen;
                    285:        if(cksum(m, len)) {
                    286:          goto drop;
                    287:        }
                    288: 
                    289:        /*
                    290:         * Check that TCP offset makes sense,
                    291:         * pull out TCP options and adjust length.              XXX
                    292:         */
                    293:        off = ti->ti_off << 2;
                    294:        if (off < sizeof (struct tcphdr) || off > tlen) {
                    295:          goto drop;
                    296:        }
                    297:        tlen -= off;
                    298:        ti->ti_len = tlen;
                    299:        if (off > sizeof (struct tcphdr)) {
                    300:          optlen = off - sizeof (struct tcphdr);
                    301:          optp = mtod(m, caddr_t) + sizeof (struct tcpiphdr);
                    302:        }
                    303:        tiflags = ti->ti_flags;
1.1.1.3   root      304: 
1.1       root      305:        /*
                    306:         * Convert TCP protocol specific fields to host format.
                    307:         */
                    308:        NTOHL(ti->ti_seq);
                    309:        NTOHL(ti->ti_ack);
                    310:        NTOHS(ti->ti_win);
                    311:        NTOHS(ti->ti_urp);
                    312: 
                    313:        /*
                    314:         * Drop TCP, IP headers and TCP options.
                    315:         */
                    316:        m->m_data += sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr);
                    317:        m->m_len  -= sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr);
1.1.1.3   root      318: 
1.1.1.5 ! root      319:     if (slirp->restricted) {
        !           320:         for (ex_ptr = slirp->exec_list; ex_ptr; ex_ptr = ex_ptr->ex_next) {
1.1.1.4   root      321:             if (ex_ptr->ex_fport == ti->ti_dport &&
1.1.1.5 ! root      322:                 ti->ti_dst.s_addr == ex_ptr->ex_addr.s_addr) {
1.1.1.4   root      323:                 break;
1.1.1.5 ! root      324:             }
        !           325:         }
1.1.1.4   root      326:         if (!ex_ptr)
                    327:             goto drop;
                    328:     }
1.1       root      329:        /*
                    330:         * Locate pcb for segment.
                    331:         */
                    332: findso:
1.1.1.5 ! root      333:        so = slirp->tcp_last_so;
1.1       root      334:        if (so->so_fport != ti->ti_dport ||
                    335:            so->so_lport != ti->ti_sport ||
                    336:            so->so_laddr.s_addr != ti->ti_src.s_addr ||
                    337:            so->so_faddr.s_addr != ti->ti_dst.s_addr) {
1.1.1.5 ! root      338:                so = solookup(&slirp->tcb, ti->ti_src, ti->ti_sport,
1.1       root      339:                               ti->ti_dst, ti->ti_dport);
                    340:                if (so)
1.1.1.5 ! root      341:                        slirp->tcp_last_so = so;
1.1       root      342:        }
                    343: 
                    344:        /*
                    345:         * If the state is CLOSED (i.e., TCB does not exist) then
                    346:         * all data in the incoming segment is discarded.
                    347:         * If the TCB exists but is in CLOSED state, it is embryonic,
                    348:         * but should either do a listen or a connect soon.
                    349:         *
                    350:         * state == CLOSED means we've done socreate() but haven't
1.1.1.3   root      351:         * attached it to a protocol yet...
                    352:         *
1.1       root      353:         * XXX If a TCB does not exist, and the TH_SYN flag is
                    354:         * the only flag set, then create a session, mark it
                    355:         * as if it was LISTENING, and continue...
                    356:         */
1.1.1.5 ! root      357:         if (so == NULL) {
1.1       root      358:          if ((tiflags & (TH_SYN|TH_FIN|TH_RST|TH_URG|TH_ACK)) != TH_SYN)
                    359:            goto dropwithreset;
1.1.1.3   root      360: 
1.1.1.5 ! root      361:          if ((so = socreate(slirp)) == NULL)
1.1       root      362:            goto dropwithreset;
                    363:          if (tcp_attach(so) < 0) {
                    364:            free(so); /* Not sofree (if it failed, it's not insqued) */
                    365:            goto dropwithreset;
                    366:          }
1.1.1.3   root      367: 
                    368:          sbreserve(&so->so_snd, TCP_SNDSPACE);
                    369:          sbreserve(&so->so_rcv, TCP_RCVSPACE);
                    370: 
1.1       root      371:          so->so_laddr = ti->ti_src;
                    372:          so->so_lport = ti->ti_sport;
                    373:          so->so_faddr = ti->ti_dst;
                    374:          so->so_fport = ti->ti_dport;
1.1.1.3   root      375: 
1.1       root      376:          if ((so->so_iptos = tcp_tos(so)) == 0)
                    377:            so->so_iptos = ((struct ip *)ti)->ip_tos;
1.1.1.3   root      378: 
1.1       root      379:          tp = sototcpcb(so);
                    380:          tp->t_state = TCPS_LISTEN;
                    381:        }
1.1.1.3   root      382: 
1.1       root      383:         /*
                    384:          * If this is a still-connecting socket, this probably
                    385:          * a retransmit of the SYN.  Whether it's a retransmit SYN
                    386:         * or something else, we nuke it.
                    387:          */
                    388:         if (so->so_state & SS_ISFCONNECTING)
                    389:                 goto drop;
                    390: 
                    391:        tp = sototcpcb(so);
1.1.1.3   root      392: 
1.1       root      393:        /* XXX Should never fail */
1.1.1.5 ! root      394:         if (tp == NULL)
1.1       root      395:                goto dropwithreset;
                    396:        if (tp->t_state == TCPS_CLOSED)
                    397:                goto drop;
1.1.1.3   root      398: 
1.1.1.5 ! root      399:        tiwin = ti->ti_win;
1.1       root      400: 
                    401:        /*
                    402:         * Segment received on connection.
                    403:         * Reset idle time and keep-alive timer.
                    404:         */
                    405:        tp->t_idle = 0;
1.1.1.3   root      406:        if (SO_OPTIONS)
                    407:           tp->t_timer[TCPT_KEEP] = TCPTV_KEEPINTVL;
1.1       root      408:        else
1.1.1.3   root      409:           tp->t_timer[TCPT_KEEP] = TCPTV_KEEP_IDLE;
1.1       root      410: 
                    411:        /*
                    412:         * Process options if not in LISTEN state,
                    413:         * else do it below (after getting remote address).
                    414:         */
                    415:        if (optp && tp->t_state != TCPS_LISTEN)
1.1.1.3   root      416:                tcp_dooptions(tp, (u_char *)optp, optlen, ti);
1.1       root      417: 
1.1.1.3   root      418:        /*
1.1       root      419:         * Header prediction: check for the two common cases
                    420:         * of a uni-directional data xfer.  If the packet has
                    421:         * no control flags, is in-sequence, the window didn't
                    422:         * change and we're not retransmitting, it's a
                    423:         * candidate.  If the length is zero and the ack moved
                    424:         * forward, we're the sender side of the xfer.  Just
                    425:         * free the data acked & wake any higher level process
                    426:         * that was blocked waiting for space.  If the length
                    427:         * is non-zero and the ack didn't move, we're the
                    428:         * receiver side.  If we're getting packets in-order
                    429:         * (the reassembly queue is empty), add the data to
                    430:         * the socket buffer and note that we need a delayed ack.
                    431:         *
                    432:         * XXX Some of these tests are not needed
                    433:         * eg: the tiwin == tp->snd_wnd prevents many more
                    434:         * predictions.. with no *real* advantage..
                    435:         */
                    436:        if (tp->t_state == TCPS_ESTABLISHED &&
                    437:            (tiflags & (TH_SYN|TH_FIN|TH_RST|TH_URG|TH_ACK)) == TH_ACK &&
                    438:            ti->ti_seq == tp->rcv_nxt &&
                    439:            tiwin && tiwin == tp->snd_wnd &&
                    440:            tp->snd_nxt == tp->snd_max) {
                    441:                if (ti->ti_len == 0) {
                    442:                        if (SEQ_GT(ti->ti_ack, tp->snd_una) &&
                    443:                            SEQ_LEQ(ti->ti_ack, tp->snd_max) &&
                    444:                            tp->snd_cwnd >= tp->snd_wnd) {
                    445:                                /*
                    446:                                 * this is a pure ack for outstanding data.
                    447:                                 */
1.1.1.5 ! root      448:                                if (tp->t_rtt &&
        !           449:                                    SEQ_GT(ti->ti_ack, tp->t_rtseq))
1.1       root      450:                                        tcp_xmit_timer(tp, tp->t_rtt);
                    451:                                acked = ti->ti_ack - tp->snd_una;
                    452:                                sbdrop(&so->so_snd, acked);
                    453:                                tp->snd_una = ti->ti_ack;
                    454:                                m_freem(m);
                    455: 
                    456:                                /*
                    457:                                 * If all outstanding data are acked, stop
                    458:                                 * retransmit timer, otherwise restart timer
                    459:                                 * using current (possibly backed-off) value.
                    460:                                 * If process is waiting for space,
                    461:                                 * wakeup/selwakeup/signal.  If data
                    462:                                 * are ready to send, let tcp_output
                    463:                                 * decide between more output or persist.
                    464:                                 */
                    465:                                if (tp->snd_una == tp->snd_max)
                    466:                                        tp->t_timer[TCPT_REXMT] = 0;
                    467:                                else if (tp->t_timer[TCPT_PERSIST] == 0)
                    468:                                        tp->t_timer[TCPT_REXMT] = tp->t_rxtcur;
                    469: 
1.1.1.3   root      470:                                /*
1.1       root      471:                                 * This is called because sowwakeup might have
                    472:                                 * put data into so_snd.  Since we don't so sowwakeup,
                    473:                                 * we don't need this.. XXX???
                    474:                                 */
                    475:                                if (so->so_snd.sb_cc)
                    476:                                        (void) tcp_output(tp);
                    477: 
                    478:                                return;
                    479:                        }
                    480:                } else if (ti->ti_ack == tp->snd_una &&
1.1.1.4   root      481:                    tcpfrag_list_empty(tp) &&
1.1       root      482:                    ti->ti_len <= sbspace(&so->so_rcv)) {
                    483:                        /*
                    484:                         * this is a pure, in-sequence data packet
                    485:                         * with nothing on the reassembly queue and
                    486:                         * we have enough buffer space to take it.
                    487:                         */
                    488:                        tp->rcv_nxt += ti->ti_len;
                    489:                        /*
                    490:                         * Add data to socket buffer.
                    491:                         */
                    492:                        if (so->so_emu) {
                    493:                                if (tcp_emu(so,m)) sbappend(so, m);
                    494:                        } else
                    495:                                sbappend(so, m);
1.1.1.3   root      496: 
                    497:                        /*
1.1       root      498:                         * If this is a short packet, then ACK now - with Nagel
                    499:                         *      congestion avoidance sender won't send more until
                    500:                         *      he gets an ACK.
1.1.1.3   root      501:                         *
1.1.1.2   root      502:                         * It is better to not delay acks at all to maximize
                    503:                         * TCP throughput.  See RFC 2581.
1.1.1.3   root      504:                         */
1.1.1.2   root      505:                        tp->t_flags |= TF_ACKNOW;
                    506:                        tcp_output(tp);
1.1       root      507:                        return;
                    508:                }
                    509:        } /* header prediction */
                    510:        /*
                    511:         * Calculate amount of space in receive window,
                    512:         * and then do TCP input processing.
                    513:         * Receive window is amount of space in rcv queue,
                    514:         * but not less than advertised window.
                    515:         */
                    516:        { int win;
                    517:           win = sbspace(&so->so_rcv);
                    518:          if (win < 0)
                    519:            win = 0;
                    520:          tp->rcv_wnd = max(win, (int)(tp->rcv_adv - tp->rcv_nxt));
                    521:        }
                    522: 
                    523:        switch (tp->t_state) {
                    524: 
                    525:        /*
                    526:         * If the state is LISTEN then ignore segment if it contains an RST.
                    527:         * If the segment contains an ACK then it is bad and send a RST.
                    528:         * If it does not contain a SYN then it is not interesting; drop it.
                    529:         * Don't bother responding if the destination was a broadcast.
                    530:         * Otherwise initialize tp->rcv_nxt, and tp->irs, select an initial
                    531:         * tp->iss, and send a segment:
                    532:         *     <SEQ=ISS><ACK=RCV_NXT><CTL=SYN,ACK>
                    533:         * Also initialize tp->snd_nxt to tp->iss+1 and tp->snd_una to tp->iss.
                    534:         * Fill in remote peer address fields if not previously specified.
                    535:         * Enter SYN_RECEIVED state, and process any other fields of this
                    536:         * segment in this state.
                    537:         */
                    538:        case TCPS_LISTEN: {
                    539: 
                    540:          if (tiflags & TH_RST)
                    541:            goto drop;
                    542:          if (tiflags & TH_ACK)
                    543:            goto dropwithreset;
                    544:          if ((tiflags & TH_SYN) == 0)
                    545:            goto drop;
1.1.1.3   root      546: 
1.1       root      547:          /*
                    548:           * This has way too many gotos...
                    549:           * But a bit of spaghetti code never hurt anybody :)
                    550:           */
1.1.1.3   root      551: 
1.1       root      552:          /*
                    553:           * If this is destined for the control address, then flag to
                    554:           * tcp_ctl once connected, otherwise connect
                    555:           */
1.1.1.5 ! root      556:          if ((so->so_faddr.s_addr & slirp->vnetwork_mask.s_addr) ==
        !           557:              slirp->vnetwork_addr.s_addr) {
        !           558:            if (so->so_faddr.s_addr != slirp->vhost_addr.s_addr &&
        !           559:                so->so_faddr.s_addr != slirp->vnameserver_addr.s_addr) {
1.1       root      560:                /* May be an add exec */
1.1.1.5 ! root      561:                for (ex_ptr = slirp->exec_list; ex_ptr;
        !           562:                     ex_ptr = ex_ptr->ex_next) {
1.1.1.3   root      563:                  if(ex_ptr->ex_fport == so->so_fport &&
1.1.1.5 ! root      564:                     so->so_faddr.s_addr == ex_ptr->ex_addr.s_addr) {
1.1       root      565:                    so->so_state |= SS_CTL;
                    566:                    break;
                    567:                  }
                    568:                }
1.1.1.5 ! root      569:                if (so->so_state & SS_CTL) {
        !           570:                    goto cont_input;
        !           571:                }
1.1       root      572:            }
                    573:            /* CTL_ALIAS: Do nothing, tcp_fconnect will be called on it */
                    574:          }
1.1.1.3   root      575: 
1.1       root      576:          if (so->so_emu & EMU_NOCONNECT) {
                    577:            so->so_emu &= ~EMU_NOCONNECT;
                    578:            goto cont_input;
                    579:          }
1.1.1.3   root      580: 
1.1       root      581:          if((tcp_fconnect(so) == -1) && (errno != EINPROGRESS) && (errno != EWOULDBLOCK)) {
                    582:            u_char code=ICMP_UNREACH_NET;
                    583:            DEBUG_MISC((dfd," tcp fconnect errno = %d-%s\n",
                    584:                        errno,strerror(errno)));
                    585:            if(errno == ECONNREFUSED) {
                    586:              /* ACK the SYN, send RST to refuse the connection */
                    587:              tcp_respond(tp, ti, m, ti->ti_seq+1, (tcp_seq)0,
1.1.1.3   root      588:                          TH_RST|TH_ACK);
1.1       root      589:            } else {
                    590:              if(errno == EHOSTUNREACH) code=ICMP_UNREACH_HOST;
                    591:              HTONL(ti->ti_seq);             /* restore tcp header */
                    592:              HTONL(ti->ti_ack);
                    593:              HTONS(ti->ti_win);
                    594:              HTONS(ti->ti_urp);
                    595:              m->m_data -= sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr);
                    596:              m->m_len  += sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr);
                    597:              *ip=save_ip;
                    598:              icmp_error(m, ICMP_UNREACH,code, 0,strerror(errno));
                    599:            }
                    600:            tp = tcp_close(tp);
                    601:            m_free(m);
                    602:          } else {
                    603:            /*
                    604:             * Haven't connected yet, save the current mbuf
                    605:             * and ti, and return
                    606:             * XXX Some OS's don't tell us whether the connect()
                    607:             * succeeded or not.  So we must time it out.
                    608:             */
                    609:            so->so_m = m;
                    610:            so->so_ti = ti;
                    611:            tp->t_timer[TCPT_KEEP] = TCPTV_KEEP_INIT;
                    612:            tp->t_state = TCPS_SYN_RECEIVED;
                    613:          }
                    614:          return;
                    615: 
1.1.1.3   root      616:        cont_conn:
                    617:          /* m==NULL
1.1       root      618:           * Check if the connect succeeded
                    619:           */
                    620:          if (so->so_state & SS_NOFDREF) {
                    621:            tp = tcp_close(tp);
                    622:            goto dropwithreset;
                    623:          }
1.1.1.3   root      624:        cont_input:
1.1       root      625:          tcp_template(tp);
1.1.1.3   root      626: 
1.1       root      627:          if (optp)
                    628:            tcp_dooptions(tp, (u_char *)optp, optlen, ti);
1.1.1.3   root      629: 
1.1       root      630:          if (iss)
                    631:            tp->iss = iss;
1.1.1.3   root      632:          else
1.1.1.5 ! root      633:            tp->iss = slirp->tcp_iss;
        !           634:          slirp->tcp_iss += TCP_ISSINCR/2;
1.1       root      635:          tp->irs = ti->ti_seq;
                    636:          tcp_sendseqinit(tp);
                    637:          tcp_rcvseqinit(tp);
                    638:          tp->t_flags |= TF_ACKNOW;
                    639:          tp->t_state = TCPS_SYN_RECEIVED;
                    640:          tp->t_timer[TCPT_KEEP] = TCPTV_KEEP_INIT;
                    641:          goto trimthenstep6;
                    642:        } /* case TCPS_LISTEN */
1.1.1.3   root      643: 
1.1       root      644:        /*
                    645:         * If the state is SYN_SENT:
                    646:         *      if seg contains an ACK, but not for our SYN, drop the input.
                    647:         *      if seg contains a RST, then drop the connection.
                    648:         *      if seg does not contain SYN, then drop it.
                    649:         * Otherwise this is an acceptable SYN segment
                    650:         *      initialize tp->rcv_nxt and tp->irs
                    651:         *      if seg contains ack then advance tp->snd_una
                    652:         *      if SYN has been acked change to ESTABLISHED else SYN_RCVD state
                    653:         *      arrange for segment to be acked (eventually)
                    654:         *      continue processing rest of data/controls, beginning with URG
                    655:         */
                    656:        case TCPS_SYN_SENT:
                    657:                if ((tiflags & TH_ACK) &&
                    658:                    (SEQ_LEQ(ti->ti_ack, tp->iss) ||
                    659:                     SEQ_GT(ti->ti_ack, tp->snd_max)))
                    660:                        goto dropwithreset;
                    661: 
                    662:                if (tiflags & TH_RST) {
                    663:                        if (tiflags & TH_ACK)
                    664:                                tp = tcp_drop(tp,0); /* XXX Check t_softerror! */
                    665:                        goto drop;
                    666:                }
                    667: 
                    668:                if ((tiflags & TH_SYN) == 0)
                    669:                        goto drop;
                    670:                if (tiflags & TH_ACK) {
                    671:                        tp->snd_una = ti->ti_ack;
                    672:                        if (SEQ_LT(tp->snd_nxt, tp->snd_una))
                    673:                                tp->snd_nxt = tp->snd_una;
                    674:                }
                    675: 
                    676:                tp->t_timer[TCPT_REXMT] = 0;
                    677:                tp->irs = ti->ti_seq;
                    678:                tcp_rcvseqinit(tp);
                    679:                tp->t_flags |= TF_ACKNOW;
                    680:                if (tiflags & TH_ACK && SEQ_GT(tp->snd_una, tp->iss)) {
                    681:                        soisfconnected(so);
                    682:                        tp->t_state = TCPS_ESTABLISHED;
1.1.1.3   root      683: 
1.1       root      684:                        (void) tcp_reass(tp, (struct tcpiphdr *)0,
                    685:                                (struct mbuf *)0);
                    686:                        /*
                    687:                         * if we didn't have to retransmit the SYN,
                    688:                         * use its rtt as our initial srtt & rtt var.
                    689:                         */
                    690:                        if (tp->t_rtt)
                    691:                                tcp_xmit_timer(tp, tp->t_rtt);
                    692:                } else
                    693:                        tp->t_state = TCPS_SYN_RECEIVED;
                    694: 
                    695: trimthenstep6:
                    696:                /*
                    697:                 * Advance ti->ti_seq to correspond to first data byte.
                    698:                 * If data, trim to stay within window,
                    699:                 * dropping FIN if necessary.
                    700:                 */
                    701:                ti->ti_seq++;
                    702:                if (ti->ti_len > tp->rcv_wnd) {
                    703:                        todrop = ti->ti_len - tp->rcv_wnd;
                    704:                        m_adj(m, -todrop);
                    705:                        ti->ti_len = tp->rcv_wnd;
                    706:                        tiflags &= ~TH_FIN;
                    707:                }
                    708:                tp->snd_wl1 = ti->ti_seq - 1;
                    709:                tp->rcv_up = ti->ti_seq;
                    710:                goto step6;
                    711:        } /* switch tp->t_state */
                    712:        /*
                    713:         * States other than LISTEN or SYN_SENT.
1.1.1.5 ! root      714:         * Check that at least some bytes of segment are within
1.1       root      715:         * receive window.  If segment begins before rcv_nxt,
                    716:         * drop leading data (and SYN); if nothing left, just ack.
                    717:         */
                    718:        todrop = tp->rcv_nxt - ti->ti_seq;
                    719:        if (todrop > 0) {
                    720:                if (tiflags & TH_SYN) {
                    721:                        tiflags &= ~TH_SYN;
                    722:                        ti->ti_seq++;
1.1.1.3   root      723:                        if (ti->ti_urp > 1)
1.1       root      724:                                ti->ti_urp--;
                    725:                        else
                    726:                                tiflags &= ~TH_URG;
                    727:                        todrop--;
                    728:                }
                    729:                /*
                    730:                 * Following if statement from Stevens, vol. 2, p. 960.
                    731:                 */
                    732:                if (todrop > ti->ti_len
                    733:                    || (todrop == ti->ti_len && (tiflags & TH_FIN) == 0)) {
                    734:                        /*
                    735:                         * Any valid FIN must be to the left of the window.
                    736:                         * At this point the FIN must be a duplicate or out
                    737:                         * of sequence; drop it.
                    738:                         */
                    739:                        tiflags &= ~TH_FIN;
1.1.1.3   root      740: 
1.1       root      741:                        /*
                    742:                         * Send an ACK to resynchronize and drop any data.
                    743:                         * But keep on processing for RST or ACK.
                    744:                         */
                    745:                        tp->t_flags |= TF_ACKNOW;
                    746:                        todrop = ti->ti_len;
                    747:                }
                    748:                m_adj(m, todrop);
                    749:                ti->ti_seq += todrop;
                    750:                ti->ti_len -= todrop;
                    751:                if (ti->ti_urp > todrop)
                    752:                        ti->ti_urp -= todrop;
                    753:                else {
                    754:                        tiflags &= ~TH_URG;
                    755:                        ti->ti_urp = 0;
                    756:                }
                    757:        }
                    758:        /*
                    759:         * If new data are received on a connection after the
                    760:         * user processes are gone, then RST the other end.
                    761:         */
                    762:        if ((so->so_state & SS_NOFDREF) &&
                    763:            tp->t_state > TCPS_CLOSE_WAIT && ti->ti_len) {
                    764:                tp = tcp_close(tp);
                    765:                goto dropwithreset;
                    766:        }
                    767: 
                    768:        /*
                    769:         * If segment ends after window, drop trailing data
                    770:         * (and PUSH and FIN); if nothing left, just ACK.
                    771:         */
                    772:        todrop = (ti->ti_seq+ti->ti_len) - (tp->rcv_nxt+tp->rcv_wnd);
                    773:        if (todrop > 0) {
                    774:                if (todrop >= ti->ti_len) {
                    775:                        /*
                    776:                         * If a new connection request is received
                    777:                         * while in TIME_WAIT, drop the old connection
                    778:                         * and start over if the sequence numbers
                    779:                         * are above the previous ones.
                    780:                         */
                    781:                        if (tiflags & TH_SYN &&
                    782:                            tp->t_state == TCPS_TIME_WAIT &&
                    783:                            SEQ_GT(ti->ti_seq, tp->rcv_nxt)) {
                    784:                                iss = tp->rcv_nxt + TCP_ISSINCR;
                    785:                                tp = tcp_close(tp);
                    786:                                goto findso;
                    787:                        }
                    788:                        /*
                    789:                         * If window is closed can only take segments at
                    790:                         * window edge, and have to drop data and PUSH from
                    791:                         * incoming segments.  Continue processing, but
                    792:                         * remember to ack.  Otherwise, drop segment
                    793:                         * and ack.
                    794:                         */
                    795:                        if (tp->rcv_wnd == 0 && ti->ti_seq == tp->rcv_nxt) {
                    796:                                tp->t_flags |= TF_ACKNOW;
1.1.1.5 ! root      797:                        } else {
1.1       root      798:                                goto dropafterack;
1.1.1.5 ! root      799:                        }
        !           800:                }
1.1       root      801:                m_adj(m, -todrop);
                    802:                ti->ti_len -= todrop;
                    803:                tiflags &= ~(TH_PUSH|TH_FIN);
                    804:        }
                    805: 
                    806:        /*
                    807:         * If the RST bit is set examine the state:
                    808:         *    SYN_RECEIVED STATE:
                    809:         *      If passive open, return to LISTEN state.
                    810:         *      If active open, inform user that connection was refused.
                    811:         *    ESTABLISHED, FIN_WAIT_1, FIN_WAIT2, CLOSE_WAIT STATES:
                    812:         *      Inform user that connection was reset, and close tcb.
                    813:         *    CLOSING, LAST_ACK, TIME_WAIT STATES
                    814:         *      Close the tcb.
                    815:         */
                    816:        if (tiflags&TH_RST) switch (tp->t_state) {
                    817: 
                    818:        case TCPS_SYN_RECEIVED:
                    819:        case TCPS_ESTABLISHED:
                    820:        case TCPS_FIN_WAIT_1:
                    821:        case TCPS_FIN_WAIT_2:
                    822:        case TCPS_CLOSE_WAIT:
                    823:                tp->t_state = TCPS_CLOSED;
                    824:                tp = tcp_close(tp);
                    825:                goto drop;
                    826: 
                    827:        case TCPS_CLOSING:
                    828:        case TCPS_LAST_ACK:
                    829:        case TCPS_TIME_WAIT:
                    830:                tp = tcp_close(tp);
                    831:                goto drop;
                    832:        }
                    833: 
                    834:        /*
                    835:         * If a SYN is in the window, then this is an
                    836:         * error and we send an RST and drop the connection.
                    837:         */
                    838:        if (tiflags & TH_SYN) {
                    839:                tp = tcp_drop(tp,0);
                    840:                goto dropwithreset;
                    841:        }
                    842: 
                    843:        /*
                    844:         * If the ACK bit is off we drop the segment and return.
                    845:         */
                    846:        if ((tiflags & TH_ACK) == 0) goto drop;
                    847: 
                    848:        /*
                    849:         * Ack processing.
                    850:         */
                    851:        switch (tp->t_state) {
                    852:        /*
                    853:         * In SYN_RECEIVED state if the ack ACKs our SYN then enter
                    854:         * ESTABLISHED state and continue processing, otherwise
                    855:         * send an RST.  una<=ack<=max
                    856:         */
                    857:        case TCPS_SYN_RECEIVED:
                    858: 
                    859:                if (SEQ_GT(tp->snd_una, ti->ti_ack) ||
                    860:                    SEQ_GT(ti->ti_ack, tp->snd_max))
                    861:                        goto dropwithreset;
                    862:                tp->t_state = TCPS_ESTABLISHED;
1.1.1.3   root      863:                /*
                    864:                 * The sent SYN is ack'ed with our sequence number +1
                    865:                 * The first data byte already in the buffer will get
1.1       root      866:                 * lost if no correction is made.  This is only needed for
                    867:                 * SS_CTL since the buffer is empty otherwise.
1.1.1.3   root      868:                 * tp->snd_una++; or:
1.1       root      869:                 */
                    870:                tp->snd_una=ti->ti_ack;
                    871:                if (so->so_state & SS_CTL) {
                    872:                  /* So tcp_ctl reports the right state */
                    873:                  ret = tcp_ctl(so);
                    874:                  if (ret == 1) {
                    875:                    soisfconnected(so);
                    876:                    so->so_state &= ~SS_CTL;   /* success XXX */
                    877:                  } else if (ret == 2) {
1.1.1.5 ! root      878:                    so->so_state &= SS_PERSISTENT_MASK;
        !           879:                    so->so_state |= SS_NOFDREF; /* CTL_CMD */
1.1       root      880:                  } else {
                    881:                    needoutput = 1;
                    882:                    tp->t_state = TCPS_FIN_WAIT_1;
                    883:                  }
                    884:                } else {
                    885:                  soisfconnected(so);
                    886:                }
1.1.1.3   root      887: 
1.1       root      888:                (void) tcp_reass(tp, (struct tcpiphdr *)0, (struct mbuf *)0);
                    889:                tp->snd_wl1 = ti->ti_seq - 1;
                    890:                /* Avoid ack processing; snd_una==ti_ack  =>  dup ack */
                    891:                goto synrx_to_est;
                    892:                /* fall into ... */
                    893: 
                    894:        /*
                    895:         * In ESTABLISHED state: drop duplicate ACKs; ACK out of range
                    896:         * ACKs.  If the ack is in the range
                    897:         *      tp->snd_una < ti->ti_ack <= tp->snd_max
                    898:         * then advance tp->snd_una to ti->ti_ack and drop
                    899:         * data from the retransmission queue.  If this ACK reflects
                    900:         * more up to date window information we update our window information.
                    901:         */
                    902:        case TCPS_ESTABLISHED:
                    903:        case TCPS_FIN_WAIT_1:
                    904:        case TCPS_FIN_WAIT_2:
                    905:        case TCPS_CLOSE_WAIT:
                    906:        case TCPS_CLOSING:
                    907:        case TCPS_LAST_ACK:
                    908:        case TCPS_TIME_WAIT:
                    909: 
                    910:                if (SEQ_LEQ(ti->ti_ack, tp->snd_una)) {
                    911:                        if (ti->ti_len == 0 && tiwin == tp->snd_wnd) {
                    912:                          DEBUG_MISC((dfd," dup ack  m = %lx  so = %lx \n",
                    913:                                      (long )m, (long )so));
                    914:                                /*
                    915:                                 * If we have outstanding data (other than
                    916:                                 * a window probe), this is a completely
                    917:                                 * duplicate ack (ie, window info didn't
                    918:                                 * change), the ack is the biggest we've
                    919:                                 * seen and we've seen exactly our rexmt
                    920:                                 * threshold of them, assume a packet
                    921:                                 * has been dropped and retransmit it.
                    922:                                 * Kludge snd_nxt & the congestion
                    923:                                 * window so we send only this one
                    924:                                 * packet.
                    925:                                 *
                    926:                                 * We know we're losing at the current
                    927:                                 * window size so do congestion avoidance
                    928:                                 * (set ssthresh to half the current window
                    929:                                 * and pull our congestion window back to
                    930:                                 * the new ssthresh).
                    931:                                 *
                    932:                                 * Dup acks mean that packets have left the
1.1.1.3   root      933:                                 * network (they're now cached at the receiver)
1.1       root      934:                                 * so bump cwnd by the amount in the receiver
                    935:                                 * to keep a constant cwnd packets in the
                    936:                                 * network.
                    937:                                 */
                    938:                                if (tp->t_timer[TCPT_REXMT] == 0 ||
                    939:                                    ti->ti_ack != tp->snd_una)
                    940:                                        tp->t_dupacks = 0;
1.1.1.3   root      941:                                else if (++tp->t_dupacks == TCPREXMTTHRESH) {
1.1       root      942:                                        tcp_seq onxt = tp->snd_nxt;
                    943:                                        u_int win =
                    944:                                            min(tp->snd_wnd, tp->snd_cwnd) / 2 /
                    945:                                                tp->t_maxseg;
                    946: 
                    947:                                        if (win < 2)
                    948:                                                win = 2;
                    949:                                        tp->snd_ssthresh = win * tp->t_maxseg;
                    950:                                        tp->t_timer[TCPT_REXMT] = 0;
                    951:                                        tp->t_rtt = 0;
                    952:                                        tp->snd_nxt = ti->ti_ack;
                    953:                                        tp->snd_cwnd = tp->t_maxseg;
                    954:                                        (void) tcp_output(tp);
                    955:                                        tp->snd_cwnd = tp->snd_ssthresh +
                    956:                                               tp->t_maxseg * tp->t_dupacks;
                    957:                                        if (SEQ_GT(onxt, tp->snd_nxt))
                    958:                                                tp->snd_nxt = onxt;
                    959:                                        goto drop;
1.1.1.3   root      960:                                } else if (tp->t_dupacks > TCPREXMTTHRESH) {
1.1       root      961:                                        tp->snd_cwnd += tp->t_maxseg;
                    962:                                        (void) tcp_output(tp);
                    963:                                        goto drop;
                    964:                                }
                    965:                        } else
                    966:                                tp->t_dupacks = 0;
                    967:                        break;
                    968:                }
                    969:        synrx_to_est:
                    970:                /*
                    971:                 * If the congestion window was inflated to account
                    972:                 * for the other side's cached packets, retract it.
                    973:                 */
1.1.1.3   root      974:                if (tp->t_dupacks > TCPREXMTTHRESH &&
1.1       root      975:                    tp->snd_cwnd > tp->snd_ssthresh)
                    976:                        tp->snd_cwnd = tp->snd_ssthresh;
                    977:                tp->t_dupacks = 0;
                    978:                if (SEQ_GT(ti->ti_ack, tp->snd_max)) {
                    979:                        goto dropafterack;
                    980:                }
                    981:                acked = ti->ti_ack - tp->snd_una;
                    982: 
                    983:                /*
1.1.1.5 ! root      984:                 * If transmit timer is running and timed sequence
1.1       root      985:                 * number was acked, update smoothed round trip time.
                    986:                 * Since we now have an rtt measurement, cancel the
                    987:                 * timer backoff (cf., Phil Karn's retransmit alg.).
                    988:                 * Recompute the initial retransmit timer.
                    989:                 */
1.1.1.5 ! root      990:                if (tp->t_rtt && SEQ_GT(ti->ti_ack, tp->t_rtseq))
1.1       root      991:                        tcp_xmit_timer(tp,tp->t_rtt);
                    992: 
                    993:                /*
                    994:                 * If all outstanding data is acked, stop retransmit
                    995:                 * timer and remember to restart (more output or persist).
                    996:                 * If there is more data to be acked, restart retransmit
                    997:                 * timer, using current (possibly backed-off) value.
                    998:                 */
                    999:                if (ti->ti_ack == tp->snd_max) {
                   1000:                        tp->t_timer[TCPT_REXMT] = 0;
                   1001:                        needoutput = 1;
                   1002:                } else if (tp->t_timer[TCPT_PERSIST] == 0)
                   1003:                        tp->t_timer[TCPT_REXMT] = tp->t_rxtcur;
                   1004:                /*
                   1005:                 * When new data is acked, open the congestion window.
                   1006:                 * If the window gives us less than ssthresh packets
                   1007:                 * in flight, open exponentially (maxseg per packet).
                   1008:                 * Otherwise open linearly: maxseg per window
                   1009:                 * (maxseg^2 / cwnd per packet).
                   1010:                 */
                   1011:                {
                   1012:                  register u_int cw = tp->snd_cwnd;
                   1013:                  register u_int incr = tp->t_maxseg;
                   1014: 
                   1015:                  if (cw > tp->snd_ssthresh)
                   1016:                    incr = incr * incr / cw;
                   1017:                  tp->snd_cwnd = min(cw + incr, TCP_MAXWIN<<tp->snd_scale);
                   1018:                }
                   1019:                if (acked > so->so_snd.sb_cc) {
                   1020:                        tp->snd_wnd -= so->so_snd.sb_cc;
                   1021:                        sbdrop(&so->so_snd, (int )so->so_snd.sb_cc);
                   1022:                        ourfinisacked = 1;
                   1023:                } else {
                   1024:                        sbdrop(&so->so_snd, acked);
                   1025:                        tp->snd_wnd -= acked;
                   1026:                        ourfinisacked = 0;
                   1027:                }
                   1028:                tp->snd_una = ti->ti_ack;
                   1029:                if (SEQ_LT(tp->snd_nxt, tp->snd_una))
                   1030:                        tp->snd_nxt = tp->snd_una;
                   1031: 
                   1032:                switch (tp->t_state) {
                   1033: 
                   1034:                /*
                   1035:                 * In FIN_WAIT_1 STATE in addition to the processing
                   1036:                 * for the ESTABLISHED state if our FIN is now acknowledged
                   1037:                 * then enter FIN_WAIT_2.
                   1038:                 */
                   1039:                case TCPS_FIN_WAIT_1:
                   1040:                        if (ourfinisacked) {
                   1041:                                /*
                   1042:                                 * If we can't receive any more
                   1043:                                 * data, then closing user can proceed.
                   1044:                                 * Starting the timer is contrary to the
                   1045:                                 * specification, but if we don't get a FIN
                   1046:                                 * we'll hang forever.
                   1047:                                 */
                   1048:                                if (so->so_state & SS_FCANTRCVMORE) {
1.1.1.3   root     1049:                                        tp->t_timer[TCPT_2MSL] = TCP_MAXIDLE;
1.1       root     1050:                                }
                   1051:                                tp->t_state = TCPS_FIN_WAIT_2;
                   1052:                        }
                   1053:                        break;
                   1054: 
                   1055:                /*
                   1056:                 * In CLOSING STATE in addition to the processing for
                   1057:                 * the ESTABLISHED state if the ACK acknowledges our FIN
                   1058:                 * then enter the TIME-WAIT state, otherwise ignore
                   1059:                 * the segment.
                   1060:                 */
                   1061:                case TCPS_CLOSING:
                   1062:                        if (ourfinisacked) {
                   1063:                                tp->t_state = TCPS_TIME_WAIT;
                   1064:                                tcp_canceltimers(tp);
                   1065:                                tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL;
                   1066:                        }
                   1067:                        break;
                   1068: 
                   1069:                /*
                   1070:                 * In LAST_ACK, we may still be waiting for data to drain
                   1071:                 * and/or to be acked, as well as for the ack of our FIN.
                   1072:                 * If our FIN is now acknowledged, delete the TCB,
                   1073:                 * enter the closed state and return.
                   1074:                 */
                   1075:                case TCPS_LAST_ACK:
                   1076:                        if (ourfinisacked) {
                   1077:                                tp = tcp_close(tp);
                   1078:                                goto drop;
                   1079:                        }
                   1080:                        break;
                   1081: 
                   1082:                /*
                   1083:                 * In TIME_WAIT state the only thing that should arrive
                   1084:                 * is a retransmission of the remote FIN.  Acknowledge
                   1085:                 * it and restart the finack timer.
                   1086:                 */
                   1087:                case TCPS_TIME_WAIT:
                   1088:                        tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL;
                   1089:                        goto dropafterack;
                   1090:                }
                   1091:        } /* switch(tp->t_state) */
                   1092: 
                   1093: step6:
                   1094:        /*
                   1095:         * Update window information.
                   1096:         * Don't look at window if no ACK: TAC's send garbage on first SYN.
                   1097:         */
                   1098:        if ((tiflags & TH_ACK) &&
1.1.1.3   root     1099:            (SEQ_LT(tp->snd_wl1, ti->ti_seq) ||
1.1       root     1100:            (tp->snd_wl1 == ti->ti_seq && (SEQ_LT(tp->snd_wl2, ti->ti_ack) ||
                   1101:            (tp->snd_wl2 == ti->ti_ack && tiwin > tp->snd_wnd))))) {
                   1102:                tp->snd_wnd = tiwin;
                   1103:                tp->snd_wl1 = ti->ti_seq;
                   1104:                tp->snd_wl2 = ti->ti_ack;
                   1105:                if (tp->snd_wnd > tp->max_sndwnd)
                   1106:                        tp->max_sndwnd = tp->snd_wnd;
                   1107:                needoutput = 1;
                   1108:        }
                   1109: 
                   1110:        /*
                   1111:         * Process segments with URG.
                   1112:         */
                   1113:        if ((tiflags & TH_URG) && ti->ti_urp &&
                   1114:            TCPS_HAVERCVDFIN(tp->t_state) == 0) {
                   1115:                /*
                   1116:                 * This is a kludge, but if we receive and accept
                   1117:                 * random urgent pointers, we'll crash in
                   1118:                 * soreceive.  It's hard to imagine someone
                   1119:                 * actually wanting to send this much urgent data.
                   1120:                 */
                   1121:                if (ti->ti_urp + so->so_rcv.sb_cc > so->so_rcv.sb_datalen) {
                   1122:                        ti->ti_urp = 0;
                   1123:                        tiflags &= ~TH_URG;
                   1124:                        goto dodata;
                   1125:                }
                   1126:                /*
                   1127:                 * If this segment advances the known urgent pointer,
                   1128:                 * then mark the data stream.  This should not happen
                   1129:                 * in CLOSE_WAIT, CLOSING, LAST_ACK or TIME_WAIT STATES since
1.1.1.3   root     1130:                 * a FIN has been received from the remote side.
1.1       root     1131:                 * In these states we ignore the URG.
                   1132:                 *
                   1133:                 * According to RFC961 (Assigned Protocols),
                   1134:                 * the urgent pointer points to the last octet
                   1135:                 * of urgent data.  We continue, however,
                   1136:                 * to consider it to indicate the first octet
1.1.1.3   root     1137:                 * of data past the urgent section as the original
1.1       root     1138:                 * spec states (in one of two places).
                   1139:                 */
                   1140:                if (SEQ_GT(ti->ti_seq+ti->ti_urp, tp->rcv_up)) {
                   1141:                        tp->rcv_up = ti->ti_seq + ti->ti_urp;
                   1142:                        so->so_urgc =  so->so_rcv.sb_cc +
                   1143:                                (tp->rcv_up - tp->rcv_nxt); /* -1; */
                   1144:                        tp->rcv_up = ti->ti_seq + ti->ti_urp;
1.1.1.3   root     1145: 
1.1       root     1146:                }
                   1147:        } else
                   1148:                /*
                   1149:                 * If no out of band data is expected,
                   1150:                 * pull receive urgent pointer along
                   1151:                 * with the receive window.
                   1152:                 */
                   1153:                if (SEQ_GT(tp->rcv_nxt, tp->rcv_up))
                   1154:                        tp->rcv_up = tp->rcv_nxt;
                   1155: dodata:
                   1156: 
                   1157:        /*
                   1158:         * Process the segment text, merging it into the TCP sequencing queue,
                   1159:         * and arranging for acknowledgment of receipt if necessary.
                   1160:         * This process logically involves adjusting tp->rcv_wnd as data
                   1161:         * is presented to the user (this happens in tcp_usrreq.c,
                   1162:         * case PRU_RCVD).  If a FIN has already been received on this
                   1163:         * connection then we just ignore the text.
                   1164:         */
                   1165:        if ((ti->ti_len || (tiflags&TH_FIN)) &&
                   1166:            TCPS_HAVERCVDFIN(tp->t_state) == 0) {
                   1167:                TCP_REASS(tp, ti, m, so, tiflags);
                   1168:                /*
                   1169:                 * Note the amount of data that peer has sent into
                   1170:                 * our window, in order to estimate the sender's
                   1171:                 * buffer size.
                   1172:                 */
                   1173:                len = so->so_rcv.sb_datalen - (tp->rcv_adv - tp->rcv_nxt);
                   1174:        } else {
                   1175:                m_free(m);
                   1176:                tiflags &= ~TH_FIN;
                   1177:        }
                   1178: 
                   1179:        /*
                   1180:         * If FIN is received ACK the FIN and let the user know
                   1181:         * that the connection is closing.
                   1182:         */
                   1183:        if (tiflags & TH_FIN) {
                   1184:                if (TCPS_HAVERCVDFIN(tp->t_state) == 0) {
                   1185:                        /*
                   1186:                         * If we receive a FIN we can't send more data,
                   1187:                         * set it SS_FDRAIN
                   1188:                          * Shutdown the socket if there is no rx data in the
                   1189:                         * buffer.
                   1190:                         * soread() is called on completion of shutdown() and
                   1191:                         * will got to TCPS_LAST_ACK, and use tcp_output()
                   1192:                         * to send the FIN.
                   1193:                         */
                   1194:                        sofwdrain(so);
1.1.1.3   root     1195: 
1.1       root     1196:                        tp->t_flags |= TF_ACKNOW;
                   1197:                        tp->rcv_nxt++;
                   1198:                }
                   1199:                switch (tp->t_state) {
                   1200: 
                   1201:                /*
                   1202:                 * In SYN_RECEIVED and ESTABLISHED STATES
                   1203:                 * enter the CLOSE_WAIT state.
                   1204:                 */
                   1205:                case TCPS_SYN_RECEIVED:
                   1206:                case TCPS_ESTABLISHED:
                   1207:                  if(so->so_emu == EMU_CTL)        /* no shutdown on socket */
                   1208:                    tp->t_state = TCPS_LAST_ACK;
1.1.1.3   root     1209:                  else
1.1       root     1210:                    tp->t_state = TCPS_CLOSE_WAIT;
                   1211:                  break;
                   1212: 
                   1213:                /*
                   1214:                 * If still in FIN_WAIT_1 STATE FIN has not been acked so
                   1215:                 * enter the CLOSING state.
                   1216:                 */
                   1217:                case TCPS_FIN_WAIT_1:
                   1218:                        tp->t_state = TCPS_CLOSING;
                   1219:                        break;
                   1220: 
                   1221:                /*
                   1222:                 * In FIN_WAIT_2 state enter the TIME_WAIT state,
1.1.1.3   root     1223:                 * starting the time-wait timer, turning off the other
1.1       root     1224:                 * standard timers.
                   1225:                 */
                   1226:                case TCPS_FIN_WAIT_2:
                   1227:                        tp->t_state = TCPS_TIME_WAIT;
                   1228:                        tcp_canceltimers(tp);
                   1229:                        tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL;
                   1230:                        break;
                   1231: 
                   1232:                /*
                   1233:                 * In TIME_WAIT state restart the 2 MSL time_wait timer.
                   1234:                 */
                   1235:                case TCPS_TIME_WAIT:
                   1236:                        tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL;
                   1237:                        break;
                   1238:                }
                   1239:        }
                   1240: 
                   1241:        /*
                   1242:         * If this is a small packet, then ACK now - with Nagel
                   1243:         *      congestion avoidance sender won't send more until
                   1244:         *      he gets an ACK.
1.1.1.3   root     1245:         *
1.1       root     1246:         * See above.
                   1247:         */
                   1248:        if (ti->ti_len && (unsigned)ti->ti_len <= 5 &&
                   1249:            ((struct tcpiphdr_2 *)ti)->first_char == (char)27) {
                   1250:                tp->t_flags |= TF_ACKNOW;
                   1251:        }
                   1252: 
                   1253:        /*
                   1254:         * Return any desired output.
                   1255:         */
                   1256:        if (needoutput || (tp->t_flags & TF_ACKNOW)) {
                   1257:                (void) tcp_output(tp);
                   1258:        }
                   1259:        return;
                   1260: 
                   1261: dropafterack:
                   1262:        /*
                   1263:         * Generate an ACK dropping incoming segment if it occupies
                   1264:         * sequence space, where the ACK reflects our state.
                   1265:         */
                   1266:        if (tiflags & TH_RST)
                   1267:                goto drop;
                   1268:        m_freem(m);
                   1269:        tp->t_flags |= TF_ACKNOW;
                   1270:        (void) tcp_output(tp);
                   1271:        return;
                   1272: 
                   1273: dropwithreset:
                   1274:        /* reuses m if m!=NULL, m_free() unnecessary */
                   1275:        if (tiflags & TH_ACK)
                   1276:                tcp_respond(tp, ti, m, (tcp_seq)0, ti->ti_ack, TH_RST);
                   1277:        else {
                   1278:                if (tiflags & TH_SYN) ti->ti_len++;
                   1279:                tcp_respond(tp, ti, m, ti->ti_seq+ti->ti_len, (tcp_seq)0,
                   1280:                    TH_RST|TH_ACK);
                   1281:        }
                   1282: 
                   1283:        return;
                   1284: 
                   1285: drop:
                   1286:        /*
                   1287:         * Drop space held by incoming segment and return.
                   1288:         */
                   1289:        m_free(m);
                   1290: 
                   1291:        return;
                   1292: }
                   1293: 
1.1.1.3   root     1294: static void
                   1295: tcp_dooptions(struct tcpcb *tp, u_char *cp, int cnt, struct tcpiphdr *ti)
1.1       root     1296: {
                   1297:        u_int16_t mss;
                   1298:        int opt, optlen;
                   1299: 
                   1300:        DEBUG_CALL("tcp_dooptions");
                   1301:        DEBUG_ARGS((dfd," tp = %lx  cnt=%i \n", (long )tp, cnt));
                   1302: 
                   1303:        for (; cnt > 0; cnt -= optlen, cp += optlen) {
                   1304:                opt = cp[0];
                   1305:                if (opt == TCPOPT_EOL)
                   1306:                        break;
                   1307:                if (opt == TCPOPT_NOP)
                   1308:                        optlen = 1;
                   1309:                else {
                   1310:                        optlen = cp[1];
                   1311:                        if (optlen <= 0)
                   1312:                                break;
                   1313:                }
                   1314:                switch (opt) {
                   1315: 
                   1316:                default:
                   1317:                        continue;
                   1318: 
                   1319:                case TCPOPT_MAXSEG:
                   1320:                        if (optlen != TCPOLEN_MAXSEG)
                   1321:                                continue;
                   1322:                        if (!(ti->ti_flags & TH_SYN))
                   1323:                                continue;
                   1324:                        memcpy((char *) &mss, (char *) cp + 2, sizeof(mss));
                   1325:                        NTOHS(mss);
                   1326:                        (void) tcp_mss(tp, mss);        /* sets t_maxseg */
                   1327:                        break;
                   1328:                }
                   1329:        }
                   1330: }
                   1331: 
                   1332: 
                   1333: /*
                   1334:  * Pull out of band byte out of a segment so
                   1335:  * it doesn't appear in the user's data queue.
                   1336:  * It is still reflected in the segment length for
                   1337:  * sequencing purposes.
                   1338:  */
                   1339: 
                   1340: #ifdef notdef
                   1341: 
                   1342: void
                   1343: tcp_pulloutofband(so, ti, m)
                   1344:        struct socket *so;
                   1345:        struct tcpiphdr *ti;
                   1346:        register struct mbuf *m;
                   1347: {
                   1348:        int cnt = ti->ti_urp - 1;
1.1.1.3   root     1349: 
1.1       root     1350:        while (cnt >= 0) {
                   1351:                if (m->m_len > cnt) {
                   1352:                        char *cp = mtod(m, caddr_t) + cnt;
                   1353:                        struct tcpcb *tp = sototcpcb(so);
                   1354: 
                   1355:                        tp->t_iobc = *cp;
                   1356:                        tp->t_oobflags |= TCPOOB_HAVEDATA;
                   1357:                        memcpy(sp, cp+1, (unsigned)(m->m_len - cnt - 1));
                   1358:                        m->m_len--;
                   1359:                        return;
                   1360:                }
                   1361:                cnt -= m->m_len;
                   1362:                m = m->m_next; /* XXX WRONG! Fix it! */
                   1363:                if (m == 0)
                   1364:                        break;
                   1365:        }
                   1366:        panic("tcp_pulloutofband");
                   1367: }
                   1368: 
                   1369: #endif /* notdef */
                   1370: 
                   1371: /*
                   1372:  * Collect new round-trip time estimate
                   1373:  * and update averages and current timeout.
                   1374:  */
                   1375: 
1.1.1.3   root     1376: static void
                   1377: tcp_xmit_timer(register struct tcpcb *tp, int rtt)
1.1       root     1378: {
                   1379:        register short delta;
                   1380: 
                   1381:        DEBUG_CALL("tcp_xmit_timer");
                   1382:        DEBUG_ARG("tp = %lx", (long)tp);
                   1383:        DEBUG_ARG("rtt = %d", rtt);
1.1.1.3   root     1384: 
1.1       root     1385:        if (tp->t_srtt != 0) {
                   1386:                /*
                   1387:                 * srtt is stored as fixed point with 3 bits after the
                   1388:                 * binary point (i.e., scaled by 8).  The following magic
                   1389:                 * is equivalent to the smoothing algorithm in rfc793 with
                   1390:                 * an alpha of .875 (srtt = rtt/8 + srtt*7/8 in fixed
                   1391:                 * point).  Adjust rtt to origin 0.
                   1392:                 */
                   1393:                delta = rtt - 1 - (tp->t_srtt >> TCP_RTT_SHIFT);
                   1394:                if ((tp->t_srtt += delta) <= 0)
                   1395:                        tp->t_srtt = 1;
                   1396:                /*
                   1397:                 * We accumulate a smoothed rtt variance (actually, a
                   1398:                 * smoothed mean difference), then set the retransmit
                   1399:                 * timer to smoothed rtt + 4 times the smoothed variance.
                   1400:                 * rttvar is stored as fixed point with 2 bits after the
                   1401:                 * binary point (scaled by 4).  The following is
                   1402:                 * equivalent to rfc793 smoothing with an alpha of .75
                   1403:                 * (rttvar = rttvar*3/4 + |delta| / 4).  This replaces
                   1404:                 * rfc793's wired-in beta.
                   1405:                 */
                   1406:                if (delta < 0)
                   1407:                        delta = -delta;
                   1408:                delta -= (tp->t_rttvar >> TCP_RTTVAR_SHIFT);
                   1409:                if ((tp->t_rttvar += delta) <= 0)
                   1410:                        tp->t_rttvar = 1;
                   1411:        } else {
1.1.1.3   root     1412:                /*
1.1       root     1413:                 * No rtt measurement yet - use the unsmoothed rtt.
                   1414:                 * Set the variance to half the rtt (so our first
                   1415:                 * retransmit happens at 3*rtt).
                   1416:                 */
                   1417:                tp->t_srtt = rtt << TCP_RTT_SHIFT;
                   1418:                tp->t_rttvar = rtt << (TCP_RTTVAR_SHIFT - 1);
                   1419:        }
                   1420:        tp->t_rtt = 0;
                   1421:        tp->t_rxtshift = 0;
                   1422: 
                   1423:        /*
                   1424:         * the retransmit should happen at rtt + 4 * rttvar.
                   1425:         * Because of the way we do the smoothing, srtt and rttvar
                   1426:         * will each average +1/2 tick of bias.  When we compute
                   1427:         * the retransmit timer, we want 1/2 tick of rounding and
                   1428:         * 1 extra tick because of +-1/2 tick uncertainty in the
                   1429:         * firing of the timer.  The bias will give us exactly the
                   1430:         * 1.5 tick we need.  But, because the bias is
                   1431:         * statistical, we have to test that we don't drop below
                   1432:         * the minimum feasible timer (which is 2 ticks).
                   1433:         */
                   1434:        TCPT_RANGESET(tp->t_rxtcur, TCP_REXMTVAL(tp),
                   1435:            (short)tp->t_rttmin, TCPTV_REXMTMAX); /* XXX */
1.1.1.3   root     1436: 
1.1       root     1437:        /*
                   1438:         * We received an ack for a packet that wasn't retransmitted;
                   1439:         * it is probably safe to discard any error indications we've
                   1440:         * received recently.  This isn't quite right, but close enough
                   1441:         * for now (a route might have failed after we sent a segment,
                   1442:         * and the return path might not be symmetrical).
                   1443:         */
                   1444:        tp->t_softerror = 0;
                   1445: }
                   1446: 
                   1447: /*
                   1448:  * Determine a reasonable value for maxseg size.
                   1449:  * If the route is known, check route for mtu.
                   1450:  * If none, use an mss that can be handled on the outgoing
                   1451:  * interface without forcing IP to fragment; if bigger than
                   1452:  * an mbuf cluster (MCLBYTES), round down to nearest multiple of MCLBYTES
                   1453:  * to utilize large mbufs.  If no route is found, route has no mtu,
                   1454:  * or the destination isn't local, use a default, hopefully conservative
                   1455:  * size (usually 512 or the default IP max size, but no more than the mtu
                   1456:  * of the interface), as we can't discover anything about intervening
                   1457:  * gateways or networks.  We also initialize the congestion/slow start
                   1458:  * window to be a single segment if the destination isn't local.
                   1459:  * While looking at the routing entry, we also initialize other path-dependent
                   1460:  * parameters from pre-set or cached values in the routing entry.
                   1461:  */
                   1462: 
                   1463: int
1.1.1.5 ! root     1464: tcp_mss(struct tcpcb *tp, u_int offer)
1.1       root     1465: {
                   1466:        struct socket *so = tp->t_socket;
                   1467:        int mss;
1.1.1.3   root     1468: 
1.1       root     1469:        DEBUG_CALL("tcp_mss");
                   1470:        DEBUG_ARG("tp = %lx", (long)tp);
                   1471:        DEBUG_ARG("offer = %d", offer);
1.1.1.3   root     1472: 
                   1473:        mss = min(IF_MTU, IF_MRU) - sizeof(struct tcpiphdr);
1.1       root     1474:        if (offer)
                   1475:                mss = min(mss, offer);
                   1476:        mss = max(mss, 32);
                   1477:        if (mss < tp->t_maxseg || offer != 0)
                   1478:           tp->t_maxseg = mss;
1.1.1.3   root     1479: 
1.1       root     1480:        tp->snd_cwnd = mss;
1.1.1.3   root     1481: 
                   1482:        sbreserve(&so->so_snd, TCP_SNDSPACE + ((TCP_SNDSPACE % mss) ?
                   1483:                                                (mss - (TCP_SNDSPACE % mss)) :
                   1484:                                                0));
                   1485:        sbreserve(&so->so_rcv, TCP_RCVSPACE + ((TCP_RCVSPACE % mss) ?
                   1486:                                                (mss - (TCP_RCVSPACE % mss)) :
                   1487:                                                0));
                   1488: 
1.1       root     1489:        DEBUG_MISC((dfd, " returning mss = %d\n", mss));
1.1.1.3   root     1490: 
1.1       root     1491:        return mss;
                   1492: }

unix.superglobalmegacorp.com