Annotation of 43BSDReno/sys/netinet/tcp_usrreq.c, revision 1.1.1.1

1.1       root        1: /*
                      2:  * Copyright (c) 1982, 1986, 1988 Regents of the University of California.
                      3:  * All rights reserved.
                      4:  *
                      5:  * Redistribution is only permitted until one year after the first shipment
                      6:  * of 4.4BSD by the Regents.  Otherwise, redistribution and use in source and
                      7:  * binary forms are permitted provided that: (1) source distributions retain
                      8:  * this entire copyright notice and comment, and (2) distributions including
                      9:  * binaries display the following acknowledgement:  This product includes
                     10:  * software developed by the University of California, Berkeley and its
                     11:  * contributors'' in the documentation or other materials provided with the
                     12:  * distribution and in all advertising materials mentioning features or use
                     13:  * of this software.  Neither the name of the University nor the names of
                     14:  * its contributors may be used to endorse or promote products derived from
                     15:  * this software without specific prior written permission.
                     16:  * THIS SOFTWARE IS PROVIDED AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED
                     17:  * WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
                     18:  * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
                     19:  *
                     20:  *     @(#)tcp_usrreq.c        7.15 (Berkeley) 6/28/90
                     21:  */
                     22: 
                     23: #include "param.h"
                     24: #include "systm.h"
                     25: #include "malloc.h"
                     26: #include "mbuf.h"
                     27: #include "socket.h"
                     28: #include "socketvar.h"
                     29: #include "protosw.h"
                     30: #include "errno.h"
                     31: #include "stat.h"
                     32: 
                     33: #include "../net/if.h"
                     34: #include "../net/route.h"
                     35: 
                     36: #include "in.h"
                     37: #include "in_systm.h"
                     38: #include "ip.h"
                     39: #include "in_pcb.h"
                     40: #include "ip_var.h"
                     41: #include "tcp.h"
                     42: #include "tcp_fsm.h"
                     43: #include "tcp_seq.h"
                     44: #include "tcp_timer.h"
                     45: #include "tcp_var.h"
                     46: #include "tcpip.h"
                     47: #include "tcp_debug.h"
                     48: 
                     49: /*
                     50:  * TCP protocol interface to socket abstraction.
                     51:  */
                     52: extern char *tcpstates[];
                     53: struct tcpcb *tcp_newtcpcb();
                     54: 
                     55: /*
                     56:  * Process a TCP user request for TCP tb.  If this is a send request
                     57:  * then m is the mbuf chain of send data.  If this is a timer expiration
                     58:  * (called from the software clock routine), then timertype tells which timer.
                     59:  */
                     60: /*ARGSUSED*/
                     61: tcp_usrreq(so, req, m, nam, control)
                     62:        struct socket *so;
                     63:        int req;
                     64:        struct mbuf *m, *nam, *control;
                     65: {
                     66:        register struct inpcb *inp;
                     67:        register struct tcpcb *tp;
                     68:        int s;
                     69:        int error = 0;
                     70:        int ostate;
                     71: 
                     72:        if (req == PRU_CONTROL)
                     73:                return (in_control(so, (int)m, (caddr_t)nam,
                     74:                        (struct ifnet *)control));
                     75:        if (control && control->m_len) {
                     76:                m_freem(control);
                     77:                if (m)
                     78:                        m_freem(m);
                     79:                return (EINVAL);
                     80:        }
                     81: 
                     82:        s = splnet();
                     83:        inp = sotoinpcb(so);
                     84:        /*
                     85:         * When a TCP is attached to a socket, then there will be
                     86:         * a (struct inpcb) pointed at by the socket, and this
                     87:         * structure will point at a subsidary (struct tcpcb).
                     88:         */
                     89:        if (inp == 0 && req != PRU_ATTACH) {
                     90:                splx(s);
                     91:                return (EINVAL);                /* XXX */
                     92:        }
                     93:        if (inp) {
                     94:                tp = intotcpcb(inp);
                     95:                /* WHAT IF TP IS 0? */
                     96: #ifdef KPROF
                     97:                tcp_acounts[tp->t_state][req]++;
                     98: #endif
                     99:                ostate = tp->t_state;
                    100:        } else
                    101:                ostate = 0;
                    102:        switch (req) {
                    103: 
                    104:        /*
                    105:         * TCP attaches to socket via PRU_ATTACH, reserving space,
                    106:         * and an internet control block.
                    107:         */
                    108:        case PRU_ATTACH:
                    109:                if (inp) {
                    110:                        error = EISCONN;
                    111:                        break;
                    112:                }
                    113:                error = tcp_attach(so);
                    114:                if (error)
                    115:                        break;
                    116:                if ((so->so_options & SO_LINGER) && so->so_linger == 0)
                    117:                        so->so_linger = TCP_LINGERTIME;
                    118:                tp = sototcpcb(so);
                    119:                break;
                    120: 
                    121:        /*
                    122:         * PRU_DETACH detaches the TCP protocol from the socket.
                    123:         * If the protocol state is non-embryonic, then can't
                    124:         * do this directly: have to initiate a PRU_DISCONNECT,
                    125:         * which may finish later; embryonic TCB's can just
                    126:         * be discarded here.
                    127:         */
                    128:        case PRU_DETACH:
                    129:                if (tp->t_state > TCPS_LISTEN)
                    130:                        tp = tcp_disconnect(tp);
                    131:                else
                    132:                        tp = tcp_close(tp);
                    133:                break;
                    134: 
                    135:        /*
                    136:         * Give the socket an address.
                    137:         */
                    138:        case PRU_BIND:
                    139:                error = in_pcbbind(inp, nam);
                    140:                if (error)
                    141:                        break;
                    142:                break;
                    143: 
                    144:        /*
                    145:         * Prepare to accept connections.
                    146:         */
                    147:        case PRU_LISTEN:
                    148:                if (inp->inp_lport == 0)
                    149:                        error = in_pcbbind(inp, (struct mbuf *)0);
                    150:                if (error == 0)
                    151:                        tp->t_state = TCPS_LISTEN;
                    152:                break;
                    153: 
                    154:        /*
                    155:         * Initiate connection to peer.
                    156:         * Create a template for use in transmissions on this connection.
                    157:         * Enter SYN_SENT state, and mark socket as connecting.
                    158:         * Start keep-alive timer, and seed output sequence space.
                    159:         * Send initial segment on connection.
                    160:         */
                    161:        case PRU_CONNECT:
                    162:                if (inp->inp_lport == 0) {
                    163:                        error = in_pcbbind(inp, (struct mbuf *)0);
                    164:                        if (error)
                    165:                                break;
                    166:                }
                    167:                error = in_pcbconnect(inp, nam);
                    168:                if (error)
                    169:                        break;
                    170:                tp->t_template = tcp_template(tp);
                    171:                if (tp->t_template == 0) {
                    172:                        in_pcbdisconnect(inp);
                    173:                        error = ENOBUFS;
                    174:                        break;
                    175:                }
                    176:                soisconnecting(so);
                    177:                tcpstat.tcps_connattempt++;
                    178:                tp->t_state = TCPS_SYN_SENT;
                    179:                tp->t_timer[TCPT_KEEP] = TCPTV_KEEP_INIT;
                    180:                tp->iss = tcp_iss; tcp_iss += TCP_ISSINCR/2;
                    181:                tcp_sendseqinit(tp);
                    182:                error = tcp_output(tp);
                    183:                break;
                    184: 
                    185:        /*
                    186:         * Create a TCP connection between two sockets.
                    187:         */
                    188:        case PRU_CONNECT2:
                    189:                error = EOPNOTSUPP;
                    190:                break;
                    191: 
                    192:        /*
                    193:         * Initiate disconnect from peer.
                    194:         * If connection never passed embryonic stage, just drop;
                    195:         * else if don't need to let data drain, then can just drop anyways,
                    196:         * else have to begin TCP shutdown process: mark socket disconnecting,
                    197:         * drain unread data, state switch to reflect user close, and
                    198:         * send segment (e.g. FIN) to peer.  Socket will be really disconnected
                    199:         * when peer sends FIN and acks ours.
                    200:         *
                    201:         * SHOULD IMPLEMENT LATER PRU_CONNECT VIA REALLOC TCPCB.
                    202:         */
                    203:        case PRU_DISCONNECT:
                    204:                tp = tcp_disconnect(tp);
                    205:                break;
                    206: 
                    207:        /*
                    208:         * Accept a connection.  Essentially all the work is
                    209:         * done at higher levels; just return the address
                    210:         * of the peer, storing through addr.
                    211:         */
                    212:        case PRU_ACCEPT: {
                    213:                struct sockaddr_in *sin = mtod(nam, struct sockaddr_in *);
                    214: 
                    215:                nam->m_len = sizeof (struct sockaddr_in);
                    216:                sin->sin_family = AF_INET;
                    217:                sin->sin_len = sizeof(*sin);
                    218:                sin->sin_port = inp->inp_fport;
                    219:                sin->sin_addr = inp->inp_faddr;
                    220:                break;
                    221:                }
                    222: 
                    223:        /*
                    224:         * Mark the connection as being incapable of further output.
                    225:         */
                    226:        case PRU_SHUTDOWN:
                    227:                socantsendmore(so);
                    228:                tp = tcp_usrclosed(tp);
                    229:                if (tp)
                    230:                        error = tcp_output(tp);
                    231:                break;
                    232: 
                    233:        /*
                    234:         * After a receive, possibly send window update to peer.
                    235:         */
                    236:        case PRU_RCVD:
                    237:                (void) tcp_output(tp);
                    238:                break;
                    239: 
                    240:        /*
                    241:         * Do a send by putting data in output queue and updating urgent
                    242:         * marker if URG set.  Possibly send more data.
                    243:         */
                    244:        case PRU_SEND:
                    245:                sbappend(&so->so_snd, m);
                    246:                error = tcp_output(tp);
                    247:                break;
                    248: 
                    249:        /*
                    250:         * Abort the TCP.
                    251:         */
                    252:        case PRU_ABORT:
                    253:                tp = tcp_drop(tp, ECONNABORTED);
                    254:                break;
                    255: 
                    256:        case PRU_SENSE:
                    257:                ((struct stat *) m)->st_blksize = so->so_snd.sb_hiwat;
                    258:                (void) splx(s);
                    259:                return (0);
                    260: 
                    261:        case PRU_RCVOOB:
                    262:                if ((so->so_oobmark == 0 &&
                    263:                    (so->so_state & SS_RCVATMARK) == 0) ||
                    264:                    so->so_options & SO_OOBINLINE ||
                    265:                    tp->t_oobflags & TCPOOB_HADDATA) {
                    266:                        error = EINVAL;
                    267:                        break;
                    268:                }
                    269:                if ((tp->t_oobflags & TCPOOB_HAVEDATA) == 0) {
                    270:                        error = EWOULDBLOCK;
                    271:                        break;
                    272:                }
                    273:                m->m_len = 1;
                    274:                *mtod(m, caddr_t) = tp->t_iobc;
                    275:                if (((int)nam & MSG_PEEK) == 0)
                    276:                        tp->t_oobflags ^= (TCPOOB_HAVEDATA | TCPOOB_HADDATA);
                    277:                break;
                    278: 
                    279:        case PRU_SENDOOB:
                    280:                if (sbspace(&so->so_snd) < -512) {
                    281:                        m_freem(m);
                    282:                        error = ENOBUFS;
                    283:                        break;
                    284:                }
                    285:                /*
                    286:                 * According to RFC961 (Assigned Protocols),
                    287:                 * the urgent pointer points to the last octet
                    288:                 * of urgent data.  We continue, however,
                    289:                 * to consider it to indicate the first octet
                    290:                 * of data past the urgent section.
                    291:                 * Otherwise, snd_up should be one lower.
                    292:                 */
                    293:                sbappend(&so->so_snd, m);
                    294:                tp->snd_up = tp->snd_una + so->so_snd.sb_cc;
                    295:                tp->t_force = 1;
                    296:                error = tcp_output(tp);
                    297:                tp->t_force = 0;
                    298:                break;
                    299: 
                    300:        case PRU_SOCKADDR:
                    301:                in_setsockaddr(inp, nam);
                    302:                break;
                    303: 
                    304:        case PRU_PEERADDR:
                    305:                in_setpeeraddr(inp, nam);
                    306:                break;
                    307: 
                    308:        /*
                    309:         * TCP slow timer went off; going through this
                    310:         * routine for tracing's sake.
                    311:         */
                    312:        case PRU_SLOWTIMO:
                    313:                tp = tcp_timers(tp, (int)nam);
                    314:                req |= (int)nam << 8;           /* for debug's sake */
                    315:                break;
                    316: 
                    317:        default:
                    318:                panic("tcp_usrreq");
                    319:        }
                    320:        if (tp && (so->so_options & SO_DEBUG))
                    321:                tcp_trace(TA_USER, ostate, tp, (struct tcpiphdr *)0, req);
                    322:        splx(s);
                    323:        return (error);
                    324: }
                    325: 
                    326: tcp_ctloutput(op, so, level, optname, mp)
                    327:        int op;
                    328:        struct socket *so;
                    329:        int level, optname;
                    330:        struct mbuf **mp;
                    331: {
                    332:        int error = 0;
                    333:        struct inpcb *inp = sotoinpcb(so);
                    334:        register struct tcpcb *tp = intotcpcb(inp);
                    335:        register struct mbuf *m;
                    336: 
                    337:        if (level != IPPROTO_TCP)
                    338:                return (ip_ctloutput(op, so, level, optname, mp));
                    339: 
                    340:        switch (op) {
                    341: 
                    342:        case PRCO_SETOPT:
                    343:                m = *mp;
                    344:                switch (optname) {
                    345: 
                    346:                case TCP_NODELAY:
                    347:                        if (m == NULL || m->m_len < sizeof (int))
                    348:                                error = EINVAL;
                    349:                        else if (*mtod(m, int *))
                    350:                                tp->t_flags |= TF_NODELAY;
                    351:                        else
                    352:                                tp->t_flags &= ~TF_NODELAY;
                    353:                        break;
                    354: 
                    355:                case TCP_MAXSEG:        /* not yet */
                    356:                default:
                    357:                        error = EINVAL;
                    358:                        break;
                    359:                }
                    360:                if (m)
                    361:                        (void) m_free(m);
                    362:                break;
                    363: 
                    364:        case PRCO_GETOPT:
                    365:                *mp = m = m_get(M_WAIT, MT_SOOPTS);
                    366:                m->m_len = sizeof(int);
                    367: 
                    368:                switch (optname) {
                    369:                case TCP_NODELAY:
                    370:                        *mtod(m, int *) = tp->t_flags & TF_NODELAY;
                    371:                        break;
                    372:                case TCP_MAXSEG:
                    373:                        *mtod(m, int *) = tp->t_maxseg;
                    374:                        break;
                    375:                default:
                    376:                        error = EINVAL;
                    377:                        break;
                    378:                }
                    379:                break;
                    380:        }
                    381:        return (error);
                    382: }
                    383: 
                    384: u_long tcp_sendspace = 1024*4;
                    385: u_long tcp_recvspace = 1024*4;
                    386: 
                    387: /*
                    388:  * Attach TCP protocol to socket, allocating
                    389:  * internet protocol control block, tcp control block,
                    390:  * bufer space, and entering LISTEN state if to accept connections.
                    391:  */
                    392: tcp_attach(so)
                    393:        struct socket *so;
                    394: {
                    395:        register struct tcpcb *tp;
                    396:        struct inpcb *inp;
                    397:        int error;
                    398: 
                    399:        if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) {
                    400:                error = soreserve(so, tcp_sendspace, tcp_recvspace);
                    401:                if (error)
                    402:                        return (error);
                    403:        }
                    404:        error = in_pcballoc(so, &tcb);
                    405:        if (error)
                    406:                return (error);
                    407:        inp = sotoinpcb(so);
                    408:        tp = tcp_newtcpcb(inp);
                    409:        if (tp == 0) {
                    410:                int nofd = so->so_state & SS_NOFDREF;   /* XXX */
                    411: 
                    412:                so->so_state &= ~SS_NOFDREF;    /* don't free the socket yet */
                    413:                in_pcbdetach(inp);
                    414:                so->so_state |= nofd;
                    415:                return (ENOBUFS);
                    416:        }
                    417:        tp->t_state = TCPS_CLOSED;
                    418:        return (0);
                    419: }
                    420: 
                    421: /*
                    422:  * Initiate (or continue) disconnect.
                    423:  * If embryonic state, just send reset (once).
                    424:  * If in ``let data drain'' option and linger null, just drop.
                    425:  * Otherwise (hard), mark socket disconnecting and drop
                    426:  * current input data; switch states based on user close, and
                    427:  * send segment to peer (with FIN).
                    428:  */
                    429: struct tcpcb *
                    430: tcp_disconnect(tp)
                    431:        register struct tcpcb *tp;
                    432: {
                    433:        struct socket *so = tp->t_inpcb->inp_socket;
                    434: 
                    435:        if (tp->t_state < TCPS_ESTABLISHED)
                    436:                tp = tcp_close(tp);
                    437:        else if ((so->so_options & SO_LINGER) && so->so_linger == 0)
                    438:                tp = tcp_drop(tp, 0);
                    439:        else {
                    440:                soisdisconnecting(so);
                    441:                sbflush(&so->so_rcv);
                    442:                tp = tcp_usrclosed(tp);
                    443:                if (tp)
                    444:                        (void) tcp_output(tp);
                    445:        }
                    446:        return (tp);
                    447: }
                    448: 
                    449: /*
                    450:  * User issued close, and wish to trail through shutdown states:
                    451:  * if never received SYN, just forget it.  If got a SYN from peer,
                    452:  * but haven't sent FIN, then go to FIN_WAIT_1 state to send peer a FIN.
                    453:  * If already got a FIN from peer, then almost done; go to LAST_ACK
                    454:  * state.  In all other cases, have already sent FIN to peer (e.g.
                    455:  * after PRU_SHUTDOWN), and just have to play tedious game waiting
                    456:  * for peer to send FIN or not respond to keep-alives, etc.
                    457:  * We can let the user exit from the close as soon as the FIN is acked.
                    458:  */
                    459: struct tcpcb *
                    460: tcp_usrclosed(tp)
                    461:        register struct tcpcb *tp;
                    462: {
                    463: 
                    464:        switch (tp->t_state) {
                    465: 
                    466:        case TCPS_CLOSED:
                    467:        case TCPS_LISTEN:
                    468:        case TCPS_SYN_SENT:
                    469:                tp->t_state = TCPS_CLOSED;
                    470:                tp = tcp_close(tp);
                    471:                break;
                    472: 
                    473:        case TCPS_SYN_RECEIVED:
                    474:        case TCPS_ESTABLISHED:
                    475:                tp->t_state = TCPS_FIN_WAIT_1;
                    476:                break;
                    477: 
                    478:        case TCPS_CLOSE_WAIT:
                    479:                tp->t_state = TCPS_LAST_ACK;
                    480:                break;
                    481:        }
                    482:        if (tp && tp->t_state >= TCPS_FIN_WAIT_2)
                    483:                soisdisconnected(tp->t_inpcb->inp_socket);
                    484:        return (tp);
                    485: }

unix.superglobalmegacorp.com

This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.