/* BSDI $Id: tcp_usrreq.c,v 1.6 1996/08/04 03:34:56 padmanab Exp hari $ */ /* * Copyright (c) 1982, 1986, 1988, 1993 * The Regents of the University of California. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. All advertising materials mentioning features or use of this software * must display the following acknowledgement: * This product includes software developed by the University of * California, Berkeley and its contributors. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)tcp_usrreq.c 8.2 (Berkeley) 1/3/94 */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef MOBILE #ifndef NOMIP #include #endif #include #include #endif /* * TCP protocol interface to socket abstraction. */ extern char *tcpstates[]; #ifdef TCP_STATS #include #endif /* * Process a TCP user request for TCP tb. If this is a send request * then m is the mbuf chain of send data. If this is a timer expiration * (called from the software clock routine), then timertype tells which timer. */ /*ARGSUSED*/ int tcp_usrreq(so, req, m, nam, control) struct socket *so; int req; struct mbuf *m, *nam, *control; { register struct inpcb *inp; register struct tcpcb *tp; int s; int error = 0; int ostate; if (req == PRU_CONTROL) return (in_control(so, (int)m, (caddr_t)nam, (struct ifnet *)control)); if (control && control->m_len) { m_freem(control); if (m) m_freem(m); return (EINVAL); } s = splnet(); inp = sotoinpcb(so); /* * When a TCP is attached to a socket, then there will be * a (struct inpcb) pointed at by the socket, and this * structure will point at a subsidary (struct tcpcb). */ if (inp == 0 && req != PRU_ATTACH) { splx(s); if (m && req != PRU_SENSE) m_freem(m); return (EINVAL); /* XXX */ } if (inp) { tp = intotcpcb(inp); /* WHAT IF TP IS 0? */ #ifdef KPROF tcp_acounts[tp->t_state][req]++; #endif ostate = tp->t_state; } else ostate = 0; switch (req) { /* * TCP attaches to socket via PRU_ATTACH, reserving space, * and an internet control block. */ case PRU_ATTACH: if (inp) { error = EISCONN; break; } error = tcp_attach(so); if (error) break; if ((so->so_options & SO_LINGER) && so->so_linger == 0) so->so_linger = TCP_LINGERTIME; tp = sototcpcb(so); break; /* * PRU_DETACH detaches the TCP protocol from the socket. * If the protocol state is non-embryonic, then can't * do this directly: have to initiate a PRU_DISCONNECT, * which may finish later; embryonic TCB's can just * be discarded here. */ case PRU_DETACH: if (tp->t_state > TCPS_LISTEN) tp = tcp_disconnect(tp); else tp = tcp_close(tp); break; /* * Give the socket an address. */ case PRU_BIND: error = in_pcbbind(inp, nam); if (error) break; break; /* * Prepare to accept connections. */ case PRU_LISTEN: if (inp->inp_lport == 0) error = in_pcbbind(inp, (struct mbuf *)0); if (error == 0) tp->t_state = TCPS_LISTEN; break; /* * Initiate connection to peer. * Create a template for use in transmissions on this connection. * Enter SYN_SENT state, and mark socket as connecting. * Start keep-alive timer, and seed output sequence space. * Send initial segment on connection. */ case PRU_CONNECT: if (inp->inp_lport == 0) { error = in_pcbbind(inp, (struct mbuf *)0); if (error) break; } error = in_pcbconnect(inp, nam); if (error) break; tp->t_template = tcp_template(tp); if (tp->t_template == 0) { in_pcbdisconnect(inp); error = ENOBUFS; break; } /* Compute window scaling to request. */ while (tp->request_r_scale < TCP_MAX_WINSHIFT && (TCP_MAXWIN << tp->request_r_scale) < so->so_rcv.sb_hiwat) tp->request_r_scale++; soisconnecting(so); tcpstat.tcps_connattempt++; tp->t_state = TCPS_SYN_SENT; tp->t_timer[TCPT_KEEP] = tcp_conntimeo; tp->iss = tcp_iss; tcp_iss += TCP_ISSINCR/4; tcp_sendseqinit(tp); error = tcp_output(tp); break; /* * Create a TCP connection between two sockets. */ case PRU_CONNECT2: error = EOPNOTSUPP; break; /* * Initiate disconnect from peer. * If connection never passed embryonic stage, just drop; * else if don't need to let data drain, then can just drop anyways, * else have to begin TCP shutdown process: mark socket disconnecting, * drain unread data, state switch to reflect user close, and * send segment (e.g. FIN) to peer. Socket will be really disconnected * when peer sends FIN and acks ours. * * SHOULD IMPLEMENT LATER PRU_CONNECT VIA REALLOC TCPCB. */ case PRU_DISCONNECT: tp = tcp_disconnect(tp); break; /* * Accept a connection. Essentially all the work is * done at higher levels; just return the address * of the peer, storing through addr. */ case PRU_ACCEPT: in_setpeeraddr(inp, nam); break; /* * Mark the connection as being incapable of further output. */ case PRU_SHUTDOWN: socantsendmore(so); tp = tcp_usrclosed(tp); if (tp) error = tcp_output(tp); break; /* * After a receive, possibly send window update to peer. */ case PRU_RCVD: (void) tcp_output(tp); break; /* * Do a send by putting data in output queue and updating urgent * marker if URG set. Possibly send more data. */ case PRU_SEND: sbappend(&so->so_snd, m); error = tcp_output(tp); break; /* * Abort the TCP. */ case PRU_ABORT: tp = tcp_drop(tp, ECONNABORTED); break; case PRU_SENSE: ((struct stat *) m)->st_blksize = so->so_snd.sb_hiwat; splx(s); return (0); case PRU_RCVOOB: if ((so->so_oobmark == 0 && (so->so_state & SS_RCVATMARK) == 0) || so->so_options & SO_OOBINLINE || tp->t_oobflags & TCPOOB_HADDATA) { error = EINVAL; break; } if ((tp->t_oobflags & TCPOOB_HAVEDATA) == 0) { error = EWOULDBLOCK; break; } m->m_len = 1; *mtod(m, caddr_t) = tp->t_iobc; if (((int)nam & MSG_PEEK) == 0) tp->t_oobflags ^= (TCPOOB_HAVEDATA | TCPOOB_HADDATA); break; case PRU_SENDOOB: if (sbspace(&so->so_snd) < -512) { m_freem(m); error = ENOBUFS; break; } /* * According to RFC961 (Assigned Protocols), * the urgent pointer points to the last octet * of urgent data. We continue, however, * to consider it to indicate the first octet * of data past the urgent section. * Otherwise, snd_up should be one lower. */ sbappend(&so->so_snd, m); tp->snd_up = tp->snd_una + so->so_snd.sb_cc; tp->t_force = 1; error = tcp_output(tp); tp->t_force = 0; break; case PRU_SOCKADDR: in_setsockaddr(inp, nam); break; case PRU_PEERADDR: in_setpeeraddr(inp, nam); break; /* * TCP slow timer went off; going through this * routine for tracing's sake. */ case PRU_SLOWTIMO: tp = tcp_timers(tp, (int)nam); req |= (int)nam << 8; /* for debug's sake */ break; default: panic("tcp_usrreq"); } if (tp && (so->so_options & SO_DEBUG)) tcp_trace(TA_USER, ostate, tp, (struct tcpiphdr *)0, req); splx(s); return (error); } #ifdef MOBILE extern struct ber_state ber; #ifdef BASESTATION extern snoop_state_t *snoopstate; #endif #ifdef LINKEMU extern int snoop_linkemu_enable; #endif #ifdef ELN int snoop_eln_enable = 0; #endif #ifdef NOSNOOPREXMT extern int snoop_rexmt_disable; #endif #ifdef SMART int smart_enable = 0; #endif #ifdef SMART_SNOOP extern int snoop_smart_enable; #endif /* SMART_SNOOP */ #ifdef PERFECT_ELN int perfect_eln_enable = 0; #endif #ifdef NEWRENO int newreno_enable = 0; #endif #endif /* MOBILE */ #ifdef ITCPEMU #include #endif #ifdef SACK extern int tcp_do_sack; #endif #ifdef DPC int advt_large_mss=0; #endif int tcp_ctloutput(op, so, level, optname, mp) int op; struct socket *so; int level, optname; struct mbuf **mp; { int error = 0, s; struct inpcb *inp; register struct tcpcb *tp; register struct mbuf *m; register int i; s = splnet(); inp = sotoinpcb(so); if (inp == NULL) { splx(s); if (op == PRCO_SETOPT && *mp) (void) m_free(*mp); return (ECONNRESET); } if (level != IPPROTO_TCP) { error = ip_ctloutput(op, so, level, optname, mp); splx(s); return (error); } tp = intotcpcb(inp); switch (op) { case PRCO_SETOPT: m = *mp; switch (optname) { case TCP_NODELAY: if (m == NULL || m->m_len < sizeof (int)) error = EINVAL; else if (*mtod(m, int *)) tp->t_flags |= TF_NODELAY; else tp->t_flags &= ~TF_NODELAY; break; case TCP_MAXSEG: if (m && (i = *mtod(m, int *)) > 0 && i <= tp->t_maxseg) tp->t_maxseg = i; else error = EINVAL; break; #ifdef MOBILE case TCP_SNOOP_ERRPROB0: i = *mtod(m, int *); ber.error_prob[0] = i; break; case TCP_SNOOP_ERRPROB1: i = *mtod(m, int *); ber.error_prob[1] = i; break; case TCP_SNOOP_BER_MODEL: i = *mtod(m, int *); ber.model = i; break; case TCP_SNOOP_TRANS0: i = *mtod(m, int *); ber.trans_perc[0] = i; break; case TCP_SNOOP_TRANS1: i = *mtod(m, int *); ber.trans_perc[1] = i; break; case TCP_SNOOP_TIMERGRAN: i = *mtod(m, int *); ber.time_granularity = i; break; case TCP_SNOOP_BURSTRATE: i = *mtod(m, int *); ber.burst_rate = i; break; case TCP_SNOOP_BER_DISABLE: i = *mtod(m, int *); ber.disable = i; break; #ifdef BASESTATION case TCP_SNOOP_DISABLE: i = *mtod(m, int *); snoopstate->disable = i; break; #endif #ifdef LINKEMU case TCP_SNOOP_LINKEMU_ENABLE: i = *mtod(m, int *); snoop_linkemu_enable = i; break; #endif #ifdef ELN case TCP_SNOOP_ELN_ENABLE: i = *mtod(m, int *); snoop_eln_enable = i; break; #endif #ifdef NOSNOOPREXMT case TCP_SNOOP_REXMT_DISABLE: i = *mtod(m, int *); snoop_rexmt_disable = i; break; #endif #ifdef SMART case TCP_SMART_ENABLE: i = *mtod(m, int *); smart_enable = i; break; #endif #ifdef SMART_SNOOP case TCP_SMART_SNOOP_ENABLE: i = *mtod(m, int *); snoop_smart_enable = i; break; #endif #ifdef PERFECT_ELN case TCP_PERFECT_ELN_ENABLE: i = *mtod(m, int *); perfect_eln_enable = i; break; #endif #ifdef NEWRENO case TCP_NEWRENO_ENABLE: i = *mtod(m, int *); newreno_enable = i; break; #endif #endif /* MOBILE */ #ifdef SACK case TCP_SACK_DISABLE: i = *mtod(m, int *); tp->sack_disable = i; break; case TCP_DO_SACK: i = *mtod(m, int *); tcp_do_sack = i; break; #endif #ifdef DPC case TCP_IGNORE_PEER: i = *mtod(m, int *); tp->ignore_peer = i; break; case TCP_PEER_MSS: i = *mtod(m, int *); tp->peer_mss = i; break; case TCP_NOTSTAMP: i = *mtod(m, int *); tp->notstamp = i; break; case TCP_ADVT_LARGE_MSS: i = *mtod(m, int *); advt_large_mss = i; break; #endif #ifdef ITCPEMU case ITCP_WIRED_SOCKET_SET: itcp_wired_socket = so; itcp_wired_conn_closed = 0; break; case ITCP_WLESS_SOCKET_SET: itcp_wless_socket = so; itcp_wired_conn_closed = 0; break; case ITCP_WIRED_SOCKET_CLR: itcp_wired_socket = NULL; itcp_wired_conn_closed = 0; break; case ITCP_WLESS_SOCKET_CLR: itcp_wless_socket = NULL; itcp_wired_conn_closed = 0; break; #endif /* ITCPEMU */ #ifdef TCP_STATS case TCP_ENABLE_STATS: { /* collection TCP-related stats */ u_long flags; if (m == NULL || m->m_len < sizeof (int)) error = EINVAL; else if (flags = *mtod(m, u_long *)) { tp->t_flags |= TF_ENABLE_STATS; tcp_stats_init(tp, flags); } else tp->t_flags &= ~TF_ENABLE_STATS; break; } #endif #ifdef DELACKS case TCP_DELACKS_ENABLE: { /* enable delayed acks */ i = *mtod(m, int *); tp->max_delack = i; break; } case TCP_SS_DELACK_OK: { /* okay to delay acks when peer is in slow start */ i = *mtod(m, int *); tp->ss_delack_ok = i; break; } #endif #ifdef TCP_USER_CONTROL case TCP_SND_CWND_SCALE: { /* scale the initial snd_cwnd */ i = *mtod(m, int *); tp->snd_cwnd_scale = i; break; } #endif default: error = ENOPROTOOPT; break; } if (m) (void) m_free(m); break; case PRCO_GETOPT: *mp = m = m_get(M_WAIT, MT_SOOPTS); m->m_len = sizeof(int); switch (optname) { case TCP_NODELAY: *mtod(m, int *) = tp->t_flags & TF_NODELAY; break; case TCP_MAXSEG: *mtod(m, int *) = tp->t_maxseg; break; #ifdef MOBILE case TCP_SNOOP_ERRPROB0: *mtod(m, int *) = ber.error_prob[0]; break; case TCP_SNOOP_ERRPROB1: *mtod(m, int *) = ber.error_prob[1]; break; case TCP_SNOOP_BER_MODEL: *mtod(m, int *) = ber.model; break; case TCP_SNOOP_TRANS0: *mtod(m, int *) = ber.trans_perc[0]; break; case TCP_SNOOP_TRANS1: *mtod(m, int *) = ber.trans_perc[1]; break; case TCP_SNOOP_TIMERGRAN: *mtod(m, int *) = ber.time_granularity; break; case TCP_SNOOP_BURSTRATE: *mtod(m, int *) = ber.burst_rate; break; case TCP_SNOOP_BER_DISABLE: *mtod(m, int *) = ber.disable; break; #ifdef BASESTATION case TCP_SNOOP_DISABLE: *mtod(m, int *) = snoopstate->disable; break; #endif #ifdef LINKEMU case TCP_SNOOP_LINKEMU_ENABLE: *mtod(m, int *) = snoop_linkemu_enable; break; #endif #ifdef ELN case TCP_SNOOP_ELN_ENABLE: *mtod(m, int *) = snoop_eln_enable; break; #endif #ifdef NOSNOOPREXMT case TCP_SNOOP_REXMT_DISABLE: *mtod(m, int *) = snoop_rexmt_disable; break; #endif #ifdef SMART case TCP_SMART_ENABLE: *mtod(m, int *) = smart_enable; break; #endif #ifdef SMART_SNOOP case TCP_SMART_SNOOP_ENABLE: *mtod(m, int *) = snoop_smart_enable; break; #endif #ifdef PERFECT_ELN case TCP_PERFECT_ELN_ENABLE: *mtod(m, int *) = perfect_eln_enable; break; #endif #ifdef NEWRENO case TCP_NEWRENO_ENABLE: *mtod(m, int *) = newreno_enable; break; #endif #endif /* MOBILE */ #ifdef SACK case TCP_SACK_DISABLE: *mtod(m, int *) = tp->sack_disable; break; case TCP_DO_SACK: *mtod(m, int *) = tcp_do_sack; break; #endif #ifdef DPC case TCP_IGNORE_PEER: *mtod(m, int *) = tp->ignore_peer; break; case TCP_PEER_MSS: *mtod(m, int *) = tp->peer_mss; break; case TCP_NOTSTAMP: *mtod(m, int *) = tp->notstamp; break; case TCP_ADVT_LARGE_MSS: *mtod(m, int *) = advt_large_mss; break; #endif #ifdef TCP_STATS case TCP_ENABLE_STATS: *mtod(m, int *) = tp->t_flags & TF_ENABLE_STATS; break; #endif #ifdef TCP_MONITOR case TCP_GET_SRTT: { int rttunit = 1000000/PR_SLOWHZ; *mtod(m, int *) = tp->t_srtt*rttunit/TCP_RTT_SCALE; break; } case TCP_GET_RTTVAR: *mtod(m, int *) = tp->t_rttvar; break; #endif #ifdef DELACKS case TCP_DELACKS_ENABLE: { /* enable delayed acks */ *mtod(m, int *) = tp->max_delack; break; } case TCP_SS_DELACK_OK: { /* okay to delay acks when peer is in slow start */ *mtod(m, int *) = tp->ss_delack_ok; break; } #endif #ifdef TCP_USER_CONTROL case TCP_SND_CWND_SCALE: { /* scale the initial snd_cwnd */ *mtod(m, int *) = tp->snd_cwnd_scale; break; } #endif default: error = ENOPROTOOPT; break; } break; } splx(s); return (error); } int *tcp_sysvars[] = TCPCTL_VARS; /* * Sysctl for tcp variables. */ tcp_sysctl(name, namelen, oldp, oldlenp, newp, newlen) int *name; u_int namelen; void *oldp; size_t *oldlenp; void *newp; size_t newlen; { if (name[0] < TCPCTL_MAXID) return (sysctl_int_arr(tcp_sysvars, name, namelen, oldp, oldlenp, newp, newlen)); return (EOPNOTSUPP); } /* * Attach TCP protocol to socket, allocating * internet protocol control block, tcp control block, * bufer space, and entering LISTEN state if to accept connections. */ int tcp_attach(so) struct socket *so; { register struct tcpcb *tp; struct inpcb *inp; int error; if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) { error = soreserve(so, (u_long)tcp_sendspace, (u_long)tcp_recvspace); if (error) return (error); } error = in_pcballoc(so, &tcb); if (error) return (error); inp = sotoinpcb(so); tp = tcp_newtcpcb(inp); if (tp == 0) { int nofd = so->so_state & SS_NOFDREF; /* XXX */ so->so_state &= ~SS_NOFDREF; /* don't free the socket yet */ in_pcbdetach(inp); so->so_state |= nofd; return (ENOBUFS); } tp->t_state = TCPS_CLOSED; return (0); } /* * Initiate (or continue) disconnect. * If embryonic state, just send reset (once). * If in ``let data drain'' option and linger null, just drop. * Otherwise (hard), mark socket disconnecting and drop * current input data; switch states based on user close, and * send segment to peer (with FIN). */ struct tcpcb * tcp_disconnect(tp) register struct tcpcb *tp; { struct socket *so = tp->t_inpcb->inp_socket; if (tp->t_state < TCPS_ESTABLISHED) tp = tcp_close(tp); else if ((so->so_options & SO_LINGER) && so->so_linger == 0) tp = tcp_drop(tp, 0); else { soisdisconnecting(so); sbflush(&so->so_rcv); tp = tcp_usrclosed(tp); if (tp) (void) tcp_output(tp); } return (tp); } /* * User issued close, and wish to trail through shutdown states: * if never received SYN, just forget it. If got a SYN from peer, * but haven't sent FIN, then go to FIN_WAIT_1 state to send peer a FIN. * If already got a FIN from peer, then almost done; go to LAST_ACK * state. In all other cases, have already sent FIN to peer (e.g. * after PRU_SHUTDOWN), and just have to play tedious game waiting * for peer to send FIN or not respond to keep-alives, etc. * We can let the user exit from the close as soon as the FIN is acked. */ struct tcpcb * tcp_usrclosed(tp) register struct tcpcb *tp; { switch (tp->t_state) { case TCPS_CLOSED: case TCPS_LISTEN: case TCPS_SYN_SENT: tp->t_state = TCPS_CLOSED; tp = tcp_close(tp); break; case TCPS_SYN_RECEIVED: case TCPS_ESTABLISHED: tp->t_state = TCPS_FIN_WAIT_1; break; case TCPS_CLOSE_WAIT: tp->t_state = TCPS_LAST_ACK; break; } if (tp && tp->t_state >= TCPS_FIN_WAIT_2) soisdisconnected(tp->t_inpcb->inp_socket); return (tp); }