/* tcp_input.c 1.43 81/12/21 */ #include "../h/param.h" #include "../h/systm.h" #include "../h/mbuf.h" #include "../h/protosw.h" #include "../h/socket.h" #include "../h/socketvar.h" #include "../net/in.h" #include "../net/in_pcb.h" #include "../net/in_systm.h" #include "../net/if.h" #include "../net/ip.h" #include "../net/ip_var.h" #include "../net/tcp.h" #include "../net/tcp_fsm.h" #include "../net/tcp_seq.h" #include "../net/tcp_timer.h" #include "../net/tcp_var.h" #include "../net/tcpip.h" #include "../net/tcp_debug.h" #include "../errno.h" int tcpcksum = 1; struct sockaddr_in tcp_in = { AF_INET }; struct tcpiphdr tcp_saveti; struct tcpcb *tcp_newtcpcb(); /* * TCP input routine, follows pages 65-76 of the * protocol specification dated September, 1981 very closely. */ tcp_input(m0) struct mbuf *m0; { register struct tcpiphdr *ti; struct inpcb *inp; register struct mbuf *m; int len, tlen, off; register struct tcpcb *tp; register int tiflags; struct socket *so; int todrop, acked; short ostate; COUNT(TCP_INPUT); /* * Get IP and TCP header together in first mbuf. * Note: IP leaves IP header in first mbuf. */ m = m0; ti = mtod(m, struct tcpiphdr *); if (((struct ip *)ti)->ip_hl > (sizeof (struct ip) >> 2)) ip_stripoptions((struct ip *)ti, (struct mbuf *)0); if (m->m_len < sizeof (struct tcpiphdr)) { if (m_pullup(m, sizeof (struct tcpiphdr)) == 0) { tcpstat.tcps_hdrops++; goto drop; } ti = mtod(m, struct tcpiphdr *); } /* * Checksum extended TCP header and data. */ tlen = ((struct ip *)ti)->ip_len; len = sizeof (struct ip) + tlen; if (tcpcksum) { ti->ti_next = ti->ti_prev = 0; ti->ti_x1 = 0; ti->ti_len = (u_short)tlen; #if vax ti->ti_len = htons(ti->ti_len); #endif if (ti->ti_sum = in_cksum(m, len)) { tcpstat.tcps_badsum++; printf("tcp cksum %x\n", ti->ti_sum); goto drop; } } /* * Check that TCP offset makes sense, * process TCP options and adjust length. */ off = ti->ti_off << 2; if (off < sizeof (struct tcphdr) || off > tlen) { tcpstat.tcps_badoff++; goto drop; } ti->ti_len = tlen - off; #if 0 if (off > sizeof (struct tcphdr)) tcp_options(ti); #endif tiflags = ti->ti_flags; #if vax /* * Convert TCP protocol specific fields to host format. */ ti->ti_seq = ntohl(ti->ti_seq); ti->ti_ack = ntohl(ti->ti_ack); ti->ti_win = ntohs(ti->ti_win); ti->ti_urp = ntohs(ti->ti_urp); #endif /* * Locate pcb for segment. */ inp = in_pcblookup (&tcb, ti->ti_src, ti->ti_sport, ti->ti_dst, ti->ti_dport); /* * If the state is CLOSED (i.e., TCB does not exist) then * all data in the incoming segment is discarded. */ if (inp == 0) { printf("cant find inp\n"); goto dropwithreset; } tp = intotcpcb(inp); if (tp == 0) { printf("tp is 0\n"); goto dropwithreset; } so = inp->inp_socket; if (so->so_options & SO_DEBUG) { ostate = tp->t_state; tcp_saveti = *ti; } /* * Segment received on connection. * Reset idle time and keep-alive timer. */ tp->t_idle = 0; tp->t_timer[TCPT_KEEP] = TCPTV_KEEP; /* * Calculate amount of space in receive window, * and then do TCP input processing. */ tp->rcv_wnd = sbspace(&so->so_rcv); if (tp->rcv_wnd < 0) tp->rcv_wnd = 0; switch (tp->t_state) { /* * If the state is LISTEN then ignore segment if it contains an RST. * If the segment contains an ACK then it is bad and send a RST. * If it does not contain a SYN then it is not interesting; drop it. * Otherwise initialize tp->rcv_nxt, and tp->irs, select an initial * tp->iss, and send a segment: * * Also initialize tp->snd_nxt to tp->iss+1 and tp->snd_una to tp->iss. * Fill in remote peer address fields if not previously specified. * Enter SYN_RECEIVED state, and process any other fields of this * segment in this state. */ case TCPS_LISTEN: if (tiflags & TH_RST) goto drop; if (tiflags & TH_ACK) { printf("contains ACK\n"); goto dropwithreset; } if ((tiflags & TH_SYN) == 0) { printf("no syn\n"); goto drop; } tcp_in.sin_addr = ti->ti_src; tcp_in.sin_port = ti->ti_sport; if (in_pcbconnect(inp, (struct sockaddr *)&tcp_in)) { printf("pcb cant connect\n"); goto drop; } tp->t_template = tcp_template(tp); if (tp->t_template == 0) { printf("can't get template\n"); in_pcbdisconnect(inp); goto drop; } tp->iss = tcp_iss; tcp_iss += TCP_ISSINCR/2; tp->irs = ti->ti_seq; tcp_sendseqinit(tp); tcp_rcvseqinit(tp); tp->t_state = TCPS_SYN_RECEIVED; tp->t_timer[TCPT_KEEP] = TCPTV_KEEP; goto trimthenstep6; /* * If the state is SYN_SENT: * if seg contains an ACK, but not for our SYN, drop the input. * if seg contains a RST, then drop the connection. * if seg does not contain SYN, then drop it. * Otherwise this is an acceptable SYN segment * initialize tp->rcv_nxt and tp->irs * if seg contains ack then advance tp->snd_una * if SYN has been acked change to ESTABLISHED else SYN_RCVD state * arrange for segment to be acked (eventually) * continue processing rest of data/controls, beginning with URG */ case TCPS_SYN_SENT: if ((tiflags & TH_ACK) && (SEQ_LEQ(ti->ti_ack, tp->iss) || SEQ_GT(ti->ti_ack, tp->snd_max))) goto dropwithreset; if (tiflags & TH_RST) { if (tiflags & TH_ACK) tcp_drop(tp, ECONNREFUSED); goto drop; } if ((tiflags & TH_SYN) == 0) goto drop; tp->snd_una = ti->ti_ack; tp->t_timer[TCPT_REXMT] = 0; tp->irs = ti->ti_seq; tcp_rcvseqinit(tp); tp->t_flags |= TF_ACKNOW; if (SEQ_GT(tp->snd_una, tp->iss)) { so->so_state |= SS_CONNAWAITING; soisconnected(so); tp->t_state = TCPS_ESTABLISHED; (void) tcp_reass(tp, (struct tcpiphdr *)0); } else tp->t_state = TCPS_SYN_RECEIVED; goto trimthenstep6; trimthenstep6: /* * Advance ti->ti_seq to correspond to first data byte. * If data, trim to stay within window, * dropping FIN if necessary. */ ti->ti_seq++; if (ti->ti_len > tp->rcv_wnd) { todrop = ti->ti_len - tp->rcv_wnd; m_adj(m, -todrop); ti->ti_len = tp->rcv_wnd; ti->ti_flags &= ~TH_FIN; } tp->snd_wl1 = ti->ti_seq - 1; goto step6; } /* * States other than LISTEN or SYN_SENT. * First check that at least some bytes of segment are within * receive window. */ if (tp->rcv_wnd == 0) { /* * If window is closed can only take segments at * window edge, and have to drop data and PUSH from * incoming segments. */ if (tp->rcv_nxt != ti->ti_seq) { printf("wnd closed, not at edge\n"); goto dropafterack; } if (ti->ti_len > 0) { ti->ti_len = 0; ti->ti_flags &= ~(TH_PUSH|TH_FIN); } } else { /* * If segment begins before rcv_nxt, drop leading * data (and SYN); if nothing left, just ack. */ if (SEQ_GT(tp->rcv_nxt, ti->ti_seq)) { todrop = tp->rcv_nxt - ti->ti_seq; if (tiflags & TH_SYN) { ti->ti_seq++; if (ti->ti_urp > 1) ti->ti_urp--; else tiflags &= ~TH_URG; todrop--; } if (todrop > ti->ti_len) { printf("window open but outside\n"); goto dropafterack; } m_adj(m, todrop); ti->ti_seq += todrop; ti->ti_len -= todrop; if (ti->ti_urp > todrop) ti->ti_urp -= todrop; else { tiflags &= ~TH_URG; /* ti->ti_flags &= ~TH_URG; */ /* ti->ti_urp = 0; */ } /* tiflags &= ~TH_SYN; */ /* ti->ti_flags &= ~TH_SYN; */ } /* * If segment ends after window, drop trailing data * (and PUSH and FIN); if nothing left, just ACK. */ if (SEQ_GT(ti->ti_seq+ti->ti_len, tp->rcv_nxt+tp->rcv_wnd)) { todrop = ti->ti_seq+ti->ti_len - (tp->rcv_nxt+tp->rcv_wnd); if (todrop > ti->ti_len) { printf("segment outside window\n"); goto dropafterack; } m_adj(m, -todrop); ti->ti_len -= todrop; ti->ti_flags &= ~(TH_PUSH|TH_FIN); } } /* * If the RST bit is set examine the state: * SYN_RECEIVED STATE: * If passive open, return to LISTEN state. * If active open, inform user that connection was refused. * ESTABLISHED, FIN_WAIT_1, FIN_WAIT2, CLOSE_WAIT STATES: * Inform user that connection was reset, and close tcb. * CLOSING, LAST_ACK, TIME_WAIT STATES * Close the tcb. */ if (tiflags&TH_RST) switch (tp->t_state) { case TCPS_SYN_RECEIVED: if (inp->inp_socket->so_options & SO_ACCEPTCONN) { /* a miniature tcp_close, but invisible to user */ (void) m_free(dtom(tp->t_template)); (void) m_free(dtom(tp)); inp->inp_ppcb = 0; tp = tcp_newtcpcb(inp); tp->t_state = TCPS_LISTEN; goto drop; } tcp_drop(tp, ECONNREFUSED); goto drop; case TCPS_ESTABLISHED: case TCPS_FIN_WAIT_1: case TCPS_FIN_WAIT_2: case TCPS_CLOSE_WAIT: tcp_drop(tp, ECONNRESET); goto drop; case TCPS_CLOSING: case TCPS_LAST_ACK: case TCPS_TIME_WAIT: tcp_close(tp); goto drop; } /* * If a SYN is in the window, then this is an * error and we send an RST and drop the connection. */ if (tiflags & TH_SYN) { tcp_drop(tp, ECONNRESET); goto dropwithreset; } /* * If the ACK bit is off we drop the segment and return. */ if ((tiflags & TH_ACK) == 0) goto drop; /* * Ack processing. */ switch (tp->t_state) { /* * In SYN_RECEIVED state if the ack ACKs our SYN then enter * ESTABLISHED state and continue processing, othewise * send an RST. */ case TCPS_SYN_RECEIVED: if (SEQ_GT(tp->snd_una, ti->ti_ack) || SEQ_GT(ti->ti_ack, tp->snd_max)) goto dropwithreset; tp->snd_una++; /* SYN acked */ tp->t_timer[TCPT_REXMT] = 0; so->so_state |= SS_CONNAWAITING; soisconnected(so); tp->t_state = TCPS_ESTABLISHED; (void) tcp_reass(tp, (struct tcpiphdr *)0); tp->snd_wl1 = ti->ti_seq - 1; /* fall into ... */ /* * In ESTABLISHED state: drop duplicate ACKs; ACK out of range * ACKs. If the ack is in the range * tp->snd_una < ti->ti_ack <= tp->snd_max * then advance tp->snd_una to ti->ti_ack and drop * data from the retransmission queue. If this ACK reflects * more up to date window information we update our window information. */ case TCPS_ESTABLISHED: case TCPS_FIN_WAIT_1: case TCPS_FIN_WAIT_2: case TCPS_CLOSE_WAIT: case TCPS_CLOSING: case TCPS_LAST_ACK: case TCPS_TIME_WAIT: #define ourfinisacked (acked > 0) if (SEQ_LEQ(ti->ti_ack, tp->snd_una)) break; if (SEQ_GT(ti->ti_ack, tp->snd_max)) { printf("ack > snd_max\n"); goto dropafterack; } acked = ti->ti_ack - tp->snd_una; if (acked >= so->so_snd.sb_cc) { acked -= so->so_snd.sb_cc; tp->snd_wnd -= so->so_snd.sb_cc; /* if acked > 0 our FIN is acked */ sbdrop(&so->so_snd, so->so_snd.sb_cc); tp->t_timer[TCPT_REXMT] = 0; } else { if (acked) { sbdrop(&so->so_snd, acked); tp->snd_wnd -= acked; acked = 0; } TCPT_RANGESET(tp->t_timer[TCPT_REXMT], tcp_beta * tp->t_srtt, TCPTV_MIN, TCPTV_MAX); } tp->snd_una = ti->ti_ack; /* * If transmit timer is running and timed sequence * number was acked, update smoothed round trip time. */ if (tp->t_rtt && SEQ_GT(ti->ti_ack, tp->t_rtseq)) { if (tp->t_srtt == 0) tp->t_srtt = tp->t_rtt; else tp->t_srtt = tcp_alpha * tp->t_srtt + (1 - tcp_alpha) * tp->t_rtt; tp->t_rtt = 0; } switch (tp->t_state) { /* * In FIN_WAIT_1 STATE in addition to the processing * for the ESTABLISHED state if our FIN is now acknowledged * then enter FIN_WAIT_2. */ case TCPS_FIN_WAIT_1: if (ourfinisacked) tp->t_state = TCPS_FIN_WAIT_2; break; /* * In CLOSING STATE in addition to the processing for * the ESTABLISHED state if the ACK acknowledges our FIN * then enter the TIME-WAIT state, otherwise ignore * the segment. */ case TCPS_CLOSING: if (ourfinisacked) { tp->t_state = TCPS_TIME_WAIT; tcp_canceltimers(tp); tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL; soisdisconnected(so); } break; /* * The only thing that can arrive in LAST_ACK state * is an acknowledgment of our FIN. If our FIN is now * acknowledged, delete the TCB, enter the closed state * and return. */ case TCPS_LAST_ACK: if (ourfinisacked) tcp_close(tp); goto drop; /* * In TIME_WAIT state the only thing that should arrive * is a retransmission of the remote FIN. Acknowledge * it and restart the finack timer. */ case TCPS_TIME_WAIT: tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL; goto dropafterack; } #undef ourfinisacked } step6: /* * Update window information. */ if (SEQ_LT(tp->snd_wl1, ti->ti_seq) || tp->snd_wl1==ti->ti_seq && SEQ_LEQ(tp->snd_wl2,ti->ti_ack)) { tp->snd_wnd = ti->ti_win; tp->snd_wl1 = ti->ti_seq; tp->snd_wl2 = ti->ti_ack; if (tp->snd_wnd > 0) tp->t_timer[TCPT_PERSIST] = 0; } /* * If an URG bit is set in the segment and is greater than the * current known urgent pointer, then signal the user that the * remote side has out of band data. This should not happen * in CLOSE_WAIT, CLOSING, LAST-ACK or TIME_WAIT STATES since * a FIN has been received from the remote side. In these states * we ignore the URG. */ if ((tiflags & TH_URG) == 0 && TCPS_HAVERCVDFIN(tp->t_state) == 0) if (SEQ_GT(ti->ti_urp, tp->rcv_up)) { tp->rcv_up = ti->ti_urp; #if 0 sohasoutofband(so); /* XXX */ #endif } /* * Process the segment text, merging it into the TCP sequencing queue, * and arranging for acknowledgment of receipt if necessary. * This process logically involves adjusting tp->rcv_wnd as data * is presented to the user (this happens in tcp_usrreq.c, * case PRU_RCVD). If a FIN has already been received on this * connection then we just ignore the text. */ if ((ti->ti_len || (tiflags&TH_FIN)) && TCPS_HAVERCVDFIN(tp->t_state) == 0) { off += sizeof (struct ip); /* drop IP header */ m->m_off += off; m->m_len -= off; tiflags = tcp_reass(tp, ti); { extern tcpdelack; if (tcpdelack) tp->t_flags |= TF_DELACK; else tp->t_flags |= TF_ACKNOW; /* XXX TF_DELACK */ } } else { m_freem(m); tiflags &= ~TH_FIN; } /* * If FIN is received ACK the FIN and let the user know * that the connection is closing. */ if (tiflags & TH_FIN) { if (TCPS_HAVERCVDFIN(tp->t_state) == 0) { socantrcvmore(so); tp->t_flags |= TF_ACKNOW; tp->rcv_nxt++; } switch (tp->t_state) { /* * In SYN_RECEIVED and ESTABLISHED STATES * enter the CLOSE_WAIT state. */ case TCPS_SYN_RECEIVED: case TCPS_ESTABLISHED: tp->t_state = TCPS_CLOSE_WAIT; break; /* * If still in FIN_WAIT_1 STATE FIN has not been acked so * enter the CLOSING state. */ case TCPS_FIN_WAIT_1: tp->t_state = TCPS_CLOSING; break; /* * In FIN_WAIT_2 state enter the TIME_WAIT state, * starting the time-wait timer, turning off the other * standard timers. */ case TCPS_FIN_WAIT_2: tp->t_state = TCPS_TIME_WAIT; tcp_canceltimers(tp); tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL; soisdisconnected(so); break; /* * In TIME_WAIT state restart the 2 MSL time_wait timer. */ case TCPS_TIME_WAIT: tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL; break; } } if (so->so_options & SO_DEBUG) tcp_trace(TA_INPUT, ostate, tp, &tcp_saveti, 0); /* * Return any desired output. */ tcp_output(tp); return; dropafterack: /* * Generate an ACK dropping incoming segment. * Make ACK reflect our state. */ if (tiflags & TH_RST) goto drop; tcp_respond(ti, tp->rcv_nxt, tp->snd_nxt, TH_ACK); return; dropwithreset: /* * Generate a RST, dropping incoming segment. * Make ACK acceptable to originator of segment. */ if (tiflags & TH_RST) goto drop; if (tiflags & TH_ACK) tcp_respond(ti, (tcp_seq)0, ti->ti_ack, TH_RST); else { if (tiflags & TH_SYN) ti->ti_len++; tcp_respond(ti, ti->ti_seq+ti->ti_len, (tcp_seq)0, TH_RST|TH_ACK); } return; drop: printf("drop\n"); /* * Drop space held by incoming segment and return. */ m_freem(m); return; } /* * Insert segment ti into reassembly queue of tcp with * control block tp. Return TH_FIN if reassembly now includes * a segment with FIN. */ tcp_reass(tp, ti) register struct tcpcb *tp; register struct tcpiphdr *ti; { register struct tcpiphdr *q; struct socket *so = tp->t_inpcb->inp_socket; struct mbuf *m; int flags; COUNT(TCP_REASS); /* * Call with ti==0 after become established to * force pre-ESTABLISHED data up to user socket. */ if (ti == 0) goto present; /* * Find a segment which begins after this one does. */ for (q = tp->seg_next; q != (struct tcpiphdr *)tp; q = (struct tcpiphdr *)q->ti_next) if (SEQ_GT(q->ti_seq, ti->ti_seq)) break; /* * If there is a preceding segment, it may provide some of * our data already. If so, drop the data from the incoming * segment. If it provides all of our data, drop us. */ if ((struct tcpiphdr *)q->ti_prev != (struct tcpiphdr *)tp) { register int i; q = (struct tcpiphdr *)(q->ti_prev); /* conversion to int (in i) handles seq wraparound */ i = q->ti_seq + q->ti_len - ti->ti_seq; if (i > 0) { if (i >= ti->ti_len) goto drop; m_adj(dtom(tp), i); ti->ti_len -= i; ti->ti_seq += i; } q = (struct tcpiphdr *)(q->ti_next); } /* * While we overlap succeeding segments trim them or, * if they are completely covered, dequeue them. */ while (q != (struct tcpiphdr *)tp && SEQ_GT(ti->ti_seq + ti->ti_len, q->ti_seq)) { register int i = (ti->ti_seq + ti->ti_len) - q->ti_seq; if (i < q->ti_len) { q->ti_len -= i; m_adj(dtom(q), i); break; } q = (struct tcpiphdr *)q->ti_next; m_freem(dtom(q->ti_prev)); remque(q->ti_prev); } /* * Stick new segment in its place. */ insque(ti, q->ti_prev); present: /* * Present data to user, advancing rcv_nxt through * completed sequence space. */ if (TCPS_HAVERCVDSYN(tp->t_state) == 0) return (0); ti = tp->seg_next; if (ti == (struct tcpiphdr *)tp || ti->ti_seq != tp->rcv_nxt) return (0); if (tp->t_state == TCPS_SYN_RECEIVED && ti->ti_len) return (0); do { tp->rcv_nxt += ti->ti_len; flags = ti->ti_flags & TH_FIN; remque(ti); m = dtom(ti); ti = (struct tcpiphdr *)ti->ti_next; if (so->so_state & SS_CANTRCVMORE) (void) m_freem(m); else sbappend(&so->so_rcv, m); } while (ti != (struct tcpiphdr *)tp && ti->ti_seq == tp->rcv_nxt); sorwakeup(so); return (flags); drop: m_freem(dtom(ti)); return (0); }