/* * $RCSfile: tcp_input.c,v $ * * Derived from: * * Copyright (c) 1982, 1986 Regents of the University of California. * All rights reserved. * * Redistribution and use in source and binary forms are permitted * provided that this notice is preserved and that due credit is given * to the University of California at Berkeley. The name of the University * may not be used to endorse or promote products derived from this * software without specific prior written permission. This software * is provided ``as is'' without express or implied warranty. * * @(#)tcp_input.c 7.13+ (Berkeley) 11/13/87 * * Modified for x-kernel v3.3 12/10/90 * Modifications Copyright (c) 1993,1991,1990,1996 Arizona Board of Regents * * $Log: tcp_input.c,v $ * Revision 1.2 1996/01/29 22:32:07 slm * Updated copyright and version. * * Revision 1.1 1995/07/28 22:15:41 slm * Initial revision * * Revision 1.25.1.4.1.2 1994/12/02 18:24:18 hkaram * David's TCP * * Revision 1.25.1.3 1994/05/02 01:09:13 davidm * (tcp_input): tcpSemSignal() changed into tcpVAll (with obvious * optimization). Otherwise, semaphore count can reach unbounded * values. * * Also changed two tracing level 2 to TR_MAJOR_EVENTS---the idea being * that with TR_GROSS_EVENTS, you get messages only in the case of * retransmission etc (abnormal events). * * Revision 1.25.1.2 1994/04/26 17:31:20 menze * Added cast to bzero argument * * Revision 1.25.1.1 1994/04/25 21:13:24 menze * Removed gratuitous remnants from tcp_reass */ #include "xkernel.h" #include "ip.h" #include "tcp_internal.h" #include "tcp_fsm.h" #include "tcp_seq.h" #include "tcp_timer.h" #include "tcp_var.h" #include "tcpip.h" #include "tcp_debug.h" typedef struct { IPpseudoHdr *pHdr; struct tcphdr tHdr; int tlen; char *options; int option_len; int try_again; } Pop_State; #define CANTRCVMORE(so) (sotoss(so)->t_flags & TF_NETCLOSED) int tcpprintfs = 0; int tcpcksum = 1; int tcprexmtthresh = 3; extern int tcpnodelack; #ifdef __STDC__ static int tcp_reass( Sessn, struct tcphdr *, Msg *, Msg * ); static void tcp_dooptions(Sessn so, char *, int, struct tcphdr *); /* static bool tcp_pulloobchar( char *ptr, long len, VOID *oobc); */ static void tcp_pulloutofband(SState *, struct tcphdr *, Msg *); #else static int tcp_reass(); static void tcp_dooptions(); /* static bool tcp_pulloobchar(); */ static void tcp_pulloutofband(); #endif /* * Insert segment ti into reassembly queue of tcp with * control block tp. Return TH_FIN if reassembly now includes * a segment with FIN. The macro form does the common case inline * (segment is the next to be received on an established connection, * and the queue is empty), avoiding linkage into and removal * from the queue and repetition of various conversions. * * Reassembled message, if one exists, is plaecd in demuxmsg */ #define TCP_REASS(ss, th, so, m, dMsg, flags) { \ if ((th)->th_seq == (ss)->rcv_nxt \ && (ss)->seg_next == (struct reass*)(&(ss)->seg_next) \ && (ss)->t_state == TCPS_ESTABLISHED \ PREDICT_TRUE) \ { \ (ss)->rcv_nxt += msgLength(m); \ flags = (th)->th_flags & TH_FIN; \ TCP_STAT(++ps->tcps_rcvpack; ps->tcps_rcvbyte += msgLength(m)); \ msgAssign((dMsg), (m)); \ } else \ (flags) = tcp_reass((so), (th), (m), (dMsg)); \ } static int tcp_reass(so, th, m, demuxmsg) Sessn so; struct tcphdr *th; Msg *m, *demuxmsg; { struct reass *q, *next; SState *ss = sotoss(so); int flags; xk_int32 i; #ifdef TCP_STATISTICS PState *ps = (PState *)xMyProtl(so)->state; #endif xIfTrace(tcpp, TR_GROSS_EVENTS) { if (th && so && th->th_seq != ss->rcv_nxt) { xTraceS2(so, TR_ALWAYS, "lost/out-of-order: expected seqno %d, got seqno %d", ss->rcv_nxt, th->th_seq); } /* if */ } /* if */ xTraceS0(so, TR_FUNCTIONAL_TRACE, "tcp_reass()"); /* * Call with th==0 after become established to * force pre-ESTABLISHED data up to user socket. */ if (th) { /* * Find a segment which begins after this one does. */ for (q = ss->seg_next; q != (struct reass *)&ss->seg_next; q = (struct reass *)q->next) { if (SEQ_GT(q->th.th_seq, th->th_seq)) { break; } /* if */ } /* for */ /* * 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 (q->prev != (struct reass *)&ss->seg_next) { q = q->prev; /* conversion to int (in i) handles seq wraparound */ i = q->th.th_seq + msgLength(&q->m) - th->th_seq; if (i > 0) { if (i >= msgLength(m)) { TCP_STAT(++ps->tcps_rcvduppack; ps->tcps_rcvdupbyte += msgLength(m)); return 0; } /* if */ msgDiscard(m, i); th->th_seq += i; } /* if */ q = q->next; } /* if */ TCP_STAT(++ps->tcps_rcvoopack; ps->tcps_rcvoobyte += msgLength(m)); /* * While we overlap succeeding segments trim them or, * if they are completely covered, dequeue them. */ while (q != (struct reass *)&ss->seg_next) { i = (th->th_seq + msgLength(m)) - q->th.th_seq; if (i <= 0) { break; } /* if */ if (i < msgLength(&q->m)) { q->th.th_seq += i; msgDiscard(&q->m, i); break; } /* if */ next = q->next; remque(q); msgDestroy(&q->m); xFree((char *) q); q = next; } /* while */ /* * Stick new segment in its place. */ { struct reass *new; new = (struct reass *)xMalloc(sizeof *new); new->th = *th; msgConstructCopy(&new->m, m); insque(new, q); } } /* if */ /* * Present data to user, advancing rcv_nxt through * completed sequence space. */ if (TCPS_HAVERCVDSYN(ss->t_state) == 0) { return 0; } /* if */ q = ss->seg_next; th = &q->th; if (q == (struct reass *)&ss->seg_next || th->th_seq != ss->rcv_nxt) { return 0; } /* if */ if (ss->t_state == TCPS_SYN_RECEIVED && msgLength(&q->m)) { return 0; } /* if */ do { ss->rcv_nxt += msgLength(&q->m); flags = th->th_flags & TH_FIN; next = q->next; remque(q); if (!CANTRCVMORE(so)) { msgJoin(demuxmsg, demuxmsg, &q->m); } msgDestroy(&q->m); xFree((char *)q); q = next; th = &q->th; } while (q != (struct reass *)&ss->seg_next && th->th_seq == ss->rcv_nxt); return flags; } /* tcp_reass */ static XkReturn drop(Msg *m) { xTrace1(tcpp, TR_EVENTS, "drop: dropping message", (u_long)m); /* * Drop space held by incoming segment and return. */ xIfTrace(tcpp, TR_MAJOR_EVENTS) { tcp_trace(TA_DROP, 0, 0, 0, 0); } /* if */ if (m) { msgDestroy(m); } /* if */ return XK_SUCCESS; } /* drop */ /* * Attention: SO argument may be NULL. */ static XkReturn dropwithreset(Sessn so, Msg *m, int tiflags, IPpseudoHdr *pHdr, struct tcphdr tHdr, int tlen, Protl tcp) { IPpseudoHdr tmpPhdr; xTrace1(tcpp, TR_EVENTS, "[tcp] dropwithreset(so=%lx)", (u_long)so); /* * Generate a RST, dropping incoming segment. * Make ACK acceptable to originator of segment. * Don't bother to respond if destination was broadcast. */ if ((tiflags & TH_RST) || IP_ADS_BCAST(pHdr->dst)) { return drop(m); } /* if */ tmpPhdr.src = pHdr->dst; tmpPhdr.dst = pHdr->src; tmpPhdr.zero = 0; tmpPhdr.prot = pHdr->prot; if (tiflags & TH_ACK) { tcp_respond(so, &tHdr, &tmpPhdr, tHdr.th_ack, (tcp_seq)0, TH_RST, tcp); } else { if (tiflags & TH_SYN) { ++tlen; } /* if */ tcp_respond(so, &tHdr, &tmpPhdr, tHdr.th_seq + tlen, (tcp_seq)0, TH_RST|TH_ACK, tcp); } /* if */ if (m) { msgDestroy(m); } /* if */ return XK_SUCCESS; } /* dropwithreset */ static XkReturn dropafterack(Sessn so, int tiflags, Msg *m) { SState *ss = sotoss(so); xTraceS1(so, TR_EVENTS, "dropafterack(so=%lx)", (u_long)so); /* * Generate an ACK dropping incoming segment if it occupies * sequence space, where the ACK reflects our state. */ if (tiflags & TH_RST) { return drop(m); } /* if */ ss->t_flags |= TF_ACKNOW; tcp_output(so); msgDestroy(m); return XK_SUCCESS; } /* dropafterack */ XkReturn tcpPop(so, lls, m, state) Sessn so; Sessn lls; Msg *m; VOID *state; { int todrop, acked, ourfinisacked, needoutput = 0; bool invoke_cantrcvmore = FALSE; SState *ss = sotoss(so); struct tcphdr tHdr; IPpseudoHdr *pHdr; Msg demuxmsg; char *options; int option_len = 0; int tiflags, tlen, len; int iss = 0; int ostate = 0; int bad; Pop_State *pop_state = state; Protl tcp = xMyProtl(so); #ifdef TCP_STATISTICS PState *ps = (PState *)xMyProtl(so)->state; #endif pHdr = pop_state->pHdr; tHdr = pop_state->tHdr; tlen = pop_state->tlen; options = pop_state->options; option_len = pop_state->option_len; xIfTrace(tcpp, 1) { ostate = ss->t_state; } /* if */ tiflags = tHdr.th_flags; msgConstructEmpty(&demuxmsg); /* * Segment received on connection. * Reset idle time and keep-alive timer. */ ss->t_idle = 0; ss->t_timer[TCPT_KEEP] = TCPTV_KEEP; /* * Process options if not in LISTEN state, * else do it below (after getting remote address). */ if (options PREDICT_FALSE) { if (ss->t_state != TCPS_LISTEN) { tcp_dooptions(so, options, option_len, &tHdr); options = 0; } /* if */ } /* if */ /* * Calculate amount of space in receive window, * and then do TCP input processing. * Receive window is amount of space in rcv queue, * but not less than advertised window. */ { int win; win = ss->rcv_space; if (win < 0) win = 0; ss->rcv_wnd = MAX(win, (int)(ss->rcv_adv - ss->rcv_nxt)); xTraceS4(so, TR_MORE_EVENTS, "New win (%x) = max(std (%x), adv (%x) - nxt (%x))", ss->rcv_wnd, win, ss->rcv_adv, ss->rcv_nxt); } xTraceS2(so, TR_EVENTS, "tcpPop: so=%lx switch %s", (u_long)so, tcpstates[ss->t_state]); if (ss->t_state == TCPS_LISTEN PREDICT_FALSE) { /* * 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. * Don't bother responding if the destination was a * broadcast. * Otherwise initialize ss->rcv_nxt, and ss->irs, select * an initial ss->iss, and send a segment: * * Also initialize ss->snd_nxt to ss->iss+1 and * ss->snd_una to ss->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. */ if (tiflags & TH_RST) return drop(&demuxmsg); if (tiflags & TH_ACK) return dropwithreset(so, &demuxmsg, tiflags, pHdr, tHdr, tlen, tcp); if (((tiflags & TH_SYN) == 0) || IP_ADS_BCAST(pHdr->dst)) return drop(&demuxmsg); xTraceS0(so, TR_MAJOR_EVENTS, "tcpPop: LISTEN"); tcp_template(so); if (options) { tcp_dooptions(so, options, option_len, &tHdr); } /* if */ if (iss) ss->iss = iss; else ss->iss = tcp_iss; tcp_iss += TCP_ISSINCR/2; ss->irs = tHdr.th_seq; tcp_sendseqinit(ss); tcp_rcvseqinit(ss); ss->t_flags |= TF_ACKNOW; ss->t_state = TCPS_SYN_RECEIVED; ss->t_timer[TCPT_KEEP] = TCPTV_KEEP; TCP_STAT(++ps->tcps_accepts); goto trimthenstep6; } else if (ss->t_state == TCPS_SYN_SENT PREDICT_FALSE) { /* * 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 ss->rcv_nxt and ss->irs * if seg contains ack then advance ss->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 */ if ((tiflags & TH_ACK) && (SEQ_LEQ(tHdr.th_ack, ss->iss) || SEQ_GT(tHdr.th_ack, ss->snd_max))) { xTraceS0(so, TR_SOFT_ERRORS, "input state SYN_SENT -- dropping with reset"); xTraceS3(so, TR_SOFT_ERRORS, " (ack==%d, iss==%d, snd_max==%d)", tHdr.th_ack, ss->iss, ss->snd_max); return dropwithreset(so, &demuxmsg, tiflags, pHdr, tHdr, tlen, tcp); } /* if */ if (tiflags & TH_RST) { if (tiflags & TH_ACK) { tcp_drop(so, ECONNREFUSED); } /* if */ xTraceS0(so, TR_SOFT_ERRORS, "peer sent reset (TH_RST)---dropping"); return drop(&demuxmsg); } /* if */ if ((tiflags & TH_SYN) == 0) { xTraceS1(so, TR_SOFT_ERRORS, "peer did not send TH_SYN---dropping (tiflags=%x)", tiflags); return drop(&demuxmsg); } /* if */ if (tiflags & TH_ACK) { ss->snd_una = tHdr.th_ack; if (SEQ_LT(ss->snd_nxt, ss->snd_una)) ss->snd_nxt = ss->snd_una; } /* if */ ss->t_timer[TCPT_REXMT] = 0; ss->irs = tHdr.th_seq; tcp_rcvseqinit(ss); ss->t_flags |= TF_ACKNOW; if (tiflags & TH_ACK && SEQ_GT(ss->snd_una, ss->iss)) { TCP_STAT(++ps->tcps_connects); soisconnected(so); ss->t_state = TCPS_ESTABLISHED; ss->t_maxseg = MIN(ss->t_maxseg, tcp_mss(so)); tcp_reass(so, NULL, NULL, &demuxmsg); /* * if we didn't have to retransmit the SYN, * use its rtt as our initial srtt & rtt var. */ if (ss->t_rtt) { ss->t_srtt = ss->t_rtt << 3; ss->t_rttvar = ss->t_rtt << 1; TCPT_RANGESET(ss->t_rxtcur, ((ss->t_srtt >> 2) + ss->t_rttvar) >> 1, TCPTV_MIN, TCPTV_REXMTMAX); ss->t_rtt = 0; } /* if */ } else { ss->t_state = TCPS_SYN_RECEIVED; } /* if */ trimthenstep6: /* * Advance ti->ti_seq to correspond to first data byte. * If data, trim to stay within window, * dropping FIN if necessary. */ tHdr.th_seq++; if (tlen > ss->rcv_wnd) { todrop = tlen - ss->rcv_wnd; xTraceS2(so, TR_MORE_EVENTS, "tlen (%d) > rcv_wnd (%d) ... truncating", tlen, ss->rcv_wnd); msgTruncate(m, todrop); tlen = ss->rcv_wnd; tiflags &= ~TH_FIN; TCP_STAT(ps->tcps_rcvpackafterwin++; ps->tcps_rcvbyteafterwin += todrop); } ss->snd_wl1 = tHdr.th_seq - 1; ss->rcv_up = tHdr.th_seq; goto step6; } /* if */ /* * States other than LISTEN or SYN_SENT. * First check that at least some bytes of segment are within * receive window. If segment begins before rcv_nxt, * drop leading data (and SYN); if nothing left, just ack. */ todrop = ss->rcv_nxt - tHdr.th_seq; if (todrop > 0 PREDICT_FALSE) { if (tiflags & TH_SYN) { tiflags &= ~TH_SYN; tHdr.th_seq++; if (tHdr.th_urp > 1) tHdr.th_urp--; else tiflags &= ~TH_URG; todrop--; } if (todrop > tlen || todrop == tlen && (tiflags & TH_FIN) == 0) { #ifdef TCP_COMPAT_42 /* * Don't toss RST in response to 4.2-style keepalive. */ if (tHdr.th_seq == ss->rcv_nxt - 1 && tiflags & TH_RST) goto do_rst; #endif TCP_STAT(++ps->tcps_rcvduppack; ps->tcps_rcvdupbyte += tlen); todrop = tlen; tiflags &= ~TH_FIN; ss->t_flags |= TF_ACKNOW; } else { TCP_STAT(++ps->tcps_rcvpartduppack; ps->tcps_rcvpartdupbyte += todrop); } xTraceS1(so, TR_MORE_EVENTS, "discarding %d bytes from front of msg", todrop); msgDiscard(m, todrop); tHdr.th_seq += todrop; tlen -= todrop; if (tHdr.th_urp > todrop) tHdr.th_urp -= todrop; else { tiflags &= ~TH_URG; tHdr.th_urp = 0; } } /* * If segment ends after window, drop trailing data * (and PUSH and FIN); if nothing left, just ACK. */ todrop = (tHdr.th_seq + tlen) - (ss->rcv_nxt + ss->rcv_wnd); if (todrop > 0 PREDICT_FALSE) { TCP_STAT(++ps->tcps_rcvpackafterwin); if (todrop >= tlen) { TCP_STAT(ps->tcps_rcvbyteafterwin += tlen); /* * If a new connection request is received * while in TIME_WAIT, drop the old connection * and start over if the sequence numbers * are above the previous ones. */ if (tiflags & TH_SYN && ss->t_state == TCPS_TIME_WAIT && SEQ_GT(tHdr.th_seq, ss->rcv_nxt)) { iss = ss->rcv_nxt + TCP_ISSINCR; /* * There can be no external references to this session * anymore (otherwise we wouldn't be in the TIME_WAIT * state), so we can proceed and directly destroy the * session: */ tcp_destroy(so); msgDestroy(&demuxmsg); pop_state->try_again = TRUE; return XK_SUCCESS; } /* if */ /* * If window is closed can only take segments at * window edge, and have to drop data and PUSH from * incoming segments. Continue processing, but * remember to ack. Otherwise, drop segment * and ack. */ if (ss->rcv_wnd == 0 && tHdr.th_seq == ss->rcv_nxt) { ss->t_flags |= TF_ACKNOW; TCP_STAT(++ps->tcps_rcvwinprobe); } else { return dropafterack(so, tiflags, &demuxmsg); } /* if */ } else { TCP_STAT(ps->tcps_rcvbyteafterwin += todrop); } /* if */ xTraceS1(so, TR_MORE_EVENTS, "segment ends after window -- truncating to %d", todrop); msgTruncate(m, todrop); tlen -= todrop; tiflags &= ~(TH_PUSH|TH_FIN); } #ifdef TCP_COMPAT_42 do_rst: #endif bad = (tiflags & (TH_RST | TH_SYN)) | (~tiflags & TH_ACK); if (bad PREDICT_FALSE) { /* * 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 (ss->t_state) { case TCPS_SYN_RECEIVED: tcp_drop(so, ECONNREFUSED); return drop(&demuxmsg); case TCPS_ESTABLISHED: case TCPS_FIN_WAIT_1: case TCPS_FIN_WAIT_2: case TCPS_CLOSE_WAIT: tcp_drop(so, ECONNRESET); return drop(&demuxmsg); case TCPS_CLOSING: case TCPS_LAST_ACK: case TCPS_TIME_WAIT: /* * There can be no external references to this session * anymore (otherwise we wouldn't be in one of the above * states), so we can proceed and directly destroy the * session: */ tcp_destroy(so); return drop(&demuxmsg); } /* switch */ } /* if */ /* * 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(so, ECONNRESET); return dropwithreset(so, &demuxmsg, tiflags, pHdr, tHdr, tlen, tcp); } /* if */ /* * If the ACK bit is off we drop the segment and return. */ if (!(tiflags & TH_ACK)) { return drop(&demuxmsg); } /* if */ } /* if */ /* * Ack processing. */ if (ss->t_state == TCPS_SYN_RECEIVED PREDICT_FALSE) { /* * In SYN_RECEIVED state if the ack ACKs our SYN then enter * ESTABLISHED state and continue processing, otherwise * send an RST. */ if (SEQ_GT(ss->snd_una, tHdr.th_ack) || SEQ_GT(tHdr.th_ack, ss->snd_max)) return dropwithreset(so, &demuxmsg, tiflags, pHdr, tHdr, tlen, tcp); TCP_STAT(++ps->tcps_connects); soisconnected(so); ss->t_state = TCPS_ESTABLISHED; ss->t_maxseg = MIN(ss->t_maxseg, tcp_mss(so)); tcp_reass(so, NULL, NULL, &demuxmsg); ss->snd_wl1 = tHdr.th_seq - 1; goto do_ack_processing; } else { switch (ss->t_state) { /* * In ESTABLISHED state: drop duplicate ACKs; ACK out of range * ACKs. If the ack is in the range * ss->snd_una < ti->ti_ack <= ss->snd_max * then advance ss->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: do_ack_processing: if (SEQ_LEQ(tHdr.th_ack, ss->snd_una)) { if (tlen == 0 && tHdr.th_win == ss->snd_wnd PREDICT_FALSE) { TCP_STAT(++ps->tcps_rcvdupack); /* * If we have outstanding data (not a * window probe), this is a completely * duplicate ack (ie, window info didn't * change), the ack is the biggest we've * seen and we've seen exactly our rexmt * threshhold of them, assume a packet * has been dropped and retransmit it. * Kludge snd_nxt & the congestion * window so we send only this one * packet. * * We know we're losing at the current * window size so do congestion avoidance * (set ssthresh to half the current window * and pull our congestion window back to * the new ssthresh). * * Dup acks mean that packets have left the * network (they're now cached at the receiver) * so bump cwnd by the amount in the receiver * to keep a constant cwnd packets in the * network. */ if (ss->t_timer[TCPT_REXMT] == 0 || tHdr.th_ack != ss->snd_una) ss->t_dupacks = 0; else if (++ss->t_dupacks == tcprexmtthresh) { tcp_seq onxt = ss->snd_nxt; u_long win = MIN(ss->snd_wnd, ss->snd_cwnd); win /= ss->t_maxseg; win >>= 1; if (win < 2) win = 2; ss->snd_ssthresh = win * ss->t_maxseg; ss->t_timer[TCPT_REXMT] = 0; ss->t_rtt = 0; ss->snd_nxt = tHdr.th_ack; ss->snd_cwnd = ss->t_maxseg; tcp_output(so); ss->snd_cwnd = ss->snd_ssthresh + ss->t_maxseg * ss->t_dupacks; if (SEQ_GT(onxt, ss->snd_nxt)) ss->snd_nxt = onxt; return drop(&demuxmsg); } else if (ss->t_dupacks > tcprexmtthresh) { ss->snd_cwnd += ss->t_maxseg; tcp_output(so); return drop(&demuxmsg); } } else ss->t_dupacks = 0; break; } /* if */ ss->t_dupacks = 0; if (SEQ_GT(tHdr.th_ack, ss->snd_max) PREDICT_FALSE) { TCP_STAT(++ps->tcps_rcvacktoomuch); return drop(&demuxmsg); } acked = tHdr.th_ack - ss->snd_una; xTraceS2(so, TR_EVENTS, "received ACK for byte %d (%d new bytes ACKED)", tHdr.th_ack, acked); TCP_STAT(++ps->tcps_rcvackpack; ps->tcps_rcvackbyte += acked); /* * If transmit timer is running and timed sequence * number was acked, update smoothed round trip time. * Since we now have an rtt measurement, cancel the * timer backoff (cf., Phil Karn's retransmit alg.). * Recompute the initial retransmit timer. */ if (ss->t_rtt && SEQ_GT(tHdr.th_ack, ss->t_rtseq)) { TCP_STAT(++ps->tcps_rttupdated); if (ss->t_srtt != 0 PREDICT_TRUE) { register short delta; /* * srtt is stored as fixed point with 3 bits * after the binary point (i.e., scaled by 8). * The following magic is equivalent * to the smoothing algorithm in rfc793 * with an alpha of .875 * (srtt = rtt/8 + srtt*7/8 in fixed point). * Adjust t_rtt to origin 0. */ ss->t_rtt--; delta = ss->t_rtt - (ss->t_srtt >> 3); if ((ss->t_srtt += delta) <= 0) ss->t_srtt = 1; /* * We accumulate a smoothed rtt variance * (actually, a smoothed mean difference), * then set the retransmit timer to smoothed * rtt + 2 times the smoothed variance. * rttvar is stored as fixed point * with 2 bits after the binary point * (scaled by 4). The following is equivalent * to rfc793 smoothing with an alpha of .75 * (rttvar = rttvar*3/4 + |delta| / 4). * This replaces rfc793's wired-in beta. */ if (delta < 0) delta = -delta; delta -= (ss->t_rttvar >> 2); if ((ss->t_rttvar += delta) <= 0) ss->t_rttvar = 1; } else { /* * No rtt measurement yet - use the * unsmoothed rtt. Set the variance * to half the rtt (so our first * retransmit happens at 2*rtt) */ ss->t_srtt = ss->t_rtt << 3; ss->t_rttvar = ss->t_rtt << 1; } ss->t_rtt = 0; ss->t_rxtshift = 0; TCPT_RANGESET(ss->t_rxtcur, ((ss->t_srtt >> 2) + ss->t_rttvar) >> 1, TCPTV_MIN, TCPTV_REXMTMAX); } /* * If all outstanding data is acked, stop retransmit * timer and remember to restart (more output or persist). * If there is more data to be acked, restart retransmit * timer, using current (possibly backed-off) value. */ if (tHdr.th_ack == ss->snd_max) { ss->t_timer[TCPT_REXMT] = 0; needoutput = 1; } else if (ss->t_timer[TCPT_PERSIST] == 0) ss->t_timer[TCPT_REXMT] = ss->t_rxtcur; /* * When new data is acked, open the congestion window. * If the window gives us less than ssthresh packets * in flight, open exponentially (maxseg per packet). * Otherwise open linearly (maxseg per window, * or maxseg^2 / cwnd per packet). */ # define MAX_CWND 65535 if (ss->snd_cwnd < MAX_CWND PREDICT_FALSE) { u_long incr = ss->t_maxseg; if (ss->snd_cwnd > ss->snd_ssthresh && incr != ss->snd_cwnd) { incr = MAX(incr * incr / ss->snd_cwnd, 1); } /* if */ xTraceS3(so, TR_MAJOR_EVENTS, "tcpPop: congestion window=%x (MIN(%x+%x,ffff))", ss->snd_cwnd + incr, ss->snd_cwnd, incr); ss->snd_cwnd = MIN(ss->snd_cwnd + incr, MAX_CWND); } /* if */ if (acked > sblength(ss->snd) PREDICT_FALSE) { ss->snd_wnd -= sblength(ss->snd); sbdrop(ss->snd, sblength(ss->snd)); ourfinisacked = 1; } else { sbdrop(ss->snd, acked); ss->snd_wnd -= acked; ourfinisacked = 0; } if (ss->waiting.waitCount) { tcpSemVAll(&ss->waiting); } /* if */ ss->snd_una = tHdr.th_ack; if (SEQ_LT(ss->snd_nxt, ss->snd_una)) ss->snd_nxt = ss->snd_una; if (ss->t_state == TCPS_FIN_WAIT_1 PREDICT_FALSE) { /* * 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. */ if (ourfinisacked) { /* * If we can't receive any more * data, then closing user can proceed. * Starting the timer is contrary to the * specification, but if we don't get a FIN * we'll hang forever. */ if (CANTRCVMORE(so)) { soisdisconnected(so); ss->t_timer[TCPT_2MSL] = TCPTV_MAXIDLE; } /* if */ ss->t_state = TCPS_FIN_WAIT_2; } /* if */ } else if (ss->t_state == TCPS_CLOSING PREDICT_FALSE) { /* * 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. */ if (ourfinisacked) { ss->t_state = TCPS_TIME_WAIT; tcp_canceltimers(ss); ss->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL; soisdisconnected(so); } /* if */ } else if (ss->t_state == TCPS_LAST_ACK PREDICT_FALSE) { /* * In LAST_ACK, we may still be waiting for data * to drain and/or to be acked, as well as for the * ack of our FIN. If our FIN is now * acknowledged, delete the TCB, enter the closed * state and return. */ if (ourfinisacked) { /* * There can be no external references to this session * anymore (otherwise we wouldn't be in the LAST_ACK * state), so we can proceed and directly destroy the * session: */ tcp_destroy(so); return drop(&demuxmsg); } /* if */ } else if (ss->t_state == TCPS_TIME_WAIT PREDICT_FALSE) { /* * In TIME_WAIT state the only thing that should * arrive is a retransmission of the remote FIN. * Acknowledge it and restart the finack timer. */ ss->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL; return dropafterack(so, tiflags, &demuxmsg); } /* if */ } /* switch */ } /* if */ step6: xTraceS1(so, TR_EVENTS, "tcpPop: step6 on so=%lx", (u_long)so); /* * Update window information. * Don't look at window if no ACK: TAC's send garbage on first SYN. */ if ((tiflags & TH_ACK) && (SEQ_LT(ss->snd_wl1, tHdr.th_seq) || ss->snd_wl1 == tHdr.th_seq && (SEQ_LT(ss->snd_wl2, tHdr.th_ack) || ss->snd_wl2 == tHdr.th_ack && tHdr.th_win > ss->snd_wnd))) { /* keep track of pure window updates */ if (tlen == 0 && ss->snd_wl2 == tHdr.th_ack && tHdr.th_win > ss->snd_wnd) { TCP_STAT(++ps->tcps_rcvwinupd); } /* if */ ss->snd_wnd = tHdr.th_win; ss->snd_wl1 = tHdr.th_seq; ss->snd_wl2 = tHdr.th_ack; if (ss->snd_wnd > ss->max_sndwnd) ss->max_sndwnd = ss->snd_wnd; needoutput = 1; } /* * Process segments with URG. */ if ((tiflags & TH_URG) && tHdr.th_urp && !TCPS_HAVERCVDFIN(ss->t_state) PREDICT_FALSE) { /* * This is a kludge, but if we receive and accept * random urgent pointers, we'll crash in * soreceive. It's hard to imagine someone * actually wanting to send this much urgent data. */ if (tHdr.th_urp > SB_MAX) { tHdr.th_urp = 0; /* XXX */ tiflags &= ~TH_URG; /* XXX */ goto dodata; /* XXX */ } /* * If this segment advances the known urgent pointer, * then mark the data stream. 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. * * 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 * as the original spec states. */ if (SEQ_GT(tHdr.th_seq+tHdr.th_urp, ss->rcv_up)) { ss->rcv_up = tHdr.th_seq + tHdr.th_urp; sohasoutofband(so, (ss->rcv_hiwat - ss->rcv_space) + (ss->rcv_up - ss->rcv_nxt) - 1); ss->t_oobflags &= ~(TCPOOB_HAVEDATA | TCPOOB_HADDATA); } /* * Remove out of band data so doesn't get presented to user. * This can happen independent of advancing the URG pointer, * but if two URG's are pending at once, some out-of-band * data may creep in... ick. */ if (tHdr.th_urp <= tlen && (ss->t_flags & TF_OOBINLINE) == 0) tcp_pulloutofband(ss, &tHdr, m); } else { /* * If no out of band data is expected, * pull receive urgent pointer along * with the receive window. */ if (SEQ_GT(ss->rcv_nxt, ss->rcv_up)) ss->rcv_up = ss->rcv_nxt; } /* if */ dodata: /* XXX */ xTraceS1(so, TR_EVENTS, "tcpPop: do data on so=%lx", (u_long)so); /* * Process the segment text, merging it into the TCP sequencing queue, * and arranging for acknowledgment of receipt if necessary. * This process logically involves adjusting ss->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 ((tlen || (tiflags & TH_FIN)) && TCPS_HAVERCVDFIN(ss->t_state) == 0) { xTrace1(tcpp, TR_EVENTS, "Calling macro reassemble with msg len %d", msgLength(m)); TCP_REASS(ss, &tHdr, so, m, &demuxmsg, tiflags); xTraceS1(so, TR_MORE_EVENTS, "after reassembly demux msg has length %d", msgLength(m)); /* * Pass received message to user: */ { int mlen = msgLength(&demuxmsg); if (mlen != 0) { if ( ! (ss->t_flags & TF_RCV_ACK_ALWAYS) ) { ss->rcv_space -= mlen; } xDemux(xGetUp(so), so, &demuxmsg); } /* if */ } if (tcpnodelack == 0) ss->t_flags |= TF_DELACK; else ss->t_flags |= TF_ACKNOW; /* * Note the amount of data that peer has sent into * our window, in order to estimate the sender's * buffer size. */ len = ss->rcv_hiwat - (ss->rcv_adv - ss->rcv_nxt); if (len > ss->max_rcvd) ss->max_rcvd = len; } else { tiflags &= ~TH_FIN; } /* * If FIN is received ACK the FIN and let the user know * that the connection is closing. */ if (tiflags & TH_FIN PREDICT_FALSE) { xTraceS1(so, TR_EVENTS, "tcpPop: got fin on so=%lx", (u_long)so); if (TCPS_HAVERCVDFIN(ss->t_state) == 0) { /* delay socantrcvmore() until after data has been processed: */ invoke_cantrcvmore = TRUE; ss->t_flags |= TF_ACKNOW; ss->rcv_nxt++; } switch (ss->t_state) { /* * In SYN_RECEIVED and ESTABLISHED STATES * enter the CLOSE_WAIT state. */ case TCPS_SYN_RECEIVED: case TCPS_ESTABLISHED: ss->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: ss->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: ss->t_state = TCPS_TIME_WAIT; tcp_canceltimers(ss); ss->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: ss->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL; break; } /* switch */ } /* if */ msgDestroy(&demuxmsg); xIfTrace(tcpp, TR_MAJOR_EVENTS) { tcp_trace(TA_INPUT, ostate, ss, 0, 0); } /* if */ /* * Return any desired output. */ if (needoutput || (ss->t_flags & TF_ACKNOW)) { tcp_output(so); } if (invoke_cantrcvmore PREDICT_FALSE) { socantrcvmore(so); } /* if */ return XK_SUCCESS; } /* tcpPop */ /* * TCP input routine, follows pages 65-76 of the * protocol specification dated September, 1981 very closely. */ XkReturn tcpDemux(self, lls, m) Protl self; Sessn lls; Msg *m; { Sessn so = 0; Protl hlp = 0, hlpType = 0; PState *ps = (PState *)self->state; SState *ss = 0; struct tcphdr tHdr; IPpseudoHdr *pHdr = NULL; char *options = 0; int option_len = 0; int tlen, off; Pop_State pop_state; char *buf; u_short check_sum; TCP_STAT(++ps->tcps_rcvtotal); tlen = msgLength(m); pHdr = (IPpseudoHdr *)msgGetAttr(m, 0); xAssert(pHdr); /* * Get IP and TCP header together in stack part of message. * x-kernel Note: IP attaches IP pseudoheader (in network byte * order) as the attribute of the message */ check_sum = inCkSum(m, (u_short *)pHdr, sizeof(IPpseudoHdr)); buf = msgPop(m, sizeof(tHdr)); if (!buf PREDICT_FALSE ) { xTraceP0(self, TR_MAJOR_EVENTS, "tcpDemux: msgPop of header failed -- dropping"); return XK_FAILURE; } /* if */ tcpHdrLoad(&tHdr, buf, sizeof(tHdr), m); msgSetAttr(m, 0, 0, 0); xTraceP4(self, TR_MAJOR_EVENTS, "tcpDemux seq %d, dlen %d, f( %s ) to port (%d)", tHdr.th_seq, msgLength(m), tcpFlagStr(tHdr.th_flags), tHdr.th_dport); /* * Check the checksum value (calculated in tcpHdrLoad): */ if (tcpcksum) { if (check_sum != 0 PREDICT_FALSE) { xTraceP1(self, TR_MAJOR_EVENTS, "tcpDemux: bad checksum (%x)", check_sum); TCP_STAT(++ps->tcps_rcvbadsum); #if BSD<=43 TCP_STAT(++ps->tcps_badsum); #endif return XK_SUCCESS; } /* if */ } /* if */ xTraceP1(self, TR_EVENTS, "received msg with sequence number %d", tHdr.th_seq); /* * Check that TCP offset makes sense, pull out TCP options and * adjust length. */ off = tHdr.th_off << 2; if (off < sizeof(struct tcphdr) || off > tlen PREDICT_FALSE) { xTraceP2(self, TR_ERRORS, "bad tcp off: src %x off %d", pHdr->src, off); TCP_STAT(++ps->tcps_rcvbadoff); #if BSD<=43 TCP_STAT(++ps->tcps_badoff); #endif return XK_SUCCESS; } /* if */ tlen -= off; option_len = off - sizeof (struct tcphdr); if (option_len > 0 PREDICT_FALSE) { xTraceP1(self, TR_MAJOR_EVENTS, "tcpDemux: %d bytes of options", option_len); /* * Re-check the length, cause option_len increases the size * of the tcp header */ if (tlen < 0) { xTraceP2(self, TR_GROSS_EVENTS, "tcpDemux: rcvd short optlen = %d, len = %d", option_len, tlen); TCP_STAT(++ps->tcps_rcvshort); return XK_SUCCESS; } /* if */ /* * Put the options somewhere reasonable */ options = xMalloc(option_len); buf = msgPop(m, option_len); xAssert(buf); tcpOptionsLoad(options, buf, option_len); } /* if */ /* * Locate session for segment. */ do { ActiveId activeId; PassiveId passiveId; bzero((char *)&activeId, sizeof(activeId)); activeId.localport = tHdr.th_dport; activeId.remoteport = tHdr.th_sport; *((int*)&activeId.remoteaddr) = *((int*)&pHdr->src); xTraceP3(self, TR_EVENTS, "looking for %d->%s:%d", activeId.localport, ipHostStr(&activeId.remoteaddr), activeId.remoteport); /* look in the active map */ if (mapResolve(ps->activeMap, &activeId, (void **)&so) != XK_SUCCESS PREDICT_FALSE) { /* * Look in the passive map */ Enable *e; passiveId = activeId.localport; xTraceP1(self, TR_SOFT_ERRORS, "looking for passive open on %d", passiveId); /* look in the map */ if (mapResolve(ps->passiveMap, &passiveId, (void **)&e) == XK_FAILURE) { xTraceP0(self, TR_SOFT_ERRORS, "No passive open object exists"); so = 0; } else { hlp = e->hlp; hlpType = e->hlpType; xTraceP1(self, TR_EVENTS, "found openenable object: %lx", (u_long)so); so = sonewconn(self, hlp, hlpType, &pHdr->src, &pHdr->dst, tHdr.th_sport, tHdr.th_dport); if (!so) { break; } /* if */ /* keep this session around until xOpenDone() or error: */ ss = sotoss(so); /* * Let TCP own a reference until the user gets chance * to take over: */ xDuplicate(so); ss->t_flags |= TF_EMBRYONIC; ss->t_template.ti_src = pHdr->dst; ss->t_template.ti_sport = tHdr.th_dport; ss->t_state = TCPS_LISTEN; } /* if */ } else { ss = sotoss(so); } /* if */ /* * If the state is CLOSED (i.e., TCB does not exist) then * all data in the incoming segment is discarded. * If the TCB exists but is in CLOSED state, it is embryonic, * but should either do a listen or a connect soon. */ if (!so PREDICT_FALSE) { xTraceP0(self, TR_GROSS_EVENTS, "tcpDemux: no so"); dropwithreset(so, 0, tHdr.th_flags, pHdr, tHdr, tlen, self); break; } /* if */ if (ss->t_state == TCPS_CLOSED PREDICT_FALSE) { xTraceP0(self, TR_GROSS_EVENTS, "tcpDemux: TCPS_CLOSED"); break; } /* if */ pop_state.pHdr = pHdr; pop_state.tHdr = tHdr; pop_state.tlen = tlen; pop_state.options = options; pop_state.option_len = option_len; pop_state.try_again = FALSE; xPop(so, lls, m, &pop_state); if (options) { xFree(options); options = 0; } /* if */ } while (pop_state.try_again); return XK_SUCCESS; } /* tcpDemux */ static void tcp_dooptions(so, options, option_len, tHdr) Sessn so; char *options; int option_len; struct tcphdr *tHdr; { SState *ss = sotoss(so); int opt, optlen, cnt; u_char *cp; cp = (u_char *)options; cnt = option_len; for (; cnt > 0; cnt -= optlen, cp += optlen) { opt = cp[0]; if (opt == TCPOPT_EOL) break; if (opt == TCPOPT_NOP) optlen = 1; else { optlen = cp[1]; if (optlen <= 0) break; } switch (opt) { default: break; case TCPOPT_MAXSEG: if (optlen != 4) continue; if (!(tHdr->th_flags & TH_SYN)) continue; ss->t_maxseg = *(u_short *)(cp + 2); ss->t_maxseg = ntohs((u_short)ss->t_maxseg); ss->t_maxseg = MIN(ss->t_maxseg, tcp_mss(so)); break; } /* switch */ } /* for */ } /* tcp_dooptions */ #if 0 static bool tcp_pulloobchar( ptr, len, oobc ) char *ptr; VOID *oobc; long len; { xAssert(len >= 1); *(char *)oobc = *ptr; return FALSE; } /* tcp_pulloobchar */ #endif /* * Pull out of band byte out of a segment so * it doesn't appear in the user's data queue. * It is still reflected in the segment length for * sequencing purposes. */ static void tcp_pulloutofband(ss, th, m) SState *ss; struct tcphdr *th; Msg *m; { Msg firstPart; int cnt = th->th_urp - 1; char *buf; ss->t_oobflags |= TCPOOB_HAVEDATA; msgConstructEmpty(&firstPart); msgBreak(m, &firstPart, cnt); buf = msgPop(m, 1); xAssert(buf); ss->t_iobc = buf[0]; msgJoin(m, &firstPart, m); msgDestroy(&firstPart); } /* tcp_pulloutofband */ /* * Determine a reasonable value for maxseg size. * If the route is known, use one that can be handled * on the given interface without forcing IP to fragment. * If bigger than an mbuf cluster (MCLBYTES), round down to nearest size * to utilize large mbufs. * If interface pointer is unavailable, or the destination isn't local, * use a conservative size (512 or the default IP max size, but no more * than the mtu of the interface through which we route), * as we can't discover anything about intervening gateways or networks. * We also initialize the congestion/slow start window to be a single * segment if the destination isn't local; this information should * probably all be saved with the routing entry at the transport level. * * This is ugly, and doesn't belong at this level, but has to happen somehow. */ int tcp_mss(so) Sessn so; { SState *ss = sotoss(so); int mss; if (xControlSessn(xGetSessnDown(so, 0), GETOPTPACKET, (char *)&mss, sizeof(int)) < sizeof(int)) { xTraceS0(so, TR_SOFT_ERRORS, "tcp_mss: GETOPTPACKET control of lls failed"); mss = 512; } /* if */ xTrace1(tcpp, 3, "tcp_mss: GETOPTPACKET control of lls returned %d", mss); mss -= sizeof(struct tcphdr); ss->snd_cwnd = mss; return mss; } /* tcp_mss */ /*** end of tcp_input.c ***/