/* * btcp_input.c * * Derived from: * * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1994 * 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_input.c 8.5 (Berkeley) 4/10/94 * * Modified for x-kernel v3.3 12/10/90 * Modifications Copyright (C) 1996,1990 Larry L. Peterson and Norman C. Hutchinson * * $Revision: 1.8 $ * $Date: 1998/05/13 02:34:24 $ */ #ifndef TUBA_INCLUDE #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" #include "btcp.h" #define CANTRCVMORE(so) (sototcpst(so)->closed & 2) #undef IN_MULTICAST #define IN_MULTICAST(i) (((i).a & 0xf0) == 0xe0) extern int Btcp_rcv_size; extern int PrintBadChksum; static int tcpcksum = 1; static int tcprexmtthresh = 3; static struct tcpiphdr tcp_saveti; /* struct inpcb *tcp_last_inpcb = &tcb; */ /* extern u_long sb_max; */ #endif /* TUBA_INCLUDE */ #define TCP_PAWS_IDLE (24 * 24 * 60 * 60 * PR_SLOWHZ) /* for modulo comparisons of timestamps */ #define TSTMP_LT(a,b) ((int)((a)-(b)) < 0) #define TSTMP_GEQ(a,b) ((int)((a)-(b)) >= 0) #ifdef __STDC__ static void tcp_dooptions(PSTATE *, struct tcpcb *, char *, int, struct tcphdr *, int *, u_long *, u_long *); static void tcp_pulloutofband(Sessn, struct tcphdr *, Msg *); static void tcp_xmit_timer(struct tcpcb *, short, PSTATE * ); #else static void tcp_dooptions(); static void tcp_pulloutofband(); static void tcp_xmit_timer(); #endif #ifdef XMEMTRACK extern int BtcpTrackId; extern char *xMallocTrack(unsigned, int); extern char *xMallocZeroTrack(unsigned, int); extern void xFreeTrack(char *, unsigned, int); #define xMalloc(s) xMallocTrack(s, BtcpTrackId) #define xMallocZero(s) xMallocZeroTrack(s, BtcpTrackId) #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. * Set DELACK for segments received in order, but ack immediately * when segments are out of order (so fast retransmit can work). * * Reassembled message, if one exists, is placed in demuxmsg */ #define TCP_REASS_ACKNOW(tp, th, so, m, dMsg, flags) { \ if ((th)->th_seq == (tp)->rcv_nxt && \ (tp)->seg_next == (struct reass *)(tp) && \ (tp)->t_state == TCPS_ESTABLISHED) { \ /* (tp)->t_flags |= TF_DELACK; */\ (tp)->rcv_nxt += msgLength(m); \ flags = (th)->th_flags & TH_FIN; \ ps->tcpstat.tcps_rcvpack++;\ ps->tcpstat.tcps_rcvbyte += msgLength(m);\ msgAssign((dMsg), (m)); \ } else \ (flags) = tcp_reass((tp), (th), (so), (m), (dMsg)); \ (tp)->t_flags |= TF_ACKNOW; \ } #define TCP_REASS_DELACK(tp, th, so, m, dMsg, flags) { \ if ((th)->th_seq == (tp)->rcv_nxt && \ (tp)->seg_next == (struct reass *)(tp) && \ (tp)->t_state == TCPS_ESTABLISHED) { \ (tp)->t_flags |= TF_DELACK; \ (tp)->rcv_nxt += msgLength(m); \ flags = (th)->th_flags & TH_FIN; \ ps->tcpstat.tcps_rcvpack++;\ ps->tcpstat.tcps_rcvbyte += msgLength(m);\ msgAssign((dMsg), (m)); \ } else { \ (flags) = tcp_reass((tp), (th), (so), (m), (dMsg)); \ (tp)->t_flags |= TF_ACKNOW; \ } \ } XkReturn btcpPop(self, lls, m, hdr) Sessn self; Sessn lls; Msg *m; void *hdr; { return xDemux(xGetUp(self), self, m); } #ifndef TUBA_INCLUDE /* * TCP input routine, follows pages 65-76 of the * protocol specification dated September, 1981 very closely. */ XkReturn btcp_input(self, transport_s, m) Protl self; Sessn transport_s; Msg *m; { /* register struct tcpiphdr *ti; */ register struct inpcb *inp; int option_buf[8]; char *options = NULL; int option_len = 0; int len, tlen, off; register struct tcpcb *tp = 0; register int tiflags; /* struct socket *so; */ Sessn so = 0; Protl hlp = 0, hlpType = 0; struct tcpstate *tcpst; int todrop, acked, ourfinisacked, needoutput = 0; short ostate; /* struct in_addr laddr; */ int dropsocket = 0; int iss = 0; int dummy1, dummy2; u_long tiwin, ts_val, ts_ecr; int iv; int ts_present = 0; Msg demuxmsg; struct tcphdr tHdr; IPpseudoHdr *pHdr; bool invoke_cantrcvmore = FALSE; PSTATE *ps = (PSTATE *)self->state; void *buffer; u_short check_sum; ps->tcpstat.tcps_rcvtotal++; tlen = msgLength(m); pHdr = (IPpseudoHdr *)msgGetAttr(m, 0); xAssert(pHdr); /* * Get IP and TCP header together in stack part of message. * UNIX Note: IP leaves IP header in first mbuf. * x-kernel Note: IP attaches IP pseudoheader (in network byte * order) as the attribute of the message */ /* ACB: Added cast to pHdr, to eliminate warning */ check_sum = inCkSum(m, (u_short *) pHdr, sizeof(IPpseudoHdr)); buffer = msgPop(m, sizeof(tHdr)); if (! buffer) { xTrace0(tcpp, 3, "btcp_input: msgPop of header failed -- dropping"); return XK_FAILURE; } tcpHdrLoad(&tHdr, buffer, sizeof(tHdr), m); msgSetAttr(m, 0, 0, 0); xTrace4(tcpp, 3, "btcp_input seq %d, dlen %d, f( %s ) to port (%d)", tHdr.th_seq, msgLength(m), tcpFlagStr(tHdr.th_flags), tHdr.th_dport); DO_TRACE(ps, TCP_EVENT_IN, tcpGetTime(), msgLength(m), GET_TIDHR(tHdr)); DO_TRACE(ps, TCP_EVENT_DATA, tHdr.th_seq, tHdr.th_win>>4, GET_TIDHR(tHdr)); DO_TRACE(ps, TCP_EVENT_DATA, tHdr.th_ack, tHdr.th_flags, GET_TIDHR(tHdr)); msgConstructEmpty(&demuxmsg); /* * Check the checksum value (calculated in tcpHdrLoad) */ if (tcpcksum) { if (check_sum) { xTrace1(tcpp, TR_ERRORS, "btcp_input: bad checksum (%x)", tHdr.th_sum); if (PrintBadChksum) printf("Bad checksum\n"); ps->tcpstat.tcps_rcvbadsum++; DO_TRACE(ps, TCP_EVENT_BADSUM, 0, 0, GET_TIDHR(tHdr)); goto drop; } } xTrace1(tcpp, 4, "Received msg with sequence number %d", tHdr.th_seq); #endif /* TUBA_INCLUDE */ /* * 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) { xTrace2(tcpp, 1, "bad tcp off: src %x off %d", pHdr->src, off); ps->tcpstat.tcps_rcvbadoff++; goto drop; } tiflags = tHdr.th_flags; tlen -= off; if ((option_len = off - sizeof (struct tcphdr)) > 0) { xTrace1(tcpp, 3, "btcp_input: %d bytes of options", option_len); /* * Re-check the length, cause option_len increases the size * of the tcp header */ if (tlen < 0) { xTrace2(tcpp, 2, "btcp_input: rcvd short optlen = %d, len = %d", option_len, tlen); ps->tcpstat.tcps_rcvshort++; /* msgDestroy(&demuxmsg); */ return XK_FAILURE; } /* * Put the options somewhere reasonable */ DO_TRACE(ps, TCP_EVENT_ROPT, option_len, 0, GET_TIDHR(tHdr)); if (option_len > 32) options = xMalloc(option_len); else options = (char *)option_buf; buffer = msgPop(m, option_len); xAssert(buffer); tcpOptionsLoad(options, buffer, option_len, 0); /* * Do quick retrieval of timestamp options ("options * prediction?"). If timestamp is the only option and it's * formatted as recommended in RFC 1323 appendix A, we * quickly get the values now and not bother calling * tcp_dooptions(), etc. */ bcopy(options, (char *)&iv, sizeof(int)); if ((option_len == TCPOLEN_TSTAMP_APPA || (option_len > TCPOLEN_TSTAMP_APPA && options[TCPOLEN_TSTAMP_APPA] == TCPOPT_EOL)) && iv == htonl(TCPOPT_TSTAMP_HDR) && (tiflags & TH_SYN) == 0) { ts_present = 1; ts_val = ntohl(*(u_long *)(options + 4)); ts_ecr = ntohl(*(u_long *)(options + 8)); DO_TRACE(ps, TCP_EVENT_OTHER9, ts_ecr, 1, GET_TIDHR(tHdr)); /* ACB: Added cast to option_buf */ if (options != (char *)option_buf) #ifdef XMEMTRACK xFreeTrack(options,(unsigned)option_len, BtcpTrackId); #else xFree(options); #endif options = NULL; /* we've parsed the options */ } } /* * Locate pcb for segment. */ findpcb: { ActiveId activeId; PassiveId passiveId; PSTATE *pstate; pstate = (PSTATE *)self->state; bzero((char *)&activeId, sizeof(activeId)); activeId.localport = tHdr.th_dport; activeId.remoteport = tHdr.th_sport; activeId.remoteaddr = pHdr->src; xTrace3(tcpp, 4, "Looking for %d->%s.%d", activeId.localport, ipHostStr(&activeId.remoteaddr), activeId.remoteport); /* look in the active map */ if (mapResolve(pstate->activeMap, &activeId, (void **)&so) == XK_FAILURE) { /* * Look in the passive map */ Enable *e; passiveId = activeId.localport; xTrace1(tcpp, 4, "Looking for passive open on %d", passiveId); /* look in the map */ if (mapResolve(pstate->passiveMap, &passiveId, (void **)&e) == XK_FAILURE) { xTrace0(tcpp, 4, "No passive open object exists"); hlp = 0; so = 0; inp = 0; tp = 0; } else { hlp = e->hlp; hlpType = e->hlpType; Btcp_rcv_size = (int)e->info; xTrace1(tcpp, 4, "Found openenable object: %lx", (u_long)so); inp = 0; tp = 0; } } else { inp = sototcpst(so)->inpcb; tp = sototcpst(so)->tcpcb; } } /* * 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 && !hlp) { xTrace0(tcpp, 2, "btcp_input: no so"); goto dropwithreset; } if (xIsProtl(hlp)) { so = btcp_sonewconn(self, hlp, hlpType, &pHdr->src, &pHdr->dst, tHdr.th_sport, tHdr.th_dport); if (so == 0) goto drop; /* * This is ugly, but .... * * Mark socket as temporary until we're * committed to keeping it. The code at * ``drop'' and ``dropwithreset'' check the * flag dropsocket to see if the temporary * socket created here should be discarded. * We mark the socket as discardable until * we're committed to it below in TCPS_LISTEN. */ dropsocket++; inp = sotoinpcb(so); bcopy((char *)&pHdr->dst, (char *)&inp->inp_laddr, sizeof(pHdr->dst)); inp->inp_lport = tHdr.th_dport; tp = intotcpcb(inp); tp->t_state = TCPS_LISTEN; /* #if BSD>=43 */ /* inp->inp_options = ip_srcroute(); */ /* #endif */ /* Compute proper scaling value from buffer space */ while (tp->request_r_scale < TCP_MAX_WINSHIFT && TCP_MAXWIN << tp->request_r_scale < sototcpst(so)->rcv_hiwat) tp->request_r_scale++; } xIfTrace(tcpp, 1) { ostate = tp->t_state; } tcpst = sototcpst(so); if (inp == 0) { xTrace0(tcpp, 2, "btcp_input: no inp"); goto dropwithreset; } if (tp == 0) { xTrace0(tcpp, 2, "btcp_input: no tp"); goto dropwithreset; } if (tp->t_state == TCPS_CLOSED) { xTrace0(tcpp, 2, "btcp_input: TCPS_CLOSED"); goto drop; } /* Unscale the window into a 32-bit value. */ if ((tiflags & TH_SYN) == 0) tiwin = tHdr.th_win << tp->snd_scale; else tiwin = tHdr.th_win; /* * Proposed tcp_demux / tcp_pop split */ /* * Segment received on connection. * Reset idle time and keep-alive timer. */ tp->t_idle = 0; /* tp->t_timer[TCPT_KEEP] = tcp_keepidle; */ tp->t_timer[TCPT_KEEP] = TCPTV_KEEP_IDLE; /* * Process options if not in LISTEN state, * else do it below (after getting remote address). */ if (options && tp->t_state != TCPS_LISTEN) { tcp_dooptions(ps, tp, options, option_len, &tHdr, &ts_present, &ts_val, &ts_ecr); /* ACB: Added cast to option_buf */ if (options != (char *)option_buf) #ifdef XMEMTRACK xFreeTrack(options, (unsigned)option_len, BtcpTrackId); #else xFree(options); #endif } /* LSB should we set options=0 as in previous implementations? */ DO_TRACE(ps, TCP_EVENT_IN0, tp->rcv_nxt, tp->t_state, tcpst->tid); /* * Header prediction: check for the two common cases * of a uni-directional data xfer. If the packet has * no control flags, is in-sequence, the window didn't * change and we're not retransmitting, it's a * candidate. If the length is zero and the ack moved * forward, we're the sender side of the xfer. Just * free the data acked & wake any higher level process * that was blocked waiting for space. If the length * is non-zero and the ack didn't move, we're the * receiver side. If we're getting packets in-order * (the reassembly queue is empty), add the data to * the socket buffer and note that we need a delayed ack. */ if (tp->t_state == TCPS_ESTABLISHED && (tiflags & (TH_SYN|TH_FIN|TH_RST|TH_URG|TH_ACK)) == TH_ACK && (!ts_present || TSTMP_GEQ(ts_val, tp->ts_recent)) && tHdr.th_seq == tp->rcv_nxt && tiwin && tiwin == tp->snd_wnd && tp->snd_nxt == tp->snd_max) { /* * If last ACK falls within this segment's sequence numbers, * record the timestamp. */ if (ts_present && SEQ_LEQ(tHdr.th_seq, tp->last_ack_sent) && SEQ_LT(tp->last_ack_sent, tHdr.th_seq + tlen)) { tp->ts_recent_age = ps->tcp_now; tp->ts_recent = ts_val; } if (tlen == 0) { if (SEQ_GT(tHdr.th_ack, tp->snd_una) && SEQ_LEQ(tHdr.th_ack, tp->snd_max) && tp->snd_cwnd >= tp->snd_wnd #ifdef FIX_1 && tp->t_dupacks < tcprexmtthresh #endif ) { /* * this is a pure ack for outstanding data. */ DO_TRACE(ps, TCP_EVENT_OTHER9, 1, 5, tcpst->tid); ++ps->tcpstat.tcps_predack; if (ts_present) tcp_xmit_timer(tp, ps->tcp_now-ts_ecr+1, ps); else if (tp->t_rtt && SEQ_GT(tHdr.th_ack, tp->t_rtseq)) tcp_xmit_timer(tp, tp->t_rtt, ps); acked = tHdr.th_ack - tp->snd_una; ps->tcpstat.tcps_rcvackpack++; ps->tcpstat.tcps_rcvackbyte += acked; sbdrop(tcpst->snd, acked, &dummy1, &dummy2); tp->snd_una = tHdr.th_ack; /* m_freem(m); */ /* * If all outstanding data are acked, stop * retransmit timer, otherwise restart timer * using current (possibly backed-off) value. * If process is waiting for space, * wakeup/selwakeup/signal. If data * are ready to send, let tcp_output * decide between more output or persist. */ if (tp->snd_una == tp->snd_max) tp->t_timer[TCPT_REXMT] = 0; else if (tp->t_timer[TCPT_PERSIST] == 0) tp->t_timer[TCPT_REXMT] = tp->t_rxtcur; if (tcpst->waiting.waitCount) { tcpSemVAll(&tcpst->waiting); } if (sblength(tcpst->snd) > 0) (void) btcp_output(tp); return XK_SUCCESS; } } else if (tHdr.th_ack == tp->snd_una && tp->seg_next == (struct reass *)tp && tlen <= tcpst->rcv_space) { /* * this is a pure, in-sequence data packet * with nothing on the reassembly queue and * we have enough buffer space to take it. */ ++ps->tcpstat.tcps_preddat; DO_TRACE(ps, TCP_EVENT_OTHER9, 2, 5, tcpst->tid); tp->rcv_nxt += tlen; ps->tcpstat.tcps_rcvpack++; ps->tcpstat.tcps_rcvbyte += tlen; /* * Drop TCP, IP headers and TCP options then add data * to socket buffer. */ /* sbappend(&so->so_rcv, m); */ /* sorwakeup(so); */ if (tp->delack) tp->t_flags |= TF_DELACK; /* * The following else causes every packet to be acked * twice, once on receipt and again when the buffer is * freed, causing a window update. DOGN. */ /* else tp->t_flags |= TF_ACKNOW; */ tcpst->rcv_space -= tlen; xPop(so, transport_s, m, 0); return XK_SUCCESS; } } /* * 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 = tcpst->rcv_space; if (win < 0) win = 0; tp->rcv_wnd = MAX(win, (int)(tp->rcv_adv - tp->rcv_nxt)); xTrace4(tcpp, 5, "New win (%x) = max(std (%x), adv (%x) - nxt (%x))", tp->rcv_wnd, win, tp->rcv_adv, tp->rcv_nxt); } xTrace2(tcpp, 4, "btcp_input: tcpcb %lx switch %s", (u_long)tp, tcpstates[tp->t_state]); 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. * Don't bother responding if the destination was a broadcast. * 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) goto dropwithreset; if ((tiflags & TH_SYN) == 0) goto drop; /* * RFC1122 4.2.3.10, p. 104: discard bcast/mcast SYN * in_broadcast() should never return true on a received * packet with M_BCAST not set. */ /* if (m->m_flags & (M_BCAST|M_MCAST) || LSB flags set by ether input (see if_ethersubr.c) */ if (in_broadcast(*(struct in_addr *)&pHdr->dst)) /* IN_MULTICAST(*(struct in_addr *)&pHdr->dst)) LSB should I?*/ goto drop; xTrace0(tcpp, 3, "btcp_input: LISTEN"); tp->t_template = btcp_template(tp); if (tp->t_template == 0) { tp = btcp_drop(tp, ENOBUFS); dropsocket = 0; /* socket is already gone */ goto drop; } if (options) { tcp_dooptions(ps, tp, options, option_len, &tHdr, &ts_present, &ts_val, &ts_ecr); /* ACB: Added cast to option_buf */ if (options != (char *)option_buf) #ifdef XMEMTRACK xFreeTrack(options,(unsigned)option_len, BtcpTrackId); #else xFree(options); #endif } if (iss) tp->iss = iss; else tp->iss = ps->tcp_iss; ps->tcp_iss += TCP_ISSINCR/2; tp->irs = tHdr.th_seq; tcp_sendseqinit(tp); tcp_rcvseqinit(tp); tp->t_flags |= TF_ACKNOW; tp->t_state = TCPS_SYN_RECEIVED; tp->t_timer[TCPT_KEEP] = TCPTV_KEEP_INIT; dropsocket = 0; /* committed to socket */ ps->tcpstat.tcps_accepts++; 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(tHdr.th_ack, tp->iss) || SEQ_GT(tHdr.th_ack, tp->snd_max))) { xTrace0(tcpp, 4, "input state SYN_SENT -- dropping with reset"); xTrace3(tcpp, 4, " (ack==%d, iss==%d, snd_max==%d)", tHdr.th_ack, tp->iss, tp->snd_max); goto dropwithreset; } if (tiflags & TH_RST) { if (tiflags & TH_ACK) tp = btcp_drop(tp, ECONNREFUSED); goto drop; } if ((tiflags & TH_SYN) == 0) goto drop; if (tiflags & TH_ACK) { tp->snd_una = tHdr.th_ack; if (SEQ_LT(tp->snd_nxt, tp->snd_una)) tp->snd_nxt = tp->snd_una; } tp->t_timer[TCPT_REXMT] = 0; tp->irs = tHdr.th_seq; tcp_rcvseqinit(tp); tp->t_flags |= TF_ACKNOW; if (tiflags & TH_ACK && SEQ_GT(tp->snd_una, tp->iss)) { ps->tcpstat.tcps_connects++; soisconnected(so); tp->t_state = TCPS_ESTABLISHED; /* Do window scaling on this connection? */ if ((tp->t_flags & (TF_RCVD_SCALE|TF_REQ_SCALE)) == (TF_RCVD_SCALE|TF_REQ_SCALE)) { tp->snd_scale = tp->requested_s_scale; tp->rcv_scale = tp->request_r_scale; } /* LSB bsd code doesn't have this */ tp->t_maxseg = MIN(tp->t_maxseg, tcp_mss(tp)); (void) tcp_reass(tp, NULL, so, NULL, &demuxmsg); /* * if we didn't have to retransmit the SYN, * use its rtt as our initial srtt & rtt var. */ if (tp->t_rtt) tcp_xmit_timer(tp, tp->t_rtt, ps); } else tp->t_state = TCPS_SYN_RECEIVED; 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 > tp->rcv_wnd) { todrop = tlen - tp->rcv_wnd; xTrace2(tcpp, 5, "tlen (%d) > rcv_wnd (%d) ... truncating", tlen, tp->rcv_wnd); msgTruncate(m, todrop); tlen = tp->rcv_wnd; tlen = tp->rcv_wnd; tiflags &= ~TH_FIN; ps->tcpstat.tcps_rcvpackafterwin++; ps->tcpstat.tcps_rcvbyteafterwin += todrop; } tp->snd_wl1 = tHdr.th_seq - 1; tp->rcv_up = tHdr.th_seq; goto step6; } /* * States other than LISTEN or SYN_SENT. * First check timestamp, if present. * Then 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. * * RFC 1323 PAWS: If we have a timestamp reply on this segment * and it's less than ts_recent, drop it. */ if (ts_present && (tiflags & TH_RST) == 0 && tp->ts_recent && TSTMP_LT(ts_val, tp->ts_recent)) { /* Check to see if ts_recent is over 24 days old. */ if ((int)(ps->tcp_now - tp->ts_recent_age) > TCP_PAWS_IDLE) { /* * Invalidate ts_recent. If this segment updates * ts_recent, the age will be reset later and ts_recent * will get a valid value. If it does not, setting * ts_recent to zero will at least satisfy the * requirement that zero be placed in the timestamp * echo reply when ts_recent isn't valid. The * age isn't reset until we get a valid ts_recent * because we don't want out-of-order segments to be * dropped when ts_recent is old. */ tp->ts_recent = 0; } else { ps->tcpstat.tcps_rcvduppack++; ps->tcpstat.tcps_rcvdupbyte += tlen; ps->tcpstat.tcps_pawsdrop++; goto dropafterack; } } todrop = tp->rcv_nxt - tHdr.th_seq; if (todrop > 0) { if (tiflags & TH_SYN) { tiflags &= ~TH_SYN; tHdr.th_seq++; if (tHdr.th_urp > 1) tHdr.th_urp--; else tiflags &= ~TH_URG; todrop--; } /* LSB next test changes from previous versions! */ if (todrop >= tlen) { ps->tcpstat.tcps_rcvduppack++; ps->tcpstat.tcps_rcvdupbyte += tlen; /* * If segment is just one to the left of the window, * check two special cases: * 1. Don't toss RST in response to 4.2-style keepalive. * 2. If the only thing to drop is a FIN, we can drop * it, but check the ACK or we will get into FIN * wars if our FINs crossed (both CLOSING). * In either case, send ACK to resynchronize, * but keep on processing for RST or ACK. */ if ((tiflags & TH_FIN && todrop == tlen + 1) #ifdef TCP_COMPAT_42 || (tiflags & TH_RST && tHdr.th_seq == tp->rcv_nxt - 1) #endif ) { todrop = tlen; tiflags &= ~TH_FIN; tp->t_flags |= TF_ACKNOW; } else { /* * Handle the case when a bound socket connects * to itself. Allow packets with a SYN and * an ACK to continue with the processing. */ if (todrop != 0 || (tiflags & TH_ACK) == 0) goto dropafterack; } } else { ps->tcpstat.tcps_rcvpartduppack++; ps->tcpstat.tcps_rcvpartdupbyte += todrop; } xTrace1(tcpp, 5, "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 new data are received on a connection after the * user processes are gone, then RST the other end. */ #define ISNOTREFERENCED(X) 0 if (ISNOTREFERENCED(so) && tp->t_state > TCPS_CLOSE_WAIT && tlen) { tp = btcp_destroy(tp); ps->tcpstat.tcps_rcvafterclose++; goto dropwithreset; } /* * If segment ends after window, drop trailing data * (and PUSH and FIN); if nothing left, just ACK. */ todrop = (tHdr.th_seq+tlen) - (tp->rcv_nxt+tp->rcv_wnd); if (todrop > 0) { ps->tcpstat.tcps_rcvpackafterwin++; if (todrop >= tlen) { ps->tcpstat.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 && tp->t_state == TCPS_TIME_WAIT && SEQ_GT(tHdr.th_seq, tp->rcv_nxt)) { iss = tp->rcv_nxt + TCP_ISSINCR; tp = btcp_destroy(tp); goto findpcb; } /* * 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 (tp->rcv_wnd == 0 && tHdr.th_seq == tp->rcv_nxt) { tp->t_flags |= TF_ACKNOW; ps->tcpstat.tcps_rcvwinprobe++; } else goto dropafterack; } else ps->tcpstat.tcps_rcvbyteafterwin += todrop; xTrace1(tcpp, 5, "segment ends after window -- truncating to %d", todrop); msgTruncate(m, todrop); tlen -= todrop; tiflags &= ~(TH_PUSH|TH_FIN); } /* * If last ACK falls within this segment's sequence numbers, * record its timestamp. */ if (ts_present && SEQ_LEQ(tHdr.th_seq, tp->last_ack_sent) && SEQ_LT(tp->last_ack_sent, tHdr.th_seq + tlen + ((tiflags & (TH_SYN|TH_FIN)) != 0))) { tp->ts_recent_age = ps->tcp_now; tp->ts_recent = ts_val; } /* * 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: tp = btcp_drop(tp, ECONNREFUSED); goto drop; case TCPS_ESTABLISHED: case TCPS_FIN_WAIT_1: case TCPS_FIN_WAIT_2: case TCPS_CLOSE_WAIT: tp = btcp_drop(tp, ECONNRESET); goto drop; case TCPS_CLOSING: case TCPS_LAST_ACK: case TCPS_TIME_WAIT: tp = btcp_destroy(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) { tp = btcp_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, otherwise * send an RST. */ case TCPS_SYN_RECEIVED: if (SEQ_GT(tp->snd_una, tHdr.th_ack) || SEQ_GT(tHdr.th_ack, tp->snd_max)) goto dropwithreset; ps->tcpstat.tcps_connects++; soisconnected(so); tp->t_state = TCPS_ESTABLISHED; /* Do window scaling? */ if ((tp->t_flags & (TF_RCVD_SCALE|TF_REQ_SCALE)) == (TF_RCVD_SCALE|TF_REQ_SCALE)) { tp->snd_scale = tp->requested_s_scale; tp->rcv_scale = tp->request_r_scale; } /* LSB should I add: tp->t_maxseg = MIN(... */ (void) tcp_reass(tp, NULL, so, NULL, &demuxmsg); tp->snd_wl1 = tHdr.th_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: if (SEQ_LEQ(tHdr.th_ack, tp->snd_una)) { if (tlen == 0 && tiwin == tp->snd_wnd) { ps->tcpstat.tcps_rcvdupack++; /* * If we have outstanding data (other than * 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 (tp->t_timer[TCPT_REXMT] == 0 || tHdr.th_ack != tp->snd_una) tp->t_dupacks = 0; else if (++tp->t_dupacks == tcprexmtthresh) { tcp_seq onxt = tp->snd_nxt; u_int win = MIN(tp->snd_wnd, tp->snd_cwnd) / 2/ tp->t_maxseg; if (win < 2) win = 2; tp->snd_ssthresh = win * tp->t_maxseg; tp->t_timer[TCPT_REXMT] = 0; tp->t_rtt = 0; tp->snd_nxt = tHdr.th_ack; tp->snd_cwnd = tp->t_maxseg; (void) btcp_output(tp); tp->snd_cwnd = tp->snd_ssthresh + tp->t_maxseg * tp->t_dupacks; if (SEQ_GT(onxt, tp->snd_nxt)) tp->snd_nxt = onxt; goto drop; } else if (tp->t_dupacks > tcprexmtthresh) { tp->snd_cwnd += tp->t_maxseg; (void) btcp_output(tp); goto drop; } } else tp->t_dupacks = 0; break; } /* * If the congestion window was inflated to account * for the other side's cached packets, retract it. */ #ifdef FIX_0 if (tp->t_dupacks >= tcprexmtthresh && #else if (tp->t_dupacks > tcprexmtthresh && #endif tp->snd_cwnd > tp->snd_ssthresh) tp->snd_cwnd = tp->snd_ssthresh; tp->t_dupacks = 0; if (SEQ_GT(tHdr.th_ack, tp->snd_max)) { ps->tcpstat.tcps_rcvacktoomuch++; goto dropafterack; } acked = tHdr.th_ack - tp->snd_una; ps->tcpstat.tcps_rcvackpack++; ps->tcpstat.tcps_rcvackbyte += acked; xTrace2(tcpp, 4, "Received ACK for byte %d (%d new bytes ACKED)", tHdr.th_ack, acked); DO_TRACE(ps, TCP_EVENT_IN1, tHdr.th_ack, acked, tcpst->tid); /* * If we have a timestamp reply, update smoothed * round trip time. If no timestamp is present but * 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 (ts_present) tcp_xmit_timer(tp, ps->tcp_now-ts_ecr+1, ps); else if (tp->t_rtt && SEQ_GT(tHdr.th_ack, tp->t_rtseq)) tcp_xmit_timer(tp,tp->t_rtt, ps); /* * 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 == tp->snd_max) { tp->t_timer[TCPT_REXMT] = 0; needoutput = 1; } else if (tp->t_timer[TCPT_PERSIST] == 0) tp->t_timer[TCPT_REXMT] = tp->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 * (maxseg^2 / cwnd per packet), plus a constant * fraction of a packet (maxseg/8) to help larger windows * open quickly enough. */ /* LSB: may need to get rid of maxseg/8 part! */ { u_int cw = tp->snd_cwnd; u_int incr = tp->t_maxseg; /* #define FIX_8 */ #define FIX_9 if (cw > tp->snd_ssthresh) #ifdef FIX_8 incr = incr / (cw / incr) #else incr = incr * incr / cw #endif #ifndef FIX_4 + incr / 8 #endif ; #ifdef FIX_9 else if ( (!tp->delack)&&(cw > tp->t_maxseg) ) /* * If using immediate acks, try increasing the slow start * window as if used delayed acks, that is, only every * other acknowlegment. */ incr = (incr+1)/2; #endif xTrace4(tcpp, TR_EVENTS, "btcp_input: congestion window=%x (MIN(%x+%x,65535<<%d))", cw + incr, cw, incr, tp->snd_scale); tp->snd_cwnd = MIN(cw + incr, TCP_MAXWIN<snd_scale); /* if ( strcmp(self->instName,"client")==0) printf( "btcp_input: congestion window=%d (MIN(%d+%d,%d<<%d))\n", tp->snd_cwnd, cw, incr, TCP_MAXWIN, tp->snd_scale); */ } if (acked > sblength(tcpst->snd)) { tp->snd_wnd -= sblength(tcpst->snd); sbdrop(tcpst->snd, sblength(tcpst->snd), &dummy1, &dummy2); ourfinisacked = 1; } else { sbdrop(tcpst->snd, acked, &dummy1, &dummy2); tp->snd_wnd -= acked; ourfinisacked = 0; } if (tcpst->waiting.waitCount) { tcpSemVAll(&tcpst->waiting); } #if 0 if (so->so_snd.sb_flags & SB_NOTIFY) sowwakeup(so); #endif tp->snd_una = tHdr.th_ack; if (SEQ_LT(tp->snd_nxt, tp->snd_una)) tp->snd_nxt = tp->snd_una; 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) { /* * 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); tp->t_timer[TCPT_2MSL]=ps->tcp_maxidle; } 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; btcp_canceltimers(tp); tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL; soisdisconnected(so); } break; /* * 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. */ case TCPS_LAST_ACK: if (ourfinisacked) { tp = btcp_destroy(tp); goto drop; } break; /* * 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; } } step6: xTrace1(tcpp, 4, "btcp_input: step6 on %lx", (u_long)tp); /* * Update window information. * Don't look at window if no ACK: TAC's send garbage on first SYN. */ if ((tiflags & TH_ACK) && (SEQ_LT(tp->snd_wl1, tHdr.th_seq) || tp->snd_wl1 == tHdr.th_seq && (SEQ_LT(tp->snd_wl2, tHdr.th_ack) || tp->snd_wl2 == tHdr.th_ack && tiwin > tp->snd_wnd))) { /* keep track of pure window updates */ if (tlen == 0 && tp->snd_wl2 == tHdr.th_ack && tiwin > tp->snd_wnd) ps->tcpstat.tcps_rcvwinupd++; tp->snd_wnd = tiwin; tp->snd_wl1 = tHdr.th_seq; tp->snd_wl2 = tHdr.th_ack; if (tp->snd_wnd > tp->max_sndwnd) tp->max_sndwnd = tp->snd_wnd; needoutput = 1; } #ifdef FIX_6 if (acked >= 3*tp->t_maxseg && (tp->snd_cwnd - (tp->snd_nxt - tp->snd_una)) > 3*tp->t_maxseg) { if (tp->snd_cwnd > tp->snd_ssthresh) tp->snd_ssthresh = tp->snd_cwnd; tp->snd_cwnd = (tp->snd_nxt - tp->snd_una) + 3*tp->t_maxseg; } #endif /* * Process segments with URG. */ if ((tiflags & TH_URG) && tHdr.th_urp && TCPS_HAVERCVDFIN(tp->t_state) == 0) { /* * 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 (in one of two places). */ if (SEQ_GT(tHdr.th_seq+tHdr.th_urp, tp->rcv_up)) { tp->rcv_up = tHdr.th_seq + tHdr.th_urp; sohasoutofband(so, (tcpst->rcv_hiwat - tcpst->rcv_space) + (tp->rcv_up - tp->rcv_nxt) - 1); tp->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 /* #ifdef SO_OOBINLINE */ && (tcpst->flags & TCPST_OOBINLINE) == 0 /* #endif */ ) tcp_pulloutofband(so, &tHdr, m); } else /* * If no out of band data is expected, * pull receive urgent pointer along * with the receive window. */ if (SEQ_GT(tp->rcv_nxt, tp->rcv_up)) tp->rcv_up = tp->rcv_nxt; dodata: /* XXX */ xTrace1(tcpp, 4, "btcp_input: do data on %lx", (u_long)tp); /* * 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 ((tlen || (tiflags&TH_FIN)) && TCPS_HAVERCVDFIN(tp->t_state) == 0) { xTrace1(tcpp, 5, "Calling macro reassemble with msg len %d", msgLength(m)); if (tp->delack) { TCP_REASS_DELACK(tp, &tHdr, so, m, &demuxmsg, tiflags); } else { TCP_REASS_ACKNOW(tp, &tHdr, so, m, &demuxmsg, tiflags); } xTrace1(tcpp, 5, "after reassembly demux msg has length %d", msgLength(m)); /* * Pass received message to user: */ { int mlen = msgLength(&demuxmsg); if (mlen != 0) { if ( ! (tcpst->flags & TCPST_RCV_ACK_ALWAYS) ) { tcpst->rcv_space -= mlen; } xPop(so, transport_s, &demuxmsg, 0); } /* if */ msgDestroy(&demuxmsg); } /* * Note the amount of data that peer has sent into * our window, in order to estimate the sender's * buffer size. */ len = tcpst->rcv_hiwat - (tp->rcv_adv - tp->rcv_nxt); } 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) { xTrace1(tcpp, 4, "btcp_input: got fin on %lx", (u_long)tp); if (TCPS_HAVERCVDFIN(tp->t_state) == 0) { invoke_cantrcvmore = TRUE; 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; btcp_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; } } xIfTrace(tcpp, 2) tcp_trace(TA_INPUT, ostate, tp, &tcp_saveti, 0); /* * Return any desired output. */ if (needoutput || (tp->t_flags & TF_ACKNOW)) (void) btcp_output(tp); if (invoke_cantrcvmore) { socantrcvmore(so); } /* if */ /* if (!demuxmsgDestroyed) LSB: check freeing */ /* msgDestroy(&demuxmsg); */ return XK_SUCCESS; dropafterack: xTrace1(tcpp, 4, "btcp_input: drop after ack %lx", (u_long)tp); DO_TRACE(ps, TCP_EVENT_DROPAA, tiflags, 0, tcpst->tid); /* * Generate an ACK dropping incoming segment if it occupies * sequence space, where the ACK reflects our state. */ if (tiflags & TH_RST) goto drop; tp->t_flags |= TF_ACKNOW; (void) btcp_output(tp); return XK_SUCCESS; dropwithreset: xTrace1(tcpp, 4, "btcp_input: drop with reset on %lx", (u_long)tp); DO_TRACE(ps, TCP_EVENT_DROPWR, tiflags, 0, tcpst->tid); /* ACB: Added cast to option_buf */ if (options != (char *)option_buf) { #ifdef XMEMTRACK (void) xFreeTrack(options, (unsigned)option_len, BtcpTrackId); #else (void)xFree(options); #endif options = 0; } /* * Generate a RST, dropping incoming segment. * Make ACK acceptable to originator of segment. * Don't bother to respond if destination was broadcast/multicast. */ if ((tiflags & TH_RST) || in_broadcast(*(struct in_addr *)&pHdr->dst) /* m->m_flags & (M_BCAST|M_MCAST) || */ || IN_MULTICAST(pHdr->src)) goto drop; { IPpseudoHdr tmpPhdr; tmpPhdr.src = pHdr->dst; tmpPhdr.dst = pHdr->src; tmpPhdr.zero = 0; tmpPhdr.prot = pHdr->prot; if (tiflags & TH_ACK) btcp_respond(self, tp, &tHdr, &tmpPhdr, tHdr.th_ack, (tcp_seq)0, TH_RST); else { if (tiflags & TH_SYN) tlen++; btcp_respond(self, tp, &tHdr, &tmpPhdr, tHdr.th_seq + tlen, (tcp_seq)0, TH_RST|TH_ACK); } } /* destroy temporarily created socket */ if (dropsocket) (void) soabort(so); msgDestroy(&demuxmsg); return XK_SUCCESS; drop: xTrace1(tcpp, 4, "btcp_input: drop on %lx", (u_long)tp); DO_TRACE(ps, TCP_EVENT_DROP, tiflags, 0, tcpst->tid); /* ACB: Added cast to option_buf */ if (options != (char *)option_buf) { #ifdef XMEMTRACK xFreeTrack(options, (unsigned)option_len, BtcpTrackId); #else xFree(options); #endif options = 0; } /* * Drop space held by incoming segment and return. */ xIfTrace(tcpp, 2) tcp_trace(TA_DROP, ostate, tp, &tcp_saveti, 0); /* destroy temporarily created socket */ if (dropsocket) (void) soabort(so); msgDestroy(&demuxmsg); return XK_SUCCESS; #ifndef TUBA_INCLUDE } static void tcp_dooptions(ps, tp, options, option_len, tHdr, ts_present, ts_val, ts_ecr) PSTATE *ps; struct tcpcb *tp; char *options; int option_len; struct tcphdr *tHdr; int *ts_present; u_long *ts_val, *ts_ecr; { register u_char *cp; int opt, optlen, cnt; struct tcpstate *tcpst; tcpst = tcpcbtoso(tp)->state; 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: continue; case TCPOPT_MAXSEG: if (optlen != TCPOLEN_MAXSEG) continue; if (!(tHdr->th_flags & TH_SYN)) continue; tp->t_maxseg = *(u_short *)(cp + 2); tp->t_maxseg = ntohs((u_short)tp->t_maxseg); DO_TRACE(ps, TCP_EVENT_OTHER9, tp->t_maxseg, 0, tcpst->tid); tp->t_maxseg = MIN(tp->t_maxseg, tcp_mss(tp)); break; case TCPOPT_WINDOW: if (optlen != TCPOLEN_WINDOW) continue; if (!(tHdr->th_flags & TH_SYN)) continue; tp->t_flags |= TF_RCVD_SCALE; tp->requested_s_scale = MIN(cp[2], TCP_MAX_WINSHIFT); DO_TRACE(ps, TCP_EVENT_OTHER9, tp->requested_s_scale, 2, tcpst->tid); break; case TCPOPT_TIMESTAMP: if (optlen != TCPOLEN_TIMESTAMP) continue; *ts_present = 1; bcopy((char *)cp + 2, (char *) ts_val, sizeof(*ts_val)); ntohl(*ts_val); bcopy((char *)cp + 6, (char *) ts_ecr, sizeof(*ts_ecr)); DO_TRACE(ps, TCP_EVENT_OTHER9, *ts_ecr, 1, tcpst->tid); ntohl(*ts_ecr); /* * A timestamp received in a SYN makes * it ok to send timestamp requests and replies. */ if (tHdr->th_flags & TH_SYN) { tp->t_flags |= TF_RCVD_TSTMP; tp->ts_recent = *ts_val; tp->ts_recent_age = ps->tcp_now; DO_TRACE(ps, TCP_EVENT_OTHER9, 3333, 1, tcpst->tid); } break; } } } /* * 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(so, th, m) Sessn so; struct tcphdr *th; Msg *m; { Msg firstPart; struct tcpcb *tp = sototcpcb(so); int cnt = th->th_urp - 1; int dataLen; MsgWalk mi; u_char *data; tp->t_oobflags |= TCPOOB_HAVEDATA; msgConstructEmpty(&firstPart); msgBreak(m, &firstPart, cnt); msgWalkInit(&mi, m); if ((data = msgWalkNext(&mi, &dataLen)) != 0) { xAssert(dataLen >= 1); tp->t_iobc = *(char *)data; } msgWalkDone(&mi); msgDiscard(m, 1); msgJoin(m, &firstPart, m); msgDestroy(&firstPart); } /* * Collect new round-trip time estimate * and update averages and current timeout. */ static void tcp_xmit_timer(tp, rtt, ps) register struct tcpcb *tp; short rtt; PSTATE *ps; { register short delta; ps->tcpstat.tcps_rttupdated++; if (tp->t_srtt != 0) { /* * 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 rtt to origin 0. */ #ifdef FIX_2 delta = ( (rtt-1) << 2 ) - ( tp->t_srtt >> 3 ); #else delta = rtt - 1 - (tp->t_srtt >> TCP_RTT_SHIFT); #endif if ((tp->t_srtt += delta) <= 0) tp->t_srtt = 1; /* * We accumulate a smoothed rtt variance (actually, a * smoothed mean difference), then set the retransmit * timer to smoothed rtt + 4 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; #ifdef FIX_2 delta -= (tp->t_rttvar >> 2); #else delta -= (tp->t_rttvar >> TCP_RTTVAR_SHIFT); #endif if ((tp->t_rttvar += delta) <= 0) tp->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 3*rtt). */ tp->t_srtt = rtt << TCP_RTT_SHIFT; tp->t_rttvar = rtt << (TCP_RTTVAR_SHIFT - 1); } tp->t_rtt = 0; tp->t_rxtshift = 0; /* * the retransmit should happen at rtt + 4 * rttvar. * Because of the way we do the smoothing, srtt and rttvar * will each average +1/2 tick of bias. When we compute * the retransmit timer, we want 1/2 tick of rounding and * 1 extra tick because of +-1/2 tick uncertainty in the * firing of the timer. The bias will give us exactly the * 1.5 tick we need. But, because the bias is * statistical, we have to test that we don't drop below * the minimum feasible timer (which is 2 ticks). */ #ifdef FIX_2 TCPT_RANGESET(tp->t_rxtcur, ( ((tp)->t_srtt >> 3) + (tp)->t_rttvar ) >> 2, rtt+1, TCPTV_REXMTMAX); #else TCPT_RANGESET(tp->t_rxtcur, TCP_REXMTVAL(tp), tp->t_rttmin, TCPTV_REXMTMAX); #endif DO_TRACE(ps, TCP_EVENT_OTHER9, tp->t_rxtcur | (rtt << 8) | (tp->t_srtt << 16) | (tp->t_rttvar << 24), 4, ((struct tcpstate *)tcpcbtoso(tp)->state)->tid); /* * We received an ack for a packet that wasn't retransmitted; * it is probably safe to discard any error indications we've * received recently. This isn't quite right, but close enough * for now (a route might have failed after we sent a segment, * and the return path might not be symmetrical). */ tp->t_softerror = 0; } #ifndef XNETSIM /* * Determine a reasonable value for maxseg size. * If the route is known, check route for mtu. * If none, use an mss that can be handled on the outgoing * interface without forcing IP to fragment; if bigger than * an mbuf cluster (MCLBYTES), round down to nearest multiple of MCLBYTES * to utilize large mbufs. If no route is found, route has no mtu, * or the destination isn't local, use a default, hopefully conservative * size (usually 512 or the default IP max size, but no more than the mtu * of the interface), 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. * While looking at the routing entry, we also initialize other path-dependent * parameters from pre-set or cached values in the routing entry. */ /* #ifdef XNETSIM extern int TcpPacket; #endif */ int tcp_mss(tp, /* offer */) register struct tcpcb *tp; /* u_int offer; */ { /* LSB: major mods to BSD code! */ Sessn tcpSessn; int mss; tcpSessn = tcpcbtoso(tp); if (xControlSessn(xGetSessnDown(tcpSessn, 0), GETOPTPACKET, (char *)&mss, sizeof(int)) < sizeof(int)) { xTrace0(tcpp, 3, "tcp_mss: GETOPTPACKET control of lls failed"); mss = 512; } /* #ifdef XNETSIM if (TcpPacket > 0) mss = TcpPacket + sizeof(struct tcphdr); #endif */ xTrace1(tcpp, 3, "tcp_mss: GETOPTPACKET control of lls returned %d", mss); mss -= sizeof(struct tcphdr); tp->snd_cwnd = tp->t_slowstart * mss; return (mss); } int tcp_reass(tp, th, so, m, demuxmsg) register struct tcpcb *tp; register struct tcphdr *th; Sessn so; Msg *m, *demuxmsg; { register struct reass *q, *next; int flags; PSTATE *ps = (PSTATE *)so->myprotl->state; xTrace0(tcpp, TR_FUNCTIONAL_TRACE, "rtcp_reass function entered"); /* * Call with th==0 after become established to * force pre-ESTABLISHED data up to user socket. */ if (th == 0) goto present; /* * Find a segment which begins after this one does. */ for (q = tp->seg_next; q != (struct reass *)tp; q = (struct reass *)q->next) if (SEQ_GT(q->th.th_seq, th->th_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 (q->prev != (struct reass *)tp) { register int i; 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)) { ps->tcpstat.tcps_rcvduppack++; ps->tcpstat.tcps_rcvdupbyte += msgLength(m); return(0); } msgDiscard(m, i); th->th_seq += i; } q = q->next; } ps->tcpstat.tcps_rcvoopack++; ps->tcpstat.tcps_rcvoobyte += msgLength(m); /* * While we overlap succeeding segments trim them or, * if they are completely covered, dequeue them. */ while (q != (struct reass *)tp) { register int i = (th->th_seq + msgLength(m)) - q->th.th_seq; if (i <= 0) break; if (i < msgLength(&q->m)) { q->th.th_seq += i; msgDiscard(&q->m, i); break; } next = q->next; remque(q); msgDestroy(&q->m); #ifdef XMEMTRACK xFreeTrack((char *) q, sizeof(struct reass), BtcpTrackId); #else xFree((char *)q); #endif q = next; } /* * 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); /* print_reass(tp, "Inserted segment"); */ } present: /* * Present data to user, advancing rcv_nxt through * completed sequence space. */ if (TCPS_HAVERCVDSYN(tp->t_state) == 0) return (0); q = tp->seg_next; th = &q->th; if (q == (struct reass *)tp || th->th_seq != tp->rcv_nxt) return (0); if (tp->t_state == TCPS_SYN_RECEIVED && msgLength(&q->m)) return (0); do { tp->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); #ifdef XMEMTRACK xFreeTrack((char *)q, sizeof(struct reass), BtcpTrackId); #else xFree((char *)q); #endif q = next; th = &q->th; } while (q != (struct reass *)tp && th->th_seq == tp->rcv_nxt); return (flags); } #endif /* TUBA_INCLUDE */ #endif /* XNETSIM */