tcp_input.c revision 490167c8721412bc56cee3521e8fd18f41762b09
* 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 * 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 * Changes and additions relating to SLiRP * Copyright (c) 1995 Danny Gasparovski. * Please read the file COPYRIGHT for the * terms and conditions of the copyright. /* for modulo comparisons of timestamps */ #
define TSTMP_LT(a, b) ((
int)((a)-(b)) < 0)
#
define TSTMP_GEQ(a, b) ((
int)((a)-(b)) >= 0)
#
else /* !TCP_ACK_HACK */#
endif /* TCP_ACK_HACK */ * tcp_reass_maxqlen = 48 (deafault) * tcp_reass_maxseg = nmbclusters/16 (nmbclusters = 1024 + maxusers * 64 from kern/kern_mbuf.c let's say 256) * XXX: tcp_reass() is rather inefficient with its data structures * and should be rewritten (see NetBSD for optimizations). While * doing that it should move to its own file tcp_reass.c. * Call with th==NULL after become established to * force pre-ESTABLISHED data up to user socket. * Limit the number of segments in the reassembly queue to prevent * holding on to too many segments (and thus running out of mbufs). * Make sure to let the missing segment through which caused this * queue. Always keep one global queue entry spare to be able to * process the missing segment. * Allocate a new queue entry. If we can't, or hit the zone limit * Find a segment which begins after this one does. * 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. /* conversion to int (in i) handles seq wraparound */ * Try to present any queued data * at the left window edge to the user. * This is needed after the 3-WHS * While we overlap succeeding segments trim them or, * if they are completely covered, dequeue them. /* Insert the new segment queue entry into place. */ * Present data to user, advancing rcv_nxt through * completed sequence space. /* XXX: This place should be checked for the same code in * original BSD code for Slirp and current BSD used SS_FCANTRCVMORE * TCP input routine, follows pages 65-76 of the * protocol specification dated September, 1981 very closely. /* int dropsocket = 0; */ /* int ts_present = 0; */ * If called with m == 0, then we're continuing the connect Log4((
"NAT: tcp_input: %R[natsock]\n",
so));
/* Re-set a few variables */ /** @todo (vvl) clarify why it might happens */ LogRel((
"NAT: ti is null. can't do any reseting connection actions\n"));
/* mbuf should be cleared in sofree called from tcp_close */ * Get IP and TCP header together in first mbuf. * Note: IP leaves IP header in first mbuf. /* XXX Check if too short */ * Save a copy of the IP header in case we want restore it * for sending an ICMP error message in response. * Checksum extended TCP header and data. /* keep checksum for ICMP reply * ti->ti_sum = cksum(m, len); * Check that TCP offset makes sense, * pull out TCP options and adjust length. XXX * 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 * Convert TCP protocol specific fields to host format. * Drop TCP, IP headers and TCP options. * Locate pcb for segment. /* @todo fix SOLOOKUP macrodefinition to be usable here */ break;
/* so is locked here */ * 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. * state == CLOSED means we've done socreate() but haven't * attached it to a protocol yet... * XXX If a TCB does not exist, and the TH_SYN flag is * the only flag set, then create a session, mark it * as if it was LISTENING, and continue... RTMemFree(
so);
/* Not sofree (if it failed, it's not insqued) */ /* tcp_last_so = so; */ /* XXX ? */ /* tp = sototcpcb(so); */ * If this is a still-connecting socket, this probably * a retransmit of the SYN. Whether it's a retransmit SYN * or something else, we nuke it. /* XXX Should never fail */ /* Unscale the window into a 32-bit value. */ /* if ((tiflags & TH_SYN) == 0) * tiwin = ti->ti_win << tp->snd_scale; * Segment received on connection. * Reset idle time and keep-alive timer. * Process options if not in LISTEN state, * else do it below (after getting remote address). /* &ts_present, &ts_val, &ts_ecr); */ * 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. * XXX Some of these tests are not needed * eg: the tiwin == tp->snd_wnd prevents many more * predictions.. with no *real* advantage.. /* && (!ts_present || TSTMP_GEQ(ts_val, tp->ts_recent)) */ * If last ACK falls within this segment's sequence numbers, * this is a pure ack for outstanding data. * If all outstanding data are acked, stop * retransmit timer, otherwise restart timer * using current (possibly backed-off) value. * If process is waiting for space, * are ready to send, let tcp_output * decide between more output or persist. * There's room in so_snd, sowwakup will read() * from the socket if we can * This is called because sowwakeup might have * put data into so_snd. Since we don't so sowwakeup, * we don't need this.. XXX??? * this is a pure, in-sequence data packet * with nothing on the reassembly queue and * we have enough buffer space to take it. * Add data to socket buffer. * XXX This is called when data arrives. Later, check * if we can actually write() to the socket * XXX Need to check? It's be NON_BLOCKING * If this is a short packet, then ACK now - with Nagel * congestion avoidance sender won't send more until * It is better to not delay acks at all to maximize * TCP throughput. See RFC 2581. }
/* header prediction */ * 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. * 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: * <SEQ=ISS><ACK=RCV_NXT><CTL=SYN,ACK> * 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 * This has way too many gotos... * But a bit of spaghetti code never hurt anybody :) /* ACK the SYN, send RST to refuse the connection */ * Haven't connected yet, save the current mbuf * XXX Some OS's don't tell us whether the connect() * succeeded or not. So we must time it out. * Check if the connect succeeded * 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 /* Do window scaling on this connection? */ * if we didn't have to retransmit the SYN, * use its rtt as our initial srtt & rtt var. * Advance ti->ti_seq to correspond to first data byte. * If data, trim to stay within window, * dropping FIN if necessary. }
/* switch tp->t_state */ * 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. /* Check to see if ts_recent is over 24 days old. */ * 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. * Following if statement from Stevens, vol. 2, p. 960. * Any valid FIN must be to the left of the window. * At this point the FIN must be a duplicate or out * Send an ACK to resynchronize and drop any data. * But keep on processing for RST or ACK. * If new data are received on a connection after the * user processes are gone, then RST the other end. * If segment ends after window, drop trailing data * (and PUSH and FIN); if nothing left, just ACK. * 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 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 * If last ACK falls within this segment's sequence numbers, * If the RST bit is set examine the 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 /* so->so_error = ECONNREFUSED; */ /* so->so_error = ECONNRESET; */ * If a SYN is in the window, then this is an * error and we send an RST and drop the connection. * If the ACK bit is off we drop the segment and return. * In SYN_RECEIVED state if the ack ACKs our SYN then enter * ESTABLISHED state and continue processing, otherwise * send an RST. una<=ack<=max * The sent SYN is ack'ed with our sequence number +1 * The first data byte already in the buffer will get * lost if no correction is made. This is only needed for * SS_CTL since the buffer is empty otherwise. /* Avoid ack processing; snd_una==ti_ack => dup ack */ * 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. * 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 * threshold of them, assume a packet * has been dropped and retransmit it. * Kludge snd_nxt & the congestion * window so we send only this one * 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 * 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 * If the congestion window was inflated to account * for the other side's cached packets, retract it. * 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 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. * 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). * XXX sowwakup is called when data is acked and there's room for * for more data... it should read() the socket * In FIN_WAIT_1 STATE in addition to the processing * for the ESTABLISHED state if our FIN is now acknowledged * 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 * 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 * 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. * In TIME_WAIT state the only thing that should arrive * is a retransmission of the remote FIN. Acknowledge * it and restart the finack timer. }
/* switch(tp->t_state) */ * Update window information. * Don't look at window if no ACK: TAC's send garbage on first SYN. /* keep track of pure window updates */ * Process segments with URG. * 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 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 no out of band data is expected, * pull receive urgent pointer along * with the receive window. * If this is a small packet, then ACK now - with Nagel * congestion avoidance sender won't send more until * 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 * case PRU_RCVD). If a FIN has already been received on this * connection then we just ignore the text. * Note the amount of data that peer has sent into * our window, in order to estimate the sender's * If FIN is received ACK the FIN and let the user know * that the connection is closing. * If we receive a FIN we can't send more data, * Shutdown the socket if there is no rx data in the * soread() is called on completion of shutdown() and * will got to TCPS_LAST_ACK, and use tcp_output() /* sofcantrcvmore(so); */ * In SYN_RECEIVED and ESTABLISHED STATES * enter the CLOSE_WAIT state. * If still in FIN_WAIT_1 STATE FIN has not been acked so * enter the CLOSING state. * In FIN_WAIT_2 state enter the TIME_WAIT state, * starting the time-wait timer, turning off the other * In TIME_WAIT state restart the 2 MSL time_wait timer. * Return any desired output. Log2((
"drop after ack\n"));
* Generate an ACK dropping incoming segment if it occupies * sequence space, where the ACK reflects our state. /* reuses m if m!=NULL, m_free() unnecessary */ * Drop space held by incoming segment and return. * A timestamp received in a SYN makes * it ok to send timestamp requests and replies. * 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 m = m->
m_next;
/* XXX WRONG! Fix it! */ panic(
"tcp_pulloutofband");
* Collect new round-trip time estimate * and update averages and current timeout. * 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. * 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. * No rtt measurement yet - use the unsmoothed rtt. * Set the variance to half the rtt (so our first * retransmit happens at 3*rtt). * 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). * 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). * 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 * 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.