tcp_output.c

VXWORKS源代码

/* tcp_output.c - TCP output routines */

/* Copyright 1984 - 2002 Wind River Systems, Inc. */
#include "copyright_wrs.h"

/*
 * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995
 *	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_output.c	8.4 (Berkeley) 5/24/95
 */

/*
modification history
--------------------
01g,05jun02,vvv  reworked previous change to fix performance degradation
01f,06mar02,vvv  fixed Nagle algorithm to handle large writes correctly
		 (SPR #72213)
01e,12oct01,rae  merge from truestack ver 01g, base 01d
01d,16mar99,spm  recovered orphaned code from tor2_0_x branch (SPR #25770)
01c,10dec98,n_s  fixed state machine update for ENOBUFs error. spr 23316
01b,22nov96,vin  added cluster support replaced m_gethdr(..) with 
		 mHdrClGet(..).
01a,03mar96,vin  created from BSD4.4lite2. Integrated with 02n of tcp_output.c
*/

/*
DESCRIPTION
*/

#include "vxWorks.h"
#include "net/mbuf.h"
#include "net/protosw.h"
#include "sys/socket.h"
#include "net/socketvar.h"
#include "errno.h"

#include "net/route.h"

#include "netinet/in.h"
#include "netinet/in_pcb.h"
#include "netinet/in_systm.h"
#include "netinet/ip.h"
#include "netinet/ip_var.h"
#include "netinet/tcp.h"
#define	TCPOUTFLAGS
#include "netinet/tcp_fsm.h"
#include "netinet/tcp_seq.h"
#include "netinet/tcp_timer.h"
#include "netinet/tcp_var.h"
#include "netinet/tcpip.h"
#include "net/systm.h"
#ifdef	BSDDEBUG
#include "netinet/tcp_debug.h"
#endif	/* BSDDEBUG */

#ifdef WV_INSTRUMENTATION
#ifdef INCLUDE_WVNET
#include "wvNetLib.h"
#endif
#endif

#ifdef notyet
extern struct mbuf *m_copypack();
#endif

#ifdef VIRTUAL_STACK
#include "netinet/vsLib.h"
#endif

#ifdef WV_INSTRUMENTATION
#ifdef INCLUDE_WVNET
    /* Set common fields of event identifiers for this module. */

LOCAL UCHAR wvNetModuleId = WV_NET_TCPOUT_MODULE; /* Value for tcp_output.c */
LOCAL UCHAR wvNetLocalFilter = WV_NET_NONE;     /* Available event filter */

LOCAL ULONG wvNetEventId;       /* Event identifier: see wvNetLib.h */
#endif    /* INCLUDE_WVNET */
#endif

#define MAX_TCPOPTLEN	32	/* max # bytes that go in options */

/*
 * Tcp output routine: figure out what should be sent and send it.
 */
int
tcp_output(tp)
	register struct tcpcb *tp;
{
	register struct socket *so = tp->t_inpcb->inp_socket;
        struct rtentry *rt;
	register long len, win;
	int off, flags, error;
	register struct mbuf *m;
	register struct tcpiphdr *ti;
	u_char opt[MAX_TCPOPTLEN];
	unsigned optlen, hdrlen;
	int idle, sendalot, sndBufLen;
	
	BOOL     pktSent = FALSE;   /* TRUE if a packet has been sent to IP */

	/*
	 * Determine length of data that should be transmitted,
	 * and flags that will be used.
	 * If there is some data or critical controls (SYN, RST)
	 * to send, then transmit; otherwise, investigate further.
	 */
	idle = (tp->snd_max == tp->snd_una);
	if (idle && tp->t_idle >= tp->t_rxtcur)
		/*
		 * We have been idle for "a while" and no acks are
		 * expected to clock out any data we send --
		 * slow start to get ack "clock" running again.
		 */
		tp->snd_cwnd = tp->t_maxseg;

	sndBufLen = so->so_snd.sb_cc;  /* number of bytes in socket buffer */
again:
	sendalot = 0;
	off = tp->snd_nxt - tp->snd_una;
	win = min(tp->snd_wnd, tp->snd_cwnd);

	flags = tcp_outflags[tp->t_state];
	/*
	 * If in persist timeout with window of 0, send 1 byte.
	 * Otherwise, if window is small but nonzero
	 * and timer expired, we will send what we can
	 * and go to transmit state.
	 */
	if (tp->t_force) {
		if (win == 0) {
			/*
			 * If we still have some data to send, then
			 * clear the FIN bit.  Usually this would
			 * happen below when it realizes that we
			 * aren't sending all the data.  However,
			 * if we have exactly 1 byte of unset data,
			 * then it won't clear the FIN bit below,
			 * and if we are in persist state, we wind
			 * up sending the packet without recording
			 * that we sent the FIN bit.
			 *
			 * We can't just blindly clear the FIN bit,
			 * because if we don't have any more data
			 * to send then the probe will be the FIN
			 * itself.
			 */
			if (off < so->so_snd.sb_cc)
				flags &= ~TH_FIN;
			win = 1;
		} else {
			tp->t_timer[TCPT_PERSIST] = 0;
			tp->t_rxtshift = 0;
		}
	}

	len = min(so->so_snd.sb_cc, win) - off;

	if (len < 0) {
		/*
		 * If FIN has been sent but not acked,
		 * but we haven't been called to retransmit,
		 * len will be -1.  Otherwise, window shrank
		 * after we sent into it.  If window shrank to 0,
		 * cancel pending retransmit and pull snd_nxt
		 * back to (closed) window.  We will enter persist
		 * state below.  If the window didn't close completely,
		 * just wait for an ACK.
		 */
		len = 0;
		if (win == 0) {
			tp->t_timer[TCPT_REXMT] = 0;
			tp->snd_nxt = tp->snd_una;
		}
	}
	if (len > tp->t_maxseg) {
		len = tp->t_maxseg;
		sendalot = 1;
	}
	if (SEQ_LT(tp->snd_nxt + len, tp->snd_una + so->so_snd.sb_cc))
		flags &= ~TH_FIN;

	win = sbspace(&so->so_rcv);

	/*
	 * Sender silly window avoidance.  If connection is idle
	 * and can send all data, a maximum segment,
	 * at least a maximum default-size segment do it,
	 * or are forced, do it; otherwise don't bother.
	 * If peer's buffer is tiny, then send
	 * when window is at least half open.
	 * If retransmitting (possibly after persist timer forced us
	 * to send into a small window), then must resend.
	 */
	if (len) {
		if (len == tp->t_maxseg)
			goto send;

		/*
		 * Modification to Nagle check - TF_EOB indicates if there
		 * is data in the output buffer which is the last part
		 * of a large write. Such data should be sent out 
		 * immediately even if it results in a small segment.
		 * (SPR #72213).
		 */

		if ((idle || tp->t_flags & TF_NODELAY ||
		     (tp->t_flags & TF_EOB)) &&
		    len + off >= so->so_snd.sb_cc)
		        goto send;

		if (tp->t_force)
			goto send;
		if (len >= tp->max_sndwnd / 2)
			goto send;
		if (SEQ_LT(tp->snd_nxt, tp->snd_max))
			goto send;
	}

	/*
	 * Compare available window to amount of window
	 * known to peer (as advertised window less
	 * next expected input).  If the difference is at least two
	 * max size segments, or at least 50% of the maximum possible
	 * window, then want to send a window update to peer.
	 */
	if (win > 0) {
		/* 
		 * "adv" is the amount we can increase the window,
		 * taking into account that we are limited by
		 * TCP_MAXWIN << tp->rcv_scale.
		 */
		long adv = min(win, (long)TCP_MAXWIN << tp->rcv_scale) -
			(tp->rcv_adv - tp->rcv_nxt);

		if (adv >= (long) (2 * tp->t_maxseg))
			goto send;
		if (2 * adv >= (long) so->so_rcv.sb_hiwat)
			goto send;
	}

	/*
	 * Send if we owe peer an ACK.
	 */
	if (tp->t_flags & TF_ACKNOW)
		goto send;
	if (flags & (TH_SYN|TH_RST))
		goto send;
	if (SEQ_GT(tp->snd_up, tp->snd_una))
		goto send;
	/*
	 * If our state indicates that FIN should be sent
	 * and we have not yet done so, or we're retransmitting the FIN,
	 * then we need to send.
	 */
	if (flags & TH_FIN &&
	    ((tp->t_flags & TF_SENTFIN) == 0 || tp->snd_nxt == tp->snd_una))
		goto send;

	/*
	 * TCP window updates are not reliable, rather a polling protocol
	 * using ``persist'' packets is used to insure receipt of window
	 * updates.  The three ``states'' for the output side are:
	 *	idle			not doing retransmits or persists
	 *	persisting		to move a small or zero window
	 *	(re)transmitting	and thereby not persisting
	 *
	 * tp->t_timer[TCPT_PERSIST]
	 *	is set when we are in persist state.
	 * tp->t_force
	 *	is set when we are called to send a persist packet.
	 * tp->t_timer[TCPT_REXMT]
	 *	is set when we are retransmitting
	 * The output side is idle when both timers are zero.
	 *
	 * If send window is too small, there is data to transmit, and no
	 * retransmit or persist is pending, then go to persist state.
	 * If nothing happens soon, send when timer expires:
	 * if window is nonzero, transmit what we can,
	 * otherwise force out a byte.
	 */
	if (so->so_snd.sb_cc && tp->t_timer[TCPT_REXMT] == 0 &&
	    tp->t_timer[TCPT_PERSIST] == 0) {
		tp->t_rxtshift = 0;
		tcp_setpersist(tp);
	}

	/*
	 * No reason to send a segment, just return.
	 */
	return (0);

send:
	/*
	 * Before ESTABLISHED, force sending of initial options
	 * unless TCP set not to do any options.
	 * NOTE: we assume that the IP/TCP header plus TCP options
	 * always fit in a single mbuf, leaving room for a maximum
	 * link header, i.e.
	 *	max_linkhdr + sizeof (struct tcpiphdr) + optlen <= MHLEN
	 */
	optlen = 0;
	hdrlen = sizeof (struct tcpiphdr);
	if (flags & TH_SYN) {
		tp->snd_nxt = tp->iss;
		if ((tp->t_flags & TF_NOOPT) == 0) {
			u_short mss;

			opt[0] = TCPOPT_MAXSEG;
			opt[1] = 4;
			mss = htons((u_short) tcp_mss(tp, 0));
			bcopy((caddr_t)&mss, (caddr_t)(opt + 2), sizeof(mss));
			optlen = 4;
	 
			if ((tp->t_flags & TF_REQ_SCALE) &&
			    ((flags & TH_ACK) == 0 ||
			    (tp->t_flags & TF_RCVD_SCALE))) {
				*((u_long *) (opt + optlen)) = htonl(
					TCPOPT_NOP << 24 |
					TCPOPT_WINDOW << 16 |
					TCPOLEN_WINDOW << 8 |
					tp->request_r_scale);
				optlen += 4;
			}
		}
 	}
 
 	/*
	 * Send a timestamp and echo-reply if this is a SYN and our side 
	 * wants to use timestamps (TF_REQ_TSTMP is set) or both our side
	 * and our peer have sent timestamps in our SYN's.
 	 */
 	if ((tp->t_flags & (TF_REQ_TSTMP|TF_NOOPT)) == TF_REQ_TSTMP &&
 	     (flags & TH_RST) == 0 &&
 	    ((flags & (TH_SYN|TH_ACK)) == TH_SYN ||
	     (tp->t_flags & TF_RCVD_TSTMP))) {
		u_long *lp = (u_long *)(opt + optlen);
 
 		/* Form timestamp option as shown in appendix A of RFC 1323. */
 		*lp++ = htonl(TCPOPT_TSTAMP_HDR);
 		*lp++ = htonl(tcp_now);
 		*lp   = htonl(tp->ts_recent);
 		optlen += TCPOLEN_TSTAMP_APPA;
 	}

 	hdrlen += optlen;
 
	/*
	 * Adjust data length if insertion of options will
	 * bump the packet length beyond the t_maxseg length.
	 */
	if (len > tp->t_maxseg - optlen) {
		len = tp->t_maxseg - optlen;
		sendalot = 1;
		flags &= ~TH_FIN;
	 }

#ifdef DIAGNOSTIC
 	if (max_linkhdr + hdrlen > CL_SIZE_128)
            {
#ifdef WV_INSTRUMENTATION
#ifdef INCLUDE_WVNET    /* WV_NET_EMERGENCY event */
        WV_NET_PORTOUT_EVENT_1 (NET_CORE_EVENT, WV_NET_EMERGENCY, 28, 1,
                                tp->t_inpcb->inp_lport, tp->t_inpcb->inp_fport, 
                                WV_NETEVENT_TCPOUT_HDRPANIC, WV_NET_SEND,
                                so->so_fd)
#endif  /* INCLUDE_WVNET */
#endif

            panic("tcphdr too big");
            }
#endif