NetBSD/sys/netinet/tcp_usrreq.c

589 lines
14 KiB
C
Raw Normal View History

/* $NetBSD: tcp_usrreq.c,v 1.23 1996/05/23 17:03:29 mycroft Exp $ */
1993-03-21 12:45:37 +03:00
/*
* Copyright (c) 1982, 1986, 1988, 1993
* The Regents of the University of California. All rights reserved.
1993-03-21 12:45:37 +03:00
*
* 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_usrreq.c 8.2 (Berkeley) 1/3/94
1993-03-21 12:45:37 +03:00
*/
1993-12-18 03:40:47 +03:00
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
1993-12-18 03:40:47 +03:00
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/protosw.h>
#include <sys/errno.h>
#include <sys/stat.h>
1996-02-14 02:40:59 +03:00
#include <sys/proc.h>
#include <sys/ucred.h>
1996-05-23 20:13:19 +04:00
1996-02-14 02:40:59 +03:00
#include <vm/vm.h>
#include <sys/sysctl.h>
1993-03-21 12:45:37 +03:00
1993-12-18 03:40:47 +03:00
#include <net/if.h>
#include <net/route.h>
1993-03-21 12:45:37 +03:00
1993-12-18 03:40:47 +03:00
#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/in_var.h>
1993-12-18 03:40:47 +03:00
#include <netinet/ip.h>
#include <netinet/in_pcb.h>
#include <netinet/ip_var.h>
#include <netinet/tcp.h>
#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 <netinet/tcp_debug.h>
1993-03-21 12:45:37 +03:00
/*
* TCP protocol interface to socket abstraction.
*/
extern char *tcpstates[];
/*
* Process a TCP user request for TCP tb. If this is a send request
* then m is the mbuf chain of send data. If this is a timer expiration
* (called from the software clock routine), then timertype tells which timer.
*/
/*ARGSUSED*/
1994-01-09 02:07:16 +03:00
int
tcp_usrreq(so, req, m, nam, control, p)
1993-03-21 12:45:37 +03:00
struct socket *so;
int req;
struct mbuf *m, *nam, *control;
struct proc *p;
1993-03-21 12:45:37 +03:00
{
register struct inpcb *inp;
1996-02-14 02:40:59 +03:00
register struct tcpcb *tp = NULL;
1993-03-21 12:45:37 +03:00
int s;
int error = 0;
int ostate;
if (req == PRU_CONTROL)
return (in_control(so, (long)m, (caddr_t)nam,
(struct ifnet *)control, p));
1993-03-21 12:45:37 +03:00
1995-08-13 03:59:09 +04:00
s = splsoftnet();
1993-03-21 12:45:37 +03:00
inp = sotoinpcb(so);
#ifdef DIAGNOSTIC
if (req != PRU_SEND && req != PRU_SENDOOB && control)
panic("tcp_usrreq: unexpected control mbuf");
#endif
1993-03-21 12:45:37 +03:00
/*
* When a TCP is attached to a socket, then there will be
* a (struct inpcb) pointed at by the socket, and this
* structure will point at a subsidary (struct tcpcb).
*/
if (inp == 0 && req != PRU_ATTACH) {
1996-05-23 20:13:19 +04:00
error = EINVAL;
goto release;
1993-03-21 12:45:37 +03:00
}
if (inp) {
tp = intotcpcb(inp);
/* WHAT IF TP IS 0? */
#ifdef KPROF
tcp_acounts[tp->t_state][req]++;
#endif
ostate = tp->t_state;
} else
ostate = 0;
1996-05-23 20:13:19 +04:00
1993-03-21 12:45:37 +03:00
switch (req) {
/*
* TCP attaches to socket via PRU_ATTACH, reserving space,
* and an internet control block.
*/
case PRU_ATTACH:
1996-05-23 20:13:19 +04:00
if (inp != 0) {
1993-03-21 12:45:37 +03:00
error = EISCONN;
break;
}
error = tcp_attach(so);
if (error)
break;
if ((so->so_options & SO_LINGER) && so->so_linger == 0)
so->so_linger = TCP_LINGERTIME * hz;
1993-03-21 12:45:37 +03:00
tp = sototcpcb(so);
break;
/*
* PRU_DETACH detaches the TCP protocol from the socket.
* If the protocol state is non-embryonic, then can't
* do this directly: have to initiate a PRU_DISCONNECT,
* which may finish later; embryonic TCB's can just
* be discarded here.
*/
case PRU_DETACH:
if (tp->t_state > TCPS_LISTEN)
tp = tcp_disconnect(tp);
else
tp = tcp_close(tp);
break;
/*
* Give the socket an address.
*/
case PRU_BIND:
error = in_pcbbind(inp, nam, p);
1993-03-21 12:45:37 +03:00
break;
/*
* Prepare to accept connections.
*/
case PRU_LISTEN:
1996-05-23 20:13:19 +04:00
if (inp->inp_lport == 0) {
error = in_pcbbind(inp, (struct mbuf *)0,
(struct proc *)0);
1996-05-23 20:13:19 +04:00
if (error)
break;
}
tp->t_state = TCPS_LISTEN;
1993-03-21 12:45:37 +03:00
break;
/*
* Initiate connection to peer.
* Create a template for use in transmissions on this connection.
* Enter SYN_SENT state, and mark socket as connecting.
* Start keep-alive timer, and seed output sequence space.
* Send initial segment on connection.
*/
case PRU_CONNECT:
if (inp->inp_lport == 0) {
error = in_pcbbind(inp, (struct mbuf *)0,
(struct proc *)0);
1993-03-21 12:45:37 +03:00
if (error)
break;
}
error = in_pcbconnect(inp, nam);
if (error)
break;
tp->t_template = tcp_template(tp);
if (tp->t_template == 0) {
in_pcbdisconnect(inp);
error = ENOBUFS;
break;
}
/* Compute window scaling to request. */
while (tp->request_r_scale < TCP_MAX_WINSHIFT &&
(TCP_MAXWIN << tp->request_r_scale) < so->so_rcv.sb_hiwat)
tp->request_r_scale++;
1993-03-21 12:45:37 +03:00
soisconnecting(so);
tcpstat.tcps_connattempt++;
tp->t_state = TCPS_SYN_SENT;
tp->t_timer[TCPT_KEEP] = TCPTV_KEEP_INIT;
tp->iss = tcp_iss; tcp_iss += TCP_ISSINCR/2;
tcp_sendseqinit(tp);
error = tcp_output(tp);
break;
/*
* Create a TCP connection between two sockets.
*/
case PRU_CONNECT2:
error = EOPNOTSUPP;
break;
/*
* Initiate disconnect from peer.
* If connection never passed embryonic stage, just drop;
* else if don't need to let data drain, then can just drop anyways,
* else have to begin TCP shutdown process: mark socket disconnecting,
* drain unread data, state switch to reflect user close, and
* send segment (e.g. FIN) to peer. Socket will be really disconnected
* when peer sends FIN and acks ours.
*
* SHOULD IMPLEMENT LATER PRU_CONNECT VIA REALLOC TCPCB.
*/
case PRU_DISCONNECT:
tp = tcp_disconnect(tp);
break;
/*
* Accept a connection. Essentially all the work is
* done at higher levels; just return the address
* of the peer, storing through addr.
*/
case PRU_ACCEPT:
in_setpeeraddr(inp, nam);
1993-03-21 12:45:37 +03:00
break;
/*
* Mark the connection as being incapable of further output.
*/
case PRU_SHUTDOWN:
socantsendmore(so);
tp = tcp_usrclosed(tp);
if (tp)
error = tcp_output(tp);
break;
/*
* After a receive, possibly send window update to peer.
*/
case PRU_RCVD:
(void) tcp_output(tp);
break;
/*
* Do a send by putting data in output queue and updating urgent
* marker if URG set. Possibly send more data.
*/
case PRU_SEND:
if (control && control->m_len) {
m_freem(control);
m_freem(m);
error = EINVAL;
break;
}
1993-03-21 12:45:37 +03:00
sbappend(&so->so_snd, m);
error = tcp_output(tp);
break;
/*
* Abort the TCP.
*/
case PRU_ABORT:
tp = tcp_drop(tp, ECONNABORTED);
break;
case PRU_SENSE:
1996-05-23 20:13:19 +04:00
/*
* stat: don't bother with a blocksize.
*/
splx(s);
1993-03-21 12:45:37 +03:00
return (0);
case PRU_RCVOOB:
if (control && control->m_len) {
m_freem(control);
m_freem(m);
error = EINVAL;
break;
}
1993-03-21 12:45:37 +03:00
if ((so->so_oobmark == 0 &&
(so->so_state & SS_RCVATMARK) == 0) ||
so->so_options & SO_OOBINLINE ||
tp->t_oobflags & TCPOOB_HADDATA) {
error = EINVAL;
break;
}
if ((tp->t_oobflags & TCPOOB_HAVEDATA) == 0) {
error = EWOULDBLOCK;
break;
}
m->m_len = 1;
*mtod(m, caddr_t) = tp->t_iobc;
if (((long)nam & MSG_PEEK) == 0)
1993-03-21 12:45:37 +03:00
tp->t_oobflags ^= (TCPOOB_HAVEDATA | TCPOOB_HADDATA);
break;
case PRU_SENDOOB:
if (sbspace(&so->so_snd) < -512) {
m_freem(m);
error = ENOBUFS;
break;
}
/*
* 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.
* Otherwise, snd_up should be one lower.
*/
sbappend(&so->so_snd, m);
tp->snd_up = tp->snd_una + so->so_snd.sb_cc;
tp->t_force = 1;
error = tcp_output(tp);
tp->t_force = 0;
break;
case PRU_SOCKADDR:
in_setsockaddr(inp, nam);
break;
case PRU_PEERADDR:
in_setpeeraddr(inp, nam);
break;
/*
* TCP slow timer went off; going through this
* routine for tracing's sake.
*/
case PRU_SLOWTIMO:
tp = tcp_timers(tp, (long)nam);
req |= (long)nam << 8; /* for debug's sake */
1993-03-21 12:45:37 +03:00
break;
default:
panic("tcp_usrreq");
}
if (tp && (so->so_options & SO_DEBUG))
tcp_trace(TA_USER, ostate, tp, (struct tcpiphdr *)0, req);
1996-05-23 20:13:19 +04:00
release:
1993-03-21 12:45:37 +03:00
splx(s);
return (error);
}
1994-01-09 02:07:16 +03:00
int
1993-03-21 12:45:37 +03:00
tcp_ctloutput(op, so, level, optname, mp)
int op;
struct socket *so;
int level, optname;
struct mbuf **mp;
{
int error = 0, s;
struct inpcb *inp;
register struct tcpcb *tp;
1993-03-21 12:45:37 +03:00
register struct mbuf *m;
register int i;
1993-03-21 12:45:37 +03:00
1995-08-13 03:59:09 +04:00
s = splsoftnet();
inp = sotoinpcb(so);
if (inp == NULL) {
splx(s);
if (op == PRCO_SETOPT && *mp)
(void) m_free(*mp);
return (ECONNRESET);
}
if (level != IPPROTO_TCP) {
error = ip_ctloutput(op, so, level, optname, mp);
splx(s);
return (error);
}
tp = intotcpcb(inp);
1993-03-21 12:45:37 +03:00
switch (op) {
case PRCO_SETOPT:
m = *mp;
switch (optname) {
case TCP_NODELAY:
if (m == NULL || m->m_len < sizeof (int))
error = EINVAL;
else if (*mtod(m, int *))
tp->t_flags |= TF_NODELAY;
else
tp->t_flags &= ~TF_NODELAY;
break;
case TCP_MAXSEG:
if (m && (i = *mtod(m, int *)) > 0 && i <= tp->t_maxseg)
tp->t_maxseg = i;
else
error = EINVAL;
break;
1993-03-21 12:45:37 +03:00
default:
error = ENOPROTOOPT;
1993-03-21 12:45:37 +03:00
break;
}
if (m)
(void) m_free(m);
break;
case PRCO_GETOPT:
*mp = m = m_get(M_WAIT, MT_SOOPTS);
m->m_len = sizeof(int);
switch (optname) {
case TCP_NODELAY:
*mtod(m, int *) = tp->t_flags & TF_NODELAY;
break;
case TCP_MAXSEG:
*mtod(m, int *) = tp->t_maxseg;
break;
default:
error = ENOPROTOOPT;
1993-03-21 12:45:37 +03:00
break;
}
break;
}
splx(s);
1993-03-21 12:45:37 +03:00
return (error);
}
#ifndef TCP_SENDSPACE
#define TCP_SENDSPACE 1024*16;
#endif
u_long tcp_sendspace = TCP_SENDSPACE;
#ifndef TCP_RECVSPACE
#define TCP_RECVSPACE 1024*16;
#endif
u_long tcp_recvspace = TCP_RECVSPACE;
1993-03-21 12:45:37 +03:00
/*
* Attach TCP protocol to socket, allocating
* internet protocol control block, tcp control block,
* bufer space, and entering LISTEN state if to accept connections.
*/
1994-01-09 02:07:16 +03:00
int
1993-03-21 12:45:37 +03:00
tcp_attach(so)
struct socket *so;
{
register struct tcpcb *tp;
struct inpcb *inp;
int error;
if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) {
error = soreserve(so, tcp_sendspace, tcp_recvspace);
if (error)
return (error);
}
error = in_pcballoc(so, &tcbtable);
1993-03-21 12:45:37 +03:00
if (error)
return (error);
inp = sotoinpcb(so);
tp = tcp_newtcpcb(inp);
if (tp == 0) {
int nofd = so->so_state & SS_NOFDREF; /* XXX */
so->so_state &= ~SS_NOFDREF; /* don't free the socket yet */
in_pcbdetach(inp);
so->so_state |= nofd;
return (ENOBUFS);
}
tp->t_state = TCPS_CLOSED;
return (0);
}
/*
* Initiate (or continue) disconnect.
* If embryonic state, just send reset (once).
* If in ``let data drain'' option and linger null, just drop.
* Otherwise (hard), mark socket disconnecting and drop
* current input data; switch states based on user close, and
* send segment to peer (with FIN).
*/
struct tcpcb *
tcp_disconnect(tp)
register struct tcpcb *tp;
{
struct socket *so = tp->t_inpcb->inp_socket;
if (TCPS_HAVEESTABLISHED(tp->t_state) == 0)
1993-03-21 12:45:37 +03:00
tp = tcp_close(tp);
else if ((so->so_options & SO_LINGER) && so->so_linger == 0)
tp = tcp_drop(tp, 0);
else {
soisdisconnecting(so);
sbflush(&so->so_rcv);
tp = tcp_usrclosed(tp);
if (tp)
(void) tcp_output(tp);
}
return (tp);
}
/*
* User issued close, and wish to trail through shutdown states:
* if never received SYN, just forget it. If got a SYN from peer,
* but haven't sent FIN, then go to FIN_WAIT_1 state to send peer a FIN.
* If already got a FIN from peer, then almost done; go to LAST_ACK
* state. In all other cases, have already sent FIN to peer (e.g.
* after PRU_SHUTDOWN), and just have to play tedious game waiting
* for peer to send FIN or not respond to keep-alives, etc.
* We can let the user exit from the close as soon as the FIN is acked.
*/
struct tcpcb *
tcp_usrclosed(tp)
register struct tcpcb *tp;
{
switch (tp->t_state) {
case TCPS_CLOSED:
case TCPS_LISTEN:
case TCPS_SYN_SENT:
tp->t_state = TCPS_CLOSED;
tp = tcp_close(tp);
break;
case TCPS_SYN_RECEIVED:
case TCPS_ESTABLISHED:
tp->t_state = TCPS_FIN_WAIT_1;
break;
case TCPS_CLOSE_WAIT:
tp->t_state = TCPS_LAST_ACK;
break;
}
if (tp && tp->t_state >= TCPS_FIN_WAIT_2) {
1993-03-21 12:45:37 +03:00
soisdisconnected(tp->t_inpcb->inp_socket);
1996-01-31 08:42:37 +03:00
/*
* If we are in FIN_WAIT_2, we arrived here because the
* application did a shutdown of the send side. Like the
* case of a transition from FIN_WAIT_1 to FIN_WAIT_2 after
* a full close, we start a timer to make sure sockets are
* not left in FIN_WAIT_2 forever.
*/
if (tp->t_state == TCPS_FIN_WAIT_2)
tp->t_timer[TCPT_2MSL] = tcp_maxidle;
}
1993-03-21 12:45:37 +03:00
return (tp);
}
/*
* Sysctl for tcp variables.
*/
int
tcp_sysctl(name, namelen, oldp, oldlenp, newp, newlen)
int *name;
u_int namelen;
void *oldp;
size_t *oldlenp;
void *newp;
size_t newlen;
{
/* All sysctl names at this level are terminal. */
if (namelen != 1)
return (ENOTDIR);
switch (name[0]) {
case TCPCTL_RFC1323:
return (sysctl_int(oldp, oldlenp, newp, newlen,
&tcp_do_rfc1323));
default:
return (ENOPROTOOPT);
}
/* NOTREACHED */
}