uipc_sock.c

完整的TCP/IP源代码

/* uipc_sock.c - uipc socket routines */

/* Copyright 1984-1996 Wind River Systems, Inc. */
#include "copyright_wrs.h"

/*
 * Copyright (c) 1982, 1986, 1988, 1990, 1993
 *	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.
 *
 *	@(#)uipc_socket.c	8.6 (Berkeley) 5/2/95
 */

/*
modification history
--------------------
01i,24sep98,ham  considered if(m->m_extSize<max_hdr) in sosend() SPR#22209.
01h,26aug98,n_s  optimized processing of uio buffer in sosend. spr #22246.
01g,26aug98,n_s  added return val check for MALLOC in socreate and for 
                 mBufClGet in soreceive. spr #22238.
01f,11aug97,vin  fixed problems in sosend, adjusted space in sosend
01e,23jan97,vin  added fix for SPR7502.
01d,03dec96,vin  replaced calloc(..) with MALLOC(..), free(..) with FREE(..)
01c,22nov96,vin  added cluster support, replaced m_get with mBufClGet and
		 m_gethdr with mHdrClGet.
01b,29aug96,vin  added zerocopy interface. 
01a,03mar96,vin  created from BSD4.4lite2. Integrated with 02u of uipc_sock.c
*/

#include "vxWorks.h"
#include "semLib.h"
#include "memLib.h"

#include "errno.h"
#include "net/mbuf.h"
#include "net/domain.h"
#include "net/protosw.h"
#include "sys/socket.h"
#include "sys/times.h"
#include "net/socketvar.h"
#include "sys/ioctl.h"
#include "net/uio.h"
#include "net/route.h"
#include "netinet/in.h"
#include "net/if.h"
#include "net/systm.h"
#include "selectLib.h"
#include "sockLib.h"

/*
 * Socket operation routines.
 * These routines are called by the routines in
 * sys_socket.c or from a system process, and
 * implement the semantics of socket operations by
 * switching out to the protocol specific routines.
 */
/*ARGSUSED*/
int
socreate(dom, aso, type, proto)
	int dom;
	struct socket **aso;
	register int type;
	int proto;
{
	register struct protosw *prp;
	register struct socket *so;
	register int error;

	if (proto)
		prp = pffindproto(dom, proto, type);
	else
		prp = pffindtype(dom, type);
	if (prp == 0 || prp->pr_usrreq == 0)
		return (EPROTONOSUPPORT);
	if (prp->pr_type != type)
		return (EPROTOTYPE);

        MALLOC(so, struct socket *, sizeof(*so), MT_SOCKET, M_WAIT);
	if (so == (struct socket *) NULL)
	    {
	    return (ENOBUFS);
	    }

        bzero((caddr_t)so, sizeof(*so));

	so->so_options = 0; 
	so->so_type = type;

	/* all sockets will be priveledged */
	so->so_state = SS_PRIV;
	so->so_proto = prp;

	/* initialize socket semaphores */

	semBInit (&so->so_timeoSem, SEM_Q_PRIORITY, SEM_EMPTY);

	semBInit (&so->so_rcv.sb_Sem, SEM_Q_PRIORITY, SEM_EMPTY);
	semBInit (&so->so_snd.sb_Sem, SEM_Q_PRIORITY, SEM_EMPTY);
	so->so_rcv.sb_want = 0;
	so->so_snd.sb_want = 0;

	/* initialize the select stuff */

	selWakeupListInit (&so->so_selWakeupList);

	error =
	    (*prp->pr_usrreq)(so, PRU_ATTACH, (struct mbuf *)0,
		(struct mbuf *)(long)proto, (struct mbuf *)0);
	if (error) {
		so->so_state |= SS_NOFDREF;
		sofree(so);
		return (error);
	}
	*aso = so;
	return (0);
}

int
sobind(so, nam)
	struct socket *so;
	struct mbuf *nam;
{
	int s = splnet();
	int error;

	error =
	    (*so->so_proto->pr_usrreq)(so, PRU_BIND,
		(struct mbuf *)0, nam, (struct mbuf *)0);
	splx(s);
	return (error);
}

int
solisten(so, backlog)
	register struct socket *so;
	int backlog;
{
	int s = splnet(), error;

	error =
	    (*so->so_proto->pr_usrreq)(so, PRU_LISTEN,
		(struct mbuf *)0, (struct mbuf *)0, (struct mbuf *)0);
	if (error) {
		splx(s);
		return (error);
	}
	if (so->so_q == 0)
		so->so_options |= SO_ACCEPTCONN;
	if (backlog < 0)
		backlog = 0;
	so->so_qlimit = min(backlog, SOMAXCONN);
	splx(s);
	return (0);
}

int
sofree(so)
	register struct socket *so;
{

	if (so->so_pcb || (so->so_state & SS_NOFDREF) == 0)
		return;
	if (so->so_head) {
		if (!soqremque(so, 0) && !soqremque(so, 1))
			panic("sofree dq");
		so->so_head = 0;
	}
	sbrelease(&so->so_snd);
	sorflush(so);
	FREE(so, MT_SOCKET); 
}

/*
 * Close a socket on last file table reference removal.
 * Initiate disconnect if connected.
 * Free socket when disconnect complete.
 */
int
soclose(so)
	register struct socket *so;
{
	int s = splnet();		/* conservative */
	int error = 0;

	if (so->so_options & SO_ACCEPTCONN) {
		while (so->so_q0)
			(void) soabort(so->so_q0);
		while (so->so_q)
			(void) soabort(so->so_q);
	}
	if (so->so_pcb == 0)
		goto discard;
	if (so->so_state & SS_ISCONNECTED) {
		if ((so->so_state & SS_ISDISCONNECTING) == 0) {
			error = sodisconnect(so);
			if (error)
				goto drop;
		}
		if (so->so_options & SO_LINGER) {
			if ((so->so_state & SS_ISDISCONNECTING) &&
			    (so->so_state & SS_NBIO))
				goto drop;
			while (so->so_state & SS_ISCONNECTED)
				ksleep(&so->so_timeoSem);
		}
	}
drop:
	if (so->so_pcb) {
		int error2 =
		    (*so->so_proto->pr_usrreq)(so, PRU_DETACH,
			(struct mbuf *)0, (struct mbuf *)0, (struct mbuf *)0);
		if (error == 0)
			error = error2;
	}
discard:
	if (so->so_state & SS_NOFDREF)
		panic("soclose: NOFDREF");
	so->so_state |= SS_NOFDREF;
	sofree(so);
	splx(s);
	return (error);
}

/*
 * Must be called at splnet...
 */
int
soabort(so)
	struct socket *so;
{

	return (
	    (*so->so_proto->pr_usrreq)(so, PRU_ABORT,
		(struct mbuf *)0, (struct mbuf *)0, (struct mbuf *)0));
}

int
soaccept(so, nam)
	register struct socket *so;
	struct mbuf *nam;
{
	int s = splnet();
	int error;

	if ((so->so_state & SS_NOFDREF) == 0)
		panic("soaccept: !NOFDREF");
	so->so_state &= ~SS_NOFDREF;
	error = (*so->so_proto->pr_usrreq)(so, PRU_ACCEPT,
	    (struct mbuf *)0, nam, (struct mbuf *)0);
	splx(s);
	return (error);
}

int
soconnect(so, nam)
	register struct socket *so;
	struct mbuf *nam;
{
	int s;
	int error;

	if (so->so_options & SO_ACCEPTCONN)
		return (EOPNOTSUPP);
	s = splnet();
	/*
	 * If protocol is connection-based, can only connect once.
	 * Otherwise, if connected, try to disconnect first.
	 * This allows user to disconnect by connecting to, e.g.,
	 * a null address.
	 */
	if (so->so_state & (SS_ISCONNECTED|SS_ISCONNECTING) &&
	    ((so->so_proto->pr_flags & PR_CONNREQUIRED) ||
	    (error = sodisconnect(so))))
		error = EISCONN;
	else
		error = (*so->so_proto->pr_usrreq)(so, PRU_CONNECT,
		    (struct mbuf *)0, nam, (struct mbuf *)0);
	splx(s);
	return (error);
}

int
soconnect2(so1, so2)
	register struct socket *so1;
	struct socket *so2;
{
	int s = splnet();
	int error;

	error = (*so1->so_proto->pr_usrreq)(so1, PRU_CONNECT2,
	    (struct mbuf *)0, (struct mbuf *)so2, (struct mbuf *)0);
	splx(s);
	return (error);
}

int
sodisconnect(so)
	register struct socket *so;
{
	int s = splnet();
	int error;

	if ((so->so_state & SS_ISCONNECTED) == 0) {
		error = ENOTCONN;
		goto bad;
	}
	if (so->so_state & SS_ISDISCONNECTING) {
		error = EALREADY;
		goto bad;
	}
	error = (*so->so_proto->pr_usrreq)(so, PRU_DISCONNECT,
	    (struct mbuf *)0, (struct mbuf *)0, (struct mbuf *)0);
bad:
	splx(s);
	return (error);
}

#define	SBLOCKWAIT(f)	(((f) & MSG_DONTWAIT) ? M_DONTWAIT : M_WAIT)
/*
 * Send on a socket.
 * If send must go all at once and message is larger than
 * send buffering, then hard error.
 * Lock against other senders.
 * If must go all at once and not enough room now, then
 * inform user that this would block and do nothing.
 * Otherwise, if nonblocking, send as much as possible.
 * The data to be sent is described by "uio" if nonzero,
 * otherwise by the mbuf chain "top" (which must be null
 * if uio is not).  Data provided in mbuf chain must be small
 * enough to send all at once.
 *
 * Returns nonzero on error, timeout or signal; callers
 * must check for short counts if EINTR/ERESTART are returned.
 * Data and control buffers are freed on return.
 *
 * WRS mods removed sblock and sbunlock and replaced with splnet and splx
 * which should serve the purpose.
 * Removed null check of uio so that blocking can be implemented for zbufs
 * also. 
 * CAVEAT: the zbuf length cannot be more than the socket high water mark.
 * The user should implement his flow control. 
 * So this call would block only if zbuf length is bigger the space available
 * in the socket buffer and less than the socket higt water mark. 
 * Added a flag canWait which is set to M_DONTWAIT if the socket option SS_NBIO
 * is set. This prevents blocking if the user chose SS_NBIO and for some reason
 * if the system runs out of mbufs. canWait defaults to M_WAIT -(vinai).
 */
int
sosend(so, addr, uio, top, control, flags)
	register struct socket *so;
	struct mbuf *addr;
	struct uio *uio;
	struct mbuf *top;
	struct mbuf *control;
	int flags;
{
	struct mbuf **mp;
	register struct mbuf *m;
	register long space, len, resid;
	int clen = 0, error, s, dontroute, mlen;
	int atomic = sosendallatonce(so) || top;
	register int canWait;

	if (uio)
		resid = uio->uio_resid;
	else
		resid = top->m_pkthdr.len;
	/*
	 * In theory resid should be unsigned.
	 * However, space must be signed, as it might be less than 0
	 * if we over-committed, and we must use a signed comparison
	 * of space and resid.  On the other hand, a negative resid
	 * causes us to loop sending 0-length segments to the protocol.
	 */
	if (resid < 0)
		return (EINVAL);
	dontroute =
	    (flags & MSG_DONTROUTE) && (so->so_options & SO_DONTROUTE) == 0 &&
	    (so->so_proto->pr_flags & PR_ATOMIC);

	if (control)
		clen = control->m_len;

	canWait = (so->so_state & SS_NBIO) ? M_DONTWAIT : M_WAIT;

#define	snderr(errno)	{ error = errno; splx(s); goto out; }

	s = splnet();
restart:
	do {
		if (so->so_state & SS_CANTSENDMORE)
			snderr(EPIPE);
		if (so->so_error)
			snderr(so->so_error);
		if ((so->so_state & SS_ISCONNECTED) == 0) {
			if (so->so_proto->pr_flags & PR_CONNREQUIRED) {
				if ((so->so_state & SS_ISCONFIRMING) == 0 &&
				    !(resid == 0 && clen != 0))
					snderr(ENOTCONN);
			} else if (addr == 0)
				snderr(EDESTADDRREQ);
		}
		space = sbspace(&so->so_snd);
		if (flags & MSG_OOB)
			space += 1024;
		if (atomic && resid > so->so_snd.sb_hiwat ||
		    clen > so->so_snd.sb_hiwat)
			snderr(EMSGSIZE);
		if (space < resid + clen && 
		    (atomic || space < so->so_snd.sb_lowat || space < clen)) {
			if (so->so_state & SS_NBIO)
			    {
			    if (flags & MSG_MBUF)
				top = NULL; /* don't free the zero copy mbuf */
			    snderr(EWOULDBLOCK);
			    }
			sbwait(&so->so_snd);
			goto restart;
		}
		mp = &top;
		space -= clen;
		do {
		    if (uio == NULL) {
			/*
			 * Data is prepackaged in "top".
			 */
			resid = 0;
			if (flags & MSG_EOR)
				top->m_flags |= M_EOR;
		    } else do {
		        len = min(resid, space);
			
			if (top == 0) {
			        m = mBufClGet(canWait, MT_DATA, len + max_hdr,
					      FALSE);
				if (m == NULL)
				    snderr(ENOBUFS);
				len = min(len, m->m_extSize);

				m->m_flags |= M_PKTHDR;
				m->m_pkthdr.len = 0;
				m->m_pkthdr.rcvif = (struct ifnet *)0;
                                /*
                                 * the assumption here is that max_hdr is
                                 * always less than the minimum cluster size
                                 * available.  Or don't set len, namely use
                                 * len which set by min() above.
                                 */
                                if (atomic && m->m_extSize > max_hdr) {
                                	len = min((m->m_extSize - max_hdr),
                                                  len);
                                        m->m_data += max_hdr;
                                }
			}
			else
			    {
			    m = mBufClGet(canWait, MT_DATA, len, FALSE);
			    if (m == NULL)
				snderr(ENOBUFS);
			    len = min(len, m->m_extSize);
			    }
                   
			space -= (m->m_extSize + MSIZE);
			error = uiomove(mtod(m, caddr_t), (int)len, uio);
			resid = uio->uio_resid;
			m->m_len = len;
			*mp = m;
			top->m_pkthdr.len += len;
			if (error)
				goto release;
			mp = &m->m_next;
			if (resid <= 0) {
				if (flags & MSG_EOR)
					top->m_flags |= M_EOR;
				break;
			}
		    } while (space > 0 && atomic);
		    if (dontroute)
			    so->so_options |= SO_DONTROUTE;
		    error = (*so->so_proto->pr_usrreq)(so,
			(flags & MSG_OOB) ? PRU_SENDOOB : PRU_SEND,
			top, addr, control);
		    if (dontroute)
			    so->so_options &= ~SO_DONTROUTE;
		    clen = 0;
		    control = 0;
		    top = 0;
		    mp = &top;
		    if (error)
			goto release;
		} while (resid && space > 0);
	} while (resid);

release:
	splx(s);	
out:
	if (top)
		m_freem(top);
	if (control)
		m_freem(control);

	if (error != 0)
	    netErrnoSet (error);

	return (error);
}

/*
 * Implement receive operations on a socket.
 * We depend on the way that records are added to the sockbuf
 * by sbappend*.  In particular, each record (mbufs linked through m_next)
 * must begin with an address if the protocol so specifies,
 * followed by an optional mbuf or mbufs containing ancillary data,
 * and then zero or more mbufs of data.
 * In order to avoid blocking network interrupts for the entire time here,
 * we splx() while doing the actual copy to user space.
 * Although the sockbuf is locked, new data may still be appended,
 * and thus we must maintain consistency of the sockbuf during that time.
 *
 * The caller may receive the data as a single mbuf chain by supplying
 * an mbuf **mp0 for use in returning the chain.  The uio is then used
 * only for the count in uio_resid.
 *
 * WRS mods: implement zero copy if out of band data is requested.
 */
int
soreceive(so, paddr, uio, mp0, controlp, flagsp)
	register struct socket *so;
	struct mbuf **paddr;
	struct uio *uio;
	struct mbuf **mp0;
	struct mbuf **controlp;
	int *flagsp;
{
	register struct mbuf *m, **mp;
	register int flags, len, error = 0, s, offset;
	struct protosw *pr = so->so_proto;
	struct mbuf *nextrecord;
	int moff, type;
	int orig_resid = uio->uio_resid;

	mp = mp0;
	if (paddr)
		*paddr = 0;