-/* uipc_socket.c 6.6 84/08/21 */
-
-#include "../h/param.h"
-#include "../h/systm.h"
-#include "../h/dir.h"
-#include "../h/user.h"
-#include "../h/proc.h"
-#include "../h/file.h"
-#include "../h/inode.h"
-#include "../h/buf.h"
-#include "../h/mbuf.h"
-#include "../h/un.h"
-#include "../h/domain.h"
-#include "../h/protosw.h"
-#include "../h/socket.h"
-#include "../h/socketvar.h"
-#include "../h/stat.h"
-#include "../h/ioctl.h"
-#include "../h/uio.h"
-#include "../net/route.h"
-#include "../netinet/in.h"
-#include "../net/if.h"
+/*
+ * Copyright (c) 1982, 1986 Regents of the University of California.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms are permitted
+ * provided that this notice is preserved and that due credit is given
+ * to the University of California at Berkeley. The name of the University
+ * may not be used to endorse or promote products derived from this
+ * software without specific prior written permission. This software
+ * is provided ``as is'' without express or implied warranty.
+ *
+ * @(#)uipc_socket.c 7.7 (Berkeley) %G%
+ */
+
+#include "param.h"
+#include "dir.h"
+#include "user.h"
+#include "proc.h"
+#include "file.h"
+#include "mbuf.h"
+#include "domain.h"
+#include "protosw.h"
+#include "socket.h"
+#include "socketvar.h"
/*
* Socket operation routines.
* switching out to the protocol specific routines.
*
* TODO:
- * sostat
* test socketpair
- * PR_RIGHTS
- * clean up select, async
+ * clean up async
* out-of-band is a kludge
*/
/*ARGSUSED*/
register int error;
if (proto)
- prp = pffindproto(dom, proto);
+ prp = pffindproto(dom, proto, type);
else
prp = pffindtype(dom, type);
if (prp == 0)
if (prp->pr_type != type)
return (EPROTOTYPE);
m = m_getclr(M_WAIT, MT_SOCKET);
- if (m == 0)
- return (ENOBUFS);
so = mtod(m, struct socket *);
so->so_options = 0;
so->so_state = 0;
so->so_proto = prp;
error =
(*prp->pr_usrreq)(so, PRU_ATTACH,
- (struct mbuf *)0, (struct mbuf *)0, (struct mbuf *)0);
+ (struct mbuf *)0, (struct mbuf *)proto, (struct mbuf *)0);
if (error) {
so->so_state |= SS_NOFDREF;
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;
}
- if (so->so_pcb || (so->so_state & SS_NOFDREF) == 0)
- return;
sbrelease(&so->so_snd);
sorflush(so);
(void) m_free(dtom(so));
goto discard;
if (so->so_state & SS_ISCONNECTED) {
if ((so->so_state & SS_ISDISCONNECTING) == 0) {
- error = sodisconnect(so, (struct mbuf *)0);
+ error = sodisconnect(so);
if (error)
goto drop;
}
register struct socket *so;
struct mbuf *nam;
{
- int s = splnet();
+ int s;
int error;
- if (so->so_state & (SS_ISCONNECTED|SS_ISCONNECTING)) {
+ 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;
- goto bad;
- }
- error = (*so->so_proto->pr_usrreq)(so, PRU_CONNECT,
- (struct mbuf *)0, nam, (struct mbuf *)0);
-bad:
+ else
+ error = (*so->so_proto->pr_usrreq)(so, PRU_CONNECT,
+ (struct mbuf *)0, nam, (struct mbuf *)0);
splx(s);
return (error);
}
return (error);
}
-sodisconnect(so, nam)
+sodisconnect(so)
register struct socket *so;
- struct mbuf *nam;
{
int s = splnet();
int error;
goto bad;
}
error = (*so->so_proto->pr_usrreq)(so, PRU_DISCONNECT,
- (struct mbuf *)0, nam, (struct mbuf *)0);
+ (struct mbuf *)0, (struct mbuf *)0, (struct mbuf *)0);
bad:
splx(s);
return (error);
struct mbuf *top = 0;
register struct mbuf *m, **mp;
register int space;
- int len, error = 0, s, dontroute, first = 1;
+ int len, rlen = 0, error = 0, s, dontroute, first = 1;
if (sosendallatonce(so) && uio->uio_resid > so->so_snd.sb_hiwat)
return (EMSGSIZE);
(flags & MSG_DONTROUTE) && (so->so_options & SO_DONTROUTE) == 0 &&
(so->so_proto->pr_flags & PR_ATOMIC);
u.u_ru.ru_msgsnd++;
+ if (rights)
+ rlen = rights->m_len;
#define snderr(errno) { error = errno; splx(s); goto release; }
restart:
sblock(&so->so_snd);
do {
s = splnet();
- if (so->so_state & SS_CANTSENDMORE) {
- psignal(u.u_procp, SIGPIPE);
+ if (so->so_state & SS_CANTSENDMORE)
snderr(EPIPE);
- }
if (so->so_error) {
error = so->so_error;
so->so_error = 0; /* ??? */
space = 1024;
else {
space = sbspace(&so->so_snd);
- if (space <= 0 ||
- (sosendallatonce(so) && space < uio->uio_resid) ||
+ if (space <= rlen ||
+ (sosendallatonce(so) &&
+ space < uio->uio_resid + rlen) ||
(uio->uio_resid >= CLBYTES && space < CLBYTES &&
so->so_snd.sb_cc >= CLBYTES &&
(so->so_state & SS_NBIO) == 0)) {
}
splx(s);
mp = ⊤
- while (uio->uio_resid > 0 && space > 0) {
- register struct iovec *iov = uio->uio_iov;
-
- if (iov->iov_len == 0) {
- uio->uio_iov++;
- uio->uio_iovcnt--;
- if (uio->uio_iovcnt < 0)
- panic("sosend");
- continue;
- }
+ space -= rlen;
+ while (space > 0) {
MGET(m, M_WAIT, MT_DATA);
- if (m == NULL) {
- error = ENOBUFS; /* SIGPIPE? */
- goto release;
- }
- if (iov->iov_len >= CLBYTES && space >= CLBYTES) {
- register struct mbuf *p;
- MCLGET(p, 1);
- if (p == 0)
+ if (uio->uio_resid >= CLBYTES / 2 && space >= CLBYTES) {
+ MCLGET(m);
+ if (m->m_len != CLBYTES)
goto nopages;
- m->m_off = (int)p - (int)m;
- len = CLBYTES;
+ len = MIN(CLBYTES, uio->uio_resid);
+ space -= CLBYTES;
} else {
nopages:
- len = MIN(MLEN, iov->iov_len);
+ len = MIN(MIN(MLEN, uio->uio_resid), space);
+ space -= len;
}
error = uiomove(mtod(m, caddr_t), len, UIO_WRITE, uio);
m->m_len = len;
if (error)
goto release;
mp = &m->m_next;
- space -= len;
- }
- if (top) {
- if (dontroute)
- so->so_options |= SO_DONTROUTE;
- s = splnet();
- error = (*so->so_proto->pr_usrreq)(so,
- (flags & MSG_OOB) ? PRU_SENDOOB : PRU_SEND,
- top, (caddr_t)nam, rights);
- splx(s);
- if (dontroute)
- so->so_options &= ~SO_DONTROUTE;
+ if (uio->uio_resid <= 0)
+ break;
}
+ if (dontroute)
+ so->so_options |= SO_DONTROUTE;
+ s = splnet(); /* XXX */
+ error = (*so->so_proto->pr_usrreq)(so,
+ (flags & MSG_OOB) ? PRU_SENDOOB : PRU_SEND,
+ top, (caddr_t)nam, rights);
+ splx(s);
+ if (dontroute)
+ so->so_options &= ~SO_DONTROUTE;
+ rights = 0;
+ rlen = 0;
top = 0;
first = 0;
if (error)
sbunlock(&so->so_snd);
if (top)
m_freem(top);
+ if (error == EPIPE)
+ psignal(u.u_procp, SIGPIPE);
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 containing access rights if supported
+ * by the protocol, 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.
+ */
soreceive(so, aname, uio, flags, rightsp)
register struct socket *so;
struct mbuf **aname;
int flags;
struct mbuf **rightsp;
{
- register struct mbuf *m, *n;
- register int len, error = 0, s, tomark;
+ register struct mbuf *m;
+ register int len, error = 0, s, offset;
struct protosw *pr = so->so_proto;
struct mbuf *nextrecord;
int moff;
*aname = 0;
if (flags & MSG_OOB) {
m = m_get(M_WAIT, MT_DATA);
- if (m == 0)
- return (ENOBUFS);
error = (*pr->pr_usrreq)(so, PRU_RCVOOB,
- m, (struct mbuf *)0, (struct mbuf *)0);
+ m, (struct mbuf *)(flags & MSG_PEEK), (struct mbuf *)0);
if (error)
goto bad;
do {
sblock(&so->so_rcv);
s = splnet();
-#define rcverr(errno) { error = errno; splx(s); goto release; }
if (so->so_rcv.sb_cc == 0) {
if (so->so_error) {
error = so->so_error;
so->so_error = 0;
- splx(s);
goto release;
}
- if (so->so_state & SS_CANTRCVMORE) {
- splx(s);
+ if (so->so_state & SS_CANTRCVMORE)
goto release;
- }
if ((so->so_state & SS_ISCONNECTED) == 0 &&
- (so->so_proto->pr_flags & PR_CONNREQUIRED))
- rcverr(ENOTCONN);
- if (so->so_state & SS_NBIO)
- rcverr(EWOULDBLOCK);
+ (so->so_proto->pr_flags & PR_CONNREQUIRED)) {
+ error = ENOTCONN;
+ goto release;
+ }
+ if (uio->uio_resid == 0)
+ goto release;
+ if (so->so_state & SS_NBIO) {
+ error = EWOULDBLOCK;
+ goto release;
+ }
sbunlock(&so->so_rcv);
sbwait(&so->so_rcv);
splx(s);
}
u.u_ru.ru_msgrcv++;
m = so->so_rcv.sb_mb;
+ if (m == 0)
+ panic("receive 1");
+ nextrecord = m->m_act;
if (pr->pr_flags & PR_ADDR) {
- if (m == 0 || m->m_type != MT_SONAME)
+ if (m->m_type != MT_SONAME)
panic("receive 1a");
if (flags & MSG_PEEK) {
if (aname)
*aname = m_copy(m, 0, m->m_len);
- else
- m = m->m_act;
+ m = m->m_next;
} else {
+ sbfree(&so->so_rcv, m);
if (aname) {
*aname = m;
- sbfree(&so->so_rcv, m);
-if(m->m_next) panic("receive 1b");
- so->so_rcv.sb_mb = m = m->m_act;
- } else
- m = sbdroprecord(&so->so_rcv);
+ m = m->m_next;
+ (*aname)->m_next = 0;
+ so->so_rcv.sb_mb = m;
+ } else {
+ MFREE(m, so->so_rcv.sb_mb);
+ m = so->so_rcv.sb_mb;
+ }
+ if (m)
+ m->m_act = nextrecord;
}
}
if (m && m->m_type == MT_RIGHTS) {
if ((pr->pr_flags & PR_RIGHTS) == 0)
- panic("receive 2a");
+ panic("receive 2");
if (flags & MSG_PEEK) {
if (rightsp)
*rightsp = m_copy(m, 0, m->m_len);
- else
- m = m->m_act;
+ m = m->m_next;
} else {
+ sbfree(&so->so_rcv, m);
if (rightsp) {
*rightsp = m;
- sbfree(&so->so_rcv, m);
-if(m->m_next) panic("receive 2b");
- so->so_rcv.sb_mb = m = m->m_act;
- } else
- m = sbdroprecord(&so->so_rcv);
+ so->so_rcv.sb_mb = m->m_next;
+ m->m_next = 0;
+ m = so->so_rcv.sb_mb;
+ } else {
+ MFREE(m, so->so_rcv.sb_mb);
+ m = so->so_rcv.sb_mb;
+ }
+ if (m)
+ m->m_act = nextrecord;
}
}
- if (m == 0)
- panic("receive 3");
moff = 0;
- tomark = so->so_oobmark;
+ offset = 0;
while (m && uio->uio_resid > 0 && error == 0) {
+ if (m->m_type != MT_DATA && m->m_type != MT_HEADER)
+ panic("receive 3");
len = uio->uio_resid;
so->so_state &= ~SS_RCVATMARK;
- if (tomark && len > tomark)
- len = tomark;
- if (moff+len > m->m_len - moff)
+ if (so->so_oobmark && len > so->so_oobmark - offset)
+ len = so->so_oobmark - offset;
+ if (len > m->m_len - moff)
len = m->m_len - moff;
splx(s);
error =
uiomove(mtod(m, caddr_t) + moff, (int)len, UIO_READ, uio);
s = splnet();
- if (len == m->m_len) {
- if ((flags & MSG_PEEK) == 0) {
+ if (len == m->m_len - moff) {
+ if (flags & MSG_PEEK) {
+ m = m->m_next;
+ moff = 0;
+ } else {
nextrecord = m->m_act;
sbfree(&so->so_rcv, m);
- MFREE(m, n);
- if (m = n)
+ MFREE(m, so->so_rcv.sb_mb);
+ m = so->so_rcv.sb_mb;
+ if (m)
m->m_act = nextrecord;
- so->so_rcv.sb_mb = m;
- } else
- m = m->m_next;
- moff = 0;
+ }
} else {
if (flags & MSG_PEEK)
moff += len;
so->so_rcv.sb_cc -= len;
}
}
- if ((flags & MSG_PEEK) == 0 && so->so_oobmark) {
- so->so_oobmark -= len;
- if (so->so_oobmark == 0) {
- so->so_state |= SS_RCVATMARK;
- break;
- }
- }
- if (tomark) {
- tomark -= len;
- if (tomark == 0)
- break;
+ if (so->so_oobmark) {
+ if ((flags & MSG_PEEK) == 0) {
+ so->so_oobmark -= len;
+ if (so->so_oobmark == 0) {
+ so->so_state |= SS_RCVATMARK;
+ break;
+ }
+ } else
+ offset += len;
}
}
if ((flags & MSG_PEEK) == 0) {
if (pr->pr_flags & PR_WANTRCVD && so->so_pcb)
(*pr->pr_usrreq)(so, PRU_RCVD, (struct mbuf *)0,
(struct mbuf *)0, (struct mbuf *)0);
+ if (error == 0 && rightsp && *rightsp &&
+ pr->pr_domain->dom_externalize)
+ error = (*pr->pr_domain->dom_externalize)(*rightsp);
}
release:
sbunlock(&so->so_rcv);
- if (error == 0 && rightsp && *rightsp && pr->pr_domain->dom_externalize)
- error = (*pr->pr_domain->dom_externalize)(*rightsp);
splx(s);
return (error);
}
sbrelease(&asb);
}
-sosetopt(so, level, optname, m)
+sosetopt(so, level, optname, m0)
register struct socket *so;
int level, optname;
- register struct mbuf *m;
+ struct mbuf *m0;
{
+ int error = 0;
+ register struct mbuf *m = m0;
+
+ if (level != SOL_SOCKET) {
+ if (so->so_proto && so->so_proto->pr_ctloutput)
+ return ((*so->so_proto->pr_ctloutput)
+ (PRCO_SETOPT, so, level, optname, &m0));
+ error = ENOPROTOOPT;
+ } else {
+ switch (optname) {
+
+ case SO_LINGER:
+ if (m == NULL || m->m_len != sizeof (struct linger)) {
+ error = EINVAL;
+ goto bad;
+ }
+ so->so_linger = mtod(m, struct linger *)->l_linger;
+ /* fall thru... */
+
+ case SO_DEBUG:
+ case SO_KEEPALIVE:
+ case SO_DONTROUTE:
+ case SO_USELOOPBACK:
+ case SO_BROADCAST:
+ case SO_REUSEADDR:
+ case SO_OOBINLINE:
+ if (m == NULL || m->m_len < sizeof (int)) {
+ error = EINVAL;
+ goto bad;
+ }
+ if (*mtod(m, int *))
+ so->so_options |= optname;
+ else
+ so->so_options &= ~optname;
+ break;
+
+ case SO_SNDBUF:
+ case SO_RCVBUF:
+ case SO_SNDLOWAT:
+ case SO_RCVLOWAT:
+ case SO_SNDTIMEO:
+ case SO_RCVTIMEO:
+ if (m == NULL || m->m_len < sizeof (int)) {
+ error = EINVAL;
+ goto bad;
+ }
+ switch (optname) {
+
+ case SO_SNDBUF:
+ case SO_RCVBUF:
+ if (sbreserve(optname == SO_SNDBUF ? &so->so_snd :
+ &so->so_rcv, *mtod(m, int *)) == 0) {
+ error = ENOBUFS;
+ goto bad;
+ }
+ break;
- if (level != SOL_SOCKET)
- return (EINVAL); /* XXX */
- switch (optname) {
-
- case SO_DEBUG:
- case SO_KEEPALIVE:
- case SO_DONTROUTE:
- case SO_USELOOPBACK:
- case SO_REUSEADDR:
- so->so_options |= optname;
- break;
-
- case SO_LINGER:
- if (m == NULL || m->m_len != sizeof (int))
- return (EINVAL);
- so->so_options |= SO_LINGER;
- so->so_linger = *mtod(m, int *);
- break;
-
- case SO_DONTLINGER:
- so->so_options &= ~SO_LINGER;
- so->so_linger = 0;
- break;
-
- default:
- return (EINVAL);
+ case SO_SNDLOWAT:
+ so->so_snd.sb_lowat = *mtod(m, int *);
+ break;
+ case SO_RCVLOWAT:
+ so->so_rcv.sb_lowat = *mtod(m, int *);
+ break;
+ case SO_SNDTIMEO:
+ so->so_snd.sb_timeo = *mtod(m, int *);
+ break;
+ case SO_RCVTIMEO:
+ so->so_rcv.sb_timeo = *mtod(m, int *);
+ break;
+ }
+ break;
+
+ default:
+ error = ENOPROTOOPT;
+ break;
+ }
}
- return (0);
+bad:
+ if (m)
+ (void) m_free(m);
+ return (error);
}
-sogetopt(so, level, optname, m)
+sogetopt(so, level, optname, mp)
register struct socket *so;
int level, optname;
- register struct mbuf *m;
+ struct mbuf **mp;
{
+ register struct mbuf *m;
- if (level != SOL_SOCKET)
- return (EINVAL); /* XXX */
- switch (optname) {
-
- case SO_USELOOPBACK:
- case SO_DONTROUTE:
- case SO_DEBUG:
- case SO_KEEPALIVE:
- case SO_LINGER:
- case SO_REUSEADDR:
- if ((so->so_options & optname) == 0)
+ if (level != SOL_SOCKET) {
+ if (so->so_proto && so->so_proto->pr_ctloutput) {
+ return ((*so->so_proto->pr_ctloutput)
+ (PRCO_GETOPT, so, level, optname, mp));
+ } else
return (ENOPROTOOPT);
- if (optname == SO_LINGER && m != NULL) {
- *mtod(m, int *) = so->so_linger;
- m->m_len = sizeof (so->so_linger);
- }
- break;
+ } else {
+ m = m_get(M_WAIT, MT_SOOPTS);
+ m->m_len = sizeof (int);
+
+ switch (optname) {
+
+ case SO_LINGER:
+ m->m_len = sizeof (struct linger);
+ mtod(m, struct linger *)->l_onoff =
+ so->so_options & SO_LINGER;
+ mtod(m, struct linger *)->l_linger = so->so_linger;
+ break;
+
+ case SO_USELOOPBACK:
+ case SO_DONTROUTE:
+ case SO_DEBUG:
+ case SO_KEEPALIVE:
+ case SO_REUSEADDR:
+ case SO_BROADCAST:
+ case SO_OOBINLINE:
+ *mtod(m, int *) = so->so_options & optname;
+ break;
+
+ case SO_TYPE:
+ *mtod(m, int *) = so->so_type;
+ break;
+
+ case SO_ERROR:
+ *mtod(m, int *) = so->so_error;
+ so->so_error = 0;
+ break;
+
+ case SO_SNDBUF:
+ *mtod(m, int *) = so->so_snd.sb_hiwat;
+ break;
+
+ case SO_RCVBUF:
+ *mtod(m, int *) = so->so_rcv.sb_hiwat;
+ break;
+
+ case SO_SNDLOWAT:
+ *mtod(m, int *) = so->so_snd.sb_lowat;
+ break;
+
+ case SO_RCVLOWAT:
+ *mtod(m, int *) = so->so_rcv.sb_lowat;
+ break;
+
+ case SO_SNDTIMEO:
+ *mtod(m, int *) = so->so_snd.sb_timeo;
+ break;
- default:
- return (EINVAL);
+ case SO_RCVTIMEO:
+ *mtod(m, int *) = so->so_rcv.sb_timeo;
+ break;
+
+ default:
+ (void)m_free(m);
+ return (ENOPROTOOPT);
+ }
+ *mp = m;
+ return (0);
}
- return (0);
}
sohasoutofband(so)
register struct socket *so;
{
-
- if (so->so_pgrp == 0)
- return;
- if (so->so_pgrp > 0)
- gsignal(so->so_pgrp, SIGURG);
- else {
- struct proc *p = pfind(-so->so_pgrp);
-
- if (p)
- psignal(p, SIGURG);
+ struct proc *p;
+
+ if (so->so_pgrp < 0)
+ gsignal(-so->so_pgrp, SIGURG);
+ else if (so->so_pgrp > 0 && (p = pfind(so->so_pgrp)) != 0)
+ psignal(p, SIGURG);
+ if (so->so_rcv.sb_sel) {
+ selwakeup(so->so_rcv.sb_sel, so->so_rcv.sb_flags & SB_COLL);
+ so->so_rcv.sb_sel = 0;
+ so->so_rcv.sb_flags &= ~SB_COLL;
}
}