void slirp_socket_recv(int addr_low_byte, int guest_port, const uint8_t *buf,
int size)
{
int ret;
struct socket *so = slirp_find_ctl_socket(addr_low_byte, guest_port);
if (!so)
return;
ret = soreadbuf(so, (const char *)buf, size);
if (ret > 0)
tcp_output(sototcpcb(so));
}
/*
* remque and free a socket, clobber cache
*/
void
sofree(PNATState pData, struct socket *so)
{
LogFlowFunc(("ENTER:%R[natsock]\n", so));
/*
* We should not remove socket when polling routine do the polling
* instead we mark it for deletion.
*/
if (so->fUnderPolling)
{
so->fShouldBeRemoved = 1;
LogFlowFunc(("LEAVE:%R[natsock] postponed deletion\n", so));
return;
}
/**
* Check that we don't freeng socket with tcbcb
*/
Assert(!sototcpcb(so));
/* udp checks */
Assert(!so->so_timeout);
Assert(!so->so_timeout_arg);
if (so == tcp_last_so)
tcp_last_so = &tcb;
else if (so == udp_last_so)
udp_last_so = &udb;
/* check if mbuf haven't been already freed */
if (so->so_m != NULL)
{
m_freem(pData, so->so_m);
so->so_m = NULL;
}
if (so->so_ohdr != NULL)
{
RTMemFree(so->so_ohdr);
so->so_ohdr = NULL;
}
if (so->so_next && so->so_prev)
{
remque(pData, so); /* crashes if so is not in a queue */
NSOCK_DEC();
}
RTMemFree(so);
LogFlowFuncLeave();
}
/*
* Tcp protocol timeout routine called every 500 ms.
* Updates the timers in all active tcb's and
* causes finite state machine actions if timers expire.
*/
void
tcp_slowtimo()
{
register struct socket *ip, *ipnxt;
register struct tcpcb *tp;
register int i;
DEBUG_CALL("tcp_slowtimo");
tcp_maxidle = TCPTV_KEEPCNT * tcp_keepintvl;
/*
* Search through tcb's and update active timers.
*/
ip = tcb.so_next;
if (ip == 0)
return;
for (; ip != &tcb; ip = ipnxt) {
ipnxt = ip->so_next;
tp = sototcpcb(ip);
if (tp == 0)
continue;
for (i = 0; i < TCPT_NTIMERS; i++) {
if (tp->t_timer[i] && --tp->t_timer[i] == 0) {
tcp_timers(tp,i);
if (ipnxt->so_prev != ip)
goto tpgone;
}
}
tp->t_idle++;
if (tp->t_rtt)
tp->t_rtt++;
tpgone:
;
}
tcp_iss += TCP_ISSINCR/PR_SLOWHZ; /* increment iss */
#ifdef TCP_COMPAT_42
if ((int)tcp_iss < 0)
tcp_iss = 0; /* XXX */
#endif
tcp_now++; /* for timestamps */
}
int soreadbuf(struct socket *so, const char *buf, int size)
{
int n, nn, copy = size;
struct sbuf *sb = &so->so_snd;
struct iovec iov[2];
DEBUG_CALL("soreadbuf");
DEBUG_ARG("so = %lx", (long )so);
/*
* No need to check if there's enough room to read.
* soread wouldn't have been called if there weren't
*/
if (sopreprbuf(so, iov, &n) < size)
goto err;
nn = MIN(iov[0].iov_len, copy);
memcpy(iov[0].iov_base, buf, nn);
copy -= nn;
buf += nn;
if (copy == 0)
goto done;
memcpy(iov[1].iov_base, buf, copy);
done:
/* Update fields */
sb->sb_cc += size;
sb->sb_wptr += size;
if (sb->sb_wptr >= (sb->sb_data + sb->sb_datalen))
sb->sb_wptr -= sb->sb_datalen;
return size;
err:
sofcantrcvmore(so);
tcp_sockclosed(sototcpcb(so));
fprintf(stderr, "soreadbuf buffer to small");
return -1;
}
/*
* Get urgent data
*
* When the socket is created, we set it SO_OOBINLINE,
* so when OOB data arrives, we soread() it and everything
* in the send buffer is sent as urgent data
*/
void sorecvoob(struct socket *so)
{
struct tcpcb *tp = sototcpcb(so);
DEBUG_CALL("sorecvoob");
DEBUG_ARG("so = %p", so);
/*
* We take a guess at how much urgent data has arrived.
* In most situations, when urgent data arrives, the next
* read() should get all the urgent data. This guess will
* be wrong however if more data arrives just after the
* urgent data, or the read() doesn't return all the
* urgent data.
*/
soread(so);
tp->snd_up = tp->snd_una + so->so_snd.sb_cc;
tp->t_force = 1;
tcp_output(tp);
tp->t_force = 0;
}
void tcp_pulloutofband(struct socket *so, struct tcpiphdr *ti, struct mbuf *m)
{
int cnt = ti->ti_urp - 1;
while (cnt >= 0) {
if (m->m_len > cnt) {
char *cp = mtod(m, caddr_t) + cnt;
struct tcpcb *tp = sototcpcb(so);
tp->t_iobc = *cp;
tp->t_oobflags |= TCPOOB_HAVEDATA;
memcpy(sp, cp+1, (unsigned)(m->m_len - cnt - 1));
m->m_len--;
return;
}
cnt -= m->m_len;
m = m->m_next; /* XXX WRONG! Fix it! */
if (m == 0)
break;
}
panic("tcp_pulloutofband");
}
/*
* Tcp protocol timeout routine called every 500 ms.
* Updates the timers in all active tcb's and
* causes finite state machine actions if timers expire.
*/
void
tcp_slowtimo(Slirp *slirp)
{
register struct socket *ip, *ipnxt;
register struct tcpcb *tp;
register int i;
DEBUG_CALL("tcp_slowtimo");
/*
* Search through tcb's and update active timers.
*/
ip = slirp->tcb.so_next;
if (ip == NULL) {
return;
}
for (; ip != &slirp->tcb; ip = ipnxt) {
ipnxt = ip->so_next;
tp = sototcpcb(ip);
if (tp == NULL) {
continue;
}
for (i = 0; i < TCPT_NTIMERS; i++) {
if (tp->t_timer[i] && --tp->t_timer[i] == 0) {
tcp_timers(tp,i);
if (ipnxt->so_prev != ip)
goto tpgone;
}
}
tp->t_idle++;
if (tp->t_rtt)
tp->t_rtt++;
tpgone:
;
}
slirp->tcp_iss += TCP_ISSINCR/PR_SLOWHZ; /* increment iss */
slirp->tcp_now++; /* for timestamps */
}
/*
* 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
* sequencing purposes.
*/
void
tcp_pulloutofband(struct usn_socket *so, struct tcpiphdr *ti, usn_mbuf_t *m)
{
int cnt = ti->ti_urp - 1;
while (cnt >= 0) {
if (m->mlen > cnt) {
char *cp = mtod(m, caddr_t) + cnt;
struct tcpcb *tp = sototcpcb(so);
tp->t_iobc = *cp;
tp->t_oobflags |= TCPOOB_HAVEDATA;
bcopy(cp+1, cp, (unsigned)(m->mlen - cnt - 1));
m->mlen--;
return;
}
cnt -= m->mlen;
m = m->next;
if (m == 0)
break;
}
DEBUG("panic: tcp_pulloutofband");
}
void slirp_select_poll(fd_set *readfds, fd_set *writefds, fd_set *xfds)
{
struct socket *so, *so_next;
int ret;
global_readfds = readfds;
global_writefds = writefds;
global_xfds = xfds;
/* Update time */
updtime();
/*
* See if anything has timed out
*/
if (link_up) {
if (time_fasttimo && ((curtime - time_fasttimo) >= FAST_TIMO)) {
tcp_fasttimo();
time_fasttimo = 0;
}
if (do_slowtimo && ((curtime - last_slowtimo) >= SLOW_TIMO)) {
ip_slowtimo();
tcp_slowtimo();
last_slowtimo = curtime;
}
}
/*
* Check sockets
*/
if (link_up) {
/*
* Check TCP sockets
*/
for (so = tcb.so_next; so != &tcb; so = so_next) {
so_next = so->so_next;
/*
* FD_ISSET is meaningless on these sockets
* (and they can crash the program)
*/
if (so->so_state & SS_NOFDREF || so->s == -1)
continue;
/*
* Check for URG data
* This will soread as well, so no need to
* test for readfds below if this succeeds
*/
if (FD_ISSET(so->s, xfds))
sorecvoob(so);
/*
* Check sockets for reading
*/
else if (FD_ISSET(so->s, readfds)) {
/*
* Check for incoming connections
*/
if (so->so_state & SS_FACCEPTCONN) {
tcp_connect(so);
continue;
} /* else */
ret = soread(so);
/* Output it if we read something */
if (ret > 0)
tcp_output(sototcpcb(so));
}
/*
* Check sockets for writing
*/
if (FD_ISSET(so->s, writefds)) {
/*
* Check for non-blocking, still-connecting sockets
*/
if (so->so_state & SS_ISFCONNECTING) {
/* Connected */
so->so_state &= ~SS_ISFCONNECTING;
ret = send(so->s, &ret, 0, 0);
if (ret < 0) {
/* XXXXX Must fix, zero bytes is a NOP */
if (errno == EAGAIN || errno == EWOULDBLOCK ||
errno == EINPROGRESS || errno == ENOTCONN)
continue;
/* else failed */
so->so_state = SS_NOFDREF;
}
/* else so->so_state &= ~SS_ISFCONNECTING; */
/*
* Continue tcp_input
*/
tcp_input((struct mbuf *)NULL, sizeof(struct ip), so);
/* continue; */
} else
ret = sowrite(so);
/*
* XXXXX If we wrote something (a lot), there
* could be a need for a window update.
* In the worst case, the remote will send
* a window probe to get things going again
*/
}
/*
* Probe a still-connecting, non-blocking socket
* to check if it's still alive
*/
#ifdef PROBE_CONN
if (so->so_state & SS_ISFCONNECTING) {
ret = recv(so->s, (char *)&ret, 0,0);
if (ret < 0) {
/* XXX */
if (errno == EAGAIN || errno == EWOULDBLOCK ||
errno == EINPROGRESS || errno == ENOTCONN)
continue; /* Still connecting, continue */
/* else failed */
so->so_state = SS_NOFDREF;
/* tcp_input will take care of it */
} else {
ret = send(so->s, &ret, 0,0);
if (ret < 0) {
/* XXX */
if (errno == EAGAIN || errno == EWOULDBLOCK ||
errno == EINPROGRESS || errno == ENOTCONN)
continue;
/* else failed */
so->so_state = SS_NOFDREF;
} else
so->so_state &= ~SS_ISFCONNECTING;
}
tcp_input((struct mbuf *)NULL, sizeof(struct ip),so);
} /* SS_ISFCONNECTING */
#endif
}
/*
* Now UDP sockets.
* Incoming packets are sent straight away, they're not buffered.
* Incoming UDP data isn't buffered either.
*/
for (so = udb.so_next; so != &udb; so = so_next) {
so_next = so->so_next;
if (so->s != -1 && FD_ISSET(so->s, readfds)) {
sorecvfrom(so);
}
}
}
/*
* See if we can start outputting
*/
if (if_queued && link_up)
if_start();
/* clear global file descriptor sets.
* these reside on the stack in vl.c
* so they're unusable if we're not in
* slirp_select_fill or slirp_select_poll.
*/
global_readfds = NULL;
global_writefds = NULL;
global_xfds = NULL;
}
/*
* Read from so's socket into sb_snd, updating all relevant sbuf fields
* NOTE: This will only be called if it is select()ed for reading, so
* a read() of 0 (or less) means it's disconnected
*/
int
soread(PNATState pData, struct socket *so)
{
int n, nn, lss, total;
struct sbuf *sb = &so->so_snd;
u_int len = sb->sb_datalen - sb->sb_cc;
struct iovec iov[2];
int mss = so->so_tcpcb->t_maxseg;
int sockerr;
STAM_PROFILE_START(&pData->StatIOread, a);
STAM_COUNTER_RESET(&pData->StatIORead_in_1);
STAM_COUNTER_RESET(&pData->StatIORead_in_2);
QSOCKET_LOCK(tcb);
SOCKET_LOCK(so);
QSOCKET_UNLOCK(tcb);
LogFlow(("soread: so = %R[natsock]\n", so));
Log2(("%s: so = %R[natsock] so->so_snd = %R[sbuf]\n", RT_GCC_EXTENSION __PRETTY_FUNCTION__, so, sb));
/*
* No need to check if there's enough room to read.
* soread wouldn't have been called if there weren't
*/
len = sb->sb_datalen - sb->sb_cc;
iov[0].iov_base = sb->sb_wptr;
iov[1].iov_base = 0;
iov[1].iov_len = 0;
if (sb->sb_wptr < sb->sb_rptr)
{
iov[0].iov_len = sb->sb_rptr - sb->sb_wptr;
/* Should never succeed, but... */
if (iov[0].iov_len > len)
iov[0].iov_len = len;
if (iov[0].iov_len > mss)
iov[0].iov_len -= iov[0].iov_len%mss;
n = 1;
}
else
{
iov[0].iov_len = (sb->sb_data + sb->sb_datalen) - sb->sb_wptr;
/* Should never succeed, but... */
if (iov[0].iov_len > len)
iov[0].iov_len = len;
len -= iov[0].iov_len;
if (len)
{
iov[1].iov_base = sb->sb_data;
iov[1].iov_len = sb->sb_rptr - sb->sb_data;
if (iov[1].iov_len > len)
iov[1].iov_len = len;
total = iov[0].iov_len + iov[1].iov_len;
if (total > mss)
{
lss = total % mss;
if (iov[1].iov_len > lss)
{
iov[1].iov_len -= lss;
n = 2;
}
else
{
lss -= iov[1].iov_len;
iov[0].iov_len -= lss;
n = 1;
}
}
else
n = 2;
}
else
{
if (iov[0].iov_len > mss)
iov[0].iov_len -= iov[0].iov_len%mss;
n = 1;
}
}
#ifdef HAVE_READV
nn = readv(so->s, (struct iovec *)iov, n);
#else
nn = recv(so->s, iov[0].iov_base, iov[0].iov_len, (so->so_tcpcb->t_force? MSG_OOB:0));
#endif
if (nn < 0)
sockerr = errno; /* save it, as it may be clobbered by logging */
else
sockerr = 0;
Log2(("%s: read(1) nn = %d bytes\n", RT_GCC_EXTENSION __PRETTY_FUNCTION__, nn));
Log2(("%s: so = %R[natsock] so->so_snd = %R[sbuf]\n", RT_GCC_EXTENSION __PRETTY_FUNCTION__, so, sb));
if (nn <= 0)
{
#ifdef RT_OS_WINDOWS
/*
* Windows reports ESHUTDOWN after SHUT_RD (SD_RECEIVE)
* instead of just returning EOF indication.
*/
if (nn < 0 && sockerr == ESHUTDOWN)
{
nn = 0;
sockerr = 0;
}
#endif
if (nn == 0) /* XXX: should this be inside #if defined(RT_OS_WINDOWS)? */
{
/*
* Special case for WSAEnumNetworkEvents: If we receive 0 bytes that
* _could_ mean that the connection is closed. But we will receive an
* FD_CLOSE event later if the connection was _really_ closed. With
* www.youtube.com I see this very often. Closing the socket too early
* would be dangerous.
*/
int status;
unsigned long pending = 0;
status = ioctlsocket(so->s, FIONREAD, &pending);
if (status < 0)
Log(("NAT:%s: error in WSAIoctl: %d\n", RT_GCC_EXTENSION __PRETTY_FUNCTION__, errno));
if (pending != 0)
{
SOCKET_UNLOCK(so);
STAM_PROFILE_STOP(&pData->StatIOread, a);
return 0;
}
}
if ( nn < 0
&& soIgnorableErrorCode(sockerr))
{
SOCKET_UNLOCK(so);
STAM_PROFILE_STOP(&pData->StatIOread, a);
return 0;
}
else
{
int fUninitializedTemplate = 0;
int shuterr;
fUninitializedTemplate = RT_BOOL(( sototcpcb(so)
&& ( sototcpcb(so)->t_template.ti_src.s_addr == INADDR_ANY
|| sototcpcb(so)->t_template.ti_dst.s_addr == INADDR_ANY)));
/* nn == 0 means peer has performed an orderly shutdown */
Log2(("%s: disconnected, nn = %d, errno = %d (%s)\n",
RT_GCC_EXTENSION __PRETTY_FUNCTION__, nn, sockerr, strerror(sockerr)));
shuterr = sofcantrcvmore(so);
if (!sockerr && !shuterr && !fUninitializedTemplate)
tcp_sockclosed(pData, sototcpcb(so));
else
{
LogRel2(("NAT: sockerr %d, shuterr %d - %R[natsock]\n", sockerr, shuterr, so));
tcp_drop(pData, sototcpcb(so), sockerr);
}
SOCKET_UNLOCK(so);
STAM_PROFILE_STOP(&pData->StatIOread, a);
return -1;
}
}
STAM_STATS(
if (n == 1)
{
STAM_COUNTER_INC(&pData->StatIORead_in_1);
STAM_COUNTER_ADD(&pData->StatIORead_in_1_bytes, nn);
}
else
{
STAM_COUNTER_INC(&pData->StatIORead_in_2);
STAM_COUNTER_ADD(&pData->StatIORead_in_2_1st_bytes, nn);
}
);
__private_extern__ void
inpcb_get_ports_used(uint32_t ifindex, int protocol, uint32_t flags,
bitstr_t *bitfield, struct inpcbinfo *pcbinfo)
{
struct inpcb *inp;
struct socket *so;
inp_gen_t gencnt;
bool iswildcard, wildcardok, nowakeok;
bool recvanyifonly, extbgidleok;
bool activeonly;
wildcardok = ((flags & INPCB_GET_PORTS_USED_WILDCARDOK) != 0);
nowakeok = ((flags & INPCB_GET_PORTS_USED_NOWAKEUPOK) != 0);
recvanyifonly = ((flags & INPCB_GET_PORTS_USED_RECVANYIFONLY) != 0);
extbgidleok = ((flags & INPCB_GET_PORTS_USED_EXTBGIDLEONLY) != 0);
activeonly = ((flags & INPCB_GET_PORTS_USED_ACTIVEONLY) != 0);
lck_rw_lock_shared(pcbinfo->ipi_lock);
gencnt = pcbinfo->ipi_gencnt;
for (inp = LIST_FIRST(pcbinfo->ipi_listhead); inp;
inp = LIST_NEXT(inp, inp_list)) {
uint16_t port;
if (inp->inp_gencnt > gencnt ||
inp->inp_state == INPCB_STATE_DEAD ||
inp->inp_wantcnt == WNT_STOPUSING)
continue;
if ((so = inp->inp_socket) == NULL ||
(so->so_state & SS_DEFUNCT) ||
(so->so_state & SS_ISDISCONNECTED))
continue;
if (!(protocol == PF_UNSPEC ||
(protocol == PF_INET && (inp->inp_vflag & INP_IPV4)) ||
(protocol == PF_INET6 && (inp->inp_vflag & INP_IPV6))))
continue;
iswildcard = (((inp->inp_vflag & INP_IPV4) &&
inp->inp_laddr.s_addr == INADDR_ANY) ||
((inp->inp_vflag & INP_IPV6) &&
IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)));
if (!wildcardok && iswildcard)
continue;
if ((so->so_options & SO_NOWAKEFROMSLEEP) &&
!nowakeok)
continue;
if (!(inp->inp_flags & INP_RECV_ANYIF) &&
recvanyifonly)
continue;
if (!(so->so_flags1 & SOF1_EXTEND_BK_IDLE_WANTED) &&
extbgidleok)
continue;
if (!iswildcard &&
!(ifindex == 0 || inp->inp_last_outifp == NULL ||
ifindex == inp->inp_last_outifp->if_index))
continue;
if (SOCK_PROTO(inp->inp_socket) == IPPROTO_UDP &&
so->so_state & SS_CANTRCVMORE)
continue;
if (SOCK_PROTO(inp->inp_socket) == IPPROTO_TCP) {
struct tcpcb *tp = sototcpcb(inp->inp_socket);
/*
* Workaround race where inp_ppcb is NULL during
* socket initialization
*/
if (tp == NULL)
continue;
switch (tp->t_state) {
case TCPS_CLOSED:
continue;
/* NOT REACHED */
case TCPS_LISTEN:
case TCPS_SYN_SENT:
case TCPS_SYN_RECEIVED:
case TCPS_ESTABLISHED:
case TCPS_FIN_WAIT_1:
/*
* Note: FIN_WAIT_1 is an active state
* because we need our FIN to be
* acknowledged
*/
break;
case TCPS_CLOSE_WAIT:
case TCPS_CLOSING:
case TCPS_LAST_ACK:
case TCPS_FIN_WAIT_2:
/*
* In the closing states, the connection
* is not idle when there is outgoing
* data having to be acknowledged
*/
if (activeonly && so->so_snd.sb_cc == 0)
continue;
break;
case TCPS_TIME_WAIT:
continue;
/* NOT REACHED */
}
}
/*
* Final safeguard to exclude unspecified local port
*/
port = ntohs(inp->inp_lport);
if (port == 0)
continue;
bit_set(bitfield, port);
}
lck_rw_done(pcbinfo->ipi_lock);
}
int
tcp_usrreq(struct socket * so,
struct mbuf * m,
struct mbuf * nam)
{
struct inpcb * inp;
struct tcpcb * tp;
int error = 0;
int req;
#ifdef DO_TCPTRACE
int ostate;
#endif
req = so->so_req; /* get request from socket struct */
inp = sotoinpcb(so);
/*
* 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)
{
return (EINVAL);
}
if (inp)
tp = intotcpcb(inp);
else /* inp and tp not set, make sure this is OK: */
{
if (req == PRU_ATTACH)
tp = NULL; /* stifle compiler warnings about using unassigned tp*/
else
{
dtrap(); /* programming error? */
return EINVAL;
}
}
switch (req)
{
/*
* TCP attaches to socket via PRU_ATTACH, reserving space,
* and an internet control block.
*/
case PRU_ATTACH:
if (inp)
{
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;
#ifdef DO_TCPTRACE
SETTP(tp, sototcpcb(so));
#endif
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)
SETTP(tp, tcp_disconnect(tp));
else
SETTP(tp, tcp_close(tp));
break;
/*
* Give the socket an address.
*/
case PRU_BIND:
/* bind is quite different for IPv4 and v6, so we use two
* seperate pcbbind routines. so_domain was checked for
* validity way up in t_bind()
*/
#ifdef IP_V4
if(inp->inp_socket->so_domain == AF_INET)
{
error = in_pcbbind(inp, nam);
break;
}
#endif /* IP_V4 */
#ifdef IP_V6
if(inp->inp_socket->so_domain == AF_INET6)
{
error = ip6_pcbbind(inp, nam);
break;
}
#endif /* IP_V6 */
dtrap(); /* not v4 or v6? */
error = EINVAL;
break;
/*
* Prepare to accept connections.
*/
case PRU_LISTEN:
if (inp->inp_lport == 0)
error = in_pcbbind(inp, (struct mbuf *)0);
if (error == 0)
tp->t_state = TCPS_LISTEN;
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)
{
#ifdef IP_V4
#ifndef IP_V6 /* v4 only */
error = in_pcbbind(inp, (struct mbuf *)0);
#else /* dual mode */
if(so->so_domain == AF_INET)
error = in_pcbbind(inp, (struct mbuf *)0);
else
error = ip6_pcbbind(inp, (struct mbuf *)0);
#endif /* end dual mode code */
#else /* no v4, v6 only */
error = ip6_pcbbind(inp, (struct mbuf *)0);
#endif /* end v6 only */
if (error)
break;
}
#ifdef IP_V4
#ifndef IP_V6 /* v4 only */
error = in_pcbconnect(inp, nam);
#else /* dual mode */
if(so->so_domain == AF_INET)
error = in_pcbconnect(inp, nam);
else
error = ip6_pcbconnect(inp, nam);
#endif /* end dual mode code */
#else /* no v4, v6 only */
error = ip6_pcbconnect(inp, nam);
#endif /* end v6 only */
if (error)
break;
tp->t_template = tcp_template(tp);
if (tp->t_template == 0)
{
#ifdef IP_V4
#ifndef IP_V6 /* v4 only */
in_pcbdisconnect(inp);
#else /* dual mode */
if(so->so_domain == AF_INET)
in_pcbdisconnect(inp);
else
ip6_pcbdisconnect(inp);
#endif /* end dual mode code */
#else /* no v4, v6 only */
ip6_pcbdisconnect(inp);
#endif /* end v6 only */
error = ENOBUFS;
break;
}
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_seq)(TCP_ISSINCR/2);
tcp_sendseqinit(tp);
error = tcp_output(tp);
if (!error)
TCP_MIB_INC(tcpActiveOpens); /* keep MIB stats */
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:
SETTP(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:
{
struct sockaddr_in * sin = mtod(nam, struct sockaddr_in *);
#ifdef IP_V6
struct sockaddr_in6 * sin6 = mtod(nam, struct sockaddr_in6 *);
#endif
#ifdef IP_V6
if (so->so_domain == AF_INET6)
{
nam->m_len = sizeof (struct sockaddr_in6);
sin6->sin6_port = inp->inp_fport;
sin6->sin6_family = AF_INET6;
IP6CPY(&sin6->sin6_addr, &inp->ip6_faddr);
}
#endif
#ifdef IP_V4
if (so->so_domain == AF_INET)
{
nam->m_len = sizeof (struct sockaddr_in);
sin->sin_family = AF_INET;
sin->sin_port = inp->inp_fport;
sin->sin_addr = inp->inp_faddr;
}
#endif
if ( !(so->so_domain == AF_INET) &&
!(so->so_domain == AF_INET6)
)
{
dprintf("*** PRU_ACCEPT bad domain = %d\n", so->so_domain);
dtrap();
}
TCP_MIB_INC(tcpPassiveOpens); /* keep MIB stats */
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 (so->so_pcb == NULL)
{ /* Return EPIPE error if socket is not connected */
error = EPIPE;
break;
}
sbappend(&so->so_snd, m);
error = tcp_output(tp);
if (error == ENOBUFS)
sbdropend(&so->so_snd,m); /* Remove data from socket buffer */
break;
/*
* Abort the TCP.
*/
case PRU_ABORT:
SETTP(tp, tcp_drop(tp, ECONNABORTED));
break;
case PRU_SENSE:
/* ((struct stat *) m)->st_blksize = so->so_snd.sb_hiwat; */
dtrap(); /* does this ever happen? */
return (0);
case PRU_RCVOOB:
if ((so->so_oobmark == 0 &&
(so->so_state & SS_RCVATMARK) == 0) ||
#ifdef SO_OOBINLINE
so->so_options & SO_OOBINLINE ||
#endif
tp->t_oobflags & TCPOOB_HADDATA)
{
error = EINVAL;
break;
}
if ((tp->t_oobflags & TCPOOB_HAVEDATA) == 0)
{
error = EWOULDBLOCK;
break;
}
m->m_len = 1;
*mtod(m, char *) = tp->t_iobc;
if ((MBUF2LONG(nam) & MSG_PEEK) == 0)
tp->t_oobflags ^= (TCPOOB_HAVEDATA | TCPOOB_HADDATA);
break;
case PRU_SENDOOB:
if (so->so_pcb == NULL)
{ /* Return EPIPE error if socket is not connected */
error = EPIPE;
break;
}
if (sbspace(&so->so_snd) == 0)
{
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);
if (error == ENOBUFS)
sbdropend(&so->so_snd,m); /* Remove data from socket buffer */
tp->t_force = 0;
break;
case PRU_SOCKADDR:
/* sockaddr and peeraddr have to switch based on IP type */
#ifdef IP_V4
#ifndef IP_V6 /* v4 only */
in_setsockaddr(inp, nam);
#else /* dual mode */
if(so->so_domain == AF_INET6)
ip6_setsockaddr(inp, nam);
else
in_setsockaddr(inp, nam);
#endif /* dual mode */
#else /* IP_V6 */
ip6_setsockaddr(inp, nam);
#endif
break;
case PRU_PEERADDR:
#ifdef IP_V4
#ifndef IP_V6 /* v4 only */
in_setpeeraddr(inp, nam);
#else /* dual mode */
if(so->so_domain == AF_INET6)
ip6_setpeeraddr(inp, nam);
else
in_setpeeraddr(inp, nam);
#endif /* dual mode */
#else /* IP_V6 */
ip6_setpeeraddr(inp, nam);
#endif
break;
case PRU_SLOWTIMO:
SETTP(tp, tcp_timers(tp, (int)MBUF2LONG(nam)));
#ifdef DO_TCPTRACE
req |= (long)nam << 8; /* for debug's sake */
#endif
break;
default:
panic("tcp_usrreq");
}
#ifdef DO_TCPTRACE
if (tp && (so->so_options & SO_DEBUG))
tcp_trace("usrreq: state: %d, tcpcb: %x, req: %d",
ostate, tp, req);
#endif
return (error);
}
/*
* Read from so's socket into sb_snd, updating all relevant sbuf fields
* NOTE: This will only be called if it is select()ed for reading, so
* a read() of 0 (or less) means it's disconnected
*/
int
soread(struct socket *so)
{
int n, nn, lss, total;
struct sbuf *sb = &so->so_snd;
int len = sb->sb_datalen - sb->sb_cc;
struct iovec iov[2];
int mss = so->so_tcpcb->t_maxseg;
DEBUG_CALL("soread");
DEBUG_ARG("so = %lx", (long )so);
/*
* No need to check if there's enough room to read.
* soread wouldn't have been called if there weren't
*/
len = sb->sb_datalen - sb->sb_cc;
iov[0].iov_base = sb->sb_wptr;
if (sb->sb_wptr < sb->sb_rptr) {
iov[0].iov_len = sb->sb_rptr - sb->sb_wptr;
/* Should never succeed, but... */
if (iov[0].iov_len > len)
iov[0].iov_len = len;
if (iov[0].iov_len > mss)
iov[0].iov_len -= iov[0].iov_len%mss;
n = 1;
} else {
iov[0].iov_len = (sb->sb_data + sb->sb_datalen) - sb->sb_wptr;
/* Should never succeed, but... */
if (iov[0].iov_len > len) iov[0].iov_len = len;
len -= iov[0].iov_len;
if (len) {
iov[1].iov_base = sb->sb_data;
iov[1].iov_len = sb->sb_rptr - sb->sb_data;
if(iov[1].iov_len > len)
iov[1].iov_len = len;
total = iov[0].iov_len + iov[1].iov_len;
if (total > mss) {
lss = total%mss;
if (iov[1].iov_len > lss) {
iov[1].iov_len -= lss;
n = 2;
} else {
lss -= iov[1].iov_len;
iov[0].iov_len -= lss;
n = 1;
}
} else
n = 2;
} else {
if (iov[0].iov_len > mss)
iov[0].iov_len -= iov[0].iov_len%mss;
n = 1;
}
}
#ifdef HAVE_READV
nn = readv(so->s, (struct iovec *)iov, n);
DEBUG_MISC((dfd, " ... read nn = %d bytes\n", nn));
#else
nn = recv(so->s, iov[0].iov_base, iov[0].iov_len,0);
#endif
if (nn <= 0) {
if (nn < 0 && (errno == EINTR || errno == EAGAIN))
return 0;
else {
DEBUG_MISC((dfd, " --- soread() disconnected, nn = %d, errno = %d-%s\n", nn, errno,strerror(errno)));
sofcantrcvmore(so);
tcp_sockclosed(sototcpcb(so));
return -1;
}
}
#ifndef HAVE_READV
/*
* If there was no error, try and read the second time round
* We read again if n = 2 (ie, there's another part of the buffer)
* and we read as much as we could in the first read
* We don't test for <= 0 this time, because there legitimately
* might not be any more data (since the socket is non-blocking),
* a close will be detected on next iteration.
* A return of -1 wont (shouldn't) happen, since it didn't happen above
*/
if (n == 2 && nn == iov[0].iov_len) {
int ret;
ret = recv(so->s, iov[1].iov_base, iov[1].iov_len,0);
if (ret > 0)
nn += ret;
}
DEBUG_MISC((dfd, " ... read nn = %d bytes\n", nn));
#endif
/* Update fields */
sb->sb_cc += nn;
sb->sb_wptr += nn;
if (sb->sb_wptr >= (sb->sb_data + sb->sb_datalen))
sb->sb_wptr -= sb->sb_datalen;
return nn;
}
void tcp_cleanup(void)
{
while (tcb.so_next != &tcb) {
tcp_close(sototcpcb(tcb.so_next));
}
}
void slirp_select_poll(fd_set *readfds, fd_set *writefds, fd_set *xfds)
{
struct socket *so, *so_next;
int ret;
global_readfds = readfds;
global_writefds = writefds;
global_xfds = xfds;
updtime();
if (link_up) {
if (time_fasttimo && ((curtime - time_fasttimo) >= 2)) {
tcp_fasttimo();
time_fasttimo = 0;
}
if (do_slowtimo && ((curtime - last_slowtimo) >= 499)) {
ip_slowtimo();
tcp_slowtimo();
last_slowtimo = curtime;
}
}
if (link_up) {
for (so = tcb.so_next; so != &tcb; so = so_next) {
so_next = so->so_next;
if (so->so_state & SS_NOFDREF || so->s == -1)
continue;
if ((so->so_state & SS_PROXIFIED) != 0)
continue;
if (FD_ISSET(so->s, xfds))
sorecvoob(so);
else if (FD_ISSET(so->s, readfds)) {
if (so->so_state & SS_FACCEPTCONN) {
tcp_connect(so);
continue;
}
ret = soread(so);
if (ret > 0)
tcp_output(sototcpcb(so));
}
if (FD_ISSET(so->s, writefds)) {
if (so->so_state & SS_ISFCONNECTING) {
so->so_state &= ~SS_ISFCONNECTING;
ret = socket_send(so->s, (const void *)&ret, 0);
if (ret < 0) {
if (errno == EAGAIN || errno == EWOULDBLOCK ||
errno == EINPROGRESS || errno == ENOTCONN)
continue;
so->so_state = SS_NOFDREF;
}
tcp_input((struct mbuf *)NULL, sizeof(struct ip), so);
} else
ret = sowrite(so);
}
#ifdef PROBE_CONN
if (so->so_state & SS_ISFCONNECTING) {
ret = socket_recv(so->s, (char *)&ret, 0);
if (ret < 0) {
if (errno == EAGAIN || errno == EWOULDBLOCK ||
errno == EINPROGRESS || errno == ENOTCONN)
continue;
so->so_state = SS_NOFDREF;
} else {
ret = socket_send(so->s, &ret, 0);
if (ret < 0) {
if (errno == EAGAIN || errno == EWOULDBLOCK ||
errno == EINPROGRESS || errno == ENOTCONN)
continue;
so->so_state = SS_NOFDREF;
} else
so->so_state &= ~SS_ISFCONNECTING;
}
tcp_input((struct mbuf *)NULL, sizeof(struct ip),so);
}
#endif
}
for (so = udb.so_next; so != &udb; so = so_next) {
so_next = so->so_next;
if ((so->so_state & SS_PROXIFIED) != 0)
continue;
if (so->s != -1 && FD_ISSET(so->s, readfds)) {
sorecvfrom(so);
}
}
}
proxy_manager_poll(readfds, writefds, xfds);
if (if_queued && link_up)
if_start();
global_readfds = NULL;
global_writefds = NULL;
global_xfds = NULL;
}
/*
* TCP input routine, follows pages 65-76 of the
* protocol specification dated September, 1981 very closely.
*/
void
tcp_input(struct mbuf *m, int iphlen, struct socket *inso)
{
struct ip save_ip, *ip;
register struct tcpiphdr *ti;
caddr_t optp = NULL;
int optlen = 0;
int len, tlen, off;
register struct tcpcb *tp = NULL;
register int tiflags;
struct socket *so = NULL;
int todrop, acked, ourfinisacked, needoutput = 0;
int iss = 0;
u_long tiwin;
int ret;
struct ex_list *ex_ptr;
Slirp *slirp;
DEBUG_CALL("tcp_input");
DEBUG_ARGS((dfd, " m = %8lx iphlen = %2d inso = %lx\n",
(long )m, iphlen, (long )inso ));
/*
* If called with m == 0, then we're continuing the connect
*/
if (m == NULL) {
so = inso;
slirp = so->slirp;
/* Re-set a few variables */
tp = sototcpcb(so);
m = so->so_m;
so->so_m = NULL;
ti = so->so_ti;
tiwin = ti->ti_win;
tiflags = ti->ti_flags;
goto cont_conn;
}
slirp = m->slirp;
/*
* Get IP and TCP header together in first mbuf.
* Note: IP leaves IP header in first mbuf.
*/
ti = mtod(m, struct tcpiphdr *);
if (iphlen > sizeof(struct ip )) {
ip_stripoptions(m, (struct mbuf *)0);
iphlen=sizeof(struct ip );
}
/* 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.
*/
ip=mtod(m, struct ip *);
save_ip = *ip;
save_ip.ip_len+= iphlen;
/*
* Checksum extended TCP header and data.
*/
tlen = ((struct ip *)ti)->ip_len;
tcpiphdr2qlink(ti)->next = tcpiphdr2qlink(ti)->prev = NULL;
memset(&ti->ti_i.ih_mbuf, 0 , sizeof(struct mbuf_ptr));
ti->ti_x1 = 0;
ti->ti_len = htons((uint16_t)tlen);
len = sizeof(struct ip ) + tlen;
if(cksum(m, len)) {
goto drop;
}
/*
* Check that TCP offset makes sense,
* pull out TCP options and adjust length. XXX
*/
off = ti->ti_off << 2;
if (off < sizeof (struct tcphdr) || off > tlen) {
goto drop;
}
tlen -= off;
ti->ti_len = tlen;
if (off > sizeof (struct tcphdr)) {
optlen = off - sizeof (struct tcphdr);
optp = mtod(m, caddr_t) + sizeof (struct tcpiphdr);
}
tiflags = ti->ti_flags;
/*
* Convert TCP protocol specific fields to host format.
*/
NTOHL(ti->ti_seq);
NTOHL(ti->ti_ack);
NTOHS(ti->ti_win);
NTOHS(ti->ti_urp);
/*
* Drop TCP, IP headers and TCP options.
*/
m->m_data += sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr);
m->m_len -= sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr);
/*
* Locate pcb for segment.
*/
findso:
so = slirp->tcp_last_so;
if (so->so_fport != ti->ti_dport ||
so->so_lport != ti->ti_sport ||
so->so_laddr.s_addr != ti->ti_src.s_addr ||
so->so_faddr.s_addr != ti->ti_dst.s_addr) {
so = solookup(&slirp->tcb, ti->ti_src, ti->ti_sport,
ti->ti_dst, ti->ti_dport);
if (so)
slirp->tcp_last_so = so;
}
/*
* 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...
*/
if (so == NULL) {
if (slirp->restricted) {
/* Any hostfwds will have an existing socket, so we only get here
* for non-hostfwd connections. These should be dropped, unless it
* happens to be a guestfwd.
*/
for (ex_ptr = slirp->exec_list; ex_ptr; ex_ptr = ex_ptr->ex_next) {
if (ex_ptr->ex_fport == ti->ti_dport &&
ti->ti_dst.s_addr == ex_ptr->ex_addr.s_addr) {
break;
}
}
if (!ex_ptr) {
goto dropwithreset;
}
}
if ((tiflags & (TH_SYN|TH_FIN|TH_RST|TH_URG|TH_ACK)) != TH_SYN)
goto dropwithreset;
if ((so = socreate(slirp)) == NULL)
goto dropwithreset;
if (tcp_attach(so) < 0) {
free(so); /* Not sofree (if it failed, it's not insqued) */
goto dropwithreset;
}
sbreserve(&so->so_snd, TCP_SNDSPACE);
sbreserve(&so->so_rcv, TCP_RCVSPACE);
so->so_laddr = ti->ti_src;
so->so_lport = ti->ti_sport;
so->so_faddr = ti->ti_dst;
so->so_fport = ti->ti_dport;
if ((so->so_iptos = tcp_tos(so)) == 0)
so->so_iptos = ((struct ip *)ti)->ip_tos;
tp = sototcpcb(so);
tp->t_state = TCPS_LISTEN;
}
/*
* 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.
*/
if (so->so_state & SS_ISFCONNECTING)
goto drop;
tp = sototcpcb(so);
/* XXX Should never fail */
if (tp == NULL)
goto dropwithreset;
if (tp->t_state == TCPS_CLOSED)
goto drop;
tiwin = ti->ti_win;
/*
* Segment received on connection.
* Reset idle time and keep-alive timer.
*/
tp->t_idle = 0;
if (SO_OPTIONS)
tp->t_timer[TCPT_KEEP] = TCPTV_KEEPINTVL;
else
tp->t_timer[TCPT_KEEP] = TCPTV_KEEP_IDLE;
/*
* Process options if not in LISTEN state,
* else do it below (after getting remote address).
*/
if (optp && tp->t_state != TCPS_LISTEN)
tcp_dooptions(tp, (u_char *)optp, optlen, ti);
/*
* 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..
*/
if (tp->t_state == TCPS_ESTABLISHED &&
(tiflags & (TH_SYN|TH_FIN|TH_RST|TH_URG|TH_ACK)) == TH_ACK &&
ti->ti_seq == tp->rcv_nxt &&
tiwin && tiwin == tp->snd_wnd &&
tp->snd_nxt == tp->snd_max) {
if (ti->ti_len == 0) {
if (SEQ_GT(ti->ti_ack, tp->snd_una) &&
SEQ_LEQ(ti->ti_ack, tp->snd_max) &&
tp->snd_cwnd >= tp->snd_wnd) {
/*
* this is a pure ack for outstanding data.
*/
if (tp->t_rtt &&
SEQ_GT(ti->ti_ack, tp->t_rtseq))
tcp_xmit_timer(tp, tp->t_rtt);
acked = ti->ti_ack - tp->snd_una;
sbdrop(&so->so_snd, acked);
tp->snd_una = ti->ti_ack;
m_free(m);
/*
* If all outstanding data are acked, stop
* retransmit timer, otherwise restart timer
* using current (possibly backed-off) value.
* If process is waiting for space,
* wakeup/selwakeup/signal. If data
* are ready to send, let tcp_output
* decide between more output or persist.
*/
if (tp->snd_una == tp->snd_max)
tp->t_timer[TCPT_REXMT] = 0;
else if (tp->t_timer[TCPT_PERSIST] == 0)
tp->t_timer[TCPT_REXMT] = tp->t_rxtcur;
/*
* This is called because sowwakeup might have
* put data into so_snd. Since we don't so sowwakeup,
* we don't need this.. XXX???
*/
if (so->so_snd.sb_cc)
(void) tcp_output(tp);
return;
}
} else if (ti->ti_ack == tp->snd_una &&
tcpfrag_list_empty(tp) &&
ti->ti_len <= sbspace(&so->so_rcv)) {
/*
* this is a pure, in-sequence data packet
* with nothing on the reassembly queue and
* we have enough buffer space to take it.
*/
tp->rcv_nxt += ti->ti_len;
/*
* Add data to socket buffer.
*/
if (so->so_emu) {
if (tcp_emu(so,m)) sbappend(so, m);
} else
sbappend(so, m);
/*
* If this is a short packet, then ACK now - with Nagel
* congestion avoidance sender won't send more until
* he gets an ACK.
*
* It is better to not delay acks at all to maximize
* TCP throughput. See RFC 2581.
*/
tp->t_flags |= TF_ACKNOW;
tcp_output(tp);
return;
}
} /* 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.
*/
{ int win;
win = sbspace(&so->so_rcv);
if (win < 0)
win = 0;
tp->rcv_wnd = max(win, (int)(tp->rcv_adv - tp->rcv_nxt));
}
switch (tp->t_state) {
/*
* 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
* segment in this state.
*/
case TCPS_LISTEN: {
if (tiflags & TH_RST)
goto drop;
if (tiflags & TH_ACK)
goto dropwithreset;
if ((tiflags & TH_SYN) == 0)
goto drop;
/*
* This has way too many gotos...
* But a bit of spaghetti code never hurt anybody :)
*/
/*
* If this is destined for the control address, then flag to
* tcp_ctl once connected, otherwise connect
*/
if ((so->so_faddr.s_addr & slirp->vnetwork_mask.s_addr) ==
slirp->vnetwork_addr.s_addr) {
if (so->so_faddr.s_addr != slirp->vhost_addr.s_addr &&
so->so_faddr.s_addr != slirp->vnameserver_addr.s_addr) {
/* May be an add exec */
for (ex_ptr = slirp->exec_list; ex_ptr;
ex_ptr = ex_ptr->ex_next) {
if(ex_ptr->ex_fport == so->so_fport &&
so->so_faddr.s_addr == ex_ptr->ex_addr.s_addr) {
so->so_state |= SS_CTL;
break;
}
}
if (so->so_state & SS_CTL) {
goto cont_input;
}
}
/* CTL_ALIAS: Do nothing, tcp_fconnect will be called on it */
}
if (so->so_emu & EMU_NOCONNECT) {
so->so_emu &= ~EMU_NOCONNECT;
goto cont_input;
}
if ((tcp_fconnect(so) == -1) &&
#if defined(_WIN32)
socket_error() != WSAEWOULDBLOCK
#else
(errno != EINPROGRESS) && (errno != EWOULDBLOCK)
#endif
) {
u_char code=ICMP_UNREACH_NET;
DEBUG_MISC((dfd, " tcp fconnect errno = %d-%s\n",
errno,strerror(errno)));
if(errno == ECONNREFUSED) {
/* ACK the SYN, send RST to refuse the connection */
tcp_respond(tp, ti, m, ti->ti_seq+1, (tcp_seq)0,
TH_RST|TH_ACK);
} else {
if(errno == EHOSTUNREACH) code=ICMP_UNREACH_HOST;
HTONL(ti->ti_seq); /* restore tcp header */
HTONL(ti->ti_ack);
HTONS(ti->ti_win);
HTONS(ti->ti_urp);
m->m_data -= sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr);
m->m_len += sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr);
*ip=save_ip;
icmp_error(m, ICMP_UNREACH,code, 0,strerror(errno));
}
tcp_close(tp);
m_free(m);
} else {
/*
* Haven't connected yet, save the current mbuf
* and ti, and return
* XXX Some OS's don't tell us whether the connect()
* succeeded or not. So we must time it out.
*/
so->so_m = m;
so->so_ti = ti;
tp->t_timer[TCPT_KEEP] = TCPTV_KEEP_INIT;
tp->t_state = TCPS_SYN_RECEIVED;
tcp_template(tp);
}
return;
cont_conn:
/* m==NULL
* Check if the connect succeeded
*/
if (so->so_state & SS_NOFDREF) {
tp = tcp_close(tp);
goto dropwithreset;
}
cont_input:
tcp_template(tp);
if (optp)
tcp_dooptions(tp, (u_char *)optp, optlen, ti);
if (iss)
tp->iss = iss;
else
tp->iss = slirp->tcp_iss;
slirp->tcp_iss += TCP_ISSINCR/2;
tp->irs = ti->ti_seq;
tcp_sendseqinit(tp);
tcp_rcvseqinit(tp);
tp->t_flags |= TF_ACKNOW;
tp->t_state = TCPS_SYN_RECEIVED;
tp->t_timer[TCPT_KEEP] = TCPTV_KEEP_INIT;
goto trimthenstep6;
} /* case TCPS_LISTEN */
/*
* 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
*/
case TCPS_SYN_SENT:
if ((tiflags & TH_ACK) &&
(SEQ_LEQ(ti->ti_ack, tp->iss) ||
SEQ_GT(ti->ti_ack, tp->snd_max)))
goto dropwithreset;
if (tiflags & TH_RST) {
if (tiflags & TH_ACK) {
tcp_drop(tp, 0); /* XXX Check t_softerror! */
}
goto drop;
}
if ((tiflags & TH_SYN) == 0)
goto drop;
if (tiflags & TH_ACK) {
tp->snd_una = ti->ti_ack;
if (SEQ_LT(tp->snd_nxt, tp->snd_una))
tp->snd_nxt = tp->snd_una;
}
tp->t_timer[TCPT_REXMT] = 0;
tp->irs = ti->ti_seq;
tcp_rcvseqinit(tp);
tp->t_flags |= TF_ACKNOW;
if (tiflags & TH_ACK && SEQ_GT(tp->snd_una, tp->iss)) {
soisfconnected(so);
tp->t_state = TCPS_ESTABLISHED;
(void) tcp_reass(tp, (struct tcpiphdr *)0,
(struct mbuf *)0);
/*
* if we didn't have to retransmit the SYN,
* use its rtt as our initial srtt & rtt var.
*/
if (tp->t_rtt)
tcp_xmit_timer(tp, tp->t_rtt);
} else
tp->t_state = TCPS_SYN_RECEIVED;
trimthenstep6:
/*
* Advance ti->ti_seq to correspond to first data byte.
* If data, trim to stay within window,
* dropping FIN if necessary.
*/
ti->ti_seq++;
if (ti->ti_len > tp->rcv_wnd) {
todrop = ti->ti_len - tp->rcv_wnd;
m_adj(m, -todrop);
ti->ti_len = tp->rcv_wnd;
tiflags &= ~TH_FIN;
}
tp->snd_wl1 = ti->ti_seq - 1;
tp->rcv_up = ti->ti_seq;
goto step6;
} /* switch tp->t_state */
/*
* States other than LISTEN or SYN_SENT.
* 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.
*/
todrop = tp->rcv_nxt - ti->ti_seq;
if (todrop > 0) {
if (tiflags & TH_SYN) {
tiflags &= ~TH_SYN;
ti->ti_seq++;
if (ti->ti_urp > 1)
ti->ti_urp--;
else
tiflags &= ~TH_URG;
todrop--;
}
/*
* Following if statement from Stevens, vol. 2, p. 960.
*/
if (todrop > ti->ti_len
|| (todrop == ti->ti_len && (tiflags & TH_FIN) == 0)) {
/*
* Any valid FIN must be to the left of the window.
* At this point the FIN must be a duplicate or out
* of sequence; drop it.
*/
tiflags &= ~TH_FIN;
/*
* Send an ACK to resynchronize and drop any data.
* But keep on processing for RST or ACK.
*/
tp->t_flags |= TF_ACKNOW;
todrop = ti->ti_len;
}
m_adj(m, todrop);
ti->ti_seq += todrop;
ti->ti_len -= todrop;
if (ti->ti_urp > todrop)
ti->ti_urp -= todrop;
else {
tiflags &= ~TH_URG;
ti->ti_urp = 0;
}
}
/*
* If new data are received on a connection after the
* user processes are gone, then RST the other end.
*/
if ((so->so_state & SS_NOFDREF) &&
tp->t_state > TCPS_CLOSE_WAIT && ti->ti_len) {
tp = tcp_close(tp);
goto dropwithreset;
}
/*
* If segment ends after window, drop trailing data
* (and PUSH and FIN); if nothing left, just ACK.
*/
todrop = (ti->ti_seq+ti->ti_len) - (tp->rcv_nxt+tp->rcv_wnd);
if (todrop > 0) {
if (todrop >= ti->ti_len) {
/*
* 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 (tiflags & TH_SYN &&
tp->t_state == TCPS_TIME_WAIT &&
SEQ_GT(ti->ti_seq, tp->rcv_nxt)) {
iss = tp->rcv_nxt + TCP_ISSINCR;
tp = tcp_close(tp);
goto findso;
}
/*
* 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
* and ack.
*/
if (tp->rcv_wnd == 0 && ti->ti_seq == tp->rcv_nxt) {
tp->t_flags |= TF_ACKNOW;
} else {
goto dropafterack;
}
}
m_adj(m, -todrop);
ti->ti_len -= todrop;
tiflags &= ~(TH_PUSH|TH_FIN);
}
/*
* If the RST bit is set examine the state:
* SYN_RECEIVED 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
* Close the tcb.
*/
if (tiflags&TH_RST) switch (tp->t_state) {
case TCPS_SYN_RECEIVED:
case TCPS_ESTABLISHED:
case TCPS_FIN_WAIT_1:
case TCPS_FIN_WAIT_2:
case TCPS_CLOSE_WAIT:
tp->t_state = TCPS_CLOSED;
tcp_close(tp);
goto drop;
case TCPS_CLOSING:
case TCPS_LAST_ACK:
case TCPS_TIME_WAIT:
tcp_close(tp);
goto drop;
}
/*
* If a SYN is in the window, then this is an
* error and we send an RST and drop the connection.
*/
if (tiflags & TH_SYN) {
tp = tcp_drop(tp,0);
goto dropwithreset;
}
/*
* If the ACK bit is off we drop the segment and return.
*/
if ((tiflags & TH_ACK) == 0) goto drop;
/*
* Ack processing.
*/
switch (tp->t_state) {
/*
* In SYN_RECEIVED state if the ack ACKs our SYN then enter
* ESTABLISHED state and continue processing, otherwise
* send an RST. una<=ack<=max
*/
case TCPS_SYN_RECEIVED:
if (SEQ_GT(tp->snd_una, ti->ti_ack) ||
SEQ_GT(ti->ti_ack, tp->snd_max))
goto dropwithreset;
tp->t_state = TCPS_ESTABLISHED;
/*
* 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.
* tp->snd_una++; or:
*/
tp->snd_una=ti->ti_ack;
if (so->so_state & SS_CTL) {
/* So tcp_ctl reports the right state */
ret = tcp_ctl(so);
if (ret == 1) {
soisfconnected(so);
so->so_state &= ~SS_CTL; /* success XXX */
} else if (ret == 2) {
so->so_state &= SS_PERSISTENT_MASK;
so->so_state |= SS_NOFDREF; /* CTL_CMD */
} else {
needoutput = 1;
tp->t_state = TCPS_FIN_WAIT_1;
}
} else {
soisfconnected(so);
}
(void) tcp_reass(tp, (struct tcpiphdr *)0, (struct mbuf *)0);
tp->snd_wl1 = ti->ti_seq - 1;
/* Avoid ack processing; snd_una==ti_ack => dup ack */
goto synrx_to_est;
/* fall into ... */
/*
* 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.
*/
case TCPS_ESTABLISHED:
case TCPS_FIN_WAIT_1:
case TCPS_FIN_WAIT_2:
case TCPS_CLOSE_WAIT:
case TCPS_CLOSING:
case TCPS_LAST_ACK:
case TCPS_TIME_WAIT:
if (SEQ_LEQ(ti->ti_ack, tp->snd_una)) {
if (ti->ti_len == 0 && tiwin == tp->snd_wnd) {
DEBUG_MISC((dfd, " dup ack m = %lx so = %lx\n",
(long )m, (long )so));
/*
* 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
* packet.
*
* 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
* the new ssthresh).
*
* 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
* network.
*/
if (tp->t_timer[TCPT_REXMT] == 0 ||
ti->ti_ack != tp->snd_una)
tp->t_dupacks = 0;
else if (++tp->t_dupacks == TCPREXMTTHRESH) {
tcp_seq onxt = tp->snd_nxt;
u_int win =
min(tp->snd_wnd, tp->snd_cwnd) / 2 /
tp->t_maxseg;
if (win < 2)
win = 2;
tp->snd_ssthresh = win * tp->t_maxseg;
tp->t_timer[TCPT_REXMT] = 0;
tp->t_rtt = 0;
tp->snd_nxt = ti->ti_ack;
tp->snd_cwnd = tp->t_maxseg;
(void) tcp_output(tp);
tp->snd_cwnd = tp->snd_ssthresh +
tp->t_maxseg * tp->t_dupacks;
if (SEQ_GT(onxt, tp->snd_nxt))
tp->snd_nxt = onxt;
goto drop;
} else if (tp->t_dupacks > TCPREXMTTHRESH) {
tp->snd_cwnd += tp->t_maxseg;
(void) tcp_output(tp);
goto drop;
}
} else
tp->t_dupacks = 0;
break;
}
synrx_to_est:
/*
* If the congestion window was inflated to account
* for the other side's cached packets, retract it.
*/
if (tp->t_dupacks > TCPREXMTTHRESH &&
tp->snd_cwnd > tp->snd_ssthresh)
tp->snd_cwnd = tp->snd_ssthresh;
tp->t_dupacks = 0;
if (SEQ_GT(ti->ti_ack, tp->snd_max)) {
goto dropafterack;
}
acked = ti->ti_ack - tp->snd_una;
/*
* If 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 (tp->t_rtt && SEQ_GT(ti->ti_ack, tp->t_rtseq))
tcp_xmit_timer(tp,tp->t_rtt);
/*
* 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.
*/
if (ti->ti_ack == tp->snd_max) {
tp->t_timer[TCPT_REXMT] = 0;
needoutput = 1;
} else if (tp->t_timer[TCPT_PERSIST] == 0)
tp->t_timer[TCPT_REXMT] = tp->t_rxtcur;
/*
* 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).
*/
{
register u_int cw = tp->snd_cwnd;
register u_int incr = tp->t_maxseg;
if (cw > tp->snd_ssthresh)
incr = incr * incr / cw;
tp->snd_cwnd = min(cw + incr, TCP_MAXWIN<<tp->snd_scale);
}
if (acked > so->so_snd.sb_cc) {
tp->snd_wnd -= so->so_snd.sb_cc;
sbdrop(&so->so_snd, (int )so->so_snd.sb_cc);
ourfinisacked = 1;
} else {
sbdrop(&so->so_snd, acked);
tp->snd_wnd -= acked;
ourfinisacked = 0;
}
tp->snd_una = ti->ti_ack;
if (SEQ_LT(tp->snd_nxt, tp->snd_una))
tp->snd_nxt = tp->snd_una;
switch (tp->t_state) {
/*
* In FIN_WAIT_1 STATE in addition to the processing
* for the ESTABLISHED state if our FIN is now acknowledged
* then enter FIN_WAIT_2.
*/
case TCPS_FIN_WAIT_1:
if (ourfinisacked) {
/*
* 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
* we'll hang forever.
*/
if (so->so_state & SS_FCANTRCVMORE) {
tp->t_timer[TCPT_2MSL] = TCP_MAXIDLE;
}
tp->t_state = TCPS_FIN_WAIT_2;
}
break;
/*
* 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
* the segment.
*/
case TCPS_CLOSING:
if (ourfinisacked) {
tp->t_state = TCPS_TIME_WAIT;
tcp_canceltimers(tp);
tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL;
}
break;
/*
* 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.
*/
case TCPS_LAST_ACK:
if (ourfinisacked) {
tcp_close(tp);
goto drop;
}
break;
/*
* In TIME_WAIT state the only thing that should arrive
* is a retransmission of the remote FIN. Acknowledge
* it and restart the finack timer.
*/
case TCPS_TIME_WAIT:
tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL;
goto dropafterack;
}
} /* switch(tp->t_state) */
step6:
/*
* Update window information.
* Don't look at window if no ACK: TAC's send garbage on first SYN.
*/
if ((tiflags & TH_ACK) &&
(SEQ_LT(tp->snd_wl1, ti->ti_seq) ||
(tp->snd_wl1 == ti->ti_seq && (SEQ_LT(tp->snd_wl2, ti->ti_ack) ||
(tp->snd_wl2 == ti->ti_ack && tiwin > tp->snd_wnd))))) {
tp->snd_wnd = tiwin;
tp->snd_wl1 = ti->ti_seq;
tp->snd_wl2 = ti->ti_ack;
if (tp->snd_wnd > tp->max_sndwnd)
tp->max_sndwnd = tp->snd_wnd;
needoutput = 1;
}
/*
* Process segments with URG.
*/
if ((tiflags & TH_URG) && ti->ti_urp &&
TCPS_HAVERCVDFIN(tp->t_state) == 0) {
/*
* 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 (ti->ti_urp + so->so_rcv.sb_cc > so->so_rcv.sb_datalen) {
ti->ti_urp = 0;
tiflags &= ~TH_URG;
goto dodata;
}
/*
* 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 (SEQ_GT(ti->ti_seq+ti->ti_urp, tp->rcv_up)) {
tp->rcv_up = ti->ti_seq + ti->ti_urp;
so->so_urgc = so->so_rcv.sb_cc +
(tp->rcv_up - tp->rcv_nxt); /* -1; */
tp->rcv_up = ti->ti_seq + ti->ti_urp;
}
} else
/*
* If no out of band data is expected,
* pull receive urgent pointer along
* with the receive window.
*/
if (SEQ_GT(tp->rcv_nxt, tp->rcv_up))
tp->rcv_up = tp->rcv_nxt;
dodata:
/*
* If this is a small packet, then ACK now - with Nagel
* congestion avoidance sender won't send more until
* he gets an ACK.
*/
if (ti->ti_len && (unsigned)ti->ti_len <= 5 &&
((struct tcpiphdr_2 *)ti)->first_char == (char)27) {
tp->t_flags |= TF_ACKNOW;
}
/*
* 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
* is presented to the user (this happens in tcp_usrreq.c,
* case PRU_RCVD). If a FIN has already been received on this
* connection then we just ignore the text.
*/
if ((ti->ti_len || (tiflags&TH_FIN)) &&
TCPS_HAVERCVDFIN(tp->t_state) == 0) {
TCP_REASS(tp, ti, m, so, tiflags);
} else {
m_free(m);
tiflags &= ~TH_FIN;
}
/*
* If FIN is received ACK the FIN and let the user know
* that the connection is closing.
*/
if (tiflags & TH_FIN) {
if (TCPS_HAVERCVDFIN(tp->t_state) == 0) {
/*
* If we receive a FIN we can't send more data,
* set it SS_FDRAIN
* Shutdown the socket if there is no rx data in the
* buffer.
* soread() is called on completion of shutdown() and
* will got to TCPS_LAST_ACK, and use tcp_output()
* to send the FIN.
*/
sofwdrain(so);
tp->t_flags |= TF_ACKNOW;
tp->rcv_nxt++;
}
switch (tp->t_state) {
/*
* In SYN_RECEIVED and ESTABLISHED STATES
* enter the CLOSE_WAIT state.
*/
case TCPS_SYN_RECEIVED:
case TCPS_ESTABLISHED:
if(so->so_emu == EMU_CTL) /* no shutdown on socket */
tp->t_state = TCPS_LAST_ACK;
else
tp->t_state = TCPS_CLOSE_WAIT;
break;
/*
* If still in FIN_WAIT_1 STATE FIN has not been acked so
* enter the CLOSING state.
*/
case TCPS_FIN_WAIT_1:
tp->t_state = TCPS_CLOSING;
break;
/*
* In FIN_WAIT_2 state enter the TIME_WAIT state,
* starting the time-wait timer, turning off the other
* standard timers.
*/
case TCPS_FIN_WAIT_2:
tp->t_state = TCPS_TIME_WAIT;
tcp_canceltimers(tp);
tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL;
break;
/*
* In TIME_WAIT state restart the 2 MSL time_wait timer.
*/
case TCPS_TIME_WAIT:
tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL;
break;
}
}
/*
* Return any desired output.
*/
if (needoutput || (tp->t_flags & TF_ACKNOW)) {
(void) tcp_output(tp);
}
return;
dropafterack:
/*
* Generate an ACK dropping incoming segment if it occupies
* sequence space, where the ACK reflects our state.
*/
if (tiflags & TH_RST)
goto drop;
m_free(m);
tp->t_flags |= TF_ACKNOW;
(void) tcp_output(tp);
return;
dropwithreset:
/* reuses m if m!=NULL, m_free() unnecessary */
if (tiflags & TH_ACK)
tcp_respond(tp, ti, m, (tcp_seq)0, ti->ti_ack, TH_RST);
else {
if (tiflags & TH_SYN) ti->ti_len++;
tcp_respond(tp, ti, m, ti->ti_seq+ti->ti_len, (tcp_seq)0,
TH_RST|TH_ACK);
}
return;
drop:
/*
* Drop space held by incoming segment and return.
*/
m_free(m);
}
void tcp_cleanup(Slirp *slirp)
{
while (slirp->tcb.so_next != &slirp->tcb) {
tcp_close(sototcpcb(slirp->tcb.so_next));
}
}
/*
* Write data from so_rcv to so's socket,
* updating all sbuf field as necessary
*/
int
sowrite(struct socket *so)
{
int n,nn;
struct sbuf *sb = &so->so_rcv;
int len = sb->sb_cc;
struct iovec iov[2];
DEBUG_CALL("sowrite");
DEBUG_ARG("so = %lx", (long)so);
if (so->so_urgc) {
sosendoob(so);
if (sb->sb_cc == 0)
return 0;
}
/*
* No need to check if there's something to write,
* sowrite wouldn't have been called otherwise
*/
iov[0].iov_base = sb->sb_rptr;
iov[1].iov_base = NULL;
iov[1].iov_len = 0;
if (sb->sb_rptr < sb->sb_wptr) {
iov[0].iov_len = sb->sb_wptr - sb->sb_rptr;
/* Should never succeed, but... */
if (iov[0].iov_len > len) iov[0].iov_len = len;
n = 1;
} else {
iov[0].iov_len = (sb->sb_data + sb->sb_datalen) - sb->sb_rptr;
if (iov[0].iov_len > len) iov[0].iov_len = len;
len -= iov[0].iov_len;
if (len) {
iov[1].iov_base = sb->sb_data;
iov[1].iov_len = sb->sb_wptr - sb->sb_data;
if (iov[1].iov_len > len) iov[1].iov_len = len;
n = 2;
} else
n = 1;
}
/* Check if there's urgent data to send, and if so, send it */
#ifdef HAVE_READV
nn = writev(so->s, (const struct iovec *)iov, n);
DEBUG_MISC((dfd, " ... wrote nn = %d bytes\n", nn));
#else
nn = slirp_send(so, iov[0].iov_base, iov[0].iov_len,0);
#endif
/* This should never happen, but people tell me it does *shrug* */
if (nn < 0 && (errno == EAGAIN || errno == EINTR))
return 0;
if (nn <= 0) {
DEBUG_MISC((dfd, " --- sowrite disconnected, so->so_state = %x, errno = %d\n",
so->so_state, errno));
sofcantsendmore(so);
tcp_sockclosed(sototcpcb(so));
return -1;
}
#ifndef HAVE_READV
if (n == 2 && nn == iov[0].iov_len) {
int ret;
ret = slirp_send(so, iov[1].iov_base, iov[1].iov_len,0);
if (ret > 0)
nn += ret;
}
DEBUG_MISC((dfd, " ... wrote nn = %d bytes\n", nn));
#endif
/* Update sbuf */
sb->sb_cc -= nn;
sb->sb_rptr += nn;
if (sb->sb_rptr >= (sb->sb_data + sb->sb_datalen))
sb->sb_rptr -= sb->sb_datalen;
/*
* If in DRAIN mode, and there's no more data, set
* it CANTSENDMORE
*/
if ((so->so_state & SS_FWDRAIN) && sb->sb_cc == 0)
sofcantsendmore(so);
return nn;
}
/*
* Read from so's socket into sb_snd, updating all relevant sbuf fields
* NOTE: This will only be called if it is select()ed for reading, so
* a read() of 0 (or less) means it's disconnected
*/
int
soread(struct socket *so)
{
int n, nn;
struct sbuf *sb = &so->so_snd;
struct iovec iov[2];
DEBUG_CALL("soread");
DEBUG_ARG("so = %p", so);
/*
* No need to check if there's enough room to read.
* soread wouldn't have been called if there weren't
*/
sopreprbuf(so, iov, &n);
#ifdef HAVE_READV
nn = readv(so->s, (struct iovec *)iov, n);
DEBUG_MISC((dfd, " ... read nn = %d bytes\n", nn));
#else
nn = qemu_recv(so->s, iov[0].iov_base, iov[0].iov_len,0);
#endif
if (nn <= 0) {
if (nn < 0 && (errno == EINTR || errno == EAGAIN))
return 0;
else {
int err;
socklen_t slen = sizeof err;
err = errno;
if (nn == 0) {
getsockopt(so->s, SOL_SOCKET, SO_ERROR,
&err, &slen);
}
DEBUG_MISC((dfd, " --- soread() disconnected, nn = %d, errno = %d-%s\n", nn, errno,strerror(errno)));
sofcantrcvmore(so);
if (err == ECONNRESET || err == ECONNREFUSED
|| err == ENOTCONN || err == EPIPE) {
tcp_drop(sototcpcb(so), err);
} else {
tcp_sockclosed(sototcpcb(so));
}
return -1;
}
}
#ifndef HAVE_READV
/*
* If there was no error, try and read the second time round
* We read again if n = 2 (ie, there's another part of the buffer)
* and we read as much as we could in the first read
* We don't test for <= 0 this time, because there legitimately
* might not be any more data (since the socket is non-blocking),
* a close will be detected on next iteration.
* A return of -1 won't (shouldn't) happen, since it didn't happen above
*/
if (n == 2 && nn == iov[0].iov_len) {
int ret;
ret = qemu_recv(so->s, iov[1].iov_base, iov[1].iov_len,0);
if (ret > 0)
nn += ret;
}
DEBUG_MISC((dfd, " ... read nn = %d bytes\n", nn));
#endif
/* Update fields */
sb->sb_cc += nn;
sb->sb_wptr += nn;
if (sb->sb_wptr >= (sb->sb_data + sb->sb_datalen))
sb->sb_wptr -= sb->sb_datalen;
return nn;
}