]> git.saurik.com Git - apple/xnu.git/blobdiff - bsd/netinet/tcp_usrreq.c
xnu-3789.1.32.tar.gz
[apple/xnu.git] / bsd / netinet / tcp_usrreq.c
index d194a867feb498ce3a8b1d2ccb54ace82b73b556..67306d65d628e7ab1af45edd27d4ae2851c616f6 100644 (file)
@@ -1,8 +1,8 @@
 /*
- * Copyright (c) 2000-2007 Apple Inc. All rights reserved.
+ * Copyright (c) 2000-2016 Apple Inc. All rights reserved.
  *
  * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
- * 
+ *
  * This file contains Original Code and/or Modifications of Original Code
  * as defined in and that are subject to the Apple Public Source License
  * Version 2.0 (the 'License'). You may not use this file except in
  * unlawful or unlicensed copies of an Apple operating system, or to
  * circumvent, violate, or enable the circumvention or violation of, any
  * terms of an Apple operating system software license agreement.
- * 
+ *
  * Please obtain a copy of the License at
  * http://www.opensource.apple.com/apsl/ and read it before using this file.
- * 
+ *
  * The Original Code and all software distributed under the License are
  * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
@@ -22,7 +22,7 @@
  * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
  * Please see the License for the specific language governing rights and
  * limitations under the License.
- * 
+ *
  * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
  */
 /*
 #if INET6
 #include <sys/domain.h>
 #endif /* INET6 */
+#include <sys/kasl.h>
 #include <sys/socket.h>
 #include <sys/socketvar.h>
 #include <sys/protosw.h>
+#include <sys/syslog.h>
 
 #include <net/if.h>
 #include <net/route.h>
+#include <net/ntstat.h>
+#include <net/content_filter.h>
 
 #include <netinet/in.h>
 #include <netinet/in_systm.h>
 #include <netinet/tcp_timer.h>
 #include <netinet/tcp_var.h>
 #include <netinet/tcpip.h>
+#include <mach/sdt.h>
 #if TCPDEBUG
 #include <netinet/tcp_debug.h>
 #endif
+#if MPTCP
+#include <netinet/mptcp_var.h>
+#endif /* MPTCP */
 
 #if IPSEC
 #include <netinet6/ipsec.h>
 #endif /*IPSEC*/
 
+#if FLOW_DIVERT
+#include <netinet/flow_divert.h>
+#endif /* FLOW_DIVERT */
+
+void   tcp_fill_info(struct tcpcb *, struct tcp_info *);
+errno_t tcp_fill_info_for_info_tuple(struct info_tuple *, struct tcp_info *);
+
+int tcp_sysctl_info(struct sysctl_oid *, void *, int , struct sysctl_req *);
+static void tcp_connection_fill_info(struct tcpcb *tp,
+    struct tcp_connection_info *tci);
+
 /*
  * TCP protocol interface to socket abstraction.
  */
@@ -114,35 +133,12 @@ static int        tcp_attach(struct socket *, struct proc *);
 static int     tcp_connect(struct tcpcb *, struct sockaddr *, struct proc *);
 #if INET6
 static int     tcp6_connect(struct tcpcb *, struct sockaddr *, struct proc *);
+static int     tcp6_usr_connect(struct socket *, struct sockaddr *,
+                   struct proc *);
 #endif /* INET6 */
-static struct tcpcb *
-               tcp_disconnect(struct tcpcb *);
-static struct tcpcb *
-               tcp_usrclosed(struct tcpcb *);
-
-__private_extern__ int tcp_win_scale = 3;
-SYSCTL_INT(_net_inet_tcp, OID_AUTO, win_scale_factor, CTLFLAG_RW,
-    &tcp_win_scale, 0, "Window scaling factor");
-
-static u_int32_t tcps_in_sw_cksum;
-SYSCTL_UINT(_net_inet_tcp, OID_AUTO, in_sw_cksum, CTLFLAG_RD,
-    &tcps_in_sw_cksum, 0,
-    "Number of received packets checksummed in software");
-
-static u_int64_t tcps_in_sw_cksum_bytes;
-SYSCTL_QUAD(_net_inet_tcp, OID_AUTO, in_sw_cksum_bytes, CTLFLAG_RD,
-    &tcps_in_sw_cksum_bytes,
-    "Amount of received data checksummed in software");
-
-static u_int32_t tcps_out_sw_cksum;
-SYSCTL_UINT(_net_inet_tcp, OID_AUTO, out_sw_cksum, CTLFLAG_RD,
-    &tcps_out_sw_cksum, 0,
-    "Number of transmitted packets checksummed in software");
-
-static u_int64_t tcps_out_sw_cksum_bytes;
-SYSCTL_QUAD(_net_inet_tcp, OID_AUTO, out_sw_cksum_bytes, CTLFLAG_RD,
-    &tcps_out_sw_cksum_bytes,
-    "Amount of transmitted data checksummed in software");
+static struct tcpcb *tcp_disconnect(struct tcpcb *);
+static struct tcpcb *tcp_usrclosed(struct tcpcb *);
+extern void tcp_sbrcv_trim(struct tcpcb *tp, struct sockbuf *sb);
 
 #if TCPDEBUG
 #define        TCPDEBUG0       int ostate = 0
@@ -155,13 +151,9 @@ SYSCTL_QUAD(_net_inet_tcp, OID_AUTO, out_sw_cksum_bytes, CTLFLAG_RD,
 #define        TCPDEBUG2(req)
 #endif
 
-#if CONFIG_USESOCKTHRESHOLD
-__private_extern__ unsigned int        tcp_sockthreshold = 64;
-#else
-__private_extern__ unsigned int        tcp_sockthreshold = 0;
-#endif
-SYSCTL_INT(_net_inet_tcp, OID_AUTO, sockthreshold, CTLFLAG_RW, 
-    &tcp_sockthreshold , 0, "TCP Socket size increased if less than threshold");
+SYSCTL_PROC(_net_inet_tcp, OID_AUTO, info,
+    CTLFLAG_RW | CTLFLAG_LOCKED | CTLFLAG_ANYBODY | CTLFLAG_KERN,
+    0 , 0, tcp_sysctl_info, "S", "TCP info per tuple");
 
 /*
  * TCP attaches to socket via pru_attach(), reserving space,
@@ -217,29 +209,43 @@ tcp_usr_detach(struct socket *so)
        if (inp == 0 || (inp->inp_state == INPCB_STATE_DEAD)) {
                return EINVAL;  /* XXX */
        }
-#if 1
-       lck_mtx_assert(((struct inpcb *)so->so_pcb)->inpcb_mtx, LCK_MTX_ASSERT_OWNED);
-#endif
+       lck_mtx_assert(&((struct inpcb *)so->so_pcb)->inpcb_mtx, LCK_MTX_ASSERT_OWNED);
        tp = intotcpcb(inp);
        /* In case we got disconnected from the peer */
-        if (tp == 0) 
-           goto out;
+        if (tp == NULL)
+               goto out;
        TCPDEBUG1();
+
+       calculate_tcp_clock();
+
        tp = tcp_disconnect(tp);
 out:
        TCPDEBUG2(PRU_DETACH);
        return error;
 }
 
-#define        COMMON_START()  TCPDEBUG0; \
-                       do { \
-                                    if (inp == 0 || (inp->inp_state == INPCB_STATE_DEAD)) { \
-                                            return EINVAL; \
-                                    } \
-                                    tp = intotcpcb(inp); \
-                                    TCPDEBUG1(); \
-                    } while(0)
-                            
+#if NECP
+#define        COMMON_START()  TCPDEBUG0;                                      \
+do {                                                                   \
+       if (inp == NULL || inp->inp_state == INPCB_STATE_DEAD)          \
+               return (EINVAL);                                        \
+       if (necp_socket_should_use_flow_divert(inp))                    \
+               return (EPROTOTYPE);                                    \
+       tp = intotcpcb(inp);                                            \
+       TCPDEBUG1();                                                    \
+       calculate_tcp_clock();                                          \
+} while (0)
+#else /* NECP */
+#define        COMMON_START()  TCPDEBUG0;                                      \
+do {                                                                   \
+       if (inp == NULL || inp->inp_state == INPCB_STATE_DEAD)          \
+               return (EINVAL);                                        \
+       tp = intotcpcb(inp);                                            \
+       TCPDEBUG1();                                                    \
+       calculate_tcp_clock();                                          \
+} while (0)
+#endif /* !NECP */
+
 #define COMMON_END(req)        out: TCPDEBUG2(req); return error; goto out
 
 
@@ -276,7 +282,7 @@ tcp_usr_bind(struct socket *so, struct sockaddr *nam, struct proc *p)
         * Must check for multicast addresses and disallow binding
         * to them.
         */
-       sinp = (struct sockaddr_in *)nam;
+       sinp = (struct sockaddr_in *)(void *)nam;
        if (sinp->sin_family == AF_INET &&
            IN_MULTICAST(ntohl(sinp->sin_addr.s_addr))) {
                error = EAFNOSUPPORT;
@@ -309,7 +315,7 @@ tcp6_usr_bind(struct socket *so, struct sockaddr *nam, struct proc *p)
         * Must check for multicast addresses and disallow binding
         * to them.
         */
-       sin6p = (struct sockaddr_in6 *)nam;
+       sin6p = (struct sockaddr_in6 *)(void *)nam;
        if (sin6p->sin6_family == AF_INET6 &&
            IN6_IS_ADDR_MULTICAST(&sin6p->sin6_addr)) {
                error = EAFNOSUPPORT;
@@ -359,7 +365,7 @@ tcp_usr_listen(struct socket *so, struct proc *p)
 
        COMMON_START();
        if (inp->inp_lport == 0)
-               error = in_pcbbind(inp, (struct sockaddr *)0, p);
+               error = in_pcbbind(inp, NULL, p);
        if (error == 0)
                tp->t_state = TCPS_LISTEN;
        COMMON_END(PRU_LISTEN);
@@ -378,7 +384,7 @@ tcp6_usr_listen(struct socket *so, struct proc *p)
                inp->inp_vflag &= ~INP_IPV4;
                if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0)
                        inp->inp_vflag |= INP_IPV4;
-               error = in6_pcbbind(inp, (struct sockaddr *)0, p);
+               error = in6_pcbbind(inp, NULL, p);
        }
        if (error == 0)
                tp->t_state = TCPS_LISTEN;
@@ -386,6 +392,27 @@ tcp6_usr_listen(struct socket *so, struct proc *p)
 }
 #endif /* INET6 */
 
+static int
+tcp_connect_complete(struct socket *so)
+{
+       struct tcpcb *tp = sototcpcb(so);
+       int error = 0;
+
+       /* TFO delays the tcp_output until later, when the app calls write() */
+       if (so->so_flags1 & SOF1_PRECONNECT_DATA) {
+               if (!necp_socket_is_allowed_to_send_recv(sotoinpcb(so), NULL, NULL))
+                       return (EHOSTUNREACH);
+
+               /* Initialize enough state so that we can actually send data */
+               tcp_mss(tp, -1, IFSCOPE_NONE);
+               tp->snd_wnd = tp->t_maxseg;
+       } else {
+               error = tcp_output(tp);
+       }
+
+       return (error);
+}
+
 /*
  * Initiate connection to peer.
  * Create a template for use in transmissions on this connection.
@@ -402,9 +429,9 @@ tcp_usr_connect(struct socket *so, struct sockaddr *nam, struct proc *p)
        struct sockaddr_in *sinp;
 
        TCPDEBUG0;
-       if (inp == 0)
+       if (inp == NULL) {
                return EINVAL;
-       else if (inp->inp_state == INPCB_STATE_DEAD) {
+       else if (inp->inp_state == INPCB_STATE_DEAD) {
                if (so->so_error) {
                        error = so->so_error;
                        so->so_error = 0;
@@ -412,9 +439,33 @@ tcp_usr_connect(struct socket *so, struct sockaddr *nam, struct proc *p)
                } else
                        return EINVAL;
        }
+#if NECP
+#if FLOW_DIVERT
+       else if (necp_socket_should_use_flow_divert(inp)) {
+               uint32_t fd_ctl_unit = necp_socket_get_flow_divert_control_unit(inp);
+               if (fd_ctl_unit > 0) {
+                       error = flow_divert_pcb_init(so, fd_ctl_unit);
+                       if (error == 0) {
+                               error = flow_divert_connect_out(so, nam, p);
+                       }
+               } else {
+                       error = ENETDOWN;
+               }
+
+               return error;
+       }
+#endif /* FLOW_DIVERT */
+#if CONTENT_FILTER
+       error = cfil_sock_attach(so);
+       if (error != 0)
+               return error;
+#endif /* CONTENT_FILTER */
+#endif /* NECP */
        tp = intotcpcb(inp);
        TCPDEBUG1();
 
+       calculate_tcp_clock();
+
        if (nam->sa_family != 0 && nam->sa_family != AF_INET) {
                error = EAFNOSUPPORT;
                goto out;
@@ -422,20 +473,165 @@ tcp_usr_connect(struct socket *so, struct sockaddr *nam, struct proc *p)
        /*
         * Must disallow TCP ``connections'' to multicast addresses.
         */
-       sinp = (struct sockaddr_in *)nam;
+       sinp = (struct sockaddr_in *)(void *)nam;
        if (sinp->sin_family == AF_INET
            && IN_MULTICAST(ntohl(sinp->sin_addr.s_addr))) {
                error = EAFNOSUPPORT;
                goto out;
        }
 
-
        if ((error = tcp_connect(tp, nam, p)) != 0)
                goto out;
-       error = tcp_output(tp);
+
+       error = tcp_connect_complete(so);
+
        COMMON_END(PRU_CONNECT);
 }
 
+static int
+tcp_usr_connectx_common(struct socket *so, int af,
+    struct sockaddr_list **src_sl, struct sockaddr_list **dst_sl,
+    struct proc *p, uint32_t ifscope, sae_associd_t aid, sae_connid_t *pcid,
+    uint32_t flags, void *arg, uint32_t arglen, struct uio *auio,
+    user_ssize_t *bytes_written)
+{
+#pragma unused(aid)
+#if !MPTCP
+#pragma unused(flags, arg, arglen)
+#endif /* !MPTCP */
+       struct sockaddr_entry *src_se = NULL, *dst_se = NULL;
+       struct inpcb *inp = sotoinpcb(so);
+       int error;
+       user_ssize_t datalen = 0;
+
+       if (inp == NULL)
+               return (EINVAL);
+
+       VERIFY(dst_sl != NULL);
+
+       /* select source (if specified) and destination addresses */
+       error = in_selectaddrs(af, src_sl, &src_se, dst_sl, &dst_se);
+       if (error != 0)
+               return (error);
+
+       VERIFY(*dst_sl != NULL && dst_se != NULL);
+       VERIFY(src_se == NULL || *src_sl != NULL);
+       VERIFY(dst_se->se_addr->sa_family == af);
+       VERIFY(src_se == NULL || src_se->se_addr->sa_family == af);
+
+#if NECP
+       inp_update_necp_policy(inp, src_se ? src_se->se_addr : NULL, dst_se ? dst_se->se_addr : NULL, ifscope);
+#endif /* NECP */
+
+       if ((so->so_flags1 & SOF1_DATA_IDEMPOTENT) &&
+           (tcp_fastopen & TCP_FASTOPEN_CLIENT))
+               sototcpcb(so)->t_flagsext |= TF_FASTOPEN;
+
+       /*
+        * We get here for 2 cases:
+        *
+        *   a. From MPTCP, to connect a subflow.  There is no need to
+        *      bind the socket to the source address and/or interface,
+        *      since everything has been taken care of by MPTCP.  We
+        *      simply check whether or not this is for the initial
+        *      MPTCP connection attempt, or to join an existing one.
+        *
+        *   b. From the socket layer, to connect a TCP.  Perform the
+        *      bind to source address and/or interface as necessary.
+        */
+#if MPTCP
+       if (flags & CONNREQF_MPTCP) {
+               struct mptsub_connreq *mpcr = arg;
+
+               /* Check to make sure this came down from MPTCP */
+               if (arg == NULL || arglen != sizeof (*mpcr))
+                       return (EOPNOTSUPP);
+
+               switch (mpcr->mpcr_type) {
+               case MPTSUB_CONNREQ_MP_ENABLE:
+                       break;
+               case MPTSUB_CONNREQ_MP_ADD:
+                       break;
+               default:
+                       return (EOPNOTSUPP);
+               }
+       } else
+#endif /* MPTCP */
+       {
+               /* bind socket to the specified interface, if requested */
+               if (ifscope != IFSCOPE_NONE &&
+                   (error = inp_bindif(inp, ifscope, NULL)) != 0)
+                       return (error);
+
+               /* if source address and/or port is specified, bind to it */
+               if (src_se != NULL) {
+                       struct sockaddr *sa = src_se->se_addr;
+                       error = sobindlock(so, sa, 0);  /* already locked */
+                       if (error != 0)
+                               return (error);
+               }
+       }
+
+       switch (af) {
+       case AF_INET:
+               error = tcp_usr_connect(so, dst_se->se_addr, p);
+               break;
+#if INET6
+       case AF_INET6:
+               error = tcp6_usr_connect(so, dst_se->se_addr, p);
+               break;
+#endif /* INET6 */
+       default:
+               VERIFY(0);
+               /* NOTREACHED */
+       }
+
+       if (error != 0)
+               return (error);
+
+       /* if there is data, copy it */
+       if (auio != NULL) {
+               socket_unlock(so, 0);
+
+               VERIFY(bytes_written != NULL);
+
+               datalen = uio_resid(auio);
+               error = so->so_proto->pr_usrreqs->pru_sosend(so, NULL,
+                   (uio_t)auio, NULL, NULL, 0);
+               socket_lock(so, 0);
+
+               if (error == 0 || error == EWOULDBLOCK)
+                       *bytes_written = datalen - uio_resid(auio);
+
+               /*
+                * sosend returns EWOULDBLOCK if it's a non-blocking
+                * socket or a timeout occured (this allows to return
+                * the amount of queued data through sendit()).
+                *
+                * However, connectx() returns EINPROGRESS in case of a
+                * blocking socket. So we change the return value here.
+                */
+               if (error == EWOULDBLOCK)
+                       error = EINPROGRESS;
+       }
+
+       if (error == 0 && pcid != NULL)
+               *pcid = 1; /* there is only one connection in regular TCP */
+
+       return (error);
+}
+
+static int
+tcp_usr_connectx(struct socket *so, struct sockaddr_list **src_sl,
+    struct sockaddr_list **dst_sl, struct proc *p, uint32_t ifscope,
+    sae_associd_t aid, sae_connid_t *pcid, uint32_t flags, void *arg,
+    uint32_t arglen, struct uio *uio, user_ssize_t *bytes_written)
+{
+       return (tcp_usr_connectx_common(so, AF_INET, src_sl, dst_sl,
+           p, ifscope, aid, pcid, flags, arg, arglen, uio,
+           bytes_written));
+}
+
 #if INET6
 static int
 tcp6_usr_connect(struct socket *so, struct sockaddr *nam, struct proc *p)
@@ -445,7 +641,44 @@ tcp6_usr_connect(struct socket *so, struct sockaddr *nam, struct proc *p)
        struct tcpcb *tp;
        struct sockaddr_in6 *sin6p;
 
-       COMMON_START();
+       TCPDEBUG0;
+       if (inp == NULL) {
+               return EINVAL;
+       } else if (inp->inp_state == INPCB_STATE_DEAD) {
+               if (so->so_error) {
+                       error = so->so_error;
+                       so->so_error = 0;
+                       return error;
+               } else
+                       return EINVAL;
+       }
+#if NECP
+#if FLOW_DIVERT
+       else if (necp_socket_should_use_flow_divert(inp)) {
+               uint32_t fd_ctl_unit = necp_socket_get_flow_divert_control_unit(inp);
+               if (fd_ctl_unit > 0) {
+                       error = flow_divert_pcb_init(so, fd_ctl_unit);
+                       if (error == 0) {
+                               error = flow_divert_connect_out(so, nam, p);
+                       }
+               } else {
+                       error = ENETDOWN;
+               }
+
+               return error;
+       }
+#endif /* FLOW_DIVERT */
+#if CONTENT_FILTER
+       error = cfil_sock_attach(so);
+       if (error != 0)
+               return error;
+#endif /* CONTENT_FILTER */
+#endif /* NECP */
+
+       tp = intotcpcb(inp);
+       TCPDEBUG1();
+
+       calculate_tcp_clock();
 
        if (nam->sa_family != 0 && nam->sa_family != AF_INET6) {
                error = EAFNOSUPPORT;
@@ -455,7 +688,7 @@ tcp6_usr_connect(struct socket *so, struct sockaddr *nam, struct proc *p)
        /*
         * Must disallow TCP ``connections'' to multicast addresses.
         */
-       sin6p = (struct sockaddr_in6 *)nam;
+       sin6p = (struct sockaddr_in6 *)(void *)nam;
        if (sin6p->sin6_family == AF_INET6
            && IN6_IS_ADDR_MULTICAST(&sin6p->sin6_addr)) {
                error = EAFNOSUPPORT;
@@ -473,18 +706,29 @@ tcp6_usr_connect(struct socket *so, struct sockaddr *nam, struct proc *p)
                inp->inp_vflag &= ~INP_IPV6;
                if ((error = tcp_connect(tp, (struct sockaddr *)&sin, p)) != 0)
                        goto out;
-               error = tcp_output(tp);
+
+               error = tcp_connect_complete(so);
                goto out;
        }
        inp->inp_vflag &= ~INP_IPV4;
        inp->inp_vflag |= INP_IPV6;
        if ((error = tcp6_connect(tp, nam, p)) != 0)
                goto out;
-       error = tcp_output(tp);
-       if (error)
-               goto out;
+
+       error = tcp_connect_complete(so);
        COMMON_END(PRU_CONNECT);
 }
+
+static int
+tcp6_usr_connectx(struct socket *so, struct sockaddr_list **src_sl,
+    struct sockaddr_list **dst_sl, struct proc *p, uint32_t ifscope,
+    sae_associd_t aid, sae_connid_t *pcid, uint32_t flags, void *arg,
+    uint32_t arglen, struct uio *uio, user_ssize_t *bytes_written)
+{
+       return (tcp_usr_connectx_common(so, AF_INET6, src_sl, dst_sl,
+           p, ifscope, aid, pcid, flags, arg, arglen, uio,
+           bytes_written));
+}
 #endif /* INET6 */
 
 /*
@@ -504,18 +748,30 @@ tcp_usr_disconnect(struct socket *so)
        int error = 0;
        struct inpcb *inp = sotoinpcb(so);
        struct tcpcb *tp;
-       
-#if 1
-       lck_mtx_assert(((struct inpcb *)so->so_pcb)->inpcb_mtx, LCK_MTX_ASSERT_OWNED);
-#endif
+
+       lck_mtx_assert(&((struct inpcb *)so->so_pcb)->inpcb_mtx,
+           LCK_MTX_ASSERT_OWNED);
        COMMON_START();
         /* In case we got disconnected from the peer */
-        if (tp == 0)
-            goto out;
+        if (tp == NULL)
+               goto out;
        tp = tcp_disconnect(tp);
        COMMON_END(PRU_DISCONNECT);
 }
 
+/*
+ * User-protocol pru_disconnectx callback.
+ */
+static int
+tcp_usr_disconnectx(struct socket *so, sae_associd_t aid, sae_connid_t cid)
+{
+#pragma unused(cid)
+       if (aid != SAE_ASSOCID_ANY && aid != SAE_ASSOCID_ALL)
+               return (EINVAL);
+
+       return (tcp_usr_disconnect(so));
+}
+
 /*
  * Accept a connection.  Essentially all the work is
  * done at higher levels; just return the address
@@ -529,16 +785,29 @@ tcp_usr_accept(struct socket *so, struct sockaddr **nam)
        struct tcpcb *tp = NULL;
        TCPDEBUG0;
 
+       in_getpeeraddr(so, nam);
+
        if (so->so_state & SS_ISDISCONNECTED) {
                error = ECONNABORTED;
                goto out;
        }
-       if (inp == 0 || (inp->inp_state == INPCB_STATE_DEAD)) {
+       if (inp == NULL || inp->inp_state == INPCB_STATE_DEAD)
                return (EINVAL);
-       }
+#if NECP
+       else if (necp_socket_should_use_flow_divert(inp))
+               return (EPROTOTYPE);
+#if CONTENT_FILTER
+       error = cfil_sock_attach(so);
+       if (error != 0)
+               return (error);
+#endif /* CONTENT_FILTER */
+#endif /* NECP */
+
        tp = intotcpcb(inp);
        TCPDEBUG1();
-       in_setpeeraddr(so, nam);
+
+       calculate_tcp_clock();
+
        COMMON_END(PRU_ACCEPT);
 }
 
@@ -555,11 +824,23 @@ tcp6_usr_accept(struct socket *so, struct sockaddr **nam)
                error = ECONNABORTED;
                goto out;
        }
-       if (inp == 0 || (inp->inp_state == INPCB_STATE_DEAD)) {
+       if (inp == NULL || inp->inp_state == INPCB_STATE_DEAD)
                return (EINVAL);
-       }
+#if NECP
+       else if (necp_socket_should_use_flow_divert(inp))
+               return (EPROTOTYPE);
+#if CONTENT_FILTER
+       error = cfil_sock_attach(so);
+       if (error != 0)
+               return (error);
+#endif /* CONTENT_FILTER */
+#endif /* NECP */
+
        tp = intotcpcb(inp);
        TCPDEBUG1();
+
+       calculate_tcp_clock();
+
        in6_mapped_peeraddr(so, nam);
        COMMON_END(PRU_ACCEPT);
 }
@@ -589,12 +870,45 @@ tcp_usr_shutdown(struct socket *so)
        struct inpcb *inp = sotoinpcb(so);
        struct tcpcb *tp;
 
-       COMMON_START();
+       TCPDEBUG0;
+       if (inp == NULL || inp->inp_state == INPCB_STATE_DEAD)
+               return (EINVAL);
+
        socantsendmore(so);
-        /* In case we got disconnected from the peer */
-        if (tp == 0)
-            goto out;
+
+        /*
+        * In case we got disconnected from the peer, or if this is
+        * a socket that is to be flow-diverted (but not yet).
+        */
+       tp = intotcpcb(inp);
+       TCPDEBUG1();
+
+       if (tp == NULL
+#if NECP
+               || (necp_socket_should_use_flow_divert(inp))
+#endif /* NECP */
+               ) {
+               if (tp != NULL)
+                       error = EPROTOTYPE;
+               goto out;
+       }
+
+       calculate_tcp_clock();
+
        tp = tcp_usrclosed(tp);
+#if MPTCP
+       /* A reset has been sent but socket exists, do not send FIN */
+       if ((so->so_flags & SOF_MP_SUBFLOW) &&
+           (tp) && (tp->t_mpflags & TMPF_RESET)) {
+               goto out;
+       }
+#endif
+#if CONTENT_FILTER
+       /* Don't send a FIN yet */
+       if (tp && !(so->so_state & SS_ISDISCONNECTED) &&
+               cfil_sock_data_pending(&so->so_snd))
+               goto out;
+#endif /* CONTENT_FILTER */
        if (tp)
                error = tcp_output(tp);
        COMMON_END(PRU_SHUTDOWN);
@@ -612,9 +926,22 @@ tcp_usr_rcvd(struct socket *so, __unused int flags)
 
        COMMON_START();
         /* In case we got disconnected from the peer */
-        if (tp == 0)
-            goto out;
-       tcp_output(tp);
+        if (tp == NULL)
+               goto out;
+       tcp_sbrcv_trim(tp, &so->so_rcv);
+
+       /*
+        * This tcp_output is solely there to trigger window-updates.
+        * However, we really do not want these window-updates while we
+        * are still in SYN_SENT or SYN_RECEIVED.
+        */
+       if (TCPS_HAVEESTABLISHED(tp->t_state))
+               tcp_output(tp);
+
+#if CONTENT_FILTER
+       cfil_sock_buf_update(&so->so_rcv);
+#endif /* CONTENT_FILTER */
+
        COMMON_END(PRU_RCVD);
 }
 
@@ -650,28 +977,39 @@ tcp_usr_rcvd(struct socket *so, __unused int flags)
  *     tcp6_connect:???                [IPV6 only]
  */
 static int
-tcp_usr_send(struct socket *so, int flags, struct mbuf *m, 
-            struct sockaddr *nam, struct mbuf *control, struct proc *p)
+tcp_usr_send(struct socket *so, int flags, struct mbuf *m,
+     struct sockaddr *nam, struct mbuf *control, struct proc *p)
 {
        int error = 0;
        struct inpcb *inp = sotoinpcb(so);
        struct tcpcb *tp;
+       uint32_t msgpri = MSG_PRI_DEFAULT;
 #if INET6
        int isipv6;
 #endif
        TCPDEBUG0;
 
-       if (inp == NULL || inp->inp_state == INPCB_STATE_DEAD) {
+       if (inp == NULL || inp->inp_state == INPCB_STATE_DEAD
+#if NECP
+               || (necp_socket_should_use_flow_divert(inp))
+#endif /* NECP */
+               ) {
                /*
                 * OOPS! we lost a race, the TCP session got reset after
                 * we checked SS_CANTSENDMORE, eg: while doing uiomove or a
                 * network interrupt in the non-splnet() section of sosend().
                 */
-               if (m)
+               if (m != NULL)
                        m_freem(m);
-               if (control)
+               if (control != NULL) {
                        m_freem(control);
-               error = ECONNRESET;     /* XXX EPIPE? */
+                       control = NULL;
+               }
+
+               if (inp == NULL)
+                       error = ECONNRESET;     /* XXX EPIPE? */
+               else
+                       error = EPROTOTYPE;
                tp = NULL;
                TCPDEBUG1();
                goto out;
@@ -681,20 +1019,56 @@ tcp_usr_send(struct socket *so, int flags, struct mbuf *m,
 #endif /* INET6 */
        tp = intotcpcb(inp);
        TCPDEBUG1();
-       if (control) {
-               /* TCP doesn't do control messages (rights, creds, etc) */
-               if (control->m_len) {
+
+       calculate_tcp_clock();
+
+       if (control != NULL) {
+               if (so->so_flags & SOF_ENABLE_MSGS) {
+                       /* Get the msg priority from control mbufs */
+                       error = tcp_get_msg_priority(control, &msgpri);
+                       if (error) {
+                               m_freem(control);
+                               if (m != NULL)
+                                       m_freem(m);
+                               control = NULL;
+                               m = NULL;
+                               goto out;
+                       }
+                       m_freem(control);
+                       control = NULL;
+               } else if (control->m_len) {
+                       /* 
+                        * if not unordered, TCP should not have 
+                        * control mbufs
+                        */
                        m_freem(control);
-                       if (m)
+                       if (m != NULL)
                                m_freem(m);
+                       control = NULL;
+                       m = NULL;
                        error = EINVAL;
                        goto out;
                }
-               m_freem(control);       /* empty control, just free it */
        }
-       if(!(flags & PRUS_OOB)) {
-               sbappendstream(&so->so_snd, m);
+
+       if (so->so_flags & SOF_ENABLE_MSGS) {
+               VERIFY(m->m_flags & M_PKTHDR);
+               m->m_pkthdr.msg_pri = msgpri;
+       }
+
+       /* MPTCP sublow socket buffers must not be compressed */
+       VERIFY(!(so->so_flags & SOF_MP_SUBFLOW) ||
+           (so->so_snd.sb_flags & SB_NOCOMPRESS));
+
+       if(!(flags & PRUS_OOB) || (so->so_flags1 & SOF1_PRECONNECT_DATA)) {
+               /* Call msg send if message delivery is enabled */
+               if (so->so_flags & SOF_ENABLE_MSGS)
+                       sbappendmsg_snd(&so->so_snd, m);
+               else
+                       sbappendstream(&so->so_snd, m);
+
                if (nam && tp->t_state < TCPS_SYN_SENT) {
+
                        /*
                         * Do implied connect if not yet connected,
                         * initialize window to default value, and
@@ -706,11 +1080,11 @@ tcp_usr_send(struct socket *so, int flags, struct mbuf *m,
                                error = tcp6_connect(tp, nam, p);
                        else
 #endif /* INET6 */
-                       error = tcp_connect(tp, nam, p);
+                               error = tcp_connect(tp, nam, p);
                        if (error)
                                goto out;
                        tp->snd_wnd = TTCP_CLIENT_SND_WND;
-                       tcp_mss(tp, -1);
+                       tcp_mss(tp, -1, IFSCOPE_NONE);
                }
 
                if (flags & PRUS_EOF) {
@@ -729,7 +1103,9 @@ tcp_usr_send(struct socket *so, int flags, struct mbuf *m,
                                tp->t_flags &= ~TF_MORETOCOME;
                }
        } else {
-               if (sbspace(&so->so_snd) < -512) {
+               if (sbspace(&so->so_snd) == 0) { 
+                       /* if no space is left in sockbuf, 
+                        * do not try to squeeze in OOB traffic */
                        m_freem(m);
                        error = ENOBUFS;
                        goto out;
@@ -759,13 +1135,26 @@ tcp_usr_send(struct socket *so, int flags, struct mbuf *m,
                        if (error)
                                goto out;
                        tp->snd_wnd = TTCP_CLIENT_SND_WND;
-                       tcp_mss(tp, -1);
+                       tcp_mss(tp, -1, IFSCOPE_NONE);
                }
                tp->snd_up = tp->snd_una + so->so_snd.sb_cc;
-               tp->t_force = 1;
+               tp->t_flagsext |= TF_FORCE;
                error = tcp_output(tp);
-               tp->t_force = 0;
+               tp->t_flagsext &= ~TF_FORCE;
+       }
+
+
+       /*
+        * We wait for the socket to successfully connect before returning.
+        * This allows us to signal a timeout to the application.
+        */
+       if (so->so_state & SS_ISCONNECTING) {
+               if (so->so_state & SS_NBIO)
+                       error = EWOULDBLOCK;
+               else
+                       error = sbwait(&so->so_snd);
        }
+
        COMMON_END((flags & PRUS_OOB) ? PRU_SENDOOB : 
                   ((flags & PRUS_EOF) ? PRU_SEND_EOF : PRU_SEND));
 }
@@ -782,8 +1171,8 @@ tcp_usr_abort(struct socket *so)
 
        COMMON_START();
         /* In case we got disconnected from the peer */
-        if (tp == 0)
-            goto out;
+        if (tp == NULL)
+               goto out;
        tp = tcp_drop(tp, ECONNABORTED);
        so->so_usecount--;
        COMMON_END(PRU_ABORT);
@@ -818,27 +1207,82 @@ tcp_usr_rcvoob(struct socket *so, struct mbuf *m, int flags)
        }
        m->m_len = 1;
        *mtod(m, caddr_t) = tp->t_iobc;
+       so->so_state &= ~SS_RCVATMARK;
        if ((flags & MSG_PEEK) == 0)
                tp->t_oobflags ^= (TCPOOB_HAVEDATA | TCPOOB_HADDATA);
        COMMON_END(PRU_RCVOOB);
 }
 
+static int
+tcp_usr_preconnect(struct socket *so)
+{
+       struct inpcb *inp = sotoinpcb(so);
+       int error = 0;
+
+#if NECP
+       if (necp_socket_should_use_flow_divert(inp)) {
+               /* May happen, if in tcp_usr_connect we did not had a chance
+                * to set the usrreqs (due to some error). So, let's get out
+                * of here.
+                */
+               goto out;
+       }
+#endif /* NECP */
+
+       error = tcp_output(sototcpcb(so));
+
+       /* One read has been done. This was enough. Get back to "normal" behavior. */
+       so->so_flags1 &= ~SOF1_PRECONNECT_DATA;
+
+       COMMON_END(PRU_PRECONNECT);
+}
+
 /* xxx - should be const */
 struct pr_usrreqs tcp_usrreqs = {
-       tcp_usr_abort, tcp_usr_accept, tcp_usr_attach, tcp_usr_bind,
-       tcp_usr_connect, pru_connect2_notsupp, in_control, tcp_usr_detach,
-       tcp_usr_disconnect, tcp_usr_listen, in_setpeeraddr, tcp_usr_rcvd,
-       tcp_usr_rcvoob, tcp_usr_send, pru_sense_null, tcp_usr_shutdown,
-       in_setsockaddr, sosend, soreceive, pru_sopoll_notsupp
+       .pru_abort =            tcp_usr_abort,
+       .pru_accept =           tcp_usr_accept,
+       .pru_attach =           tcp_usr_attach,
+       .pru_bind =             tcp_usr_bind,
+       .pru_connect =          tcp_usr_connect,
+       .pru_connectx =         tcp_usr_connectx,
+       .pru_control =          in_control,
+       .pru_detach =           tcp_usr_detach,
+       .pru_disconnect =       tcp_usr_disconnect,
+       .pru_disconnectx =      tcp_usr_disconnectx,
+       .pru_listen =           tcp_usr_listen,
+       .pru_peeraddr =         in_getpeeraddr,
+       .pru_rcvd =             tcp_usr_rcvd,
+       .pru_rcvoob =           tcp_usr_rcvoob,
+       .pru_send =             tcp_usr_send,
+       .pru_shutdown =         tcp_usr_shutdown,
+       .pru_sockaddr =         in_getsockaddr,
+       .pru_sosend =           sosend,
+       .pru_soreceive =        soreceive,
+       .pru_preconnect =       tcp_usr_preconnect,
 };
 
 #if INET6
 struct pr_usrreqs tcp6_usrreqs = {
-       tcp_usr_abort, tcp6_usr_accept, tcp_usr_attach, tcp6_usr_bind,
-       tcp6_usr_connect, pru_connect2_notsupp, in6_control, tcp_usr_detach,
-       tcp_usr_disconnect, tcp6_usr_listen, in6_mapped_peeraddr, tcp_usr_rcvd,
-       tcp_usr_rcvoob, tcp_usr_send, pru_sense_null, tcp_usr_shutdown,
-       in6_mapped_sockaddr, sosend, soreceive, pru_sopoll_notsupp
+       .pru_abort =            tcp_usr_abort,
+       .pru_accept =           tcp6_usr_accept,
+       .pru_attach =           tcp_usr_attach,
+       .pru_bind =             tcp6_usr_bind,
+       .pru_connect =          tcp6_usr_connect,
+       .pru_connectx =         tcp6_usr_connectx,
+       .pru_control =          in6_control,
+       .pru_detach =           tcp_usr_detach,
+       .pru_disconnect =       tcp_usr_disconnect,
+       .pru_disconnectx =      tcp_usr_disconnectx,
+       .pru_listen =           tcp6_usr_listen,
+       .pru_peeraddr =         in6_mapped_peeraddr,
+       .pru_rcvd =             tcp_usr_rcvd,
+       .pru_rcvoob =           tcp_usr_rcvoob,
+       .pru_send =             tcp_usr_send,
+       .pru_shutdown =         tcp_usr_shutdown,
+       .pru_sockaddr =         in6_mapped_sockaddr,
+       .pru_sosend =           sosend,
+       .pru_soreceive =        soreceive,
+       .pru_preconnect =       tcp_usr_preconnect,
 };
 #endif /* INET6 */
 
@@ -867,24 +1311,20 @@ struct pr_usrreqs tcp6_usrreqs = {
  *     in_pcbladdr:EADDRNOTAVAIL       Address not available
  */
 static int
-tcp_connect(tp, nam, p)
-       register struct tcpcb *tp;
-       struct sockaddr *nam;
-       struct proc *p;
+tcp_connect(struct tcpcb *tp, struct sockaddr *nam, struct proc *p)
 {
        struct inpcb *inp = tp->t_inpcb, *oinp;
        struct socket *so = inp->inp_socket;
        struct tcpcb *otp;
-       struct sockaddr_in *sin = (struct sockaddr_in *)nam;
-       struct sockaddr_in *ifaddr;
-       struct rmxp_tao *taop;
-       struct rmxp_tao tao_noncached;
-       int error;
+       struct sockaddr_in *sin = (struct sockaddr_in *)(void *)nam;
+       struct in_addr laddr;
+       int error = 0;
+       struct ifnet *outif = NULL;
 
        if (inp->inp_lport == 0) {
-               error = in_pcbbind(inp, (struct sockaddr *)0, p);
+               error = in_pcbbind(inp, NULL, p);
                if (error)
-                       return error;
+                       goto done;
        }
 
        /*
@@ -892,15 +1332,14 @@ tcp_connect(tp, nam, p)
         * earlier incarnation of this same connection still in
         * TIME_WAIT state, creating an ADDRINUSE error.
         */
-       error = in_pcbladdr(inp, nam, &ifaddr);
+       error = in_pcbladdr(inp, nam, &laddr, IFSCOPE_NONE, &outif, 0);
        if (error)
-               return error;
+               goto done;
 
        tcp_unlock(inp->inp_socket, 0, 0);
        oinp = in_pcblookup_hash(inp->inp_pcbinfo,
            sin->sin_addr, sin->sin_port,
-           inp->inp_laddr.s_addr != INADDR_ANY ? inp->inp_laddr
-                                               : ifaddr->sin_addr,
+           inp->inp_laddr.s_addr != INADDR_ANY ? inp->inp_laddr : laddr,
            inp->inp_lport,  0, NULL);
 
        tcp_lock(inp->inp_socket, 0, 0);
@@ -914,113 +1353,96 @@ tcp_connect(tp, nam, p)
                }
 
                if (oinp != inp && (otp = intotcpcb(oinp)) != NULL &&
-               otp->t_state == TCPS_TIME_WAIT &&
-                   otp->t_starttime < (u_long)tcp_msl &&
-                   (otp->t_flags & TF_RCVD_CC))
+                   otp->t_state == TCPS_TIME_WAIT &&
+                   ((int)(tcp_now - otp->t_starttime)) < tcp_msl &&
+                   (otp->t_flags & TF_RCVD_CC)) {
                        otp = tcp_close(otp);
-               else {
-                       printf("tcp_connect: inp=%p err=EADDRINUSE\n", inp);
+               } else {
+                       printf("tcp_connect: inp=0x%llx err=EADDRINUSE\n",
+                           (uint64_t)VM_KERNEL_ADDRPERM(inp));
                        if (oinp != inp)
                                tcp_unlock(oinp->inp_socket, 1, 0);
-                       return EADDRINUSE;
+                       error = EADDRINUSE;
+                       goto done;
                }
                if (oinp != inp)
                        tcp_unlock(oinp->inp_socket, 1, 0);
        }
 skip_oinp:
-       if ((inp->inp_laddr.s_addr == INADDR_ANY ? ifaddr->sin_addr.s_addr :
-                inp->inp_laddr.s_addr) == sin->sin_addr.s_addr &&
-           inp->inp_lport == sin->sin_port)
-                       return EINVAL;
-       if (!lck_rw_try_lock_exclusive(inp->inp_pcbinfo->mtx)) {
+       if ((inp->inp_laddr.s_addr == INADDR_ANY ? laddr.s_addr :
+           inp->inp_laddr.s_addr) == sin->sin_addr.s_addr &&
+           inp->inp_lport == sin->sin_port) {
+               error = EINVAL;
+               goto done;
+       }
+       if (!lck_rw_try_lock_exclusive(inp->inp_pcbinfo->ipi_lock)) {
                /*lock inversion issue, mostly with udp multicast packets */
                socket_unlock(inp->inp_socket, 0);
-               lck_rw_lock_exclusive(inp->inp_pcbinfo->mtx);
+               lck_rw_lock_exclusive(inp->inp_pcbinfo->ipi_lock);
                socket_lock(inp->inp_socket, 0);
        }
-       if (inp->inp_laddr.s_addr == INADDR_ANY)
-               inp->inp_laddr = ifaddr->sin_addr;
+       if (inp->inp_laddr.s_addr == INADDR_ANY) {
+               inp->inp_laddr = laddr;
+               /* no reference needed */
+               inp->inp_last_outifp = outif;
+               inp->inp_flags |= INP_INADDR_ANY;
+       }
        inp->inp_faddr = sin->sin_addr;
        inp->inp_fport = sin->sin_port;
        in_pcbrehash(inp);
-       lck_rw_done(inp->inp_pcbinfo->mtx);
+       lck_rw_done(inp->inp_pcbinfo->ipi_lock);
 
-       /* Compute window scaling to requesti according to sb_hiwat
-        * or leave us some room to increase potentially increase the window size depending
-        * on the default win scale
-        */
-       while (tp->request_r_scale < TCP_MAX_WINSHIFT &&
-        (TCP_MAXWIN << tp->request_r_scale) < so->so_rcv.sb_hiwat)
-               tp->request_r_scale++;
-
-       /*
-        * Inflate window size only if no setsockopt was performed on the recv sockbuf and
-        * if we're not over our number of active pcbs.
-        */
+       if (inp->inp_flowhash == 0)
+               inp->inp_flowhash = inp_calc_flowhash(inp);
 
-       if (((so->so_rcv.sb_flags & SB_USRSIZE) == 0) && (inp->inp_pcbinfo->ipi_count < tcp_sockthreshold)) {
-               tp->request_r_scale = max(tcp_win_scale, tp->request_r_scale);
-               so->so_rcv.sb_hiwat = min(TCP_MAXWIN << tp->request_r_scale, (sb_max / (MSIZE+MCLBYTES)) * MCLBYTES);  
-       }
+       tcp_set_max_rwinscale(tp, so, TCP_AUTORCVBUF_MAX(outif));
 
        soisconnecting(so);
        tcpstat.tcps_connattempt++;
        tp->t_state = TCPS_SYN_SENT;
-       tp->t_timer[TCPT_KEEP] = tcp_keepinit;
+       tp->t_timer[TCPT_KEEP] = OFFSET_FROM_START(tp, TCP_CONN_KEEPINIT(tp));
        tp->iss = tcp_new_isn(tp);
        tcp_sendseqinit(tp);
+       if (nstat_collect)
+               nstat_route_connect_attempt(inp->inp_route.ro_rt);
 
-       /*
-        * Generate a CC value for this connection and
-        * check whether CC or CCnew should be used.
-        */
-       if ((taop = tcp_gettaocache(tp->t_inpcb)) == NULL) {
-               taop = &tao_noncached;
-               bzero(taop, sizeof(*taop));
-       }
-
-       tp->cc_send = CC_INC(tcp_ccgen);
-       if (taop->tao_ccsent != 0 &&
-           CC_GEQ(tp->cc_send, taop->tao_ccsent)) {
-               taop->tao_ccsent = tp->cc_send;
-       } else {
-               taop->tao_ccsent = 0;
-               tp->t_flags |= TF_SENDCCNEW;
-       }
+done:
+       if (outif != NULL)
+               ifnet_release(outif);
 
-       return 0;
+       return (error);
 }
 
 #if INET6
 static int
-tcp6_connect(tp, nam, p)
-       register struct tcpcb *tp;
-       struct sockaddr *nam;
-       struct proc *p;
+tcp6_connect(struct tcpcb *tp, struct sockaddr *nam, struct proc *p)
 {
        struct inpcb *inp = tp->t_inpcb, *oinp;
        struct socket *so = inp->inp_socket;
        struct tcpcb *otp;
-       struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)nam;
+       struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)(void *)nam;
        struct in6_addr addr6;
-       struct rmxp_tao *taop;
-       struct rmxp_tao tao_noncached;
-       int error;
+       int error = 0;
+       struct ifnet *outif = NULL;
 
        if (inp->inp_lport == 0) {
-               error = in6_pcbbind(inp, (struct sockaddr *)0, p);
+               error = in6_pcbbind(inp, NULL, p);
                if (error)
-                       return error;
+                       goto done;
        }
 
        /*
         * Cannot simply call in_pcbconnect, because there might be an
         * earlier incarnation of this same connection still in
         * TIME_WAIT state, creating an ADDRINUSE error.
+        *
+        * in6_pcbladdr() might return an ifp with its reference held
+        * even in the error case, so make sure that it's released
+        * whenever it's non-NULL.
         */
-       error = in6_pcbladdr(inp, nam, &addr6);
+       error = in6_pcbladdr(inp, nam, &addr6, &outif);
        if (error)
-               return error;
+               goto done;
        tcp_unlock(inp->inp_socket, 0, 0);
        oinp = in6_pcblookup_hash(inp->inp_pcbinfo,
                                  &sin6->sin6_addr, sin6->sin6_port,
@@ -1032,72 +1454,431 @@ tcp6_connect(tp, nam, p)
        if (oinp) {
                if (oinp != inp && (otp = intotcpcb(oinp)) != NULL &&
                    otp->t_state == TCPS_TIME_WAIT &&
-                   otp->t_starttime < (u_long)tcp_msl &&
-                   (otp->t_flags & TF_RCVD_CC))
+                   ((int)(tcp_now - otp->t_starttime)) < tcp_msl &&
+                   (otp->t_flags & TF_RCVD_CC)) {
                        otp = tcp_close(otp);
-               else
-                       return EADDRINUSE;
+               } else {
+                       error = EADDRINUSE;
+                       goto done;
+               }
        }
-       if (!lck_rw_try_lock_exclusive(inp->inp_pcbinfo->mtx)) {
+       if (!lck_rw_try_lock_exclusive(inp->inp_pcbinfo->ipi_lock)) {
                /*lock inversion issue, mostly with udp multicast packets */
                socket_unlock(inp->inp_socket, 0);
-               lck_rw_lock_exclusive(inp->inp_pcbinfo->mtx);
+               lck_rw_lock_exclusive(inp->inp_pcbinfo->ipi_lock);
                socket_lock(inp->inp_socket, 0);
        }
-       if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr))
+       if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)) {
                inp->in6p_laddr = addr6;
+               inp->in6p_last_outifp = outif;  /* no reference needed */
+               inp->in6p_flags |= INP_IN6ADDR_ANY;
+       }
        inp->in6p_faddr = sin6->sin6_addr;
        inp->inp_fport = sin6->sin6_port;
        if ((sin6->sin6_flowinfo & IPV6_FLOWINFO_MASK) != 0)
-               inp->in6p_flowinfo = sin6->sin6_flowinfo;
+               inp->inp_flow = sin6->sin6_flowinfo;
        in_pcbrehash(inp);
-       lck_rw_done(inp->inp_pcbinfo->mtx);
+       lck_rw_done(inp->inp_pcbinfo->ipi_lock);
+
+       if (inp->inp_flowhash == 0)
+               inp->inp_flowhash = inp_calc_flowhash(inp);
+       /* update flowinfo - RFC 6437 */
+       if (inp->inp_flow == 0 && inp->in6p_flags & IN6P_AUTOFLOWLABEL) {
+               inp->inp_flow &= ~IPV6_FLOWLABEL_MASK;
+               inp->inp_flow |=
+                   (htonl(inp->inp_flowhash) & IPV6_FLOWLABEL_MASK);
+       }
 
-       /* 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++;
+       tcp_set_max_rwinscale(tp, so, TCP_AUTORCVBUF_MAX(outif));
 
        soisconnecting(so);
        tcpstat.tcps_connattempt++;
        tp->t_state = TCPS_SYN_SENT;
-       tp->t_timer[TCPT_KEEP] = tcp_keepinit;
+       tp->t_timer[TCPT_KEEP] = OFFSET_FROM_START(tp, 
+               TCP_CONN_KEEPINIT(tp));
        tp->iss = tcp_new_isn(tp);
        tcp_sendseqinit(tp);
+       if (nstat_collect)
+               nstat_route_connect_attempt(inp->inp_route.ro_rt);
 
-       /*
-        * Generate a CC value for this connection and
-        * check whether CC or CCnew should be used.
-        */
-       if ((taop = tcp_gettaocache(tp->t_inpcb)) == NULL) {
-               taop = &tao_noncached;
-               bzero(taop, sizeof(*taop));
-       }
-
-       tp->cc_send = CC_INC(tcp_ccgen);
-       if (taop->tao_ccsent != 0 &&
-           CC_GEQ(tp->cc_send, taop->tao_ccsent)) {
-               taop->tao_ccsent = tp->cc_send;
-       } else {
-               taop->tao_ccsent = 0;
-               tp->t_flags |= TF_SENDCCNEW;
-       }
+done:
+       if (outif != NULL)
+               ifnet_release(outif);
 
-       return 0;
+       return (error);
 }
 #endif /* INET6 */
 
 /*
- * The new sockopt interface makes it possible for us to block in the
- * copyin/out step (if we take a page fault).  Taking a page fault at
- * splnet() is probably a Bad Thing.  (Since sockets and pcbs both now
- * use TSM, there probably isn't any need for this function to run at
- * splnet() any more.  This needs more examination.)
+ * Export TCP internal state information via a struct tcp_info
  */
-int
-tcp_ctloutput(so, sopt)
-       struct socket *so;
-       struct sockopt *sopt;
+__private_extern__ void
+tcp_fill_info(struct tcpcb *tp, struct tcp_info *ti)
+{
+       struct inpcb *inp = tp->t_inpcb;
+       
+       bzero(ti, sizeof(*ti));
+
+       ti->tcpi_state = tp->t_state;
+       ti->tcpi_flowhash = inp->inp_flowhash;
+
+       if (tp->t_state > TCPS_LISTEN) {
+               if (TSTMP_SUPPORTED(tp))
+                       ti->tcpi_options |= TCPI_OPT_TIMESTAMPS;
+               if (SACK_ENABLED(tp))
+                       ti->tcpi_options |= TCPI_OPT_SACK;
+               if (TCP_WINDOW_SCALE_ENABLED(tp)) {
+                       ti->tcpi_options |= TCPI_OPT_WSCALE;
+                       ti->tcpi_snd_wscale = tp->snd_scale;
+                       ti->tcpi_rcv_wscale = tp->rcv_scale;
+               }
+               if (TCP_ECN_ENABLED(tp))
+                       ti->tcpi_options |= TCPI_OPT_ECN;
+
+               /* Are we in retranmission episode */
+               if (IN_FASTRECOVERY(tp) || tp->t_rxtshift > 0)
+                       ti->tcpi_flags |= TCPI_FLAG_LOSSRECOVERY;
+
+               if (tp->t_flags & TF_STREAMING_ON)
+                       ti->tcpi_flags |= TCPI_FLAG_STREAMING_ON;
+
+               ti->tcpi_rto = tp->t_timer[TCPT_REXMT] ? tp->t_rxtcur : 0;
+               ti->tcpi_snd_mss = tp->t_maxseg;
+               ti->tcpi_rcv_mss = tp->t_maxseg;
+
+               ti->tcpi_rttcur = tp->t_rttcur;
+               ti->tcpi_srtt = tp->t_srtt >> TCP_RTT_SHIFT;
+               ti->tcpi_rttvar = tp->t_rttvar >> TCP_RTTVAR_SHIFT;
+               ti->tcpi_rttbest = tp->t_rttbest >> TCP_RTT_SHIFT;
+
+               ti->tcpi_snd_ssthresh = tp->snd_ssthresh;
+               ti->tcpi_snd_cwnd = tp->snd_cwnd;
+               ti->tcpi_snd_sbbytes = inp->inp_socket->so_snd.sb_cc;
+       
+               ti->tcpi_rcv_space = tp->rcv_wnd;
+
+               ti->tcpi_snd_wnd = tp->snd_wnd;
+               ti->tcpi_snd_nxt = tp->snd_nxt;
+               ti->tcpi_rcv_nxt = tp->rcv_nxt;
+
+               /* convert bytes/msec to bits/sec */
+               if ((tp->t_flagsext & TF_MEASURESNDBW) != 0 &&
+                       tp->t_bwmeas != NULL) {
+                       ti->tcpi_snd_bw = (tp->t_bwmeas->bw_sndbw * 8000);
+               }
+               
+               ti->tcpi_last_outif = (tp->t_inpcb->inp_last_outifp == NULL) ? 0 :
+                   tp->t_inpcb->inp_last_outifp->if_index;
+
+               //atomic_get_64(ti->tcpi_txbytes, &inp->inp_stat->txbytes);
+               ti->tcpi_txpackets = inp->inp_stat->txpackets;
+               ti->tcpi_txbytes = inp->inp_stat->txbytes;
+               ti->tcpi_txretransmitbytes = tp->t_stat.txretransmitbytes;
+               ti->tcpi_txunacked = tp->snd_max - tp->snd_una;
+               
+               //atomic_get_64(ti->tcpi_rxbytes, &inp->inp_stat->rxbytes);
+               ti->tcpi_rxpackets = inp->inp_stat->rxpackets;
+               ti->tcpi_rxbytes = inp->inp_stat->rxbytes;
+               ti->tcpi_rxduplicatebytes = tp->t_stat.rxduplicatebytes;
+               ti->tcpi_rxoutoforderbytes = tp->t_stat.rxoutoforderbytes;
+
+               if (tp->t_state > TCPS_LISTEN) {
+                       ti->tcpi_synrexmits = tp->t_stat.synrxtshift;
+               }
+               ti->tcpi_cell_rxpackets = inp->inp_cstat->rxpackets;
+               ti->tcpi_cell_rxbytes = inp->inp_cstat->rxbytes;
+               ti->tcpi_cell_txpackets = inp->inp_cstat->txpackets;
+               ti->tcpi_cell_txbytes = inp->inp_cstat->txbytes;
+
+               ti->tcpi_wifi_rxpackets = inp->inp_wstat->rxpackets;
+               ti->tcpi_wifi_rxbytes = inp->inp_wstat->rxbytes;
+               ti->tcpi_wifi_txpackets = inp->inp_wstat->txpackets;
+               ti->tcpi_wifi_txbytes = inp->inp_wstat->txbytes;
+
+               ti->tcpi_wired_rxpackets = inp->inp_Wstat->rxpackets;
+               ti->tcpi_wired_rxbytes = inp->inp_Wstat->rxbytes;
+               ti->tcpi_wired_txpackets = inp->inp_Wstat->txpackets;
+               ti->tcpi_wired_txbytes = inp->inp_Wstat->txbytes;
+               tcp_get_connectivity_status(tp, &ti->tcpi_connstatus);
+
+               ti->tcpi_tfo_syn_data_rcv = !!(tp->t_tfo_stats & TFO_S_SYNDATA_RCV);
+               ti->tcpi_tfo_cookie_req_rcv = !!(tp->t_tfo_stats & TFO_S_COOKIEREQ_RECV);
+               ti->tcpi_tfo_cookie_sent = !!(tp->t_tfo_stats & TFO_S_COOKIE_SENT);
+               ti->tcpi_tfo_cookie_invalid = !!(tp->t_tfo_stats & TFO_S_COOKIE_INVALID);
+
+               ti->tcpi_tfo_cookie_req = !!(tp->t_tfo_stats & TFO_S_COOKIE_REQ);
+               ti->tcpi_tfo_cookie_rcv = !!(tp->t_tfo_stats & TFO_S_COOKIE_RCV);
+               ti->tcpi_tfo_syn_data_sent = !!(tp->t_tfo_stats & TFO_S_SYN_DATA_SENT);
+               ti->tcpi_tfo_syn_data_acked = !!(tp->t_tfo_stats & TFO_S_SYN_DATA_ACKED);
+               ti->tcpi_tfo_syn_loss = !!(tp->t_tfo_stats & TFO_S_SYN_LOSS);
+               ti->tcpi_tfo_cookie_wrong = !!(tp->t_tfo_stats & TFO_S_COOKIE_WRONG);
+               ti->tcpi_tfo_no_cookie_rcv = !!(tp->t_tfo_stats & TFO_S_NO_COOKIE_RCV);
+               ti->tcpi_tfo_heuristics_disable = !!(tp->t_tfo_stats & TFO_S_HEURISTICS_DISABLE);
+               ti->tcpi_tfo_send_blackhole = !!(tp->t_tfo_stats & TFO_S_SEND_BLACKHOLE);
+               ti->tcpi_tfo_recv_blackhole = !!(tp->t_tfo_stats & TFO_S_RECV_BLACKHOLE);
+
+               ti->tcpi_ecn_client_setup = !!(tp->ecn_flags & TE_SETUPSENT);
+               ti->tcpi_ecn_server_setup = !!(tp->ecn_flags & TE_SETUPRECEIVED);
+               ti->tcpi_ecn_success = (tp->ecn_flags & TE_ECN_ON) == TE_ECN_ON ? 1 : 0;
+               ti->tcpi_ecn_lost_syn = !!(tp->ecn_flags & TE_LOST_SYN);
+               ti->tcpi_ecn_lost_synack = !!(tp->ecn_flags & TE_LOST_SYNACK);
+
+               ti->tcpi_local_peer = !!(tp->t_flags & TF_LOCAL);
+
+               if (tp->t_inpcb->inp_last_outifp != NULL) {
+                       if (IFNET_IS_CELLULAR(tp->t_inpcb->inp_last_outifp))
+                               ti->tcpi_if_cell = 1;
+                       else if (IFNET_IS_WIFI(tp->t_inpcb->inp_last_outifp))
+                               ti->tcpi_if_wifi = 1;
+               }
+
+               ti->tcpi_ecn_recv_ce = tp->t_ecn_recv_ce;
+               ti->tcpi_ecn_recv_cwr = tp->t_ecn_recv_cwr;
+
+               ti->tcpi_rcvoopack = tp->t_rcvoopack;
+               ti->tcpi_pawsdrop = tp->t_pawsdrop;
+               ti->tcpi_sack_recovery_episode = tp->t_sack_recovery_episode;
+               ti->tcpi_reordered_pkts = tp->t_reordered_pkts;
+               ti->tcpi_dsack_sent = tp->t_dsack_sent;
+               ti->tcpi_dsack_recvd = tp->t_dsack_recvd;
+       }
+}
+
+__private_extern__ errno_t
+tcp_fill_info_for_info_tuple(struct info_tuple *itpl, struct tcp_info *ti)
+{
+       struct inpcbinfo *pcbinfo = NULL;
+       struct inpcb *inp = NULL;
+       struct socket *so;
+       struct tcpcb *tp;
+       
+       if (itpl->itpl_proto == IPPROTO_TCP)
+               pcbinfo = &tcbinfo;
+       else
+               return EINVAL;
+       
+       if (itpl->itpl_local_sa.sa_family == AF_INET &&
+               itpl->itpl_remote_sa.sa_family == AF_INET) {
+               inp = in_pcblookup_hash(pcbinfo, 
+                       itpl->itpl_remote_sin.sin_addr,
+                       itpl->itpl_remote_sin.sin_port,
+                       itpl->itpl_local_sin.sin_addr,
+                       itpl->itpl_local_sin.sin_port,
+                       0, NULL);
+       } else if (itpl->itpl_local_sa.sa_family == AF_INET6 &&
+               itpl->itpl_remote_sa.sa_family == AF_INET6) {
+               struct in6_addr ina6_local;
+               struct in6_addr ina6_remote;
+               
+               ina6_local = itpl->itpl_local_sin6.sin6_addr;
+               if (IN6_IS_SCOPE_LINKLOCAL(&ina6_local) && 
+                       itpl->itpl_local_sin6.sin6_scope_id)
+                       ina6_local.s6_addr16[1] = htons(itpl->itpl_local_sin6.sin6_scope_id);
+
+               ina6_remote = itpl->itpl_remote_sin6.sin6_addr;
+               if (IN6_IS_SCOPE_LINKLOCAL(&ina6_remote) && 
+                       itpl->itpl_remote_sin6.sin6_scope_id)
+                       ina6_remote.s6_addr16[1] = htons(itpl->itpl_remote_sin6.sin6_scope_id);
+               
+               inp = in6_pcblookup_hash(pcbinfo, 
+                       &ina6_remote,
+                       itpl->itpl_remote_sin6.sin6_port,
+                       &ina6_local,
+                       itpl->itpl_local_sin6.sin6_port,
+                       0, NULL);
+       } else {
+               return EINVAL;
+       }
+       if (inp == NULL || (so = inp->inp_socket) == NULL)
+               return ENOENT;
+
+       socket_lock(so, 0);
+       if (in_pcb_checkstate(inp, WNT_RELEASE, 1) == WNT_STOPUSING) {
+               socket_unlock(so, 0);
+               return ENOENT;
+       }
+       tp = intotcpcb(inp);
+
+       tcp_fill_info(tp, ti);
+       socket_unlock(so, 0);
+
+       return 0;
+}
+
+static void
+tcp_connection_fill_info(struct tcpcb *tp, struct tcp_connection_info *tci)
+{
+       struct inpcb *inp = tp->t_inpcb;
+
+       bzero(tci, sizeof(*tci));
+       tci->tcpi_state = tp->t_state;
+       if (tp->t_state > TCPS_LISTEN) {
+               if (TSTMP_SUPPORTED(tp))
+                       tci->tcpi_options |= TCPCI_OPT_TIMESTAMPS;
+               if (SACK_ENABLED(tp))
+                       tci->tcpi_options |= TCPCI_OPT_SACK;
+               if (TCP_WINDOW_SCALE_ENABLED(tp)) {
+                       tci->tcpi_options |= TCPCI_OPT_WSCALE;
+                       tci->tcpi_snd_wscale = tp->snd_scale;
+                       tci->tcpi_rcv_wscale = tp->rcv_scale;
+               }
+               if (TCP_ECN_ENABLED(tp))
+                       tci->tcpi_options |= TCPCI_OPT_ECN;
+               if (IN_FASTRECOVERY(tp) || tp->t_rxtshift > 0)
+                       tci->tcpi_flags |= TCPCI_FLAG_LOSSRECOVERY;
+               if (tp->t_flagsext & TF_PKTS_REORDERED)
+                       tci->tcpi_flags |= TCPCI_FLAG_REORDERING_DETECTED;
+               tci->tcpi_rto = (tp->t_timer[TCPT_REXMT] > 0) ?
+                       tp->t_rxtcur : 0;
+               tci->tcpi_maxseg = tp->t_maxseg;
+               tci->tcpi_snd_ssthresh = tp->snd_ssthresh;
+               tci->tcpi_snd_cwnd = tp->snd_cwnd;
+               tci->tcpi_snd_wnd = tp->snd_wnd;
+               tci->tcpi_snd_sbbytes = inp->inp_socket->so_snd.sb_cc;
+               tci->tcpi_rcv_wnd = tp->rcv_wnd;
+               tci->tcpi_rttcur = tp->t_rttcur;
+               tci->tcpi_srtt = (tp->t_srtt >> TCP_RTT_SHIFT);
+               tci->tcpi_rttvar = (tp->t_rttvar >> TCP_RTTVAR_SHIFT);
+               tci->tcpi_txpackets = inp->inp_stat->txpackets;
+               tci->tcpi_txbytes = inp->inp_stat->txbytes;
+               tci->tcpi_txretransmitbytes = tp->t_stat.txretransmitbytes;
+               tci->tcpi_rxpackets = inp->inp_stat->rxpackets;
+               tci->tcpi_rxbytes = inp->inp_stat->rxbytes;
+               tci->tcpi_rxoutoforderbytes = tp->t_stat.rxoutoforderbytes;
+
+               tci->tcpi_tfo_syn_data_rcv = !!(tp->t_tfo_stats & TFO_S_SYNDATA_RCV);
+               tci->tcpi_tfo_cookie_req_rcv = !!(tp->t_tfo_stats & TFO_S_COOKIEREQ_RECV);
+               tci->tcpi_tfo_cookie_sent = !!(tp->t_tfo_stats & TFO_S_COOKIE_SENT);
+               tci->tcpi_tfo_cookie_invalid = !!(tp->t_tfo_stats & TFO_S_COOKIE_INVALID);
+               tci->tcpi_tfo_cookie_req = !!(tp->t_tfo_stats & TFO_S_COOKIE_REQ);
+               tci->tcpi_tfo_cookie_rcv = !!(tp->t_tfo_stats & TFO_S_COOKIE_RCV);
+               tci->tcpi_tfo_syn_data_sent = !!(tp->t_tfo_stats & TFO_S_SYN_DATA_SENT);
+               tci->tcpi_tfo_syn_data_acked = !!(tp->t_tfo_stats & TFO_S_SYN_DATA_ACKED);
+               tci->tcpi_tfo_syn_loss = !!(tp->t_tfo_stats & TFO_S_SYN_LOSS);
+               tci->tcpi_tfo_cookie_wrong = !!(tp->t_tfo_stats & TFO_S_COOKIE_WRONG);
+               tci->tcpi_tfo_no_cookie_rcv = !!(tp->t_tfo_stats & TFO_S_NO_COOKIE_RCV);
+               tci->tcpi_tfo_heuristics_disable = !!(tp->t_tfo_stats & TFO_S_HEURISTICS_DISABLE);
+               tci->tcpi_tfo_send_blackhole = !!(tp->t_tfo_stats & TFO_S_SEND_BLACKHOLE);
+               tci->tcpi_tfo_recv_blackhole = !!(tp->t_tfo_stats & TFO_S_RECV_BLACKHOLE);
+       }
+}
+
+
+__private_extern__ int 
+tcp_sysctl_info(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       int error;
+       struct tcp_info ti;
+       struct info_tuple itpl;
+       proc_t caller = PROC_NULL;
+       proc_t caller_parent = PROC_NULL;
+       char command_name[MAXCOMLEN + 1] = "";
+       char parent_name[MAXCOMLEN + 1] = "";
+
+       if ((caller = proc_self()) != PROC_NULL) {
+               /* get process name */
+               strlcpy(command_name, caller->p_comm, sizeof(command_name));
+
+               /* get parent process name if possible */
+               if ((caller_parent = proc_find(caller->p_ppid)) != PROC_NULL) {
+                       strlcpy(parent_name, caller_parent->p_comm,
+                           sizeof(parent_name));
+                       proc_rele(caller_parent);
+               }
+
+               if ((escape_str(command_name, strlen(command_name),
+                   sizeof(command_name)) == 0) &&
+                   (escape_str(parent_name, strlen(parent_name),
+                   sizeof(parent_name)) == 0)) {
+                       kern_asl_msg(LOG_DEBUG, "messagetracer",
+                           5,
+                           "com.apple.message.domain",
+                           "com.apple.kernel.tcpstat", /* 1 */
+                           "com.apple.message.signature",
+                           "tcpinfo", /* 2 */
+                           "com.apple.message.signature2", command_name, /* 3 */
+                           "com.apple.message.signature3", parent_name, /* 4 */
+                           "com.apple.message.summarize", "YES", /* 5 */
+                           NULL);
+               }
+       }
+
+       if (caller != PROC_NULL)
+               proc_rele(caller);
+
+       if (req->newptr == USER_ADDR_NULL) {
+               return EINVAL;
+       }
+       if (req->newlen < sizeof(struct info_tuple)) {
+               return EINVAL;
+       }
+       error = SYSCTL_IN(req, &itpl, sizeof(struct info_tuple));
+       if (error != 0) {
+               return error;
+       }
+       error = tcp_fill_info_for_info_tuple(&itpl, &ti);
+       if (error != 0) {
+               return error;
+       }
+       error = SYSCTL_OUT(req, &ti, sizeof(struct tcp_info));
+       if (error != 0) {
+               return error;
+       }
+       
+       return 0;
+}
+
+static int
+tcp_lookup_peer_pid_locked(struct socket *so, pid_t *out_pid)
+{
+       int error = EHOSTUNREACH;
+       *out_pid = -1;
+       if ((so->so_state & SS_ISCONNECTED) == 0) return ENOTCONN;
+       
+       struct inpcb    *inp = (struct inpcb*)so->so_pcb;
+       uint16_t                lport = inp->inp_lport;
+       uint16_t                fport = inp->inp_fport;
+       struct inpcb    *finp = NULL;
+       
+       if (inp->inp_vflag & INP_IPV6) {
+               struct  in6_addr        laddr6 = inp->in6p_laddr;
+               struct  in6_addr        faddr6 = inp->in6p_faddr;
+               socket_unlock(so, 0);
+               finp = in6_pcblookup_hash(&tcbinfo, &laddr6, lport, &faddr6, fport, 0, NULL);
+               socket_lock(so, 0);
+       } else if (inp->inp_vflag & INP_IPV4) {
+               struct  in_addr laddr4 = inp->inp_laddr;
+               struct  in_addr faddr4 = inp->inp_faddr;
+               socket_unlock(so, 0);
+               finp = in_pcblookup_hash(&tcbinfo, laddr4, lport, faddr4, fport, 0, NULL);
+               socket_lock(so, 0);
+       }
+       
+       if (finp) {
+               *out_pid = finp->inp_socket->last_pid;
+               error = 0;
+               in_pcb_checkstate(finp, WNT_RELEASE, 0);
+       }
+       
+       return error;
+}
+
+void
+tcp_getconninfo(struct socket *so, struct conninfo_tcp *tcp_ci)
+{
+       (void) tcp_lookup_peer_pid_locked(so, &tcp_ci->tcpci_peer_pid);
+       tcp_fill_info(sototcpcb(so), &tcp_ci->tcpci_tcp_info);
+}
+
+/*
+ * The new sockopt interface makes it possible for us to block in the
+ * copyin/out step (if we take a page fault).  Taking a page fault at
+ * splnet() is probably a Bad Thing.  (Since sockets and pcbs both now
+ * use TSM, there probably isn't any need for this function to run at
+ * splnet() any more.  This needs more examination.)
+ */
+int
+tcp_ctloutput(struct socket *so, struct sockopt *sopt)
 {
        int     error, opt, optval;
        struct  inpcb *inp;
@@ -1108,9 +1889,12 @@ tcp_ctloutput(so, sopt)
        if (inp == NULL) {
                return (ECONNRESET);
        }
-       if (sopt->sopt_level != IPPROTO_TCP) {
+       /* Allow <SOL_SOCKET,SO_FLUSH/SO_TRAFFIC_MGT_BACKGROUND> at this level */
+       if (sopt->sopt_level != IPPROTO_TCP &&
+           !(sopt->sopt_level == SOL_SOCKET && (sopt->sopt_name == SO_FLUSH ||
+           sopt->sopt_name == SO_TRAFFIC_MGT_BACKGROUND))) {
 #if INET6
-               if (INP_CHECK_SOCKAF(so, AF_INET6))
+               if (SOCK_CHECK_DOM(so, PF_INET6))
                        error = ip6_ctloutput(so, sopt);
                else
 #endif /* INET6 */
@@ -1118,9 +1902,11 @@ tcp_ctloutput(so, sopt)
                return (error);
        }
        tp = intotcpcb(inp);
-        if (tp == NULL) {
-                return (ECONNRESET);
-        }
+       if (tp == NULL) {
+               return (ECONNRESET);
+       }
+
+       calculate_tcp_clock();
 
        switch (sopt->sopt_dir) {
        case SOPT_SET:
@@ -1153,7 +1939,80 @@ tcp_ctloutput(so, sopt)
                        else
                                tp->t_flags &= ~opt;
                        break;
+               case TCP_RXT_FINDROP:
+               case TCP_NOTIMEWAIT:
+                       error = sooptcopyin(sopt, &optval, sizeof optval,
+                               sizeof optval);
+                       if (error)
+                               break;
+                       switch (sopt->sopt_name) {
+                       case TCP_RXT_FINDROP:
+                               opt = TF_RXTFINDROP;
+                               break;
+                       case TCP_NOTIMEWAIT:
+                               opt = TF_NOTIMEWAIT;
+                               break;
+                       default:
+                               opt = 0;
+                               break;
+                       }
+                       if (optval)
+                               tp->t_flagsext |= opt;
+                       else
+                               tp->t_flagsext &= ~opt;
+                       break;
+               case TCP_MEASURE_SND_BW:
+                       error = sooptcopyin(sopt, &optval, sizeof optval,
+                               sizeof optval);
+                       if (error)
+                               break;
+                       opt = TF_MEASURESNDBW;
+                       if (optval) {
+                               if (tp->t_bwmeas == NULL) {
+                                       tp->t_bwmeas = tcp_bwmeas_alloc(tp);
+                                       if (tp->t_bwmeas == NULL) {
+                                               error = ENOMEM;
+                                               break;
+                                       }
+                               }
+                               tp->t_flagsext |= opt;
+                       } else {
+                               tp->t_flagsext &= ~opt;
+                               /* Reset snd bw measurement state */
+                               tp->t_flagsext &= ~(TF_BWMEAS_INPROGRESS);
+                               if (tp->t_bwmeas != NULL) {
+                                       tcp_bwmeas_free(tp);
+                               }
+                       }
+                       break;
+               case TCP_MEASURE_BW_BURST: {
+                       struct tcp_measure_bw_burst in;
+                       uint32_t minpkts, maxpkts;
+                       bzero(&in, sizeof(in));
 
+                       error = sooptcopyin(sopt, &in, sizeof(in),
+                               sizeof(in));
+                       if (error)
+                               break;
+                       if ((tp->t_flagsext & TF_MEASURESNDBW) == 0 ||
+                               tp->t_bwmeas == NULL) {
+                               error = EINVAL;
+                               break;
+                       }
+                       minpkts = (in.min_burst_size != 0) ? in.min_burst_size : 
+                               tp->t_bwmeas->bw_minsizepkts;
+                       maxpkts = (in.max_burst_size != 0) ? in.max_burst_size :
+                               tp->t_bwmeas->bw_maxsizepkts;
+                       if (minpkts > maxpkts) {
+                               error = EINVAL;
+                               break;
+                       }
+                       tp->t_bwmeas->bw_minsizepkts = minpkts;
+                       tp->t_bwmeas->bw_maxsizepkts = maxpkts;
+                       tp->t_bwmeas->bw_minsize = (minpkts * tp->t_maxseg);
+                       tp->t_bwmeas->bw_maxsize = (maxpkts * tp->t_maxseg);
+                       break;
+               }
                case TCP_MAXSEG:
                        error = sooptcopyin(sopt, &optval, sizeof optval,
                                            sizeof optval);
@@ -1172,14 +2031,323 @@ tcp_ctloutput(so, sopt)
                                                sizeof optval);
                        if (error)
                                break;
-                       if (optval < 0)
+                       if (optval < 0 || optval > UINT32_MAX/TCP_RETRANSHZ) {
                                error = EINVAL;
-                       else {
+                       else {
                                tp->t_keepidle = optval * TCP_RETRANSHZ;
-                               tp->t_timer[TCPT_KEEP] = TCP_KEEPIDLE(tp); /* reset the timer to new value */
+                               /* reset the timer to new value */
+                               tp->t_timer[TCPT_KEEP] = OFFSET_FROM_START(tp, 
+                                       TCP_CONN_KEEPIDLE(tp));
+                               tcp_check_timer_state(tp);
                        }
                         break;
-               
+
+               case TCP_CONNECTIONTIMEOUT:
+                       error = sooptcopyin(sopt, &optval, sizeof optval,
+                                               sizeof optval);
+                       if (error)
+                               break;
+                       if (optval < 0 || optval > UINT32_MAX/TCP_RETRANSHZ) {
+                               error = EINVAL;
+                       } else {
+                               tp->t_keepinit = optval * TCP_RETRANSHZ;
+                               if (tp->t_state == TCPS_SYN_RECEIVED ||
+                                       tp->t_state == TCPS_SYN_SENT) {
+                                       tp->t_timer[TCPT_KEEP] = OFFSET_FROM_START(tp,
+                                               TCP_CONN_KEEPINIT(tp));
+                                       tcp_check_timer_state(tp);
+                               }
+                       }
+                       break;
+
+               case TCP_KEEPINTVL:
+                       error = sooptcopyin(sopt, &optval, sizeof(optval),
+                               sizeof(optval));
+                       if (error)
+                               break;
+                       if (optval < 0 || optval > UINT32_MAX/TCP_RETRANSHZ) {
+                               error = EINVAL;
+                       } else {
+                               tp->t_keepintvl = optval * TCP_RETRANSHZ;
+                               if (tp->t_state == TCPS_FIN_WAIT_2 &&
+                                       TCP_CONN_MAXIDLE(tp) > 0) {
+                                       tp->t_timer[TCPT_2MSL] = OFFSET_FROM_START(tp,
+                                               TCP_CONN_MAXIDLE(tp));
+                                       tcp_check_timer_state(tp);
+                               }
+                       }
+                       break;
+
+               case TCP_KEEPCNT:
+                       error = sooptcopyin(sopt, &optval, sizeof(optval),
+                               sizeof(optval));
+                       if (error)
+                               break;
+                       if (optval < 0 || optval > INT32_MAX) {
+                               error = EINVAL;
+                       } else {
+                               tp->t_keepcnt = optval;
+                               if (tp->t_state == TCPS_FIN_WAIT_2 &&
+                                       TCP_CONN_MAXIDLE(tp) > 0) {
+                                       tp->t_timer[TCPT_2MSL] = OFFSET_FROM_START(tp,
+                                               TCP_CONN_MAXIDLE(tp));
+                                       tcp_check_timer_state(tp);
+                               }
+                       }
+                       break;
+
+               case TCP_KEEPALIVE_OFFLOAD:
+                       error = sooptcopyin(sopt, &optval, sizeof(optval),
+                               sizeof(optval));
+                       if (error)
+                               break;
+                       if (optval < 0 || optval > INT32_MAX) {
+                               error = EINVAL;
+                               break;
+                       } 
+                       if (optval != 0)
+                               inp->inp_flags2 |= INP2_KEEPALIVE_OFFLOAD;
+                       else
+                               inp->inp_flags2 &= ~INP2_KEEPALIVE_OFFLOAD;
+                       break;
+
+               case PERSIST_TIMEOUT:
+                       error = sooptcopyin(sopt, &optval, sizeof optval,
+                                               sizeof optval);
+                       if (error)
+                               break;
+                       if (optval < 0)
+                               error = EINVAL;
+                       else 
+                               tp->t_persist_timeout = optval * TCP_RETRANSHZ;
+                       break;
+               case TCP_RXT_CONNDROPTIME:
+                       error = sooptcopyin(sopt, &optval, sizeof(optval),
+                                       sizeof(optval));
+                       if (error)
+                               break;
+                       if (optval < 0)
+                               error = EINVAL;
+                       else
+                               tp->t_rxt_conndroptime = optval * TCP_RETRANSHZ;
+                       break;
+               case TCP_NOTSENT_LOWAT:
+                       error = sooptcopyin(sopt, &optval, sizeof(optval),
+                               sizeof(optval));
+                       if (error)
+                               break;
+                       if (optval < 0) {
+                               error = EINVAL;
+                               break;
+                       } else {
+                               if (optval == 0) {
+                                       so->so_flags &= ~(SOF_NOTSENT_LOWAT);
+                                       tp->t_notsent_lowat = 0;
+                               } else { 
+                                       so->so_flags |= SOF_NOTSENT_LOWAT;
+                                       tp->t_notsent_lowat = optval;
+                               }
+                       }
+                       break;
+               case TCP_ADAPTIVE_READ_TIMEOUT:
+                       error = sooptcopyin(sopt, &optval, sizeof (optval),
+                           sizeof(optval));
+                       if (error)
+                               break;
+                       if (optval < 0 || 
+                           optval > TCP_ADAPTIVE_TIMEOUT_MAX) {
+                               error = EINVAL;
+                               break;
+                       } else if (optval == 0) {
+                               tp->t_adaptive_rtimo = 0;
+                               tcp_keepalive_reset(tp);
+                       } else {
+                               tp->t_adaptive_rtimo = optval;
+                       }
+                       break;
+               case TCP_ADAPTIVE_WRITE_TIMEOUT:
+                       error = sooptcopyin(sopt, &optval, sizeof (optval),
+                           sizeof (optval));
+                       if (error)
+                               break;
+                       if (optval < 0 || 
+                           optval > TCP_ADAPTIVE_TIMEOUT_MAX) {
+                               error = EINVAL;
+                               break;
+                       } else {
+                               tp->t_adaptive_wtimo = optval;
+                       }
+                       break;
+               case TCP_ENABLE_MSGS:
+                       error = sooptcopyin(sopt, &optval, sizeof(optval),
+                               sizeof(optval));
+                       if (error)
+                               break;
+                       if (optval < 0 || optval > 1) {
+                               error = EINVAL;
+                       } else if (optval == 1) {
+                               /*
+                                * Check if messages option is already
+                                * enabled, if so return.
+                                */
+                               if (so->so_flags & SOF_ENABLE_MSGS) {
+                                       VERIFY(so->so_msg_state != NULL);
+                                       break;
+                               }
+
+                               /*
+                                * allocate memory for storing message
+                                * related state
+                                */
+                               VERIFY(so->so_msg_state == NULL);
+                               MALLOC(so->so_msg_state,
+                                       struct msg_state *,
+                                       sizeof(struct msg_state),
+                                       M_TEMP, M_WAITOK | M_ZERO);
+                               if (so->so_msg_state == NULL) {
+                                       error = ENOMEM;
+                                       break;
+                               }
+
+                               /* Enable message delivery */
+                               so->so_flags |= SOF_ENABLE_MSGS;
+                       } else {
+                               /*
+                                * Can't disable message delivery on socket
+                                * because of restrictions imposed by
+                                * encoding/decoding
+                                */
+                               error = EINVAL;
+                       }
+                       break;
+               case TCP_SENDMOREACKS:
+                       error = sooptcopyin(sopt, &optval, sizeof(optval),
+                               sizeof(optval));
+                       if (error)
+                               break;
+                       if (optval < 0 || optval > 1) {
+                               error = EINVAL;
+                       } else if (optval == 0) {
+                               tp->t_flagsext &= ~(TF_NOSTRETCHACK);
+                       } else {
+                               tp->t_flagsext |= TF_NOSTRETCHACK;
+                       }
+                       break;
+               case TCP_DISABLE_BLACKHOLE_DETECTION:
+                       error = sooptcopyin(sopt, &optval, sizeof(optval),
+                               sizeof(optval));
+                       if (error)
+                               break;
+                       if (optval < 0 || optval > 1) {
+                               error = EINVAL;
+                       } else if (optval == 0) {
+                               tp->t_flagsext &= ~TF_NOBLACKHOLE_DETECTION;
+                       } else {
+                               tp->t_flagsext |= TF_NOBLACKHOLE_DETECTION;
+                               if ((tp->t_flags & TF_BLACKHOLE) &&
+                                   tp->t_pmtud_saved_maxopd > 0)
+                                       tcp_pmtud_revert_segment_size(tp);
+                       }
+                       break;
+               case TCP_FASTOPEN:
+                       if (!(tcp_fastopen & TCP_FASTOPEN_SERVER)) {
+                               error = ENOTSUP;
+                               break;
+                       }
+
+                       error = sooptcopyin(sopt, &optval, sizeof(optval),
+                               sizeof(optval));
+                       if (error)
+                               break;
+                       if (optval < 0 || optval > 1) {
+                               error = EINVAL;
+                               break;
+                       }
+                       if (tp->t_state != TCPS_LISTEN) {
+                               error =  EINVAL;
+                               break;
+                       }
+                       if (optval)
+                               tp->t_flagsext |= TF_FASTOPEN;
+                       else
+                               tcp_disable_tfo(tp);
+                       break;
+               case TCP_ENABLE_ECN:
+                       error = sooptcopyin(sopt, &optval, sizeof optval,
+                                           sizeof optval);
+                       if (error)
+                               break;
+                       if (optval) {
+                               tp->ecn_flags |= TE_ECN_MODE_ENABLE;
+                               tp->ecn_flags &= ~TE_ECN_MODE_DISABLE;
+                       } else {
+                               tp->ecn_flags &= ~TE_ECN_MODE_ENABLE;
+                       }
+                       break;
+               case TCP_ECN_MODE:
+                       error = sooptcopyin(sopt, &optval, sizeof optval,
+                                           sizeof optval);
+                       if (error)
+                               break;
+                       if (optval == ECN_MODE_DEFAULT) {
+                               tp->ecn_flags &= ~TE_ECN_MODE_ENABLE;
+                               tp->ecn_flags &= ~TE_ECN_MODE_DISABLE;
+                       } else if (optval == ECN_MODE_ENABLE) {
+                               tp->ecn_flags |= TE_ECN_MODE_ENABLE;
+                               tp->ecn_flags &= ~TE_ECN_MODE_DISABLE;
+                       } else if (optval == ECN_MODE_DISABLE) {
+                               tp->ecn_flags &= ~TE_ECN_MODE_ENABLE;
+                               tp->ecn_flags |= TE_ECN_MODE_DISABLE;
+                       } else {
+                               error = EINVAL;
+                       }
+                       break;
+               case TCP_NOTIFY_ACKNOWLEDGEMENT:
+                       error = sooptcopyin(sopt, &optval,
+                           sizeof(optval), sizeof(optval));
+                       if (error)
+                               break;
+                       if (optval <= 0) {
+                               error = EINVAL;
+                               break;
+                       }
+                       if (tp->t_notify_ack_count >= TCP_MAX_NOTIFY_ACK) {
+                               error = ETOOMANYREFS;
+                               break;
+                       }
+
+                       /*
+                        * validate that the given marker id is not
+                        * a duplicate to avoid ambiguity
+                        */
+                       if ((error = tcp_notify_ack_id_valid(tp, so,
+                           optval)) != 0) {
+                               break;
+                       }
+                       error = tcp_add_notify_ack_marker(tp, optval);
+                       break;
+               case SO_FLUSH:
+                       if ((error = sooptcopyin(sopt, &optval, sizeof (optval),
+                           sizeof (optval))) != 0)
+                               break;
+
+                       error = inp_flush(inp, optval);
+                       break;
+
+               case SO_TRAFFIC_MGT_BACKGROUND:
+                       if ((error = sooptcopyin(sopt, &optval, sizeof (optval),
+                           sizeof (optval))) != 0)
+                               break;
+
+                       if (optval) {
+                               socket_set_traffic_mgt_flags_locked(so,
+                                   TRAFFIC_MGT_SO_BACKGROUND);
+                       } else {
+                               socket_clear_traffic_mgt_flags_locked(so,
+                                   TRAFFIC_MGT_SO_BACKGROUND);
+                       }
+                       break;
+
                default:
                        error = ENOPROTOOPT;
                        break;
@@ -1195,7 +2363,25 @@ tcp_ctloutput(so, sopt)
                        optval = tp->t_maxseg;
                        break;
                case TCP_KEEPALIVE:
-                       optval = tp->t_keepidle / TCP_RETRANSHZ;
+                       if (tp->t_keepidle > 0)
+                               optval = tp->t_keepidle / TCP_RETRANSHZ;
+                       else
+                               optval = tcp_keepidle  / TCP_RETRANSHZ;
+                       break;
+               case TCP_KEEPINTVL:
+                       if (tp->t_keepintvl > 0)
+                               optval = tp->t_keepintvl / TCP_RETRANSHZ;
+                       else
+                               optval = tcp_keepintvl / TCP_RETRANSHZ;
+                       break;
+               case TCP_KEEPCNT:
+                       if (tp->t_keepcnt > 0)
+                               optval = tp->t_keepcnt;
+                       else
+                               optval = tcp_keepcnt;
+                       break;
+               case TCP_KEEPALIVE_OFFLOAD:
+                       optval = !!(inp->inp_flags2 & INP2_KEEPALIVE_OFFLOAD);
                        break;
                case TCP_NOOPT:
                        optval = tp->t_flags & TF_NOOPT;
@@ -1203,6 +2389,129 @@ tcp_ctloutput(so, sopt)
                case TCP_NOPUSH:
                        optval = tp->t_flags & TF_NOPUSH;
                        break;
+               case TCP_ENABLE_ECN:
+                       optval = (tp->ecn_flags & TE_ECN_MODE_ENABLE) ? 1 : 0;
+                       break;
+               case TCP_ECN_MODE:
+                       if (tp->ecn_flags & TE_ECN_MODE_ENABLE)
+                               optval = ECN_MODE_ENABLE;
+                       else if (tp->ecn_flags & TE_ECN_MODE_DISABLE)
+                               optval = ECN_MODE_DISABLE;
+                       else
+                               optval = ECN_MODE_DEFAULT;
+                       break;
+               case TCP_CONNECTIONTIMEOUT:
+                       optval = tp->t_keepinit / TCP_RETRANSHZ;
+                       break;
+               case PERSIST_TIMEOUT:
+                       optval = tp->t_persist_timeout / TCP_RETRANSHZ;
+                       break;
+               case TCP_RXT_CONNDROPTIME:
+                       optval = tp->t_rxt_conndroptime / TCP_RETRANSHZ;
+                       break;
+               case TCP_RXT_FINDROP:
+                       optval = tp->t_flagsext & TF_RXTFINDROP;
+                       break; 
+               case TCP_NOTIMEWAIT:
+                       optval = (tp->t_flagsext & TF_NOTIMEWAIT) ? 1 : 0;
+                       break;
+               case TCP_FASTOPEN:
+                       if (tp->t_state != TCPS_LISTEN ||
+                           !(tcp_fastopen & TCP_FASTOPEN_SERVER)) {
+                               error = ENOTSUP;
+                               break;
+                       }
+                       optval = tfo_enabled(tp);
+                       break;
+               case TCP_MEASURE_SND_BW:
+                       optval = tp->t_flagsext & TF_MEASURESNDBW;
+                       break;
+               case TCP_INFO: {
+                       struct tcp_info ti;
+
+                       tcp_fill_info(tp, &ti);
+                       error = sooptcopyout(sopt, &ti, sizeof(struct tcp_info));
+                       goto done;
+                       /* NOT REACHED */
+               }
+               case TCP_CONNECTION_INFO: {
+                       struct tcp_connection_info tci;
+                       tcp_connection_fill_info(tp, &tci);
+                       error = sooptcopyout(sopt, &tci,
+                           sizeof(struct tcp_connection_info));
+                       goto done;
+               }
+               case TCP_MEASURE_BW_BURST: {
+                       struct tcp_measure_bw_burst out;
+                       if ((tp->t_flagsext & TF_MEASURESNDBW) == 0 ||
+                               tp->t_bwmeas == NULL) {
+                               error = EINVAL;
+                               break;
+                       }
+                       out.min_burst_size = tp->t_bwmeas->bw_minsizepkts;
+                       out.max_burst_size = tp->t_bwmeas->bw_maxsizepkts;
+                       error = sooptcopyout(sopt, &out, sizeof(out));
+                       goto done;
+               }
+               case TCP_NOTSENT_LOWAT:
+                       if ((so->so_flags & SOF_NOTSENT_LOWAT) != 0) {
+                               optval = tp->t_notsent_lowat;
+                       } else {
+                               optval = 0;
+                       }
+                       break;
+
+               case TCP_ENABLE_MSGS:
+                       if (so->so_flags & SOF_ENABLE_MSGS) {
+                               optval = 1;
+                       } else {
+                               optval = 0;
+                       }
+                       break;
+               case TCP_SENDMOREACKS:
+                       if (tp->t_flagsext & TF_NOSTRETCHACK)
+                               optval = 1;
+                       else
+                               optval = 0;
+                       break;
+               case TCP_DISABLE_BLACKHOLE_DETECTION:
+                       if (tp->t_flagsext & TF_NOBLACKHOLE_DETECTION)
+                               optval = 1;
+                       else
+                               optval = 0;
+                       break;
+               case TCP_PEER_PID: {
+                       pid_t   pid;
+                       error = tcp_lookup_peer_pid_locked(so, &pid);
+                       if (error == 0)
+                               error = sooptcopyout(sopt, &pid, sizeof(pid));
+                       goto done;
+               }
+               case TCP_ADAPTIVE_READ_TIMEOUT:
+                       optval = tp->t_adaptive_rtimo;
+                       break;
+               case TCP_ADAPTIVE_WRITE_TIMEOUT:
+                       optval = tp->t_adaptive_wtimo;
+                       break;
+               case SO_TRAFFIC_MGT_BACKGROUND:
+                       optval = (so->so_flags1 &
+                           SOF1_TRAFFIC_MGT_SO_BACKGROUND) ? 1 : 0;
+                       break;
+               case TCP_NOTIFY_ACKNOWLEDGEMENT: {
+                       struct tcp_notify_ack_complete retid;
+
+                       if (sopt->sopt_valsize != sizeof (retid)) {
+                               error = EINVAL;
+                               break;
+                       }
+                       bzero(&retid, sizeof (retid));
+                       tcp_get_notify_ack_count(tp, &retid);
+                       if (retid.notify_complete_count > 0)
+                               tcp_get_notify_ack_ids(tp, &retid);
+
+                       error = sooptcopyout(sopt, &retid, sizeof (retid));
+                       goto done;
+               }
                default:
                        error = ENOPROTOOPT;
                        break;
@@ -1211,6 +2520,7 @@ tcp_ctloutput(so, sopt)
                        error = sooptcopyout(sopt, &optval, sizeof optval);
                break;
        }
+done:
        return (error);
 }
 
@@ -1219,13 +2529,47 @@ tcp_ctloutput(so, sopt)
  * sizes, respectively.  These are obsolescent (this information should
  * be set by the route).
  */
-u_long tcp_sendspace = 1448*256;
-SYSCTL_INT(_net_inet_tcp, TCPCTL_SENDSPACE, sendspace, CTLFLAG_RW, 
-    &tcp_sendspace , 0, "Maximum outgoing TCP datagram size");
-u_long tcp_recvspace = 1448*384;
-SYSCTL_INT(_net_inet_tcp, TCPCTL_RECVSPACE, recvspace, CTLFLAG_RW, 
-    &tcp_recvspace , 0, "Maximum incoming TCP datagram size");
+u_int32_t      tcp_sendspace = 1448*256;
+u_int32_t      tcp_recvspace = 1448*384;
+
+/* During attach, the size of socket buffer allocated is limited to
+ * sb_max in sbreserve. Disallow setting the tcp send and recv space
+ * to be more than sb_max because that will cause tcp_attach to fail
+ * (see radar 5713060)
+ */  
+static int
+sysctl_tcp_sospace(struct sysctl_oid *oidp, __unused void *arg1,
+       __unused int arg2, struct sysctl_req *req) {
+       u_int32_t new_value = 0, *space_p = NULL;
+       int changed = 0, error = 0;
+       u_quad_t sb_effective_max = (sb_max / (MSIZE+MCLBYTES)) * MCLBYTES;
+
+       switch (oidp->oid_number) {
+               case TCPCTL_SENDSPACE:
+                       space_p = &tcp_sendspace;
+                       break;
+               case TCPCTL_RECVSPACE:
+                       space_p = &tcp_recvspace;
+                       break;
+               default:
+                       return EINVAL;
+       }
+       error = sysctl_io_number(req, *space_p, sizeof(u_int32_t),
+               &new_value, &changed);
+       if (changed) {
+               if (new_value > 0 && new_value <= sb_effective_max) {
+                       *space_p = new_value;
+               } else {
+                       error = ERANGE;
+               }
+       }
+       return error;
+}
 
+SYSCTL_PROC(_net_inet_tcp, TCPCTL_SENDSPACE, sendspace, CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED,
+    &tcp_sendspace , 0, &sysctl_tcp_sospace, "IU", "Maximum outgoing TCP datagram size");
+SYSCTL_PROC(_net_inet_tcp, TCPCTL_RECVSPACE, recvspace, CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED,
+    &tcp_recvspace , 0, &sysctl_tcp_sospace, "IU", "Maximum incoming TCP datagram size");
 
 /*
  * Attach TCP protocol to socket, allocating
@@ -1239,16 +2583,13 @@ SYSCTL_INT(_net_inet_tcp, TCPCTL_RECVSPACE, recvspace, CTLFLAG_RW,
  *     soreserve:ENOBUFS
  */
 static int
-tcp_attach(so, p)
-       struct socket *so;
-       struct proc *p;
+tcp_attach(struct socket *so, struct proc *p)
 {
-       register struct tcpcb *tp;
+       struct tcpcb *tp;
        struct inpcb *inp;
        int error;
-       u_long sb_effective_max;
 #if INET6
-       int isipv6 = INP_CHECK_SOCKAF(so, AF_INET6) != 0;
+       int isipv6 = SOCK_CHECK_DOM(so, PF_INET6) != 0;
 #endif
 
        error = in_pcballoc(so, &tcbinfo, p);
@@ -1258,28 +2599,20 @@ tcp_attach(so, p)
        inp = sotoinpcb(so);
 
        if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) {
-               /*
-                * The goal is to let clients machines use large send/rcv default windows to compensate for link
-                * latency and make sure the receiver is not constraining the sender window.
-                * But we doon't want to have a few connections use all our mbuf space for servers.
-                * This is done by watching a threshold of tcpcbs in use and bumping the default send and rcvspace
-                * only if that threshold isn't reached.
-                * We're also advertising a much bigger window size (tuneable by sysctl) in correlation with                             * the max socket buffer size if 
-                * we consider that we have enough ressources for it. This window will be adjusted depending on the
-                * global socket layer buffer use with the use of tcp_sbpace
-                */
-
-               if (inp->inp_pcbinfo->ipi_count < tcp_sockthreshold) {
-                       sb_effective_max = (sb_max / (MSIZE+MCLBYTES)) * MCLBYTES;  
-                       error = soreserve(so, max(min((TCP_MAXWIN << tcp_win_scale)/4, sb_effective_max), tcp_sendspace),
-                                       max(min((TCP_MAXWIN << tcp_win_scale)/2, sb_effective_max), tcp_recvspace));
-               }
-               else    
-                       error = soreserve(so, tcp_sendspace, tcp_recvspace);
+               error = soreserve(so, tcp_sendspace, tcp_recvspace);
                if (error)
                        return (error);
        }
 
+       if (so->so_snd.sb_preconn_hiwat == 0) {
+               soreserve_preconnect(so, 2048);
+       }
+
+       if ((so->so_rcv.sb_flags & SB_USRSIZE) == 0)
+               so->so_rcv.sb_flags |= SB_AUTOSIZE;
+       if ((so->so_snd.sb_flags & SB_USRSIZE) == 0)
+               so->so_snd.sb_flags |= SB_AUTOSIZE;
+
 #if INET6
        if (isipv6) {
                inp->inp_vflag |= INP_IPV6;
@@ -1289,7 +2622,7 @@ tcp_attach(so, p)
 #endif /* INET6 */
        inp->inp_vflag |= INP_IPV4;
        tp = tcp_newtcpcb(inp);
-       if (tp == 0) {
+       if (tp == NULL) {
                int nofd = so->so_state & SS_NOFDREF;   /* XXX */
 
                so->so_state &= ~SS_NOFDREF;    /* don't free the socket yet */
@@ -1302,6 +2635,8 @@ tcp_attach(so, p)
                so->so_state |= nofd;
                return (ENOBUFS);
        }
+       if (nstat_collect)
+               nstat_tcp_new_pcb(inp);
        tp->t_state = TCPS_CLOSED;
        return (0);
 }
@@ -1315,8 +2650,7 @@ tcp_attach(so, p)
  * send segment to peer (with FIN).
  */
 static struct tcpcb *
-tcp_disconnect(tp)
-       register struct tcpcb *tp;
+tcp_disconnect(struct tcpcb *tp)
 {
        struct socket *so = tp->t_inpcb->inp_socket;
 
@@ -1328,6 +2662,12 @@ tcp_disconnect(tp)
                soisdisconnecting(so);
                sbflush(&so->so_rcv);
                tp = tcp_usrclosed(tp);
+#if MPTCP
+               /* A reset has been sent but socket exists, do not send FIN */
+               if ((so->so_flags & SOF_MP_SUBFLOW) &&
+                   (tp) && (tp->t_mpflags & TMPF_RESET))
+                       return (tp);
+#endif
                if (tp)
                        (void) tcp_output(tp);
        }
@@ -1345,15 +2685,12 @@ tcp_disconnect(tp)
  * We can let the user exit from the close as soon as the FIN is acked.
  */
 static struct tcpcb *
-tcp_usrclosed(tp)
-       register struct tcpcb *tp;
+tcp_usrclosed(struct tcpcb *tp)
 {
-
        switch (tp->t_state) {
 
        case TCPS_CLOSED:
        case TCPS_LISTEN:
-               tp->t_state = TCPS_CLOSED;
                tp = tcp_close(tp);
                break;
 
@@ -1363,10 +2700,18 @@ tcp_usrclosed(tp)
                break;
 
        case TCPS_ESTABLISHED:
+               DTRACE_TCP4(state__change, void, NULL, 
+                       struct inpcb *, tp->t_inpcb,
+                       struct tcpcb *, tp, 
+                       int32_t, TCPS_FIN_WAIT_1);
                tp->t_state = TCPS_FIN_WAIT_1;
                break;
 
        case TCPS_CLOSE_WAIT:
+               DTRACE_TCP4(state__change, void, NULL, 
+                       struct inpcb *, tp->t_inpcb,
+                       struct tcpcb *, tp, 
+                       int32_t, TCPS_LAST_ACK);
                tp->t_state = TCPS_LAST_ACK;
                break;
        }
@@ -1374,7 +2719,8 @@ tcp_usrclosed(tp)
                soisdisconnected(tp->t_inpcb->inp_socket);
                /* To prevent the connection hanging in FIN_WAIT_2 forever. */
                if (tp->t_state == TCPS_FIN_WAIT_2)
-                       tp->t_timer[TCPT_2MSL] = tcp_maxidle;
+                       tp->t_timer[TCPT_2MSL] = OFFSET_FROM_START(tp, 
+                               TCP_CONN_MAXIDLE(tp));
        }
        return (tp);
 }
@@ -1382,13 +2728,57 @@ tcp_usrclosed(tp)
 void
 tcp_in_cksum_stats(u_int32_t len)
 {
-       tcps_in_sw_cksum++;
-       tcps_in_sw_cksum_bytes += len;
+       tcpstat.tcps_rcv_swcsum++;
+       tcpstat.tcps_rcv_swcsum_bytes += len;
 }
 
 void
 tcp_out_cksum_stats(u_int32_t len)
 {
-       tcps_out_sw_cksum++;
-       tcps_out_sw_cksum_bytes += len;
+       tcpstat.tcps_snd_swcsum++;
+       tcpstat.tcps_snd_swcsum_bytes += len;
 }
+
+#if INET6
+void
+tcp_in6_cksum_stats(u_int32_t len)
+{
+       tcpstat.tcps_rcv6_swcsum++;
+       tcpstat.tcps_rcv6_swcsum_bytes += len;
+}
+
+void
+tcp_out6_cksum_stats(u_int32_t len)
+{
+       tcpstat.tcps_snd6_swcsum++;
+       tcpstat.tcps_snd6_swcsum_bytes += len;
+}
+
+/* 
+ * When messages are enabled on a TCP socket, the message priority
+ * is sent as a control message. This function will extract it.
+ */
+int
+tcp_get_msg_priority(struct mbuf *control, uint32_t *msgpri) 
+{
+       struct cmsghdr *cm;
+       if (control == NULL)
+               return(EINVAL);
+
+       for (cm = M_FIRST_CMSGHDR(control); cm;
+               cm = M_NXT_CMSGHDR(control, cm)) {
+               if (cm->cmsg_len < sizeof(struct cmsghdr) ||
+                       cm->cmsg_len > control->m_len) {
+                       return (EINVAL);
+               }
+               if (cm->cmsg_level == SOL_SOCKET &&
+                       cm->cmsg_type == SCM_MSG_PRIORITY) {
+                       *msgpri = *(unsigned int *)(void *)CMSG_DATA(cm);
+                       break;
+               }
+       }
+
+       VERIFY(*msgpri >= MSG_PRI_MIN && *msgpri <= MSG_PRI_MAX); 
+       return (0);
+}
+#endif /* INET6 */