]> git.saurik.com Git - apple/xnu.git/blobdiff - bsd/netinet/raw_ip.c
xnu-792.18.15.tar.gz
[apple/xnu.git] / bsd / netinet / raw_ip.c
index 60f2ce01bc2d8743689ad7b8c8c8a3985290769a..6219b925037d94d7a3b791a944cfefcbe625830f 100644 (file)
@@ -1,23 +1,29 @@
 /*
  * Copyright (c) 2000 Apple Computer, Inc. All rights reserved.
  *
- * @APPLE_LICENSE_HEADER_START@
+ * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
  * 
- * The contents of this file constitute Original Code as defined in and
- * are subject to the Apple Public Source License Version 1.1 (the
- * "License").  You may not use this file except in compliance with the
- * License.  Please obtain a copy of the License at
- * http://www.apple.com/publicsource and read it before using this file.
+ * 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
+ * compliance with the License. The rights granted to you under the License
+ * may not be used to create, or enable the creation or redistribution of,
+ * 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.
  * 
- * This Original Code and all software distributed under the License are
- * distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY KIND, EITHER
+ * 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,
  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE OR NON-INFRINGEMENT.  Please see the
- * License for the specific language governing rights and limitations
- * under the License.
+ * 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_LICENSE_HEADER_END@
+ * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
  */
 /*
  * Copyright (c) 1982, 1986, 1988, 1993
@@ -60,6 +66,7 @@
 #include <sys/malloc.h>
 #include <sys/mbuf.h>
 #include <sys/proc.h>
+#include <sys/domain.h>
 #include <sys/protosw.h>
 #include <sys/socket.h>
 #include <sys/socketvar.h>
 
 #if IPSEC
 extern int ipsec_bypass;
+extern lck_mtx_t *sadb_mutex;
 #endif
 
+extern u_long  route_generation;
 struct inpcbhead ripcb;
 struct inpcbinfo ripcbinfo;
 
+/* control hooks for ipfw and dummynet */
+ip_fw_ctl_t *ip_fw_ctl_ptr;
+#if DUMMYNET
+ip_dn_ctl_t *ip_dn_ctl_ptr;
+#endif /* DUMMYNET */
+
 /*
  * Nominal space allocated to a raw ip socket.
  */
@@ -114,6 +129,8 @@ struct      inpcbinfo ripcbinfo;
 void
 rip_init()
 {
+       struct inpcbinfo *pcbinfo;
+
        LIST_INIT(&ripcb);
        ripcbinfo.listhead = &ripcb;
        /*
@@ -128,6 +145,22 @@ rip_init()
                                            (4096 * sizeof(struct inpcb)), 
                                            4096, "ripzone");
 
+       pcbinfo = &ripcbinfo;
+        /*
+        * allocate lock group attribute and group for udp pcb mutexes
+        */
+       pcbinfo->mtx_grp_attr = lck_grp_attr_alloc_init();
+
+       pcbinfo->mtx_grp = lck_grp_alloc_init("ripcb", pcbinfo->mtx_grp_attr);
+               
+       /*
+        * allocate the lock attribute for udp pcb mutexes
+        */
+       pcbinfo->mtx_attr = lck_attr_alloc_init();
+
+       if ((pcbinfo->mtx = lck_rw_alloc_init(pcbinfo->mtx_grp, pcbinfo->mtx_attr)) == NULL)
+               return; /* pretty much dead if this fails... */
+
 }
 
 static struct  sockaddr_in ripsrc = { sizeof(ripsrc), AF_INET };
@@ -145,8 +178,10 @@ rip_input(m, iphlen)
        register struct inpcb *inp;
        struct inpcb *last = 0;
        struct mbuf *opts = 0;
+       int skipit;
 
        ripsrc.sin_addr = ip->ip_src;
+       lck_rw_lock_shared(ripcbinfo.mtx);
        LIST_FOREACH(inp, &ripcb, inp_list) {
 #if INET6
                if ((inp->inp_vflag & INP_IPV4) == 0)
@@ -162,16 +197,23 @@ rip_input(m, iphlen)
                        continue;
                if (last) {
                        struct mbuf *n = m_copy(m, 0, (int)M_COPYALL);
-
+               
 #if IPSEC
                        /* check AH/ESP integrity. */
-                       if (ipsec_bypass == 0 && n && ipsec4_in_reject_so(n, last->inp_socket)) {
-                               m_freem(n);
-                               ipsecstat.in_polvio++;
-                               /* do not inject data to pcb */
-                       } else
+                       skipit = 0;
+                       if (ipsec_bypass == 0 && n) {
+                               lck_mtx_lock(sadb_mutex);
+                               if (ipsec4_in_reject_so(n, last->inp_socket)) {
+                                       m_freem(n);
+                                       ipsecstat.in_polvio++;
+                                       /* do not inject data to pcb */
+                                       skipit = 1;
+                               }
+                               lck_mtx_unlock(sadb_mutex);
+                       } 
 #endif /*IPSEC*/
-                       if (n) {
+                       if (n && skipit == 0) {
+                               int error = 0;
                                if (last->inp_flags & INP_CONTROLOPTS ||
                                    last->inp_socket->so_options & SO_TIMESTAMP)
                                    ip_savecontrol(last, &opts, ip, n);
@@ -180,51 +222,60 @@ rip_input(m, iphlen)
                                        n->m_pkthdr.len -= iphlen;
                                        n->m_data += iphlen;
                                }
+// ###LOCK need to lock that socket?
                                if (sbappendaddr(&last->inp_socket->so_rcv,
                                    (struct sockaddr *)&ripsrc, n,
-                                   opts) == 0) {
-                                       /* should notify about lost packet */
-                                   kprintf("rip_input can't append to socket\n");
-                                       m_freem(n);
-                                       if (opts)
-                                           m_freem(opts);
-                               } else
+                                   opts, &error) != 0) {
                                        sorwakeup(last->inp_socket);
+                               }
+                               else {
+                                       if (error) {
+                                               /* should notify about lost packet */
+                                               kprintf("rip_input can't append to socket\n");
+                                       }
+                               }
                                opts = 0;
                        }
                }
                last = inp;
        }
+       lck_rw_done(ripcbinfo.mtx);
 #if IPSEC
        /* check AH/ESP integrity. */
-       if (ipsec_bypass == 0 && last && ipsec4_in_reject_so(m, last->inp_socket)) {
-               m_freem(m);
-               ipsecstat.in_polvio++;
-               ipstat.ips_delivered--;
-               /* do not inject data to pcb */
-       } else
+       skipit = 0;
+       if (ipsec_bypass == 0 && last) {
+               lck_mtx_lock(sadb_mutex);
+               if (ipsec4_in_reject_so(m, last->inp_socket)) {
+                       m_freem(m);
+                       ipsecstat.in_polvio++;
+                       ipstat.ips_delivered--;
+                       /* do not inject data to pcb */
+                       skipit = 1;
+               }
+               lck_mtx_unlock(sadb_mutex);
+       } 
 #endif /*IPSEC*/
-       if (last) {
-               if (last->inp_flags & INP_CONTROLOPTS ||
-                   last->inp_socket->so_options & SO_TIMESTAMP)
-                       ip_savecontrol(last, &opts, ip, m);
-        if (last->inp_flags & INP_STRIPHDR) {
-            m->m_len -= iphlen;
-            m->m_pkthdr.len -= iphlen;
-            m->m_data += iphlen;
-        }
-               if (sbappendaddr(&last->inp_socket->so_rcv,
-                   (struct sockaddr *)&ripsrc, m, opts) == 0) {
-                   kprintf("rip_input(2) can't append to socket\n");
+       if (skipit == 0) {
+               if (last) {
+                       if (last->inp_flags & INP_CONTROLOPTS ||
+                               last->inp_socket->so_options & SO_TIMESTAMP)
+                               ip_savecontrol(last, &opts, ip, m);
+                       if (last->inp_flags & INP_STRIPHDR) {
+                               m->m_len -= iphlen;
+                               m->m_pkthdr.len -= iphlen;
+                               m->m_data += iphlen;
+                       }
+                       if (sbappendaddr(&last->inp_socket->so_rcv,
+                               (struct sockaddr *)&ripsrc, m, opts, NULL) != 0) {
+                               sorwakeup(last->inp_socket);
+                       } else {
+                               kprintf("rip_input(2) can't append to socket\n");
+                       }
+               } else {
                        m_freem(m);
-                       if (opts)
-                           m_freem(opts);
-               } else
-                       sorwakeup(last->inp_socket);
-       } else {
-               m_freem(m);
-               ipstat.ips_noproto++;
-               ipstat.ips_delivered--;
+                       ipstat.ips_noproto++;
+                       ipstat.ips_delivered--;
+               }
        }
 }
 
@@ -293,23 +344,27 @@ rip_output(m, so, dst)
        }
 #endif /*IPSEC*/
 
-       return (ip_output(m, inp->inp_options, &inp->inp_route, flags,
+       if (inp->inp_route.ro_rt && inp->inp_route.ro_rt->generation_id != route_generation) {
+               rtfree(inp->inp_route.ro_rt);
+               inp->inp_route.ro_rt = (struct rtentry *)0;
+       }
+
+       return (ip_output_list(m, 0, inp->inp_options, &inp->inp_route, flags,
                          inp->inp_moptions));
 }
 
-int
+extern int
 load_ipfw()
 {
        kern_return_t   err;
        
-       /* Load the kext by the identifier */
-       err = kmod_load_extension("com.apple.nke.IPFirewall");
-       if (err) return err;
+       ipfw_init();
        
-       if (ip_fw_ctl_ptr == NULL) {
-               /* Wait for the kext to finish loading */
-               err = tsleep(&ip_fw_ctl_ptr, PWAIT | PCATCH, "load_ipfw_kext", 5 * 60 /* 5 seconds */);
-       }
+#if DUMMYNET
+       if (!DUMMYNET_LOADED)
+               ip_dn_init();
+#endif /* DUMMYNET */
+       err = 0;
        
        return err == 0 && ip_fw_ctl_ptr == NULL ? -1 : err;
 }
@@ -357,10 +412,10 @@ rip_ctloutput(so, sopt)
 
 #if DUMMYNET
                case IP_DUMMYNET_GET:
-                       if (ip_dn_ctl_ptr == NULL)
-                               error = ENOPROTOOPT ;
-                       else
+                       if (DUMMYNET_LOADED)
                                error = ip_dn_ctl_ptr(sopt);
+                       else
+                               error = ENOPROTOOPT;
                        break ;
 #endif /* DUMMYNET */
 
@@ -428,10 +483,10 @@ rip_ctloutput(so, sopt)
                case IP_DUMMYNET_CONFIGURE:
                case IP_DUMMYNET_DEL:
                case IP_DUMMYNET_FLUSH:
-                       if (ip_dn_ctl_ptr == NULL)
-                               error = ENOPROTOOPT ;
-                       else
+                       if (DUMMYNET_LOADED)
                                error = ip_dn_ctl_ptr(sopt);
+                       else
+                               error = ENOPROTOOPT ;
                        break ;
 #endif
 
@@ -493,6 +548,7 @@ rip_ctlinput(cmd, sa, vip)
 
        switch (cmd) {
        case PRC_IFDOWN:
+               lck_mtx_lock(rt_mtx);
                for (ia = in_ifaddrhead.tqh_first; ia;
                     ia = ia->ia_link.tqe_next) {
                        if (ia->ia_ifa.ifa_addr == sa
@@ -500,7 +556,7 @@ rip_ctlinput(cmd, sa, vip)
                                /*
                                 * in_ifscrub kills the interface route.
                                 */
-                               in_ifscrub(ia->ia_ifp, ia);
+                               in_ifscrub(ia->ia_ifp, ia, 1);
                                /*
                                 * in_ifadown gets rid of all the rest of
                                 * the routes.  This is not quite the right
@@ -511,16 +567,20 @@ rip_ctlinput(cmd, sa, vip)
                                break;
                        }
                }
+               lck_mtx_unlock(rt_mtx);
                break;
 
        case PRC_IFUP:
+               lck_mtx_lock(rt_mtx);
                for (ia = in_ifaddrhead.tqh_first; ia;
                     ia = ia->ia_link.tqe_next) {
                        if (ia->ia_ifa.ifa_addr == sa)
                                break;
                }
-               if (ia == 0 || (ia->ia_flags & IFA_ROUTE))
+               if (ia == 0 || (ia->ia_flags & IFA_ROUTE)) {
+                       lck_mtx_unlock(rt_mtx);
                        return;
+               }
                flags = RTF_UP;
                ifp = ia->ia_ifa.ifa_ifp;
 
@@ -528,7 +588,8 @@ rip_ctlinput(cmd, sa, vip)
                    || (ifp->if_flags & IFF_POINTOPOINT))
                        flags |= RTF_HOST;
 
-               err = rtinit(&ia->ia_ifa, RTM_ADD, flags);
+               err = rtinit_locked(&ia->ia_ifa, RTM_ADD, flags);
+               lck_mtx_unlock(rt_mtx);
                if (err == 0)
                        ia->ia_flags |= IFA_ROUTE;
                break;
@@ -612,15 +673,21 @@ rip_bind(struct socket *so, struct sockaddr *nam, struct proc *p)
 {
        struct inpcb *inp = sotoinpcb(so);
        struct sockaddr_in *addr = (struct sockaddr_in *)nam;
+       struct ifaddr *ifa = NULL;
 
        if (nam->sa_len != sizeof(*addr))
                return EINVAL;
 
-       if (TAILQ_EMPTY(&ifnet) || ((addr->sin_family != AF_INET) &&
+       if (TAILQ_EMPTY(&ifnet_head) || ((addr->sin_family != AF_INET) &&
                                    (addr->sin_family != AF_IMPLINK)) ||
            (addr->sin_addr.s_addr &&
-            ifa_ifwithaddr((struct sockaddr *)addr) == 0))
+            (ifa = ifa_ifwithaddr((struct sockaddr *)addr)) == 0)) {
                return EADDRNOTAVAIL;
+       }
+       else if (ifa) {
+               ifafree(ifa);
+               ifa = NULL;
+       }
        inp->inp_laddr = addr->sin_addr;
        return 0;
 }
@@ -633,7 +700,7 @@ rip_connect(struct socket *so, struct sockaddr *nam, struct proc *p)
 
        if (nam->sa_len != sizeof(*addr))
                return EINVAL;
-       if (TAILQ_EMPTY(&ifnet))
+       if (TAILQ_EMPTY(&ifnet_head))
                return EADDRNOTAVAIL;
        if ((addr->sin_family != AF_INET) &&
            (addr->sin_family != AF_IMPLINK))
@@ -673,6 +740,38 @@ rip_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
        return rip_output(m, so, dst);
 }
 
+/* note: rip_unlock is called from different protos  instead of the generic socket_unlock,
+ * it will handle the socket dealloc on last reference 
+ * */
+int
+rip_unlock(struct socket *so, int refcount, int debug)
+{
+       int lr_saved;
+       struct inpcb *inp = sotoinpcb(so);
+
+       if (debug == 0) 
+               lr_saved = (unsigned int) __builtin_return_address(0);
+       else lr_saved = debug;
+
+       if (refcount) {
+               if (so->so_usecount <= 0)
+                       panic("rip_unlock: bad refoucnt so=%x val=%x\n", so, so->so_usecount);
+               so->so_usecount--;
+               if (so->so_usecount == 0 && (inp->inp_wantcnt == WNT_STOPUSING)) {
+                       /* cleanup after last reference */
+                       lck_mtx_unlock(so->so_proto->pr_domain->dom_mtx);
+                       lck_rw_lock_exclusive(ripcbinfo.mtx);
+                       in_pcbdispose(inp);
+                       lck_rw_done(ripcbinfo.mtx);
+                       return(0);
+               }
+       }
+       so->unlock_lr[so->next_unlock_lr] = (u_int *)lr_saved;
+       so->next_unlock_lr = (so->next_unlock_lr+1) % SO_LCKDBG_MAX;
+       lck_mtx_unlock(so->so_proto->pr_domain->dom_mtx);
+       return(0);
+}
+
 static int
 rip_pcblist SYSCTL_HANDLER_ARGS
 {
@@ -685,58 +784,67 @@ rip_pcblist SYSCTL_HANDLER_ARGS
         * The process of preparing the TCB list is too time-consuming and
         * resource-intensive to repeat twice on every request.
         */
-       if (req->oldptr == 0) {
+       lck_rw_lock_exclusive(ripcbinfo.mtx);
+       if (req->oldptr == USER_ADDR_NULL) {
                n = ripcbinfo.ipi_count;
                req->oldidx = 2 * (sizeof xig)
                        + (n + n/8) * sizeof(struct xinpcb);
+               lck_rw_done(ripcbinfo.mtx);
                return 0;
        }
 
-       if (req->newptr != 0)
+       if (req->newptr != USER_ADDR_NULL) {
+               lck_rw_done(ripcbinfo.mtx);
                return EPERM;
+       }
 
        /*
         * OK, now we're committed to doing something.
         */
-       s = splnet();
        gencnt = ripcbinfo.ipi_gencnt;
        n = ripcbinfo.ipi_count;
-       splx(s);
-
+       
+       bzero(&xig, sizeof(xig));
        xig.xig_len = sizeof xig;
        xig.xig_count = n;
        xig.xig_gen = gencnt;
        xig.xig_sogen = so_gencnt;
        error = SYSCTL_OUT(req, &xig, sizeof xig);
-       if (error)
+       if (error) {
+               lck_rw_done(ripcbinfo.mtx);
                return error;
+       }
     /*
      * We are done if there is no pcb
      */
-    if (n == 0)  
+    if (n == 0) {
+       lck_rw_done(ripcbinfo.mtx);
         return 0; 
+    }
 
        inp_list = _MALLOC(n * sizeof *inp_list, M_TEMP, M_WAITOK);
-       if (inp_list == 0)
+       if (inp_list == 0) {
+               lck_rw_done(ripcbinfo.mtx);
                return ENOMEM;
+       }
        
-       s = splnet();
        for (inp = ripcbinfo.listhead->lh_first, i = 0; inp && i < n;
             inp = inp->inp_list.le_next) {
-               if (inp->inp_gencnt <= gencnt)
+               if (inp->inp_gencnt <= gencnt && inp->inp_state != INPCB_STATE_DEAD)
                        inp_list[i++] = inp;
        }
-       splx(s);
        n = i;
 
        error = 0;
        for (i = 0; i < n; i++) {
                inp = inp_list[i];
-               if (inp->inp_gencnt <= gencnt) {
+               if (inp->inp_gencnt <= gencnt && inp->inp_state != INPCB_STATE_DEAD) {
                        struct xinpcb xi;
+
+                       bzero(&xi, sizeof(xi));
                        xi.xi_len = sizeof xi;
                        /* XXX should avoid extra copy */
-                       bcopy(inp, &xi.xi_inp, sizeof *inp);
+                       inpcb_to_compat(inp, &xi.xi_inp);
                        if (inp->inp_socket)
                                sotoxsocket(inp->inp_socket, &xi.xi_socket);
                        error = SYSCTL_OUT(req, &xi, sizeof xi);
@@ -750,14 +858,15 @@ rip_pcblist SYSCTL_HANDLER_ARGS
                 * while we were processing this request, and it
                 * might be necessary to retry.
                 */
-               s = splnet();
+               bzero(&xig, sizeof(xig));
+               xig.xig_len = sizeof xig;
                xig.xig_gen = ripcbinfo.ipi_gencnt;
                xig.xig_sogen = so_gencnt;
                xig.xig_count = ripcbinfo.ipi_count;
-               splx(s);
                error = SYSCTL_OUT(req, &xig, sizeof xig);
        }
        FREE(inp_list, M_TEMP);
+       lck_rw_done(ripcbinfo.mtx);
        return error;
 }
 
@@ -769,5 +878,5 @@ struct pr_usrreqs rip_usrreqs = {
        pru_connect2_notsupp, in_control, rip_detach, rip_disconnect,
        pru_listen_notsupp, in_setpeeraddr, pru_rcvd_notsupp,
        pru_rcvoob_notsupp, rip_send, pru_sense_null, rip_shutdown,
-       in_setsockaddr, sosend, soreceive, sopoll
+       in_setsockaddr, sosend, soreceive, pru_sopoll_notsupp
 };