vserver 1.9.3
[linux-2.6.git] / net / ipv4 / udp.c
index 05dcb54..49a9cd0 100644 (file)
@@ -174,14 +174,12 @@ gotit:
                        struct inet_opt *inet2 = inet_sk(sk2);
 
                        if (inet2->num == snum &&
-                           sk2 != sk &&
-                           !ipv6_only_sock(sk2) &&
+                           sk2 != sk && !ipv6_only_sock(sk2) &&
                            (!sk2->sk_bound_dev_if ||
                             !sk->sk_bound_dev_if ||
                             sk2->sk_bound_dev_if == sk->sk_bound_dev_if) &&
-                           (!inet2->rcv_saddr ||
-                            !inet->rcv_saddr ||
-                            inet2->rcv_saddr == inet->rcv_saddr) &&
+                           nx_addr_conflict(sk->sk_nx_info,
+                            tcp_v4_rcv_saddr(sk), sk2) &&
                            (!sk2->sk_reuse || !sk->sk_reuse))
                                goto fail;
                }
@@ -216,6 +214,17 @@ static void udp_v4_unhash(struct sock *sk)
        write_unlock_bh(&udp_hash_lock);
 }
 
+static inline int udp_in_list(struct nx_info *nx_info, u32 addr)
+{
+       int n = nx_info->nbipv4;
+       int i;
+
+       for (i=0; i<n; i++)
+               if (nx_info->ipv4[i] == addr)
+                       return 1;
+       return 0;
+}
+
 /* UDP is nearly always wildcards out the wazoo, it makes no sense to try
  * harder than this. -DaveM
  */
@@ -235,6 +244,11 @@ struct sock *udp_v4_lookup_longway(u32 saddr, u16 sport, u32 daddr, u16 dport, i
                                if (inet->rcv_saddr != daddr)
                                        continue;
                                score+=2;
+                       } else if (sk->sk_nx_info) {
+                               if (udp_in_list(sk->sk_nx_info, daddr))
+                                       score+=2;
+                               else
+                                       continue;
                        }
                        if (inet->daddr) {
                                if (inet->daddr != saddr)
@@ -290,11 +304,12 @@ static inline struct sock *udp_v4_mcast_next(struct sock *sk,
                if (inet->num != hnum                                   ||
                    (inet->daddr && inet->daddr != rmt_addr)            ||
                    (inet->dport != rmt_port && inet->dport)            ||
-                   (inet->rcv_saddr && inet->rcv_saddr != loc_addr)    ||
+                   (inet->rcv_saddr && inet->rcv_saddr != loc_addr &&
+                    inet->rcv_saddr2 && inet->rcv_saddr2 != loc_addr)  ||
                    ipv6_only_sock(s)                                   ||
                    (s->sk_bound_dev_if && s->sk_bound_dev_if != dif))
                        continue;
-               if (!ip_mc_sf_allow(sk, loc_addr, rmt_addr, dif))
+               if (!ip_mc_sf_allow(s, loc_addr, rmt_addr, dif))
                        continue;
                goto found;
        }
@@ -327,7 +342,7 @@ void udp_err(struct sk_buff *skb, u32 info)
 
        sk = udp_v4_lookup(iph->daddr, uh->dest, iph->saddr, uh->source, skb->dev->ifindex);
        if (sk == NULL) {
-               ICMP_INC_STATS_BH(IcmpInErrors);
+               ICMP_INC_STATS_BH(ICMP_MIB_INERRORS);
                return; /* No socket for error */
        }
 
@@ -599,6 +614,15 @@ int udp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
                                    .uli_u = { .ports =
                                               { .sport = inet->sport,
                                                 .dport = dport } } };
+               struct nx_info *nxi = sk->sk_nx_info;
+
+               if (nxi) {
+                       err = ip_find_src(nxi, &rt, &fl);
+                       if (err)
+                               goto out;
+                       if (daddr == IPI_LOOPBACK && !vx_check(0, VX_ADMIN))
+                               daddr = fl.fl4_dst = nxi->ipv4[0];
+               }
                err = ip_route_output_flow(&rt, &fl, sk, !(msg->msg_flags&MSG_DONTWAIT));
                if (err)
                        goto out;
@@ -654,7 +678,7 @@ out:
        if (free)
                kfree(ipc.opt);
        if (!err) {
-               UDP_INC_STATS_USER(UdpOutDatagrams);
+               UDP_INC_STATS_USER(UDP_MIB_OUTDATAGRAMS);
                return len;
        }
        return err;
@@ -725,7 +749,7 @@ int udp_ioctl(struct sock *sk, int cmd, unsigned long arg)
                case SIOCOUTQ:
                {
                        int amount = atomic_read(&sk->sk_wmem_alloc);
-                       return put_user(amount, (int *)arg);
+                       return put_user(amount, (int __user *)arg);
                }
 
                case SIOCINQ:
@@ -745,7 +769,7 @@ int udp_ioctl(struct sock *sk, int cmd, unsigned long arg)
                                amount = skb->len - sizeof(struct udphdr);
                        }
                        spin_unlock_irq(&sk->sk_receive_queue.lock);
-                       return put_user(amount, (int *)arg);
+                       return put_user(amount, (int __user *)arg);
                }
 
                default:
@@ -828,7 +852,10 @@ try_again:
        }
        if (inet->cmsg_flags)
                ip_cmsg_recv(msg, skb);
+
        err = copied;
+       if (flags & MSG_TRUNC)
+               err = skb->len - sizeof(struct udphdr);
   
 out_free:
        skb_free_datagram(sk, skb);
@@ -836,7 +863,7 @@ out:
        return err;
 
 csum_copy_err:
-       UDP_INC_STATS_BH(UdpInErrors);
+       UDP_INC_STATS_BH(UDP_MIB_INERRORS);
 
        /* Clear queue. */
        if (flags&MSG_PEEK) {
@@ -858,54 +885,6 @@ csum_copy_err:
        goto try_again;
 }
 
-int udp_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
-{
-       struct inet_opt *inet = inet_sk(sk);
-       struct sockaddr_in *usin = (struct sockaddr_in *) uaddr;
-       struct rtable *rt;
-       u32 saddr;
-       int oif;
-       int err;
-
-       
-       if (addr_len < sizeof(*usin)) 
-               return -EINVAL;
-
-       if (usin->sin_family != AF_INET) 
-               return -EAFNOSUPPORT;
-
-       sk_dst_reset(sk);
-
-       oif = sk->sk_bound_dev_if;
-       saddr = inet->saddr;
-       if (MULTICAST(usin->sin_addr.s_addr)) {
-               if (!oif)
-                       oif = inet->mc_index;
-               if (!saddr)
-                       saddr = inet->mc_addr;
-       }
-       err = ip_route_connect(&rt, usin->sin_addr.s_addr, saddr,
-                              RT_CONN_FLAGS(sk), oif,
-                              IPPROTO_UDP,
-                              inet->sport, usin->sin_port, sk);
-       if (err)
-               return err;
-       if ((rt->rt_flags & RTCF_BROADCAST) && !sock_flag(sk, SOCK_BROADCAST)) {
-               ip_rt_put(rt);
-               return -EACCES;
-       }
-       if (!inet->saddr)
-               inet->saddr = rt->rt_src;       /* Update source address */
-       if (!inet->rcv_saddr)
-               inet->rcv_saddr = rt->rt_src;
-       inet->daddr = rt->rt_dst;
-       inet->dport = usin->sin_port;
-       sk->sk_state = TCP_ESTABLISHED;
-       inet->id = jiffies;
-
-       sk_dst_set(sk, &rt->u.dst);
-       return(0);
-}
 
 int udp_disconnect(struct sock *sk, int flags)
 {
@@ -931,7 +910,7 @@ int udp_disconnect(struct sock *sk, int flags)
 
 static void udp_close(struct sock *sk, long timeout)
 {
-       inet_sock_release(sk);
+       sk_common_release(sk);
 }
 
 /* return:
@@ -964,6 +943,7 @@ static int udp_encap_rcv(struct sock * sk, struct sk_buff *skb)
        len = skb->tail - udpdata;
 
        switch (encap_type) {
+       default:
        case UDP_ENCAP_ESPINUDP:
                /* Check if this is a keepalive packet.  If so, eat it. */
                if (len == 1 && udpdata[0] == 0xff) {
@@ -974,35 +954,7 @@ static int udp_encap_rcv(struct sock * sk, struct sk_buff *skb)
                } else
                        /* Must be an IKE packet.. pass it through */
                        return 1;
-
-       decaps:
-               /* At this point we are sure that this is an ESPinUDP packet,
-                * so we need to remove 'len' bytes from the packet (the UDP
-                * header and optional ESP marker bytes) and then modify the
-                * protocol to ESP, and then call into the transform receiver.
-                */
-
-               /* Now we can update and verify the packet length... */
-               iph = skb->nh.iph;
-               iphlen = iph->ihl << 2;
-               iph->tot_len = htons(ntohs(iph->tot_len) - len);
-               if (skb->len < iphlen + len) {
-                       /* packet is too small!?! */
-                       return 0;
-               }
-
-               /* pull the data buffer up to the ESP header and set the
-                * transport header to point to ESP.  Keep UDP on the stack
-                * for later.
-                */
-               skb->h.raw = skb_pull(skb, len);
-
-               /* modify the protocol (it's ESP!) */
-               iph->protocol = IPPROTO_ESP;
-
-               /* and let the caller know to send this into the ESP processor... */
-               return -1;
-
+               break;
        case UDP_ENCAP_ESPINUDP_NON_IKE:
                /* Check if this is a keepalive packet.  If so, eat it. */
                if (len == 1 && udpdata[0] == 0xff) {
@@ -1012,17 +964,38 @@ static int udp_encap_rcv(struct sock * sk, struct sk_buff *skb)
                        
                        /* ESP Packet with Non-IKE marker */
                        len = sizeof(struct udphdr) + 2 * sizeof(u32);
-                       goto decaps;
                } else
                        /* Must be an IKE packet.. pass it through */
                        return 1;
+               break;
+       }
 
-       default:
-               if (net_ratelimit())
-                       printk(KERN_INFO "udp_encap_rcv(): Unhandled UDP encap type: %u\n",
-                              encap_type);
-               return 1;
+       /* At this point we are sure that this is an ESPinUDP packet,
+        * so we need to remove 'len' bytes from the packet (the UDP
+        * header and optional ESP marker bytes) and then modify the
+        * protocol to ESP, and then call into the transform receiver.
+        */
+
+       /* Now we can update and verify the packet length... */
+       iph = skb->nh.iph;
+       iphlen = iph->ihl << 2;
+       iph->tot_len = htons(ntohs(iph->tot_len) - len);
+       if (skb->len < iphlen + len) {
+               /* packet is too small!?! */
+               return 0;
        }
+
+       /* pull the data buffer up to the ESP header and set the
+        * transport header to point to ESP.  Keep UDP on the stack
+        * for later.
+        */
+       skb->h.raw = skb_pull(skb, len);
+
+       /* modify the protocol (it's ESP!) */
+       iph->protocol = IPPROTO_ESP;
+
+       /* and let the caller know to send this into the ESP processor... */
+       return -1;
 #endif
 }
 
@@ -1067,7 +1040,7 @@ static int udp_queue_rcv_skb(struct sock * sk, struct sk_buff *skb)
                if (ret < 0) {
                        /* process the ESP packet */
                        ret = xfrm4_rcv_encap(skb, up->encap_type);
-                       UDP_INC_STATS_BH(UdpInDatagrams);
+                       UDP_INC_STATS_BH(UDP_MIB_INDATAGRAMS);
                        return -ret;
                }
                /* FALLTHROUGH -- it's a UDP Packet */
@@ -1075,7 +1048,7 @@ static int udp_queue_rcv_skb(struct sock * sk, struct sk_buff *skb)
 
        if (sk->sk_filter && skb->ip_summed != CHECKSUM_UNNECESSARY) {
                if (__udp_checksum_complete(skb)) {
-                       UDP_INC_STATS_BH(UdpInErrors);
+                       UDP_INC_STATS_BH(UDP_MIB_INERRORS);
                        kfree_skb(skb);
                        return -1;
                }
@@ -1083,11 +1056,11 @@ static int udp_queue_rcv_skb(struct sock * sk, struct sk_buff *skb)
        }
 
        if (sock_queue_rcv_skb(sk,skb)<0) {
-               UDP_INC_STATS_BH(UdpInErrors);
+               UDP_INC_STATS_BH(UDP_MIB_INERRORS);
                kfree_skb(skb);
                return -1;
        }
-       UDP_INC_STATS_BH(UdpInDatagrams);
+       UDP_INC_STATS_BH(UDP_MIB_INDATAGRAMS);
        return 0;
 }
 
@@ -1215,7 +1188,7 @@ int udp_rcv(struct sk_buff *skb)
        if (udp_checksum_complete(skb))
                goto csum_error;
 
-       UDP_INC_STATS_BH(UdpNoPorts);
+       UDP_INC_STATS_BH(UDP_MIB_NOPORTS);
        icmp_send(skb, ICMP_DEST_UNREACH, ICMP_PORT_UNREACH, 0);
 
        /*
@@ -1235,7 +1208,7 @@ short_packet:
                        NIPQUAD(daddr),
                        ntohs(uh->dest)));
 no_header:
-       UDP_INC_STATS_BH(UdpInErrors);
+       UDP_INC_STATS_BH(UDP_MIB_INERRORS);
        kfree_skb(skb);
        return(0);
 
@@ -1252,7 +1225,7 @@ csum_error:
                        ntohs(uh->dest),
                        ulen));
 drop:
-       UDP_INC_STATS_BH(UdpInErrors);
+       UDP_INC_STATS_BH(UDP_MIB_INERRORS);
        kfree_skb(skb);
        return(0);
 }
@@ -1269,7 +1242,7 @@ static int udp_destroy_sock(struct sock *sk)
  *     Socket option code for UDP
  */
 static int udp_setsockopt(struct sock *sk, int level, int optname, 
-                         char *optval, int optlen)
+                         char __user *optval, int optlen)
 {
        struct udp_opt *up = udp_sk(sk);
        int val;
@@ -1281,7 +1254,7 @@ static int udp_setsockopt(struct sock *sk, int level, int optname,
        if(optlen<sizeof(int))
                return -EINVAL;
 
-       if (get_user(val, (int *)optval))
+       if (get_user(val, (int __user *)optval))
                return -EFAULT;
 
        switch(optname) {
@@ -1297,7 +1270,16 @@ static int udp_setsockopt(struct sock *sk, int level, int optname,
                break;
                
        case UDP_ENCAP:
-               up->encap_type = val;
+               switch (val) {
+               case 0:
+               case UDP_ENCAP_ESPINUDP:
+               case UDP_ENCAP_ESPINUDP_NON_IKE:
+                       up->encap_type = val;
+                       break;
+               default:
+                       err = -ENOPROTOOPT;
+                       break;
+               }
                break;
 
        default:
@@ -1309,7 +1291,7 @@ static int udp_setsockopt(struct sock *sk, int level, int optname,
 }
 
 static int udp_getsockopt(struct sock *sk, int level, int optname, 
-                         char *optval, int *optlen)
+                         char __user *optval, int __user *optlen)
 {
        struct udp_opt *up = udp_sk(sk);
        int val, len;
@@ -1349,7 +1331,7 @@ static int udp_getsockopt(struct sock *sk, int level, int optname,
 struct proto udp_prot = {
        .name =         "UDP",
        .close =        udp_close,
-       .connect =      udp_connect,
+       .connect =      ip4_datagram_connect,
        .disconnect =   udp_disconnect,
        .ioctl =        udp_ioctl,
        .destroy =      udp_destroy_sock,
@@ -1362,6 +1344,7 @@ struct proto udp_prot = {
        .hash =         udp_v4_hash,
        .unhash =       udp_v4_unhash,
        .get_port =     udp_v4_get_port,
+       .slab_obj_size = sizeof(struct udp_sock),
 };
 
 /* ------------------------------------------------------------------------ */
@@ -1374,8 +1357,10 @@ static struct sock *udp_get_first(struct seq_file *seq)
 
        for (state->bucket = 0; state->bucket < UDP_HTABLE_SIZE; ++state->bucket) {
                struct hlist_node *node;
+
                sk_for_each(sk, node, &udp_hash[state->bucket]) {
-                       if (sk->sk_family == state->family)
+                       if (sk->sk_family == state->family &&
+                               vx_check(sk->sk_xid, VX_WATCH|VX_IDENT))
                                goto found;
                }
        }
@@ -1392,7 +1377,8 @@ static struct sock *udp_get_next(struct seq_file *seq, struct sock *sk)
                sk = sk_next(sk);
 try_again:
                ;
-       } while (sk && sk->sk_family != state->family);
+       } while (sk && (sk->sk_family != state->family ||
+               !vx_check(sk->sk_xid, VX_WATCH|VX_IDENT)));
 
        if (!sk && ++state->bucket < UDP_HTABLE_SIZE) {
                sk = sk_head(&udp_hash[state->bucket]);
@@ -1550,7 +1536,6 @@ void udp4_proc_exit(void)
 }
 #endif /* CONFIG_PROC_FS */
 
-EXPORT_SYMBOL(udp_connect);
 EXPORT_SYMBOL(udp_disconnect);
 EXPORT_SYMBOL(udp_hash);
 EXPORT_SYMBOL(udp_hash_lock);