Merge branch 'master' of github.com:davem330/net
[linux-2.6.git] / net / ipv6 / raw.c
index e99307f..3486f62 100644 (file)
@@ -21,6 +21,7 @@
 #include <linux/errno.h>
 #include <linux/types.h>
 #include <linux/socket.h>
+#include <linux/slab.h>
 #include <linux/sockios.h>
 #include <linux/net.h>
 #include <linux/in6.h>
@@ -30,6 +31,7 @@
 #include <linux/netfilter.h>
 #include <linux/netfilter_ipv6.h>
 #include <linux/skbuff.h>
+#include <linux/compat.h>
 #include <asm/uaccess.h>
 #include <asm/ioctls.h>
 
@@ -65,14 +67,14 @@ static struct raw_hashinfo raw_v6_hashinfo = {
 };
 
 static struct sock *__raw_v6_lookup(struct net *net, struct sock *sk,
-               unsigned short num, struct in6_addr *loc_addr,
-               struct in6_addr *rmt_addr, int dif)
+               unsigned short num, const struct in6_addr *loc_addr,
+               const struct in6_addr *rmt_addr, int dif)
 {
        struct hlist_node *node;
        int is_multicast = ipv6_addr_is_multicast(loc_addr);
 
        sk_for_each_from(sk, node)
-               if (inet_sk(sk)->num == num) {
+               if (inet_sk(sk)->inet_num == num) {
                        struct ipv6_pinfo *np = inet6_sk(sk);
 
                        if (!net_eq(sock_net(sk), net))
@@ -122,20 +124,20 @@ static __inline__ int icmpv6_filter(struct sock *sk, struct sk_buff *skb)
 }
 
 #if defined(CONFIG_IPV6_MIP6) || defined(CONFIG_IPV6_MIP6_MODULE)
-static int (*mh_filter)(struct sock *sock, struct sk_buff *skb);
+typedef int mh_filter_t(struct sock *sock, struct sk_buff *skb);
 
-int rawv6_mh_filter_register(int (*filter)(struct sock *sock,
-                                          struct sk_buff *skb))
+static mh_filter_t __rcu *mh_filter __read_mostly;
+
+int rawv6_mh_filter_register(mh_filter_t filter)
 {
-       rcu_assign_pointer(mh_filter, filter);
+       RCU_INIT_POINTER(mh_filter, filter);
        return 0;
 }
 EXPORT_SYMBOL(rawv6_mh_filter_register);
 
-int rawv6_mh_filter_unregister(int (*filter)(struct sock *sock,
-                                            struct sk_buff *skb))
+int rawv6_mh_filter_unregister(mh_filter_t filter)
 {
-       rcu_assign_pointer(mh_filter, NULL);
+       RCU_INIT_POINTER(mh_filter, NULL);
        synchronize_rcu();
        return 0;
 }
@@ -152,8 +154,8 @@ EXPORT_SYMBOL(rawv6_mh_filter_unregister);
  */
 static int ipv6_raw_deliver(struct sk_buff *skb, int nexthdr)
 {
-       struct in6_addr *saddr;
-       struct in6_addr *daddr;
+       const struct in6_addr *saddr;
+       const struct in6_addr *daddr;
        struct sock *sk;
        int delivered = 0;
        __u8 hash;
@@ -191,10 +193,10 @@ static int ipv6_raw_deliver(struct sk_buff *skb, int nexthdr)
                         * policy is placed in rawv6_rcv() because it is
                         * required for each socket.
                         */
-                       int (*filter)(struct sock *sock, struct sk_buff *skb);
+                       mh_filter_t *filter;
 
                        filter = rcu_dereference(mh_filter);
-                       filtered = filter ? filter(sk, skb) : 0;
+                       filtered = filter ? (*filter)(sk, skb) : 0;
                        break;
                }
 #endif
@@ -249,7 +251,7 @@ static int rawv6_bind(struct sock *sk, struct sockaddr *uaddr, int addr_len)
 
        /* Raw sockets are IPv6 only */
        if (addr_type == IPV6_ADDR_MAPPED)
-               return(-EADDRNOTAVAIL);
+               return -EADDRNOTAVAIL;
 
        lock_sock(sk);
 
@@ -257,6 +259,7 @@ static int rawv6_bind(struct sock *sk, struct sockaddr *uaddr, int addr_len)
        if (sk->sk_state != TCP_CLOSE)
                goto out;
 
+       rcu_read_lock();
        /* Check if the address belongs to the host. */
        if (addr_type != IPV6_ADDR_ANY) {
                struct net_device *dev = NULL;
@@ -272,13 +275,13 @@ static int rawv6_bind(struct sock *sk, struct sockaddr *uaddr, int addr_len)
 
                        /* Binding to link-local address requires an interface */
                        if (!sk->sk_bound_dev_if)
-                               goto out;
+                               goto out_unlock;
 
-                       dev = dev_get_by_index(sock_net(sk), sk->sk_bound_dev_if);
-                       if (!dev) {
-                               err = -ENODEV;
-                               goto out;
-                       }
+                       err = -ENODEV;
+                       dev = dev_get_by_index_rcu(sock_net(sk),
+                                                  sk->sk_bound_dev_if);
+                       if (!dev)
+                               goto out_unlock;
                }
 
                /* ipv4 addr of the socket is invalid.  Only the
@@ -289,20 +292,18 @@ static int rawv6_bind(struct sock *sk, struct sockaddr *uaddr, int addr_len)
                        err = -EADDRNOTAVAIL;
                        if (!ipv6_chk_addr(sock_net(sk), &addr->sin6_addr,
                                           dev, 0)) {
-                               if (dev)
-                                       dev_put(dev);
-                               goto out;
+                               goto out_unlock;
                        }
                }
-               if (dev)
-                       dev_put(dev);
        }
 
-       inet->rcv_saddr = inet->saddr = v4addr;
+       inet->inet_rcv_saddr = inet->inet_saddr = v4addr;
        ipv6_addr_copy(&np->rcv_saddr, &addr->sin6_addr);
        if (!(addr_type & IPV6_ADDR_MULTICAST))
                ipv6_addr_copy(&np->saddr, &addr->sin6_addr);
        err = 0;
+out_unlock:
+       rcu_read_unlock();
 out:
        release_sock(sk);
        return err;
@@ -310,7 +311,7 @@ out:
 
 static void rawv6_err(struct sock *sk, struct sk_buff *skb,
               struct inet6_skb_parm *opt,
-              int type, int code, int offset, __be32 info)
+              u8 type, u8 code, int offset, __be32 info)
 {
        struct inet_sock *inet = inet_sk(sk);
        struct ipv6_pinfo *np = inet6_sk(sk);
@@ -343,11 +344,11 @@ static void rawv6_err(struct sock *sk, struct sk_buff *skb,
 }
 
 void raw6_icmp_error(struct sk_buff *skb, int nexthdr,
-               int type, int code, int inner_offset, __be32 info)
+               u8 type, u8 code, int inner_offset, __be32 info)
 {
        struct sock *sk;
        int hash;
-       struct in6_addr *saddr, *daddr;
+       const struct in6_addr *saddr, *daddr;
        struct net *net;
 
        hash = nexthdr & (RAW_HTABLE_SIZE - 1);
@@ -356,7 +357,7 @@ void raw6_icmp_error(struct sk_buff *skb, int nexthdr,
        sk = sk_head(&raw_v6_hashinfo.ht[hash]);
        if (sk != NULL) {
                /* Note: ipv6_hdr(skb) != skb->data */
-               struct ipv6hdr *ip6h = (struct ipv6hdr *)skb->data;
+               const struct ipv6hdr *ip6h = (const struct ipv6hdr *)skb->data;
                saddr = &ip6h->saddr;
                daddr = &ip6h->daddr;
                net = dev_net(skb->dev);
@@ -371,9 +372,9 @@ void raw6_icmp_error(struct sk_buff *skb, int nexthdr,
        read_unlock(&raw_v6_hashinfo.lock);
 }
 
-static inline int rawv6_rcv_skb(struct sock * sk, struct sk_buff * skb)
+static inline int rawv6_rcv_skb(struct sock *sk, struct sk_buff *skb)
 {
-       if ((raw6_sk(sk)->checksum || sk->sk_filter) &&
+       if ((raw6_sk(sk)->checksum || rcu_access_pointer(sk->sk_filter)) &&
            skb_checksum_complete(skb)) {
                atomic_inc(&sk->sk_drops);
                kfree_skb(skb);
@@ -381,8 +382,7 @@ static inline int rawv6_rcv_skb(struct sock * sk, struct sk_buff * skb)
        }
 
        /* Charge it to the socket. */
-       if (sock_queue_rcv_skb(sk,skb)<0) {
-               atomic_inc(&sk->sk_drops);
+       if (ip_queue_rcv_skb(sk, skb) < 0) {
                kfree_skb(skb);
                return NET_RX_DROP;
        }
@@ -416,14 +416,14 @@ int rawv6_rcv(struct sock *sk, struct sk_buff *skb)
                                   skb_network_header_len(skb));
                if (!csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
                                     &ipv6_hdr(skb)->daddr,
-                                    skb->len, inet->num, skb->csum))
+                                    skb->len, inet->inet_num, skb->csum))
                        skb->ip_summed = CHECKSUM_UNNECESSARY;
        }
        if (!skb_csum_unnecessary(skb))
                skb->csum = ~csum_unfold(csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
                                                         &ipv6_hdr(skb)->daddr,
                                                         skb->len,
-                                                        inet->num, 0));
+                                                        inet->inet_num, 0));
 
        if (inet->hdrincl) {
                if (skb_checksum_complete(skb)) {
@@ -462,6 +462,9 @@ static int rawv6_recvmsg(struct kiocb *iocb, struct sock *sk,
        if (flags & MSG_ERRQUEUE)
                return ipv6_recv_error(sk, msg, len);
 
+       if (np->rxpmtu && np->rxopt.bits.rxpmtu)
+               return ipv6_recv_rxpmtu(sk, msg, len);
+
        skb = skb_recv_datagram(sk, flags, noblock, &err);
        if (!skb)
                goto out;
@@ -497,7 +500,7 @@ static int rawv6_recvmsg(struct kiocb *iocb, struct sock *sk,
                        sin6->sin6_scope_id = IP6CB(skb)->iif;
        }
 
-       sock_recv_timestamp(msg, sk, skb);
+       sock_recv_ts_and_drops(msg, sk, skb);
 
        if (np->rxopt.all)
                datagram_recv_ctl(sk, msg, skb);
@@ -518,11 +521,10 @@ csum_copy_err:
           as some normal condition.
         */
        err = (flags&MSG_DONTWAIT) ? -EAGAIN : -EHOSTUNREACH;
-       atomic_inc(&sk->sk_drops);
        goto out;
 }
 
-static int rawv6_push_pending_frames(struct sock *sk, struct flowi *fl,
+static int rawv6_push_pending_frames(struct sock *sk, struct flowi6 *fl6,
                                     struct raw6_sock *rp)
 {
        struct sk_buff *skb;
@@ -540,8 +542,8 @@ static int rawv6_push_pending_frames(struct sock *sk, struct flowi *fl,
                goto out;
 
        offset = rp->offset;
-       total_len = inet_sk(sk)->cork.length - (skb_network_header(skb) -
-                                               skb->data);
+       total_len = inet_sk(sk)->cork.base.length - (skb_network_header(skb) -
+                                                    skb->data);
        if (offset >= total_len - 1) {
                err = -EINVAL;
                ip6_flush_pending_frames(sk);
@@ -584,11 +586,10 @@ static int rawv6_push_pending_frames(struct sock *sk, struct flowi *fl,
        if (unlikely(csum))
                tmp_csum = csum_sub(tmp_csum, csum_unfold(csum));
 
-       csum = csum_ipv6_magic(&fl->fl6_src,
-                                  &fl->fl6_dst,
-                                  total_len, fl->proto, tmp_csum);
+       csum = csum_ipv6_magic(&fl6->saddr, &fl6->daddr,
+                              total_len, fl6->flowi6_proto, tmp_csum);
 
-       if (csum == 0 && fl->proto == IPPROTO_UDP)
+       if (csum == 0 && fl6->flowi6_proto == IPPROTO_UDP)
                csum = CSUM_MANGLED_0;
 
        if (skb_store_bits(skb, offset, &csum, 2))
@@ -601,31 +602,33 @@ out:
 }
 
 static int rawv6_send_hdrinc(struct sock *sk, void *from, int length,
-                       struct flowi *fl, struct rt6_info *rt,
+                       struct flowi6 *fl6, struct dst_entry **dstp,
                        unsigned int flags)
 {
        struct ipv6_pinfo *np = inet6_sk(sk);
        struct ipv6hdr *iph;
        struct sk_buff *skb;
        int err;
+       struct rt6_info *rt = (struct rt6_info *)*dstp;
 
-       if (length > rt->u.dst.dev->mtu) {
-               ipv6_local_error(sk, EMSGSIZE, fl, rt->u.dst.dev->mtu);
+       if (length > rt->dst.dev->mtu) {
+               ipv6_local_error(sk, EMSGSIZE, fl6, rt->dst.dev->mtu);
                return -EMSGSIZE;
        }
        if (flags&MSG_PROBE)
                goto out;
 
        skb = sock_alloc_send_skb(sk,
-                                 length + LL_ALLOCATED_SPACE(rt->u.dst.dev) + 15,
+                                 length + LL_ALLOCATED_SPACE(rt->dst.dev) + 15,
                                  flags & MSG_DONTWAIT, &err);
        if (skb == NULL)
                goto error;
-       skb_reserve(skb, LL_RESERVED_SPACE(rt->u.dst.dev));
+       skb_reserve(skb, LL_RESERVED_SPACE(rt->dst.dev));
 
        skb->priority = sk->sk_priority;
        skb->mark = sk->sk_mark;
-       skb->dst = dst_clone(&rt->u.dst);
+       skb_dst_set(skb, &rt->dst);
+       *dstp = NULL;
 
        skb_put(skb, length);
        skb_reset_network_header(skb);
@@ -639,10 +642,10 @@ static int rawv6_send_hdrinc(struct sock *sk, void *from, int length,
                goto error_fault;
 
        IP6_UPD_PO_STATS(sock_net(sk), rt->rt6i_idev, IPSTATS_MIB_OUT, skb->len);
-       err = NF_HOOK(PF_INET6, NF_INET_LOCAL_OUT, skb, NULL, rt->u.dst.dev,
-                     dst_output);
+       err = NF_HOOK(NFPROTO_IPV6, NF_INET_LOCAL_OUT, skb, NULL,
+                     rt->dst.dev, dst_output);
        if (err > 0)
-               err = np->recverr ? net_xmit_errno(err) : 0;
+               err = net_xmit_errno(err);
        if (err)
                goto error;
 out:
@@ -653,10 +656,12 @@ error_fault:
        kfree_skb(skb);
 error:
        IP6_INC_STATS(sock_net(sk), rt->rt6i_idev, IPSTATS_MIB_OUTDISCARDS);
+       if (err == -ENOBUFS && !np->recverr)
+               err = 0;
        return err;
 }
 
-static int rawv6_probe_proto_opt(struct flowi *fl, struct msghdr *msg)
+static int rawv6_probe_proto_opt(struct flowi6 *fl6, struct msghdr *msg)
 {
        struct iovec *iov;
        u8 __user *type = NULL;
@@ -673,7 +678,7 @@ static int rawv6_probe_proto_opt(struct flowi *fl, struct msghdr *msg)
                if (!iov)
                        continue;
 
-               switch (fl->proto) {
+               switch (fl6->flowi6_proto) {
                case IPPROTO_ICMPV6:
                        /* check if one-byte field is readable or not. */
                        if (iov->iov_base && iov->iov_len < 1)
@@ -688,8 +693,8 @@ static int rawv6_probe_proto_opt(struct flowi *fl, struct msghdr *msg)
                                code = iov->iov_base;
 
                        if (type && code) {
-                               if (get_user(fl->fl_icmp_type, type) ||
-                                   get_user(fl->fl_icmp_code, code))
+                               if (get_user(fl6->fl6_icmp_type, type) ||
+                                   get_user(fl6->fl6_icmp_code, code))
                                        return -EFAULT;
                                probed = 1;
                        }
@@ -700,7 +705,7 @@ static int rawv6_probe_proto_opt(struct flowi *fl, struct msghdr *msg)
                        /* check if type field is readable or not. */
                        if (iov->iov_len > 2 - len) {
                                u8 __user *p = iov->iov_base;
-                               if (get_user(fl->fl_mh_type, &p[2 - len]))
+                               if (get_user(fl6->fl6_mh_type, &p[2 - len]))
                                        return -EFAULT;
                                probed = 1;
                        } else
@@ -722,17 +727,18 @@ static int rawv6_sendmsg(struct kiocb *iocb, struct sock *sk,
 {
        struct ipv6_txoptions opt_space;
        struct sockaddr_in6 * sin6 = (struct sockaddr_in6 *) msg->msg_name;
-       struct in6_addr *daddr, *final_p = NULL, final;
+       struct in6_addr *daddr, *final_p, final;
        struct inet_sock *inet = inet_sk(sk);
        struct ipv6_pinfo *np = inet6_sk(sk);
        struct raw6_sock *rp = raw6_sk(sk);
        struct ipv6_txoptions *opt = NULL;
        struct ip6_flowlabel *flowlabel = NULL;
        struct dst_entry *dst = NULL;
-       struct flowi fl;
+       struct flowi6 fl6;
        int addr_len = msg->msg_namelen;
        int hlimit = -1;
        int tclass = -1;
+       int dontfrag = -1;
        u16 proto;
        int err;
 
@@ -749,33 +755,33 @@ static int rawv6_sendmsg(struct kiocb *iocb, struct sock *sk,
        /*
         *      Get and verify the address.
         */
-       memset(&fl, 0, sizeof(fl));
+       memset(&fl6, 0, sizeof(fl6));
 
-       fl.mark = sk->sk_mark;
+       fl6.flowi6_mark = sk->sk_mark;
 
        if (sin6) {
                if (addr_len < SIN6_LEN_RFC2133)
                        return -EINVAL;
 
                if (sin6->sin6_family && sin6->sin6_family != AF_INET6)
-                       return(-EAFNOSUPPORT);
+                       return -EAFNOSUPPORT;
 
                /* port is the proto value [0..255] carried in nexthdr */
                proto = ntohs(sin6->sin6_port);
 
                if (!proto)
-                       proto = inet->num;
-               else if (proto != inet->num)
-                       return(-EINVAL);
+                       proto = inet->inet_num;
+               else if (proto != inet->inet_num)
+                       return -EINVAL;
 
                if (proto > 255)
-                       return(-EINVAL);
+                       return -EINVAL;
 
                daddr = &sin6->sin6_addr;
                if (np->sndflow) {
-                       fl.fl6_flowlabel = sin6->sin6_flowinfo&IPV6_FLOWINFO_MASK;
-                       if (fl.fl6_flowlabel&IPV6_FLOWLABEL_MASK) {
-                               flowlabel = fl6_sock_lookup(sk, fl.fl6_flowlabel);
+                       fl6.flowlabel = sin6->sin6_flowinfo&IPV6_FLOWINFO_MASK;
+                       if (fl6.flowlabel&IPV6_FLOWLABEL_MASK) {
+                               flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
                                if (flowlabel == NULL)
                                        return -EINVAL;
                                daddr = &flowlabel->dst;
@@ -793,31 +799,32 @@ static int rawv6_sendmsg(struct kiocb *iocb, struct sock *sk,
                if (addr_len >= sizeof(struct sockaddr_in6) &&
                    sin6->sin6_scope_id &&
                    ipv6_addr_type(daddr)&IPV6_ADDR_LINKLOCAL)
-                       fl.oif = sin6->sin6_scope_id;
+                       fl6.flowi6_oif = sin6->sin6_scope_id;
        } else {
                if (sk->sk_state != TCP_ESTABLISHED)
                        return -EDESTADDRREQ;
 
-               proto = inet->num;
+               proto = inet->inet_num;
                daddr = &np->daddr;
-               fl.fl6_flowlabel = np->flow_label;
+               fl6.flowlabel = np->flow_label;
        }
 
-       if (fl.oif == 0)
-               fl.oif = sk->sk_bound_dev_if;
+       if (fl6.flowi6_oif == 0)
+               fl6.flowi6_oif = sk->sk_bound_dev_if;
 
        if (msg->msg_controllen) {
                opt = &opt_space;
                memset(opt, 0, sizeof(struct ipv6_txoptions));
                opt->tot_len = sizeof(struct ipv6_txoptions);
 
-               err = datagram_send_ctl(sock_net(sk), msg, &fl, opt, &hlimit, &tclass);
+               err = datagram_send_ctl(sock_net(sk), sk, msg, &fl6, opt,
+                                       &hlimit, &tclass, &dontfrag);
                if (err < 0) {
                        fl6_sock_release(flowlabel);
                        return err;
                }
-               if ((fl.fl6_flowlabel&IPV6_FLOWLABEL_MASK) && !flowlabel) {
-                       flowlabel = fl6_sock_lookup(sk, fl.fl6_flowlabel);
+               if ((fl6.flowlabel&IPV6_FLOWLABEL_MASK) && !flowlabel) {
+                       flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
                        if (flowlabel == NULL)
                                return -EINVAL;
                }
@@ -830,46 +837,31 @@ static int rawv6_sendmsg(struct kiocb *iocb, struct sock *sk,
                opt = fl6_merge_options(&opt_space, flowlabel, opt);
        opt = ipv6_fixup_options(&opt_space, opt);
 
-       fl.proto = proto;
-       err = rawv6_probe_proto_opt(&fl, msg);
+       fl6.flowi6_proto = proto;
+       err = rawv6_probe_proto_opt(&fl6, msg);
        if (err)
                goto out;
 
        if (!ipv6_addr_any(daddr))
-               ipv6_addr_copy(&fl.fl6_dst, daddr);
+               ipv6_addr_copy(&fl6.daddr, daddr);
        else
-               fl.fl6_dst.s6_addr[15] = 0x1; /* :: means loopback (BSD'ism) */
-       if (ipv6_addr_any(&fl.fl6_src) && !ipv6_addr_any(&np->saddr))
-               ipv6_addr_copy(&fl.fl6_src, &np->saddr);
-
-       /* merge ip6_build_xmit from ip6_output */
-       if (opt && opt->srcrt) {
-               struct rt0_hdr *rt0 = (struct rt0_hdr *) opt->srcrt;
-               ipv6_addr_copy(&final, &fl.fl6_dst);
-               ipv6_addr_copy(&fl.fl6_dst, rt0->addr);
-               final_p = &final;
-       }
+               fl6.daddr.s6_addr[15] = 0x1; /* :: means loopback (BSD'ism) */
+       if (ipv6_addr_any(&fl6.saddr) && !ipv6_addr_any(&np->saddr))
+               ipv6_addr_copy(&fl6.saddr, &np->saddr);
 
-       if (!fl.oif && ipv6_addr_is_multicast(&fl.fl6_dst))
-               fl.oif = np->mcast_oif;
-       security_sk_classify_flow(sk, &fl);
+       final_p = fl6_update_dst(&fl6, opt, &final);
 
-       err = ip6_dst_lookup(sk, &dst, &fl);
-       if (err)
+       if (!fl6.flowi6_oif && ipv6_addr_is_multicast(&fl6.daddr))
+               fl6.flowi6_oif = np->mcast_oif;
+       security_sk_classify_flow(sk, flowi6_to_flowi(&fl6));
+
+       dst = ip6_dst_lookup_flow(sk, &fl6, final_p, true);
+       if (IS_ERR(dst)) {
+               err = PTR_ERR(dst);
                goto out;
-       if (final_p)
-               ipv6_addr_copy(&fl.fl6_dst, final_p);
-
-       err = __xfrm_lookup(sock_net(sk), &dst, &fl, sk, XFRM_LOOKUP_WAIT);
-       if (err < 0) {
-               if (err == -EREMOTE)
-                       err = ip6_dst_blackhole(sk, &dst, &fl);
-               if (err < 0)
-                       goto out;
        }
-
        if (hlimit < 0) {
-               if (ipv6_addr_is_multicast(&fl.fl6_dst))
+               if (ipv6_addr_is_multicast(&fl6.daddr))
                        hlimit = np->mcast_hops;
                else
                        hlimit = np->hop_limit;
@@ -877,28 +869,28 @@ static int rawv6_sendmsg(struct kiocb *iocb, struct sock *sk,
                        hlimit = ip6_dst_hoplimit(dst);
        }
 
-       if (tclass < 0) {
+       if (tclass < 0)
                tclass = np->tclass;
-               if (tclass < 0)
-                       tclass = 0;
-       }
+
+       if (dontfrag < 0)
+               dontfrag = np->dontfrag;
 
        if (msg->msg_flags&MSG_CONFIRM)
                goto do_confirm;
 
 back_from_confirm:
-       if (inet->hdrincl) {
-               err = rawv6_send_hdrinc(sk, msg->msg_iov, len, &fl, (struct rt6_info*)dst, msg->msg_flags);
-       } else {
+       if (inet->hdrincl)
+               err = rawv6_send_hdrinc(sk, msg->msg_iov, len, &fl6, &dst, msg->msg_flags);
+       else {
                lock_sock(sk);
                err = ip6_append_data(sk, ip_generic_getfrag, msg->msg_iov,
-                       len, 0, hlimit, tclass, opt, &fl, (struct rt6_info*)dst,
-                       msg->msg_flags);
+                       len, 0, hlimit, tclass, opt, &fl6, (struct rt6_info*)dst,
+                       msg->msg_flags, dontfrag);
 
                if (err)
                        ip6_flush_pending_frames(sk);
                else if (!(msg->msg_flags & MSG_MORE))
-                       err = rawv6_push_pending_frames(sk, &fl, rp);
+                       err = rawv6_push_pending_frames(sk, &fl6, rp);
                release_sock(sk);
        }
 done:
@@ -958,7 +950,7 @@ static int rawv6_geticmpfilter(struct sock *sk, int level, int optname,
 
 
 static int do_rawv6_setsockopt(struct sock *sk, int level, int optname,
-                           char __user *optval, int optlen)
+                           char __user *optval, unsigned int optlen)
 {
        struct raw6_sock *rp = raw6_sk(sk);
        int val;
@@ -967,57 +959,54 @@ static int do_rawv6_setsockopt(struct sock *sk, int level, int optname,
                return -EFAULT;
 
        switch (optname) {
-               case IPV6_CHECKSUM:
-                       if (inet_sk(sk)->num == IPPROTO_ICMPV6 &&
-                           level == IPPROTO_IPV6) {
-                               /*
-                                * RFC3542 tells that IPV6_CHECKSUM socket
-                                * option in the IPPROTO_IPV6 level is not
-                                * allowed on ICMPv6 sockets.
-                                * If you want to set it, use IPPROTO_RAW
-                                * level IPV6_CHECKSUM socket option
-                                * (Linux extension).
-                                */
-                               return -EINVAL;
-                       }
+       case IPV6_CHECKSUM:
+               if (inet_sk(sk)->inet_num == IPPROTO_ICMPV6 &&
+                   level == IPPROTO_IPV6) {
+                       /*
+                        * RFC3542 tells that IPV6_CHECKSUM socket
+                        * option in the IPPROTO_IPV6 level is not
+                        * allowed on ICMPv6 sockets.
+                        * If you want to set it, use IPPROTO_RAW
+                        * level IPV6_CHECKSUM socket option
+                        * (Linux extension).
+                        */
+                       return -EINVAL;
+               }
 
-                       /* You may get strange result with a positive odd offset;
-                          RFC2292bis agrees with me. */
-                       if (val > 0 && (val&1))
-                               return(-EINVAL);
-                       if (val < 0) {
-                               rp->checksum = 0;
-                       } else {
-                               rp->checksum = 1;
-                               rp->offset = val;
-                       }
+               /* You may get strange result with a positive odd offset;
+                  RFC2292bis agrees with me. */
+               if (val > 0 && (val&1))
+                       return -EINVAL;
+               if (val < 0) {
+                       rp->checksum = 0;
+               } else {
+                       rp->checksum = 1;
+                       rp->offset = val;
+               }
 
-                       return 0;
-                       break;
+               return 0;
 
-               default:
-                       return(-ENOPROTOOPT);
+       default:
+               return -ENOPROTOOPT;
        }
 }
 
 static int rawv6_setsockopt(struct sock *sk, int level, int optname,
-                         char __user *optval, int optlen)
+                         char __user *optval, unsigned int optlen)
 {
-       switch(level) {
-               case SOL_RAW:
-                       break;
+       switch (level) {
+       case SOL_RAW:
+               break;
 
-               case SOL_ICMPV6:
-                       if (inet_sk(sk)->num != IPPROTO_ICMPV6)
-                               return -EOPNOTSUPP;
-                       return rawv6_seticmpfilter(sk, level, optname, optval,
-                                                  optlen);
-               case SOL_IPV6:
-                       if (optname == IPV6_CHECKSUM)
-                               break;
-               default:
-                       return ipv6_setsockopt(sk, level, optname, optval,
-                                              optlen);
+       case SOL_ICMPV6:
+               if (inet_sk(sk)->inet_num != IPPROTO_ICMPV6)
+                       return -EOPNOTSUPP;
+               return rawv6_seticmpfilter(sk, level, optname, optval, optlen);
+       case SOL_IPV6:
+               if (optname == IPV6_CHECKSUM)
+                       break;
+       default:
+               return ipv6_setsockopt(sk, level, optname, optval, optlen);
        }
 
        return do_rawv6_setsockopt(sk, level, optname, optval, optlen);
@@ -1025,13 +1014,13 @@ static int rawv6_setsockopt(struct sock *sk, int level, int optname,
 
 #ifdef CONFIG_COMPAT
 static int compat_rawv6_setsockopt(struct sock *sk, int level, int optname,
-                                  char __user *optval, int optlen)
+                                  char __user *optval, unsigned int optlen)
 {
        switch (level) {
        case SOL_RAW:
                break;
        case SOL_ICMPV6:
-               if (inet_sk(sk)->num != IPPROTO_ICMPV6)
+               if (inet_sk(sk)->inet_num != IPPROTO_ICMPV6)
                        return -EOPNOTSUPP;
                return rawv6_seticmpfilter(sk, level, optname, optval, optlen);
        case SOL_IPV6:
@@ -1083,21 +1072,19 @@ static int do_rawv6_getsockopt(struct sock *sk, int level, int optname,
 static int rawv6_getsockopt(struct sock *sk, int level, int optname,
                          char __user *optval, int __user *optlen)
 {
-       switch(level) {
-               case SOL_RAW:
-                       break;
+       switch (level) {
+       case SOL_RAW:
+               break;
 
-               case SOL_ICMPV6:
-                       if (inet_sk(sk)->num != IPPROTO_ICMPV6)
-                               return -EOPNOTSUPP;
-                       return rawv6_geticmpfilter(sk, level, optname, optval,
-                                                  optlen);
-               case SOL_IPV6:
-                       if (optname == IPV6_CHECKSUM)
-                               break;
-               default:
-                       return ipv6_getsockopt(sk, level, optname, optval,
-                                              optlen);
+       case SOL_ICMPV6:
+               if (inet_sk(sk)->inet_num != IPPROTO_ICMPV6)
+                       return -EOPNOTSUPP;
+               return rawv6_geticmpfilter(sk, level, optname, optval, optlen);
+       case SOL_IPV6:
+               if (optname == IPV6_CHECKSUM)
+                       break;
+       default:
+               return ipv6_getsockopt(sk, level, optname, optval, optlen);
        }
 
        return do_rawv6_getsockopt(sk, level, optname, optval, optlen);
@@ -1111,7 +1098,7 @@ static int compat_rawv6_getsockopt(struct sock *sk, int level, int optname,
        case SOL_RAW:
                break;
        case SOL_ICMPV6:
-               if (inet_sk(sk)->num != IPPROTO_ICMPV6)
+               if (inet_sk(sk)->inet_num != IPPROTO_ICMPV6)
                        return -EOPNOTSUPP;
                return rawv6_geticmpfilter(sk, level, optname, optval, optlen);
        case SOL_IPV6:
@@ -1127,37 +1114,53 @@ static int compat_rawv6_getsockopt(struct sock *sk, int level, int optname,
 
 static int rawv6_ioctl(struct sock *sk, int cmd, unsigned long arg)
 {
-       switch(cmd) {
-               case SIOCOUTQ:
-               {
-                       int amount = atomic_read(&sk->sk_wmem_alloc);
-                       return put_user(amount, (int __user *)arg);
-               }
-               case SIOCINQ:
-               {
-                       struct sk_buff *skb;
-                       int amount = 0;
-
-                       spin_lock_bh(&sk->sk_receive_queue.lock);
-                       skb = skb_peek(&sk->sk_receive_queue);
-                       if (skb != NULL)
-                               amount = skb->tail - skb->transport_header;
-                       spin_unlock_bh(&sk->sk_receive_queue.lock);
-                       return put_user(amount, (int __user *)arg);
-               }
+       switch (cmd) {
+       case SIOCOUTQ: {
+               int amount = sk_wmem_alloc_get(sk);
 
-               default:
+               return put_user(amount, (int __user *)arg);
+       }
+       case SIOCINQ: {
+               struct sk_buff *skb;
+               int amount = 0;
+
+               spin_lock_bh(&sk->sk_receive_queue.lock);
+               skb = skb_peek(&sk->sk_receive_queue);
+               if (skb != NULL)
+                       amount = skb->tail - skb->transport_header;
+               spin_unlock_bh(&sk->sk_receive_queue.lock);
+               return put_user(amount, (int __user *)arg);
+       }
+
+       default:
+#ifdef CONFIG_IPV6_MROUTE
+               return ip6mr_ioctl(sk, cmd, (void __user *)arg);
+#else
+               return -ENOIOCTLCMD;
+#endif
+       }
+}
+
+#ifdef CONFIG_COMPAT
+static int compat_rawv6_ioctl(struct sock *sk, unsigned int cmd, unsigned long arg)
+{
+       switch (cmd) {
+       case SIOCOUTQ:
+       case SIOCINQ:
+               return -ENOIOCTLCMD;
+       default:
 #ifdef CONFIG_IPV6_MROUTE
-                       return ip6mr_ioctl(sk, cmd, (void __user *)arg);
+               return ip6mr_compat_ioctl(sk, cmd, compat_ptr(arg));
 #else
-                       return -ENOIOCTLCMD;
+               return -ENOIOCTLCMD;
 #endif
        }
 }
+#endif
 
 static void rawv6_close(struct sock *sk, long timeout)
 {
-       if (inet_sk(sk)->num == IPPROTO_RAW)
+       if (inet_sk(sk)->inet_num == IPPROTO_RAW)
                ip6_ra_control(sk, -1);
        ip6mr_sk_done(sk);
        sk_common_release(sk);
@@ -1176,7 +1179,7 @@ static int rawv6_init_sk(struct sock *sk)
 {
        struct raw6_sock *rp = raw6_sk(sk);
 
-       switch (inet_sk(sk)->num) {
+       switch (inet_sk(sk)->inet_num) {
        case IPPROTO_ICMPV6:
                rp->checksum = 1;
                rp->offset   = 2;
@@ -1188,7 +1191,7 @@ static int rawv6_init_sk(struct sock *sk)
        default:
                break;
        }
-       return(0);
+       return 0;
 }
 
 struct proto rawv6_prot = {
@@ -1213,6 +1216,7 @@ struct proto rawv6_prot = {
 #ifdef CONFIG_COMPAT
        .compat_setsockopt = compat_rawv6_setsockopt,
        .compat_getsockopt = compat_rawv6_getsockopt,
+       .compat_ioctl      = compat_rawv6_ioctl,
 #endif
 };
 
@@ -1220,24 +1224,24 @@ struct proto rawv6_prot = {
 static void raw6_sock_seq_show(struct seq_file *seq, struct sock *sp, int i)
 {
        struct ipv6_pinfo *np = inet6_sk(sp);
-       struct in6_addr *dest, *src;
+       const struct in6_addr *dest, *src;
        __u16 destp, srcp;
 
        dest  = &np->daddr;
        src   = &np->rcv_saddr;
        destp = 0;
-       srcp  = inet_sk(sp)->num;
+       srcp  = inet_sk(sp)->inet_num;
        seq_printf(seq,
                   "%4d: %08X%08X%08X%08X:%04X %08X%08X%08X%08X:%04X "
-                  "%02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %p %d\n",
+                  "%02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %pK %d\n",
                   i,
                   src->s6_addr32[0], src->s6_addr32[1],
                   src->s6_addr32[2], src->s6_addr32[3], srcp,
                   dest->s6_addr32[0], dest->s6_addr32[1],
                   dest->s6_addr32[2], dest->s6_addr32[3], destp,
                   sp->sk_state,
-                  atomic_read(&sp->sk_wmem_alloc),
-                  atomic_read(&sp->sk_rmem_alloc),
+                  sk_wmem_alloc_get(sp),
+                  sk_rmem_alloc_get(sp),
                   0, 0L, 0,
                   sock_i_uid(sp), 0,
                   sock_i_ino(sp),
@@ -1278,7 +1282,7 @@ static const struct file_operations raw6_seq_fops = {
        .release =      seq_release_net,
 };
 
-static int raw6_init_net(struct net *net)
+static int __net_init raw6_init_net(struct net *net)
 {
        if (!proc_net_fops_create(net, "raw6", S_IRUGO, &raw6_seq_fops))
                return -ENOMEM;
@@ -1286,7 +1290,7 @@ static int raw6_init_net(struct net *net)
        return 0;
 }
 
-static void raw6_exit_net(struct net *net)
+static void __net_exit raw6_exit_net(struct net *net)
 {
        proc_net_remove(net, "raw6");
 }
@@ -1338,7 +1342,6 @@ static struct inet_protosw rawv6_protosw = {
        .protocol       = IPPROTO_IP,   /* wild card */
        .prot           = &rawv6_prot,
        .ops            = &inet6_sockraw_ops,
-       .capability     = CAP_NET_RAW,
        .no_check       = UDP_CSUM_DEFAULT,
        .flags          = INET_PROTOSW_REUSE,
 };