ndisc.c

ipv6地址转换器

		 *	If we where using an "all destinations on link" route
		 *	delete it
		 */

		rt6_purge_dflt_routers(RTF_ALLONLINK);
	}

	if (rt)
		rt->rt6i_expires = jiffies + (HZ * lifetime);

	if (ra_msg->icmph.icmp6_hop_limit)
		in6_dev->cnf.hop_limit = ra_msg->icmph.icmp6_hop_limit;

	/*
	 *	Update Reachable Time and Retrans Timer
	 */

	if (in6_dev->nd_parms) {
		if (ra_msg->retrans_timer)
			in6_dev->nd_parms->retrans_time = (ntohl(ra_msg->retrans_timer)*HZ)/1000;

		if (ra_msg->reachable_time) {
			__u32 rtime = (ntohl(ra_msg->reachable_time)*HZ)/1000;

			if (rtime != in6_dev->nd_parms->base_reachable_time) {
				in6_dev->nd_parms->base_reachable_time = rtime;
				in6_dev->nd_parms->gc_staletime = 3 * rtime;
				in6_dev->nd_parms->reachable_time = neigh_rand_reach_time(rtime);
			}
		}
	}

	/*
	 *	Process options.
	 */

        while (optlen > 0) {
                int len = (opt[1] << 3);

		if (len == 0) {
			ND_PRINTK0("RA: opt has 0 len\n");
			break;
		}

                switch(*opt) {
                case ND_OPT_SOURCE_LL_ADDR:

			if (rt == NULL)
				break;
			
			if ((neigh = rt->rt6i_nexthop) != NULL &&
			    skb->dev->addr_len + 2 >= len)
				neigh_update(neigh, opt+2, NUD_STALE, 1, 1);
			break;

                case ND_OPT_PREFIX_INFO:
			addrconf_prefix_rcv(skb->dev, opt, len);
                        break;

                case ND_OPT_MTU:
			{
				int mtu;
				
				mtu = htonl(*(__u32 *)(opt+4));

				if (mtu < IPV6_MIN_MTU || mtu > skb->dev->mtu) {
					ND_PRINTK0("NDISC: router "
						   "announcement with mtu = %d\n",
						   mtu);
					break;
				}

				if (in6_dev->cnf.mtu6 != mtu) {
					in6_dev->cnf.mtu6 = mtu;

					if (rt)
						rt->u.dst.pmtu = mtu;

					rt6_mtu_change(skb->dev, mtu);
				}
			}
                        break;

		case ND_OPT_TARGET_LL_ADDR:
		case ND_OPT_REDIRECT_HDR:
			ND_PRINTK0("got illegal option with RA");
			break;
		default:
			ND_PRINTK0("unkown option in RA\n");
                };
                optlen -= len;
                opt += len;
        }
	if (rt)
		dst_release(&rt->u.dst);
}

static void ndisc_redirect_rcv(struct sk_buff *skb)
{
	struct inet6_dev *in6_dev;
	struct icmp6hdr *icmph;
	struct in6_addr *dest;
	struct in6_addr *target;	/* new first hop to destination */
	struct neighbour *neigh;
	int on_link = 0;
	int optlen;

	if (skb->nh.ipv6h->hop_limit != 255) {
		printk(KERN_WARNING "NDISC: fake ICMP redirect received\n");
		return;
	}

	if (!(ipv6_addr_type(&skb->nh.ipv6h->saddr) & IPV6_ADDR_LINKLOCAL)) {
		printk(KERN_WARNING "ICMP redirect: source address is not linklocal\n");
		return;
	}

	optlen = skb->tail - skb->h.raw;
	optlen -= sizeof(struct icmp6hdr) + 2 * sizeof(struct in6_addr);

	if (optlen < 0) {
		printk(KERN_WARNING "ICMP redirect: packet too small\n");
		return;
	}

	icmph = (struct icmp6hdr *) skb->h.raw;
	target = (struct in6_addr *) (icmph + 1);
	dest = target + 1;

	if (ipv6_addr_type(dest) & IPV6_ADDR_MULTICAST) {
		printk(KERN_WARNING "ICMP redirect for multicast addr\n");
		return;
	}

	if (ipv6_addr_cmp(dest, target) == 0) {
		on_link = 1;
	} else if (!(ipv6_addr_type(target) & IPV6_ADDR_LINKLOCAL)) {
		printk(KERN_WARNING "ICMP redirect: target address is not linklocal\n");
		return;
	}

	in6_dev = ipv6_get_idev(skb->dev);
	if (!in6_dev || in6_dev->cnf.forwarding || !in6_dev->cnf.accept_redirects)
		return;

	/* passed validation tests */

	/*
	   We install redirect only if nexthop state is valid.
	 */

	neigh = __neigh_lookup(&nd_tbl, target, skb->dev, 1);
	if (neigh) {
		ndisc_update(neigh, (u8*)(dest + 1), optlen, ND_OPT_TARGET_LL_ADDR);
		if (neigh->nud_state&NUD_VALID)
			rt6_redirect(dest, &skb->nh.ipv6h->saddr, neigh, on_link);
		else
			__neigh_event_send(neigh, NULL);
		neigh_release(neigh);
	}
}

void ndisc_send_redirect(struct sk_buff *skb, struct neighbour *neigh,
			 struct in6_addr *target)
{
	struct sock *sk = ndisc_socket->sk;
	int len = sizeof(struct icmp6hdr) + 2 * sizeof(struct in6_addr);
	struct sk_buff *buff;
	struct inet6_ifaddr *ifp;
	struct icmp6hdr *icmph;
	struct in6_addr *addrp;
	struct device *dev;
	struct rt6_info *rt;
	u8 *opt;
	int rd_len;
	int err;
	int hlen;

	dev = skb->dev;
	rt = rt6_lookup(&skb->nh.ipv6h->saddr, NULL, dev->ifindex, 1);

	if (rt == NULL)
		return;

	if (rt->rt6i_flags & RTF_GATEWAY) {
		ND_PRINTK1("ndisc_send_redirect: not a neighbour\n");
		dst_release(&rt->u.dst);
		return;
	}
	if (!xrlim_allow(&rt->u.dst, 1*HZ)) {
		dst_release(&rt->u.dst);
		return;
	}
	dst_release(&rt->u.dst);

	if (dev->addr_len) {
		if (neigh->nud_state&NUD_VALID) {
			len  += NDISC_OPT_SPACE(dev->addr_len);
		} else {
			/* If nexthop is not valid, do not redirect!
			   We will make it later, when will be sure,
			   that it is alive.
			 */
			return;
		}
	}

	rd_len = min(IPV6_MIN_MTU-sizeof(struct ipv6hdr)-len, skb->len + 8);
	rd_len &= ~0x7;
	len += rd_len;

	ifp = ipv6_get_lladdr(dev);

	if (ifp == NULL) {
		ND_PRINTK1("redirect: no link_local addr for dev\n");
		return;
	}

	buff = sock_alloc_send_skb(sk, MAX_HEADER + len + dev->hard_header_len + 15,
				   0, 0, &err);
	if (buff == NULL) {
		ND_PRINTK1("ndisc_send_redirect: alloc_skb failed\n");
		return;
	}

	hlen = 0;

	if (ndisc_build_ll_hdr(buff, dev, &skb->nh.ipv6h->saddr, NULL, len) == 0) {
		kfree_skb(buff);
		return;
	}

	ip6_nd_hdr(sk, buff, dev, &ifp->addr, &skb->nh.ipv6h->saddr,
		   IPPROTO_ICMPV6, len);

	icmph = (struct icmp6hdr *) skb_put(buff, len);

	memset(icmph, 0, sizeof(struct icmp6hdr));
	icmph->icmp6_type = NDISC_REDIRECT;

	/*
	 *	copy target and destination addresses
	 */

	addrp = (struct in6_addr *)(icmph + 1);
	ipv6_addr_copy(addrp, target);
	addrp++;
	ipv6_addr_copy(addrp, &skb->nh.ipv6h->daddr);

	opt = (u8*) (addrp + 1);

	/*
	 *	include target_address option
	 */

	if (dev->addr_len)
		opt = ndisc_fill_option(opt, ND_OPT_TARGET_LL_ADDR, neigh->ha, dev->addr_len);

	/*
	 *	build redirect option and copy skb over to the new packet.
	 */

	memset(opt, 0, 8);	
	*(opt++) = ND_OPT_REDIRECT_HDR;
	*(opt++) = (rd_len >> 3);
	opt += 6;

	memcpy(opt, skb->nh.ipv6h, rd_len - 8);

	icmph->icmp6_cksum = csum_ipv6_magic(&ifp->addr, &skb->nh.ipv6h->saddr,
					     len, IPPROTO_ICMPV6,
					     csum_partial((u8 *) icmph, len, 0));

	dev_queue_xmit(buff);

	icmpv6_statistics.Icmp6OutRedirects++;
	icmpv6_statistics.Icmp6OutMsgs++;
}

static __inline__ struct neighbour *
ndisc_recv_ns(struct in6_addr *saddr, struct sk_buff *skb)
{
	u8 *opt;

	opt = skb->h.raw;
	opt += sizeof(struct icmp6hdr) + sizeof(struct in6_addr);
	opt = ndisc_find_option(opt, skb->dev->addr_len+2, skb->tail - opt, ND_OPT_SOURCE_LL_ADDR);

	return neigh_event_ns(&nd_tbl, opt, saddr, skb->dev);
}

static __inline__ int ndisc_recv_na(struct neighbour *neigh, struct sk_buff *skb)
{
	struct nd_msg *msg = (struct nd_msg *) skb->h.raw;
	u8 *opt;

	opt = skb->h.raw;
	opt += sizeof(struct icmp6hdr) + sizeof(struct in6_addr);
	opt = ndisc_find_option(opt, skb->dev->addr_len+2, skb->tail - opt, ND_OPT_TARGET_LL_ADDR);

	return neigh_update(neigh, opt,
			    msg->icmph.icmp6_solicited ? NUD_REACHABLE : NUD_STALE,
			    msg->icmph.icmp6_override, 1);
}

static void pndisc_redo(struct sk_buff *skb)
{
	ndisc_rcv(skb, skb->len);
	kfree_skb(skb);
}

int ndisc_rcv(struct sk_buff *skb, unsigned long len)
{
	struct device *dev = skb->dev;
	struct in6_addr *saddr = &skb->nh.ipv6h->saddr;
	struct in6_addr *daddr = &skb->nh.ipv6h->daddr;
	struct nd_msg *msg = (struct nd_msg *) skb->h.raw;
	struct neighbour *neigh;
	struct inet6_ifaddr *ifp;

	switch (msg->icmph.icmp6_type) {
	case NDISC_NEIGHBOUR_SOLICITATION:
		if ((ifp = ipv6_chk_addr(&msg->target, dev, 1)) != NULL) {
			int addr_type = ipv6_addr_type(saddr);

			if (ifp->flags & ADDR_INVALID)
				return 0;
			if (ifp->flags & DAD_INCOMPLETE) {
				/* Address is tentative. If the source
				   is unspecified address, it is someone
				   does DAD, otherwise we ignore solicitations
				   until DAD timer expires.
				 */
				if (addr_type == IPV6_ADDR_ANY)
					addrconf_dad_failure(ifp);
				return 0;
			}

			if (addr_type == IPV6_ADDR_ANY) {
				struct in6_addr maddr;

				ipv6_addr_all_nodes(&maddr);
				ndisc_send_na(dev, NULL, &maddr, &ifp->addr, 
					      ifp->idev->cnf.forwarding, 0, 1, 1);
				return 0;
			}

			if (addr_type & IPV6_ADDR_UNICAST) {
				int inc = ipv6_addr_type(daddr)&IPV6_ADDR_MULTICAST;

				if (inc)
					nd_tbl.stats.rcv_probes_mcast++;
				else
					nd_tbl.stats.rcv_probes_ucast++;

				/* 
				 *	update / create cache entry
				 *	for the source adddress
				 */

				neigh = ndisc_recv_ns(saddr, skb);

				if (neigh) {
					ndisc_send_na(dev, neigh, saddr, &ifp->addr, 
						      ifp->idev->cnf.forwarding, 1, inc, inc);
					neigh_release(neigh);
				}
			}
		} else {
			struct inet6_dev *in6_dev = ipv6_get_idev(dev);
			int addr_type = ipv6_addr_type(saddr);

			if (in6_dev && in6_dev->cnf.forwarding &&
			    (addr_type & IPV6_ADDR_UNICAST) &&
			    pneigh_lookup(&nd_tbl, &msg->target, dev, 0)) {
				int inc = ipv6_addr_type(daddr)&IPV6_ADDR_MULTICAST;

				if (skb->stamp.tv_sec == 0 ||
				    skb->pkt_type == PACKET_HOST ||
				    inc == 0 ||
				    in6_dev->nd_parms->proxy_delay == 0) {
					if (inc)
						nd_tbl.stats.rcv_probes_mcast++;
					else
						nd_tbl.stats.rcv_probes_ucast++;

					neigh = ndisc_recv_ns(saddr, skb);

					if (neigh) {
						ndisc_send_na(dev, neigh, saddr, &msg->target,
							      0, 1, 0, inc);
						neigh_release(neigh);
					}
				} else {
					/* Hack. It will be freed upon exit from
					   ndisc_rcv
					 */
					atomic_inc(&skb->users);
					pneigh_enqueue(&nd_tbl, in6_dev->nd_parms, skb);
					return 0;
				}
			}
		}
		return 0;

	case NDISC_NEIGHBOUR_ADVERTISEMENT:
		if ((ipv6_addr_type(saddr)&IPV6_ADDR_MULTICAST) &&
		    msg->icmph.icmp6_solicited) {
			ND_PRINTK0("NDISC: solicited NA is multicasted\n");
			return 0;
		}
		/* BUG! Target can be link-local on ANOTHER interface. Fixed. */
		if ((ifp = ipv6_chk_addr(&msg->target, dev, 1))) {
			if (ifp->flags & ADDR_INVALID)
				return 0;
			if (ifp->flags & DAD_INCOMPLETE) {
				addrconf_dad_failure(ifp);
				return 0;
			}
			/* What should we make now? The advertisement
			   is invalid, but ndisc specs say nothing
			   about it. It could be misconfiguration, or
			   an smart proxy agent tries to help us :-)
			 */
			ND_PRINTK0("%s: someone avertise our address!\n",
				   ifp->idev->dev->name);
			return 0;
		}
		neigh = __neigh_lookup(&nd_tbl, &msg->target, skb->dev, 0);

		if (neigh) {
			if (neigh->flags & NTF_ROUTER) {
				if (msg->icmph.icmp6_router == 0) {
					/*
					 *	Change: router to host
					 */
					struct rt6_info *rt;
					rt = rt6_get_dflt_router(saddr, skb->dev);
					if (rt) {
						/* It is safe only because
						   we aer in BH */
						dst_release(&rt->u.dst);
						ip6_del_rt(rt);
					}
				}
			} else {
				if (msg->icmph.icmp6_router)
					neigh->flags |= NTF_ROUTER;
			}

			ndisc_recv_na(neigh, skb);
			neigh_release(neigh);
		}
		break;

	case NDISC_ROUTER_ADVERTISEMENT:
		ndisc_router_discovery(skb);
		break;

	case NDISC_REDIRECT:
		ndisc_redirect_rcv(skb);
		break;
	};

	return 0;
}

#ifdef CONFIG_PROC_FS
#ifndef CONFIG_RTNETLINK
int ndisc_get_info(char *buffer, char **start, off_t offset, int length, int dummy)
{
	int len=0;
	off_t pos=0;
	int size;
	unsigned long now = jiffies;
	int i;

	neigh_table_lock(&nd_tbl);

	for (i = 0; i <= NEIGH_HASHMASK; i++) {
		struct neighbour *neigh;

		for (neigh = nd_tbl.hash_buckets[i]; neigh; neigh = neigh->next) {
			int j;

			size = 0;
			for (j=0; j<16; j++) {
				sprintf(buffer+len+size, "%02x", neigh->primary_key[j]);
				size += 2;
			}

			size += sprintf(buffer+len+size,
				       " %02x %02x %02x %02x %08lx %08lx %08x %04x %04x %04x %8s ", i,
				       128,
				       neigh->type,
				       neigh->nud_state,
				       now - neigh->used,
				       now - neigh->confirmed,
				       neigh->parms->reachable_time,
				       neigh->parms->gc_staletime,
				       atomic_read(&neigh->refcnt),
				       neigh->flags | (!neigh->hh ? 0 : (neigh->hh->hh_output==dev_queue_xmit ? 4 : 2)),
				       neigh->dev->name);

			if ((neigh->nud_state&NUD_VALID) && neigh->dev->addr_len) {
				for (j=0; j < neigh->dev->addr_len; j++) {
					sprintf(buffer+len+size, "%02x", neigh->ha[j]);
					size += 2;
				}
			} else {
                                size += sprintf(buffer+len+size, "000000000000");
			}
			size += sprintf(buffer+len+size, "\n");
			len += size;
			pos += size;
		  
			if (pos <= offset)
				len=0;
			if (pos >= offset+length)
				goto done;
		}
	}

done:
	neigh_table_unlock(&nd_tbl);

	*start = buffer+len-(pos-offset);	/* Start of wanted data */
	len = pos-offset;			/* Start slop */
	if (len>length)
		len = length;			/* Ending slop */
	if (len<0)
		len = 0;
	return len;
}

struct proc_dir_entry ndisc_proc_entry =
{
        PROC_NET_NDISC, 5, "ndisc",
        S_IFREG | S_IRUGO, 1, 0, 0,
        0, NULL,
        &ndisc_get_info
};
#endif
#endif	/* CONFIG_PROC_FS */



__initfunc(int ndisc_init(struct net_proto_family *ops))
{
	struct sock *sk;
        int err;

	ndisc_socket = sock_alloc();
	if (ndisc_socket == NULL) {
		printk(KERN_ERR
		       "Failed to create the NDISC control socket.\n");
		return -1;
	}
	ndisc_socket->inode->i_uid = 0;
	ndisc_socket->inode->i_gid = 0;
	ndisc_socket->type = SOCK_RAW;

	if((err = ops->create(ndisc_socket, IPPROTO_ICMPV6)) < 0) {
		printk(KERN_DEBUG 
		       "Failed to initializee the NDISC control socket (err %d).\n",
		       err);
		sock_release(ndisc_socket);
		ndisc_socket = NULL; /* For safety. */
		return err;
	}

	sk = ndisc_socket->sk;
	sk->allocation = GFP_ATOMIC;
	sk->net_pinfo.af_inet6.hop_limit = 255;
	/* Do not loopback ndisc messages */
	sk->net_pinfo.af_inet6.mc_loop = 0;
	sk->num = 256;

        /*
         * Initialize the neighbour table
         */
	
	neigh_table_init(&nd_tbl);

#ifdef CONFIG_PROC_FS
#ifndef CONFIG_RTNETLINK
	proc_net_register(&ndisc_proc_entry);
#endif
#endif
#ifdef CONFIG_SYSCTL
	neigh_sysctl_register(NULL, &nd_tbl.parms, NET_IPV6, NET_IPV6_NEIGH, "ipv6");
#endif

	return 0;
}

void ndisc_cleanup(void)
{
#ifdef CONFIG_PROC_FS
#ifndef CONFIG_RTNETLINK
        proc_net_unregister(ndisc_proc_entry.low_ino);
#endif
#endif
	neigh_table_clear(&nd_tbl);
	sock_release(ndisc_socket);
	ndisc_socket = NULL; /* For safety. */
}