tcp_ipv6.c

Linux Kernel 2.6.9 for OMAP1710

/*
 *	TCP over IPv6
 *	Linux INET6 implementation 
 *
 *	Authors:
 *	Pedro Roque		<[e-mail]>	
 *
 *	$Id: tcp_ipv6.c,v 1.144 2002/02/01 22:01:04 davem Exp $
 *
 *	Based on: 
 *	linux/net/ipv4/tcp.c
 *	linux/net/ipv4/tcp_input.c
 *	linux/net/ipv4/tcp_output.c
 *
 *	Fixes:
 *	Hideaki YOSHIFUJI	:	sin6_scope_id support
 *	YOSHIFUJI Hideaki @USAGI and:	Support IPV6_V6ONLY socket option, which
 *	Alexey Kuznetsov		allow both IPv4 and IPv6 sockets to bind
 *					a single port at the same time.
 *	YOSHIFUJI Hideaki @USAGI:	convert /proc/net/tcp6 to seq_file.
 *
 *	This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

#include <linux/module.h>
#include <linux/config.h>
#include <linux/errno.h>
#include <linux/types.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/net.h>
#include <linux/jiffies.h>
#include <linux/in.h>
#include <linux/in6.h>
#include <linux/netdevice.h>
#include <linux/init.h>
#include <linux/jhash.h>
#include <linux/ipsec.h>
#include <linux/times.h>

#include <linux/ipv6.h>
#include <linux/icmpv6.h>
#include <linux/random.h>

#include <net/tcp.h>
#include <net/ndisc.h>
#include <net/ipv6.h>
#include <net/transp_v6.h>
#include <net/addrconf.h>
#include <net/ip6_route.h>
#include <net/ip6_checksum.h>
#include <net/inet_ecn.h>
#include <net/protocol.h>
#include <net/xfrm.h>
#include <net/addrconf.h>
#include <net/snmp.h>
#include <net/dsfield.h>

#include <asm/uaccess.h>

#include <linux/proc_fs.h>
#include <linux/seq_file.h>

static void	tcp_v6_send_reset(struct sk_buff *skb);
static void	tcp_v6_or_send_ack(struct sk_buff *skb, struct open_request *req);
static void	tcp_v6_send_check(struct sock *sk, struct tcphdr *th, int len, 
				  struct sk_buff *skb);

static int	tcp_v6_do_rcv(struct sock *sk, struct sk_buff *skb);
static int	tcp_v6_xmit(struct sk_buff *skb, int ipfragok);

static struct tcp_func ipv6_mapped;
static struct tcp_func ipv6_specific;

/* I have no idea if this is a good hash for v6 or not. -DaveM */
static __inline__ int tcp_v6_hashfn(struct in6_addr *laddr, u16 lport,
				    struct in6_addr *faddr, u16 fport)
{
	int hashent = (lport ^ fport);

	hashent ^= (laddr->s6_addr32[3] ^ faddr->s6_addr32[3]);
	hashent ^= hashent>>16;
	hashent ^= hashent>>8;
	return (hashent & (tcp_ehash_size - 1));
}

static __inline__ int tcp_v6_sk_hashfn(struct sock *sk)
{
	struct inet_opt *inet = inet_sk(sk);
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct in6_addr *laddr = &np->rcv_saddr;
	struct in6_addr *faddr = &np->daddr;
	__u16 lport = inet->num;
	__u16 fport = inet->dport;
	return tcp_v6_hashfn(laddr, lport, faddr, fport);
}

static inline int tcp_v6_bind_conflict(struct sock *sk,
				       struct tcp_bind_bucket *tb)
{
	struct sock *sk2;
	struct hlist_node *node;

	/* We must walk the whole port owner list in this case. -DaveM */
	sk_for_each_bound(sk2, node, &tb->owners) {
		if (sk != sk2 &&
		    (!sk->sk_bound_dev_if ||
		     !sk2->sk_bound_dev_if ||
		     sk->sk_bound_dev_if == sk2->sk_bound_dev_if) &&
		    (!sk->sk_reuse || !sk2->sk_reuse ||
		     sk2->sk_state == TCP_LISTEN) &&
		     ipv6_rcv_saddr_equal(sk, sk2))
			break;
	}

	return node != NULL;
}

/* Grrr, addr_type already calculated by caller, but I don't want
 * to add some silly "cookie" argument to this method just for that.
 * But it doesn't matter, the recalculation is in the rarest path
 * this function ever takes.
 */
static int tcp_v6_get_port(struct sock *sk, unsigned short snum)
{
	struct tcp_bind_hashbucket *head;
	struct tcp_bind_bucket *tb;
	struct hlist_node *node;
	int ret;

	local_bh_disable();
	if (snum == 0) {
		int low = sysctl_local_port_range[0];
		int high = sysctl_local_port_range[1];
		int remaining = (high - low) + 1;
		int rover;

		spin_lock(&tcp_portalloc_lock);
		rover = tcp_port_rover;
		do {	rover++;
			if ((rover < low) || (rover > high))
				rover = low;
			head = &tcp_bhash[tcp_bhashfn(rover)];
			spin_lock(&head->lock);
			tb_for_each(tb, node, &head->chain)
				if (tb->port == rover)
					goto next;
			break;
		next:
			spin_unlock(&head->lock);
		} while (--remaining > 0);
		tcp_port_rover = rover;
		spin_unlock(&tcp_portalloc_lock);

		/* Exhausted local port range during search? */
		ret = 1;
		if (remaining <= 0)
			goto fail;

		/* OK, here is the one we will use. */
		snum = rover;
	} else {
		head = &tcp_bhash[tcp_bhashfn(snum)];
		spin_lock(&head->lock);
		tb_for_each(tb, node, &head->chain)
			if (tb->port == snum)
				goto tb_found;
	}
	tb = NULL;
	goto tb_not_found;
tb_found:
	if (tb && !hlist_empty(&tb->owners)) {
		if (tb->fastreuse > 0 && sk->sk_reuse &&
		    sk->sk_state != TCP_LISTEN) {
			goto success;
		} else {
			ret = 1;
			if (tcp_v6_bind_conflict(sk, tb))
				goto fail_unlock;
		}
	}
tb_not_found:
	ret = 1;
	if (!tb && (tb = tcp_bucket_create(head, snum)) == NULL)
		goto fail_unlock;
	if (hlist_empty(&tb->owners)) {
		if (sk->sk_reuse && sk->sk_state != TCP_LISTEN)
			tb->fastreuse = 1;
		else
			tb->fastreuse = 0;
	} else if (tb->fastreuse &&
		   (!sk->sk_reuse || sk->sk_state == TCP_LISTEN))
		tb->fastreuse = 0;

success:
	if (!tcp_sk(sk)->bind_hash)
		tcp_bind_hash(sk, tb, snum);
	BUG_TRAP(tcp_sk(sk)->bind_hash == tb);
	ret = 0;

fail_unlock:
	spin_unlock(&head->lock);
fail:
	local_bh_enable();
	return ret;
}

static __inline__ void __tcp_v6_hash(struct sock *sk)
{
	struct hlist_head *list;
	rwlock_t *lock;

	BUG_TRAP(sk_unhashed(sk));

	if (sk->sk_state == TCP_LISTEN) {
		list = &tcp_listening_hash[tcp_sk_listen_hashfn(sk)];
		lock = &tcp_lhash_lock;
		tcp_listen_wlock();
	} else {
		sk->sk_hashent = tcp_v6_sk_hashfn(sk);
		list = &tcp_ehash[sk->sk_hashent].chain;
		lock = &tcp_ehash[sk->sk_hashent].lock;
		write_lock(lock);
	}

	__sk_add_node(sk, list);
	sock_prot_inc_use(sk->sk_prot);
	write_unlock(lock);
}


static void tcp_v6_hash(struct sock *sk)
{
	if (sk->sk_state != TCP_CLOSE) {
		struct tcp_opt *tp = tcp_sk(sk);

		if (tp->af_specific == &ipv6_mapped) {
			tcp_prot.hash(sk);
			return;
		}
		local_bh_disable();
		__tcp_v6_hash(sk);
		local_bh_enable();
	}
}

static struct sock *tcp_v6_lookup_listener(struct in6_addr *daddr, unsigned short hnum, int dif)
{
	struct sock *sk;
	struct hlist_node *node;
	struct sock *result = NULL;
	int score, hiscore;

	hiscore=0;
	read_lock(&tcp_lhash_lock);
	sk_for_each(sk, node, &tcp_listening_hash[tcp_lhashfn(hnum)]) {
		if (inet_sk(sk)->num == hnum && sk->sk_family == PF_INET6) {
			struct ipv6_pinfo *np = inet6_sk(sk);
			
			score = 1;
			if (!ipv6_addr_any(&np->rcv_saddr)) {
				if (ipv6_addr_cmp(&np->rcv_saddr, daddr))
					continue;
				score++;
			}
			if (sk->sk_bound_dev_if) {
				if (sk->sk_bound_dev_if != dif)
					continue;
				score++;
			}
			if (score == 3) {
				result = sk;
				break;
			}
			if (score > hiscore) {
				hiscore = score;
				result = sk;
			}
		}
	}
	if (result)
		sock_hold(result);
	read_unlock(&tcp_lhash_lock);
	return result;
}

/* Sockets in TCP_CLOSE state are _always_ taken out of the hash, so
 * we need not check it for TCP lookups anymore, thanks Alexey. -DaveM
 *
 * The sockhash lock must be held as a reader here.
 */

static inline struct sock *__tcp_v6_lookup_established(struct in6_addr *saddr, u16 sport,
						       struct in6_addr *daddr, u16 hnum,
						       int dif)
{
	struct tcp_ehash_bucket *head;
	struct sock *sk;
	struct hlist_node *node;
	__u32 ports = TCP_COMBINED_PORTS(sport, hnum);
	int hash;

	/* Optimize here for direct hit, only listening connections can
	 * have wildcards anyways.
	 */
	hash = tcp_v6_hashfn(daddr, hnum, saddr, sport);
	head = &tcp_ehash[hash];
	read_lock(&head->lock);
	sk_for_each(sk, node, &head->chain) {
		/* For IPV6 do the cheaper port and family tests first. */
		if(TCP_IPV6_MATCH(sk, saddr, daddr, ports, dif))
			goto hit; /* You sunk my battleship! */
	}
	/* Must check for a TIME_WAIT'er before going to listener hash. */
	sk_for_each(sk, node, &(head + tcp_ehash_size)->chain) {
		/* FIXME: acme: check this... */
		struct tcp_tw_bucket *tw = (struct tcp_tw_bucket *)sk;

		if(*((__u32 *)&(tw->tw_dport))	== ports	&&
		   sk->sk_family		== PF_INET6) {
			if(!ipv6_addr_cmp(&tw->tw_v6_daddr, saddr)	&&
			   !ipv6_addr_cmp(&tw->tw_v6_rcv_saddr, daddr)	&&
			   (!sk->sk_bound_dev_if || sk->sk_bound_dev_if == dif))
				goto hit;
		}
	}
	read_unlock(&head->lock);
	return NULL;

hit:
	sock_hold(sk);
	read_unlock(&head->lock);
	return sk;
}


static inline struct sock *__tcp_v6_lookup(struct in6_addr *saddr, u16 sport,
					   struct in6_addr *daddr, u16 hnum,
					   int dif)
{
	struct sock *sk;

	sk = __tcp_v6_lookup_established(saddr, sport, daddr, hnum, dif);

	if (sk)
		return sk;

	return tcp_v6_lookup_listener(daddr, hnum, dif);
}

inline struct sock *tcp_v6_lookup(struct in6_addr *saddr, u16 sport,
				  struct in6_addr *daddr, u16 dport,
				  int dif)
{
	struct sock *sk;

	local_bh_disable();
	sk = __tcp_v6_lookup(saddr, sport, daddr, ntohs(dport), dif);
	local_bh_enable();

	return sk;
}


/*
 * Open request hash tables.
 */

static u32 tcp_v6_synq_hash(struct in6_addr *raddr, u16 rport, u32 rnd)
{
	u32 a, b, c;

	a = raddr->s6_addr32[0];
	b = raddr->s6_addr32[1];
	c = raddr->s6_addr32[2];

	a += JHASH_GOLDEN_RATIO;
	b += JHASH_GOLDEN_RATIO;
	c += rnd;
	__jhash_mix(a, b, c);

	a += raddr->s6_addr32[3];
	b += (u32) rport;
	__jhash_mix(a, b, c);

	return c & (TCP_SYNQ_HSIZE - 1);
}

static struct open_request *tcp_v6_search_req(struct tcp_opt *tp,
					      struct open_request ***prevp,
					      __u16 rport,
					      struct in6_addr *raddr,
					      struct in6_addr *laddr,
					      int iif)
{
	struct tcp_listen_opt *lopt = tp->listen_opt;
	struct open_request *req, **prev;  

	for (prev = &lopt->syn_table[tcp_v6_synq_hash(raddr, rport, lopt->hash_rnd)];
	     (req = *prev) != NULL;
	     prev = &req->dl_next) {
		if (req->rmt_port == rport &&
		    req->class->family == AF_INET6 &&
		    !ipv6_addr_cmp(&req->af.v6_req.rmt_addr, raddr) &&
		    !ipv6_addr_cmp(&req->af.v6_req.loc_addr, laddr) &&
		    (!req->af.v6_req.iif || req->af.v6_req.iif == iif)) {
			BUG_TRAP(req->sk == NULL);
			*prevp = prev;
			return req;
		}
	}

	return NULL;
}

static __inline__ u16 tcp_v6_check(struct tcphdr *th, int len,
				   struct in6_addr *saddr, 
				   struct in6_addr *daddr, 
				   unsigned long base)
{
	return csum_ipv6_magic(saddr, daddr, len, IPPROTO_TCP, base);
}

static __u32 tcp_v6_init_sequence(struct sock *sk, struct sk_buff *skb)
{
	if (skb->protocol == htons(ETH_P_IPV6)) {
		return secure_tcpv6_sequence_number(skb->nh.ipv6h->daddr.s6_addr32,
						    skb->nh.ipv6h->saddr.s6_addr32,
						    skb->h.th->dest,
						    skb->h.th->source);
	} else {
		return secure_tcp_sequence_number(skb->nh.iph->daddr,
						  skb->nh.iph->saddr,
						  skb->h.th->dest,
						  skb->h.th->source);
	}
}

static int tcp_v6_check_established(struct sock *sk)
{
	struct inet_opt *inet = inet_sk(sk);
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct in6_addr *daddr = &np->rcv_saddr;
	struct in6_addr *saddr = &np->daddr;
	int dif = sk->sk_bound_dev_if;
	u32 ports = TCP_COMBINED_PORTS(inet->dport, inet->num);
	int hash = tcp_v6_hashfn(daddr, inet->num, saddr, inet->dport);
	struct tcp_ehash_bucket *head = &tcp_ehash[hash];
	struct sock *sk2;
	struct hlist_node *node;
	struct tcp_tw_bucket *tw;

	write_lock_bh(&head->lock);

	/* Check TIME-WAIT sockets first. */
	sk_for_each(sk2, node, &(head + tcp_ehash_size)->chain) {
		tw = (struct tcp_tw_bucket*)sk2;

		if(*((__u32 *)&(tw->tw_dport))	== ports	&&
		   sk2->sk_family		== PF_INET6	&&
		   !ipv6_addr_cmp(&tw->tw_v6_daddr, saddr)	&&
		   !ipv6_addr_cmp(&tw->tw_v6_rcv_saddr, daddr)	&&
		   sk2->sk_bound_dev_if == sk->sk_bound_dev_if) {
			struct tcp_opt *tp = tcp_sk(sk);

			if (tw->tw_ts_recent_stamp) {
				/* See comment in tcp_ipv4.c */
				tp->write_seq = tw->tw_snd_nxt + 65535 + 2;
				if (!tp->write_seq)
					tp->write_seq = 1;
				tp->ts_recent = tw->tw_ts_recent;
				tp->ts_recent_stamp = tw->tw_ts_recent_stamp;
				sock_hold(sk2);
				goto unique;
			} else
				goto not_unique;
		}
	}
	tw = NULL;

	/* And established part... */
	sk_for_each(sk2, node, &head->chain) {
		if(TCP_IPV6_MATCH(sk2, saddr, daddr, ports, dif))
			goto not_unique;
	}

unique:
	BUG_TRAP(sk_unhashed(sk));
	__sk_add_node(sk, &head->chain);
	sk->sk_hashent = hash;
	sock_prot_inc_use(sk->sk_prot);
	write_unlock_bh(&head->lock);

	if (tw) {
		/* Silly. Should hash-dance instead... */
		local_bh_disable();
		tcp_tw_deschedule(tw);
		NET_INC_STATS_BH(LINUX_MIB_TIMEWAITRECYCLED);
		local_bh_enable();

		tcp_tw_put(tw);
	}
	return 0;

not_unique:
	write_unlock_bh(&head->lock);
	return -EADDRNOTAVAIL;
}

static int tcp_v6_hash_connect(struct sock *sk)
{
	struct tcp_bind_hashbucket *head;
	struct tcp_bind_bucket *tb;

	/* XXX */
	if (inet_sk(sk)->num == 0) { 
		int err = tcp_v6_get_port(sk, inet_sk(sk)->num);
		if (err)
			return err;
		inet_sk(sk)->sport = htons(inet_sk(sk)->num);
	}

	head = &tcp_bhash[tcp_bhashfn(inet_sk(sk)->num)];
	tb = tb_head(head);

	spin_lock_bh(&head->lock);

	if (sk_head(&tb->owners) == sk && !sk->sk_bind_node.next) {
		__tcp_v6_hash(sk);
		spin_unlock_bh(&head->lock);
		return 0;
	} else {
		spin_unlock_bh(&head->lock);
		return tcp_v6_check_established(sk);
	}
}

static __inline__ int tcp_v6_iif(struct sk_buff *skb)
{
	return IP6CB(skb)->iif;
}