socket.c

	if (asoc) {
		info.sinfo_stream = asoc->default_stream;
		info.sinfo_flags = asoc->default_flags;
		info.sinfo_ppid = asoc->default_ppid;
		info.sinfo_context = asoc->default_context;
		info.sinfo_timetolive = asoc->default_timetolive;
	} else {
		info.sinfo_stream = sp->default_stream;
		info.sinfo_flags = sp->default_flags;
		info.sinfo_ppid = sp->default_ppid;
		info.sinfo_context = sp->default_context;
		info.sinfo_timetolive = sp->default_timetolive;
	}

	if (copy_to_user(optval, &info, sizeof(struct sctp_sndrcvinfo)))
		return -EFAULT;

	return 0;
}

/*
 *
 * 7.1.5 SCTP_NODELAY
 *
 * Turn on/off any Nagle-like algorithm.  This means that packets are
 * generally sent as soon as possible and no unnecessary delays are
 * introduced, at the cost of more packets in the network.  Expects an
 * integer boolean flag.
 */

static int sctp_getsockopt_nodelay(struct sock *sk, int len,
				   char __user *optval, int __user *optlen)
{
	int val;

	if (len < sizeof(int))
		return -EINVAL;

	len = sizeof(int);
	val = (sctp_sk(sk)->nodelay == 1);
	if (put_user(len, optlen))
		return -EFAULT;
	if (copy_to_user(optval, &val, len))
		return -EFAULT;
	return 0;
}

/*
 *
 * 7.1.1 SCTP_RTOINFO
 *
 * The protocol parameters used to initialize and bound retransmission
 * timeout (RTO) are tunable. sctp_rtoinfo structure is used to access
 * and modify these parameters.
 * All parameters are time values, in milliseconds.  A value of 0, when
 * modifying the parameters, indicates that the current value should not
 * be changed.
 *
 */
static int sctp_getsockopt_rtoinfo(struct sock *sk, int len,
				char __user *optval,
				int __user *optlen) {
	struct sctp_rtoinfo rtoinfo;
	struct sctp_association *asoc;

	if (len != sizeof (struct sctp_rtoinfo))
		return -EINVAL;

	if (copy_from_user(&rtoinfo, optval, sizeof (struct sctp_rtoinfo)))
		return -EFAULT;

	asoc = sctp_id2assoc(sk, rtoinfo.srto_assoc_id);

	if (!asoc && rtoinfo.srto_assoc_id && sctp_style(sk, UDP))
		return -EINVAL;

	/* Values corresponding to the specific association. */
	if (asoc) {
		rtoinfo.srto_initial = jiffies_to_msecs(asoc->rto_initial);
		rtoinfo.srto_max = jiffies_to_msecs(asoc->rto_max);
		rtoinfo.srto_min = jiffies_to_msecs(asoc->rto_min);
	} else {
		/* Values corresponding to the endpoint. */
		struct sctp_sock *sp = sctp_sk(sk);

		rtoinfo.srto_initial = sp->rtoinfo.srto_initial;
		rtoinfo.srto_max = sp->rtoinfo.srto_max;
		rtoinfo.srto_min = sp->rtoinfo.srto_min;
	}

	if (put_user(len, optlen))
		return -EFAULT;

	if (copy_to_user(optval, &rtoinfo, len))
		return -EFAULT;

	return 0;
}

/*
 *
 * 7.1.2 SCTP_ASSOCINFO
 *
 * This option is used to tune the the maximum retransmission attempts
 * of the association.
 * Returns an error if the new association retransmission value is
 * greater than the sum of the retransmission value  of the peer.
 * See [SCTP] for more information.
 *
 */
static int sctp_getsockopt_associnfo(struct sock *sk, int len,
				     char __user *optval,
				     int __user *optlen)
{

	struct sctp_assocparams assocparams;
	struct sctp_association *asoc;
	struct list_head *pos;
	int cnt = 0;

	if (len != sizeof (struct sctp_assocparams))
		return -EINVAL;

	if (copy_from_user(&assocparams, optval,
			sizeof (struct sctp_assocparams)))
		return -EFAULT;

	asoc = sctp_id2assoc(sk, assocparams.sasoc_assoc_id);

	if (!asoc && assocparams.sasoc_assoc_id && sctp_style(sk, UDP))
		return -EINVAL;

	/* Values correspoinding to the specific association */
	if (asoc) {
		assocparams.sasoc_asocmaxrxt = asoc->max_retrans;
		assocparams.sasoc_peer_rwnd = asoc->peer.rwnd;
		assocparams.sasoc_local_rwnd = asoc->a_rwnd;
		assocparams.sasoc_cookie_life = (asoc->cookie_life.tv_sec
						* 1000) +
						(asoc->cookie_life.tv_usec
						/ 1000);

		list_for_each(pos, &asoc->peer.transport_addr_list) {
			cnt ++;
		}

		assocparams.sasoc_number_peer_destinations = cnt;
	} else {
		/* Values corresponding to the endpoint */
		struct sctp_sock *sp = sctp_sk(sk);

		assocparams.sasoc_asocmaxrxt = sp->assocparams.sasoc_asocmaxrxt;
		assocparams.sasoc_peer_rwnd = sp->assocparams.sasoc_peer_rwnd;
		assocparams.sasoc_local_rwnd = sp->assocparams.sasoc_local_rwnd;
		assocparams.sasoc_cookie_life =
					sp->assocparams.sasoc_cookie_life;
		assocparams.sasoc_number_peer_destinations =
					sp->assocparams.
					sasoc_number_peer_destinations;
	}

	if (put_user(len, optlen))
		return -EFAULT;

	if (copy_to_user(optval, &assocparams, len))
		return -EFAULT;

	return 0;
}

/*
 * 7.1.16 Set/clear IPv4 mapped addresses (SCTP_I_WANT_MAPPED_V4_ADDR)
 *
 * This socket option is a boolean flag which turns on or off mapped V4
 * addresses.  If this option is turned on and the socket is type
 * PF_INET6, then IPv4 addresses will be mapped to V6 representation.
 * If this option is turned off, then no mapping will be done of V4
 * addresses and a user will receive both PF_INET6 and PF_INET type
 * addresses on the socket.
 */
static int sctp_getsockopt_mappedv4(struct sock *sk, int len,
				    char __user *optval, int __user *optlen)
{
	int val;
	struct sctp_sock *sp = sctp_sk(sk);

	if (len < sizeof(int))
		return -EINVAL;

	len = sizeof(int);
	val = sp->v4mapped;
	if (put_user(len, optlen))
		return -EFAULT;
	if (copy_to_user(optval, &val, len))
		return -EFAULT;

	return 0;
}

/*
 * 7.1.17 Set the maximum fragrmentation size (SCTP_MAXSEG)
 *
 * This socket option specifies the maximum size to put in any outgoing
 * SCTP chunk.  If a message is larger than this size it will be
 * fragmented by SCTP into the specified size.  Note that the underlying
 * SCTP implementation may fragment into smaller sized chunks when the
 * PMTU of the underlying association is smaller than the value set by
 * the user.
 */
static int sctp_getsockopt_maxseg(struct sock *sk, int len,
				  char __user *optval, int __user *optlen)
{
	int val;

	if (len < sizeof(int))
		return -EINVAL;

	len = sizeof(int);

	val = sctp_sk(sk)->user_frag;
	if (put_user(len, optlen))
		return -EFAULT;
	if (copy_to_user(optval, &val, len))
		return -EFAULT;

	return 0;
}

SCTP_STATIC int sctp_getsockopt(struct sock *sk, int level, int optname,
				char __user *optval, int __user *optlen)
{
	int retval = 0;
	int len;

	SCTP_DEBUG_PRINTK("sctp_getsockopt(sk: %p... optname: %d)\n",
			  sk, optname);

	/* I can hardly begin to describe how wrong this is.  This is
	 * so broken as to be worse than useless.  The API draft
	 * REALLY is NOT helpful here...  I am not convinced that the
	 * semantics of getsockopt() with a level OTHER THAN SOL_SCTP
	 * are at all well-founded.
	 */
	if (level != SOL_SCTP) {
		struct sctp_af *af = sctp_sk(sk)->pf->af;

		retval = af->getsockopt(sk, level, optname, optval, optlen);
		return retval;
	}

	if (get_user(len, optlen))
		return -EFAULT;

	sctp_lock_sock(sk);

	switch (optname) {
	case SCTP_STATUS:
		retval = sctp_getsockopt_sctp_status(sk, len, optval, optlen);
		break;
	case SCTP_DISABLE_FRAGMENTS:
		retval = sctp_getsockopt_disable_fragments(sk, len, optval,
							   optlen);
		break;
	case SCTP_EVENTS:
		retval = sctp_getsockopt_events(sk, len, optval, optlen);
		break;
	case SCTP_AUTOCLOSE:
		retval = sctp_getsockopt_autoclose(sk, len, optval, optlen);
		break;
	case SCTP_SOCKOPT_PEELOFF:
		retval = sctp_getsockopt_peeloff(sk, len, optval, optlen);
		break;
	case SCTP_PEER_ADDR_PARAMS:
		retval = sctp_getsockopt_peer_addr_params(sk, len, optval,
							  optlen);
		break;
	case SCTP_INITMSG:
		retval = sctp_getsockopt_initmsg(sk, len, optval, optlen);
		break;
	case SCTP_GET_PEER_ADDRS_NUM_OLD:
		retval = sctp_getsockopt_peer_addrs_num_old(sk, len, optval,
							    optlen);
		break;
	case SCTP_GET_LOCAL_ADDRS_NUM_OLD:
		retval = sctp_getsockopt_local_addrs_num_old(sk, len, optval,
							     optlen);
		break;
	case SCTP_GET_PEER_ADDRS_OLD:
		retval = sctp_getsockopt_peer_addrs_old(sk, len, optval,
							optlen);
		break;
	case SCTP_GET_LOCAL_ADDRS_OLD:
		retval = sctp_getsockopt_local_addrs_old(sk, len, optval,
							 optlen);
		break;
	case SCTP_GET_PEER_ADDRS:
		retval = sctp_getsockopt_peer_addrs(sk, len, optval,
						    optlen);
		break;
	case SCTP_GET_LOCAL_ADDRS:
		retval = sctp_getsockopt_local_addrs(sk, len, optval,
						     optlen);
		break;
	case SCTP_DEFAULT_SEND_PARAM:
		retval = sctp_getsockopt_default_send_param(sk, len,
							    optval, optlen);
		break;
	case SCTP_PRIMARY_ADDR:
		retval = sctp_getsockopt_primary_addr(sk, len, optval, optlen);
		break;
	case SCTP_NODELAY:
		retval = sctp_getsockopt_nodelay(sk, len, optval, optlen);
		break;
	case SCTP_RTOINFO:
		retval = sctp_getsockopt_rtoinfo(sk, len, optval, optlen);
		break;
	case SCTP_ASSOCINFO:
		retval = sctp_getsockopt_associnfo(sk, len, optval, optlen);
		break;
	case SCTP_I_WANT_MAPPED_V4_ADDR:
		retval = sctp_getsockopt_mappedv4(sk, len, optval, optlen);
		break;
	case SCTP_MAXSEG:
		retval = sctp_getsockopt_maxseg(sk, len, optval, optlen);
		break;
	case SCTP_GET_PEER_ADDR_INFO:
		retval = sctp_getsockopt_peer_addr_info(sk, len, optval,
							optlen);
		break;
	case SCTP_ADAPTION_LAYER:
		retval = sctp_getsockopt_adaption_layer(sk, len, optval,
							optlen);
		break;
	default:
		retval = -ENOPROTOOPT;
		break;
	};

	sctp_release_sock(sk);
	return retval;
}

static void sctp_hash(struct sock *sk)
{
	/* STUB */
}

static void sctp_unhash(struct sock *sk)
{
	/* STUB */
}

/* Check if port is acceptable.  Possibly find first available port.
 *
 * The port hash table (contained in the 'global' SCTP protocol storage
 * returned by struct sctp_protocol *sctp_get_protocol()). The hash
 * table is an array of 4096 lists (sctp_bind_hashbucket). Each
 * list (the list number is the port number hashed out, so as you
 * would expect from a hash function, all the ports in a given list have
 * such a number that hashes out to the same list number; you were
 * expecting that, right?); so each list has a set of ports, with a
 * link to the socket (struct sock) that uses it, the port number and
 * a fastreuse flag (FIXME: NPI ipg).
 */
static struct sctp_bind_bucket *sctp_bucket_create(
	struct sctp_bind_hashbucket *head, unsigned short snum);

static long sctp_get_port_local(struct sock *sk, union sctp_addr *addr)
{
	struct sctp_bind_hashbucket *head; /* hash list */
	struct sctp_bind_bucket *pp; /* hash list port iterator */
	unsigned short snum;
	int ret;

	/* NOTE:  Remember to put this back to net order. */
	addr->v4.sin_port = ntohs(addr->v4.sin_port);
	snum = addr->v4.sin_port;

	SCTP_DEBUG_PRINTK("sctp_get_port() begins, snum=%d\n", snum);
	sctp_local_bh_disable();

	if (snum == 0) {
		/* Search for an available port.
		 *
		 * 'sctp_port_rover' was the last port assigned, so
		 * we start to search from 'sctp_port_rover +
		 * 1'. What we do is first check if port 'rover' is
		 * already in the hash table; if not, we use that; if
		 * it is, we try next.
		 */
		int low = sysctl_local_port_range[0];
		int high = sysctl_local_port_range[1];
		int remaining = (high - low) + 1;
		int rover;
		int index;

		sctp_spin_lock(&sctp_port_alloc_lock);
		rover = sctp_port_rover;
		do {
			rover++;
			if ((rover < low) || (rover > high))
				rover = low;
			index = sctp_phashfn(rover);
			head = &sctp_port_hashtable[index];
			sctp_spin_lock(&head->lock);
			for (pp = head->chain; pp; pp = pp->next)
				if (pp->port == rover)
					goto next;
			break;
		next:
			sctp_spin_unlock(&head->lock);
		} while (--remaining > 0);
		sctp_port_rover = rover;
		sctp_spin_unlock(&sctp_port_alloc_lock);

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

		/* OK, here is the one we will use.  HEAD (the port
		 * hash table list entry) is non-NULL and we hold it's
		 * mutex.
		 */
		snum = rover;
	} else {
		/* We are given an specific port number; we verify
		 * that it is not being used. If it is used, we will
		 * exahust the search in the hash list corresponding
		 * to the port number (snum) - we detect that with the
		 * port iterator, pp being NULL.
		 */
		head = &sctp_port_hashtable[sctp_phashfn(snum)];
		sctp_spin_lock(&head->lock);
		for (pp = head->chain; pp; pp = pp->next) {
			if (pp->port == snum)
				goto pp_found;
		}
	}
	pp = NULL;
	goto pp_not_found;
pp_found:
	if (!hlist_empty(&pp->owner)) {
		/* We had a port hash table hit - there is an
		 * available port (pp != NULL) and it is being
		 * used by other socket (pp->owner not empty); that other
		 * socket is going to be sk2.
		 */
		int reuse = sk->sk_reuse;
		struct sock *sk2;
		struct hlist_node *node;

		SCTP_DEBUG_PRINTK("sctp_get_port() found a possible match\n");
		if (pp->fastreuse && sk->sk_reuse)
			goto success;

		/* Run through the list of sockets bound to the port
		 * (pp->port) [via the pointers bind_next and
		 * bind_pprev in the struct sock *sk2 (pp->sk)]. On each one,
		 * we get the endpoint they describe and run through
		 * the endpoint's list of IP (v4 or v6) addresses,
		 * comparing each of the addresses with the address of
		 * the socket sk. If we find a match, then that means
		 * that this port/socket (sk) combination are already
		 * in an endpoint.
		 */
		sk_for_each_bound(sk2, node, &pp->owner) {
			struct sctp_endpoint *ep2;
			ep2 = sctp_sk(sk2)->ep;

			if (reuse && sk2->sk_reuse)
				continue;

			if (sctp_bind_addr_match(&ep2->base.bind_addr, addr,
						 sctp_sk(sk))) {
				ret = (long)sk2;
				goto fail_unlock;
			}
		}
		SCTP_DEBUG_PRINTK("sctp_get_port(): Found a match\n");
	}
pp_not_found:
	/* If there was a hash table miss, create a new port.  */
	ret = 1;
	if (!pp && !(pp = sctp_bucket_create(head, snum)))
		goto fail_unlock;

	/* In either case (hit or miss), make sure fastreuse is 1 only
	 * if sk->sk_reuse is too (that is, if the caller requested
	 * SO_REUSEADDR on this socket -sk-).
	 */
	if (hlist_empty(&pp->owner))
		pp->fastreuse = sk->sk_reuse ? 1 : 0;
	else if (pp->fastreuse && !sk->sk_reuse)
		pp->fastreuse = 0;

	/* We are set, so fill up all the data in the hash table
	 * entry, tie the socket list information with the rest of the
	 * sockets FIXME: Blurry, NPI (ipg).
	 */
success:
	inet_sk(sk)->num = snum;
	if (!sctp_sk(sk)->bind_hash) {
		sk_add_bind_node(sk, &pp->owner);
		sctp_sk(sk)->bind_hash = pp;
	}
	ret = 0;

fail_unlock:
	sctp_spin_unlock(&head->lock);

fail:
	sctp_local_bh_enable();
	addr->v4.sin_port = htons(addr->v4.sin_port);
	return ret;
}

/* Assign a 'snum' port to the socket.  If snum == 0, an ephemeral
 * port is requested.
 */
static int sctp_get_port(struct sock *sk, unsigned short snum)
{
	long ret;
	union sctp_addr addr;
	struct sctp_af *af = sctp_sk(sk)->pf->af;

	/* Set up a dummy address struct from the sk. */
	af->from_sk(&addr, sk);
	addr.v4.sin_port = htons(snum);

	/* Note: sk->sk_num gets filled in if ephemeral port request. */
	ret = sctp_get_port_local(sk, &addr);

	return (ret ? 1 : 0);
}

/*
 * 3.1.3 listen() - UDP Style Syntax
 *
 *   By default, new associations are not accepted for UDP style sockets.
 *   An application uses listen() to mark a socket as being able to
 *   accept new associations.
 */
SCTP_STATIC int sctp_seqpacket_listen(struct sock *sk, int backlog)
{
	struct sctp_sock *sp = sctp_sk(sk);
	struct sctp_endpoint *ep = sp->ep;

	/* Only UDP style sockets that are not peeled off are allowed to
	 * listen().
	 */
	if (!sctp_style(sk, UDP))
		return -EINVAL;

	/* If backlog is zero, disable listening. */
	if (!backlog) {
		if (sctp_sstate(sk, CLOSED))
			return 0;
		
		sctp_unhash_endpoint(ep);
		sk->sk_state = SCTP_SS_CLOSED;
	}

	/* Return if we are already listening. */
	if (sctp_sstate(sk, LISTENING))
		return 0;
		
	/*
	 * If a bind() or sctp_bindx() is not called prior to a listen()
	 * call that allows new associations to be accepted, the system
	 * picks an ephemeral port and will choose an address set equivalent
	 * to binding with a wildcard address.
	 *
	 * This is not currently spelled out in the SCTP sockets
	 * extensions draft, but follows the practice as seen in TCP
	 * sockets.
	 */
	if (!ep->base.bind_addr.port) {
		if (sctp_autobind(sk))
			return -EAGAIN;
	}
	sk->sk_state = SCTP_SS_LISTENING;
	sctp_hash_endpoint(ep);
	return 0;
}

/*
 * 4.1.3 listen() - TCP Style Syntax
 *
 *   Applications uses listen() to ready the SCTP endpoint for accepting
 *   inbound associations.
 */
SCTP_STATIC int sctp_stream_listen(struct sock *sk, int backlog)
{
	struct sctp_sock *sp = sctp_sk(sk);
	struct sctp_endpoint *ep = sp->ep;

	/* If backlog is zero, disable listening. */
	if (!backlog) {
		if (sctp_sstate(sk, CLOSED))
			return 0;
		
		sctp_unhash_endpoint(ep);
		sk->sk_state = SCTP_SS_CLOSED;
	}

	if (sctp_sstate(sk, LISTENING))
		return 0;

	/*
	 * If a bind() or sctp_bindx() is not called prior to a listen()
	 * call that allows new associations to be accepted, the system
	 * picks an ephemeral port and will choose an address set equivalent
	 * to binding with a wildcard address.
	 *
	 * This is not currently spelled out in the SCTP sockets
	 * extensions draft, but follows the practice as seen in TCP
	 * sockets.
	 */
	if (!ep->base.bind_addr.port) {
		if (sctp_autobind(sk))
			return -EAGAIN;
	}
	sk->sk_state = SCTP_SS_LISTENING;
	sk->sk_max_ack_backlog = backlog;
	sctp_hash_endpoint(ep);
	return 0;
}

/*
 *  Move a socket to LISTENING state.
 */
int sctp_inet_listen(struct socket *sock, int backlog)
{
	struct sock *sk = sock->sk;
	struct crypto_tfm *tfm=NULL;
	int err = -EINVAL;

	if (unlikely(backlog < 0))
		goto out;

	sctp_lock_sock(sk);

	if (sock->state != SS_UNCONNECTED)
		goto out;

	/* Allocate HMAC for generating cookie. */
	if (sctp_hmac_alg) {
		tfm = sctp_crypto_alloc_tfm(sctp_hmac_alg, 0);
		if (!tfm) {
			err = -ENOSYS;
			goto out;
		}
	}

	switch (sock->type) {
	case SOCK_SEQPACKET:
		err = sctp_seqpacket_listen(sk, backlog);
		break;
	case SOCK_STREAM:
		err = sctp_stream_listen(sk, backlog);
		break;
	default:
		break;
	};
	if (err)
		goto cleanup;

	/* Store away the transform reference. */
	sctp_sk(sk)->hmac = tfm;
out:
	sctp_release_sock(sk);
	return err;
cleanup:
	sctp_crypto_free_tfm(tfm);
	goto out;
}

/*
 * This function is done by modeling the current datagram_poll() and the
 * tcp_poll().  Note that, based on these implementations, we don't
 * lock the socket in this function, even though it seems that,
 * ideally, locking or some other mechanisms can be used to ensure
 * the integrity of the counters (sndbuf and wmem_alloc) used
 * in this place.  We assume that we don't need locks either until proven
 * otherwise.
 *
 * Another thing to note is that we include the Async I/O support
 * here, again, by modeling the current TCP/UDP code.  We don't have
 * a good way to test with it yet.
 */
unsigned int sctp_poll(struct file *file, struct socket *sock, poll_table *wait)
{
	struct sock *sk = sock->sk;
	struct sctp_sock *sp = sctp_sk(sk);
	unsigned int mask;

	poll_wait(file, sk->sk_sleep, wait);

	/* A TCP-style listening socket becomes readable when the accept queue
	 * is not empty.
	 */
	if (sctp_style(sk, TCP) && sctp_sstate(sk, LISTENING))
		return (!list_empty(&sp->ep->asocs)) ?
		       	(POLLIN | POLLRDNORM) : 0;

	mask = 0;

	/* Is there any exceptional events?  */
	if (sk->sk_err || !skb_queue_empty(&sk->sk_error_queue))
		mask |= POLLERR;
	if (sk->sk_shutdown == SHUTDOWN_MASK)
		mask |= POLLHUP;

	/* Is it readable?  Reconsider this code with TCP-style support.  */
	if (!skb_queue_empty(&sk->sk_receive_queue) ||
	    (sk->sk_shutdown & RCV_SHUTDOWN))
		mask |= POLLIN | POLLRDNORM;

	/* The association is either gone or not ready.  */
	if (!sctp_style(sk, UDP) && sctp_sstate(sk, CLOSED))
		return mask;

	/* Is it writable?  */
	if (sctp_writeable(sk)) {
		mask |= POLLOUT | POLLWRNORM;
	} else {
		set_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags);
		/*
		 * Since the socket is not locked, the buffer
		 * might be made available after the writeable check and
		 * before the bit is set.  This could cause a lost I/O
		 * signal.  tcp_poll() has a race breaker for this race
		 * condition.  Based on their implementation, we put
		 * in the following code to cover it as well.
		 */
		if (sctp_writeable(sk))
			mask |= POLLOUT | POLLWRNORM;
	}
	return mask;
}

/********************************************************************
 * 2nd Level Abstractions
 ********************************************************************/

static struct sctp_bind_bucket *sctp_bucket_create(
	struct sctp_bind_hashbucket *head, unsigned short snum)
{
	struct sctp_bind_bucket *pp;

	pp = kmem_cache_alloc(sctp_bucket_cachep, SLAB_ATOMIC);
	SCTP_DBG_OBJCNT_INC(bind_bucket);
	if (pp) {
		pp->port = snum;
		pp->fastreuse = 0;
		INIT_HLIST_HEAD(&pp->owner);
		if ((pp->next = head->chain) != NULL)
			pp->next->pprev = &pp->next;
		head->chain = pp;
		pp->pprev = &head->chain;
	}
	return pp;
}

/* Caller must hold hashbucket lock for this tb with local BH disabled */
static void sctp_bucket_destroy(struct sctp_bind_bucket *pp)
{
	if (hlist_empty(&pp->owner)) {
		if (pp->next)
			pp->next->pprev = pp->pprev;
		*(pp->pprev) = pp->next;
		kmem_cache_free(sctp_bucket_cachep, pp);
		SCTP_DBG_OBJCNT_DEC(bind_bucket);
	}
}

/* Release this socket's reference to a local port.  */
static inline void __sctp_put_port(struct sock *sk)
{
	struct sctp_bind_hashbucket *head =
		&sctp_port_hashtable[sctp_phashfn(inet_sk(sk)->num)];
	struct sctp_bind_bucket *pp;

	sctp_spin_lock(&head->lock);
	pp = sctp_sk(sk)->bind_hash;
	__sk_del_bind_node(sk);
	sctp_sk(sk)->bind_hash = NULL;
	inet_sk(sk)->num = 0;
	sctp_bucket_destroy(pp);
	sctp_spin_unlock(&head->lock);
}

void sctp_put_port(struct sock *sk)
{
	sctp_local_bh_disable();
	__sctp_put_port(sk);
	sctp_local_bh_enable();
}

/*
 * The system picks an ephemeral port and choose an address set equivalent
 * to binding with a wildcard address.
 * One of those addresses will be the primary address for the association.
 * This automatically enables the multihoming capability of SCTP.
 */
static int sctp_autobind(struct sock *sk)
{
	union sctp_addr autoaddr;
	struct sctp_af *af;
	unsigned short port;

	/* Initialize a local sockaddr structure to INADDR_ANY. */
	af = sctp_sk(sk)->pf->af;

	port = htons(inet_sk(sk)->num);
	af->inaddr_any(&autoaddr, port);

	return sctp_do_bind(sk, &autoaddr, af->sockaddr_len);
}

/* Parse out IPPROTO_SCTP CMSG headers.  Perform only minimal validation.
 *
 * From RFC 2292
 * 4.2 The cmsghdr Structure *
 *
 * When ancillary data is sent or received, any number of ancillary data
 * objects can be specified by the msg_control and msg_controllen members of
 * the msghdr structure, because each object is preceded by
 * a cmsghdr structure defining the object's length (the cmsg_len member).
 * Historically Berkeley-derived implementations have passed only one object
 * at a time, but this API allows multiple objects to be
 * passed in a single call to sendmsg() or recvmsg(). The following example
 * shows two ancillary data objects in a control buffer.
 *
 *   |<--------------------------- msg_controllen -------------------------->|
 *   |                                                                       |
 *
 *   |<----- ancillary data object ----->|<----- ancillary data object ----->|
 *
 *   |<---------- CMSG_SPACE() --------->|<---------- CMSG_SPACE() --------->|
 *   |                                   |                                   |
 *
 *   |<---------- cmsg_len ---------->|  |<--------- cmsg_len ----------->|  |
 *
 *   |<--------- CMSG_LEN() --------->|  |<-------- CMSG_LEN() ---------->|  |
 *   |                                |  |                                |  |
 *
 *   +-----+-----+-----+--+-----------+--+-----+-----+-----+--+-----------+--+
 *   |cmsg_|cmsg_|cmsg_|XX|           |XX|cmsg_|cmsg_|cmsg_|XX|           |XX|
 *
 *   |len  |level|type |XX|cmsg_data[]|XX|len  |level|type |XX|cmsg_data[]|XX|
 *
 *   +-----+-----+-----+--+-----------+--+-----+-----+-----+--+-----------+--+
 *    ^
 *    |
 *
 * msg_control
 * points here
 */
SCTP_STATIC int sctp_msghdr_parse(const struct msghdr *msg,
				  sctp_cmsgs_t *cmsgs)
{
	struct cmsghdr *cmsg;

	for (cmsg = CMSG_FIRSTHDR(msg);
	     cmsg != NULL;
	     cmsg = CMSG_NXTHDR((struct msghdr*)msg, cmsg)) {
		if (!CMSG_OK(msg, cmsg))
			return -EINVAL;

		/* Should we parse this header or ignore?  */
		if (cmsg->cmsg_level != IPPROTO_SCTP)
			continue;

		/* Strictly check lengths following example in SCM code.  */
		switch (cmsg->cmsg_type) {
		case SCTP_INIT:
			/* SCTP Socket API Extension
			 * 5.2.1 SCTP Initiation Structure (SCTP_INIT)
			 *
			 * This cmsghdr structure provides information for
			 * initializing new SCTP associations with sendmsg().
			 * The SCTP_INITMSG socket option uses this same data
			 * structure.  This structure is not used for
			 * recvmsg().
			 *
			 * cmsg_level    cmsg_type      cmsg_data[]
			 * ------------  ------------   ----------------------
			 * IPPROTO_SCTP  SCTP_INIT      struct sctp_initmsg
			 */
			if (cmsg->cmsg_len !=
			    CMSG_LEN(sizeof(struct sctp_initmsg)))
				return -EINVAL;
			cmsgs->init = (struct sctp_initmsg *)CMSG_DATA(cmsg);
			break;

		case SCTP_SNDRCV:
			/* SCTP Socket API Extension
			 * 5.2.2 SCTP Header Information Structure(SCTP_SNDRCV)
			 *
			 * This cmsghdr structure specifies SCTP options for
			 * sendmsg() and describes SCTP header information
			 * about a received message through recvmsg().
			 *
			 * cmsg_level    cmsg_type      cmsg_data[]
			 * ------------  ------------   ----------------------
			 * IPPROTO_SCTP  SCTP_SNDRCV    struct sctp_sndrcvinfo
			 */
			if (cmsg->cmsg_len !=
			    CMSG_LEN(sizeof(struct sctp_sndrcvinfo)))
				return -EINVAL;

			cmsgs->info =
				(struct sctp_sndrcvinfo *)CMSG_DATA(cmsg);

			/* Minimally, validate the sinfo_flags. */
			if (cmsgs->info->sinfo_flags &
			    ~(SCTP_UNORDERED | SCTP_ADDR_OVER |
			      SCTP_ABORT | SCTP_EOF))
				return -EINVAL;
			break;

		default:
			return -EINVAL;
		};
	}
	return 0;
}

/*
 * Wait for a packet..
 * Note: This function is the same function as in core/datagram.c
 * with a few modifications to make lksctp work.
 */
static int sctp_wait_for_packet(struct sock * sk, int *err, long *timeo_p)
{
	int error;
	DEFINE_WAIT(wait);

	prepare_to_wait_exclusive(sk->sk_sleep, &wait, TASK_INTERRUPTIBLE);

	/* Socket errors? */
	error = sock_error(sk);
	if (error)
		goto out;

	if (!skb_queue_empty(&sk->sk_receive_queue))
		goto ready;

	/* Socket shut down?  */
	if (sk->sk_shutdown & RCV_SHUTDOWN)
		goto out;

	/* Sequenced packets can come disconnected.  If so we report the
	 * problem.
	 */
	error = -ENOTCONN;

	/* Is there a good reason to think that we may receive some data?  */
	if (list_empty(&sctp_sk(sk)->ep->asocs) && !sctp_sstate(sk, LISTENING))
		goto out;

	/* Handle signals.  */
	if (signal_pending(current))
		goto interrupted;

	/* Let another process have a go.  Since we are going to sleep
	 * anyway.  Note: This may cause odd behaviors if the message
	 * does not fit in the user's buffer, but this seems to be the
	 * only way to honor MSG_DONTWAIT realistically.
	 */
	sctp_release_sock(sk);
	*timeo_p = schedule_timeout(*timeo_p);
	sctp_lock_sock(sk);

ready:
	finish_wait(sk->sk_sleep, &wait);
	return 0;

interrupted:
	error = sock_intr_errno(*timeo_p);

out:
	finish_wait(sk->sk_sleep, &wait);
	*err = error;
	return error;
}

/* Receive a datagram.
 * Note: This is pretty much the same routine as in core/datagram.c
 * with a few changes to make lksctp work.
 */
static struct sk_buff *sctp_skb_recv_datagram(struct sock *sk, int flags,
					      int noblock, int *err)
{
	int error;
	struct sk_buff *skb;
	long timeo;