socket.c

	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;

	snum = ntohs(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 &&
			sk->sk_state != SCTP_SS_LISTENING)
			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 &&
			    sk2->sk_state != SCTP_SS_LISTENING)
				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)) {
		if (sk->sk_reuse && sk->sk_state != SCTP_SS_LISTENING)
			pp->fastreuse = 1;
		else
			pp->fastreuse = 0;
	} else if (pp->fastreuse &&
		(!sk->sk_reuse || sk->sk_state == SCTP_SS_LISTENING))
		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:
	if (!sctp_sk(sk)->bind_hash) {
		inet_sk(sk)->num = snum;
		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();
	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.
	 *
	 * Additionally, turn off fastreuse flag since we are not listening
	 */
	sk->sk_state = SCTP_SS_LISTENING;
	if (!ep->base.bind_addr.port) {
		if (sctp_autobind(sk))
			return -EAGAIN;
	} else
		sctp_sk(sk)->bind_hash->fastreuse = 0;

	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.
	 */
	sk->sk_state = SCTP_SS_LISTENING;
	if (!ep->base.bind_addr.port) {
		if (sctp_autobind(sk))
			return -EAGAIN;
	} else
		sctp_sk(sk)->bind_hash->fastreuse = 0;

	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_hash *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 = crypto_alloc_hash(sctp_hmac_alg, 0, CRYPTO_ALG_ASYNC);
		if (IS_ERR(tfm)) {
			if (net_ratelimit()) {
				printk(KERN_INFO
				       "SCTP: failed to load transform for %s: %ld\n",
					sctp_hmac_alg, PTR_ERR(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:
	crypto_free_hash(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 & RCV_SHUTDOWN)
		mask |= POLLRDHUP;
	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, GFP_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 (pp && 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;
	__be16 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;

	timeo = sock_rcvtimeo(sk, noblock);

	SCTP_DEBUG_PRINTK("Timeout: timeo: %ld, MAX: %ld.\n",
			  timeo, MAX_SCHEDULE_TIMEOUT);

	do {
		/* Again only user level code calls this function,
		 * so nothing interrupt level
		 * will suddenly eat the receive_queue.
		 *
		 *  Look at current nfs client by the way...
		 *  However, this function was corrent in any case. 8)
		 */
		if (flags & MSG_PEEK) {
			spin_lock_bh(&sk->sk_receive_queue.lock);
			skb = skb_peek(&sk->sk_receive_queue);
			if (skb)
				atomic_inc(&skb->users);
			spin_unlock_bh(&sk->sk_receive_queue.lock);
		} else {
			skb = skb_dequeue(&sk->sk_receive_queue);
		}

		if (skb)
			return skb;

		/* Caller is allowed not to check sk->sk_err before calling. */
		error = sock_error(sk);
		if (error)
			goto no_packet;

		if (sk->sk_shutdown & RCV_SHUTDOWN)
			break;

		/* User doesn't want to wait.  */
		error = -EAGAIN;
		if (!timeo)
			goto no_packet;
	} while (sctp_wait_for_packet(sk, err, &timeo) == 0);

	return NULL;

no_packet:
	*err = error;
	return NULL;
}

/* If sndbuf has changed, wake up per association sndbuf waiters.  */
static void __sctp_write_space(struct sctp_association *asoc)
{
	struct sock *sk = asoc->base.sk;
	struct socket *sock = sk->sk_socket;

	if ((sctp_wspace(asoc) > 0) && sock) {
		if (waitqueue_active(&asoc->wait))
			wake_up_interruptible(&asoc->wait);

		if (sctp_writeable(sk)) {
			if (sk->sk_sleep && waitqueue_active(sk->sk_sleep))
				wake_up_interruptible(sk->sk_sleep);

			/* Note that we try to include the Async I/O support
			 * here by modeling from the current TCP/UDP code.
			 * We have not tested with it yet.
			 */
			if (sock->fasync_list &&
			    !(sk->sk_shutdown & SEND_SHUTDOWN))
				sock_wake_async(sock, 2, POLL_OUT);
		}
	}
}

/* Do accounting for the sndbuf space.
 * Decrement the used sndbuf space of the corresponding association by the
 * data size which was just transmitted(freed).
 */
static void sctp_wfree(struct sk_buff *skb)
{
	struct sctp_association *asoc;
	struct sctp_chunk *chunk;
	struct sock *sk;

	/* Get the saved chunk pointer.  */
	chunk = *((struct sctp_chunk **)(skb->cb));
	asoc = chunk->asoc;
	sk = asoc->base.sk;
	asoc->sndbuf_used -= SCTP_DATA_SNDSIZE(chunk) +
				sizeof(struct sk_buff) +
				sizeof(struct sctp_chunk);

	atomic_sub(sizeof(struct sctp_chunk), &sk->sk_wmem_alloc);

	sock_wfree(skb);
	__sctp_write_space(asoc);

	sctp_association_put(asoc);
}

/* Do accounting for the receive space on the socket.
 * Accounting for the association is done in ulpevent.c
 * We set this as a destructor for the cloned data skbs so that
 * accounting is done at the correct time.
 */
void sctp_sock_rfree(struct sk_buff *skb)
{
	struct sock *sk = skb->sk;
	struct sctp_ulpevent *event = sctp_skb2event(skb);

	atomic_sub(event->rmem_len, &sk->sk_rmem_alloc);
}


/* Helper function to wait for space in the sndbuf.  */
static int sctp_wait_for_sndbuf(struct sctp_association *asoc, long *timeo_p,
				size_t msg_len)
{
	struct sock *sk = asoc->base.sk;
	int err = 0;
	long current_timeo = *timeo_p;
	DEFINE_WAIT(wait);

	SCTP_DEBUG_PRINTK("wait_for_sndbuf: asoc=%p, timeo=%ld, msg_len=%zu\n",
			  asoc, (long)(*timeo_p), msg_len);

	/* Increment the association's refcnt.  */
	sctp_association_hold(asoc);

	/* Wait on the association specific sndbuf space. */
	for (;;) {
		prepare_to_wait_exclusive(&asoc->wait, &wait,
					  TASK_INTERRUPTIBLE);
		if (!*timeo_p)
			goto do_nonblock;
		if (sk->sk_err || asoc->state >= SCTP_STATE_SHUTDOWN_PENDING ||
		    asoc->base.dead)
			goto do_error;
		if (signal_pending(current))
			goto do_interrupted;
		if (msg_len <= sctp_wspace(asoc))
			break;

		/* Let another process have a go.  Since we are going
		 * to sleep anyway.
		 */
		sctp_release_sock(sk);
		current_timeo = schedule_timeout(current_timeo);
		BUG_ON(sk != asoc->base.sk);
		sctp_lock_sock(sk);

		*timeo_p = current_timeo;
	}

out:
	finish_wait(&asoc->wait, &wait);

	/* Release the association's refcnt.  */
	sctp_association_put(asoc);

	return err;

do_error:
	err = -EPIPE;
	goto out;

do_interrupted:
	err = sock_intr_errno(*timeo_p);
	goto out;

do_nonblock:
	err = -EAGAIN;
	goto out;
}

/* If socket sndbuf has changed, wake up all per association waiters.  */
void sctp_write_space(struct sock *sk)
{
	struct sctp_association *asoc;
	struct list_head *pos;

	/* Wake up the tasks in each wait queue.  */
	list_for_each(pos, &((sctp_sk(sk))->ep->asocs)) {
		asoc = list_entry(pos, struct sctp_association, asocs);
		__sctp_write_space(asoc);
	}
}

/* Is there any sndbuf space available on the socket?
 *
 * Note that sk_wmem_alloc is the sum of the send buffers on all of the
 * associations on the same socket.  For a UDP-style socket with
 * multiple associations, it is possible for it to be "unwriteable"
 * prematurely.  I assume that this is acceptable because
 * a premature "unwriteable" is better than an accidental "writeable" which
 * would cause an unwanted block under certain circumstances.  For the 1-1
 * UDP-style sockets or TCP-style sockets, this code should work.
 *  - Daisy
 */
static int sctp_writeable(struct sock *sk)
{
	int amt = 0;

	amt = sk->sk_sndbuf - atomic_read(&sk->sk_wmem_alloc);
	if (amt < 0)
		amt = 0;
	return amt;
}

/* Wait for an association to go into ESTABLISHED state. If timeout is 0,
 * returns immediately with EINPROGRESS.
 */
static int sctp_wait_for_connect(struct sctp_association *asoc, long *timeo_p)
{
	struct sock *sk = asoc->base.sk;
	int err = 0;
	long current_timeo = *timeo_p;
	DEFINE_WAIT(wait);

	SCTP_DEBUG_PRINTK("%s: asoc=%p, timeo=%ld\n", __FUNCTION__, asoc,
			  (long)(*timeo_p));

	/* Increment the association's refcnt.  */
	sctp_association_hold(asoc);

	for (;;) {
		prepare_to_wait_exclusive(&asoc->wait, &wait,
					  TASK_INTERRUPTIBLE);
		if (!*timeo_p)
			goto do_nonblock;
		if (sk->sk_shutdown & RCV_SHUTDOWN)
			break;
		if (sk->sk_err || asoc->state >= SCTP_STATE_SHUTDOWN_PENDING ||
		    asoc->base.dead)
			goto do_error;
		if (signal_pending(current))
			goto do_interrupted;

		if (sctp_state(asoc, ESTABLISHED))
			break;

		/* Let another process have a go.  Since we are going
		 * to sleep anyway.
		 */
		sctp_release_sock(sk);
		current_timeo = schedule_timeout(current_timeo);
		sctp_lock_sock(sk);

		*timeo_p = current_timeo;
	}

out:
	finish_wait(&asoc->wait, &wait);

	/* Release the association's refcnt.  */
	sctp_association_put(asoc);

	return err;

do_error:
	if (asoc->init_err_counter + 1 > asoc->max_init_attempts)
		err = -ETIMEDOUT;
	else
		err = -ECONNREFUSED;
	goto out;

do_interrupted:
	err = sock_intr_errno(*timeo_p);
	goto out;

do_nonblock:
	err = -EINPROGRESS;
	goto out;
}

static int sctp_wait_for_accept(struct sock *sk, long timeo)
{
	struct sctp_endpoint *ep;
	int err = 0;
	DEFINE_WAIT(wait);

	ep = sctp_sk(sk)->ep;


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

		if (list_empty(&ep->asocs)) {
			sctp_release_sock(sk);
			timeo = schedule_timeout(timeo);
			sctp_lock_sock(sk);
		}

		err = -EINVAL;
		if (!sctp_sstate(sk, LISTENING))
			break;

		err = 0;
		if (!list_empty(&ep->asocs))
			break;

		err = sock_intr_errno(timeo);
		if (signal_pending(current))
			break;

		err = -EAGAIN;
		if (!timeo)
			break;
	}

	finish_wait(sk->sk_sleep, &wait);

	return err;
}

void sctp_wait_for_close(struct sock *sk, long timeout)
{
	DEFINE_WAIT(wait);

	do {
		prepare_to_wait(sk->sk_sleep, &wait, TASK_INTERRUPTIBLE);
		if (list_empty(&sctp_sk(sk)->ep->asocs))
			break;
		sctp_release_sock(sk);
		timeout = schedule_timeout(timeout);
		sctp_lock_sock(sk);
	} while (!signal_pending(current) && timeout);

	finish_wait(sk->sk_sleep, &wait);
}

static void sctp_sock_rfree_frag(struct sk_buff *skb)
{
	struct sk_buff *frag;

	if (!skb->data_len)
		goto done;

	/* Don't forget the fragments. */
	for (frag = skb_shinfo(skb)->frag_list; frag; frag = frag->next)
		sctp_sock_rfree_frag(frag);

done:
	sctp_sock_rfree(skb);
}

static void sctp_skb_set_owner_r_frag(struct sk_buff *skb, struct sock *sk)
{
	struct sk_buff *frag;