socket.c

linux 内核源代码

		if (likely(res != -ELINKCONG)) {
exit:
			up(&tsock->sem);
			return res;
		}
		if (m->msg_flags & MSG_DONTWAIT) {
			res = -EWOULDBLOCK;
			goto exit;
		}
		if (wait_event_interruptible(*sock->sk->sk_sleep,
					     !tsock->p->congested)) {
		    res = -ERESTARTSYS;
		    goto exit;
		}
	} while (1);
}

/**
 * send_stream - send stream-oriented data
 * @iocb: (unused)
 * @sock: socket structure
 * @m: data to send
 * @total_len: total length of data to be sent
 *
 * Used for SOCK_STREAM data.
 *
 * Returns the number of bytes sent on success (or partial success),
 * or errno if no data sent
 */


static int send_stream(struct kiocb *iocb, struct socket *sock,
		       struct msghdr *m, size_t total_len)
{
	struct tipc_port *tport;
	struct msghdr my_msg;
	struct iovec my_iov;
	struct iovec *curr_iov;
	int curr_iovlen;
	char __user *curr_start;
	u32 hdr_size;
	int curr_left;
	int bytes_to_send;
	int bytes_sent;
	int res;

	/* Handle special cases where there is no connection */

	if (unlikely(sock->state != SS_CONNECTED)) {
		if (sock->state == SS_UNCONNECTED)
			return send_packet(iocb, sock, m, total_len);
		else if (sock->state == SS_DISCONNECTING)
			return -EPIPE;
		else
			return -ENOTCONN;
	}

	if (unlikely(m->msg_name))
		return -EISCONN;

	/*
	 * Send each iovec entry using one or more messages
	 *
	 * Note: This algorithm is good for the most likely case
	 * (i.e. one large iovec entry), but could be improved to pass sets
	 * of small iovec entries into send_packet().
	 */

	curr_iov = m->msg_iov;
	curr_iovlen = m->msg_iovlen;
	my_msg.msg_iov = &my_iov;
	my_msg.msg_iovlen = 1;
	my_msg.msg_flags = m->msg_flags;
	my_msg.msg_name = NULL;
	bytes_sent = 0;

	tport = tipc_sk(sock->sk)->p;
	hdr_size = msg_hdr_sz(&tport->phdr);

	while (curr_iovlen--) {
		curr_start = curr_iov->iov_base;
		curr_left = curr_iov->iov_len;

		while (curr_left) {
			bytes_to_send = tport->max_pkt - hdr_size;
			if (bytes_to_send > TIPC_MAX_USER_MSG_SIZE)
				bytes_to_send = TIPC_MAX_USER_MSG_SIZE;
			if (curr_left < bytes_to_send)
				bytes_to_send = curr_left;
			my_iov.iov_base = curr_start;
			my_iov.iov_len = bytes_to_send;
			if ((res = send_packet(iocb, sock, &my_msg, 0)) < 0) {
				if (bytes_sent != 0)
					res = bytes_sent;
				return res;
			}
			curr_left -= bytes_to_send;
			curr_start += bytes_to_send;
			bytes_sent += bytes_to_send;
		}

		curr_iov++;
	}

	return bytes_sent;
}

/**
 * auto_connect - complete connection setup to a remote port
 * @sock: socket structure
 * @tsock: TIPC-specific socket structure
 * @msg: peer's response message
 *
 * Returns 0 on success, errno otherwise
 */

static int auto_connect(struct socket *sock, struct tipc_sock *tsock,
			struct tipc_msg *msg)
{
	struct tipc_portid peer;

	if (msg_errcode(msg)) {
		sock->state = SS_DISCONNECTING;
		return -ECONNREFUSED;
	}

	peer.ref = msg_origport(msg);
	peer.node = msg_orignode(msg);
	tipc_connect2port(tsock->p->ref, &peer);
	tipc_set_portimportance(tsock->p->ref, msg_importance(msg));
	sock->state = SS_CONNECTED;
	return 0;
}

/**
 * set_orig_addr - capture sender's address for received message
 * @m: descriptor for message info
 * @msg: received message header
 *
 * Note: Address is not captured if not requested by receiver.
 */

static void set_orig_addr(struct msghdr *m, struct tipc_msg *msg)
{
	struct sockaddr_tipc *addr = (struct sockaddr_tipc *)m->msg_name;

	if (addr) {
		addr->family = AF_TIPC;
		addr->addrtype = TIPC_ADDR_ID;
		addr->addr.id.ref = msg_origport(msg);
		addr->addr.id.node = msg_orignode(msg);
		addr->addr.name.domain = 0;   	/* could leave uninitialized */
		addr->scope = 0;   		/* could leave uninitialized */
		m->msg_namelen = sizeof(struct sockaddr_tipc);
	}
}

/**
 * anc_data_recv - optionally capture ancillary data for received message
 * @m: descriptor for message info
 * @msg: received message header
 * @tport: TIPC port associated with message
 *
 * Note: Ancillary data is not captured if not requested by receiver.
 *
 * Returns 0 if successful, otherwise errno
 */

static int anc_data_recv(struct msghdr *m, struct tipc_msg *msg,
				struct tipc_port *tport)
{
	u32 anc_data[3];
	u32 err;
	u32 dest_type;
	int has_name;
	int res;

	if (likely(m->msg_controllen == 0))
		return 0;

	/* Optionally capture errored message object(s) */

	err = msg ? msg_errcode(msg) : 0;
	if (unlikely(err)) {
		anc_data[0] = err;
		anc_data[1] = msg_data_sz(msg);
		if ((res = put_cmsg(m, SOL_TIPC, TIPC_ERRINFO, 8, anc_data)))
			return res;
		if (anc_data[1] &&
		    (res = put_cmsg(m, SOL_TIPC, TIPC_RETDATA, anc_data[1],
				    msg_data(msg))))
			return res;
	}

	/* Optionally capture message destination object */

	dest_type = msg ? msg_type(msg) : TIPC_DIRECT_MSG;
	switch (dest_type) {
	case TIPC_NAMED_MSG:
		has_name = 1;
		anc_data[0] = msg_nametype(msg);
		anc_data[1] = msg_namelower(msg);
		anc_data[2] = msg_namelower(msg);
		break;
	case TIPC_MCAST_MSG:
		has_name = 1;
		anc_data[0] = msg_nametype(msg);
		anc_data[1] = msg_namelower(msg);
		anc_data[2] = msg_nameupper(msg);
		break;
	case TIPC_CONN_MSG:
		has_name = (tport->conn_type != 0);
		anc_data[0] = tport->conn_type;
		anc_data[1] = tport->conn_instance;
		anc_data[2] = tport->conn_instance;
		break;
	default:
		has_name = 0;
	}
	if (has_name &&
	    (res = put_cmsg(m, SOL_TIPC, TIPC_DESTNAME, 12, anc_data)))
		return res;

	return 0;
}

/**
 * recv_msg - receive packet-oriented message
 * @iocb: (unused)
 * @m: descriptor for message info
 * @buf_len: total size of user buffer area
 * @flags: receive flags
 *
 * Used for SOCK_DGRAM, SOCK_RDM, and SOCK_SEQPACKET messages.
 * If the complete message doesn't fit in user area, truncate it.
 *
 * Returns size of returned message data, errno otherwise
 */

static int recv_msg(struct kiocb *iocb, struct socket *sock,
		    struct msghdr *m, size_t buf_len, int flags)
{
	struct tipc_sock *tsock = tipc_sk(sock->sk);
	struct sk_buff *buf;
	struct tipc_msg *msg;
	unsigned int q_len;
	unsigned int sz;
	u32 err;
	int res;

	/* Currently doesn't support receiving into multiple iovec entries */

	if (m->msg_iovlen != 1)
		return -EOPNOTSUPP;

	/* Catch invalid receive attempts */

	if (unlikely(!buf_len))
		return -EINVAL;

	if (sock->type == SOCK_SEQPACKET) {
		if (unlikely(sock->state == SS_UNCONNECTED))
			return -ENOTCONN;
		if (unlikely((sock->state == SS_DISCONNECTING) &&
			     (skb_queue_len(&sock->sk->sk_receive_queue) == 0)))
			return -ENOTCONN;
	}

	/* Look for a message in receive queue; wait if necessary */

	if (unlikely(down_interruptible(&tsock->sem)))
		return -ERESTARTSYS;

restart:
	if (unlikely((skb_queue_len(&sock->sk->sk_receive_queue) == 0) &&
		     (flags & MSG_DONTWAIT))) {
		res = -EWOULDBLOCK;
		goto exit;
	}

	if ((res = wait_event_interruptible(
		*sock->sk->sk_sleep,
		((q_len = skb_queue_len(&sock->sk->sk_receive_queue)) ||
		 (sock->state == SS_DISCONNECTING))) )) {
		goto exit;
	}

	/* Catch attempt to receive on an already terminated connection */
	/* [THIS CHECK MAY OVERLAP WITH AN EARLIER CHECK] */

	if (!q_len) {
		res = -ENOTCONN;
		goto exit;
	}

	/* Get access to first message in receive queue */

	buf = skb_peek(&sock->sk->sk_receive_queue);
	msg = buf_msg(buf);
	sz = msg_data_sz(msg);
	err = msg_errcode(msg);

	/* Complete connection setup for an implied connect */

	if (unlikely(sock->state == SS_CONNECTING)) {
		if ((res = auto_connect(sock, tsock, msg)))
			goto exit;
	}

	/* Discard an empty non-errored message & try again */

	if ((!sz) && (!err)) {
		advance_queue(tsock);
		goto restart;
	}

	/* Capture sender's address (optional) */

	set_orig_addr(m, msg);

	/* Capture ancillary data (optional) */

	if ((res = anc_data_recv(m, msg, tsock->p)))
		goto exit;

	/* Capture message data (if valid) & compute return value (always) */

	if (!err) {
		if (unlikely(buf_len < sz)) {
			sz = buf_len;
			m->msg_flags |= MSG_TRUNC;
		}
		if (unlikely(copy_to_user(m->msg_iov->iov_base, msg_data(msg),
					  sz))) {
			res = -EFAULT;
			goto exit;
		}
		res = sz;
	} else {
		if ((sock->state == SS_READY) ||
		    ((err == TIPC_CONN_SHUTDOWN) || m->msg_control))
			res = 0;
		else
			res = -ECONNRESET;
	}

	/* Consume received message (optional) */

	if (likely(!(flags & MSG_PEEK))) {
		if (unlikely(++tsock->p->conn_unacked >= TIPC_FLOW_CONTROL_WIN))
			tipc_acknowledge(tsock->p->ref, tsock->p->conn_unacked);
		advance_queue(tsock);
	}
exit:
	up(&tsock->sem);
	return res;
}

/**
 * recv_stream - receive stream-oriented data
 * @iocb: (unused)
 * @m: descriptor for message info
 * @buf_len: total size of user buffer area
 * @flags: receive flags
 *
 * Used for SOCK_STREAM messages only.  If not enough data is available
 * will optionally wait for more; never truncates data.
 *
 * Returns size of returned message data, errno otherwise
 */

static int recv_stream(struct kiocb *iocb, struct socket *sock,
		       struct msghdr *m, size_t buf_len, int flags)
{
	struct tipc_sock *tsock = tipc_sk(sock->sk);
	struct sk_buff *buf;
	struct tipc_msg *msg;
	unsigned int q_len;
	unsigned int sz;
	int sz_to_copy;
	int sz_copied = 0;
	int needed;
	char __user *crs = m->msg_iov->iov_base;
	unsigned char *buf_crs;
	u32 err;
	int res;

	/* Currently doesn't support receiving into multiple iovec entries */

	if (m->msg_iovlen != 1)
		return -EOPNOTSUPP;

	/* Catch invalid receive attempts */

	if (unlikely(!buf_len))
		return -EINVAL;

	if (unlikely(sock->state == SS_DISCONNECTING)) {
		if (skb_queue_len(&sock->sk->sk_receive_queue) == 0)
			return -ENOTCONN;
	} else if (unlikely(sock->state != SS_CONNECTED))
		return -ENOTCONN;

	/* Look for a message in receive queue; wait if necessary */

	if (unlikely(down_interruptible(&tsock->sem)))
		return -ERESTARTSYS;

restart:
	if (unlikely((skb_queue_len(&sock->sk->sk_receive_queue) == 0) &&
		     (flags & MSG_DONTWAIT))) {
		res = -EWOULDBLOCK;
		goto exit;
	}

	if ((res = wait_event_interruptible(
		*sock->sk->sk_sleep,
		((q_len = skb_queue_len(&sock->sk->sk_receive_queue)) ||
		 (sock->state == SS_DISCONNECTING))) )) {
		goto exit;
	}

	/* Catch attempt to receive on an already terminated connection */
	/* [THIS CHECK MAY OVERLAP WITH AN EARLIER CHECK] */

	if (!q_len) {
		res = -ENOTCONN;
		goto exit;
	}

	/* Get access to first message in receive queue */

	buf = skb_peek(&sock->sk->sk_receive_queue);
	msg = buf_msg(buf);
	sz = msg_data_sz(msg);
	err = msg_errcode(msg);

	/* Discard an empty non-errored message & try again */

	if ((!sz) && (!err)) {
		advance_queue(tsock);
		goto restart;
	}

	/* Optionally capture sender's address & ancillary data of first msg */

	if (sz_copied == 0) {
		set_orig_addr(m, msg);
		if ((res = anc_data_recv(m, msg, tsock->p)))
			goto exit;
	}

	/* Capture message data (if valid) & compute return value (always) */

	if (!err) {
		buf_crs = (unsigned char *)(TIPC_SKB_CB(buf)->handle);
		sz = skb_tail_pointer(buf) - buf_crs;

		needed = (buf_len - sz_copied);
		sz_to_copy = (sz <= needed) ? sz : needed;
		if (unlikely(copy_to_user(crs, buf_crs, sz_to_copy))) {
			res = -EFAULT;
			goto exit;
		}
		sz_copied += sz_to_copy;

		if (sz_to_copy < sz) {
			if (!(flags & MSG_PEEK))
				TIPC_SKB_CB(buf)->handle = buf_crs + sz_to_copy;
			goto exit;
		}

		crs += sz_to_copy;
	} else {
		if (sz_copied != 0)
			goto exit; /* can't add error msg to valid data */

		if ((err == TIPC_CONN_SHUTDOWN) || m->msg_control)
			res = 0;
		else
			res = -ECONNRESET;
	}

	/* Consume received message (optional) */

	if (likely(!(flags & MSG_PEEK))) {
		if (unlikely(++tsock->p->conn_unacked >= TIPC_FLOW_CONTROL_WIN))
			tipc_acknowledge(tsock->p->ref, tsock->p->conn_unacked);
		advance_queue(tsock);
	}

	/* Loop around if more data is required */

	if ((sz_copied < buf_len)    /* didn't get all requested data */
	    && (flags & MSG_WAITALL) /* ... and need to wait for more */
	    && (!(flags & MSG_PEEK)) /* ... and aren't just peeking at data */
	    && (!err)                /* ... and haven't reached a FIN */
	    )
		goto restart;

exit:
	up(&tsock->sem);
	return sz_copied ? sz_copied : res;
}

/**
 * queue_overloaded - test if queue overload condition exists
 * @queue_size: current size of queue
 * @base: nominal maximum size of queue
 * @msg: message to be added to queue
 *
 * Returns 1 if queue is currently overloaded, 0 otherwise
 */

static int queue_overloaded(u32 queue_size, u32 base, struct tipc_msg *msg)
{
	u32 threshold;
	u32 imp = msg_importance(msg);

	if (imp == TIPC_LOW_IMPORTANCE)
		threshold = base;
	else if (imp == TIPC_MEDIUM_IMPORTANCE)
		threshold = base * 2;
	else if (imp == TIPC_HIGH_IMPORTANCE)
		threshold = base * 100;
	else
		return 0;

	if (msg_connected(msg))
		threshold *= 4;

	return (queue_size > threshold);
}

/**
 * async_disconnect - wrapper function used to disconnect port
 * @portref: TIPC port reference (passed as pointer-sized value)
 */

static void async_disconnect(unsigned long portref)
{
	tipc_disconnect((u32)portref);
}

/**
 * dispatch - handle arriving message
 * @tport: TIPC port that received message
 * @buf: message
 *
 * Called with port locked.  Must not take socket lock to avoid deadlock risk.
 *
 * Returns TIPC error status code (TIPC_OK if message is not to be rejected)
 */

static u32 dispatch(struct tipc_port *tport, struct sk_buff *buf)
{
	struct tipc_msg *msg = buf_msg(buf);
	struct tipc_sock *tsock = (struct tipc_sock *)tport->usr_handle;
	struct socket *sock;
	u32 recv_q_len;

	/* Reject message if socket is closing */

	if (!tsock)
		return TIPC_ERR_NO_PORT;

	/* Reject message if it is wrong sort of message for socket */

	/*
	 * WOULD IT BE BETTER TO JUST DISCARD THESE MESSAGES INSTEAD?
	 * "NO PORT" ISN'T REALLY THE RIGHT ERROR CODE, AND THERE MAY
	 * BE SECURITY IMPLICATIONS INHERENT IN REJECTING INVALID TRAFFIC
	 */
	sock = tsock->sk.sk_socket;
	if (sock->state == SS_READY) {
		if (msg_connected(msg)) {
			msg_dbg(msg, "dispatch filter 1\n");
			return TIPC_ERR_NO_PORT;