svcsock.c

Linux内核源代码 为压缩文件 是<<Linux内核>>一书中的源代码

	struct proto_ops *ops;
	struct svc_sock	*newsvsk;
	int		err, slen;

	dprintk("svc: tcp_accept %p sock %p\n", svsk, sock);
	if (!sock)
		return;

	if (!(newsock = sock_alloc())) {
		printk(KERN_WARNING "%s: no more sockets!\n", serv->sv_name);
		return;
	}
	dprintk("svc: tcp_accept %p allocated\n", newsock);

	newsock->type = sock->type;
	newsock->ops = ops = sock->ops;

	if ((err = ops->accept(sock, newsock, O_NONBLOCK)) < 0) {
		if (net_ratelimit())
			printk(KERN_WARNING "%s: accept failed (err %d)!\n",
				   serv->sv_name, -err);
		goto failed;		/* aborted connection or whatever */
	}

	slen = sizeof(sin);
	err = ops->getname(newsock, (struct sockaddr *) &sin, &slen, 1);
	if (err < 0) {
		if (net_ratelimit())
			printk(KERN_WARNING "%s: peername failed (err %d)!\n",
				   serv->sv_name, -err);
		goto failed;		/* aborted connection or whatever */
	}

	/* Ideally, we would want to reject connections from unauthorized
	 * hosts here, but when we get encription, the IP of the host won't
	 * tell us anything. For now just warn about unpriv connections.
	 */
	if (ntohs(sin.sin_port) >= 1024) {
		if (net_ratelimit())
			printk(KERN_WARNING
				   "%s: connect from unprivileged port: %u.%u.%u.%u:%d\n",
				   serv->sv_name, 
				   NIPQUAD(sin.sin_addr.s_addr), ntohs(sin.sin_port));
	}

	dprintk("%s: connect from %u.%u.%u.%u:%04x\n", serv->sv_name,
			NIPQUAD(sin.sin_addr.s_addr), ntohs(sin.sin_port));

	if (!(newsvsk = svc_setup_socket(serv, newsock, &err, 0)))
		goto failed;

	/* Precharge. Data may have arrived on the socket before we
	 * installed the data_ready callback. 
	 */
	spin_lock_bh(&newsvsk->sk_lock);
	newsvsk->sk_data = 1;
	newsvsk->sk_temp = 1;
	svc_sock_enqueue(newsvsk);
	spin_unlock_bh(&newsvsk->sk_lock);

	if (serv->sv_stats)
		serv->sv_stats->nettcpconn++;

	return;

failed:
	sock_release(newsock);
	return;
}

/*
 * Receive data from a TCP socket.
 */
static int
svc_tcp_recvfrom(struct svc_rqst *rqstp)
{
	struct svc_sock	*svsk = rqstp->rq_sock;
	struct svc_serv	*serv = svsk->sk_server;
	struct svc_buf	*bufp = &rqstp->rq_argbuf;
	int		len, ready, used;

	dprintk("svc: tcp_recv %p data %d conn %d close %d\n",
			svsk, svsk->sk_data, svsk->sk_conn, svsk->sk_close);

	if (svsk->sk_close) {
		svc_delete_socket(svsk);
		return 0;
	}

	if (svsk->sk_conn) {
		svc_tcp_accept(svsk);
		svc_sock_accepted(svsk);
		return 0;
	}

	ready = svsk->sk_data;

	/* Receive data. If we haven't got the record length yet, get
	 * the next four bytes. Otherwise try to gobble up as much as
	 * possible up to the complete record length.
	 */
	if (svsk->sk_tcplen < 4) {
		unsigned long	want = 4 - svsk->sk_tcplen;
		struct iovec	iov;

		iov.iov_base = ((char *) &svsk->sk_reclen) + svsk->sk_tcplen;
		iov.iov_len  = want;
		if ((len = svc_recvfrom(rqstp, &iov, 1, want)) < 0)
			goto error;
		svsk->sk_tcplen += len;

		svsk->sk_reclen = ntohl(svsk->sk_reclen);
		if (!(svsk->sk_reclen & 0x80000000)) {
			/* FIXME: technically, a record can be fragmented,
			 *  and non-terminal fragments will not have the top
			 *  bit set in the fragment length header.
			 *  But apparently no known nfs clients send fragmented
			 *  records. */
			/* FIXME: shutdown socket */
			printk(KERN_NOTICE "RPC: bad TCP reclen %08lx",
			       (unsigned long) svsk->sk_reclen);
			return -EIO;
		}
		svsk->sk_reclen &= 0x7fffffff;
		dprintk("svc: TCP record, %d bytes\n", svsk->sk_reclen);
	}

	/* Check whether enough data is available */
	len = svc_recv_available(svsk);
	if (len < 0)
		goto error;

	if (len < svsk->sk_reclen) {
		/* FIXME: if sk_reclen > window-size, then we will
		 * never be able to receive the record, so should
		 * shutdown the connection
		 */
		dprintk("svc: incomplete TCP record (%d of %d)\n",
			len, svsk->sk_reclen);
		svc_sock_received(svsk, ready);
		return -EAGAIN;	/* record not complete */
	}
	/* if we think there is only one more record to read, but
	 * it is bigger than we expect, then two records must have arrived
	 * together, so pretend we aren't using the record.. */
	if (len > svsk->sk_reclen && ready == 1)
		used = 0;
	else	used = 1;

	/* Frob argbuf */
	bufp->iov[0].iov_base += 4;
	bufp->iov[0].iov_len  -= 4;

	/* Now receive data */
	len = svc_recvfrom(rqstp, bufp->iov, bufp->nriov, svsk->sk_reclen);
	if (len < 0)
		goto error;

	dprintk("svc: TCP complete record (%d bytes)\n", len);

	/* Position reply write pointer immediately after
	 * record length */
	rqstp->rq_resbuf.buf += 1;
	rqstp->rq_resbuf.len  = 1;

	rqstp->rq_skbuff      = 0;
	rqstp->rq_argbuf.buf += 1;
	rqstp->rq_argbuf.len  = (len >> 2);
	rqstp->rq_prot	      = IPPROTO_TCP;

	/* Reset TCP read info */
	svsk->sk_reclen = 0;
	svsk->sk_tcplen = 0;

	svc_sock_received(svsk, used);
	if (serv->sv_stats)
		serv->sv_stats->nettcpcnt++;

	return len;

error:
	if (len == -EAGAIN) {
		dprintk("RPC: TCP recvfrom got EAGAIN\n");
		svc_sock_received(svsk, ready); /* Clear data ready */
	} else {
		printk(KERN_NOTICE "%s: recvfrom returned errno %d\n",
					svsk->sk_server->sv_name, -len);
		svc_sock_received(svsk, 0);
	}

	return len;
}

/*
 * Send out data on TCP socket.
 * FIXME: Make the sendto call non-blocking in order not to hang
 * a daemon on a dead client. Requires write queue maintenance.
 */
static int
svc_tcp_sendto(struct svc_rqst *rqstp)
{
	struct svc_buf	*bufp = &rqstp->rq_resbuf;
	int sent;

	/* Set up the first element of the reply iovec.
	 * Any other iovecs that may be in use have been taken
	 * care of by the server implementation itself.
	 */
	bufp->iov[0].iov_base = bufp->base;
	bufp->iov[0].iov_len  = bufp->len << 2;
	bufp->base[0] = htonl(0x80000000|((bufp->len << 2) - 4));

	sent = svc_sendto(rqstp, bufp->iov, bufp->nriov);
	if (sent != bufp->len<<2) {
		printk(KERN_NOTICE "rpc-srv/tcp: %s: sent only %d bytes of %d - should shutdown socket\n",
		       rqstp->rq_sock->sk_server->sv_name,
		       sent, bufp->len << 2);
		/* FIXME: should shutdown the socket, or allocate more memort
		 * or wait and try again or something.  Otherwise
		 * client will get confused
		 */
	}
	return sent;
}

static int
svc_tcp_init(struct svc_sock *svsk)
{
	struct sock	*sk = svsk->sk_sk;

	svsk->sk_recvfrom = svc_tcp_recvfrom;
	svsk->sk_sendto = svc_tcp_sendto;

	if (sk->state == TCP_LISTEN) {
		dprintk("setting up TCP socket for listening\n");
		sk->state_change = svc_tcp_state_change1;
	} else {
		dprintk("setting up TCP socket for reading\n");
		sk->state_change = svc_tcp_state_change2;
		sk->data_ready = svc_tcp_data_ready;

		svsk->sk_reclen = 0;
		svsk->sk_tcplen = 0;
	}

	return 0;
}

/*
 * Receive the next request on any socket.
 */
int
svc_recv(struct svc_serv *serv, struct svc_rqst *rqstp, long timeout)
{
	struct svc_sock		*svsk;
	int			len;
	DECLARE_WAITQUEUE(wait, current);

	dprintk("svc: server %p waiting for data (to = %ld)\n",
		rqstp, timeout);

	if (rqstp->rq_sock)
		printk(KERN_ERR 
			"svc_recv: service %p, socket not NULL!\n",
			 rqstp);
	if (waitqueue_active(&rqstp->rq_wait))
		printk(KERN_ERR 
			"svc_recv: service %p, wait queue active!\n",
			 rqstp);

	/* Initialize the buffers */
	rqstp->rq_argbuf = rqstp->rq_defbuf;
	rqstp->rq_resbuf = rqstp->rq_defbuf;

	if (signalled())
		return -EINTR;

	spin_lock_bh(&serv->sv_lock);
	if ((svsk = svc_sock_dequeue(serv)) != NULL) {
		rqstp->rq_sock = svsk;
		svsk->sk_inuse++;
	} else {
		/* No data pending. Go to sleep */
		svc_serv_enqueue(serv, rqstp);

		/*
		 * We have to be able to interrupt this wait
		 * to bring down the daemons ...
		 */
		set_current_state(TASK_INTERRUPTIBLE);
		add_wait_queue(&rqstp->rq_wait, &wait);
		spin_unlock_bh(&serv->sv_lock);

		schedule_timeout(timeout);

		spin_lock_bh(&serv->sv_lock);
		remove_wait_queue(&rqstp->rq_wait, &wait);

		if (!(svsk = rqstp->rq_sock)) {
			svc_serv_dequeue(serv, rqstp);
			spin_unlock_bh(&serv->sv_lock);
			dprintk("svc: server %p, no data yet\n", rqstp);
			return signalled()? -EINTR : -EAGAIN;
		}
	}
	spin_unlock_bh(&serv->sv_lock);

	dprintk("svc: server %p, socket %p, inuse=%d\n",
		 rqstp, svsk, svsk->sk_inuse);
	len = svsk->sk_recvfrom(rqstp);
	dprintk("svc: got len=%d\n", len);

	/* No data, incomplete (TCP) read, or accept() */
	if (len == 0 || len == -EAGAIN) {
		svc_sock_release(rqstp);
		return -EAGAIN;
	}

	rqstp->rq_secure  = ntohs(rqstp->rq_addr.sin_port) < 1024;
	rqstp->rq_userset = 0;
	rqstp->rq_verfed  = 0;

	svc_getlong(&rqstp->rq_argbuf, rqstp->rq_xid);
	svc_putlong(&rqstp->rq_resbuf, rqstp->rq_xid);

	/* Assume that the reply consists of a single buffer. */
	rqstp->rq_resbuf.nriov = 1;

	if (serv->sv_stats)
		serv->sv_stats->netcnt++;
	return len;
}

/* 
 * Drop request
 */
void
svc_drop(struct svc_rqst *rqstp)
{
	dprintk("svc: socket %p dropped request\n", rqstp->rq_sock);
	svc_sock_release(rqstp);
}

/*
 * Return reply to client.
 */
int
svc_send(struct svc_rqst *rqstp)
{
	struct svc_sock	*svsk;
	int		len;

	if ((svsk = rqstp->rq_sock) == NULL) {
		printk(KERN_WARNING "NULL socket pointer in %s:%d\n",
				__FILE__, __LINE__);
		return -EFAULT;
	}

	/* release the receive skb before sending the reply */
	svc_release_skb(rqstp);

	len = svsk->sk_sendto(rqstp);
	svc_sock_release(rqstp);

	if (len == -ECONNREFUSED || len == -ENOTCONN || len == -EAGAIN)
		return 0;
	return len;
}

/*
 * Initialize socket for RPC use and create svc_sock struct
 * XXX: May want to setsockopt SO_SNDBUF and SO_RCVBUF.
 */
static struct svc_sock *
svc_setup_socket(struct svc_serv *serv, struct socket *sock,
					int *errp, int pmap_register)
{
	struct svc_sock	*svsk;
	struct sock	*inet;

	dprintk("svc: svc_setup_socket %p\n", sock);
	if (!(svsk = kmalloc(sizeof(*svsk), GFP_KERNEL))) {
		*errp = -ENOMEM;
		return NULL;
	}
	memset(svsk, 0, sizeof(*svsk));

	inet = sock->sk;
	inet->user_data = svsk;
	svsk->sk_sock = sock;
	svsk->sk_sk = inet;
	svsk->sk_ostate = inet->state_change;
	svsk->sk_odata = inet->data_ready;
	svsk->sk_server = serv;
	spin_lock_init(&svsk->sk_lock);

	/* Initialize the socket */
	if (sock->type == SOCK_DGRAM)
		*errp = svc_udp_init(svsk);
	else
		*errp = svc_tcp_init(svsk);
if (svsk->sk_sk == NULL)
	printk(KERN_WARNING "svsk->sk_sk == NULL after svc_prot_init!\n");

	/* Register socket with portmapper */
	if (*errp >= 0 && pmap_register)
		*errp = svc_register(serv, inet->protocol, ntohs(inet->sport));

	if (*errp < 0) {
		inet->user_data = NULL;
		kfree(svsk);
		return NULL;
	}

	spin_lock_bh(&serv->sv_lock);
	svsk->sk_list = serv->sv_allsocks;
	serv->sv_allsocks = svsk;
	spin_unlock_bh(&serv->sv_lock);

	dprintk("svc: svc_setup_socket created %p (inet %p)\n",
				svsk, svsk->sk_sk);
	return svsk;
}

/*
 * Create socket for RPC service.
 */
static int
svc_create_socket(struct svc_serv *serv, int protocol, struct sockaddr_in *sin)
{
	struct svc_sock	*svsk;
	struct socket	*sock;
	int		error;
	int		type;

	dprintk("svc: svc_create_socket(%s, %d, %u.%u.%u.%u:%d)\n",
				serv->sv_program->pg_name, protocol,
				NIPQUAD(sin->sin_addr.s_addr),
				ntohs(sin->sin_port));

	if (protocol != IPPROTO_UDP && protocol != IPPROTO_TCP) {
		printk(KERN_WARNING "svc: only UDP and TCP "
				"sockets supported\n");
		return -EINVAL;
	}
	type = (protocol == IPPROTO_UDP)? SOCK_DGRAM : SOCK_STREAM;

	if ((error = sock_create(PF_INET, type, protocol, &sock)) < 0)
		return error;

	if (sin != NULL) {
		error = sock->ops->bind(sock, (struct sockaddr *) sin,
						sizeof(*sin));
		if (error < 0)
			goto bummer;
	}

	if (protocol == IPPROTO_TCP) {
		if ((error = sock->ops->listen(sock, 5)) < 0)
			goto bummer;
	}

	if ((svsk = svc_setup_socket(serv, sock, &error, 1)) != NULL)
		return 0;

bummer:
	dprintk("svc: svc_create_socket error = %d\n", -error);
	sock_release(sock);
	return error;
}

/*
 * Remove a dead socket
 */
void
svc_delete_socket(struct svc_sock *svsk)
{
	struct svc_sock	**rsk;
	struct svc_serv	*serv;
	struct sock	*sk;

	dprintk("svc: svc_delete_socket(%p)\n", svsk);

	serv = svsk->sk_server;
	sk = svsk->sk_sk;

	sk->state_change = svsk->sk_ostate;
	sk->data_ready = svsk->sk_odata;

	spin_lock_bh(&serv->sv_lock);

	for (rsk = &serv->sv_allsocks; *rsk; rsk = &(*rsk)->sk_list) {
		if (*rsk == svsk)
			break;
	}
	if (!*rsk) {
		spin_unlock_bh(&serv->sv_lock);
		return;
	}
	*rsk = svsk->sk_list;
	if (svsk->sk_qued)
		rpc_remove_list(&serv->sv_sockets, svsk);

	spin_unlock_bh(&serv->sv_lock);

	svsk->sk_dead = 1;

	if (!svsk->sk_inuse) {
		sock_release(svsk->sk_sock);
		kfree(svsk);
	} else {
		printk(KERN_NOTICE "svc: server socket destroy delayed\n");
		/* svsk->sk_server = NULL; */
	}
}

/*
 * Make a socket for nfsd and lockd
 */
int
svc_makesock(struct svc_serv *serv, int protocol, unsigned short port)
{
	struct sockaddr_in	sin;

	dprintk("svc: creating socket proto = %d\n", protocol);
	sin.sin_family      = AF_INET;
	sin.sin_addr.s_addr = INADDR_ANY;
	sin.sin_port        = htons(port);
	return svc_create_socket(serv, protocol, &sin);
}