xprtsock.c

linux 内核源代码

	} else
		base -= xdr->head[0].iov_len;

	if (base < xdr->page_len) {
		unsigned int len = xdr->page_len - base;
		remainder -= len;
		err = xs_send_pagedata(sock, xdr, base, remainder != 0);
		if (remainder == 0 || err != len)
			goto out;
		sent += err;
		base = 0;
	} else
		base -= xdr->page_len;

	if (base >= xdr->tail[0].iov_len)
		return sent;
	err = xs_send_kvec(sock, NULL, 0, &xdr->tail[0], base, 0);
out:
	if (sent == 0)
		return err;
	if (err > 0)
		sent += err;
	return sent;
}

/**
 * xs_nospace - place task on wait queue if transmit was incomplete
 * @task: task to put to sleep
 *
 */
static void xs_nospace(struct rpc_task *task)
{
	struct rpc_rqst *req = task->tk_rqstp;
	struct rpc_xprt *xprt = req->rq_xprt;
	struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);

	dprintk("RPC: %5u xmit incomplete (%u left of %u)\n",
			task->tk_pid, req->rq_slen - req->rq_bytes_sent,
			req->rq_slen);

	if (test_bit(SOCK_ASYNC_NOSPACE, &transport->sock->flags)) {
		/* Protect against races with write_space */
		spin_lock_bh(&xprt->transport_lock);

		/* Don't race with disconnect */
		if (!xprt_connected(xprt))
			task->tk_status = -ENOTCONN;
		else if (test_bit(SOCK_NOSPACE, &transport->sock->flags))
			xprt_wait_for_buffer_space(task);

		spin_unlock_bh(&xprt->transport_lock);
	} else
		/* Keep holding the socket if it is blocked */
		rpc_delay(task, HZ>>4);
}

/**
 * xs_udp_send_request - write an RPC request to a UDP socket
 * @task: address of RPC task that manages the state of an RPC request
 *
 * Return values:
 *        0:	The request has been sent
 *   EAGAIN:	The socket was blocked, please call again later to
 *		complete the request
 * ENOTCONN:	Caller needs to invoke connect logic then call again
 *    other:	Some other error occured, the request was not sent
 */
static int xs_udp_send_request(struct rpc_task *task)
{
	struct rpc_rqst *req = task->tk_rqstp;
	struct rpc_xprt *xprt = req->rq_xprt;
	struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
	struct xdr_buf *xdr = &req->rq_snd_buf;
	int status;

	xs_pktdump("packet data:",
				req->rq_svec->iov_base,
				req->rq_svec->iov_len);

	req->rq_xtime = jiffies;
	status = xs_sendpages(transport->sock,
			      xs_addr(xprt),
			      xprt->addrlen, xdr,
			      req->rq_bytes_sent);

	dprintk("RPC:       xs_udp_send_request(%u) = %d\n",
			xdr->len - req->rq_bytes_sent, status);

	if (status >= 0) {
		task->tk_bytes_sent += status;
		if (status >= req->rq_slen)
			return 0;
		/* Still some bytes left; set up for a retry later. */
		status = -EAGAIN;
	}

	switch (status) {
	case -ENETUNREACH:
	case -EPIPE:
	case -ECONNREFUSED:
		/* When the server has died, an ICMP port unreachable message
		 * prompts ECONNREFUSED. */
		break;
	case -EAGAIN:
		xs_nospace(task);
		break;
	default:
		dprintk("RPC:       sendmsg returned unrecognized error %d\n",
			-status);
		break;
	}

	return status;
}

static inline void xs_encode_tcp_record_marker(struct xdr_buf *buf)
{
	u32 reclen = buf->len - sizeof(rpc_fraghdr);
	rpc_fraghdr *base = buf->head[0].iov_base;
	*base = htonl(RPC_LAST_STREAM_FRAGMENT | reclen);
}

/**
 * xs_tcp_send_request - write an RPC request to a TCP socket
 * @task: address of RPC task that manages the state of an RPC request
 *
 * Return values:
 *        0:	The request has been sent
 *   EAGAIN:	The socket was blocked, please call again later to
 *		complete the request
 * ENOTCONN:	Caller needs to invoke connect logic then call again
 *    other:	Some other error occured, the request was not sent
 *
 * XXX: In the case of soft timeouts, should we eventually give up
 *	if sendmsg is not able to make progress?
 */
static int xs_tcp_send_request(struct rpc_task *task)
{
	struct rpc_rqst *req = task->tk_rqstp;
	struct rpc_xprt *xprt = req->rq_xprt;
	struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
	struct xdr_buf *xdr = &req->rq_snd_buf;
	int status;
	unsigned int retry = 0;

	xs_encode_tcp_record_marker(&req->rq_snd_buf);

	xs_pktdump("packet data:",
				req->rq_svec->iov_base,
				req->rq_svec->iov_len);

	/* Continue transmitting the packet/record. We must be careful
	 * to cope with writespace callbacks arriving _after_ we have
	 * called sendmsg(). */
	while (1) {
		req->rq_xtime = jiffies;
		status = xs_sendpages(transport->sock,
					NULL, 0, xdr, req->rq_bytes_sent);

		dprintk("RPC:       xs_tcp_send_request(%u) = %d\n",
				xdr->len - req->rq_bytes_sent, status);

		if (unlikely(status < 0))
			break;

		/* If we've sent the entire packet, immediately
		 * reset the count of bytes sent. */
		req->rq_bytes_sent += status;
		task->tk_bytes_sent += status;
		if (likely(req->rq_bytes_sent >= req->rq_slen)) {
			req->rq_bytes_sent = 0;
			return 0;
		}

		status = -EAGAIN;
		if (retry++ > XS_SENDMSG_RETRY)
			break;
	}

	switch (status) {
	case -EAGAIN:
		xs_nospace(task);
		break;
	case -ECONNREFUSED:
	case -ECONNRESET:
	case -ENOTCONN:
	case -EPIPE:
		status = -ENOTCONN;
		break;
	default:
		dprintk("RPC:       sendmsg returned unrecognized error %d\n",
			-status);
		xprt_disconnect(xprt);
		break;
	}

	return status;
}

/**
 * xs_tcp_release_xprt - clean up after a tcp transmission
 * @xprt: transport
 * @task: rpc task
 *
 * This cleans up if an error causes us to abort the transmission of a request.
 * In this case, the socket may need to be reset in order to avoid confusing
 * the server.
 */
static void xs_tcp_release_xprt(struct rpc_xprt *xprt, struct rpc_task *task)
{
	struct rpc_rqst *req;

	if (task != xprt->snd_task)
		return;
	if (task == NULL)
		goto out_release;
	req = task->tk_rqstp;
	if (req->rq_bytes_sent == 0)
		goto out_release;
	if (req->rq_bytes_sent == req->rq_snd_buf.len)
		goto out_release;
	set_bit(XPRT_CLOSE_WAIT, &task->tk_xprt->state);
out_release:
	xprt_release_xprt(xprt, task);
}

/**
 * xs_close - close a socket
 * @xprt: transport
 *
 * This is used when all requests are complete; ie, no DRC state remains
 * on the server we want to save.
 */
static void xs_close(struct rpc_xprt *xprt)
{
	struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
	struct socket *sock = transport->sock;
	struct sock *sk = transport->inet;

	if (!sk)
		goto clear_close_wait;

	dprintk("RPC:       xs_close xprt %p\n", xprt);

	write_lock_bh(&sk->sk_callback_lock);
	transport->inet = NULL;
	transport->sock = NULL;

	sk->sk_user_data = NULL;
	sk->sk_data_ready = transport->old_data_ready;
	sk->sk_state_change = transport->old_state_change;
	sk->sk_write_space = transport->old_write_space;
	write_unlock_bh(&sk->sk_callback_lock);

	sk->sk_no_check = 0;

	sock_release(sock);
clear_close_wait:
	smp_mb__before_clear_bit();
	clear_bit(XPRT_CLOSE_WAIT, &xprt->state);
	smp_mb__after_clear_bit();
}

/**
 * xs_destroy - prepare to shutdown a transport
 * @xprt: doomed transport
 *
 */
static void xs_destroy(struct rpc_xprt *xprt)
{
	struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);

	dprintk("RPC:       xs_destroy xprt %p\n", xprt);

	cancel_rearming_delayed_work(&transport->connect_worker);

	xprt_disconnect(xprt);
	xs_close(xprt);
	xs_free_peer_addresses(xprt);
	kfree(xprt->slot);
	kfree(xprt);
	module_put(THIS_MODULE);
}

static inline struct rpc_xprt *xprt_from_sock(struct sock *sk)
{
	return (struct rpc_xprt *) sk->sk_user_data;
}

/**
 * xs_udp_data_ready - "data ready" callback for UDP sockets
 * @sk: socket with data to read
 * @len: how much data to read
 *
 */
static void xs_udp_data_ready(struct sock *sk, int len)
{
	struct rpc_task *task;
	struct rpc_xprt *xprt;
	struct rpc_rqst *rovr;
	struct sk_buff *skb;
	int err, repsize, copied;
	u32 _xid;
	__be32 *xp;

	read_lock(&sk->sk_callback_lock);
	dprintk("RPC:       xs_udp_data_ready...\n");
	if (!(xprt = xprt_from_sock(sk)))
		goto out;

	if ((skb = skb_recv_datagram(sk, 0, 1, &err)) == NULL)
		goto out;

	if (xprt->shutdown)
		goto dropit;

	repsize = skb->len - sizeof(struct udphdr);
	if (repsize < 4) {
		dprintk("RPC:       impossible RPC reply size %d!\n", repsize);
		goto dropit;
	}

	/* Copy the XID from the skb... */
	xp = skb_header_pointer(skb, sizeof(struct udphdr),
				sizeof(_xid), &_xid);
	if (xp == NULL)
		goto dropit;

	/* Look up and lock the request corresponding to the given XID */
	spin_lock(&xprt->transport_lock);
	rovr = xprt_lookup_rqst(xprt, *xp);
	if (!rovr)
		goto out_unlock;
	task = rovr->rq_task;

	if ((copied = rovr->rq_private_buf.buflen) > repsize)
		copied = repsize;

	/* Suck it into the iovec, verify checksum if not done by hw. */
	if (csum_partial_copy_to_xdr(&rovr->rq_private_buf, skb))
		goto out_unlock;

	/* Something worked... */
	dst_confirm(skb->dst);

	xprt_adjust_cwnd(task, copied);
	xprt_update_rtt(task);
	xprt_complete_rqst(task, copied);

 out_unlock:
	spin_unlock(&xprt->transport_lock);
 dropit:
	skb_free_datagram(sk, skb);
 out:
	read_unlock(&sk->sk_callback_lock);
}

static inline void xs_tcp_read_fraghdr(struct rpc_xprt *xprt, struct xdr_skb_reader *desc)
{
	struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
	size_t len, used;
	char *p;

	p = ((char *) &transport->tcp_fraghdr) + transport->tcp_offset;
	len = sizeof(transport->tcp_fraghdr) - transport->tcp_offset;
	used = xdr_skb_read_bits(desc, p, len);
	transport->tcp_offset += used;
	if (used != len)
		return;

	transport->tcp_reclen = ntohl(transport->tcp_fraghdr);
	if (transport->tcp_reclen & RPC_LAST_STREAM_FRAGMENT)
		transport->tcp_flags |= TCP_RCV_LAST_FRAG;
	else
		transport->tcp_flags &= ~TCP_RCV_LAST_FRAG;
	transport->tcp_reclen &= RPC_FRAGMENT_SIZE_MASK;

	transport->tcp_flags &= ~TCP_RCV_COPY_FRAGHDR;
	transport->tcp_offset = 0;

	/* Sanity check of the record length */
	if (unlikely(transport->tcp_reclen < 4)) {
		dprintk("RPC:       invalid TCP record fragment length\n");
		xprt_disconnect(xprt);
		return;
	}
	dprintk("RPC:       reading TCP record fragment of length %d\n",
			transport->tcp_reclen);
}

static void xs_tcp_check_fraghdr(struct sock_xprt *transport)
{
	if (transport->tcp_offset == transport->tcp_reclen) {
		transport->tcp_flags |= TCP_RCV_COPY_FRAGHDR;
		transport->tcp_offset = 0;
		if (transport->tcp_flags & TCP_RCV_LAST_FRAG) {
			transport->tcp_flags &= ~TCP_RCV_COPY_DATA;
			transport->tcp_flags |= TCP_RCV_COPY_XID;
			transport->tcp_copied = 0;
		}
	}
}

static inline void xs_tcp_read_xid(struct sock_xprt *transport, struct xdr_skb_reader *desc)
{
	size_t len, used;
	char *p;

	len = sizeof(transport->tcp_xid) - transport->tcp_offset;
	dprintk("RPC:       reading XID (%Zu bytes)\n", len);
	p = ((char *) &transport->tcp_xid) + transport->tcp_offset;
	used = xdr_skb_read_bits(desc, p, len);
	transport->tcp_offset += used;
	if (used != len)
		return;
	transport->tcp_flags &= ~TCP_RCV_COPY_XID;
	transport->tcp_flags |= TCP_RCV_COPY_DATA;
	transport->tcp_copied = 4;
	dprintk("RPC:       reading reply for XID %08x\n",
			ntohl(transport->tcp_xid));
	xs_tcp_check_fraghdr(transport);
}

static inline void xs_tcp_read_request(struct rpc_xprt *xprt, struct xdr_skb_reader *desc)
{
	struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
	struct rpc_rqst *req;
	struct xdr_buf *rcvbuf;
	size_t len;
	ssize_t r;

	/* Find and lock the request corresponding to this xid */
	spin_lock(&xprt->transport_lock);
	req = xprt_lookup_rqst(xprt, transport->tcp_xid);
	if (!req) {
		transport->tcp_flags &= ~TCP_RCV_COPY_DATA;
		dprintk("RPC:       XID %08x request not found!\n",
				ntohl(transport->tcp_xid));
		spin_unlock(&xprt->transport_lock);
		return;
	}

	rcvbuf = &req->rq_private_buf;
	len = desc->count;
	if (len > transport->tcp_reclen - transport->tcp_offset) {
		struct xdr_skb_reader my_desc;

		len = transport->tcp_reclen - transport->tcp_offset;
		memcpy(&my_desc, desc, sizeof(my_desc));
		my_desc.count = len;
		r = xdr_partial_copy_from_skb(rcvbuf, transport->tcp_copied,
					  &my_desc, xdr_skb_read_bits);
		desc->count -= r;
		desc->offset += r;
	} else
		r = xdr_partial_copy_from_skb(rcvbuf, transport->tcp_copied,
					  desc, xdr_skb_read_bits);

	if (r > 0) {
		transport->tcp_copied += r;
		transport->tcp_offset += r;
	}
	if (r != len) {
		/* Error when copying to the receive buffer,
		 * usually because we weren't able to allocate
		 * additional buffer pages. All we can do now
		 * is turn off TCP_RCV_COPY_DATA, so the request
		 * will not receive any additional updates,
		 * and time out.
		 * Any remaining data from this record will
		 * be discarded.
		 */
		transport->tcp_flags &= ~TCP_RCV_COPY_DATA;
		dprintk("RPC:       XID %08x truncated request\n",
				ntohl(transport->tcp_xid));
		dprintk("RPC:       xprt = %p, tcp_copied = %lu, "
				"tcp_offset = %u, tcp_reclen = %u\n",
				xprt, transport->tcp_copied,
				transport->tcp_offset, transport->tcp_reclen);
		goto out;
	}

	dprintk("RPC:       XID %08x read %Zd bytes\n",
			ntohl(transport->tcp_xid), r);
	dprintk("RPC:       xprt = %p, tcp_copied = %lu, tcp_offset = %u, "
			"tcp_reclen = %u\n", xprt, transport->tcp_copied,
			transport->tcp_offset, transport->tcp_reclen);

	if (transport->tcp_copied == req->rq_private_buf.buflen)
		transport->tcp_flags &= ~TCP_RCV_COPY_DATA;
	else if (transport->tcp_offset == transport->tcp_reclen) {
		if (transport->tcp_flags & TCP_RCV_LAST_FRAG)
			transport->tcp_flags &= ~TCP_RCV_COPY_DATA;
	}

out:
	if (!(transport->tcp_flags & TCP_RCV_COPY_DATA))
		xprt_complete_rqst(req->rq_task, transport->tcp_copied);
	spin_unlock(&xprt->transport_lock);
	xs_tcp_check_fraghdr(transport);
}