Merge branch 'work.sendmsg' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs

Pull vfs sendmsg updates from Al Viro:
"More sendmsg work.

This is a fairly separate isolated stuff (there's a continuation
around lustre, but that one was too late to soak in -next), thus the
separate pull request"

* 'work.sendmsg' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
ncpfs: switch to sock_sendmsg()
ncpfs: don't mess with manually advancing iovec on send
ncpfs: sendmsg does *not* bugger iovec these days
ceph_tcp_sendpage(): use ITER_BVEC sendmsg
afs_send_pages(): use ITER_BVEC
rds: remove dead code
ceph: switch to sock_recvmsg()
usbip_recv(): switch to sock_recvmsg()
iscsi_target: deal with short writes on the tx side
[nbd] pass iov_iter to nbd_xmit()
[nbd] switch sock_xmit() to sock_{send,recv}msg()
[drbd] use sock_sendmsg()

drivers/block/drbd/drbd_main.c

···
 int drbd_send(struct drbd_connection *connection, struct socket *sock,
 	      void *buf, size_t size, unsigned msg_flags)
 {
-	struct kvec iov;
+	struct kvec iov = {.iov_base = buf, .iov_len = size};
 	struct msghdr msg;
 	int rv, sent = 0;
 
···
 
 	/* THINK  if (signal_pending) return ... ? */
 
-	iov.iov_base = buf;
-	iov.iov_len  = size;
-
 	msg.msg_name       = NULL;
 	msg.msg_namelen    = 0;
 	msg.msg_control    = NULL;
 	msg.msg_controllen = 0;
 	msg.msg_flags      = msg_flags | MSG_NOSIGNAL;
+
+	iov_iter_kvec(&msg.msg_iter, WRITE | ITER_KVEC, &iov, 1, size);
 
 	if (sock == connection->data.socket) {
 		rcu_read_lock();
···
 		drbd_update_congested(connection);
 	}
 	do {
-		rv = kernel_sendmsg(sock, &msg, &iov, 1, iov.iov_len);
+		rv = sock_sendmsg(sock, &msg);
 		if (rv == -EAGAIN) {
 			if (we_should_drop_the_connection(connection, sock))
 				break;
···
 		if (rv < 0)
 			break;
 		sent += rv;
-		iov.iov_base += rv;
-		iov.iov_len  -= rv;
 	} while (sent < size);
 
 	if (sock == connection->data.socket)

drivers/block/nbd.c

···
 /*
  *  Send or receive packet.
  */
-static int sock_xmit(struct nbd_device *nbd, int index, int send, void *buf,
-		     int size, int msg_flags)
+static int sock_xmit(struct nbd_device *nbd, int index, int send,
+		     struct iov_iter *iter, int msg_flags)
 {
 	struct socket *sock = nbd->socks[index]->sock;
 	int result;
 	struct msghdr msg;
-	struct kvec iov;
 	unsigned long pflags = current->flags;
 
 	if (unlikely(!sock)) {
···
 		return -EINVAL;
 	}
 
+	msg.msg_iter = *iter;
+
 	current->flags |= PF_MEMALLOC;
 	do {
 		sock->sk->sk_allocation = GFP_NOIO | __GFP_MEMALLOC;
-		iov.iov_base = buf;
-		iov.iov_len = size;
 		msg.msg_name = NULL;
 		msg.msg_namelen = 0;
 		msg.msg_control = NULL;
···
 		msg.msg_flags = msg_flags | MSG_NOSIGNAL;
 
 		if (send)
-			result = kernel_sendmsg(sock, &msg, &iov, 1, size);
+			result = sock_sendmsg(sock, &msg);
 		else
-			result = kernel_recvmsg(sock, &msg, &iov, 1, size,
-						msg.msg_flags);
+			result = sock_recvmsg(sock, &msg, msg.msg_flags);
 
 		if (result <= 0) {
 			if (result == 0)
 				result = -EPIPE; /* short read */
 			break;
 		}
-		size -= result;
-		buf += result;
-	} while (size > 0);
+	} while (msg_data_left(&msg));
 
 	tsk_restore_flags(current, pflags, PF_MEMALLOC);
 
-	return result;
-}
-
-static inline int sock_send_bvec(struct nbd_device *nbd, int index,
-				 struct bio_vec *bvec, int flags)
-{
-	int result;
-	void *kaddr = kmap(bvec->bv_page);
-	result = sock_xmit(nbd, index, 1, kaddr + bvec->bv_offset,
-			   bvec->bv_len, flags);
-	kunmap(bvec->bv_page);
 	return result;
 }
 
···
 {
 	struct request *req = blk_mq_rq_from_pdu(cmd);
 	int result;
-	struct nbd_request request;
+	struct nbd_request request = {.magic = htonl(NBD_REQUEST_MAGIC)};
+	struct kvec iov = {.iov_base = &request, .iov_len = sizeof(request)};
+	struct iov_iter from;
 	unsigned long size = blk_rq_bytes(req);
 	struct bio *bio;
 	u32 type;
 	u32 tag = blk_mq_unique_tag(req);
+
+	iov_iter_kvec(&from, WRITE | ITER_KVEC, &iov, 1, sizeof(request));
 
 	switch (req_op(req)) {
 	case REQ_OP_DISCARD:
···
 		return -EIO;
 	}
 
-	memset(&request, 0, sizeof(request));
-	request.magic = htonl(NBD_REQUEST_MAGIC);
 	request.type = htonl(type);
 	if (type != NBD_CMD_FLUSH) {
 		request.from = cpu_to_be64((u64)blk_rq_pos(req) << 9);
···
 	dev_dbg(nbd_to_dev(nbd), "request %p: sending control (%s@%llu,%uB)\n",
 		cmd, nbdcmd_to_ascii(type),
 		(unsigned long long)blk_rq_pos(req) << 9, blk_rq_bytes(req));
-	result = sock_xmit(nbd, index, 1, &request, sizeof(request),
+	result = sock_xmit(nbd, index, 1, &from,
 			(type == NBD_CMD_WRITE) ? MSG_MORE : 0);
 	if (result <= 0) {
 		dev_err_ratelimited(disk_to_dev(nbd->disk),
···
 
 			dev_dbg(nbd_to_dev(nbd), "request %p: sending %d bytes data\n",
 				cmd, bvec.bv_len);
-			result = sock_send_bvec(nbd, index, &bvec, flags);
+			iov_iter_bvec(&from, ITER_BVEC | WRITE,
+				      &bvec, 1, bvec.bv_len);
+			result = sock_xmit(nbd, index, 1, &from, flags);
 			if (result <= 0) {
 				dev_err(disk_to_dev(nbd->disk),
 					"Send data failed (result %d)\n",
···
 	return 0;
 }
 
-static inline int sock_recv_bvec(struct nbd_device *nbd, int index,
-				 struct bio_vec *bvec)
-{
-	int result;
-	void *kaddr = kmap(bvec->bv_page);
-	result = sock_xmit(nbd, index, 0, kaddr + bvec->bv_offset,
-			   bvec->bv_len, MSG_WAITALL);
-	kunmap(bvec->bv_page);
-	return result;
-}
-
 /* NULL returned = something went wrong, inform userspace */
 static struct nbd_cmd *nbd_read_stat(struct nbd_device *nbd, int index)
 {
···
 	struct request *req = NULL;
 	u16 hwq;
 	u32 tag;
+	struct kvec iov = {.iov_base = &reply, .iov_len = sizeof(reply)};
+	struct iov_iter to;
 
 	reply.magic = 0;
-	result = sock_xmit(nbd, index, 0, &reply, sizeof(reply), MSG_WAITALL);
+	iov_iter_kvec(&to, READ | ITER_KVEC, &iov, 1, sizeof(reply));
+	result = sock_xmit(nbd, index, 0, &to, MSG_WAITALL);
 	if (result <= 0) {
 		if (!test_bit(NBD_DISCONNECTED, &nbd->runtime_flags) &&
 		    !test_bit(NBD_DISCONNECT_REQUESTED, &nbd->runtime_flags))
···
 		struct bio_vec bvec;
 
 		rq_for_each_segment(bvec, req, iter) {
-			result = sock_recv_bvec(nbd, index, &bvec);
+			iov_iter_bvec(&to, ITER_BVEC | READ,
+				      &bvec, 1, bvec.bv_len);
+			result = sock_xmit(nbd, index, 0, &to, MSG_WAITALL);
 			if (result <= 0) {
 				dev_err(disk_to_dev(nbd->disk), "Receive data failed (result %d)\n",
 					result);
···
 
 static void send_disconnects(struct nbd_device *nbd)
 {
-	struct nbd_request request = {};
+	struct nbd_request request = {
+		.magic = htonl(NBD_REQUEST_MAGIC),
+		.type = htonl(NBD_CMD_DISC),
+	};
+	struct kvec iov = {.iov_base = &request, .iov_len = sizeof(request)};
+	struct iov_iter from;
 	int i, ret;
 
-	request.magic = htonl(NBD_REQUEST_MAGIC);
-	request.type = htonl(NBD_CMD_DISC);
-
 	for (i = 0; i < nbd->num_connections; i++) {
-		ret = sock_xmit(nbd, i, 1, &request, sizeof(request), 0);
+		iov_iter_kvec(&from, WRITE | ITER_KVEC, &iov, 1, sizeof(request));
+		ret = sock_xmit(nbd, i, 1, &from, 0);
 		if (ret <= 0)
 			dev_err(disk_to_dev(nbd->disk),
 				"Send disconnect failed %d\n", ret);

drivers/target/iscsi/iscsi_target_util.c

···
 	return total_rx;
 }
 
-static int iscsit_do_tx_data(
-	struct iscsi_conn *conn,
-	struct iscsi_data_count *count)
-{
-	int ret, iov_len;
-	struct kvec *iov_p;
-	struct msghdr msg;
-
-	if (!conn || !conn->sock || !conn->conn_ops)
-		return -1;
-
-	if (count->data_length <= 0) {
-		pr_err("Data length is: %d\n", count->data_length);
-		return -1;
-	}
-
-	memset(&msg, 0, sizeof(struct msghdr));
-
-	iov_p = count->iov;
-	iov_len = count->iov_count;
-
-	ret = kernel_sendmsg(conn->sock, &msg, iov_p, iov_len,
-			     count->data_length);
-	if (ret != count->data_length) {
-		pr_err("Unexpected ret: %d send data %d\n",
-		       ret, count->data_length);
-		return -EPIPE;
-	}
-	pr_debug("ret: %d, sent data: %d\n", ret, count->data_length);
-
-	return ret;
-}
-
 int rx_data(
 	struct iscsi_conn *conn,
 	struct kvec *iov,
···
 	int iov_count,
 	int data)
 {
-	struct iscsi_data_count c;
+	struct msghdr msg;
+	int total_tx = 0;
 
 	if (!conn || !conn->sock || !conn->conn_ops)
 		return -1;
 
-	memset(&c, 0, sizeof(struct iscsi_data_count));
-	c.iov = iov;
-	c.iov_count = iov_count;
-	c.data_length = data;
-	c.type = ISCSI_TX_DATA;
+	if (data <= 0) {
+		pr_err("Data length is: %d\n", data);
+		return -1;
+	}
 
-	return iscsit_do_tx_data(conn, &c);
+	memset(&msg, 0, sizeof(struct msghdr));
+
+	iov_iter_kvec(&msg.msg_iter, WRITE | ITER_KVEC,
+		      iov, iov_count, data);
+
+	while (msg_data_left(&msg)) {
+		int tx_loop = sock_sendmsg(conn->sock, &msg);
+		if (tx_loop <= 0) {
+			pr_debug("tx_loop: %d total_tx %d\n",
+				tx_loop, total_tx);
+			return tx_loop;
+		}
+		total_tx += tx_loop;
+		pr_debug("tx_loop: %d, total_tx: %d, data: %d\n",
+					tx_loop, total_tx, data);
+	}
+
+	return total_tx;
 }
 
 void iscsit_collect_login_stats(

drivers/usb/usbip/usbip_common.c

···
 int usbip_recv(struct socket *sock, void *buf, int size)
 {
 	int result;
-	struct msghdr msg;
-	struct kvec iov;
+	struct kvec iov = {.iov_base = buf, .iov_len = size};
+	struct msghdr msg = {.msg_flags = MSG_NOSIGNAL};
 	int total = 0;
 
-	/* for blocks of if (usbip_dbg_flag_xmit) */
-	char *bp = buf;
-	int osize = size;
+	iov_iter_kvec(&msg.msg_iter, READ|ITER_KVEC, &iov, 1, size);
 
 	usbip_dbg_xmit("enter\n");
 
···
 	}
 
 	do {
+		int sz = msg_data_left(&msg);
 		sock->sk->sk_allocation = GFP_NOIO;
-		iov.iov_base    = buf;
-		iov.iov_len     = size;
-		msg.msg_name    = NULL;
-		msg.msg_namelen = 0;
-		msg.msg_control = NULL;
-		msg.msg_controllen = 0;
-		msg.msg_flags      = MSG_NOSIGNAL;
 
-		result = kernel_recvmsg(sock, &msg, &iov, 1, size, MSG_WAITALL);
+		result = sock_recvmsg(sock, &msg, MSG_WAITALL);
 		if (result <= 0) {
 			pr_debug("receive sock %p buf %p size %u ret %d total %d\n",
-				 sock, buf, size, result, total);
+				 sock, buf + total, sz, result, total);
 			goto err;
 		}
 
-		size -= result;
-		buf += result;
 		total += result;
-	} while (size > 0);
+	} while (msg_data_left(&msg));
 
 	if (usbip_dbg_flag_xmit) {
 		if (!in_interrupt())
···
 			pr_debug("interrupt  :");
 
 		pr_debug("receiving....\n");
-		usbip_dump_buffer(bp, osize);
-		pr_debug("received, osize %d ret %d size %d total %d\n",
-			 osize, result, size, total);
+		usbip_dump_buffer(buf, size);
+		pr_debug("received, osize %d ret %d size %zd total %d\n",
+			 size, result, msg_data_left(&msg), total);
 	}
 
 	return total;

fs/afs/rxrpc.c

···
 /*
  * attach the data from a bunch of pages on an inode to a call
  */
-static int afs_send_pages(struct afs_call *call, struct msghdr *msg,
-			  struct kvec *iov)
+static int afs_send_pages(struct afs_call *call, struct msghdr *msg)
 {
 	struct page *pages[8];
 	unsigned count, n, loop, offset, to;
···
 
 		loop = 0;
 		do {
+			struct bio_vec bvec = {.bv_page = pages[loop],
+					       .bv_offset = offset};
 			msg->msg_flags = 0;
 			to = PAGE_SIZE;
 			if (first + loop >= last)
 				to = call->last_to;
 			else
 				msg->msg_flags = MSG_MORE;
-			iov->iov_base = kmap(pages[loop]) + offset;
-			iov->iov_len = to - offset;
+			bvec.bv_len = to - offset;
 			offset = 0;
 
 			_debug("- range %u-%u%s",
 			       offset, to, msg->msg_flags ? " [more]" : "");
-			iov_iter_kvec(&msg->msg_iter, WRITE | ITER_KVEC,
-				      iov, 1, to - offset);
+			iov_iter_bvec(&msg->msg_iter, WRITE | ITER_BVEC,
+				      &bvec, 1, to - offset);
 
 			/* have to change the state *before* sending the last
 			 * packet as RxRPC might give us the reply before it
···
 				call->state = AFS_CALL_AWAIT_REPLY;
 			ret = rxrpc_kernel_send_data(afs_socket, call->rxcall,
 						     msg, to - offset);
-			kunmap(pages[loop]);
 			if (ret < 0)
 				break;
 		} while (++loop < count);
···
 		goto error_do_abort;
 
 	if (call->send_pages) {
-		ret = afs_send_pages(call, &msg, iov);
+		ret = afs_send_pages(call, &msg);
 		if (ret < 0)
 			goto error_do_abort;
 	}

fs/ncpfs/sock.c

···
 	return kernel_recvmsg(sock, &msg, &iov, 1, size, flags);
 }
 
-static inline int do_send(struct socket *sock, struct kvec *vec, int count,
-			  int len, unsigned flags)
-{
-	struct msghdr msg = { .msg_flags = flags };
-	return kernel_sendmsg(sock, &msg, vec, count, len);
-}
-
 static int _send(struct socket *sock, const void *buff, int len)
 {
-	struct kvec vec;
-	vec.iov_base = (void *) buff;
-	vec.iov_len = len;
-	return do_send(sock, &vec, 1, len, 0);
+	struct msghdr msg = { .msg_flags = 0 };
+	struct kvec vec = {.iov_base = (void *)buff, .iov_len = len};
+	iov_iter_kvec(&msg.msg_iter, WRITE | ITER_KVEC, &vec, 1, len);
+	return sock_sendmsg(sock, &msg);
 }
 
 struct ncp_request_reply {
···
 	size_t datalen;
 	int result;
 	enum { RQ_DONE, RQ_INPROGRESS, RQ_QUEUED, RQ_IDLE, RQ_ABANDONED } status;
-	struct kvec* tx_ciov;
-	size_t tx_totallen;
-	size_t tx_iovlen;
+	struct iov_iter from;
 	struct kvec tx_iov[3];
 	u_int16_t tx_type;
 	u_int32_t sign[6];
···
 
 static int ncpdgram_send(struct socket *sock, struct ncp_request_reply *req)
 {
-	struct kvec vec[3];
-	/* sock_sendmsg updates iov pointers for us :-( */
-	memcpy(vec, req->tx_ciov, req->tx_iovlen * sizeof(vec[0]));
-	return do_send(sock, vec, req->tx_iovlen,
-		       req->tx_totallen, MSG_DONTWAIT);
+	struct msghdr msg = { .msg_iter = req->from, .msg_flags = MSG_DONTWAIT };
+	return sock_sendmsg(sock, &msg);
 }
 
 static void __ncptcp_try_send(struct ncp_server *server)
 {
 	struct ncp_request_reply *rq;
-	struct kvec *iov;
-	struct kvec iovc[3];
+	struct msghdr msg = { .msg_flags = MSG_NOSIGNAL | MSG_DONTWAIT };
 	int result;
 
 	rq = server->tx.creq;
 	if (!rq)
 		return;
 
-	/* sock_sendmsg updates iov pointers for us :-( */
-	memcpy(iovc, rq->tx_ciov, rq->tx_iovlen * sizeof(iov[0]));
-	result = do_send(server->ncp_sock, iovc, rq->tx_iovlen,
-			 rq->tx_totallen, MSG_NOSIGNAL | MSG_DONTWAIT);
+	msg.msg_iter = rq->from;
+	result = sock_sendmsg(server->ncp_sock, &msg);
 
 	if (result == -EAGAIN)
 		return;
···
 		__ncp_abort_request(server, rq, result);
 		return;
 	}
-	if (result >= rq->tx_totallen) {
+	if (!msg_data_left(&msg)) {
 		server->rcv.creq = rq;
 		server->tx.creq = NULL;
 		return;
 	}
-	rq->tx_totallen -= result;
-	iov = rq->tx_ciov;
-	while (iov->iov_len <= result) {
-		result -= iov->iov_len;
-		iov++;
-		rq->tx_iovlen--;
-	}
-	iov->iov_base += result;
-	iov->iov_len -= result;
-	rq->tx_ciov = iov;
+	rq->from = msg.msg_iter;
 }
 
 static inline void ncp_init_header(struct ncp_server *server, struct ncp_request_reply *req, struct ncp_request_header *h)
···
 	
 static void ncpdgram_start_request(struct ncp_server *server, struct ncp_request_reply *req)
 {
-	size_t signlen;
-	struct ncp_request_header* h;
+	size_t signlen, len = req->tx_iov[1].iov_len;
+	struct ncp_request_header *h = req->tx_iov[1].iov_base;
 	
-	req->tx_ciov = req->tx_iov + 1;
-
-	h = req->tx_iov[1].iov_base;
 	ncp_init_header(server, req, h);
-	signlen = sign_packet(server, req->tx_iov[1].iov_base + sizeof(struct ncp_request_header) - 1, 
-			req->tx_iov[1].iov_len - sizeof(struct ncp_request_header) + 1,
-			cpu_to_le32(req->tx_totallen), req->sign);
+	signlen = sign_packet(server,
+			req->tx_iov[1].iov_base + sizeof(struct ncp_request_header) - 1, 
+			len - sizeof(struct ncp_request_header) + 1,
+			cpu_to_le32(len), req->sign);
 	if (signlen) {
-		req->tx_ciov[1].iov_base = req->sign;
-		req->tx_ciov[1].iov_len = signlen;
-		req->tx_iovlen += 1;
-		req->tx_totallen += signlen;
+		/* NCP over UDP appends signature */
+		req->tx_iov[2].iov_base = req->sign;
+		req->tx_iov[2].iov_len = signlen;
 	}
+	iov_iter_kvec(&req->from, WRITE | ITER_KVEC,
+			req->tx_iov + 1, signlen ? 2 : 1, len + signlen);
 	server->rcv.creq = req;
 	server->timeout_last = server->m.time_out;
 	server->timeout_retries = server->m.retry_count;
···
 
 static void ncptcp_start_request(struct ncp_server *server, struct ncp_request_reply *req)
 {
-	size_t signlen;
-	struct ncp_request_header* h;
+	size_t signlen, len = req->tx_iov[1].iov_len;
+	struct ncp_request_header *h = req->tx_iov[1].iov_base;
 
-	req->tx_ciov = req->tx_iov;
-	h = req->tx_iov[1].iov_base;
 	ncp_init_header(server, req, h);
 	signlen = sign_packet(server, req->tx_iov[1].iov_base + sizeof(struct ncp_request_header) - 1,
-			req->tx_iov[1].iov_len - sizeof(struct ncp_request_header) + 1,
-			cpu_to_be32(req->tx_totallen + 24), req->sign + 4) + 16;
+			len - sizeof(struct ncp_request_header) + 1,
+			cpu_to_be32(len + 24), req->sign + 4) + 16;
 
 	req->sign[0] = htonl(NCP_TCP_XMIT_MAGIC);
-	req->sign[1] = htonl(req->tx_totallen + signlen);
+	req->sign[1] = htonl(len + signlen);
 	req->sign[2] = htonl(NCP_TCP_XMIT_VERSION);
 	req->sign[3] = htonl(req->datalen + 8);
+	/* NCP over TCP prepends signature */
 	req->tx_iov[0].iov_base = req->sign;
 	req->tx_iov[0].iov_len = signlen;
-	req->tx_iovlen += 1;
-	req->tx_totallen += signlen;
+	iov_iter_kvec(&req->from, WRITE | ITER_KVEC,
+			req->tx_iov, 2, len + signlen);
 
 	server->tx.creq = req;
 	__ncptcp_try_send(server);
···
 static void info_server(struct ncp_server *server, unsigned int id, const void * data, size_t len)
 {
 	if (server->info_sock) {
-		struct kvec iov[2];
-		__be32 hdr[2];
-	
-		hdr[0] = cpu_to_be32(len + 8);
-		hdr[1] = cpu_to_be32(id);
-	
-		iov[0].iov_base = hdr;
-		iov[0].iov_len = 8;
-		iov[1].iov_base = (void *) data;
-		iov[1].iov_len = len;
+		struct msghdr msg = { .msg_flags = MSG_NOSIGNAL };
+		__be32 hdr[2] = {cpu_to_be32(len + 8), cpu_to_be32(id)};
+		struct kvec iov[2] = {
+			{.iov_base = hdr, .iov_len = 8},
+			{.iov_base = (void *)data, .iov_len = len},
+		};
 
-		do_send(server->info_sock, iov, 2, len + 8, MSG_NOSIGNAL);
+		iov_iter_kvec(&msg.msg_iter, ITER_KVEC | WRITE,
+				iov, 2, len + 8);
+
+		sock_sendmsg(server->info_sock, &msg);
 	}
 }
 
···
 	req->datalen = max_reply_size;
 	req->tx_iov[1].iov_base = server->packet;
 	req->tx_iov[1].iov_len = size;
-	req->tx_iovlen = 1;
-	req->tx_totallen = size;
 	req->tx_type = *(u_int16_t*)server->packet;
 
 	result = ncp_add_request(server, req);

net/ceph/messenger.c

···
 	struct msghdr msg = { .msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL };
 	int r;
 
-	r = kernel_recvmsg(sock, &msg, &iov, 1, len, msg.msg_flags);
+	iov_iter_kvec(&msg.msg_iter, READ | ITER_KVEC, &iov, 1, len);
+	r = sock_recvmsg(sock, &msg, msg.msg_flags);
 	if (r == -EAGAIN)
 		r = 0;
 	return r;
···
 static int ceph_tcp_recvpage(struct socket *sock, struct page *page,
 		     int page_offset, size_t length)
 {
-	void *kaddr;
-	int ret;
+	struct bio_vec bvec = {
+		.bv_page = page,
+		.bv_offset = page_offset,
+		.bv_len = length
+	};
+	struct msghdr msg = { .msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL };
+	int r;
 
 	BUG_ON(page_offset + length > PAGE_SIZE);
-
-	kaddr = kmap(page);
-	BUG_ON(!kaddr);
-	ret = ceph_tcp_recvmsg(sock, kaddr + page_offset, length);
-	kunmap(page);
-
-	return ret;
+	iov_iter_bvec(&msg.msg_iter, READ | ITER_BVEC, &bvec, 1, length);
+	r = sock_recvmsg(sock, &msg, msg.msg_flags);
+	if (r == -EAGAIN)
+		r = 0;
+	return r;
 }
 
 /*
···
 static int ceph_tcp_sendpage(struct socket *sock, struct page *page,
 		     int offset, size_t size, bool more)
 {
+	struct msghdr msg = { .msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL };
+	struct bio_vec bvec;
 	int ret;
-	struct kvec iov;
 
 	/* sendpage cannot properly handle pages with page_count == 0,
 	 * we need to fallback to sendmsg if that's the case */
 	if (page_count(page) >= 1)
 		return __ceph_tcp_sendpage(sock, page, offset, size, more);
 
-	iov.iov_base = kmap(page) + offset;
-	iov.iov_len = size;
-	ret = ceph_tcp_sendmsg(sock, &iov, 1, size, more);
-	kunmap(page);
+	bvec.bv_page = page;
+	bvec.bv_offset = offset;
+	bvec.bv_len = size;
+
+	if (more)
+		msg.msg_flags |= MSG_MORE;
+	else
+		msg.msg_flags |= MSG_EOR;  /* superfluous, but what the hell */
+
+	iov_iter_bvec(&msg.msg_iter, WRITE | ITER_BVEC, &bvec, 1, size);
+	ret = sock_sendmsg(sock, &msg);
+	if (ret == -EAGAIN)
+		ret = 0;
 
 	return ret;
 }

net/rds/page.c

···
 static
 DEFINE_PER_CPU_SHARED_ALIGNED(struct rds_page_remainder, rds_page_remainders);
 
-/*
- * returns 0 on success or -errno on failure.
- *
- * We don't have to worry about flush_dcache_page() as this only works
- * with private pages.  If, say, we were to do directed receive to pinned
- * user pages we'd have to worry more about cache coherence.  (Though
- * the flush_dcache_page() in get_user_pages() would probably be enough).
- */
-int rds_page_copy_user(struct page *page, unsigned long offset,
-		       void __user *ptr, unsigned long bytes,
-		       int to_user)
-{
-	unsigned long ret;
-	void *addr;
-
-	addr = kmap(page);
-	if (to_user) {
-		rds_stats_add(s_copy_to_user, bytes);
-		ret = copy_to_user(ptr, addr + offset, bytes);
-	} else {
-		rds_stats_add(s_copy_from_user, bytes);
-		ret = copy_from_user(addr + offset, ptr, bytes);
-	}
-	kunmap(page);
-
-	return ret ? -EFAULT : 0;
-}
-EXPORT_SYMBOL_GPL(rds_page_copy_user);
-
 /**
  * rds_page_remainder_alloc - build up regions of a message.
  *

net/rds/rds.h

···
 /* page.c */
 int rds_page_remainder_alloc(struct scatterlist *scat, unsigned long bytes,
 			     gfp_t gfp);
-int rds_page_copy_user(struct page *page, unsigned long offset,
-		       void __user *ptr, unsigned long bytes,
-		       int to_user);
-#define rds_page_copy_to_user(page, offset, ptr, bytes) \
-	rds_page_copy_user(page, offset, ptr, bytes, 1)
-#define rds_page_copy_from_user(page, offset, ptr, bytes) \
-	rds_page_copy_user(page, offset, ptr, bytes, 0)
 void rds_page_exit(void);
 
 /* recv.c */