drivers/scsi/iscsi_tcp.c

/*
 * iSCSI Initiator over TCP/IP Data-Path
 *
 * Copyright (C) 2004 Dmitry Yusupov
 * Copyright (C) 2004 Alex Aizman
 * Copyright (C) 2005 - 2006 Mike Christie
 * Copyright (C) 2006 Red Hat, Inc.  All rights reserved.
 * maintained by open-iscsi@googlegroups.com
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published
 * by the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
 * General Public License for more details.
 *
 * See the file COPYING included with this distribution for more details.
 *
 * Credits:
 *	Christoph Hellwig
 *	FUJITA Tomonori
 *	Arne Redlich
 *	Zhenyu Wang
 */

#include <linux/types.h>
#include <linux/inet.h>
#include <linux/file.h>
#include <linux/blkdev.h>
#include <linux/crypto.h>
#include <linux/delay.h>
#include <linux/kfifo.h>
#include <linux/scatterlist.h>
#include <net/tcp.h>
#include <scsi/scsi_cmnd.h>
#include <scsi/scsi_device.h>
#include <scsi/scsi_host.h>
#include <scsi/scsi.h>
#include <scsi/scsi_transport_iscsi.h>

#include "iscsi_tcp.h"

MODULE_AUTHOR("Mike Christie <michaelc@cs.wisc.edu>, "
	      "Dmitry Yusupov <dmitry_yus@yahoo.com>, "
	      "Alex Aizman <itn780@yahoo.com>");
MODULE_DESCRIPTION("iSCSI/TCP data-path");
MODULE_LICENSE("GPL");

static struct scsi_transport_template *iscsi_sw_tcp_scsi_transport;
static struct scsi_host_template iscsi_sw_tcp_sht;
static struct iscsi_transport iscsi_sw_tcp_transport;

static unsigned int iscsi_max_lun = 512;
module_param_named(max_lun, iscsi_max_lun, uint, S_IRUGO);

static int iscsi_sw_tcp_dbg;
module_param_named(debug_iscsi_tcp, iscsi_sw_tcp_dbg, int,
		   S_IRUGO | S_IWUSR);
MODULE_PARM_DESC(debug_iscsi_tcp, "Turn on debugging for iscsi_tcp module "
		 "Set to 1 to turn on, and zero to turn off. Default is off.");

#define ISCSI_SW_TCP_DBG(_conn, dbg_fmt, arg...)		\
	do {							\
		if (iscsi_sw_tcp_dbg)				\
			iscsi_conn_printk(KERN_INFO, _conn,	\
					     "%s " dbg_fmt,	\
					     __func__, ##arg);	\
	} while (0);


/**
 * iscsi_sw_tcp_recv - TCP receive in sendfile fashion
 * @rd_desc: read descriptor
 * @skb: socket buffer
 * @offset: offset in skb
 * @len: skb->len - offset
 */
static int iscsi_sw_tcp_recv(read_descriptor_t *rd_desc, struct sk_buff *skb,
			     unsigned int offset, size_t len)
{
	struct iscsi_conn *conn = rd_desc->arg.data;
	unsigned int consumed, total_consumed = 0;
	int status;

	ISCSI_SW_TCP_DBG(conn, "in %d bytes\n", skb->len - offset);

	do {
		status = 0;
		consumed = iscsi_tcp_recv_skb(conn, skb, offset, 0, &status);
		offset += consumed;
		total_consumed += consumed;
	} while (consumed != 0 && status != ISCSI_TCP_SKB_DONE);

	ISCSI_SW_TCP_DBG(conn, "read %d bytes status %d\n",
			 skb->len - offset, status);
	return total_consumed;
}

static void iscsi_sw_tcp_data_ready(struct sock *sk, int flag)
{
	struct iscsi_conn *conn = sk->sk_user_data;
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	read_descriptor_t rd_desc;

	read_lock(&sk->sk_callback_lock);

	/*
	 * Use rd_desc to pass 'conn' to iscsi_tcp_recv.
	 * We set count to 1 because we want the network layer to
	 * hand us all the skbs that are available. iscsi_tcp_recv
	 * handled pdus that cross buffers or pdus that still need data.
	 */
	rd_desc.arg.data = conn;
	rd_desc.count = 1;
	tcp_read_sock(sk, &rd_desc, iscsi_sw_tcp_recv);

	read_unlock(&sk->sk_callback_lock);

	/* If we had to (atomically) map a highmem page,
	 * unmap it now. */
	iscsi_tcp_segment_unmap(&tcp_conn->in.segment);
}

static void iscsi_sw_tcp_state_change(struct sock *sk)
{
	struct iscsi_tcp_conn *tcp_conn;
	struct iscsi_sw_tcp_conn *tcp_sw_conn;
	struct iscsi_conn *conn;
	struct iscsi_session *session;
	void (*old_state_change)(struct sock *);

	read_lock(&sk->sk_callback_lock);

	conn = (struct iscsi_conn*)sk->sk_user_data;
	session = conn->session;

	if ((sk->sk_state == TCP_CLOSE_WAIT ||
	     sk->sk_state == TCP_CLOSE) &&
	    !atomic_read(&sk->sk_rmem_alloc)) {
		ISCSI_SW_TCP_DBG(conn, "iscsi_tcp_state_change: "
				 "TCP_CLOSE|TCP_CLOSE_WAIT\n");
		iscsi_conn_failure(conn, ISCSI_ERR_CONN_FAILED);
	}

	tcp_conn = conn->dd_data;
	tcp_sw_conn = tcp_conn->dd_data;
	old_state_change = tcp_sw_conn->old_state_change;

	read_unlock(&sk->sk_callback_lock);

	old_state_change(sk);
}

/**
 * iscsi_write_space - Called when more output buffer space is available
 * @sk: socket space is available for
 **/
static void iscsi_sw_tcp_write_space(struct sock *sk)
{
	struct iscsi_conn *conn = (struct iscsi_conn*)sk->sk_user_data;
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;

	tcp_sw_conn->old_write_space(sk);
	ISCSI_SW_TCP_DBG(conn, "iscsi_write_space\n");
	iscsi_conn_queue_work(conn);
}

static void iscsi_sw_tcp_conn_set_callbacks(struct iscsi_conn *conn)
{
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;
	struct sock *sk = tcp_sw_conn->sock->sk;

	/* assign new callbacks */
	write_lock_bh(&sk->sk_callback_lock);
	sk->sk_user_data = conn;
	tcp_sw_conn->old_data_ready = sk->sk_data_ready;
	tcp_sw_conn->old_state_change = sk->sk_state_change;
	tcp_sw_conn->old_write_space = sk->sk_write_space;
	sk->sk_data_ready = iscsi_sw_tcp_data_ready;
	sk->sk_state_change = iscsi_sw_tcp_state_change;
	sk->sk_write_space = iscsi_sw_tcp_write_space;
	write_unlock_bh(&sk->sk_callback_lock);
}

static void
iscsi_sw_tcp_conn_restore_callbacks(struct iscsi_sw_tcp_conn *tcp_sw_conn)
{
	struct sock *sk = tcp_sw_conn->sock->sk;

	/* restore socket callbacks, see also: iscsi_conn_set_callbacks() */
	write_lock_bh(&sk->sk_callback_lock);
	sk->sk_user_data    = NULL;
	sk->sk_data_ready   = tcp_sw_conn->old_data_ready;
	sk->sk_state_change = tcp_sw_conn->old_state_change;
	sk->sk_write_space  = tcp_sw_conn->old_write_space;
	sk->sk_no_check	 = 0;
	write_unlock_bh(&sk->sk_callback_lock);
}

/**
 * iscsi_sw_tcp_xmit_segment - transmit segment
 * @tcp_conn: the iSCSI TCP connection
 * @segment: the buffer to transmnit
 *
 * This function transmits as much of the buffer as
 * the network layer will accept, and returns the number of
 * bytes transmitted.
 *
 * If CRC hashing is enabled, the function will compute the
 * hash as it goes. When the entire segment has been transmitted,
 * it will retrieve the hash value and send it as well.
 */
static int iscsi_sw_tcp_xmit_segment(struct iscsi_tcp_conn *tcp_conn,
				     struct iscsi_segment *segment)
{
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;
	struct socket *sk = tcp_sw_conn->sock;
	unsigned int copied = 0;
	int r = 0;

	while (!iscsi_tcp_segment_done(tcp_conn, segment, 0, r)) {
		struct scatterlist *sg;
		unsigned int offset, copy;
		int flags = 0;

		r = 0;
		offset = segment->copied;
		copy = segment->size - offset;

		if (segment->total_copied + segment->size < segment->total_size)
			flags |= MSG_MORE;

		/* Use sendpage if we can; else fall back to sendmsg */
		if (!segment->data) {
			sg = segment->sg;
			offset += segment->sg_offset + sg->offset;
			r = tcp_sw_conn->sendpage(sk, sg_page(sg), offset,
						  copy, flags);
		} else {
			struct msghdr msg = { .msg_flags = flags };
			struct kvec iov = {
				.iov_base = segment->data + offset,
				.iov_len = copy
			};

			r = kernel_sendmsg(sk, &msg, &iov, 1, copy);
		}

		if (r < 0) {
			iscsi_tcp_segment_unmap(segment);
			if (copied || r == -EAGAIN)
				break;
			return r;
		}
		copied += r;
	}
	return copied;
}

/**
 * iscsi_sw_tcp_xmit - TCP transmit
 **/
static int iscsi_sw_tcp_xmit(struct iscsi_conn *conn)
{
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;
	struct iscsi_segment *segment = &tcp_sw_conn->out.segment;
	unsigned int consumed = 0;
	int rc = 0;

	while (1) {
		rc = iscsi_sw_tcp_xmit_segment(tcp_conn, segment);
		if (rc < 0) {
			rc = ISCSI_ERR_XMIT_FAILED;
			goto error;
		}
		if (rc == 0)
			break;

		consumed += rc;

		if (segment->total_copied >= segment->total_size) {
			if (segment->done != NULL) {
				rc = segment->done(tcp_conn, segment);
				if (rc != 0)
					goto error;
			}
		}
	}

	ISCSI_SW_TCP_DBG(conn, "xmit %d bytes\n", consumed);

	conn->txdata_octets += consumed;
	return consumed;

error:
	/* Transmit error. We could initiate error recovery
	 * here. */
	ISCSI_SW_TCP_DBG(conn, "Error sending PDU, errno=%d\n", rc);
	iscsi_conn_failure(conn, rc);
	return -EIO;
}

/**
 * iscsi_tcp_xmit_qlen - return the number of bytes queued for xmit
 */
static inline int iscsi_sw_tcp_xmit_qlen(struct iscsi_conn *conn)
{
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;
	struct iscsi_segment *segment = &tcp_sw_conn->out.segment;

	return segment->total_copied - segment->total_size;
}

static int iscsi_sw_tcp_pdu_xmit(struct iscsi_task *task)
{
	struct iscsi_conn *conn = task->conn;
	int rc;

	while (iscsi_sw_tcp_xmit_qlen(conn)) {
		rc = iscsi_sw_tcp_xmit(conn);
		if (rc == 0)
			return -EAGAIN;
		if (rc < 0)
			return rc;
	}

	return 0;
}

/*
 * This is called when we're done sending the header.
 * Simply copy the data_segment to the send segment, and return.
 */
static int iscsi_sw_tcp_send_hdr_done(struct iscsi_tcp_conn *tcp_conn,
				      struct iscsi_segment *segment)
{
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;

	tcp_sw_conn->out.segment = tcp_sw_conn->out.data_segment;
	ISCSI_SW_TCP_DBG(tcp_conn->iscsi_conn,
			 "Header done. Next segment size %u total_size %u\n",
			 tcp_sw_conn->out.segment.size,
			 tcp_sw_conn->out.segment.total_size);
	return 0;
}

static void iscsi_sw_tcp_send_hdr_prep(struct iscsi_conn *conn, void *hdr,
				       size_t hdrlen)
{
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;

	ISCSI_SW_TCP_DBG(conn, "%s\n", conn->hdrdgst_en ?
			 "digest enabled" : "digest disabled");

	/* Clear the data segment - needs to be filled in by the
	 * caller using iscsi_tcp_send_data_prep() */
	memset(&tcp_sw_conn->out.data_segment, 0,
	       sizeof(struct iscsi_segment));

	/* If header digest is enabled, compute the CRC and
	 * place the digest into the same buffer. We make
	 * sure that both iscsi_tcp_task and mtask have
	 * sufficient room.
	 */
	if (conn->hdrdgst_en) {
		iscsi_tcp_dgst_header(&tcp_sw_conn->tx_hash, hdr, hdrlen,
				      hdr + hdrlen);
		hdrlen += ISCSI_DIGEST_SIZE;
	}

	/* Remember header pointer for later, when we need
	 * to decide whether there's a payload to go along
	 * with the header. */
	tcp_sw_conn->out.hdr = hdr;

	iscsi_segment_init_linear(&tcp_sw_conn->out.segment, hdr, hdrlen,
				  iscsi_sw_tcp_send_hdr_done, NULL);
}

/*
 * Prepare the send buffer for the payload data.
 * Padding and checksumming will all be taken care
 * of by the iscsi_segment routines.
 */
static int
iscsi_sw_tcp_send_data_prep(struct iscsi_conn *conn, struct scatterlist *sg,
			    unsigned int count, unsigned int offset,
			    unsigned int len)
{
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;
	struct hash_desc *tx_hash = NULL;
	unsigned int hdr_spec_len;

	ISCSI_SW_TCP_DBG(conn, "offset=%d, datalen=%d %s\n", offset, len,
			 conn->datadgst_en ?
			 "digest enabled" : "digest disabled");

	/* Make sure the datalen matches what the caller
	   said he would send. */
	hdr_spec_len = ntoh24(tcp_sw_conn->out.hdr->dlength);
	WARN_ON(iscsi_padded(len) != iscsi_padded(hdr_spec_len));

	if (conn->datadgst_en)
		tx_hash = &tcp_sw_conn->tx_hash;

	return iscsi_segment_seek_sg(&tcp_sw_conn->out.data_segment,
				     sg, count, offset, len,
				     NULL, tx_hash);
}

static void
iscsi_sw_tcp_send_linear_data_prep(struct iscsi_conn *conn, void *data,
				   size_t len)
{
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;
	struct hash_desc *tx_hash = NULL;
	unsigned int hdr_spec_len;

	ISCSI_SW_TCP_DBG(conn, "datalen=%zd %s\n", len, conn->datadgst_en ?
			 "digest enabled" : "digest disabled");

	/* Make sure the datalen matches what the caller
	   said he would send. */
	hdr_spec_len = ntoh24(tcp_sw_conn->out.hdr->dlength);
	WARN_ON(iscsi_padded(len) != iscsi_padded(hdr_spec_len));

	if (conn->datadgst_en)
		tx_hash = &tcp_sw_conn->tx_hash;

	iscsi_segment_init_linear(&tcp_sw_conn->out.data_segment,
				data, len, NULL, tx_hash);
}

static int iscsi_sw_tcp_pdu_init(struct iscsi_task *task,
				 unsigned int offset, unsigned int count)
{
	struct iscsi_conn *conn = task->conn;
	int err = 0;

	iscsi_sw_tcp_send_hdr_prep(conn, task->hdr, task->hdr_len);

	if (!count)
		return 0;

	if (!task->sc)
		iscsi_sw_tcp_send_linear_data_prep(conn, task->data, count);
	else {
		struct scsi_data_buffer *sdb = scsi_out(task->sc);

		err = iscsi_sw_tcp_send_data_prep(conn, sdb->table.sgl,
						  sdb->table.nents, offset,
						  count);
	}

	if (err) {
		iscsi_conn_failure(conn, err);
		return -EIO;
	}
	return 0;
}

static int iscsi_sw_tcp_pdu_alloc(struct iscsi_task *task, uint8_t opcode)
{
	struct iscsi_tcp_task *tcp_task = task->dd_data;

	task->hdr = task->dd_data + sizeof(*tcp_task);
	task->hdr_max = sizeof(struct iscsi_sw_tcp_hdrbuf) - ISCSI_DIGEST_SIZE;
	return 0;
}

static struct iscsi_cls_conn *
iscsi_sw_tcp_conn_create(struct iscsi_cls_session *cls_session,
			 uint32_t conn_idx)
{
	struct iscsi_conn *conn;
	struct iscsi_cls_conn *cls_conn;
	struct iscsi_tcp_conn *tcp_conn;
	struct iscsi_sw_tcp_conn *tcp_sw_conn;

	cls_conn = iscsi_tcp_conn_setup(cls_session, sizeof(*tcp_sw_conn),
					conn_idx);
	if (!cls_conn)
		return NULL;
	conn = cls_conn->dd_data;
	tcp_conn = conn->dd_data;
	tcp_sw_conn = tcp_conn->dd_data;

	tcp_sw_conn->tx_hash.tfm = crypto_alloc_hash("crc32c", 0,
						     CRYPTO_ALG_ASYNC);
	tcp_sw_conn->tx_hash.flags = 0;
	if (IS_ERR(tcp_sw_conn->tx_hash.tfm))
		goto free_conn;

	tcp_sw_conn->rx_hash.tfm = crypto_alloc_hash("crc32c", 0,
						     CRYPTO_ALG_ASYNC);
	tcp_sw_conn->rx_hash.flags = 0;
	if (IS_ERR(tcp_sw_conn->rx_hash.tfm))
		goto free_tx_tfm;
	tcp_conn->rx_hash = &tcp_sw_conn->rx_hash;

	return cls_conn;

free_tx_tfm:
	crypto_free_hash(tcp_sw_conn->tx_hash.tfm);
free_conn:
	iscsi_conn_printk(KERN_ERR, conn,
			  "Could not create connection due to crc32c "
			  "loading error. Make sure the crc32c "
			  "module is built as a module or into the "
			  "kernel\n");
	iscsi_tcp_conn_teardown(cls_conn);
	return NULL;
}

static void iscsi_sw_tcp_release_conn(struct iscsi_conn *conn)
{
	struct iscsi_session *session = conn->session;
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;
	struct socket *sock = tcp_sw_conn->sock;

	if (!sock)
		return;

	sock_hold(sock->sk);
	iscsi_sw_tcp_conn_restore_callbacks(tcp_sw_conn);
	sock_put(sock->sk);

	spin_lock_bh(&session->lock);
	tcp_sw_conn->sock = NULL;
	spin_unlock_bh(&session->lock);
	sockfd_put(sock);
}

static void iscsi_sw_tcp_conn_destroy(struct iscsi_cls_conn *cls_conn)
{
	struct iscsi_conn *conn = cls_conn->dd_data;
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;

	iscsi_sw_tcp_release_conn(conn);

	if (tcp_sw_conn->tx_hash.tfm)
		crypto_free_hash(tcp_sw_conn->tx_hash.tfm);
	if (tcp_sw_conn->rx_hash.tfm)
		crypto_free_hash(tcp_sw_conn->rx_hash.tfm);

	iscsi_tcp_conn_teardown(cls_conn);
}

static void iscsi_sw_tcp_conn_stop(struct iscsi_cls_conn *cls_conn, int flag)
{
	struct iscsi_conn *conn = cls_conn->dd_data;
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;

	/* userspace may have goofed up and not bound us */
	if (!tcp_sw_conn->sock)
		return;
	/*
	 * Make sure our recv side is stopped.
	 * Older tools called conn stop before ep_disconnect
	 * so IO could still be coming in.
	 */
	write_lock_bh(&tcp_sw_conn->sock->sk->sk_callback_lock);
	set_bit(ISCSI_SUSPEND_BIT, &conn->suspend_rx);
	write_unlock_bh(&tcp_sw_conn->sock->sk->sk_callback_lock);

	iscsi_conn_stop(cls_conn, flag);
	iscsi_sw_tcp_release_conn(conn);
}

static int iscsi_sw_tcp_get_addr(struct iscsi_conn *conn, struct socket *sock,
				 char *buf, int *port,
				 int (*getname)(struct socket *,
						struct sockaddr *,
						int *addrlen))
{
	struct sockaddr_storage *addr;
	struct sockaddr_in6 *sin6;
	struct sockaddr_in *sin;
	int rc = 0, len;

	addr = kmalloc(sizeof(*addr), GFP_KERNEL);
	if (!addr)
		return -ENOMEM;

	if (getname(sock, (struct sockaddr *) addr, &len)) {
		rc = -ENODEV;
		goto free_addr;
	}

	switch (addr->ss_family) {
	case AF_INET:
		sin = (struct sockaddr_in *)addr;
		spin_lock_bh(&conn->session->lock);
		sprintf(buf, "%pI4", &sin->sin_addr.s_addr);
		*port = be16_to_cpu(sin->sin_port);
		spin_unlock_bh(&conn->session->lock);
		break;
	case AF_INET6:
		sin6 = (struct sockaddr_in6 *)addr;
		spin_lock_bh(&conn->session->lock);
		sprintf(buf, "%pI6", &sin6->sin6_addr);
		*port = be16_to_cpu(sin6->sin6_port);
		spin_unlock_bh(&conn->session->lock);
		break;
	}
free_addr:
	kfree(addr);
	return rc;
}

static int
iscsi_sw_tcp_conn_bind(struct iscsi_cls_session *cls_session,
		       struct iscsi_cls_conn *cls_conn, uint64_t transport_eph,
		       int is_leading)
{
	struct Scsi_Host *shost = iscsi_session_to_shost(cls_session);
	struct iscsi_host *ihost = shost_priv(shost);
	struct iscsi_conn *conn = cls_conn->dd_data;
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;
	struct sock *sk;
	struct socket *sock;
	int err;

	/* lookup for existing socket */
	sock = sockfd_lookup((int)transport_eph, &err);
	if (!sock) {
		iscsi_conn_printk(KERN_ERR, conn,
				  "sockfd_lookup failed %d\n", err);
		return -EEXIST;
	}
	/*
	 * copy these values now because if we drop the session
	 * userspace may still want to query the values since we will
	 * be using them for the reconnect
	 */
	err = iscsi_sw_tcp_get_addr(conn, sock, conn->portal_address,
				    &conn->portal_port, kernel_getpeername);
	if (err)
		goto free_socket;

	err = iscsi_sw_tcp_get_addr(conn, sock, ihost->local_address,
				    &ihost->local_port, kernel_getsockname);
	if (err)
		goto free_socket;

	err = iscsi_conn_bind(cls_session, cls_conn, is_leading);
	if (err)
		goto free_socket;

	/* bind iSCSI connection and socket */
	tcp_sw_conn->sock = sock;

	/* setup Socket parameters */
	sk = sock->sk;
	sk->sk_reuse = 1;
	sk->sk_sndtimeo = 15 * HZ; /* FIXME: make it configurable */
	sk->sk_allocation = GFP_ATOMIC;

	iscsi_sw_tcp_conn_set_callbacks(conn);
	tcp_sw_conn->sendpage = tcp_sw_conn->sock->ops->sendpage;
	/*
	 * set receive state machine into initial state
	 */
	iscsi_tcp_hdr_recv_prep(tcp_conn);
	return 0;

free_socket:
	sockfd_put(sock);
	return err;
}

static int iscsi_sw_tcp_conn_set_param(struct iscsi_cls_conn *cls_conn,
				       enum iscsi_param param, char *buf,
				       int buflen)
{
	struct iscsi_conn *conn = cls_conn->dd_data;
	struct iscsi_session *session = conn->session;
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;
	int value;

	switch(param) {
	case ISCSI_PARAM_HDRDGST_EN:
		iscsi_set_param(cls_conn, param, buf, buflen);
		break;
	case ISCSI_PARAM_DATADGST_EN:
		iscsi_set_param(cls_conn, param, buf, buflen);
		tcp_sw_conn->sendpage = conn->datadgst_en ?
			sock_no_sendpage : tcp_sw_conn->sock->ops->sendpage;
		break;
	case ISCSI_PARAM_MAX_R2T:
		sscanf(buf, "%d", &value);
		if (value <= 0 || !is_power_of_2(value))
			return -EINVAL;
		if (session->max_r2t == value)
			break;
		iscsi_tcp_r2tpool_free(session);
		iscsi_set_param(cls_conn, param, buf, buflen);
		if (iscsi_tcp_r2tpool_alloc(session))
			return -ENOMEM;
		break;
	default:
		return iscsi_set_param(cls_conn, param, buf, buflen);
	}

	return 0;
}

static int iscsi_sw_tcp_conn_get_param(struct iscsi_cls_conn *cls_conn,
				       enum iscsi_param param, char *buf)
{
	struct iscsi_conn *conn = cls_conn->dd_data;
	int len;

	switch(param) {
	case ISCSI_PARAM_CONN_PORT:
		spin_lock_bh(&conn->session->lock);
		len = sprintf(buf, "%hu\n", conn->portal_port);
		spin_unlock_bh(&conn->session->lock);
		break;
	case ISCSI_PARAM_CONN_ADDRESS:
		spin_lock_bh(&conn->session->lock);
		len = sprintf(buf, "%s\n", conn->portal_address);
		spin_unlock_bh(&conn->session->lock);
		break;
	default:
		return iscsi_conn_get_param(cls_conn, param, buf);
	}

	return len;
}

static void
iscsi_sw_tcp_conn_get_stats(struct iscsi_cls_conn *cls_conn,
			    struct iscsi_stats *stats)
{
	struct iscsi_conn *conn = cls_conn->dd_data;
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_sw_tcp_conn *tcp_sw_conn = tcp_conn->dd_data;

	stats->custom_length = 3;
	strcpy(stats->custom[0].desc, "tx_sendpage_failures");
	stats->custom[0].value = tcp_sw_conn->sendpage_failures_cnt;
	strcpy(stats->custom[1].desc, "rx_discontiguous_hdr");
	stats->custom[1].value = tcp_sw_conn->discontiguous_hdr_cnt;
	strcpy(stats->custom[2].desc, "eh_abort_cnt");
	stats->custom[2].value = conn->eh_abort_cnt;

	iscsi_tcp_conn_get_stats(cls_conn, stats);
}

static struct iscsi_cls_session *
iscsi_sw_tcp_session_create(struct iscsi_endpoint *ep, uint16_t cmds_max,
			    uint16_t qdepth, uint32_t initial_cmdsn,
			    uint32_t *hostno)
{
	struct iscsi_cls_session *cls_session;
	struct iscsi_session *session;
	struct Scsi_Host *shost;

	if (ep) {
		printk(KERN_ERR "iscsi_tcp: invalid ep %p.\n", ep);
		return NULL;
	}

	shost = iscsi_host_alloc(&iscsi_sw_tcp_sht, 0, 1);
	if (!shost)
		return NULL;
	shost->transportt = iscsi_sw_tcp_scsi_transport;
	shost->cmd_per_lun = qdepth;
	shost->max_lun = iscsi_max_lun;
	shost->max_id = 0;
	shost->max_channel = 0;
	shost->max_cmd_len = SCSI_MAX_VARLEN_CDB_SIZE;

	if (iscsi_host_add(shost, NULL))
		goto free_host;
	*hostno = shost->host_no;

	cls_session = iscsi_session_setup(&iscsi_sw_tcp_transport, shost,
					  cmds_max,
					  sizeof(struct iscsi_tcp_task) +
					  sizeof(struct iscsi_sw_tcp_hdrbuf),
					  initial_cmdsn, 0);
	if (!cls_session)
		goto remove_host;
	session = cls_session->dd_data;

	shost->can_queue = session->scsi_cmds_max;
	if (iscsi_tcp_r2tpool_alloc(session))
		goto remove_session;
	return cls_session;

remove_session:
	iscsi_session_teardown(cls_session);
remove_host:
	iscsi_host_remove(shost);
free_host:
	iscsi_host_free(shost);
	return NULL;
}

static void iscsi_sw_tcp_session_destroy(struct iscsi_cls_session *cls_session)
{
	struct Scsi_Host *shost = iscsi_session_to_shost(cls_session);

	iscsi_tcp_r2tpool_free(cls_session->dd_data);
	iscsi_session_teardown(cls_session);

	iscsi_host_remove(shost);
	iscsi_host_free(shost);
}

static int iscsi_sw_tcp_slave_alloc(struct scsi_device *sdev)
{
	set_bit(QUEUE_FLAG_BIDI, &sdev->request_queue->queue_flags);
	return 0;
}

static int iscsi_sw_tcp_slave_configure(struct scsi_device *sdev)
{
	blk_queue_bounce_limit(sdev->request_queue, BLK_BOUNCE_ANY);
	blk_queue_dma_alignment(sdev->request_queue, 0);
	return 0;
}

static struct scsi_host_template iscsi_sw_tcp_sht = {
	.module			= THIS_MODULE,
	.name			= "iSCSI Initiator over TCP/IP",
	.queuecommand           = iscsi_queuecommand,
	.change_queue_depth	= iscsi_change_queue_depth,
	.can_queue		= ISCSI_DEF_XMIT_CMDS_MAX - 1,
	.sg_tablesize		= 4096,
	.max_sectors		= 0xFFFF,
	.cmd_per_lun		= ISCSI_DEF_CMD_PER_LUN,
	.eh_abort_handler       = iscsi_eh_abort,
	.eh_device_reset_handler= iscsi_eh_device_reset,
	.eh_target_reset_handler= iscsi_eh_target_reset,
	.use_clustering         = DISABLE_CLUSTERING,
	.slave_alloc            = iscsi_sw_tcp_slave_alloc,
	.slave_configure        = iscsi_sw_tcp_slave_configure,
	.proc_name		= "iscsi_tcp",
	.this_id		= -1,
};

static struct iscsi_transport iscsi_sw_tcp_transport = {
	.owner			= THIS_MODULE,
	.name			= "tcp",
	.caps			= CAP_RECOVERY_L0 | CAP_MULTI_R2T | CAP_HDRDGST
				  | CAP_DATADGST,
	.param_mask		= ISCSI_MAX_RECV_DLENGTH |
				  ISCSI_MAX_XMIT_DLENGTH |
				  ISCSI_HDRDGST_EN |
				  ISCSI_DATADGST_EN |
				  ISCSI_INITIAL_R2T_EN |
				  ISCSI_MAX_R2T |
				  ISCSI_IMM_DATA_EN |
				  ISCSI_FIRST_BURST |
				  ISCSI_MAX_BURST |
				  ISCSI_PDU_INORDER_EN |
				  ISCSI_DATASEQ_INORDER_EN |
				  ISCSI_ERL |
				  ISCSI_CONN_PORT |
				  ISCSI_CONN_ADDRESS |
				  ISCSI_EXP_STATSN |
				  ISCSI_PERSISTENT_PORT |
				  ISCSI_PERSISTENT_ADDRESS |
				  ISCSI_TARGET_NAME | ISCSI_TPGT |
				  ISCSI_USERNAME | ISCSI_PASSWORD |
				  ISCSI_USERNAME_IN | ISCSI_PASSWORD_IN |
				  ISCSI_FAST_ABORT | ISCSI_ABORT_TMO |
				  ISCSI_LU_RESET_TMO |
				  ISCSI_PING_TMO | ISCSI_RECV_TMO |
				  ISCSI_IFACE_NAME | ISCSI_INITIATOR_NAME,
	.host_param_mask	= ISCSI_HOST_HWADDRESS | ISCSI_HOST_IPADDRESS |
				  ISCSI_HOST_INITIATOR_NAME |
				  ISCSI_HOST_NETDEV_NAME,
	/* session management */
	.create_session		= iscsi_sw_tcp_session_create,
	.destroy_session	= iscsi_sw_tcp_session_destroy,
	/* connection management */
	.create_conn		= iscsi_sw_tcp_conn_create,
	.bind_conn		= iscsi_sw_tcp_conn_bind,
	.destroy_conn		= iscsi_sw_tcp_conn_destroy,
	.set_param		= iscsi_sw_tcp_conn_set_param,
	.get_conn_param		= iscsi_sw_tcp_conn_get_param,
	.get_session_param	= iscsi_session_get_param,
	.start_conn		= iscsi_conn_start,
	.stop_conn		= iscsi_sw_tcp_conn_stop,
	/* iscsi host params */
	.get_host_param		= iscsi_host_get_param,
	.set_host_param		= iscsi_host_set_param,
	/* IO */
	.send_pdu		= iscsi_conn_send_pdu,
	.get_stats		= iscsi_sw_tcp_conn_get_stats,
	/* iscsi task/cmd helpers */
	.init_task		= iscsi_tcp_task_init,
	.xmit_task		= iscsi_tcp_task_xmit,
	.cleanup_task		= iscsi_tcp_cleanup_task,
	/* low level pdu helpers */
	.xmit_pdu		= iscsi_sw_tcp_pdu_xmit,
	.init_pdu		= iscsi_sw_tcp_pdu_init,
	.alloc_pdu		= iscsi_sw_tcp_pdu_alloc,
	/* recovery */
	.session_recovery_timedout = iscsi_session_recovery_timedout,
};

static int __init iscsi_sw_tcp_init(void)
{
	if (iscsi_max_lun < 1) {
		printk(KERN_ERR "iscsi_tcp: Invalid max_lun value of %u\n",
		       iscsi_max_lun);
		return -EINVAL;
	}

	iscsi_sw_tcp_scsi_transport = iscsi_register_transport(
						&iscsi_sw_tcp_transport);
	if (!iscsi_sw_tcp_scsi_transport)
		return -ENODEV;

	return 0;
}

static void __exit iscsi_sw_tcp_exit(void)
{
	iscsi_unregister_transport(&iscsi_sw_tcp_transport);
}

module_init(iscsi_sw_tcp_init);
module_exit(iscsi_sw_tcp_exit);