iscsi_tcp.c

static void
iscsi_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;

	tcp_conn->old_write_space(sk);
	debug_tcp("iscsi_write_space: cid %d\n", conn->id);
	clear_bit(ISCSI_SUSPEND_BIT, &conn->suspend_tx);
	scsi_queue_work(conn->session->host, &conn->xmitwork);
}

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

	/* assign new callbacks */
	write_lock_bh(&sk->sk_callback_lock);
	sk->sk_user_data = conn;
	tcp_conn->old_data_ready = sk->sk_data_ready;
	tcp_conn->old_state_change = sk->sk_state_change;
	tcp_conn->old_write_space = sk->sk_write_space;
	sk->sk_data_ready = iscsi_tcp_data_ready;
	sk->sk_state_change = iscsi_tcp_state_change;
	sk->sk_write_space = iscsi_write_space;
	write_unlock_bh(&sk->sk_callback_lock);
}

static void
iscsi_conn_restore_callbacks(struct iscsi_conn *conn)
{
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct sock *sk = tcp_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_conn->old_data_ready;
	sk->sk_state_change = tcp_conn->old_state_change;
	sk->sk_write_space  = tcp_conn->old_write_space;
	sk->sk_no_check	 = 0;
	write_unlock_bh(&sk->sk_callback_lock);
}

/**
 * iscsi_send - generic send routine
 * @sk: kernel's socket
 * @buf: buffer to write from
 * @size: actual size to write
 * @flags: socket's flags
 */
static inline int
iscsi_send(struct iscsi_conn *conn, struct iscsi_buf *buf, int size, int flags)
{
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct socket *sk = tcp_conn->sock;
	int offset = buf->sg.offset + buf->sent;

	/*
	 * if we got use_sg=0 or are sending something we kmallocd
	 * then we did not have to do kmap (kmap returns page_address)
	 *
	 * if we got use_sg > 0, but had to drop down, we do not
	 * set clustering so this should only happen for that
	 * slab case.
	 */
	if (buf->use_sendmsg)
		return sock_no_sendpage(sk, buf->sg.page, offset, size, flags);
	else
		return tcp_conn->sendpage(sk, buf->sg.page, offset, size,
					  flags);
}

/**
 * iscsi_sendhdr - send PDU Header via tcp_sendpage()
 * @conn: iscsi connection
 * @buf: buffer to write from
 * @datalen: lenght of data to be sent after the header
 *
 * Notes:
 *	(Tx, Fast Path)
 **/
static inline int
iscsi_sendhdr(struct iscsi_conn *conn, struct iscsi_buf *buf, int datalen)
{
	struct iscsi_tcp_conn *tcp_conn;
	int flags = 0; /* MSG_DONTWAIT; */
	int res, size;

	size = buf->sg.length - buf->sent;
	BUG_ON(buf->sent + size > buf->sg.length);
	if (buf->sent + size != buf->sg.length || datalen)
		flags |= MSG_MORE;

	res = iscsi_send(conn, buf, size, flags);
	debug_tcp("sendhdr %d bytes, sent %d res %d\n", size, buf->sent, res);
	if (res >= 0) {
		conn->txdata_octets += res;
		buf->sent += res;
		if (size != res)
			return -EAGAIN;
		return 0;
	} else if (res == -EAGAIN) {
		tcp_conn = conn->dd_data;
		tcp_conn->sendpage_failures_cnt++;
		set_bit(ISCSI_SUSPEND_BIT, &conn->suspend_tx);
	} else if (res == -EPIPE)
		iscsi_conn_failure(conn, ISCSI_ERR_CONN_FAILED);

	return res;
}

/**
 * iscsi_sendpage - send one page of iSCSI Data-Out.
 * @conn: iscsi connection
 * @buf: buffer to write from
 * @count: remaining data
 * @sent: number of bytes sent
 *
 * Notes:
 *	(Tx, Fast Path)
 **/
static inline int
iscsi_sendpage(struct iscsi_conn *conn, struct iscsi_buf *buf,
	       int *count, int *sent)
{
	struct iscsi_tcp_conn *tcp_conn;
	int flags = 0; /* MSG_DONTWAIT; */
	int res, size;

	size = buf->sg.length - buf->sent;
	BUG_ON(buf->sent + size > buf->sg.length);
	if (size > *count)
		size = *count;
	if (buf->sent + size != buf->sg.length || *count != size)
		flags |= MSG_MORE;

	res = iscsi_send(conn, buf, size, flags);
	debug_tcp("sendpage: %d bytes, sent %d left %d sent %d res %d\n",
		  size, buf->sent, *count, *sent, res);
	if (res >= 0) {
		conn->txdata_octets += res;
		buf->sent += res;
		*count -= res;
		*sent += res;
		if (size != res)
			return -EAGAIN;
		return 0;
	} else if (res == -EAGAIN) {
		tcp_conn = conn->dd_data;
		tcp_conn->sendpage_failures_cnt++;
		set_bit(ISCSI_SUSPEND_BIT, &conn->suspend_tx);
	} else if (res == -EPIPE)
		iscsi_conn_failure(conn, ISCSI_ERR_CONN_FAILED);

	return res;
}

static inline void
iscsi_data_digest_init(struct iscsi_tcp_conn *tcp_conn,
		      struct iscsi_cmd_task *ctask)
{
	struct iscsi_tcp_cmd_task *tcp_ctask = ctask->dd_data;

	BUG_ON(!tcp_conn->data_tx_tfm);
	crypto_digest_init(tcp_conn->data_tx_tfm);
	tcp_ctask->digest_count = 4;
}

static int
iscsi_digest_final_send(struct iscsi_conn *conn, struct iscsi_cmd_task *ctask,
			struct iscsi_buf *buf, uint32_t *digest, int final)
{
	struct iscsi_tcp_cmd_task *tcp_ctask = ctask->dd_data;
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	int rc = 0;
	int sent = 0;

	if (final)
		crypto_digest_final(tcp_conn->data_tx_tfm, (u8*)digest);

	iscsi_buf_init_iov(buf, (char*)digest, 4);
	rc = iscsi_sendpage(conn, buf, &tcp_ctask->digest_count, &sent);
	if (rc) {
		tcp_ctask->datadigest = *digest;
		tcp_ctask->xmstate |= XMSTATE_DATA_DIGEST;
	} else
		tcp_ctask->digest_count = 4;
	return rc;
}

/**
 * iscsi_solicit_data_cont - initialize next Data-Out
 * @conn: iscsi connection
 * @ctask: scsi command task
 * @r2t: R2T info
 * @left: bytes left to transfer
 *
 * Notes:
 *	Initialize next Data-Out within this R2T sequence and continue
 *	to process next Scatter-Gather element(if any) of this SCSI command.
 *
 *	Called under connection lock.
 **/
static void
iscsi_solicit_data_cont(struct iscsi_conn *conn, struct iscsi_cmd_task *ctask,
			struct iscsi_r2t_info *r2t, int left)
{
	struct iscsi_tcp_cmd_task *tcp_ctask = ctask->dd_data;
	struct iscsi_data *hdr;
	struct scsi_cmnd *sc = ctask->sc;
	int new_offset;

	hdr = &r2t->dtask.hdr;
	memset(hdr, 0, sizeof(struct iscsi_data));
	hdr->ttt = r2t->ttt;
	hdr->datasn = cpu_to_be32(r2t->solicit_datasn);
	r2t->solicit_datasn++;
	hdr->opcode = ISCSI_OP_SCSI_DATA_OUT;
	memcpy(hdr->lun, ctask->hdr->lun, sizeof(hdr->lun));
	hdr->itt = ctask->hdr->itt;
	hdr->exp_statsn = r2t->exp_statsn;
	new_offset = r2t->data_offset + r2t->sent;
	hdr->offset = cpu_to_be32(new_offset);
	if (left > conn->max_xmit_dlength) {
		hton24(hdr->dlength, conn->max_xmit_dlength);
		r2t->data_count = conn->max_xmit_dlength;
	} else {
		hton24(hdr->dlength, left);
		r2t->data_count = left;
		hdr->flags = ISCSI_FLAG_CMD_FINAL;
	}
	conn->dataout_pdus_cnt++;

	iscsi_buf_init_iov(&r2t->headbuf, (char*)hdr,
			   sizeof(struct iscsi_hdr));

	if (sc->use_sg && !iscsi_buf_left(&r2t->sendbuf)) {
		BUG_ON(tcp_ctask->bad_sg == r2t->sg);
		iscsi_buf_init_sg(&r2t->sendbuf, r2t->sg);
		r2t->sg += 1;
	} else
		iscsi_buf_init_iov(&tcp_ctask->sendbuf,
			    (char*)sc->request_buffer + new_offset,
			    r2t->data_count);
}

static void
iscsi_unsolicit_data_init(struct iscsi_conn *conn, struct iscsi_cmd_task *ctask)
{
	struct iscsi_tcp_cmd_task *tcp_ctask = ctask->dd_data;
	struct iscsi_data_task *dtask;

	dtask = tcp_ctask->dtask = &tcp_ctask->unsol_dtask;
	iscsi_prep_unsolicit_data_pdu(ctask, &dtask->hdr,
				      tcp_ctask->r2t_data_count);
	iscsi_buf_init_iov(&tcp_ctask->headbuf, (char*)&dtask->hdr,
			   sizeof(struct iscsi_hdr));
}

/**
 * iscsi_tcp_cmd_init - Initialize iSCSI SCSI_READ or SCSI_WRITE commands
 * @conn: iscsi connection
 * @ctask: scsi command task
 * @sc: scsi command
 **/
static void
iscsi_tcp_cmd_init(struct iscsi_cmd_task *ctask)
{
	struct scsi_cmnd *sc = ctask->sc;
	struct iscsi_tcp_cmd_task *tcp_ctask = ctask->dd_data;

	BUG_ON(__kfifo_len(tcp_ctask->r2tqueue));

	tcp_ctask->sent = 0;
	tcp_ctask->sg_count = 0;

	if (sc->sc_data_direction == DMA_TO_DEVICE) {
		tcp_ctask->xmstate = XMSTATE_W_HDR;
		tcp_ctask->exp_r2tsn = 0;
		BUG_ON(ctask->total_length == 0);

		if (sc->use_sg) {
			struct scatterlist *sg = sc->request_buffer;

			iscsi_buf_init_sg(&tcp_ctask->sendbuf,
					  &sg[tcp_ctask->sg_count++]);
			tcp_ctask->sg = sg;
			tcp_ctask->bad_sg = sg + sc->use_sg;
		} else
			iscsi_buf_init_iov(&tcp_ctask->sendbuf,
					   sc->request_buffer,
					   sc->request_bufflen);

		if (ctask->imm_count)
			tcp_ctask->xmstate |= XMSTATE_IMM_DATA;

		tcp_ctask->pad_count = ctask->total_length & (ISCSI_PAD_LEN-1);
		if (tcp_ctask->pad_count) {
			tcp_ctask->pad_count = ISCSI_PAD_LEN -
							tcp_ctask->pad_count;
			debug_scsi("write padding %d bytes\n",
				   tcp_ctask->pad_count);
			tcp_ctask->xmstate |= XMSTATE_W_PAD;
		}

		if (ctask->unsol_count)
			tcp_ctask->xmstate |= XMSTATE_UNS_HDR |
						XMSTATE_UNS_INIT;
		tcp_ctask->r2t_data_count = ctask->total_length -
				    ctask->imm_count -
				    ctask->unsol_count;

		debug_scsi("cmd [itt %x total %d imm %d imm_data %d "
			   "r2t_data %d]\n",
			   ctask->itt, ctask->total_length, ctask->imm_count,
			   ctask->unsol_count, tcp_ctask->r2t_data_count);
	} else
		tcp_ctask->xmstate = XMSTATE_R_HDR;

	iscsi_buf_init_iov(&tcp_ctask->headbuf, (char*)ctask->hdr,
			    sizeof(struct iscsi_hdr));
}

/**
 * iscsi_tcp_mtask_xmit - xmit management(immediate) task
 * @conn: iscsi connection
 * @mtask: task management task
 *
 * Notes:
 *	The function can return -EAGAIN in which case caller must
 *	call it again later, or recover. '0' return code means successful
 *	xmit.
 *
 *	Management xmit state machine consists of two states:
 *		IN_PROGRESS_IMM_HEAD - PDU Header xmit in progress
 *		IN_PROGRESS_IMM_DATA - PDU Data xmit in progress
 **/
static int
iscsi_tcp_mtask_xmit(struct iscsi_conn *conn, struct iscsi_mgmt_task *mtask)
{
	struct iscsi_tcp_mgmt_task *tcp_mtask = mtask->dd_data;

	debug_scsi("mtask deq [cid %d state %x itt 0x%x]\n",
		conn->id, tcp_mtask->xmstate, mtask->itt);

	if (tcp_mtask->xmstate & XMSTATE_IMM_HDR) {
		tcp_mtask->xmstate &= ~XMSTATE_IMM_HDR;
		if (mtask->data_count)
			tcp_mtask->xmstate |= XMSTATE_IMM_DATA;
		if (conn->c_stage != ISCSI_CONN_INITIAL_STAGE &&
		    conn->stop_stage != STOP_CONN_RECOVER &&
		    conn->hdrdgst_en)
			iscsi_hdr_digest(conn, &tcp_mtask->headbuf,
					(u8*)tcp_mtask->hdrext);
		if (iscsi_sendhdr(conn, &tcp_mtask->headbuf,
				  mtask->data_count)) {
			tcp_mtask->xmstate |= XMSTATE_IMM_HDR;
			if (mtask->data_count)
				tcp_mtask->xmstate &= ~XMSTATE_IMM_DATA;
			return -EAGAIN;
		}
	}

	if (tcp_mtask->xmstate & XMSTATE_IMM_DATA) {
		BUG_ON(!mtask->data_count);
		tcp_mtask->xmstate &= ~XMSTATE_IMM_DATA;
		/* FIXME: implement.
		 * Virtual buffer could be spreaded across multiple pages...
		 */
		do {
			if (iscsi_sendpage(conn, &tcp_mtask->sendbuf,
				   &mtask->data_count, &tcp_mtask->sent)) {
				tcp_mtask->xmstate |= XMSTATE_IMM_DATA;
				return -EAGAIN;
			}
		} while (mtask->data_count);
	}

	BUG_ON(tcp_mtask->xmstate != XMSTATE_IDLE);
	if (mtask->hdr->itt == cpu_to_be32(ISCSI_RESERVED_TAG)) {
		struct iscsi_session *session = conn->session;

		spin_lock_bh(&session->lock);
		list_del(&conn->mtask->running);
		__kfifo_put(session->mgmtpool.queue, (void*)&conn->mtask,
			    sizeof(void*));
		spin_unlock_bh(&session->lock);
	}
	return 0;
}

static inline int
handle_xmstate_r_hdr(struct iscsi_conn *conn,
		     struct iscsi_tcp_cmd_task *tcp_ctask)
{
	tcp_ctask->xmstate &= ~XMSTATE_R_HDR;
	if (conn->hdrdgst_en)
		iscsi_hdr_digest(conn, &tcp_ctask->headbuf,
				 (u8*)tcp_ctask->hdrext);
	if (!iscsi_sendhdr(conn, &tcp_ctask->headbuf, 0)) {
		BUG_ON(tcp_ctask->xmstate != XMSTATE_IDLE);
		return 0; /* wait for Data-In */
	}
	tcp_ctask->xmstate |= XMSTATE_R_HDR;
	return -EAGAIN;
}

static inline int
handle_xmstate_w_hdr(struct iscsi_conn *conn,
		     struct iscsi_cmd_task *ctask)
{
	struct iscsi_tcp_cmd_task *tcp_ctask = ctask->dd_data;

	tcp_ctask->xmstate &= ~XMSTATE_W_HDR;
	if (conn->hdrdgst_en)
		iscsi_hdr_digest(conn, &tcp_ctask->headbuf,
				 (u8*)tcp_ctask->hdrext);
	if (iscsi_sendhdr(conn, &tcp_ctask->headbuf, ctask->imm_count)) {
		tcp_ctask->xmstate |= XMSTATE_W_HDR;
		return -EAGAIN;
	}
	return 0;
}

static inline int
handle_xmstate_data_digest(struct iscsi_conn *conn,
			   struct iscsi_cmd_task *ctask)
{
	struct iscsi_tcp_cmd_task *tcp_ctask = ctask->dd_data;

	tcp_ctask->xmstate &= ~XMSTATE_DATA_DIGEST;
	debug_tcp("resent data digest 0x%x\n", tcp_ctask->datadigest);
	if (iscsi_digest_final_send(conn, ctask, &tcp_ctask->immbuf,
				    &tcp_ctask->datadigest, 0)) {
		tcp_ctask->xmstate |= XMSTATE_DATA_DIGEST;
		debug_tcp("resent data digest 0x%x fail!\n",
			  tcp_ctask->datadigest);
		return -EAGAIN;
	}
	return 0;
}

static inline int
handle_xmstate_imm_data(struct iscsi_conn *conn, struct iscsi_cmd_task *ctask)
{
	struct iscsi_tcp_cmd_task *tcp_ctask = ctask->dd_data;
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;

	BUG_ON(!ctask->imm_count);
	tcp_ctask->xmstate &= ~XMSTATE_IMM_DATA;

	if (conn->datadgst_en) {
		iscsi_data_digest_init(tcp_conn, ctask);
		tcp_ctask->immdigest = 0;
	}

	for (;;) {
		if (iscsi_sendpage(conn, &tcp_ctask->sendbuf, &ctask->imm_count,
				   &tcp_ctask->sent)) {
			tcp_ctask->xmstate |= XMSTATE_IMM_DATA;
			if (conn->datadgst_en) {
				crypto_digest_final(tcp_conn->data_tx_tfm,
						(u8*)&tcp_ctask->immdigest);
				debug_tcp("tx imm sendpage fail 0x%x\n",
					  tcp_ctask->datadigest);
			}
			return -EAGAIN;
		}
		if (conn->datadgst_en)
			crypto_digest_update(tcp_conn->data_tx_tfm,
					     &tcp_ctask->sendbuf.sg, 1);

		if (!ctask->imm_count)
			break;
		iscsi_buf_init_sg(&tcp_ctask->sendbuf,
				  &tcp_ctask->sg[tcp_ctask->sg_count++]);
	}

	if (conn->datadgst_en && !(tcp_ctask->xmstate & XMSTATE_W_PAD)) {
		if (iscsi_digest_final_send(conn, ctask, &tcp_ctask->immbuf,
				            &tcp_ctask->immdigest, 1)) {
			debug_tcp("sending imm digest 0x%x fail!\n",
				  tcp_ctask->immdigest);
			return -EAGAIN;
		}
		debug_tcp("sending imm digest 0x%x\n", tcp_ctask->immdigest);
	}

	return 0;
}

static inline int
handle_xmstate_uns_hdr(struct iscsi_conn *conn, struct iscsi_cmd_task *ctask)
{
	struct iscsi_tcp_cmd_task *tcp_ctask = ctask->dd_data;
	struct iscsi_data_task *dtask;

	tcp_ctask->xmstate |= XMSTATE_UNS_DATA;
	if (tcp_ctask->xmstate & XMSTATE_UNS_INIT) {
		iscsi_unsolicit_data_init(conn, ctask);
		dtask = tcp_ctask->dtask;
		if (conn->hdrdgst_en)
			iscsi_hdr_digest(conn, &tcp_ctask->headbuf,
					(u8*)dtask->hdrext);
		tcp_ctask->xmstate &= ~XMSTATE_UNS_INIT;
	}
	if (iscsi_sendhdr(conn, &tcp_ctask->headbuf, ctask->data_count)) {
		tcp_ctask->xmstate &= ~XMSTATE_UNS_DATA;
		tcp_ctask->xmstate |= XMSTATE_UNS_HDR;
		return -EAGAIN;
	}

	debug_scsi("uns dout [itt 0x%x dlen %d sent %d]\n",
		   ctask->itt, ctask->unsol_count, tcp_ctask->sent);
	return 0;
}

static inline int
handle_xmstate_uns_data(struct iscsi_conn *conn, struct iscsi_cmd_task *ctask)
{
	struct iscsi_tcp_cmd_task *tcp_ctask = ctask->dd_data;
	struct iscsi_data_task *dtask = tcp_ctask->dtask;
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;

	BUG_ON(!ctask->data_count);
	tcp_ctask->xmstate &= ~XMSTATE_UNS_DATA;

	if (conn->datadgst_en) {
		iscsi_data_digest_init(tcp_conn, ctask);
		dtask->digest = 0;
	}

	for (;;) {
		int start = tcp_ctask->sent;

		if (iscsi_sendpage(conn, &tcp_ctask->sendbuf,
				   &ctask->data_count, &tcp_ctask->sent)) {
			ctask->unsol_count -= tcp_ctask->sent - start;
			tcp_ctask->xmstate |= XMSTATE_UNS_DATA;
			/* will continue with this ctask later.. */
			if (conn->datadgst_en) {
				crypto_digest_final(tcp_conn->data_tx_tfm,
						(u8 *)&dtask->digest);
				debug_tcp("tx uns data fail 0x%x\n",
					  dtask->digest);
			}
			return -EAGAIN;
		}

		BUG_ON(tcp_ctask->sent > ctask->total_length);
		ctask->unsol_count -= tcp_ctask->sent - start;

		/*
		 * XXX:we may run here with un-initial sendbuf.
		 * so pass it
		 */
		if (conn->datadgst_en && tcp_ctask->sent - start > 0)
			crypto_digest_update(tcp_conn->data_tx_tfm,
					     &tcp_ctask->sendbuf.sg, 1);

		if (!ctask->data_count)
			break;
		iscsi_buf_init_sg(&tcp_ctask->sendbuf,
				  &tcp_ctask->sg[tcp_ctask->sg_count++]);
	}
	BUG_ON(ctask->unsol_count < 0);

	/*
	 * Done with the Data-Out. Next, check if we need
	 * to send another unsolicited Data-Out.
	 */
	if (ctask->unsol_count) {
		if (conn->datadgst_en) {
			if (iscsi_digest_final_send(conn, ctask,
						    &dtask->digestbuf,
						    &dtask->digest, 1)) {
				debug_tcp("send uns digest 0x%x fail\n",
					  dtask->digest);
				return -EAGAIN;
			}
			debug_tcp("sending uns digest 0x%x, more uns\n",
				  dtask->digest);
		}
		tcp_ctask->xmstate |= XMSTATE_UNS_INIT;
		return 1;
	}

	if (conn->datadgst_en && !(tcp_ctask->xmstate & XMSTATE_W_PAD)) {
		if (iscsi_digest_final_send(conn, ctask,
					    &dtask->digestbuf,
					    &dtask->digest, 1)) {
			debug_tcp("send last uns digest 0x%x fail\n",
				   dtask->digest);
			return -EAGAIN;
		}
		debug_tcp("sending uns digest 0x%x\n",dtask->digest);
	}

	return 0;
}

static inline int
handle_xmstate_sol_data(struct iscsi_conn *conn, struct iscsi_cmd_task *ctask)
{
	struct iscsi_session *session = conn->session;
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_tcp_cmd_task *tcp_ctask = ctask->dd_data;
	struct iscsi_r2t_info *r2t = tcp_ctask->r2t;
	struct iscsi_data_task *dtask = &r2t->dtask;
	int left;

	tcp_ctask->xmstate &= ~XMSTATE_SOL_DATA;
	tcp_ctask->dtask = dtask;

	if (conn->datadgst_en) {
		iscsi_data_digest_init(tcp_conn, ctask);
		dtask->digest = 0;
	}
solicit_again:
	/*
	 * send Data-Out whitnin this R2T sequence.
	 */
	if (!r2t->data_count)
		goto data_out_done;

	if (iscsi_sendpage(conn, &r2t->sendbuf, &r2t->data_count, &r2t->sent)) {
		tcp_ctask->xmstate |= XMSTATE_SOL_DATA;
		/* will continue with this ctask later.. */
		if (conn->datadgst_en) {
			crypto_digest_final(tcp_conn->data_tx_tfm,
					  (u8 *)&dtask->digest);
			debug_tcp("r2t data send fail 0x%x\n", dtask->digest);
		}
		return -EAGAIN;
	}

	BUG_ON(r2t->data_count < 0);
	if (conn->datadgst_en)
		crypto_digest_update(tcp_conn->data_tx_tfm, &r2t->sendbuf.sg,
				     1);

	if (r2t->data_count) {
		BUG_ON(ctask->sc->use_sg == 0);
		if (!iscsi_buf_left(&r2t->sendbuf)) {
			BUG_ON(tcp_ctask->bad_sg == r2t->sg);
			iscsi_buf_init_sg(&r2t->sendbuf, r2t->sg);
			r2t->sg += 1;
		}
		goto solicit_again;
	}

data_out_done:
	/*
	 * Done with this Data-Out. Next, check if we have
	 * to send another Data-Out for this R2T.
	 */
	BUG_ON(r2t->data_length - r2t->sent < 0);
	left = r2t->data_length - r2t->sent;
	if (left) {
		if (conn->datadgst_en) {
			if (iscsi_digest_final_send(conn, ctask,
						    &dtask->digestbuf,
						    &dtask->digest, 1)) {
				debug_tcp("send r2t data digest 0x%x"
					  "fail\n", dtask->digest);
				return -EAGAIN;
			}
			debug_tcp("r2t data send digest 0x%x\n",
				  dtask->digest);
		}
		iscsi_solicit_data_cont(conn, ctask, r2t, left);
		tcp_ctask->xmstate |= XMSTATE_SOL_DATA;
		tcp_ctask->xmstate &= ~XMSTATE_SOL_HDR;
		return 1;
	}

	/*
	 * Done with this R2T. Check if there are more
	 * outstanding R2Ts ready to be processed.
	 */
	BUG_ON(tcp_ctask->r2t_data_count - r2t->data_length < 0);
	if (conn->datadgst_en) {
		if (iscsi_digest_final_send(conn, ctask, &dtask->digestbuf,
					    &dtask->digest, 1)) {
			debug_tcp("send last r2t data digest 0x%x"
				  "fail\n", dtask->digest);
			return -EAGAIN;
		}
		debug_tcp("r2t done dout digest 0x%x\n", dtask->digest);
	}

	tcp_ctask->r2t_data_count -= r2t->data_length;
	tcp_ctask->r2t = NULL;
	spin_lock_bh(&session->lock);
	__kfifo_put(tcp_ctask->r2tpool.queue, (void*)&r2t, sizeof(void*));
	spin_unlock_bh(&session->lock);
	if (__kfifo_get(tcp_ctask->r2tqueue, (void*)&r2t, sizeof(void*))) {
		tcp_ctask->r2t = r2t;
		tcp_ctask->xmstate |= XMSTATE_SOL_DATA;
		tcp_ctask->xmstate &= ~XMSTATE_SOL_HDR;
		return 1;
	}

	return 0;
}

static inline int
handle_xmstate_w_pad(struct iscsi_conn *conn, struct iscsi_cmd_task *ctask)
{
	struct iscsi_tcp_cmd_task *tcp_ctask = ctask->dd_data;
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct iscsi_data_task *dtask = tcp_ctask->dtask;
	int sent;

	tcp_ctask->xmstate &= ~XMSTATE_W_PAD;
	iscsi_buf_init_iov(&tcp_ctask->sendbuf, (char*)&tcp_ctask->pad,
			    tcp_ctask->pad_count);
	if (iscsi_sendpage(conn, &tcp_ctask->sendbuf, &tcp_ctask->pad_count,
			   &sent)) {
		tcp_ctask->xmstate |= XMSTATE_W_PAD;
		return -EAGAIN;
	}

	if (conn->datadgst_en) {
		crypto_digest_update(tcp_conn->data_tx_tfm,
				     &tcp_ctask->sendbuf.sg, 1);
		/* imm data? */
		if (!dtask) {
			if (iscsi_digest_final_send(conn, ctask,
						    &tcp_ctask->immbuf,
						    &tcp_ctask->immdigest, 1)) {
				debug_tcp("send padding digest 0x%x"
					  "fail!\n", tcp_ctask->immdigest);
				return -EAGAIN;
			}
			debug_tcp("done with padding, digest 0x%x\n",
				  tcp_ctask->datadigest);
		} else {
			if (iscsi_digest_final_send(conn, ctask,
						    &dtask->digestbuf,
						    &dtask->digest, 1)) {
				debug_tcp("send padding digest 0x%x"
				          "fail\n", dtask->digest);
				return -EAGAIN;
			}
			debug_tcp("done with padding, digest 0x%x\n",
				  dtask->digest);
		}
	}

	return 0;
}

static int
iscsi_tcp_ctask_xmit(struct iscsi_conn *conn, struct iscsi_cmd_task *ctask)
{
	struct iscsi_tcp_cmd_task *tcp_ctask = ctask->dd_data;
	int rc = 0;

	debug_scsi("ctask deq [cid %d xmstate %x itt 0x%x]\n",
		conn->id, tcp_ctask->xmstate, ctask->itt);

	/*
	 * serialize with TMF AbortTask
	 */
	if (ctask->mtask)
		return rc;

	if (tcp_ctask->xmstate & XMSTATE_R_HDR) {
		rc = handle_xmstate_r_hdr(conn, tcp_ctask);
		return rc;
	}

	if (tcp_ctask->xmstate & XMSTATE_W_HDR) {
		rc = handle_xmstate_w_hdr(conn, ctask);
		if (rc)
			return rc;
	}

	/* XXX: for data digest xmit recover */
	if (tcp_ctask->xmstate & XMSTATE_DATA_DIGEST) {
		rc = handle_xmstate_data_digest(conn, ctask);
		if (rc)
			return rc;
	}

	if (tcp_ctask->xmstate & XMSTATE_IMM_DATA) {
		rc = handle_xmstate_imm_data(conn, ctask);
		if (rc)
			return rc;
	}

	if (tcp_ctask->xmstate & XMSTATE_UNS_HDR) {
		BUG_ON(!ctask->unsol_count);
		tcp_ctask->xmstate &= ~XMSTATE_UNS_HDR;
unsolicit_head_again:
		rc = handle_xmstate_uns_hdr(conn, ctask);
		if (rc)
			return rc;
	}

	if (tcp_ctask->xmstate & XMSTATE_UNS_DATA) {
		rc = handle_xmstate_uns_data(conn, ctask);
		if (rc == 1)
			goto unsolicit_head_again;
		else if (rc)
			return rc;
		goto done;
	}

	if (tcp_ctask->xmstate & XMSTATE_SOL_HDR) {
		struct iscsi_r2t_info *r2t;

		tcp_ctask->xmstate &= ~XMSTATE_SOL_HDR;
		tcp_ctask->xmstate |= XMSTATE_SOL_DATA;
		if (!tcp_ctask->r2t)
			__kfifo_get(tcp_ctask->r2tqueue, (void*)&tcp_ctask->r2t,
				    sizeof(void*));
solicit_head_again:
		r2t = tcp_ctask->r2t;
		if (conn->hdrdgst_en)
			iscsi_hdr_digest(conn, &r2t->headbuf,
					(u8*)r2t->dtask.hdrext);
		if (iscsi_sendhdr(conn, &r2t->headbuf, r2t->data_count)) {
			tcp_ctask->xmstate &= ~XMSTATE_SOL_DATA;
			tcp_ctask->xmstate |= XMSTATE_SOL_HDR;
			return -EAGAIN;
		}

		debug_scsi("sol dout [dsn %d itt 0x%x dlen %d sent %d]\n",
			r2t->solicit_datasn - 1, ctask->itt, r2t->data_count,
			r2t->sent);
	}

	if (tcp_ctask->xmstate & XMSTATE_SOL_DATA) {
		rc = handle_xmstate_sol_data(conn, ctask);
		if (rc == 1)
			goto solicit_head_again;
		if (rc)
			return rc;
	}

done:
	/*
	 * Last thing to check is whether we need to send write
	 * padding. Note that we check for xmstate equality, not just the bit.
	 */
	if (tcp_ctask->xmstate == XMSTATE_W_PAD)
		rc = handle_xmstate_w_pad(conn, ctask);

	return rc;
}

static struct iscsi_cls_conn *
iscsi_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;

	cls_conn = iscsi_conn_setup(cls_session, conn_idx);
	if (!cls_conn)
		return NULL;
	conn = cls_conn->dd_data;
	/*
	 * due to strange issues with iser these are not set
	 * in iscsi_conn_setup
	 */
	conn->max_recv_dlength = DEFAULT_MAX_RECV_DATA_SEGMENT_LENGTH;

	tcp_conn = kzalloc(sizeof(*tcp_conn), GFP_KERNEL);
	if (!tcp_conn)
		goto tcp_conn_alloc_fail;

	conn->dd_data = tcp_conn;
	tcp_conn->iscsi_conn = conn;
	tcp_conn->in_progress = IN_PROGRESS_WAIT_HEADER;
	/* initial operational parameters */
	tcp_conn->hdr_size = sizeof(struct iscsi_hdr);
	tcp_conn->data_size = DEFAULT_MAX_RECV_DATA_SEGMENT_LENGTH;

	/* allocate initial PDU receive place holder */
	if (tcp_conn->data_size <= PAGE_SIZE)
		tcp_conn->data = kmalloc(tcp_conn->data_size, GFP_KERNEL);
	else
		tcp_conn->data = (void*)__get_free_pages(GFP_KERNEL,
					get_order(tcp_conn->data_size));
	if (!tcp_conn->data)
		goto max_recv_dlenght_alloc_fail;

	return cls_conn;

max_recv_dlenght_alloc_fail:
	kfree(tcp_conn);
tcp_conn_alloc_fail:
	iscsi_conn_teardown(cls_conn);
	return NULL;
}

static void
iscsi_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;
	int digest = 0;

	if (conn->hdrdgst_en || conn->datadgst_en)
		digest = 1;

	iscsi_conn_teardown(cls_conn);

	/* now free tcp_conn */
	if (digest) {
		if (tcp_conn->tx_tfm)
			crypto_free_tfm(tcp_conn->tx_tfm);
		if (tcp_conn->rx_tfm)
			crypto_free_tfm(tcp_conn->rx_tfm);
		if (tcp_conn->data_tx_tfm)
			crypto_free_tfm(tcp_conn->data_tx_tfm);
		if (tcp_conn->data_rx_tfm)
			crypto_free_tfm(tcp_conn->data_rx_tfm);
	}

	/* free conn->data, size = MaxRecvDataSegmentLength */
	if (tcp_conn->data_size <= PAGE_SIZE)
		kfree(tcp_conn->data);
	else
		free_pages((unsigned long)tcp_conn->data,
			   get_order(tcp_conn->data_size));
	kfree(tcp_conn);
}

static int
iscsi_tcp_conn_bind(struct iscsi_cls_session *cls_session,
		    struct iscsi_cls_conn *cls_conn, uint64_t transport_eph,
		    int is_leading)
{
	struct iscsi_conn *conn = cls_conn->dd_data;
	struct iscsi_tcp_conn *tcp_conn = conn->dd_data;
	struct sock *sk;
	struct socket *sock;
	int err;

	/* lookup for existing socket */
	sock = sockfd_lookup((int)transport_eph, &err);
	if (!sock) {
		printk(KERN_ERR "iscsi_tcp: sockfd_lookup failed %d\n", err);
		return -EEXIST;
	}

	err = iscsi_conn_bind(cls_session, cls_conn, is_leading);
	if (err)
		return err;

	/* bind iSCSI connection and socket */
	tcp_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;

	/* FIXME: disable Nagle's algorithm */

	/*
	 * Intercept TCP callbacks for sendfile like receive
	 * processing.
	 */
	conn->recv_lock = &sk->sk_callback_lock;
	iscsi_conn_set_callbacks(conn);
	tcp_conn->sendpage = tcp_conn->sock->ops->sendpage;
	/*
	 * set receive state machine into initial state
	 */
	tcp_conn->in_progress = IN_PROGRESS_WAIT_HEADER;

	return 0;
}

static void
iscsi_tcp_cleanup_ctask(struct iscsi_conn *conn, struct iscsi_cmd_task *ctask)
{
	struct iscsi_tcp_cmd_task *tcp_ctask = ctask->dd_data;
	struct iscsi_r2t_info *r2t;

	/* flush ctask's r2t queues */
	while (__kfifo_get(tcp_ctask->r2tqueue, (void*)&r2t, sizeof(void*)))
		__kfifo_put(tcp_ctask->r2tpool.queue, (void*)&r2t,
			    sizeof(void*));

	__iscsi_ctask_cleanup(conn, ctask);
}

static void
iscsi_tcp_suspend_conn_rx(struct iscsi_conn *conn)
{