target: More core cleanups from AGrover (round 2)

This patch contains the squashed version of second round of target core
cleanups and simplifications and Andy and Co.   It also contains a handful
of fixes to address bugs the original series and other minor cleanups.

Here is the condensed shortlog:

target: Remove unneeded casts to void*
target: Rename get_lun_for_{cmd,tmr} to lookup_{cmd,tmr}_lun
target: Make t_task a member of se_cmd, not a pointer
target: Handle functions returning "-2"
target: Use cmd->se_dev over cmd->se_lun->lun_se_dev
target: Embed qr in struct se_cmd
target: Replace embedded struct se_queue_req with a list_head
target: Rename list_heads that are nodes in struct se_cmd to "*_node"
target: Fold transport_device_setup_cmd() into lookup_{tmr,cmd}_lun()
target: Make t_mem_list and t_mem_list_bidi members of t_task
target: Add comment & cleanup transport_map_sg_to_mem()
target: Remove unneeded checks in transport_free_pages()

(Roland: Fix se_queue_req removal leftovers OOPs)
(nab: Fix transport_lookup_tmr_lun failure case)
(nab: Fix list_empty(&cmd->t_task.t_mem_bidi_list) inversion bugs)

Signed-off-by: Andy Grover <agrover@redhat.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>

19 files changed, 761 insertions, 941 deletions

diff --git a/drivers/target/loopback/tcm_loop.c b/drivers/target/loopback/tcm_loop.c
index 2f19e19..eeb7ee7 100644
--- a/drivers/target/loopback/tcm_loop.c
+++ b/drivers/target/loopback/tcm_loop.c
@@ -118,17 +118,16 @@ static struct se_cmd *tcm_loop_allocate_core_cmd(
 	 * Signal BIDI usage with T_TASK(cmd)->t_tasks_bidi
 	 */
 	if (scsi_bidi_cmnd(sc))
-		se_cmd->t_task->t_tasks_bidi = 1;
+		se_cmd->t_task.t_tasks_bidi = 1;
 	/*
 	 * Locate the struct se_lun pointer and attach it to struct se_cmd
 	 */
-	if (transport_get_lun_for_cmd(se_cmd, tl_cmd->sc->device->lun) < 0) {
+	if (transport_lookup_cmd_lun(se_cmd, tl_cmd->sc->device->lun) < 0) {
 		kmem_cache_free(tcm_loop_cmd_cache, tl_cmd);
 		set_host_byte(sc, DID_NO_CONNECT);
 		return NULL;
 	}
 
-	transport_device_setup_cmd(se_cmd);
 	return se_cmd;
 }
 
@@ -143,17 +142,17 @@ static int tcm_loop_new_cmd_map(struct se_cmd *se_cmd)
 	struct tcm_loop_cmd *tl_cmd = container_of(se_cmd,
 				struct tcm_loop_cmd, tl_se_cmd);
 	struct scsi_cmnd *sc = tl_cmd->sc;
-	void *mem_ptr, *mem_bidi_ptr = NULL;
-	u32 sg_no_bidi = 0;
+	struct scatterlist *sgl_bidi = NULL;
+	u32 sgl_bidi_count = 0;
 	int ret;
 	/*
 	 * Allocate the necessary tasks to complete the received CDB+data
 	 */
-	ret = transport_generic_allocate_tasks(se_cmd, tl_cmd->sc->cmnd);
-	if (ret == -1) {
+	ret = transport_generic_allocate_tasks(se_cmd, sc->cmnd);
+	if (ret == -ENOMEM) {
 		/* Out of Resources */
 		return PYX_TRANSPORT_LU_COMM_FAILURE;
-	} else if (ret == -2) {
+	} else if (ret == -EINVAL) {
 		/*
 		 * Handle case for SAM_STAT_RESERVATION_CONFLICT
 		 */
@@ -165,35 +164,24 @@ static int tcm_loop_new_cmd_map(struct se_cmd *se_cmd)
 		 */
 		return PYX_TRANSPORT_USE_SENSE_REASON;
 	}
+
 	/*
-	 * Setup the struct scatterlist memory from the received
-	 * struct scsi_cmnd.
+	 * For BIDI commands, pass in the extra READ buffer
+	 * to transport_generic_map_mem_to_cmd() below..
 	 */
-	if (scsi_sg_count(sc)) {
-		se_cmd->se_cmd_flags |= SCF_PASSTHROUGH_SG_TO_MEM;
-		mem_ptr = (void *)scsi_sglist(sc);
-		/*
-		 * For BIDI commands, pass in the extra READ buffer
-		 * to transport_generic_map_mem_to_cmd() below..
-		 */
-		if (se_cmd->t_task->t_tasks_bidi) {
-			struct scsi_data_buffer *sdb = scsi_in(sc);
+	if (se_cmd->t_task.t_tasks_bidi) {
+		struct scsi_data_buffer *sdb = scsi_in(sc);
 
-			mem_bidi_ptr = (void *)sdb->table.sgl;
-			sg_no_bidi = sdb->table.nents;
-		}
-	} else {
-		/*
-		 * Used for DMA_NONE
-		 */
-		mem_ptr = NULL;
+		sgl_bidi = sdb->table.sgl;
+		sgl_bidi_count = sdb->table.nents;
 	}
+
 	/*
 	 * Map the SG memory into struct se_mem->page linked list using the same
 	 * physical memory at sg->page_link.
 	 */
-	ret = transport_generic_map_mem_to_cmd(se_cmd, mem_ptr,
-			scsi_sg_count(sc), mem_bidi_ptr, sg_no_bidi);
+	ret = transport_generic_map_mem_to_cmd(se_cmd, scsi_sglist(sc),
+			scsi_sg_count(sc), sgl_bidi, sgl_bidi_count);
 	if (ret < 0)
 		return PYX_TRANSPORT_LU_COMM_FAILURE;
 
@@ -384,14 +372,14 @@ static int tcm_loop_device_reset(struct scsi_cmnd *sc)
 	/*
 	 * Allocate the LUN_RESET TMR
 	 */
-	se_cmd->se_tmr_req = core_tmr_alloc_req(se_cmd, (void *)tl_tmr,
+	se_cmd->se_tmr_req = core_tmr_alloc_req(se_cmd, tl_tmr,
 				TMR_LUN_RESET);
 	if (IS_ERR(se_cmd->se_tmr_req))
 		goto release;
 	/*
 	 * Locate the underlying TCM struct se_lun from sc->device->lun
 	 */
-	if (transport_get_lun_for_tmr(se_cmd, sc->device->lun) < 0)
+	if (transport_lookup_tmr_lun(se_cmd, sc->device->lun) < 0)
 		goto release;
 	/*
 	 * Queue the TMR to TCM Core and sleep waiting for tcm_loop_queue_tm_rsp()
@@ -904,7 +892,7 @@ static int tcm_loop_queue_status(struct se_cmd *se_cmd)
 	   ((se_cmd->se_cmd_flags & SCF_TRANSPORT_TASK_SENSE) ||
 	    (se_cmd->se_cmd_flags & SCF_EMULATED_TASK_SENSE))) {
 
-		memcpy((void *)sc->sense_buffer, (void *)se_cmd->sense_buffer,
+		memcpy(sc->sense_buffer, se_cmd->sense_buffer,
 				SCSI_SENSE_BUFFERSIZE);
 		sc->result = SAM_STAT_CHECK_CONDITION;
 		set_driver_byte(sc, DRIVER_SENSE);
@@ -1054,7 +1042,7 @@ static int tcm_loop_make_nexus(
 	 * transport_register_session()
 	 */
 	__transport_register_session(se_tpg, tl_nexus->se_sess->se_node_acl,
-			tl_nexus->se_sess, (void *)tl_nexus);
+			tl_nexus->se_sess, tl_nexus);
 	tl_tpg->tl_hba->tl_nexus = tl_nexus;
 	printk(KERN_INFO "TCM_Loop_ConfigFS: Established I_T Nexus to emulated"
 		" %s Initiator Port: %s\n", tcm_loop_dump_proto_id(tl_hba),
@@ -1242,7 +1230,7 @@ struct se_portal_group *tcm_loop_make_naa_tpg(
 	 * Register the tl_tpg as a emulated SAS TCM Target Endpoint
 	 */
 	ret = core_tpg_register(&tcm_loop_fabric_configfs->tf_ops,
-			wwn, &tl_tpg->tl_se_tpg, (void *)tl_tpg,
+			wwn, &tl_tpg->tl_se_tpg, tl_tpg,
 			TRANSPORT_TPG_TYPE_NORMAL);
 	if (ret < 0)
 		return ERR_PTR(-ENOMEM);
diff --git a/drivers/target/target_core_alua.c b/drivers/target/target_core_alua.c
index bfc42ad..76abd86 100644
--- a/drivers/target/target_core_alua.c
+++ b/drivers/target/target_core_alua.c
@@ -61,11 +61,11 @@ struct t10_alua_lu_gp *default_lu_gp;
  */
 int core_emulate_report_target_port_groups(struct se_cmd *cmd)
 {
-	struct se_subsystem_dev *su_dev = cmd->se_lun->lun_se_dev->se_sub_dev;
+	struct se_subsystem_dev *su_dev = cmd->se_dev->se_sub_dev;
 	struct se_port *port;
 	struct t10_alua_tg_pt_gp *tg_pt_gp;
 	struct t10_alua_tg_pt_gp_member *tg_pt_gp_mem;
-	unsigned char *buf = (unsigned char *)cmd->t_task->t_task_buf;
+	unsigned char *buf = (unsigned char *)cmd->t_task.t_task_buf;
 	u32 rd_len = 0, off = 4; /* Skip over RESERVED area to first
 				    Target port group descriptor */
 
@@ -151,13 +151,13 @@ int core_emulate_report_target_port_groups(struct se_cmd *cmd)
  */
 int core_emulate_set_target_port_groups(struct se_cmd *cmd)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	struct se_subsystem_dev *su_dev = dev->se_sub_dev;
 	struct se_port *port, *l_port = cmd->se_lun->lun_sep;
 	struct se_node_acl *nacl = cmd->se_sess->se_node_acl;
 	struct t10_alua_tg_pt_gp *tg_pt_gp = NULL, *l_tg_pt_gp;
 	struct t10_alua_tg_pt_gp_member *tg_pt_gp_mem, *l_tg_pt_gp_mem;
-	unsigned char *buf = (unsigned char *)cmd->t_task->t_task_buf;
+	unsigned char *buf = (unsigned char *)cmd->t_task.t_task_buf;
 	unsigned char *ptr = &buf[4]; /* Skip over RESERVED area in header */
 	u32 len = 4; /* Skip over RESERVED area in header */
 	int alua_access_state, primary = 0, rc;
diff --git a/drivers/target/target_core_cdb.c b/drivers/target/target_core_cdb.c
index 7d9ccf3..95195d7 100644
--- a/drivers/target/target_core_cdb.c
+++ b/drivers/target/target_core_cdb.c
@@ -65,8 +65,8 @@ static int
 target_emulate_inquiry_std(struct se_cmd *cmd)
 {
 	struct se_lun *lun = cmd->se_lun;
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
-	unsigned char *buf = cmd->t_task->t_task_buf;
+	struct se_device *dev = cmd->se_dev;
+	unsigned char *buf = cmd->t_task.t_task_buf;
 
 	/*
 	 * Make sure we at least have 6 bytes of INQUIRY response
@@ -128,7 +128,7 @@ target_emulate_evpd_00(struct se_cmd *cmd, unsigned char *buf)
 	 * Registered Extended LUN WWN has been set via ConfigFS
 	 * during device creation/restart.
 	 */
-	if (cmd->se_lun->lun_se_dev->se_sub_dev->su_dev_flags &
+	if (cmd->se_dev->se_sub_dev->su_dev_flags &
 			SDF_EMULATED_VPD_UNIT_SERIAL) {
 		buf[3] = 3;
 		buf[5] = 0x80;
@@ -143,7 +143,7 @@ target_emulate_evpd_00(struct se_cmd *cmd, unsigned char *buf)
 static int
 target_emulate_evpd_80(struct se_cmd *cmd, unsigned char *buf)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	u16 len = 0;
 
 	buf[1] = 0x80;
@@ -176,7 +176,7 @@ target_emulate_evpd_80(struct se_cmd *cmd, unsigned char *buf)
 static int
 target_emulate_evpd_83(struct se_cmd *cmd, unsigned char *buf)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	struct se_lun *lun = cmd->se_lun;
 	struct se_port *port = NULL;
 	struct se_portal_group *tpg = NULL;
@@ -477,7 +477,7 @@ target_emulate_evpd_86(struct se_cmd *cmd, unsigned char *buf)
 	buf[5] = 0x07;
 
 	/* If WriteCache emulation is enabled, set V_SUP */
-	if (cmd->se_lun->lun_se_dev->se_sub_dev->se_dev_attrib.emulate_write_cache > 0)
+	if (cmd->se_dev->se_sub_dev->se_dev_attrib.emulate_write_cache > 0)
 		buf[6] = 0x01;
 	return 0;
 }
@@ -486,7 +486,7 @@ target_emulate_evpd_86(struct se_cmd *cmd, unsigned char *buf)
 static int
 target_emulate_evpd_b0(struct se_cmd *cmd, unsigned char *buf)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	int have_tp = 0;
 
 	/*
@@ -568,7 +568,7 @@ target_emulate_evpd_b0(struct se_cmd *cmd, unsigned char *buf)
 static int
 target_emulate_evpd_b2(struct se_cmd *cmd, unsigned char *buf)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 
 	/*
 	 * From sbc3r22 section 6.5.4 Thin Provisioning VPD page:
@@ -620,9 +620,9 @@ target_emulate_evpd_b2(struct se_cmd *cmd, unsigned char *buf)
 static int
 target_emulate_inquiry(struct se_cmd *cmd)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
-	unsigned char *buf = cmd->t_task->t_task_buf;
-	unsigned char *cdb = cmd->t_task->t_task_cdb;
+	struct se_device *dev = cmd->se_dev;
+	unsigned char *buf = cmd->t_task.t_task_buf;
+	unsigned char *cdb = cmd->t_task.t_task_cdb;
 
 	if (!(cdb[1] & 0x1))
 		return target_emulate_inquiry_std(cmd);
@@ -665,8 +665,8 @@ target_emulate_inquiry(struct se_cmd *cmd)
 static int
 target_emulate_readcapacity(struct se_cmd *cmd)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
-	unsigned char *buf = cmd->t_task->t_task_buf;
+	struct se_device *dev = cmd->se_dev;
+	unsigned char *buf = cmd->t_task.t_task_buf;
 	unsigned long long blocks_long = dev->transport->get_blocks(dev);
 	u32 blocks;
 
@@ -695,8 +695,8 @@ target_emulate_readcapacity(struct se_cmd *cmd)
 static int
 target_emulate_readcapacity_16(struct se_cmd *cmd)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
-	unsigned char *buf = cmd->t_task->t_task_buf;
+	struct se_device *dev = cmd->se_dev;
+	unsigned char *buf = cmd->t_task.t_task_buf;
 	unsigned long long blocks = dev->transport->get_blocks(dev);
 
 	buf[0] = (blocks >> 56) & 0xff;
@@ -830,9 +830,9 @@ target_modesense_dpofua(unsigned char *buf, int type)
 static int
 target_emulate_modesense(struct se_cmd *cmd, int ten)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
-	char *cdb = cmd->t_task->t_task_cdb;
-	unsigned char *rbuf = cmd->t_task->t_task_buf;
+	struct se_device *dev = cmd->se_dev;
+	char *cdb = cmd->t_task.t_task_cdb;
+	unsigned char *rbuf = cmd->t_task.t_task_buf;
 	int type = dev->transport->get_device_type(dev);
 	int offset = (ten) ? 8 : 4;
 	int length = 0;
@@ -903,8 +903,8 @@ target_emulate_modesense(struct se_cmd *cmd, int ten)
 static int
 target_emulate_request_sense(struct se_cmd *cmd)
 {
-	unsigned char *cdb = cmd->t_task->t_task_cdb;
-	unsigned char *buf = cmd->t_task->t_task_buf;
+	unsigned char *cdb = cmd->t_task.t_task_cdb;
+	unsigned char *buf = cmd->t_task.t_task_buf;
 	u8 ua_asc = 0, ua_ascq = 0;
 
 	if (cdb[1] & 0x01) {
@@ -964,9 +964,9 @@ static int
 target_emulate_unmap(struct se_task *task)
 {
 	struct se_cmd *cmd = task->task_se_cmd;
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
-	unsigned char *buf = cmd->t_task->t_task_buf, *ptr = NULL;
-	unsigned char *cdb = &cmd->t_task->t_task_cdb[0];
+	struct se_device *dev = cmd->se_dev;
+	unsigned char *buf = cmd->t_task.t_task_buf, *ptr = NULL;
+	unsigned char *cdb = &cmd->t_task.t_task_cdb[0];
 	sector_t lba;
 	unsigned int size = cmd->data_length, range;
 	int ret, offset;
@@ -1011,8 +1011,8 @@ static int
 target_emulate_write_same(struct se_task *task)
 {
 	struct se_cmd *cmd = task->task_se_cmd;
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
-	sector_t lba = cmd->t_task->t_task_lba;
+	struct se_device *dev = cmd->se_dev;
+	sector_t lba = cmd->t_task.t_task_lba;
 	unsigned int range;
 	int ret;
 
@@ -1036,11 +1036,11 @@ int
 transport_emulate_control_cdb(struct se_task *task)
 {
 	struct se_cmd *cmd = task->task_se_cmd;
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	unsigned short service_action;
 	int ret = 0;
 
-	switch (cmd->t_task->t_task_cdb[0]) {
+	switch (cmd->t_task.t_task_cdb[0]) {
 	case INQUIRY:
 		ret = target_emulate_inquiry(cmd);
 		break;
@@ -1054,13 +1054,13 @@ transport_emulate_control_cdb(struct se_task *task)
 		ret = target_emulate_modesense(cmd, 1);
 		break;
 	case SERVICE_ACTION_IN:
-		switch (cmd->t_task->t_task_cdb[1] & 0x1f) {
+		switch (cmd->t_task.t_task_cdb[1] & 0x1f) {
 		case SAI_READ_CAPACITY_16:
 			ret = target_emulate_readcapacity_16(cmd);
 			break;
 		default:
 			printk(KERN_ERR "Unsupported SA: 0x%02x\n",
-				cmd->t_task->t_task_cdb[1] & 0x1f);
+				cmd->t_task.t_task_cdb[1] & 0x1f);
 			return PYX_TRANSPORT_UNKNOWN_SAM_OPCODE;
 		}
 		break;
@@ -1085,7 +1085,7 @@ transport_emulate_control_cdb(struct se_task *task)
 		break;
 	case VARIABLE_LENGTH_CMD:
 		service_action =
-			get_unaligned_be16(&cmd->t_task->t_task_cdb[8]);
+			get_unaligned_be16(&cmd->t_task.t_task_cdb[8]);
 		switch (service_action) {
 		case WRITE_SAME_32:
 			if (!dev->transport->do_discard) {
@@ -1124,7 +1124,7 @@ transport_emulate_control_cdb(struct se_task *task)
 		break;
 	default:
 		printk(KERN_ERR "Unsupported SCSI Opcode: 0x%02x for %s\n",
-			cmd->t_task->t_task_cdb[0], dev->transport->name);
+			cmd->t_task.t_task_cdb[0], dev->transport->name);
 		return PYX_TRANSPORT_UNKNOWN_SAM_OPCODE;
 	}
 
diff --git a/drivers/target/target_core_configfs.c b/drivers/target/target_core_configfs.c
index 64418ef..ac7f765 100644
--- a/drivers/target/target_core_configfs.c
+++ b/drivers/target/target_core_configfs.c
@@ -2037,7 +2037,7 @@ static ssize_t target_core_dev_show(struct config_item *item,
 	if (!(tc_attr->show))
 		return -EINVAL;
 
-	return tc_attr->show((void *)se_dev, page);
+	return tc_attr->show(se_dev, page);
 }
 
 static ssize_t target_core_dev_store(struct config_item *item,
@@ -2053,7 +2053,7 @@ static ssize_t target_core_dev_store(struct config_item *item,
 	if (!(tc_attr->store))
 		return -EINVAL;
 
-	return tc_attr->store((void *)se_dev, page, count);
+	return tc_attr->store(se_dev, page, count);
 }
 
 static struct configfs_item_operations target_core_dev_item_ops = {
diff --git a/drivers/target/target_core_device.c b/drivers/target/target_core_device.c
index fd92385..ea92f75 100644
--- a/drivers/target/target_core_device.c
+++ b/drivers/target/target_core_device.c
@@ -59,15 +59,12 @@ static struct se_subsystem_dev *lun0_su_dev;
 /* not static, needed by tpg.c */
 struct se_device *g_lun0_dev;
 
-int transport_get_lun_for_cmd(
-	struct se_cmd *se_cmd,
-	u32 unpacked_lun)
+int transport_lookup_cmd_lun(struct se_cmd *se_cmd, u32 unpacked_lun)
 {
-	struct se_dev_entry *deve;
 	struct se_lun *se_lun = NULL;
 	struct se_session *se_sess = se_cmd->se_sess;
+	struct se_device *dev;
 	unsigned long flags;
-	int read_only = 0;
 
 	if (unpacked_lun >= TRANSPORT_MAX_LUNS_PER_TPG) {
 		se_cmd->scsi_sense_reason = TCM_NON_EXISTENT_LUN;
@@ -76,91 +73,87 @@ int transport_get_lun_for_cmd(
 	}
 
 	spin_lock_irq(&se_sess->se_node_acl->device_list_lock);
-	deve = se_cmd->se_deve =
-			&se_sess->se_node_acl->device_list[unpacked_lun];
-	if (deve->lun_flags & TRANSPORT_LUNFLAGS_INITIATOR_ACCESS) {
-		if (se_cmd) {
-			deve->total_cmds++;
-			deve->total_bytes += se_cmd->data_length;
-
-			if (se_cmd->data_direction == DMA_TO_DEVICE) {
-				if (deve->lun_flags &
-						TRANSPORT_LUNFLAGS_READ_ONLY) {
-					read_only = 1;
-					goto out;
-				}
-				deve->write_bytes += se_cmd->data_length;
-			} else if (se_cmd->data_direction ==
-				   DMA_FROM_DEVICE) {
-				deve->read_bytes += se_cmd->data_length;
-			}
+	se_cmd->se_deve = &se_sess->se_node_acl->device_list[unpacked_lun];
+	if (se_cmd->se_deve->lun_flags & TRANSPORT_LUNFLAGS_INITIATOR_ACCESS) {
+		struct se_dev_entry *deve = se_cmd->se_deve;
+
+		deve->total_cmds++;
+		deve->total_bytes += se_cmd->data_length;
+
+		if ((se_cmd->data_direction == DMA_TO_DEVICE) &&
+		    (deve->lun_flags & TRANSPORT_LUNFLAGS_READ_ONLY)) {
+			se_cmd->scsi_sense_reason = TCM_WRITE_PROTECTED;
+			se_cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
+			printk("TARGET_CORE[%s]: Detected WRITE_PROTECTED LUN"
+				" Access for 0x%08x\n",
+				se_cmd->se_tfo->get_fabric_name(),
+				unpacked_lun);
+			spin_unlock_irq(&se_sess->se_node_acl->device_list_lock);
+			return -EACCES;
 		}
+
+		if (se_cmd->data_direction == DMA_TO_DEVICE)
+			deve->write_bytes += se_cmd->data_length;
+		else if (se_cmd->data_direction == DMA_FROM_DEVICE)
+			deve->read_bytes += se_cmd->data_length;
+
 		deve->deve_cmds++;
 
-		se_lun = se_cmd->se_lun = deve->se_lun;
+		se_lun = deve->se_lun;
+		se_cmd->se_lun = deve->se_lun;
 		se_cmd->pr_res_key = deve->pr_res_key;
 		se_cmd->orig_fe_lun = unpacked_lun;
 		se_cmd->se_orig_obj_ptr = se_cmd->se_lun->lun_se_dev;
 		se_cmd->se_cmd_flags |= SCF_SE_LUN_CMD;
 	}
-out:
 	spin_unlock_irq(&se_sess->se_node_acl->device_list_lock);
 
 	if (!se_lun) {
-		if (read_only) {
-			se_cmd->scsi_sense_reason = TCM_WRITE_PROTECTED;
+		/*
+		 * Use the se_portal_group->tpg_virt_lun0 to allow for
+		 * REPORT_LUNS, et al to be returned when no active
+		 * MappedLUN=0 exists for this Initiator Port.
+		 */
+		if (unpacked_lun != 0) {
+			se_cmd->scsi_sense_reason = TCM_NON_EXISTENT_LUN;
 			se_cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
-			printk("TARGET_CORE[%s]: Detected WRITE_PROTECTED LUN"
+			printk("TARGET_CORE[%s]: Detected NON_EXISTENT_LUN"
 				" Access for 0x%08x\n",
 				se_cmd->se_tfo->get_fabric_name(),
 				unpacked_lun);
+			return -ENODEV;
+		}
+		/*
+		 * Force WRITE PROTECT for virtual LUN 0
+		 */
+		if ((se_cmd->data_direction != DMA_FROM_DEVICE) &&
+		    (se_cmd->data_direction != DMA_NONE)) {
+			se_cmd->scsi_sense_reason = TCM_WRITE_PROTECTED;
+			se_cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
 			return -EACCES;
-		} else {
-			/*
-			 * Use the se_portal_group->tpg_virt_lun0 to allow for
-			 * REPORT_LUNS, et al to be returned when no active
-			 * MappedLUN=0 exists for this Initiator Port.
-			 */
-			if (unpacked_lun != 0) {
-				se_cmd->scsi_sense_reason = TCM_NON_EXISTENT_LUN;
-				se_cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
-				printk("TARGET_CORE[%s]: Detected NON_EXISTENT_LUN"
-					" Access for 0x%08x\n",
-					se_cmd->se_tfo->get_fabric_name(),
-					unpacked_lun);
-				return -ENODEV;
-			}
-			/*
-			 * Force WRITE PROTECT for virtual LUN 0
-			 */
-			if ((se_cmd->data_direction != DMA_FROM_DEVICE) &&
-			    (se_cmd->data_direction != DMA_NONE)) {
-				se_cmd->scsi_sense_reason = TCM_WRITE_PROTECTED;
-				se_cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
-				return -EACCES;
-			}
-#if 0
-			printk("TARGET_CORE[%s]: Using virtual LUN0! :-)\n",
-				se_cmd->se_tfo->get_fabric_name());
-#endif
-			se_lun = se_cmd->se_lun = &se_sess->se_tpg->tpg_virt_lun0;
-			se_cmd->orig_fe_lun = 0;
-			se_cmd->se_orig_obj_ptr = se_cmd->se_lun->lun_se_dev;
-			se_cmd->se_cmd_flags |= SCF_SE_LUN_CMD;
 		}
+
+		se_lun = &se_sess->se_tpg->tpg_virt_lun0;
+		se_cmd->se_lun = &se_sess->se_tpg->tpg_virt_lun0;
+		se_cmd->orig_fe_lun = 0;
+		se_cmd->se_orig_obj_ptr = se_cmd->se_lun->lun_se_dev;
+		se_cmd->se_cmd_flags |= SCF_SE_LUN_CMD;
 	}
 	/*
 	 * Determine if the struct se_lun is online.
+	 * FIXME: Check for LUN_RESET + UNIT Attention
 	 */
-/* #warning FIXME: Check for LUN_RESET + UNIT Attention */
 	if (se_dev_check_online(se_lun->lun_se_dev) != 0) {
 		se_cmd->scsi_sense_reason = TCM_NON_EXISTENT_LUN;
 		se_cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
 		return -ENODEV;
 	}
 
-	{
-	struct se_device *dev = se_lun->lun_se_dev;
+	/* Directly associate cmd with se_dev */
+	se_cmd->se_dev = se_lun->lun_se_dev;
+
+	/* TODO: get rid of this and use atomics for stats */
+	dev = se_lun->lun_se_dev;
 	spin_lock_irq(&dev->stats_lock);
 	dev->num_cmds++;
 	if (se_cmd->data_direction == DMA_TO_DEVICE)
@@ -168,30 +161,22 @@ out:
 	else if (se_cmd->data_direction == DMA_FROM_DEVICE)
 		dev->read_bytes += se_cmd->data_length;
 	spin_unlock_irq(&dev->stats_lock);
-	}
 
 	/*
 	 * Add the iscsi_cmd_t to the struct se_lun's cmd list.  This list is used
 	 * for tracking state of struct se_cmds during LUN shutdown events.
 	 */
 	spin_lock_irqsave(&se_lun->lun_cmd_lock, flags);
-	list_add_tail(&se_cmd->se_lun_list, &se_lun->lun_cmd_list);
-	atomic_set(&se_cmd->t_task->transport_lun_active, 1);
-#if 0
-	printk(KERN_INFO "Adding ITT: 0x%08x to LUN LIST[%d]\n",
-		se_cmd->se_tfo->get_task_tag(se_cmd), se_lun->unpacked_lun);
-#endif
+	list_add_tail(&se_cmd->se_lun_node, &se_lun->lun_cmd_list);
+	atomic_set(&se_cmd->t_task.transport_lun_active, 1);
 	spin_unlock_irqrestore(&se_lun->lun_cmd_lock, flags);
 
 	return 0;
 }
-EXPORT_SYMBOL(transport_get_lun_for_cmd);
+EXPORT_SYMBOL(transport_lookup_cmd_lun);
 
-int transport_get_lun_for_tmr(
-	struct se_cmd *se_cmd,
-	u32 unpacked_lun)
+int transport_lookup_tmr_lun(struct se_cmd *se_cmd, u32 unpacked_lun)
 {
-	struct se_device *dev = NULL;
 	struct se_dev_entry *deve;
 	struct se_lun *se_lun = NULL;
 	struct se_session *se_sess = se_cmd->se_sess;
@@ -204,15 +189,16 @@ int transport_get_lun_for_tmr(
 	}
 
 	spin_lock_irq(&se_sess->se_node_acl->device_list_lock);
-	deve = se_cmd->se_deve =
-			&se_sess->se_node_acl->device_list[unpacked_lun];
+	se_cmd->se_deve = &se_sess->se_node_acl->device_list[unpacked_lun];
+	deve = se_cmd->se_deve;
+
 	if (deve->lun_flags & TRANSPORT_LUNFLAGS_INITIATOR_ACCESS) {
-		se_lun = se_cmd->se_lun = se_tmr->tmr_lun = deve->se_lun;
-		dev = se_lun->lun_se_dev;
+		se_tmr->tmr_lun = deve->se_lun;
+		se_cmd->se_lun = deve->se_lun;
+		se_lun = deve->se_lun;
 		se_cmd->pr_res_key = deve->pr_res_key;
 		se_cmd->orig_fe_lun = unpacked_lun;
-		se_cmd->se_orig_obj_ptr = se_cmd->se_lun->lun_se_dev;
-/*		se_cmd->se_cmd_flags |= SCF_SE_LUN_CMD; */
+		se_cmd->se_orig_obj_ptr = se_cmd->se_dev;
 	}
 	spin_unlock_irq(&se_sess->se_node_acl->device_list_lock);
 
@@ -226,21 +212,24 @@ int transport_get_lun_for_tmr(
 	}
 	/*
 	 * Determine if the struct se_lun is online.
+	 * FIXME: Check for LUN_RESET + UNIT Attention
 	 */
-/* #warning FIXME: Check for LUN_RESET + UNIT Attention */
 	if (se_dev_check_online(se_lun->lun_se_dev) != 0) {
 		se_cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
 		return -ENODEV;
 	}
-	se_tmr->tmr_dev = dev;
 
-	spin_lock(&dev->se_tmr_lock);
-	list_add_tail(&se_tmr->tmr_list, &dev->dev_tmr_list);
-	spin_unlock(&dev->se_tmr_lock);
+	/* Directly associate cmd with se_dev */
+	se_cmd->se_dev = se_lun->lun_se_dev;
+	se_tmr->tmr_dev = se_lun->lun_se_dev;
+
+	spin_lock(&se_tmr->tmr_dev->se_tmr_lock);
+	list_add_tail(&se_tmr->tmr_list, &se_tmr->tmr_dev->dev_tmr_list);
+	spin_unlock(&se_tmr->tmr_dev->se_tmr_lock);
 
 	return 0;
 }
-EXPORT_SYMBOL(transport_get_lun_for_tmr);
+EXPORT_SYMBOL(transport_lookup_tmr_lun);
 
 /*
  * This function is called from core_scsi3_emulate_pro_register_and_move()
@@ -667,10 +656,10 @@ int transport_core_report_lun_response(struct se_cmd *se_cmd)
 	struct se_lun *se_lun;
 	struct se_session *se_sess = se_cmd->se_sess;
 	struct se_task *se_task;
-	unsigned char *buf = se_cmd->t_task->t_task_buf;
+	unsigned char *buf = se_cmd->t_task.t_task_buf;
 	u32 cdb_offset = 0, lun_count = 0, offset = 8, i;
 
-	list_for_each_entry(se_task, &se_cmd->t_task->t_task_list, t_list)
+	list_for_each_entry(se_task, &se_cmd->t_task.t_task_list, t_list)
 		break;
 
 	if (!(se_task)) {
diff --git a/drivers/target/target_core_file.c b/drivers/target/target_core_file.c
index 0c44bc0..2e7ea74 100644
--- a/drivers/target/target_core_file.c
+++ b/drivers/target/target_core_file.c
@@ -223,7 +223,7 @@ static struct se_device *fd_create_virtdevice(
 	dev_limits.queue_depth = FD_DEVICE_QUEUE_DEPTH;
 
 	dev = transport_add_device_to_core_hba(hba, &fileio_template,
-				se_dev, dev_flags, (void *)fd_dev,
+				se_dev, dev_flags, fd_dev,
 				&dev_limits, "FILEIO", FD_VERSION);
 	if (!(dev))
 		goto fail;
@@ -279,7 +279,7 @@ fd_alloc_task(struct se_cmd *cmd)
 		return NULL;
 	}
 
-	fd_req->fd_dev = cmd->se_lun->lun_se_dev->dev_ptr;
+	fd_req->fd_dev = cmd->se_dev->dev_ptr;
 
 	return &fd_req->fd_task;
 }
@@ -377,7 +377,7 @@ static void fd_emulate_sync_cache(struct se_task *task)
 	struct se_cmd *cmd = task->task_se_cmd;
 	struct se_device *dev = cmd->se_dev;
 	struct fd_dev *fd_dev = dev->dev_ptr;
-	int immed = (cmd->t_task->t_task_cdb[1] & 0x2);
+	int immed = (cmd->t_task.t_task_cdb[1] & 0x2);
 	loff_t start, end;
 	int ret;
 
@@ -391,11 +391,11 @@ static void fd_emulate_sync_cache(struct se_task *task)
 	/*
 	 * Determine if we will be flushing the entire device.
 	 */
-	if (cmd->t_task->t_task_lba == 0 && cmd->data_length == 0) {
+	if (cmd->t_task.t_task_lba == 0 && cmd->data_length == 0) {
 		start = 0;
 		end = LLONG_MAX;
 	} else {
-		start = cmd->t_task->t_task_lba * dev->se_sub_dev->se_dev_attrib.block_size;
+		start = cmd->t_task.t_task_lba * dev->se_sub_dev->se_dev_attrib.block_size;
 		if (cmd->data_length)
 			end = start + cmd->data_length;
 		else
@@ -475,7 +475,7 @@ static int fd_do_task(struct se_task *task)
 		if (ret > 0 &&
 		    dev->se_sub_dev->se_dev_attrib.emulate_write_cache > 0 &&
 		    dev->se_sub_dev->se_dev_attrib.emulate_fua_write > 0 &&
-		    cmd->t_task->t_tasks_fua) {
+		    cmd->t_task.t_tasks_fua) {
 			/*
 			 * We might need to be a bit smarter here
 			 * and return some sense data to let the initiator
diff --git a/drivers/target/target_core_iblock.c b/drivers/target/target_core_iblock.c
index fb15987..c73baef 100644
--- a/drivers/target/target_core_iblock.c
+++ b/drivers/target/target_core_iblock.c
@@ -74,7 +74,7 @@ static int iblock_attach_hba(struct se_hba *hba, u32 host_id)
 
 	ib_host->iblock_host_id = host_id;
 
-	hba->hba_ptr = (void *) ib_host;
+	hba->hba_ptr = ib_host;
 
 	printk(KERN_INFO "CORE_HBA[%d] - TCM iBlock HBA Driver %s on"
 		" Generic Target Core Stack %s\n", hba->hba_id,
@@ -172,7 +172,7 @@ static struct se_device *iblock_create_virtdevice(
 	ib_dev->ibd_bd = bd;
 
 	dev = transport_add_device_to_core_hba(hba,
-			&iblock_template, se_dev, dev_flags, (void *)ib_dev,
+			&iblock_template, se_dev, dev_flags, ib_dev,
 			&dev_limits, "IBLOCK", IBLOCK_VERSION);
 	if (!(dev))
 		goto failed;
@@ -240,7 +240,7 @@ iblock_alloc_task(struct se_cmd *cmd)
 		return NULL;
 	}
 
-	ib_req->ib_dev = cmd->se_lun->lun_se_dev->dev_ptr;
+	ib_req->ib_dev = cmd->se_dev->dev_ptr;
 	atomic_set(&ib_req->ib_bio_cnt, 0);
 	return &ib_req->ib_task;
 }
@@ -331,7 +331,7 @@ static void iblock_emulate_sync_cache(struct se_task *task)
 {
 	struct se_cmd *cmd = task->task_se_cmd;
 	struct iblock_dev *ib_dev = cmd->se_dev->dev_ptr;
-	int immed = (cmd->t_task->t_task_cdb[1] & 0x2);
+	int immed = (cmd->t_task.t_task_cdb[1] & 0x2);
 	sector_t error_sector;
 	int ret;
 
@@ -400,7 +400,7 @@ static int iblock_do_task(struct se_task *task)
 		 */
 		if (dev->se_sub_dev->se_dev_attrib.emulate_write_cache == 0 ||
 		    (dev->se_sub_dev->se_dev_attrib.emulate_fua_write > 0 &&
-		     task->task_se_cmd->t_task->t_tasks_fua))
+		     task->task_se_cmd->t_task.t_tasks_fua))
 			rw = WRITE_FUA;
 		else
 			rw = WRITE;
@@ -593,7 +593,7 @@ static struct bio *iblock_get_bio(
 	DEBUG_IBLOCK("Allocated bio: %p task_size: %u\n", bio, task->task_size);
 
 	bio->bi_bdev = ib_dev->ibd_bd;
-	bio->bi_private = (void *) task;
+	bio->bi_private = task;
 	bio->bi_destructor = iblock_bio_destructor;
 	bio->bi_end_io = &iblock_bio_done;
 	bio->bi_sector = lba;
@@ -608,7 +608,7 @@ static struct bio *iblock_get_bio(
 static int iblock_map_task_SG(struct se_task *task)
 {
 	struct se_cmd *cmd = task->task_se_cmd;
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	struct iblock_dev *ib_dev = task->se_dev->dev_ptr;
 	struct iblock_req *ib_req = IBLOCK_REQ(task);
 	struct bio *bio = NULL, *hbio = NULL, *tbio = NULL;
diff --git a/drivers/target/target_core_pr.c b/drivers/target/target_core_pr.c
index 27a7525..19406a3 100644
--- a/drivers/target/target_core_pr.c
+++ b/drivers/target/target_core_pr.c
@@ -157,8 +157,8 @@ static int core_scsi2_reservation_reserve(struct se_cmd *cmd)
 	struct se_session *sess = cmd->se_sess;
 	struct se_portal_group *tpg = sess->se_tpg;
 
-	if ((cmd->t_task->t_task_cdb[1] & 0x01) &&
-	    (cmd->t_task->t_task_cdb[1] & 0x02)) {
+	if ((cmd->t_task.t_task_cdb[1] & 0x01) &&
+	    (cmd->t_task.t_task_cdb[1] & 0x02)) {
 		printk(KERN_ERR "LongIO and Obselete Bits set, returning"
 				" ILLEGAL_REQUEST\n");
 		return PYX_TRANSPORT_ILLEGAL_REQUEST;
@@ -216,7 +216,7 @@ int core_scsi2_emulate_crh(struct se_cmd *cmd)
 	struct se_subsystem_dev *su_dev = cmd->se_dev->se_sub_dev;
 	struct t10_pr_registration *pr_reg;
 	struct t10_reservation *pr_tmpl = &su_dev->t10_pr;
-	unsigned char *cdb = &cmd->t_task->t_task_cdb[0];
+	unsigned char *cdb = &cmd->t_task.t_task_cdb[0];
 	int crh = (su_dev->t10_pr.res_type == SPC3_PERSISTENT_RESERVATIONS);
 	int conflict = 0;
 
@@ -1471,7 +1471,7 @@ static int core_scsi3_decode_spec_i_port(
 	int all_tg_pt,
 	int aptpl)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	struct se_port *tmp_port;
 	struct se_portal_group *dest_tpg = NULL, *tmp_tpg;
 	struct se_session *se_sess = cmd->se_sess;
@@ -1482,7 +1482,7 @@ static int core_scsi3_decode_spec_i_port(
 	struct list_head tid_dest_list;
 	struct pr_transport_id_holder *tidh_new, *tidh, *tidh_tmp;
 	struct target_core_fabric_ops *tmp_tf_ops;
-	unsigned char *buf = (unsigned char *)cmd->t_task->t_task_buf;
+	unsigned char *buf = (unsigned char *)cmd->t_task.t_task_buf;
 	unsigned char *ptr, *i_str = NULL, proto_ident, tmp_proto_ident;
 	char *iport_ptr = NULL, dest_iport[64], i_buf[PR_REG_ISID_ID_LEN];
 	u32 tpdl, tid_len = 0;
@@ -1509,7 +1509,7 @@ static int core_scsi3_decode_spec_i_port(
 	tidh_new->dest_node_acl = se_sess->se_node_acl;
 	tidh_new->dest_se_deve = local_se_deve;
 
-	local_pr_reg = __core_scsi3_alloc_registration(cmd->se_lun->lun_se_dev,
+	local_pr_reg = __core_scsi3_alloc_registration(cmd->se_dev,
 				se_sess->se_node_acl, local_se_deve, l_isid,
 				sa_res_key, all_tg_pt, aptpl);
 	if (!(local_pr_reg)) {
@@ -1741,7 +1741,7 @@ static int core_scsi3_decode_spec_i_port(
 		 * and then call __core_scsi3_add_registration() in the
 		 * 2nd loop which will never fail.
 		 */
-		dest_pr_reg = __core_scsi3_alloc_registration(cmd->se_lun->lun_se_dev,
+		dest_pr_reg = __core_scsi3_alloc_registration(cmd->se_dev,
 				dest_node_acl, dest_se_deve, iport_ptr,
 				sa_res_key, all_tg_pt, aptpl);
 		if (!(dest_pr_reg)) {
@@ -1787,7 +1787,7 @@ static int core_scsi3_decode_spec_i_port(
 		prf_isid = core_pr_dump_initiator_port(dest_pr_reg, &i_buf[0],
 						PR_REG_ISID_ID_LEN);
 
-		__core_scsi3_add_registration(cmd->se_lun->lun_se_dev, dest_node_acl,
+		__core_scsi3_add_registration(cmd->se_dev, dest_node_acl,
 					dest_pr_reg, 0, 0);
 
 		printk(KERN_INFO "SPC-3 PR [%s] SPEC_I_PT: Successfully"
@@ -2071,7 +2071,7 @@ static int core_scsi3_emulate_pro_register(
 	int ignore_key)
 {
 	struct se_session *se_sess = cmd->se_sess;
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	struct se_dev_entry *se_deve;
 	struct se_lun *se_lun = cmd->se_lun;
 	struct se_portal_group *se_tpg;
@@ -2117,7 +2117,7 @@ static int core_scsi3_emulate_pro_register(
 			 * Port Endpoint that the PRO was received from on the
 			 * Logical Unit of the SCSI device server.
 			 */
-			ret = core_scsi3_alloc_registration(cmd->se_lun->lun_se_dev,
+			ret = core_scsi3_alloc_registration(cmd->se_dev,
 					se_sess->se_node_acl, se_deve, isid_ptr,
 					sa_res_key, all_tg_pt, aptpl,
 					ignore_key, 0);
@@ -2145,7 +2145,7 @@ static int core_scsi3_emulate_pro_register(
 		 */
 		if (!(aptpl)) {
 			pr_tmpl->pr_aptpl_active = 0;
-			core_scsi3_update_and_write_aptpl(cmd->se_lun->lun_se_dev, NULL, 0);
+			core_scsi3_update_and_write_aptpl(cmd->se_dev, NULL, 0);
 			printk("SPC-3 PR: Set APTPL Bit Deactivated for"
 					" REGISTER\n");
 			return 0;
@@ -2155,10 +2155,10 @@ static int core_scsi3_emulate_pro_register(
 		 * update the APTPL metadata information using its
 		 * preallocated *pr_reg->pr_aptpl_buf.
 		 */
-		pr_reg = core_scsi3_locate_pr_reg(cmd->se_lun->lun_se_dev,
+		pr_reg = core_scsi3_locate_pr_reg(cmd->se_dev,
 				se_sess->se_node_acl, se_sess);
 
-		ret = core_scsi3_update_and_write_aptpl(cmd->se_lun->lun_se_dev,
+		ret = core_scsi3_update_and_write_aptpl(cmd->se_dev,
 				&pr_reg->pr_aptpl_buf[0],
 				pr_tmpl->pr_aptpl_buf_len);
 		if (!(ret)) {
@@ -2223,7 +2223,7 @@ static int core_scsi3_emulate_pro_register(
 		 */
 		if (!(sa_res_key)) {
 			pr_holder = core_scsi3_check_implict_release(
-					cmd->se_lun->lun_se_dev, pr_reg);
+					cmd->se_dev, pr_reg);
 			if (pr_holder < 0) {
 				kfree(pr_aptpl_buf);
 				core_scsi3_put_pr_reg(pr_reg);
@@ -2260,7 +2260,7 @@ static int core_scsi3_emulate_pro_register(
 			/*
 			 * Release the calling I_T Nexus registration now..
 			 */
-			__core_scsi3_free_registration(cmd->se_lun->lun_se_dev, pr_reg,
+			__core_scsi3_free_registration(cmd->se_dev, pr_reg,
 							NULL, 1);
 			/*
 			 * From spc4r17, section 5.7.11.3 Unregistering
@@ -2315,7 +2315,7 @@ static int core_scsi3_emulate_pro_register(
 			 * READ_KEYS service action.
 			 */
 			pr_reg->pr_res_generation = core_scsi3_pr_generation(
-							cmd->se_lun->lun_se_dev);
+							cmd->se_dev);
 			pr_reg->pr_res_key = sa_res_key;
 			printk("SPC-3 PR [%s] REGISTER%s: Changed Reservation"
 				" Key for %s to: 0x%016Lx PRgeneration:"
@@ -2398,7 +2398,7 @@ static int core_scsi3_pro_reserve(
 	/*
 	 * Locate the existing *pr_reg via struct se_node_acl pointers
 	 */
-	pr_reg = core_scsi3_locate_pr_reg(cmd->se_lun->lun_se_dev, se_sess->se_node_acl,
+	pr_reg = core_scsi3_locate_pr_reg(cmd->se_dev, se_sess->se_node_acl,
 				se_sess);
 	if (!(pr_reg)) {
 		printk(KERN_ERR "SPC-3 PR: Unable to locate"
@@ -2527,7 +2527,7 @@ static int core_scsi3_pro_reserve(
 	spin_unlock(&dev->dev_reservation_lock);
 
 	if (pr_tmpl->pr_aptpl_active) {
-		ret = core_scsi3_update_and_write_aptpl(cmd->se_lun->lun_se_dev,
+		ret = core_scsi3_update_and_write_aptpl(cmd->se_dev,
 				&pr_reg->pr_aptpl_buf[0],
 				pr_tmpl->pr_aptpl_buf_len);
 		if (!(ret))
@@ -2758,7 +2758,7 @@ static int core_scsi3_emulate_pro_release(
 
 write_aptpl:
 	if (pr_tmpl->pr_aptpl_active) {
-		ret = core_scsi3_update_and_write_aptpl(cmd->se_lun->lun_se_dev,
+		ret = core_scsi3_update_and_write_aptpl(cmd->se_dev,
 				&pr_reg->pr_aptpl_buf[0],
 				pr_tmpl->pr_aptpl_buf_len);
 		if (!(ret))
@@ -2783,7 +2783,7 @@ static int core_scsi3_emulate_pro_clear(
 	/*
 	 * Locate the existing *pr_reg via struct se_node_acl pointers
 	 */
-	pr_reg_n = core_scsi3_locate_pr_reg(cmd->se_lun->lun_se_dev,
+	pr_reg_n = core_scsi3_locate_pr_reg(cmd->se_dev,
 			se_sess->se_node_acl, se_sess);
 	if (!(pr_reg_n)) {
 		printk(KERN_ERR "SPC-3 PR: Unable to locate"
@@ -2849,7 +2849,7 @@ static int core_scsi3_emulate_pro_clear(
 		cmd->se_tfo->get_fabric_name());
 
 	if (pr_tmpl->pr_aptpl_active) {
-		core_scsi3_update_and_write_aptpl(cmd->se_lun->lun_se_dev, NULL, 0);
+		core_scsi3_update_and_write_aptpl(cmd->se_dev, NULL, 0);
 		printk(KERN_INFO "SPC-3 PR: Updated APTPL metadata"
 				" for CLEAR\n");
 	}
@@ -2954,7 +2954,7 @@ static int core_scsi3_pro_preempt(
 	u64 sa_res_key,
 	int abort)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	struct se_dev_entry *se_deve;
 	struct se_node_acl *pr_reg_nacl;
 	struct se_session *se_sess = cmd->se_sess;
@@ -2969,7 +2969,7 @@ static int core_scsi3_pro_preempt(
 		return PYX_TRANSPORT_LU_COMM_FAILURE;
 
 	se_deve = &se_sess->se_node_acl->device_list[cmd->orig_fe_lun];
-	pr_reg_n = core_scsi3_locate_pr_reg(cmd->se_lun->lun_se_dev, se_sess->se_node_acl,
+	pr_reg_n = core_scsi3_locate_pr_reg(cmd->se_dev, se_sess->se_node_acl,
 				se_sess);
 	if (!(pr_reg_n)) {
 		printk(KERN_ERR "SPC-3 PR: Unable to locate"
@@ -3111,7 +3111,7 @@ static int core_scsi3_pro_preempt(
 		spin_unlock(&dev->dev_reservation_lock);
 
 		if (pr_tmpl->pr_aptpl_active) {
-			ret = core_scsi3_update_and_write_aptpl(cmd->se_lun->lun_se_dev,
+			ret = core_scsi3_update_and_write_aptpl(cmd->se_dev,
 					&pr_reg_n->pr_aptpl_buf[0],
 					pr_tmpl->pr_aptpl_buf_len);
 			if (!(ret))
@@ -3121,7 +3121,7 @@ static int core_scsi3_pro_preempt(
 		}
 
 		core_scsi3_put_pr_reg(pr_reg_n);
-		core_scsi3_pr_generation(cmd->se_lun->lun_se_dev);
+		core_scsi3_pr_generation(cmd->se_dev);
 		return 0;
 	}
 	/*
@@ -3247,7 +3247,7 @@ static int core_scsi3_pro_preempt(
 	}
 
 	if (pr_tmpl->pr_aptpl_active) {
-		ret = core_scsi3_update_and_write_aptpl(cmd->se_lun->lun_se_dev,
+		ret = core_scsi3_update_and_write_aptpl(cmd->se_dev,
 				&pr_reg_n->pr_aptpl_buf[0],
 				pr_tmpl->pr_aptpl_buf_len);
 		if (!(ret))
@@ -3256,7 +3256,7 @@ static int core_scsi3_pro_preempt(
 	}
 
 	core_scsi3_put_pr_reg(pr_reg_n);
-	core_scsi3_pr_generation(cmd->se_lun->lun_se_dev);
+	core_scsi3_pr_generation(cmd->se_dev);
 	return 0;
 }
 
@@ -3298,7 +3298,7 @@ static int core_scsi3_emulate_pro_register_and_move(
 	int unreg)
 {
 	struct se_session *se_sess = cmd->se_sess;
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	struct se_dev_entry *se_deve, *dest_se_deve = NULL;
 	struct se_lun *se_lun = cmd->se_lun;
 	struct se_node_acl *pr_res_nacl, *pr_reg_nacl, *dest_node_acl = NULL;
@@ -3307,7 +3307,7 @@ static int core_scsi3_emulate_pro_register_and_move(
 	struct target_core_fabric_ops *dest_tf_ops = NULL, *tf_ops;
 	struct t10_pr_registration *pr_reg, *pr_res_holder, *dest_pr_reg;
 	struct t10_reservation *pr_tmpl = &dev->se_sub_dev->t10_pr;
-	unsigned char *buf = (unsigned char *)cmd->t_task->t_task_buf;
+	unsigned char *buf = (unsigned char *)cmd->t_task.t_task_buf;
 	unsigned char *initiator_str;
 	char *iport_ptr = NULL, dest_iport[64], i_buf[PR_REG_ISID_ID_LEN];
 	u32 tid_len, tmp_tid_len;
@@ -3330,7 +3330,7 @@ static int core_scsi3_emulate_pro_register_and_move(
 	 *
 	 * Locate the existing *pr_reg via struct se_node_acl pointers
 	 */
-	pr_reg = core_scsi3_locate_pr_reg(cmd->se_lun->lun_se_dev, se_sess->se_node_acl,
+	pr_reg = core_scsi3_locate_pr_reg(cmd->se_dev, se_sess->se_node_acl,
 				se_sess);
 	if (!(pr_reg)) {
 		printk(KERN_ERR "SPC-3 PR: Unable to locate PR_REGISTERED"
@@ -3612,7 +3612,7 @@ after_iport_check:
 	dest_pr_reg = __core_scsi3_locate_pr_reg(dev, dest_node_acl,
 					iport_ptr);
 	if (!(dest_pr_reg)) {
-		ret = core_scsi3_alloc_registration(cmd->se_lun->lun_se_dev,
+		ret = core_scsi3_alloc_registration(cmd->se_dev,
 				dest_node_acl, dest_se_deve, iport_ptr,
 				sa_res_key, 0, aptpl, 2, 1);
 		if (ret != 0) {
@@ -3683,12 +3683,12 @@ after_iport_check:
 	 */
 	if (!(aptpl)) {
 		pr_tmpl->pr_aptpl_active = 0;
-		core_scsi3_update_and_write_aptpl(cmd->se_lun->lun_se_dev, NULL, 0);
+		core_scsi3_update_and_write_aptpl(cmd->se_dev, NULL, 0);
 		printk("SPC-3 PR: Set APTPL Bit Deactivated for"
 				" REGISTER_AND_MOVE\n");
 	} else {
 		pr_tmpl->pr_aptpl_active = 1;
-		ret = core_scsi3_update_and_write_aptpl(cmd->se_lun->lun_se_dev,
+		ret = core_scsi3_update_and_write_aptpl(cmd->se_dev,
 				&dest_pr_reg->pr_aptpl_buf[0],
 				pr_tmpl->pr_aptpl_buf_len);
 		if (!(ret))
@@ -3723,7 +3723,7 @@ static unsigned long long core_scsi3_extract_reservation_key(unsigned char *cdb)
  */
 static int core_scsi3_emulate_pr_out(struct se_cmd *cmd, unsigned char *cdb)
 {
-	unsigned char *buf = (unsigned char *)cmd->t_task->t_task_buf;
+	unsigned char *buf = (unsigned char *)cmd->t_task.t_task_buf;
 	u64 res_key, sa_res_key;
 	int sa, scope, type, aptpl;
 	int spec_i_pt = 0, all_tg_pt = 0, unreg = 0;
@@ -3827,10 +3827,10 @@ static int core_scsi3_emulate_pr_out(struct se_cmd *cmd, unsigned char *cdb)
  */
 static int core_scsi3_pri_read_keys(struct se_cmd *cmd)
 {
-	struct se_device *se_dev = cmd->se_lun->lun_se_dev;
+	struct se_device *se_dev = cmd->se_dev;
 	struct se_subsystem_dev *su_dev = se_dev->se_sub_dev;
 	struct t10_pr_registration *pr_reg;
-	unsigned char *buf = (unsigned char *)cmd->t_task->t_task_buf;
+	unsigned char *buf = (unsigned char *)cmd->t_task.t_task_buf;
 	u32 add_len = 0, off = 8;
 
 	if (cmd->data_length < 8) {
@@ -3882,10 +3882,10 @@ static int core_scsi3_pri_read_keys(struct se_cmd *cmd)
  */
 static int core_scsi3_pri_read_reservation(struct se_cmd *cmd)
 {
-	struct se_device *se_dev = cmd->se_lun->lun_se_dev;
+	struct se_device *se_dev = cmd->se_dev;
 	struct se_subsystem_dev *su_dev = se_dev->se_sub_dev;
 	struct t10_pr_registration *pr_reg;
-	unsigned char *buf = (unsigned char *)cmd->t_task->t_task_buf;
+	unsigned char *buf = (unsigned char *)cmd->t_task.t_task_buf;
 	u64 pr_res_key;
 	u32 add_len = 16; /* Hardcoded to 16 when a reservation is held. */
 
@@ -3963,9 +3963,9 @@ static int core_scsi3_pri_read_reservation(struct se_cmd *cmd)
  */
 static int core_scsi3_pri_report_capabilities(struct se_cmd *cmd)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	struct t10_reservation *pr_tmpl = &dev->se_sub_dev->t10_pr;
-	unsigned char *buf = (unsigned char *)cmd->t_task->t_task_buf;
+	unsigned char *buf = (unsigned char *)cmd->t_task.t_task_buf;
 	u16 add_len = 8; /* Hardcoded to 8. */
 
 	if (cmd->data_length < 6) {
@@ -4014,13 +4014,13 @@ static int core_scsi3_pri_report_capabilities(struct se_cmd *cmd)
  */
 static int core_scsi3_pri_read_full_status(struct se_cmd *cmd)
 {
-	struct se_device *se_dev = cmd->se_lun->lun_se_dev;
+	struct se_device *se_dev = cmd->se_dev;
 	struct se_node_acl *se_nacl;
 	struct se_subsystem_dev *su_dev = se_dev->se_sub_dev;
 	struct se_portal_group *se_tpg;
 	struct t10_pr_registration *pr_reg, *pr_reg_tmp;
 	struct t10_reservation *pr_tmpl = &se_dev->se_sub_dev->t10_pr;
-	unsigned char *buf = (unsigned char *)cmd->t_task->t_task_buf;
+	unsigned char *buf = (unsigned char *)cmd->t_task.t_task_buf;
 	u32 add_desc_len = 0, add_len = 0, desc_len, exp_desc_len;
 	u32 off = 8; /* off into first Full Status descriptor */
 	int format_code = 0;
@@ -4174,7 +4174,7 @@ static int core_scsi3_emulate_pr_in(struct se_cmd *cmd, unsigned char *cdb)
 
 int core_scsi3_emulate_pr(struct se_cmd *cmd)
 {
-	unsigned char *cdb = &cmd->t_task->t_task_cdb[0];
+	unsigned char *cdb = &cmd->t_task.t_task_cdb[0];
 	struct se_device *dev = cmd->se_dev;
 	/*
 	 * Following spc2r20 5.5.1 Reservations overview:
diff --git a/drivers/target/target_core_pscsi.c b/drivers/target/target_core_pscsi.c
index 44a79a5..ecfe889 100644
--- a/drivers/target/target_core_pscsi.c
+++ b/drivers/target/target_core_pscsi.c
@@ -72,7 +72,7 @@ static int pscsi_attach_hba(struct se_hba *hba, u32 host_id)
 	phv->phv_host_id = host_id;
 	phv->phv_mode = PHV_VIRUTAL_HOST_ID;
 
-	hba->hba_ptr = (void *)phv;
+	hba->hba_ptr = phv;
 
 	printk(KERN_INFO "CORE_HBA[%d] - TCM SCSI HBA Driver %s on"
 		" Generic Target Core Stack %s\n", hba->hba_id,
@@ -355,7 +355,7 @@ static struct se_device *pscsi_add_device_to_list(
 	pdv->pdv_sd = sd;
 
 	dev = transport_add_device_to_core_hba(hba, &pscsi_template,
-				se_dev, dev_flags, (void *)pdv,
+				se_dev, dev_flags, pdv,
 				&dev_limits, NULL, NULL);
 	if (!(dev)) {
 		pdv->pdv_sd = NULL;
@@ -394,7 +394,7 @@ static void *pscsi_allocate_virtdevice(struct se_hba *hba, const char *name)
 	pdv->pdv_se_hba = hba;
 
 	printk(KERN_INFO "PSCSI: Allocated pdv: %p for %s\n", pdv, name);
-	return (void *)pdv;
+	return pdv;
 }
 
 /*
@@ -697,7 +697,7 @@ static int pscsi_transport_complete(struct se_task *task)
 
 		if (task->task_se_cmd->se_deve->lun_flags &
 				TRANSPORT_LUNFLAGS_READ_ONLY) {
-			unsigned char *buf = task->task_se_cmd->t_task->t_task_buf;
+			unsigned char *buf = task->task_se_cmd->t_task.t_task_buf;
 
 			if (cdb[0] == MODE_SENSE_10) {
 				if (!(buf[3] & 0x80))
@@ -763,7 +763,7 @@ static struct se_task *
 pscsi_alloc_task(struct se_cmd *cmd)
 {
 	struct pscsi_plugin_task *pt;
-	unsigned char *cdb = cmd->t_task->t_task_cdb;
+	unsigned char *cdb = cmd->t_task.t_task_cdb;
 
 	pt = kzalloc(sizeof(struct pscsi_plugin_task), GFP_KERNEL);
 	if (!pt) {
@@ -776,7 +776,7 @@ pscsi_alloc_task(struct se_cmd *cmd)
 	 * allocate the extended CDB buffer for per struct se_task context
 	 * pt->pscsi_cdb now.
 	 */
-	if (cmd->t_task->t_task_cdb != cmd->t_task->__t_task_cdb) {
+	if (cmd->t_task.t_task_cdb != cmd->t_task.__t_task_cdb) {
 
 		pt->pscsi_cdb = kzalloc(scsi_command_size(cdb), GFP_KERNEL);
 		if (!(pt->pscsi_cdb)) {
@@ -812,7 +812,7 @@ static inline void pscsi_blk_init_request(
 	 * also set the end_io_data pointer.to struct se_task.
 	 */
 	req->end_io = pscsi_req_done;
-	req->end_io_data = (void *)task;
+	req->end_io_data = task;
 	/*
 	 * Load the referenced struct se_task's SCSI CDB into
 	 * include/linux/blkdev.h:struct request->cmd
@@ -822,7 +822,7 @@ static inline void pscsi_blk_init_request(
 	/*
 	 * Setup pointer for outgoing sense data.
 	 */
-	req->sense = (void *)&pt->pscsi_sense[0];
+	req->sense = &pt->pscsi_sense[0];
 	req->sense_len = 0;
 }
 
@@ -889,7 +889,7 @@ static void pscsi_free_task(struct se_task *task)
 	 * Release the extended CDB allocation from pscsi_alloc_task()
 	 * if one exists.
 	 */
-	if (cmd->t_task->t_task_cdb != cmd->t_task->__t_task_cdb)
+	if (cmd->t_task.t_task_cdb != cmd->t_task.__t_task_cdb)
 		kfree(pt->pscsi_cdb);
 	/*
 	 * We do not release the bio(s) here associated with this task, as
@@ -1266,7 +1266,7 @@ static int pscsi_map_task_non_SG(struct se_task *task)
 		return 0;
 
 	ret = blk_rq_map_kern(pdv->pdv_sd->request_queue,
-			pt->pscsi_req, cmd->t_task->t_task_buf,
+			pt->pscsi_req, cmd->t_task.t_task_buf,
 			task->task_size, GFP_KERNEL);
 	if (ret < 0) {
 		printk(KERN_ERR "PSCSI: blk_rq_map_kern() failed: %d\n", ret);
diff --git a/drivers/target/target_core_rd.c b/drivers/target/target_core_rd.c
index fbf06c3..384a8e2 100644
--- a/drivers/target/target_core_rd.c
+++ b/drivers/target/target_core_rd.c
@@ -66,7 +66,7 @@ static int rd_attach_hba(struct se_hba *hba, u32 host_id)
 
 	rd_host->rd_host_id = host_id;
 
-	hba->hba_ptr = (void *) rd_host;
+	hba->hba_ptr = rd_host;
 
 	printk(KERN_INFO "CORE_HBA[%d] - TCM Ramdisk HBA Driver %s on"
 		" Generic Target Core Stack %s\n", hba->hba_id,
@@ -271,7 +271,7 @@ static struct se_device *rd_create_virtdevice(
 
 	dev = transport_add_device_to_core_hba(hba,
 			(rd_dev->rd_direct) ? &rd_dr_template :
-			&rd_mcp_template, se_dev, dev_flags, (void *)rd_dev,
+			&rd_mcp_template, se_dev, dev_flags, rd_dev,
 			&dev_limits, prod, rev);
 	if (!(dev))
 		goto fail;
@@ -336,7 +336,7 @@ rd_alloc_task(struct se_cmd *cmd)
 		printk(KERN_ERR "Unable to allocate struct rd_request\n");
 		return NULL;
 	}
-	rd_req->rd_dev = cmd->se_lun->lun_se_dev->dev_ptr;
+	rd_req->rd_dev = cmd->se_dev->dev_ptr;
 
 	return &rd_req->rd_task;
 }
@@ -737,7 +737,7 @@ check_eot:
 	}
 
 out:
-	task->task_se_cmd->t_task->t_tasks_se_num += *se_mem_cnt;
+	task->task_se_cmd->t_task.t_tasks_se_num += *se_mem_cnt;
 #ifdef DEBUG_RAMDISK_DR
 	printk(KERN_INFO "RD_DR - Allocated %u struct se_mem segments for task\n",
 			*se_mem_cnt);
@@ -819,7 +819,7 @@ static int rd_DIRECT_without_offset(
 	}
 
 out:
-	task->task_se_cmd->t_task->t_tasks_se_num += *se_mem_cnt;
+	task->task_se_cmd->t_task.t_tasks_se_num += *se_mem_cnt;
 #ifdef DEBUG_RAMDISK_DR
 	printk(KERN_INFO "RD_DR - Allocated %u struct se_mem segments for task\n",
 			*se_mem_cnt);
@@ -880,14 +880,14 @@ static int rd_DIRECT_do_se_mem_map(
 	 * across multiple struct se_task->task_sg[].
 	 */
 	ret = transport_init_task_sg(task,
-			list_entry(cmd->t_task->t_mem_list->next,
+			list_first_entry(&cmd->t_task.t_mem_list,
 				   struct se_mem, se_list),
 			task_offset);
 	if (ret <= 0)
 		return ret;
 
 	return transport_map_mem_to_sg(task, se_mem_list, task->task_sg,
-			list_entry(cmd->t_task->t_mem_list->next,
+			list_first_entry(&cmd->t_task.t_mem_list,
 				   struct se_mem, se_list),
 			out_se_mem, se_mem_cnt, task_offset_in);
 }
diff --git a/drivers/target/target_core_tmr.c b/drivers/target/target_core_tmr.c
index 2f73749..e1f99f7 100644
--- a/drivers/target/target_core_tmr.c
+++ b/drivers/target/target_core_tmr.c
@@ -113,15 +113,14 @@ int core_tmr_lun_reset(
 	struct list_head *preempt_and_abort_list,
 	struct se_cmd *prout_cmd)
 {
-	struct se_cmd *cmd;
-	struct se_queue_req *qr, *qr_tmp;
+	struct se_cmd *cmd, *tcmd;
 	struct se_node_acl *tmr_nacl = NULL;
 	struct se_portal_group *tmr_tpg = NULL;
 	struct se_queue_obj *qobj = &dev->dev_queue_obj;
 	struct se_tmr_req *tmr_p, *tmr_pp;
 	struct se_task *task, *task_tmp;
 	unsigned long flags;
-	int fe_count, state, tas;
+	int fe_count, tas;
 	/*
 	 * TASK_ABORTED status bit, this is configurable via ConfigFS
 	 * struct se_device attributes.  spc4r17 section 7.4.6 Control mode page
@@ -179,14 +178,14 @@ int core_tmr_lun_reset(
 			continue;
 		spin_unlock(&dev->se_tmr_lock);
 
-		spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
-		if (!(atomic_read(&cmd->t_task->t_transport_active))) {
-			spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
+		if (!(atomic_read(&cmd->t_task.t_transport_active))) {
+			spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 			spin_lock(&dev->se_tmr_lock);
 			continue;
 		}
 		if (cmd->t_state == TRANSPORT_ISTATE_PROCESSING) {
-			spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+			spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 			spin_lock(&dev->se_tmr_lock);
 			continue;
 		}
@@ -194,7 +193,7 @@ int core_tmr_lun_reset(
 			" Response: 0x%02x, t_state: %d\n",
 			(preempt_and_abort_list) ? "Preempt" : "", tmr_p,
 			tmr_p->function, tmr_p->response, cmd->t_state);
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 		transport_cmd_finish_abort_tmr(cmd);
 		spin_lock(&dev->se_tmr_lock);
@@ -230,12 +229,6 @@ int core_tmr_lun_reset(
 		}
 		cmd = task->task_se_cmd;
 
-		if (!cmd->t_task) {
-			printk(KERN_ERR "cmd->t_task is NULL for task: %p cmd:"
-				" %p ITT: 0x%08x\n", task, cmd,
-				cmd->se_tfo->get_task_tag(cmd));
-			continue;
-		}
 		/*
 		 * For PREEMPT_AND_ABORT usage, only process commands
 		 * with a matching reservation key.
@@ -254,38 +247,38 @@ int core_tmr_lun_reset(
 		atomic_set(&task->task_state_active, 0);
 		spin_unlock_irqrestore(&dev->execute_task_lock, flags);
 
-		spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+		spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 		DEBUG_LR("LUN_RESET: %s cmd: %p task: %p"
 			" ITT/CmdSN: 0x%08x/0x%08x, i_state: %d, t_state/"
 			"def_t_state: %d/%d cdb: 0x%02x\n",
 			(preempt_and_abort_list) ? "Preempt" : "", cmd, task,
 			cmd->se_tfo->get_task_tag(cmd), 0,
 			cmd->se_tfo->get_cmd_state(cmd), cmd->t_state,
-			cmd->deferred_t_state, cmd->t_task->t_task_cdb[0]);
+			cmd->deferred_t_state, cmd->t_task.t_task_cdb[0]);
 		DEBUG_LR("LUN_RESET: ITT[0x%08x] - pr_res_key: 0x%016Lx"
 			" t_task_cdbs: %d t_task_cdbs_left: %d"
 			" t_task_cdbs_sent: %d -- t_transport_active: %d"
 			" t_transport_stop: %d t_transport_sent: %d\n",
 			cmd->se_tfo->get_task_tag(cmd), cmd->pr_res_key,
-			cmd->t_task->t_task_cdbs,
-			atomic_read(&cmd->t_task->t_task_cdbs_left),
-			atomic_read(&cmd->t_task->t_task_cdbs_sent),
-			atomic_read(&cmd->t_task->t_transport_active),
-			atomic_read(&cmd->t_task->t_transport_stop),
-			atomic_read(&cmd->t_task->t_transport_sent));
+			cmd->t_task.t_task_cdbs,
+			atomic_read(&cmd->t_task.t_task_cdbs_left),
+			atomic_read(&cmd->t_task.t_task_cdbs_sent),
+			atomic_read(&cmd->t_task.t_transport_active),
+			atomic_read(&cmd->t_task.t_transport_stop),
+			atomic_read(&cmd->t_task.t_transport_sent));
 
 		if (atomic_read(&task->task_active)) {
 			atomic_set(&task->task_stop, 1);
 			spin_unlock_irqrestore(
-				&cmd->t_task->t_state_lock, flags);
+				&cmd->t_task.t_state_lock, flags);
 
 			DEBUG_LR("LUN_RESET: Waiting for task: %p to shutdown"
 				" for dev: %p\n", task, dev);
 			wait_for_completion(&task->task_stop_comp);
 			DEBUG_LR("LUN_RESET Completed task: %p shutdown for"
 				" dev: %p\n", task, dev);
-			spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
-			atomic_dec(&cmd->t_task->t_task_cdbs_left);
+			spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
+			atomic_dec(&cmd->t_task.t_task_cdbs_left);
 
 			atomic_set(&task->task_active, 0);
 			atomic_set(&task->task_stop, 0);
@@ -295,24 +288,24 @@ int core_tmr_lun_reset(
 		}
 		__transport_stop_task_timer(task, &flags);
 
-		if (!(atomic_dec_and_test(&cmd->t_task->t_task_cdbs_ex_left))) {
+		if (!(atomic_dec_and_test(&cmd->t_task.t_task_cdbs_ex_left))) {
 			spin_unlock_irqrestore(
-					&cmd->t_task->t_state_lock, flags);
+					&cmd->t_task.t_state_lock, flags);
 			DEBUG_LR("LUN_RESET: Skipping task: %p, dev: %p for"
 				" t_task_cdbs_ex_left: %d\n", task, dev,
-				atomic_read(&cmd->t_task->t_task_cdbs_ex_left));
+				atomic_read(&cmd->t_task.t_task_cdbs_ex_left));
 
 			spin_lock_irqsave(&dev->execute_task_lock, flags);
 			continue;
 		}
-		fe_count = atomic_read(&cmd->t_task->t_fe_count);
+		fe_count = atomic_read(&cmd->t_task.t_fe_count);
 
-		if (atomic_read(&cmd->t_task->t_transport_active)) {
+		if (atomic_read(&cmd->t_task.t_transport_active)) {
 			DEBUG_LR("LUN_RESET: got t_transport_active = 1 for"
 				" task: %p, t_fe_count: %d dev: %p\n", task,
 				fe_count, dev);
-			atomic_set(&cmd->t_task->t_transport_aborted, 1);
-			spin_unlock_irqrestore(&cmd->t_task->t_state_lock,
+			atomic_set(&cmd->t_task.t_transport_aborted, 1);
+			spin_unlock_irqrestore(&cmd->t_task.t_state_lock,
 						flags);
 			core_tmr_handle_tas_abort(tmr_nacl, cmd, tas, fe_count);
 
@@ -321,8 +314,8 @@ int core_tmr_lun_reset(
 		}
 		DEBUG_LR("LUN_RESET: Got t_transport_active = 0 for task: %p,"
 			" t_fe_count: %d dev: %p\n", task, fe_count, dev);
-		atomic_set(&cmd->t_task->t_transport_aborted, 1);
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		atomic_set(&cmd->t_task.t_transport_aborted, 1);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		core_tmr_handle_tas_abort(tmr_nacl, cmd, tas, fe_count);
 
 		spin_lock_irqsave(&dev->execute_task_lock, flags);
@@ -337,20 +330,7 @@ int core_tmr_lun_reset(
 	 * reference, otherwise the struct se_cmd is released.
 	 */
 	spin_lock_irqsave(&qobj->cmd_queue_lock, flags);
-	list_for_each_entry_safe(qr, qr_tmp, &qobj->qobj_list, qr_list) {
-		cmd = (struct se_cmd *)qr->cmd;
-		if (!(cmd)) {
-			/*
-			 * Skip these for non PREEMPT_AND_ABORT usage..
-			 */
-			if (preempt_and_abort_list != NULL)
-				continue;
-
-			atomic_dec(&qobj->queue_cnt);
-			list_del(&qr->qr_list);
-			kfree(qr);
-			continue;
-		}
+	list_for_each_entry_safe(cmd, tcmd, &qobj->qobj_list, se_queue_node) {
 		/*
 		 * For PREEMPT_AND_ABORT usage, only process commands
 		 * with a matching reservation key.
@@ -365,18 +345,15 @@ int core_tmr_lun_reset(
 		if (prout_cmd == cmd)
 			continue;
 
-		atomic_dec(&cmd->t_task->t_transport_queue_active);
+		atomic_dec(&cmd->t_task.t_transport_queue_active);
 		atomic_dec(&qobj->queue_cnt);
-		list_del(&qr->qr_list);
+		list_del(&cmd->se_queue_node);
 		spin_unlock_irqrestore(&qobj->cmd_queue_lock, flags);
 
-		state = qr->state;
-		kfree(qr);
-
 		DEBUG_LR("LUN_RESET: %s from Device Queue: cmd: %p t_state:"
 			" %d t_fe_count: %d\n", (preempt_and_abort_list) ?
-			"Preempt" : "", cmd, state,
-			atomic_read(&cmd->t_task->t_fe_count));
+			"Preempt" : "", cmd, cmd->t_state,
+			atomic_read(&cmd->t_task.t_fe_count));
 		/*
 		 * Signal that the command has failed via cmd->se_cmd_flags,
 		 * and call TFO->new_cmd_failure() to wakeup any fabric
@@ -388,7 +365,7 @@ int core_tmr_lun_reset(
 		transport_new_cmd_failure(cmd);
 
 		core_tmr_handle_tas_abort(tmr_nacl, cmd, tas,
-				atomic_read(&cmd->t_task->t_fe_count));
+				atomic_read(&cmd->t_task.t_fe_count));
 		spin_lock_irqsave(&qobj->cmd_queue_lock, flags);
 	}
 	spin_unlock_irqrestore(&qobj->cmd_queue_lock, flags);
diff --git a/drivers/target/target_core_transport.c b/drivers/target/target_core_transport.c
index 6f5d4df..d0cd601 100644
--- a/drivers/target/target_core_transport.c
+++ b/drivers/target/target_core_transport.c
@@ -201,7 +201,7 @@ struct kmem_cache *t10_alua_tg_pt_gp_mem_cache;
 typedef int (*map_func_t)(struct se_task *, u32);
 
 static int transport_generic_write_pending(struct se_cmd *);
-static int transport_processing_thread(void *);
+static int transport_processing_thread(void *param);
 static int __transport_execute_tasks(struct se_device *dev);
 static void transport_complete_task_attr(struct se_cmd *cmd);
 static void transport_direct_request_timeout(struct se_cmd *cmd);
@@ -215,9 +215,8 @@ static int transport_generic_get_mem(struct se_cmd *cmd, u32 length,
 static int transport_generic_remove(struct se_cmd *cmd,
 		int release_to_pool, int session_reinstatement);
 static int transport_get_sectors(struct se_cmd *cmd);
-static struct list_head *transport_init_se_mem_list(void);
 static int transport_map_sg_to_mem(struct se_cmd *cmd,
-		struct list_head *se_mem_list, void *in_mem,
+		struct list_head *se_mem_list, struct scatterlist *sgl,
 		u32 *se_mem_cnt);
 static void transport_memcpy_se_mem_read_contig(struct se_cmd *cmd,
 		unsigned char *dst, struct list_head *se_mem_list);
@@ -574,7 +573,7 @@ void transport_deregister_session(struct se_session *se_sess)
 EXPORT_SYMBOL(transport_deregister_session);
 
 /*
- * Called with cmd->t_task->t_state_lock held.
+ * Called with cmd->t_task.t_state_lock held.
  */
 static void transport_all_task_dev_remove_state(struct se_cmd *cmd)
 {
@@ -582,10 +581,7 @@ static void transport_all_task_dev_remove_state(struct se_cmd *cmd)
 	struct se_task *task;
 	unsigned long flags;
 
-	if (!cmd->t_task)
-		return;
-
-	list_for_each_entry(task, &cmd->t_task->t_task_list, t_list) {
+	list_for_each_entry(task, &cmd->t_task.t_task_list, t_list) {
 		dev = task->se_dev;
 		if (!(dev))
 			continue;
@@ -603,7 +599,7 @@ static void transport_all_task_dev_remove_state(struct se_cmd *cmd)
 		spin_unlock_irqrestore(&dev->execute_task_lock, flags);
 
 		atomic_set(&task->task_state_active, 0);
-		atomic_dec(&cmd->t_task->t_task_cdbs_ex_left);
+		atomic_dec(&cmd->t_task.t_task_cdbs_ex_left);
 	}
 }
 
@@ -622,32 +618,32 @@ static int transport_cmd_check_stop(
 {
 	unsigned long flags;
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 	/*
 	 * Determine if IOCTL context caller in requesting the stopping of this
 	 * command for LUN shutdown purposes.
 	 */
-	if (atomic_read(&cmd->t_task->transport_lun_stop)) {
-		DEBUG_CS("%s:%d atomic_read(&cmd->t_task->transport_lun_stop)"
+	if (atomic_read(&cmd->t_task.transport_lun_stop)) {
+		DEBUG_CS("%s:%d atomic_read(&cmd->t_task.transport_lun_stop)"
 			" == TRUE for ITT: 0x%08x\n", __func__, __LINE__,
 			cmd->se_tfo->get_task_tag(cmd));
 
 		cmd->deferred_t_state = cmd->t_state;
 		cmd->t_state = TRANSPORT_DEFERRED_CMD;
-		atomic_set(&cmd->t_task->t_transport_active, 0);
+		atomic_set(&cmd->t_task.t_transport_active, 0);
 		if (transport_off == 2)
 			transport_all_task_dev_remove_state(cmd);
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
-		complete(&cmd->t_task->transport_lun_stop_comp);
+		complete(&cmd->t_task.transport_lun_stop_comp);
 		return 1;
 	}
 	/*
 	 * Determine if frontend context caller is requesting the stopping of
 	 * this command for frontend exceptions.
 	 */
-	if (atomic_read(&cmd->t_task->t_transport_stop)) {
-		DEBUG_CS("%s:%d atomic_read(&cmd->t_task->t_transport_stop) =="
+	if (atomic_read(&cmd->t_task.t_transport_stop)) {
+		DEBUG_CS("%s:%d atomic_read(&cmd->t_task.t_transport_stop) =="
 			" TRUE for ITT: 0x%08x\n", __func__, __LINE__,
 			cmd->se_tfo->get_task_tag(cmd));
 
@@ -662,13 +658,13 @@ static int transport_cmd_check_stop(
 		 */
 		if (transport_off == 2)
 			cmd->se_lun = NULL;
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
-		complete(&cmd->t_task->t_transport_stop_comp);
+		complete(&cmd->t_task.t_transport_stop_comp);
 		return 1;
 	}
 	if (transport_off) {
-		atomic_set(&cmd->t_task->t_transport_active, 0);
+		atomic_set(&cmd->t_task.t_transport_active, 0);
 		if (transport_off == 2) {
 			transport_all_task_dev_remove_state(cmd);
 			/*
@@ -683,18 +679,18 @@ static int transport_cmd_check_stop(
 			 */
 			if (cmd->se_tfo->check_stop_free != NULL) {
 				spin_unlock_irqrestore(
-					&cmd->t_task->t_state_lock, flags);
+					&cmd->t_task.t_state_lock, flags);
 
 				cmd->se_tfo->check_stop_free(cmd);
 				return 1;
 			}
 		}
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 		return 0;
 	} else if (t_state)
 		cmd->t_state = t_state;
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 	return 0;
 }
@@ -712,21 +708,21 @@ static void transport_lun_remove_cmd(struct se_cmd *cmd)
 	if (!lun)
 		return;
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
-	if (!(atomic_read(&cmd->t_task->transport_dev_active))) {
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
+	if (!(atomic_read(&cmd->t_task.transport_dev_active))) {
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		goto check_lun;
 	}
-	atomic_set(&cmd->t_task->transport_dev_active, 0);
+	atomic_set(&cmd->t_task.transport_dev_active, 0);
 	transport_all_task_dev_remove_state(cmd);
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 
 check_lun:
 	spin_lock_irqsave(&lun->lun_cmd_lock, flags);
-	if (atomic_read(&cmd->t_task->transport_lun_active)) {
-		list_del(&cmd->se_lun_list);
-		atomic_set(&cmd->t_task->transport_lun_active, 0);
+	if (atomic_read(&cmd->t_task.transport_lun_active)) {
+		list_del(&cmd->se_lun_node);
+		atomic_set(&cmd->t_task.transport_lun_active, 0);
 #if 0
 		printk(KERN_INFO "Removed ITT: 0x%08x from LUN LIST[%d]\n"
 			cmd->se_tfo->get_task_tag(cmd), lun->unpacked_lun);
@@ -737,7 +733,7 @@ check_lun:
 
 void transport_cmd_finish_abort(struct se_cmd *cmd, int remove)
 {
-	transport_remove_cmd_from_queue(cmd, &cmd->se_lun->lun_se_dev->dev_queue_obj);
+	transport_remove_cmd_from_queue(cmd, &cmd->se_dev->dev_queue_obj);
 	transport_lun_remove_cmd(cmd);
 
 	if (transport_cmd_check_stop_to_fabric(cmd))
@@ -748,7 +744,7 @@ void transport_cmd_finish_abort(struct se_cmd *cmd, int remove)
 
 void transport_cmd_finish_abort_tmr(struct se_cmd *cmd)
 {
-	transport_remove_cmd_from_queue(cmd, &cmd->se_lun->lun_se_dev->dev_queue_obj);
+	transport_remove_cmd_from_queue(cmd, &cmd->se_dev->dev_queue_obj);
 
 	if (transport_cmd_check_stop_to_fabric(cmd))
 		return;
@@ -756,50 +752,36 @@ void transport_cmd_finish_abort_tmr(struct se_cmd *cmd)
 	transport_generic_remove(cmd, 0, 0);
 }
 
-static int transport_add_cmd_to_queue(
+static void transport_add_cmd_to_queue(
 	struct se_cmd *cmd,
 	int t_state)
 {
 	struct se_device *dev = cmd->se_dev;
 	struct se_queue_obj *qobj = &dev->dev_queue_obj;
-	struct se_queue_req *qr;
 	unsigned long flags;
 
-	qr = kzalloc(sizeof(struct se_queue_req), GFP_ATOMIC);
-	if (!(qr)) {
-		printk(KERN_ERR "Unable to allocate memory for"
-				" struct se_queue_req\n");
-		return -ENOMEM;
-	}
-	INIT_LIST_HEAD(&qr->qr_list);
-
-	qr->cmd = cmd;
-	qr->state = t_state;
+	INIT_LIST_HEAD(&cmd->se_queue_node);
 
 	if (t_state) {
-		spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+		spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 		cmd->t_state = t_state;
-		atomic_set(&cmd->t_task->t_transport_active, 1);
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		atomic_set(&cmd->t_task.t_transport_active, 1);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 	}
 
 	spin_lock_irqsave(&qobj->cmd_queue_lock, flags);
-	list_add_tail(&qr->qr_list, &qobj->qobj_list);
-	atomic_inc(&cmd->t_task->t_transport_queue_active);
+	list_add_tail(&cmd->se_queue_node, &qobj->qobj_list);
+	atomic_inc(&cmd->t_task.t_transport_queue_active);
 	spin_unlock_irqrestore(&qobj->cmd_queue_lock, flags);
 
 	atomic_inc(&qobj->queue_cnt);
 	wake_up_interruptible(&qobj->thread_wq);
-	return 0;
 }
 
-/*
- * Called with struct se_queue_obj->cmd_queue_lock held.
- */
-static struct se_queue_req *
-transport_get_qr_from_queue(struct se_queue_obj *qobj)
+static struct se_cmd *
+transport_get_cmd_from_queue(struct se_queue_obj *qobj)
 {
-	struct se_queue_req *qr;
+	struct se_cmd *cmd;
 	unsigned long flags;
 
 	spin_lock_irqsave(&qobj->cmd_queue_lock, flags);
@@ -807,47 +789,42 @@ transport_get_qr_from_queue(struct se_queue_obj *qobj)
 		spin_unlock_irqrestore(&qobj->cmd_queue_lock, flags);
 		return NULL;
 	}
+	cmd = list_first_entry(&qobj->qobj_list, struct se_cmd, se_queue_node);
 
-	list_for_each_entry(qr, &qobj->qobj_list, qr_list)
-		break;
+	atomic_dec(&cmd->t_task.t_transport_queue_active);
 
-	if (qr->cmd)
-		atomic_dec(&qr->cmd->t_task->t_transport_queue_active);
-
-	list_del(&qr->qr_list);
+	list_del(&cmd->se_queue_node);
 	atomic_dec(&qobj->queue_cnt);
 	spin_unlock_irqrestore(&qobj->cmd_queue_lock, flags);
 
-	return qr;
+	return cmd;
 }
 
 static void transport_remove_cmd_from_queue(struct se_cmd *cmd,
 		struct se_queue_obj *qobj)
 {
-	struct se_queue_req *qr = NULL, *qr_p = NULL;
+	struct se_cmd *t;
 	unsigned long flags;
 
 	spin_lock_irqsave(&qobj->cmd_queue_lock, flags);
-	if (!(atomic_read(&cmd->t_task->t_transport_queue_active))) {
+	if (!(atomic_read(&cmd->t_task.t_transport_queue_active))) {
 		spin_unlock_irqrestore(&qobj->cmd_queue_lock, flags);
 		return;
 	}
 
-	list_for_each_entry_safe(qr, qr_p, &qobj->qobj_list, qr_list) {
-		if (qr->cmd != cmd)
-			continue;
-
-		atomic_dec(&qr->cmd->t_task->t_transport_queue_active);
-		atomic_dec(&qobj->queue_cnt);
-		list_del(&qr->qr_list);
-		kfree(qr);
-	}
+	list_for_each_entry(t, &qobj->qobj_list, se_queue_node)
+		if (t == cmd) {
+			atomic_dec(&cmd->t_task.t_transport_queue_active);
+			atomic_dec(&qobj->queue_cnt);
+			list_del(&cmd->se_queue_node);
+			break;
+		}
 	spin_unlock_irqrestore(&qobj->cmd_queue_lock, flags);
 
-	if (atomic_read(&cmd->t_task->t_transport_queue_active)) {
+	if (atomic_read(&cmd->t_task.t_transport_queue_active)) {
 		printk(KERN_ERR "ITT: 0x%08x t_transport_queue_active: %d\n",
 			cmd->se_tfo->get_task_tag(cmd),
-			atomic_read(&cmd->t_task->t_transport_queue_active));
+			atomic_read(&cmd->t_task.t_transport_queue_active));
 	}
 }
 
@@ -857,7 +834,7 @@ static void transport_remove_cmd_from_queue(struct se_cmd *cmd,
  */
 void transport_complete_sync_cache(struct se_cmd *cmd, int good)
 {
-	struct se_task *task = list_entry(cmd->t_task->t_task_list.next,
+	struct se_task *task = list_entry(cmd->t_task.t_task_list.next,
 				struct se_task, t_list);
 
 	if (good) {
@@ -887,12 +864,12 @@ void transport_complete_task(struct se_task *task, int success)
 	unsigned long flags;
 #if 0
 	printk(KERN_INFO "task: %p CDB: 0x%02x obj_ptr: %p\n", task,
-			cmd->t_task->t_task_cdb[0], dev);
+			cmd->t_task.t_task_cdb[0], dev);
 #endif
 	if (dev)
 		atomic_inc(&dev->depth_left);
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 	atomic_set(&task->task_active, 0);
 
 	/*
@@ -914,14 +891,14 @@ void transport_complete_task(struct se_task *task, int success)
 	 */
 	if (atomic_read(&task->task_stop)) {
 		/*
-		 * Decrement cmd->t_task->t_se_count if this task had
+		 * Decrement cmd->t_task.t_se_count if this task had
 		 * previously thrown its timeout exception handler.
 		 */
 		if (atomic_read(&task->task_timeout)) {
-			atomic_dec(&cmd->t_task->t_se_count);
+			atomic_dec(&cmd->t_task.t_se_count);
 			atomic_set(&task->task_timeout, 0);
 		}
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 		complete(&task->task_stop_comp);
 		return;
@@ -933,33 +910,33 @@ void transport_complete_task(struct se_task *task, int success)
 	 */
 	if (atomic_read(&task->task_timeout)) {
 		if (!(atomic_dec_and_test(
-				&cmd->t_task->t_task_cdbs_timeout_left))) {
-			spin_unlock_irqrestore(&cmd->t_task->t_state_lock,
+				&cmd->t_task.t_task_cdbs_timeout_left))) {
+			spin_unlock_irqrestore(&cmd->t_task.t_state_lock,
 				flags);
 			return;
 		}
 		t_state = TRANSPORT_COMPLETE_TIMEOUT;
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 		transport_add_cmd_to_queue(cmd, t_state);
 		return;
 	}
-	atomic_dec(&cmd->t_task->t_task_cdbs_timeout_left);
+	atomic_dec(&cmd->t_task.t_task_cdbs_timeout_left);
 
 	/*
 	 * Decrement the outstanding t_task_cdbs_left count.  The last
 	 * struct se_task from struct se_cmd will complete itself into the
 	 * device queue depending upon int success.
 	 */
-	if (!(atomic_dec_and_test(&cmd->t_task->t_task_cdbs_left))) {
+	if (!(atomic_dec_and_test(&cmd->t_task.t_task_cdbs_left))) {
 		if (!success)
-			cmd->t_task->t_tasks_failed = 1;
+			cmd->t_task.t_tasks_failed = 1;
 
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		return;
 	}
 
-	if (!success || cmd->t_task->t_tasks_failed) {
+	if (!success || cmd->t_task.t_tasks_failed) {
 		t_state = TRANSPORT_COMPLETE_FAILURE;
 		if (!task->task_error_status) {
 			task->task_error_status =
@@ -968,10 +945,10 @@ void transport_complete_task(struct se_task *task, int success)
 				PYX_TRANSPORT_UNKNOWN_SAM_OPCODE;
 		}
 	} else {
-		atomic_set(&cmd->t_task->t_transport_complete, 1);
+		atomic_set(&cmd->t_task.t_transport_complete, 1);
 		t_state = TRANSPORT_COMPLETE_OK;
 	}
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 	transport_add_cmd_to_queue(cmd, t_state);
 }
@@ -1064,8 +1041,8 @@ static void transport_add_tasks_to_state_queue(struct se_cmd *cmd)
 	struct se_task *task;
 	unsigned long flags;
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
-	list_for_each_entry(task, &cmd->t_task->t_task_list, t_list) {
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
+	list_for_each_entry(task, &cmd->t_task.t_task_list, t_list) {
 		dev = task->se_dev;
 
 		if (atomic_read(&task->task_state_active))
@@ -1081,17 +1058,17 @@ static void transport_add_tasks_to_state_queue(struct se_cmd *cmd)
 
 		spin_unlock(&dev->execute_task_lock);
 	}
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 }
 
 static void transport_add_tasks_from_cmd(struct se_cmd *cmd)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	struct se_task *task, *task_prev = NULL;
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev->execute_task_lock, flags);
-	list_for_each_entry(task, &cmd->t_task->t_task_list, t_list) {
+	list_for_each_entry(task, &cmd->t_task.t_task_list, t_list) {
 		if (atomic_read(&task->task_execute_queue))
 			continue;
 		/*
@@ -1184,19 +1161,15 @@ void transport_dump_dev_state(
  */
 static void transport_release_all_cmds(struct se_device *dev)
 {
-	struct se_cmd *cmd = NULL;
-	struct se_queue_req *qr = NULL, *qr_p = NULL;
+	struct se_cmd *cmd, *tcmd;
 	int bug_out = 0, t_state;
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev->dev_queue_obj.cmd_queue_lock, flags);
-	list_for_each_entry_safe(qr, qr_p, &dev->dev_queue_obj.qobj_list,
-				qr_list) {
-
-		cmd = qr->cmd;
-		t_state = qr->state;
-		list_del(&qr->qr_list);
-		kfree(qr);
+	list_for_each_entry_safe(cmd, tcmd, &dev->dev_queue_obj.qobj_list,
+				se_queue_node) {
+		t_state = cmd->t_state;
+		list_del(&cmd->se_queue_node);
 		spin_unlock_irqrestore(&dev->dev_queue_obj.cmd_queue_lock,
 				flags);
 
@@ -1548,7 +1521,7 @@ struct se_device *transport_add_device_to_core_hba(
 	transport_init_queue_obj(&dev->dev_queue_obj);
 	dev->dev_flags		= device_flags;
 	dev->dev_status		|= TRANSPORT_DEVICE_DEACTIVATED;
-	dev->dev_ptr		= (void *) transport_dev;
+	dev->dev_ptr		= transport_dev;
 	dev->se_hba		= hba;
 	dev->se_sub_dev		= se_dev;
 	dev->transport		= transport;
@@ -1684,7 +1657,7 @@ transport_generic_get_task(struct se_cmd *cmd,
 		enum dma_data_direction data_direction)
 {
 	struct se_task *task;
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	unsigned long flags;
 
 	task = dev->transport->alloc_task(cmd);
@@ -1697,26 +1670,20 @@ transport_generic_get_task(struct se_cmd *cmd,
 	INIT_LIST_HEAD(&task->t_execute_list);
 	INIT_LIST_HEAD(&task->t_state_list);
 	init_completion(&task->task_stop_comp);
-	task->task_no = cmd->t_task->t_tasks_no++;
+	task->task_no = cmd->t_task.t_tasks_no++;
 	task->task_se_cmd = cmd;
 	task->se_dev = dev;
 	task->task_data_direction = data_direction;
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
-	list_add_tail(&task->t_list, &cmd->t_task->t_task_list);
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
+	list_add_tail(&task->t_list, &cmd->t_task.t_task_list);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 	return task;
 }
 
 static int transport_generic_cmd_sequencer(struct se_cmd *, unsigned char *);
 
-void transport_device_setup_cmd(struct se_cmd *cmd)
-{
-	cmd->se_dev = cmd->se_lun->lun_se_dev;
-}
-EXPORT_SYMBOL(transport_device_setup_cmd);
-
 /*
  * Used by fabric modules containing a local struct se_cmd within their
  * fabric dependent per I/O descriptor.
@@ -1730,20 +1697,18 @@ void transport_init_se_cmd(
 	int task_attr,
 	unsigned char *sense_buffer)
 {
-	INIT_LIST_HEAD(&cmd->se_lun_list);
-	INIT_LIST_HEAD(&cmd->se_delayed_list);
-	INIT_LIST_HEAD(&cmd->se_ordered_list);
-	/*
-	 * Setup t_task pointer to t_task_backstore
-	 */
-	cmd->t_task = &cmd->t_task_backstore;
+	INIT_LIST_HEAD(&cmd->se_lun_node);
+	INIT_LIST_HEAD(&cmd->se_delayed_node);
+	INIT_LIST_HEAD(&cmd->se_ordered_node);
 
-	INIT_LIST_HEAD(&cmd->t_task->t_task_list);
-	init_completion(&cmd->t_task->transport_lun_fe_stop_comp);
-	init_completion(&cmd->t_task->transport_lun_stop_comp);
-	init_completion(&cmd->t_task->t_transport_stop_comp);
-	spin_lock_init(&cmd->t_task->t_state_lock);
-	atomic_set(&cmd->t_task->transport_dev_active, 1);
+	INIT_LIST_HEAD(&cmd->t_task.t_mem_list);
+	INIT_LIST_HEAD(&cmd->t_task.t_mem_bidi_list);
+	INIT_LIST_HEAD(&cmd->t_task.t_task_list);
+	init_completion(&cmd->t_task.transport_lun_fe_stop_comp);
+	init_completion(&cmd->t_task.transport_lun_stop_comp);
+	init_completion(&cmd->t_task.t_transport_stop_comp);
+	spin_lock_init(&cmd->t_task.t_state_lock);
+	atomic_set(&cmd->t_task.transport_dev_active, 1);
 
 	cmd->se_tfo = tfo;
 	cmd->se_sess = se_sess;
@@ -1760,7 +1725,7 @@ static int transport_check_alloc_task_attr(struct se_cmd *cmd)
 	 * Check if SAM Task Attribute emulation is enabled for this
 	 * struct se_device storage object
 	 */
-	if (cmd->se_lun->lun_se_dev->dev_task_attr_type != SAM_TASK_ATTR_EMULATED)
+	if (cmd->se_dev->dev_task_attr_type != SAM_TASK_ATTR_EMULATED)
 		return 0;
 
 	if (cmd->sam_task_attr == MSG_ACA_TAG) {
@@ -1772,7 +1737,7 @@ static int transport_check_alloc_task_attr(struct se_cmd *cmd)
 	 * Used to determine when ORDERED commands should go from
 	 * Dormant to Active status.
 	 */
-	cmd->se_ordered_id = atomic_inc_return(&cmd->se_lun->lun_se_dev->dev_ordered_id);
+	cmd->se_ordered_id = atomic_inc_return(&cmd->se_dev->dev_ordered_id);
 	smp_mb__after_atomic_inc();
 	DEBUG_STA("Allocated se_ordered_id: %u for Task Attr: 0x%02x on %s\n",
 			cmd->se_ordered_id, cmd->sam_task_attr,
@@ -1788,8 +1753,8 @@ void transport_free_se_cmd(
 	/*
 	 * Check and free any extended CDB buffer that was allocated
 	 */
-	if (se_cmd->t_task->t_task_cdb != se_cmd->t_task->__t_task_cdb)
-		kfree(se_cmd->t_task->t_task_cdb);
+	if (se_cmd->t_task.t_task_cdb != se_cmd->t_task.__t_task_cdb)
+		kfree(se_cmd->t_task.t_task_cdb);
 }
 EXPORT_SYMBOL(transport_free_se_cmd);
 
@@ -1812,7 +1777,6 @@ int transport_generic_allocate_tasks(
 	 */
 	cmd->transport_wait_for_tasks = &transport_generic_wait_for_tasks;
 
-	transport_device_setup_cmd(cmd);
 	/*
 	 * Ensure that the received CDB is less than the max (252 + 8) bytes
 	 * for VARIABLE_LENGTH_CMD
@@ -1828,26 +1792,26 @@ int transport_generic_allocate_tasks(
 	 * allocate the additional extended CDB buffer now..  Otherwise
 	 * setup the pointer from __t_task_cdb to t_task_cdb.
 	 */
-	if (scsi_command_size(cdb) > sizeof(cmd->t_task->__t_task_cdb)) {
-		cmd->t_task->t_task_cdb = kzalloc(scsi_command_size(cdb),
+	if (scsi_command_size(cdb) > sizeof(cmd->t_task.__t_task_cdb)) {
+		cmd->t_task.t_task_cdb = kzalloc(scsi_command_size(cdb),
 						GFP_KERNEL);
-		if (!(cmd->t_task->t_task_cdb)) {
-			printk(KERN_ERR "Unable to allocate cmd->t_task->t_task_cdb"
-				" %u > sizeof(cmd->t_task->__t_task_cdb): %lu ops\n",
+		if (!(cmd->t_task.t_task_cdb)) {
+			printk(KERN_ERR "Unable to allocate cmd->t_task.t_task_cdb"
+				" %u > sizeof(cmd->t_task.__t_task_cdb): %lu ops\n",
 				scsi_command_size(cdb),
-				(unsigned long)sizeof(cmd->t_task->__t_task_cdb));
+				(unsigned long)sizeof(cmd->t_task.__t_task_cdb));
 			return -ENOMEM;
 		}
 	} else
-		cmd->t_task->t_task_cdb = &cmd->t_task->__t_task_cdb[0];
+		cmd->t_task.t_task_cdb = &cmd->t_task.__t_task_cdb[0];
 	/*
 	 * Copy the original CDB into cmd->t_task.
 	 */
-	memcpy(cmd->t_task->t_task_cdb, cdb, scsi_command_size(cdb));
+	memcpy(cmd->t_task.t_task_cdb, cdb, scsi_command_size(cdb));
 	/*
 	 * Setup the received CDB based on SCSI defined opcodes and
 	 * perform unit attention, persistent reservations and ALUA
-	 * checks for virtual device backends.  The cmd->t_task->t_task_cdb
+	 * checks for virtual device backends.  The cmd->t_task.t_task_cdb
 	 * pointer is expected to be setup before we reach this point.
 	 */
 	ret = transport_generic_cmd_sequencer(cmd, cdb);
@@ -1859,7 +1823,7 @@ int transport_generic_allocate_tasks(
 	if (transport_check_alloc_task_attr(cmd) < 0) {
 		cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
 		cmd->scsi_sense_reason = TCM_INVALID_CDB_FIELD;
-		return -2;
+		return -EINVAL;
 	}
 	spin_lock(&cmd->se_lun->lun_sep_lock);
 	if (cmd->se_lun->lun_sep)
@@ -1947,7 +1911,6 @@ int transport_generic_handle_tmr(
 	 * This is needed for early exceptions.
 	 */
 	cmd->transport_wait_for_tasks = &transport_generic_wait_for_tasks;
-	transport_device_setup_cmd(cmd);
 
 	transport_add_cmd_to_queue(cmd, TRANSPORT_PROCESS_TMR);
 	return 0;
@@ -1973,9 +1936,9 @@ static int transport_stop_tasks_for_cmd(struct se_cmd *cmd)
 	/*
 	 * No tasks remain in the execution queue
 	 */
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 	list_for_each_entry_safe(task, task_tmp,
-				&cmd->t_task->t_task_list, t_list) {
+				&cmd->t_task.t_task_list, t_list) {
 		DEBUG_TS("task_no[%d] - Processing task %p\n",
 				task->task_no, task);
 		/*
@@ -1984,14 +1947,14 @@ static int transport_stop_tasks_for_cmd(struct se_cmd *cmd)
 		 */
 		if (!atomic_read(&task->task_sent) &&
 		    !atomic_read(&task->task_active)) {
-			spin_unlock_irqrestore(&cmd->t_task->t_state_lock,
+			spin_unlock_irqrestore(&cmd->t_task.t_state_lock,
 					flags);
 			transport_remove_task_from_execute_queue(task,
 					task->se_dev);
 
 			DEBUG_TS("task_no[%d] - Removed from execute queue\n",
 				task->task_no);
-			spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+			spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 			continue;
 		}
 
@@ -2001,7 +1964,7 @@ static int transport_stop_tasks_for_cmd(struct se_cmd *cmd)
 		 */
 		if (atomic_read(&task->task_active)) {
 			atomic_set(&task->task_stop, 1);
-			spin_unlock_irqrestore(&cmd->t_task->t_state_lock,
+			spin_unlock_irqrestore(&cmd->t_task.t_state_lock,
 					flags);
 
 			DEBUG_TS("task_no[%d] - Waiting to complete\n",
@@ -2010,8 +1973,8 @@ static int transport_stop_tasks_for_cmd(struct se_cmd *cmd)
 			DEBUG_TS("task_no[%d] - Stopped successfully\n",
 				task->task_no);
 
-			spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
-			atomic_dec(&cmd->t_task->t_task_cdbs_left);
+			spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
+			atomic_dec(&cmd->t_task.t_task_cdbs_left);
 
 			atomic_set(&task->task_active, 0);
 			atomic_set(&task->task_stop, 0);
@@ -2022,7 +1985,7 @@ static int transport_stop_tasks_for_cmd(struct se_cmd *cmd)
 
 		__transport_stop_task_timer(task, &flags);
 	}
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 	return ret;
 }
@@ -2038,7 +2001,7 @@ static void transport_generic_request_failure(
 {
 	DEBUG_GRF("-----[ Storage Engine Exception for cmd: %p ITT: 0x%08x"
 		" CDB: 0x%02x\n", cmd, cmd->se_tfo->get_task_tag(cmd),
-		cmd->t_task->t_task_cdb[0]);
+		cmd->t_task.t_task_cdb[0]);
 	DEBUG_GRF("-----[ i_state: %d t_state/def_t_state:"
 		" %d/%d transport_error_status: %d\n",
 		cmd->se_tfo->get_cmd_state(cmd),
@@ -2047,13 +2010,13 @@ static void transport_generic_request_failure(
 	DEBUG_GRF("-----[ t_task_cdbs: %d t_task_cdbs_left: %d"
 		" t_task_cdbs_sent: %d t_task_cdbs_ex_left: %d --"
 		" t_transport_active: %d t_transport_stop: %d"
-		" t_transport_sent: %d\n", cmd->t_task->t_task_cdbs,
-		atomic_read(&cmd->t_task->t_task_cdbs_left),
-		atomic_read(&cmd->t_task->t_task_cdbs_sent),
-		atomic_read(&cmd->t_task->t_task_cdbs_ex_left),
-		atomic_read(&cmd->t_task->t_transport_active),
-		atomic_read(&cmd->t_task->t_transport_stop),
-		atomic_read(&cmd->t_task->t_transport_sent));
+		" t_transport_sent: %d\n", cmd->t_task.t_task_cdbs,
+		atomic_read(&cmd->t_task.t_task_cdbs_left),
+		atomic_read(&cmd->t_task.t_task_cdbs_sent),
+		atomic_read(&cmd->t_task.t_task_cdbs_ex_left),
+		atomic_read(&cmd->t_task.t_transport_active),
+		atomic_read(&cmd->t_task.t_transport_stop),
+		atomic_read(&cmd->t_task.t_transport_sent));
 
 	transport_stop_all_task_timers(cmd);
 
@@ -2135,7 +2098,7 @@ static void transport_generic_request_failure(
 		break;
 	default:
 		printk(KERN_ERR "Unknown transport error for CDB 0x%02x: %d\n",
-			cmd->t_task->t_task_cdb[0],
+			cmd->t_task.t_task_cdb[0],
 			cmd->transport_error_status);
 		cmd->scsi_sense_reason = TCM_UNSUPPORTED_SCSI_OPCODE;
 		break;
@@ -2156,19 +2119,19 @@ static void transport_direct_request_timeout(struct se_cmd *cmd)
 {
 	unsigned long flags;
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
-	if (!(atomic_read(&cmd->t_task->t_transport_timeout))) {
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
+	if (!(atomic_read(&cmd->t_task.t_transport_timeout))) {
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		return;
 	}
-	if (atomic_read(&cmd->t_task->t_task_cdbs_timeout_left)) {
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	if (atomic_read(&cmd->t_task.t_task_cdbs_timeout_left)) {
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		return;
 	}
 
-	atomic_sub(atomic_read(&cmd->t_task->t_transport_timeout),
-		   &cmd->t_task->t_se_count);
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	atomic_sub(atomic_read(&cmd->t_task.t_transport_timeout),
+		   &cmd->t_task.t_se_count);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 }
 
 static void transport_generic_request_timeout(struct se_cmd *cmd)
@@ -2176,16 +2139,16 @@ static void transport_generic_request_timeout(struct se_cmd *cmd)
 	unsigned long flags;
 
 	/*
-	 * Reset cmd->t_task->t_se_count to allow transport_generic_remove()
+	 * Reset cmd->t_task.t_se_count to allow transport_generic_remove()
 	 * to allow last call to free memory resources.
 	 */
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
-	if (atomic_read(&cmd->t_task->t_transport_timeout) > 1) {
-		int tmp = (atomic_read(&cmd->t_task->t_transport_timeout) - 1);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
+	if (atomic_read(&cmd->t_task.t_transport_timeout) > 1) {
+		int tmp = (atomic_read(&cmd->t_task.t_transport_timeout) - 1);
 
-		atomic_sub(tmp, &cmd->t_task->t_se_count);
+		atomic_sub(tmp, &cmd->t_task.t_se_count);
 	}
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 	transport_generic_remove(cmd, 0, 0);
 }
@@ -2201,8 +2164,8 @@ transport_generic_allocate_buf(struct se_cmd *cmd, u32 data_length)
 		return -ENOMEM;
 	}
 
-	cmd->t_task->t_tasks_se_num = 0;
-	cmd->t_task->t_task_buf = buf;
+	cmd->t_task.t_tasks_se_num = 0;
+	cmd->t_task.t_task_buf = buf;
 
 	return 0;
 }
@@ -2244,9 +2207,9 @@ static void transport_set_supported_SAM_opcode(struct se_cmd *se_cmd)
 {
 	unsigned long flags;
 
-	spin_lock_irqsave(&se_cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&se_cmd->t_task.t_state_lock, flags);
 	se_cmd->se_cmd_flags |= SCF_SUPPORTED_SAM_OPCODE;
-	spin_unlock_irqrestore(&se_cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&se_cmd->t_task.t_state_lock, flags);
 }
 
 /*
@@ -2260,9 +2223,9 @@ static void transport_task_timeout_handler(unsigned long data)
 
 	DEBUG_TT("transport task timeout fired! task: %p cmd: %p\n", task, cmd);
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 	if (task->task_flags & TF_STOP) {
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		return;
 	}
 	task->task_flags &= ~TF_RUNNING;
@@ -2273,13 +2236,13 @@ static void transport_task_timeout_handler(unsigned long data)
 	if (!(atomic_read(&task->task_active))) {
 		DEBUG_TT("transport task: %p cmd: %p timeout task_active"
 				" == 0\n", task, cmd);
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		return;
 	}
 
-	atomic_inc(&cmd->t_task->t_se_count);
-	atomic_inc(&cmd->t_task->t_transport_timeout);
-	cmd->t_task->t_tasks_failed = 1;
+	atomic_inc(&cmd->t_task.t_se_count);
+	atomic_inc(&cmd->t_task.t_transport_timeout);
+	cmd->t_task.t_tasks_failed = 1;
 
 	atomic_set(&task->task_timeout, 1);
 	task->task_error_status = PYX_TRANSPORT_TASK_TIMEOUT;
@@ -2288,28 +2251,28 @@ static void transport_task_timeout_handler(unsigned long data)
 	if (atomic_read(&task->task_stop)) {
 		DEBUG_TT("transport task: %p cmd: %p timeout task_stop"
 				" == 1\n", task, cmd);
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		complete(&task->task_stop_comp);
 		return;
 	}
 
-	if (!(atomic_dec_and_test(&cmd->t_task->t_task_cdbs_left))) {
+	if (!(atomic_dec_and_test(&cmd->t_task.t_task_cdbs_left))) {
 		DEBUG_TT("transport task: %p cmd: %p timeout non zero"
 				" t_task_cdbs_left\n", task, cmd);
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		return;
 	}
 	DEBUG_TT("transport task: %p cmd: %p timeout ZERO t_task_cdbs_left\n",
 			task, cmd);
 
 	cmd->t_state = TRANSPORT_COMPLETE_FAILURE;
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 	transport_add_cmd_to_queue(cmd, TRANSPORT_COMPLETE_FAILURE);
 }
 
 /*
- * Called with cmd->t_task->t_state_lock held.
+ * Called with cmd->t_task.t_state_lock held.
  */
 static void transport_start_task_timer(struct se_task *task)
 {
@@ -2339,7 +2302,7 @@ static void transport_start_task_timer(struct se_task *task)
 }
 
 /*
- * Called with spin_lock_irq(&cmd->t_task->t_state_lock) held.
+ * Called with spin_lock_irq(&cmd->t_task.t_state_lock) held.
  */
 void __transport_stop_task_timer(struct se_task *task, unsigned long *flags)
 {
@@ -2349,11 +2312,11 @@ void __transport_stop_task_timer(struct se_task *task, unsigned long *flags)
 		return;
 
 	task->task_flags |= TF_STOP;
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, *flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, *flags);
 
 	del_timer_sync(&task->task_timer);
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, *flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, *flags);
 	task->task_flags &= ~TF_RUNNING;
 	task->task_flags &= ~TF_STOP;
 }
@@ -2363,11 +2326,11 @@ static void transport_stop_all_task_timers(struct se_cmd *cmd)
 	struct se_task *task = NULL, *task_tmp;
 	unsigned long flags;
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 	list_for_each_entry_safe(task, task_tmp,
-				&cmd->t_task->t_task_list, t_list)
+				&cmd->t_task.t_task_list, t_list)
 		__transport_stop_task_timer(task, &flags);
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 }
 
 static inline int transport_tcq_window_closed(struct se_device *dev)
@@ -2391,14 +2354,14 @@ static inline int transport_tcq_window_closed(struct se_device *dev)
  */
 static inline int transport_execute_task_attr(struct se_cmd *cmd)
 {
-	if (cmd->se_lun->lun_se_dev->dev_task_attr_type != SAM_TASK_ATTR_EMULATED)
+	if (cmd->se_dev->dev_task_attr_type != SAM_TASK_ATTR_EMULATED)
 		return 1;
 	/*
 	 * Check for the existence of HEAD_OF_QUEUE, and if true return 1
 	 * to allow the passed struct se_cmd list of tasks to the front of the list.
 	 */
 	 if (cmd->sam_task_attr == MSG_HEAD_TAG) {
-		atomic_inc(&cmd->se_lun->lun_se_dev->dev_hoq_count);
+		atomic_inc(&cmd->se_dev->dev_hoq_count);
 		smp_mb__after_atomic_inc();
 		DEBUG_STA("Added HEAD_OF_QUEUE for CDB:"
 			" 0x%02x, se_ordered_id: %u\n",
@@ -2406,30 +2369,30 @@ static inline int transport_execute_task_attr(struct se_cmd *cmd)
 			cmd->se_ordered_id);
 		return 1;
 	} else if (cmd->sam_task_attr == MSG_ORDERED_TAG) {
-		spin_lock(&cmd->se_lun->lun_se_dev->ordered_cmd_lock);
-		list_add_tail(&cmd->se_ordered_list,
-				&cmd->se_lun->lun_se_dev->ordered_cmd_list);
-		spin_unlock(&cmd->se_lun->lun_se_dev->ordered_cmd_lock);
+		spin_lock(&cmd->se_dev->ordered_cmd_lock);
+		list_add_tail(&cmd->se_ordered_node,
+				&cmd->se_dev->ordered_cmd_list);
+		spin_unlock(&cmd->se_dev->ordered_cmd_lock);
 
-		atomic_inc(&cmd->se_lun->lun_se_dev->dev_ordered_sync);
+		atomic_inc(&cmd->se_dev->dev_ordered_sync);
 		smp_mb__after_atomic_inc();
 
 		DEBUG_STA("Added ORDERED for CDB: 0x%02x to ordered"
 				" list, se_ordered_id: %u\n",
-				cmd->t_task->t_task_cdb[0],
+				cmd->t_task.t_task_cdb[0],
 				cmd->se_ordered_id);
 		/*
 		 * Add ORDERED command to tail of execution queue if
 		 * no other older commands exist that need to be
 		 * completed first.
 		 */
-		if (!(atomic_read(&cmd->se_lun->lun_se_dev->simple_cmds)))
+		if (!(atomic_read(&cmd->se_dev->simple_cmds)))
 			return 1;
 	} else {
 		/*
 		 * For SIMPLE and UNTAGGED Task Attribute commands
 		 */
-		atomic_inc(&cmd->se_lun->lun_se_dev->simple_cmds);
+		atomic_inc(&cmd->se_dev->simple_cmds);
 		smp_mb__after_atomic_inc();
 	}
 	/*
@@ -2437,20 +2400,20 @@ static inline int transport_execute_task_attr(struct se_cmd *cmd)
 	 * add the dormant task(s) built for the passed struct se_cmd to the
 	 * execution queue and become in Active state for this struct se_device.
 	 */
-	if (atomic_read(&cmd->se_lun->lun_se_dev->dev_ordered_sync) != 0) {
+	if (atomic_read(&cmd->se_dev->dev_ordered_sync) != 0) {
 		/*
 		 * Otherwise, add cmd w/ tasks to delayed cmd queue that
 		 * will be drained upon completion of HEAD_OF_QUEUE task.
 		 */
-		spin_lock(&cmd->se_lun->lun_se_dev->delayed_cmd_lock);
+		spin_lock(&cmd->se_dev->delayed_cmd_lock);
 		cmd->se_cmd_flags |= SCF_DELAYED_CMD_FROM_SAM_ATTR;
-		list_add_tail(&cmd->se_delayed_list,
-				&cmd->se_lun->lun_se_dev->delayed_cmd_list);
-		spin_unlock(&cmd->se_lun->lun_se_dev->delayed_cmd_lock);
+		list_add_tail(&cmd->se_delayed_node,
+				&cmd->se_dev->delayed_cmd_list);
+		spin_unlock(&cmd->se_dev->delayed_cmd_lock);
 
 		DEBUG_STA("Added CDB: 0x%02x Task Attr: 0x%02x to"
 			" delayed CMD list, se_ordered_id: %u\n",
-			cmd->t_task->t_task_cdb[0], cmd->sam_task_attr,
+			cmd->t_task.t_task_cdb[0], cmd->sam_task_attr,
 			cmd->se_ordered_id);
 		/*
 		 * Return zero to let transport_execute_tasks() know
@@ -2505,7 +2468,7 @@ static int transport_execute_tasks(struct se_cmd *cmd)
 	 * storage object.
 	 */
 execute_tasks:
-	__transport_execute_tasks(cmd->se_lun->lun_se_dev);
+	__transport_execute_tasks(cmd->se_dev);
 	return 0;
 }
 
@@ -2548,17 +2511,17 @@ check_depth:
 
 	cmd = task->task_se_cmd;
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 	atomic_set(&task->task_active, 1);
 	atomic_set(&task->task_sent, 1);
-	atomic_inc(&cmd->t_task->t_task_cdbs_sent);
+	atomic_inc(&cmd->t_task.t_task_cdbs_sent);
 
-	if (atomic_read(&cmd->t_task->t_task_cdbs_sent) ==
-	    cmd->t_task->t_task_cdbs)
+	if (atomic_read(&cmd->t_task.t_task_cdbs_sent) ==
+	    cmd->t_task.t_task_cdbs)
 		atomic_set(&cmd->transport_sent, 1);
 
 	transport_start_task_timer(task);
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 	/*
 	 * The struct se_cmd->transport_emulate_cdb() function pointer is used
 	 * to grab REPORT_LUNS and other CDBs we want to handle before they hit the
@@ -2623,10 +2586,10 @@ void transport_new_cmd_failure(struct se_cmd *se_cmd)
 	 * Any unsolicited data will get dumped for failed command inside of
 	 * the fabric plugin
 	 */
-	spin_lock_irqsave(&se_cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&se_cmd->t_task.t_state_lock, flags);
 	se_cmd->se_cmd_flags |= SCF_SE_CMD_FAILED;
 	se_cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
-	spin_unlock_irqrestore(&se_cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&se_cmd->t_task.t_state_lock, flags);
 
 	se_cmd->se_tfo->new_cmd_failure(se_cmd);
 }
@@ -2638,7 +2601,7 @@ static inline u32 transport_get_sectors_6(
 	struct se_cmd *cmd,
 	int *ret)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 
 	/*
 	 * Assume TYPE_DISK for non struct se_device objects.
@@ -2666,7 +2629,7 @@ static inline u32 transport_get_sectors_10(
 	struct se_cmd *cmd,
 	int *ret)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 
 	/*
 	 * Assume TYPE_DISK for non struct se_device objects.
@@ -2696,7 +2659,7 @@ static inline u32 transport_get_sectors_12(
 	struct se_cmd *cmd,
 	int *ret)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 
 	/*
 	 * Assume TYPE_DISK for non struct se_device objects.
@@ -2726,7 +2689,7 @@ static inline u32 transport_get_sectors_16(
 	struct se_cmd *cmd,
 	int *ret)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 
 	/*
 	 * Assume TYPE_DISK for non struct se_device objects.
@@ -2768,7 +2731,7 @@ static inline u32 transport_get_size(
 	unsigned char *cdb,
 	struct se_cmd *cmd)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 
 	if (dev->transport->get_device_type(dev) == TYPE_TAPE) {
 		if (cdb[1] & 1) { /* sectors */
@@ -2836,17 +2799,17 @@ static void transport_xor_callback(struct se_cmd *cmd)
 		return;
 	}
 	/*
-	 * Copy the scatterlist WRITE buffer located at cmd->t_task->t_mem_list
+	 * Copy the scatterlist WRITE buffer located at cmd->t_task.t_mem_list
 	 * into the locally allocated *buf
 	 */
-	transport_memcpy_se_mem_read_contig(cmd, buf, cmd->t_task->t_mem_list);
+	transport_memcpy_se_mem_read_contig(cmd, buf, &cmd->t_task.t_mem_list);
 	/*
 	 * Now perform the XOR against the BIDI read memory located at
-	 * cmd->t_task->t_mem_bidi_list
+	 * cmd->t_task.t_mem_bidi_list
 	 */
 
 	offset = 0;
-	list_for_each_entry(se_mem, cmd->t_task->t_mem_bidi_list, se_list) {
+	list_for_each_entry(se_mem, &cmd->t_task.t_mem_bidi_list, se_list) {
 		addr = (unsigned char *)kmap_atomic(se_mem->se_page, KM_USER0);
 		if (!(addr))
 			goto out;
@@ -2874,14 +2837,14 @@ static int transport_get_sense_data(struct se_cmd *cmd)
 
 	WARN_ON(!cmd->se_lun);
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 	if (cmd->se_cmd_flags & SCF_SENT_CHECK_CONDITION) {
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		return 0;
 	}
 
 	list_for_each_entry_safe(task, task_tmp,
-				&cmd->t_task->t_task_list, t_list) {
+				&cmd->t_task.t_task_list, t_list) {
 
 		if (!task->task_sense)
 			continue;
@@ -2903,12 +2866,12 @@ static int transport_get_sense_data(struct se_cmd *cmd)
 				cmd->se_tfo->get_task_tag(cmd), task->task_no);
 			continue;
 		}
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 		offset = cmd->se_tfo->set_fabric_sense_len(cmd,
 				TRANSPORT_SENSE_BUFFER);
 
-		memcpy((void *)&buffer[offset], (void *)sense_buffer,
+		memcpy(&buffer[offset], sense_buffer,
 				TRANSPORT_SENSE_BUFFER);
 		cmd->scsi_status = task->task_scsi_status;
 		/* Automatically padded */
@@ -2921,7 +2884,7 @@ static int transport_get_sense_data(struct se_cmd *cmd)
 				cmd->scsi_status);
 		return 0;
 	}
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 	return -1;
 }
@@ -2958,7 +2921,7 @@ transport_handle_reservation_conflict(struct se_cmd *cmd)
 		core_scsi3_ua_allocate(cmd->se_sess->se_node_acl,
 			cmd->orig_fe_lun, 0x2C,
 			ASCQ_2CH_PREVIOUS_RESERVATION_CONFLICT_STATUS);
-	return -2;
+	return -EINVAL;
 }
 
 /*	transport_generic_cmd_sequencer():
@@ -2975,7 +2938,7 @@ static int transport_generic_cmd_sequencer(
 	struct se_cmd *cmd,
 	unsigned char *cdb)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	struct se_subsystem_dev *su_dev = dev->se_sub_dev;
 	int ret = 0, sector_ret = 0, passthrough;
 	u32 sectors = 0, size = 0, pr_reg_type = 0;
@@ -2989,7 +2952,7 @@ static int transport_generic_cmd_sequencer(
 				&transport_nop_wait_for_tasks;
 		cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
 		cmd->scsi_sense_reason = TCM_CHECK_CONDITION_UNIT_ATTENTION;
-		return -2;
+		return -EINVAL;
 	}
 	/*
 	 * Check status of Asymmetric Logical Unit Assignment port
@@ -3011,7 +2974,7 @@ static int transport_generic_cmd_sequencer(
 			transport_set_sense_codes(cmd, 0x04, alua_ascq);
 			cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
 			cmd->scsi_sense_reason = TCM_CHECK_CONDITION_NOT_READY;
-			return -2;
+			return -EINVAL;
 		}
 		goto out_invalid_cdb_field;
 	}
@@ -3036,7 +2999,7 @@ static int transport_generic_cmd_sequencer(
 			goto out_unsupported_cdb;
 		size = transport_get_size(sectors, cdb, cmd);
 		cmd->transport_split_cdb = &split_cdb_XX_6;
-		cmd->t_task->t_task_lba = transport_lba_21(cdb);
+		cmd->t_task.t_task_lba = transport_lba_21(cdb);
 		cmd->se_cmd_flags |= SCF_SCSI_DATA_SG_IO_CDB;
 		break;
 	case READ_10:
@@ -3045,7 +3008,7 @@ static int transport_generic_cmd_sequencer(
 			goto out_unsupported_cdb;
 		size = transport_get_size(sectors, cdb, cmd);
 		cmd->transport_split_cdb = &split_cdb_XX_10;
-		cmd->t_task->t_task_lba = transport_lba_32(cdb);
+		cmd->t_task.t_task_lba = transport_lba_32(cdb);
 		cmd->se_cmd_flags |= SCF_SCSI_DATA_SG_IO_CDB;
 		break;
 	case READ_12:
@@ -3054,7 +3017,7 @@ static int transport_generic_cmd_sequencer(
 			goto out_unsupported_cdb;
 		size = transport_get_size(sectors, cdb, cmd);
 		cmd->transport_split_cdb = &split_cdb_XX_12;
-		cmd->t_task->t_task_lba = transport_lba_32(cdb);
+		cmd->t_task.t_task_lba = transport_lba_32(cdb);
 		cmd->se_cmd_flags |= SCF_SCSI_DATA_SG_IO_CDB;
 		break;
 	case READ_16:
@@ -3063,7 +3026,7 @@ static int transport_generic_cmd_sequencer(
 			goto out_unsupported_cdb;
 		size = transport_get_size(sectors, cdb, cmd);
 		cmd->transport_split_cdb = &split_cdb_XX_16;
-		cmd->t_task->t_task_lba = transport_lba_64(cdb);
+		cmd->t_task.t_task_lba = transport_lba_64(cdb);
 		cmd->se_cmd_flags |= SCF_SCSI_DATA_SG_IO_CDB;
 		break;
 	case WRITE_6:
@@ -3072,7 +3035,7 @@ static int transport_generic_cmd_sequencer(
 			goto out_unsupported_cdb;
 		size = transport_get_size(sectors, cdb, cmd);
 		cmd->transport_split_cdb = &split_cdb_XX_6;
-		cmd->t_task->t_task_lba = transport_lba_21(cdb);
+		cmd->t_task.t_task_lba = transport_lba_21(cdb);
 		cmd->se_cmd_flags |= SCF_SCSI_DATA_SG_IO_CDB;
 		break;
 	case WRITE_10:
@@ -3081,8 +3044,8 @@ static int transport_generic_cmd_sequencer(
 			goto out_unsupported_cdb;
 		size = transport_get_size(sectors, cdb, cmd);
 		cmd->transport_split_cdb = &split_cdb_XX_10;
-		cmd->t_task->t_task_lba = transport_lba_32(cdb);
-		cmd->t_task->t_tasks_fua = (cdb[1] & 0x8);
+		cmd->t_task.t_task_lba = transport_lba_32(cdb);
+		cmd->t_task.t_tasks_fua = (cdb[1] & 0x8);
 		cmd->se_cmd_flags |= SCF_SCSI_DATA_SG_IO_CDB;
 		break;
 	case WRITE_12:
@@ -3091,8 +3054,8 @@ static int transport_generic_cmd_sequencer(
 			goto out_unsupported_cdb;
 		size = transport_get_size(sectors, cdb, cmd);
 		cmd->transport_split_cdb = &split_cdb_XX_12;
-		cmd->t_task->t_task_lba = transport_lba_32(cdb);
-		cmd->t_task->t_tasks_fua = (cdb[1] & 0x8);
+		cmd->t_task.t_task_lba = transport_lba_32(cdb);
+		cmd->t_task.t_tasks_fua = (cdb[1] & 0x8);
 		cmd->se_cmd_flags |= SCF_SCSI_DATA_SG_IO_CDB;
 		break;
 	case WRITE_16:
@@ -3101,20 +3064,20 @@ static int transport_generic_cmd_sequencer(
 			goto out_unsupported_cdb;
 		size = transport_get_size(sectors, cdb, cmd);
 		cmd->transport_split_cdb = &split_cdb_XX_16;
-		cmd->t_task->t_task_lba = transport_lba_64(cdb);
-		cmd->t_task->t_tasks_fua = (cdb[1] & 0x8);
+		cmd->t_task.t_task_lba = transport_lba_64(cdb);
+		cmd->t_task.t_tasks_fua = (cdb[1] & 0x8);
 		cmd->se_cmd_flags |= SCF_SCSI_DATA_SG_IO_CDB;
 		break;
 	case XDWRITEREAD_10:
 		if ((cmd->data_direction != DMA_TO_DEVICE) ||
-		    !(cmd->t_task->t_tasks_bidi))
+		    !(cmd->t_task.t_tasks_bidi))
 			goto out_invalid_cdb_field;
 		sectors = transport_get_sectors_10(cdb, cmd, &sector_ret);
 		if (sector_ret)
 			goto out_unsupported_cdb;
 		size = transport_get_size(sectors, cdb, cmd);
 		cmd->transport_split_cdb = &split_cdb_XX_10;
-		cmd->t_task->t_task_lba = transport_lba_32(cdb);
+		cmd->t_task.t_task_lba = transport_lba_32(cdb);
 		cmd->se_cmd_flags |= SCF_SCSI_DATA_SG_IO_CDB;
 		passthrough = (dev->transport->transport_type ==
 				TRANSPORT_PLUGIN_PHBA_PDEV);
@@ -3127,7 +3090,7 @@ static int transport_generic_cmd_sequencer(
 		 * Setup BIDI XOR callback to be run during transport_generic_complete_ok()
 		 */
 		cmd->transport_complete_callback = &transport_xor_callback;
-		cmd->t_task->t_tasks_fua = (cdb[1] & 0x8);
+		cmd->t_task.t_tasks_fua = (cdb[1] & 0x8);
 		break;
 	case VARIABLE_LENGTH_CMD:
 		service_action = get_unaligned_be16(&cdb[8]);
@@ -3149,7 +3112,7 @@ static int transport_generic_cmd_sequencer(
 			 * XDWRITE_READ_32 logic.
 			 */
 			cmd->transport_split_cdb = &split_cdb_XX_32;
-			cmd->t_task->t_task_lba = transport_lba_64_ext(cdb);
+			cmd->t_task.t_task_lba = transport_lba_64_ext(cdb);
 			cmd->se_cmd_flags |= SCF_SCSI_DATA_SG_IO_CDB;
 
 			/*
@@ -3163,14 +3126,14 @@ static int transport_generic_cmd_sequencer(
 			 * transport_generic_complete_ok()
 			 */
 			cmd->transport_complete_callback = &transport_xor_callback;
-			cmd->t_task->t_tasks_fua = (cdb[10] & 0x8);
+			cmd->t_task.t_tasks_fua = (cdb[10] & 0x8);
 			break;
 		case WRITE_SAME_32:
 			sectors = transport_get_sectors_32(cdb, cmd, &sector_ret);
 			if (sector_ret)
 				goto out_unsupported_cdb;
 			size = transport_get_size(sectors, cdb, cmd);
-			cmd->t_task->t_task_lba = get_unaligned_be64(&cdb[12]);
+			cmd->t_task.t_task_lba = get_unaligned_be64(&cdb[12]);
 			cmd->se_cmd_flags |= SCF_SCSI_CONTROL_SG_IO_CDB;
 
 			/*
@@ -3299,7 +3262,7 @@ static int transport_generic_cmd_sequencer(
 		 * Do implict HEAD_OF_QUEUE processing for INQUIRY.
 		 * See spc4r17 section 5.3
 		 */
-		if (cmd->se_lun->lun_se_dev->dev_task_attr_type == SAM_TASK_ATTR_EMULATED)
+		if (cmd->se_dev->dev_task_attr_type == SAM_TASK_ATTR_EMULATED)
 			cmd->sam_task_attr = MSG_HEAD_TAG;
 		cmd->se_cmd_flags |= SCF_SCSI_CONTROL_NONSG_IO_CDB;
 		break;
@@ -3405,10 +3368,10 @@ static int transport_generic_cmd_sequencer(
 		 */
 		if (cdb[0] == SYNCHRONIZE_CACHE) {
 			sectors = transport_get_sectors_10(cdb, cmd, &sector_ret);
-			cmd->t_task->t_task_lba = transport_lba_32(cdb);
+			cmd->t_task.t_task_lba = transport_lba_32(cdb);
 		} else {
 			sectors = transport_get_sectors_16(cdb, cmd, &sector_ret);
-			cmd->t_task->t_task_lba = transport_lba_64(cdb);
+			cmd->t_task.t_task_lba = transport_lba_64(cdb);
 		}
 		if (sector_ret)
 			goto out_unsupported_cdb;
@@ -3454,7 +3417,7 @@ static int transport_generic_cmd_sequencer(
 		if (sector_ret)
 			goto out_unsupported_cdb;
 		size = transport_get_size(sectors, cdb, cmd);
-		cmd->t_task->t_task_lba = get_unaligned_be16(&cdb[2]);
+		cmd->t_task.t_task_lba = get_unaligned_be16(&cdb[2]);
 		passthrough = (dev->transport->transport_type ==
 				TRANSPORT_PLUGIN_PHBA_PDEV);
 		/*
@@ -3507,7 +3470,7 @@ static int transport_generic_cmd_sequencer(
 		 * Do implict HEAD_OF_QUEUE processing for REPORT_LUNS
 		 * See spc4r17 section 5.3
 		 */
-		if (cmd->se_lun->lun_se_dev->dev_task_attr_type == SAM_TASK_ATTR_EMULATED)
+		if (cmd->se_dev->dev_task_attr_type == SAM_TASK_ATTR_EMULATED)
 			cmd->sam_task_attr = MSG_HEAD_TAG;
 		cmd->se_cmd_flags |= SCF_SCSI_CONTROL_NONSG_IO_CDB;
 		break;
@@ -3560,11 +3523,11 @@ static int transport_generic_cmd_sequencer(
 out_unsupported_cdb:
 	cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
 	cmd->scsi_sense_reason = TCM_UNSUPPORTED_SCSI_OPCODE;
-	return -2;
+	return -EINVAL;
 out_invalid_cdb_field:
 	cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
 	cmd->scsi_sense_reason = TCM_INVALID_CDB_FIELD;
-	return -2;
+	return -EINVAL;
 }
 
 static inline void transport_release_tasks(struct se_cmd *);
@@ -3662,7 +3625,7 @@ static void transport_memcpy_se_mem_read_contig(
  */
 static void transport_complete_task_attr(struct se_cmd *cmd)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	struct se_cmd *cmd_p, *cmd_tmp;
 	int new_active_tasks = 0;
 
@@ -3682,7 +3645,7 @@ static void transport_complete_task_attr(struct se_cmd *cmd)
 			cmd->se_ordered_id);
 	} else if (cmd->sam_task_attr == MSG_ORDERED_TAG) {
 		spin_lock(&dev->ordered_cmd_lock);
-		list_del(&cmd->se_ordered_list);
+		list_del(&cmd->se_ordered_node);
 		atomic_dec(&dev->dev_ordered_sync);
 		smp_mb__after_atomic_dec();
 		spin_unlock(&dev->ordered_cmd_lock);
@@ -3698,9 +3661,9 @@ static void transport_complete_task_attr(struct se_cmd *cmd)
 	 */
 	spin_lock(&dev->delayed_cmd_lock);
 	list_for_each_entry_safe(cmd_p, cmd_tmp,
-			&dev->delayed_cmd_list, se_delayed_list) {
+			&dev->delayed_cmd_list, se_delayed_node) {
 
-		list_del(&cmd_p->se_delayed_list);
+		list_del(&cmd_p->se_delayed_node);
 		spin_unlock(&dev->delayed_cmd_lock);
 
 		DEBUG_STA("Calling add_tasks() for"
@@ -3733,7 +3696,7 @@ static void transport_generic_complete_ok(struct se_cmd *cmd)
 	 * delayed execution list after a HEAD_OF_QUEUE or ORDERED Task
 	 * Attribute.
 	 */
-	if (cmd->se_lun->lun_se_dev->dev_task_attr_type == SAM_TASK_ATTR_EMULATED)
+	if (cmd->se_dev->dev_task_attr_type == SAM_TASK_ATTR_EMULATED)
 		transport_complete_task_attr(cmd);
 	/*
 	 * Check if we need to retrieve a sense buffer from
@@ -3777,8 +3740,8 @@ static void transport_generic_complete_ok(struct se_cmd *cmd)
 		 */
 		if (cmd->se_cmd_flags & SCF_PASSTHROUGH_CONTIG_TO_SG)
 			transport_memcpy_write_contig(cmd,
-				 cmd->t_task->t_task_pt_sgl,
-				 cmd->t_task->t_task_buf);
+				 cmd->t_task.t_task_pt_sgl,
+				 cmd->t_task.t_task_buf);
 
 		cmd->se_tfo->queue_data_in(cmd);
 		break;
@@ -3792,7 +3755,7 @@ static void transport_generic_complete_ok(struct se_cmd *cmd)
 		/*
 		 * Check if we need to send READ payload for BIDI-COMMAND
 		 */
-		if (cmd->t_task->t_mem_bidi_list != NULL) {
+		if (!list_empty(&cmd->t_task.t_mem_bidi_list)) {
 			spin_lock(&cmd->se_lun->lun_sep_lock);
 			if (cmd->se_lun->lun_sep) {
 				cmd->se_lun->lun_sep->sep_stats.tx_data_octets +=
@@ -3819,9 +3782,9 @@ static void transport_free_dev_tasks(struct se_cmd *cmd)
 	struct se_task *task, *task_tmp;
 	unsigned long flags;
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 	list_for_each_entry_safe(task, task_tmp,
-				&cmd->t_task->t_task_list, t_list) {
+				&cmd->t_task.t_task_list, t_list) {
 		if (atomic_read(&task->task_active))
 			continue;
 
@@ -3830,15 +3793,15 @@ static void transport_free_dev_tasks(struct se_cmd *cmd)
 
 		list_del(&task->t_list);
 
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		if (task->se_dev)
 			task->se_dev->transport->free_task(task);
 		else
 			printk(KERN_ERR "task[%u] - task->se_dev is NULL\n",
 				task->task_no);
-		spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+		spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 	}
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 }
 
 static inline void transport_free_pages(struct se_cmd *cmd)
@@ -3851,9 +3814,9 @@ static inline void transport_free_pages(struct se_cmd *cmd)
 	if (cmd->se_dev->transport->do_se_mem_map)
 		free_page = 0;
 
-	if (cmd->t_task->t_task_buf) {
-		kfree(cmd->t_task->t_task_buf);
-		cmd->t_task->t_task_buf = NULL;
+	if (cmd->t_task.t_task_buf) {
+		kfree(cmd->t_task.t_task_buf);
+		cmd->t_task.t_task_buf = NULL;
 		return;
 	}
 
@@ -3863,11 +3826,8 @@ static inline void transport_free_pages(struct se_cmd *cmd)
 	if (cmd->se_cmd_flags & SCF_CMD_PASSTHROUGH_NOALLOC)
 		return;
 
-	if (!(cmd->t_task->t_tasks_se_num))
-		return;
-
 	list_for_each_entry_safe(se_mem, se_mem_tmp,
-			cmd->t_task->t_mem_list, se_list) {
+			&cmd->t_task.t_mem_list, se_list) {
 		/*
 		 * We only release call __free_page(struct se_mem->se_page) when
 		 * SCF_PASSTHROUGH_SG_TO_MEM_NOALLOC is NOT in use,
@@ -3878,27 +3838,21 @@ static inline void transport_free_pages(struct se_cmd *cmd)
 		list_del(&se_mem->se_list);
 		kmem_cache_free(se_mem_cache, se_mem);
 	}
+	cmd->t_task.t_tasks_se_num = 0;
 
-	if (cmd->t_task->t_mem_bidi_list && cmd->t_task->t_tasks_se_bidi_num) {
-		list_for_each_entry_safe(se_mem, se_mem_tmp,
-				cmd->t_task->t_mem_bidi_list, se_list) {
-			/*
-			 * We only release call __free_page(struct se_mem->se_page) when
-			 * SCF_PASSTHROUGH_SG_TO_MEM_NOALLOC is NOT in use,
-			 */
-			if (free_page)
-				__free_page(se_mem->se_page);
+	list_for_each_entry_safe(se_mem, se_mem_tmp,
+				 &cmd->t_task.t_mem_bidi_list, se_list) {
+		/*
+		 * We only release call __free_page(struct se_mem->se_page) when
+		 * SCF_PASSTHROUGH_SG_TO_MEM_NOALLOC is NOT in use,
+		 */
+		if (free_page)
+			__free_page(se_mem->se_page);
 
-			list_del(&se_mem->se_list);
-			kmem_cache_free(se_mem_cache, se_mem);
-		}
+		list_del(&se_mem->se_list);
+		kmem_cache_free(se_mem_cache, se_mem);
 	}
-
-	kfree(cmd->t_task->t_mem_bidi_list);
-	cmd->t_task->t_mem_bidi_list = NULL;
-	kfree(cmd->t_task->t_mem_list);
-	cmd->t_task->t_mem_list = NULL;
-	cmd->t_task->t_tasks_se_num = 0;
+	cmd->t_task.t_tasks_se_bidi_num = 0;
 }
 
 static inline void transport_release_tasks(struct se_cmd *cmd)
@@ -3910,23 +3864,23 @@ static inline int transport_dec_and_check(struct se_cmd *cmd)
 {
 	unsigned long flags;
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
-	if (atomic_read(&cmd->t_task->t_fe_count)) {
-		if (!(atomic_dec_and_test(&cmd->t_task->t_fe_count))) {
-			spin_unlock_irqrestore(&cmd->t_task->t_state_lock,
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
+	if (atomic_read(&cmd->t_task.t_fe_count)) {
+		if (!(atomic_dec_and_test(&cmd->t_task.t_fe_count))) {
+			spin_unlock_irqrestore(&cmd->t_task.t_state_lock,
 					flags);
 			return 1;
 		}
 	}
 
-	if (atomic_read(&cmd->t_task->t_se_count)) {
-		if (!(atomic_dec_and_test(&cmd->t_task->t_se_count))) {
-			spin_unlock_irqrestore(&cmd->t_task->t_state_lock,
+	if (atomic_read(&cmd->t_task.t_se_count)) {
+		if (!(atomic_dec_and_test(&cmd->t_task.t_se_count))) {
+			spin_unlock_irqrestore(&cmd->t_task.t_state_lock,
 					flags);
 			return 1;
 		}
 	}
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 	return 0;
 }
@@ -3938,14 +3892,14 @@ static void transport_release_fe_cmd(struct se_cmd *cmd)
 	if (transport_dec_and_check(cmd))
 		return;
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
-	if (!(atomic_read(&cmd->t_task->transport_dev_active))) {
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
+	if (!(atomic_read(&cmd->t_task.transport_dev_active))) {
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		goto free_pages;
 	}
-	atomic_set(&cmd->t_task->transport_dev_active, 0);
+	atomic_set(&cmd->t_task.transport_dev_active, 0);
 	transport_all_task_dev_remove_state(cmd);
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 	transport_release_tasks(cmd);
 free_pages:
@@ -3961,33 +3915,30 @@ static int transport_generic_remove(
 {
 	unsigned long flags;
 
-	if (!(cmd->t_task))
-		goto release_cmd;
-
 	if (transport_dec_and_check(cmd)) {
 		if (session_reinstatement) {
-			spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+			spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 			transport_all_task_dev_remove_state(cmd);
-			spin_unlock_irqrestore(&cmd->t_task->t_state_lock,
+			spin_unlock_irqrestore(&cmd->t_task.t_state_lock,
 					flags);
 		}
 		return 1;
 	}
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
-	if (!(atomic_read(&cmd->t_task->transport_dev_active))) {
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
+	if (!(atomic_read(&cmd->t_task.transport_dev_active))) {
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		goto free_pages;
 	}
-	atomic_set(&cmd->t_task->transport_dev_active, 0);
+	atomic_set(&cmd->t_task.transport_dev_active, 0);
 	transport_all_task_dev_remove_state(cmd);
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 	transport_release_tasks(cmd);
+
 free_pages:
 	transport_free_pages(cmd);
 
-release_cmd:
 	if (release_to_pool) {
 		transport_release_cmd_to_pool(cmd);
 	} else {
@@ -4011,35 +3962,19 @@ release_cmd:
  */
 int transport_generic_map_mem_to_cmd(
 	struct se_cmd *cmd,
-	struct scatterlist *mem,
-	u32 sg_mem_num,
-	struct scatterlist *mem_bidi_in,
-	u32 sg_mem_bidi_num)
+	struct scatterlist *sgl,
+	u32 sgl_count,
+	struct scatterlist *sgl_bidi,
+	u32 sgl_bidi_count)
 {
-	u32 se_mem_cnt_out = 0;
+	u32 mapped_sg_count = 0;
 	int ret;
 
-	if (!(mem) || !(sg_mem_num))
+	if (!sgl || !sgl_count)
 		return 0;
-	/*
-	 * Passed *mem will contain a list_head containing preformatted
-	 * struct se_mem elements...
-	 */
-	if (!(cmd->se_cmd_flags & SCF_PASSTHROUGH_SG_TO_MEM)) {
-		if ((mem_bidi_in) || (sg_mem_bidi_num)) {
-			printk(KERN_ERR "SCF_CMD_PASSTHROUGH_NOALLOC not supported"
-				" with BIDI-COMMAND\n");
-			return -ENOSYS;
-		}
 
-		cmd->t_task->t_mem_list = (struct list_head *)mem;
-		cmd->t_task->t_tasks_se_num = sg_mem_num;
-		cmd->se_cmd_flags |= SCF_CMD_PASSTHROUGH_NOALLOC;
-		return 0;
-	}
 	/*
-	 * Otherwise, assume the caller is passing a struct scatterlist
-	 * array from include/linux/scatterlist.h
+	 * Convert sgls (sgl, sgl_bidi) to list of se_mems
 	 */
 	if ((cmd->se_cmd_flags & SCF_SCSI_DATA_SG_IO_CDB) ||
 	    (cmd->se_cmd_flags & SCF_SCSI_CONTROL_SG_IO_CDB)) {
@@ -4048,41 +3983,29 @@ int transport_generic_map_mem_to_cmd(
 		 * processed into a TCM struct se_subsystem_dev, we do the mapping
 		 * from the passed physical memory to struct se_mem->se_page here.
 		 */
-		cmd->t_task->t_mem_list = transport_init_se_mem_list();
-		if (!(cmd->t_task->t_mem_list))
-			return -ENOMEM;
-
 		ret = transport_map_sg_to_mem(cmd,
-			cmd->t_task->t_mem_list, mem, &se_mem_cnt_out);
+			&cmd->t_task.t_mem_list, sgl, &mapped_sg_count);
 		if (ret < 0)
 			return -ENOMEM;
 
-		cmd->t_task->t_tasks_se_num = se_mem_cnt_out;
+		cmd->t_task.t_tasks_se_num = mapped_sg_count;
 		/*
 		 * Setup BIDI READ list of struct se_mem elements
 		 */
-		if ((mem_bidi_in) && (sg_mem_bidi_num)) {
-			cmd->t_task->t_mem_bidi_list = transport_init_se_mem_list();
-			if (!(cmd->t_task->t_mem_bidi_list)) {
-				kfree(cmd->t_task->t_mem_list);
-				return -ENOMEM;
-			}
-			se_mem_cnt_out = 0;
-
+		if (sgl_bidi && sgl_bidi_count) {
+			mapped_sg_count = 0;
 			ret = transport_map_sg_to_mem(cmd,
-				cmd->t_task->t_mem_bidi_list, mem_bidi_in,
-				&se_mem_cnt_out);
-			if (ret < 0) {
-				kfree(cmd->t_task->t_mem_list);
+				&cmd->t_task.t_mem_bidi_list, sgl_bidi,
+				&mapped_sg_count);
+			if (ret < 0)
 				return -ENOMEM;
-			}
 
-			cmd->t_task->t_tasks_se_bidi_num = se_mem_cnt_out;
+			cmd->t_task.t_tasks_se_bidi_num = mapped_sg_count;
 		}
 		cmd->se_cmd_flags |= SCF_PASSTHROUGH_SG_TO_MEM_NOALLOC;
 
 	} else if (cmd->se_cmd_flags & SCF_SCSI_CONTROL_NONSG_IO_CDB) {
-		if (mem_bidi_in || sg_mem_bidi_num) {
+		if (sgl_bidi || sgl_bidi_count) {
 			printk(KERN_ERR "BIDI-Commands not supported using "
 				"SCF_SCSI_CONTROL_NONSG_IO_CDB\n");
 			return -ENOSYS;
@@ -4097,7 +4020,8 @@ int transport_generic_map_mem_to_cmd(
 		 * struct scatterlist format.
 		 */
 		cmd->se_cmd_flags |= SCF_PASSTHROUGH_CONTIG_TO_SG;
-		cmd->t_task->t_task_pt_sgl = mem;
+		cmd->t_task.t_task_pt_sgl = sgl;
+		/* don't need sgl count? We assume it contains cmd->data_length data */
 	}
 
 	return 0;
@@ -4112,21 +4036,21 @@ static inline long long transport_dev_end_lba(struct se_device *dev)
 
 static int transport_get_sectors(struct se_cmd *cmd)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 
-	cmd->t_task->t_tasks_sectors =
+	cmd->t_task.t_tasks_sectors =
 		(cmd->data_length / dev->se_sub_dev->se_dev_attrib.block_size);
-	if (!(cmd->t_task->t_tasks_sectors))
-		cmd->t_task->t_tasks_sectors = 1;
+	if (!(cmd->t_task.t_tasks_sectors))
+		cmd->t_task.t_tasks_sectors = 1;
 
 	if (dev->transport->get_device_type(dev) != TYPE_DISK)
 		return 0;
 
-	if ((cmd->t_task->t_task_lba + cmd->t_task->t_tasks_sectors) >
+	if ((cmd->t_task.t_task_lba + cmd->t_task.t_tasks_sectors) >
 	     transport_dev_end_lba(dev)) {
 		printk(KERN_ERR "LBA: %llu Sectors: %u exceeds"
 			" transport_dev_end_lba(): %llu\n",
-			cmd->t_task->t_task_lba, cmd->t_task->t_tasks_sectors,
+			cmd->t_task.t_task_lba, cmd->t_task.t_tasks_sectors,
 			transport_dev_end_lba(dev));
 		cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
 		cmd->scsi_sense_reason = TCM_SECTOR_COUNT_TOO_MANY;
@@ -4138,26 +4062,26 @@ static int transport_get_sectors(struct se_cmd *cmd)
 
 static int transport_new_cmd_obj(struct se_cmd *cmd)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	u32 task_cdbs = 0, rc;
 
 	if (!(cmd->se_cmd_flags & SCF_SCSI_DATA_SG_IO_CDB)) {
 		task_cdbs++;
-		cmd->t_task->t_task_cdbs++;
+		cmd->t_task.t_task_cdbs++;
 	} else {
 		int set_counts = 1;
 
 		/*
 		 * Setup any BIDI READ tasks and memory from
-		 * cmd->t_task->t_mem_bidi_list so the READ struct se_tasks
+		 * cmd->t_task.t_mem_bidi_list so the READ struct se_tasks
 		 * are queued first for the non pSCSI passthrough case.
 		 */
-		if ((cmd->t_task->t_mem_bidi_list != NULL) &&
+		if (!list_empty(&cmd->t_task.t_mem_bidi_list) &&
 		    (dev->transport->transport_type != TRANSPORT_PLUGIN_PHBA_PDEV)) {
 			rc = transport_generic_get_cdb_count(cmd,
-				cmd->t_task->t_task_lba,
-				cmd->t_task->t_tasks_sectors,
-				DMA_FROM_DEVICE, cmd->t_task->t_mem_bidi_list,
+				cmd->t_task.t_task_lba,
+				cmd->t_task.t_tasks_sectors,
+				DMA_FROM_DEVICE, &cmd->t_task.t_mem_bidi_list,
 				set_counts);
 			if (!(rc)) {
 				cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
@@ -4168,13 +4092,13 @@ static int transport_new_cmd_obj(struct se_cmd *cmd)
 			set_counts = 0;
 		}
 		/*
-		 * Setup the tasks and memory from cmd->t_task->t_mem_list
+		 * Setup the tasks and memory from cmd->t_task.t_mem_list
 		 * Note for BIDI transfers this will contain the WRITE payload
 		 */
 		task_cdbs = transport_generic_get_cdb_count(cmd,
-				cmd->t_task->t_task_lba,
-				cmd->t_task->t_tasks_sectors,
-				cmd->data_direction, cmd->t_task->t_mem_list,
+				cmd->t_task.t_task_lba,
+				cmd->t_task.t_tasks_sectors,
+				cmd->data_direction, &cmd->t_task.t_mem_list,
 				set_counts);
 		if (!(task_cdbs)) {
 			cmd->se_cmd_flags |= SCF_SCSI_CDB_EXCEPTION;
@@ -4182,63 +4106,34 @@ static int transport_new_cmd_obj(struct se_cmd *cmd)
 					TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
 			return PYX_TRANSPORT_LU_COMM_FAILURE;
 		}
-		cmd->t_task->t_task_cdbs += task_cdbs;
+		cmd->t_task.t_task_cdbs += task_cdbs;
 
 #if 0
 		printk(KERN_INFO "data_length: %u, LBA: %llu t_tasks_sectors:"
 			" %u, t_task_cdbs: %u\n", obj_ptr, cmd->data_length,
-			cmd->t_task->t_task_lba, cmd->t_task->t_tasks_sectors,
-			cmd->t_task->t_task_cdbs);
+			cmd->t_task.t_task_lba, cmd->t_task.t_tasks_sectors,
+			cmd->t_task.t_task_cdbs);
 #endif
 	}
 
-	atomic_set(&cmd->t_task->t_task_cdbs_left, task_cdbs);
-	atomic_set(&cmd->t_task->t_task_cdbs_ex_left, task_cdbs);
-	atomic_set(&cmd->t_task->t_task_cdbs_timeout_left, task_cdbs);
+	atomic_set(&cmd->t_task.t_task_cdbs_left, task_cdbs);
+	atomic_set(&cmd->t_task.t_task_cdbs_ex_left, task_cdbs);
+	atomic_set(&cmd->t_task.t_task_cdbs_timeout_left, task_cdbs);
 	return 0;
 }
 
-static struct list_head *transport_init_se_mem_list(void)
-{
-	struct list_head *se_mem_list;
-
-	se_mem_list = kzalloc(sizeof(struct list_head), GFP_KERNEL);
-	if (!(se_mem_list)) {
-		printk(KERN_ERR "Unable to allocate memory for se_mem_list\n");
-		return NULL;
-	}
-	INIT_LIST_HEAD(se_mem_list);
-
-	return se_mem_list;
-}
-
 static int
 transport_generic_get_mem(struct se_cmd *cmd, u32 length, u32 dma_size)
 {
 	unsigned char *buf;
 	struct se_mem *se_mem;
 
-	cmd->t_task->t_mem_list = transport_init_se_mem_list();
-	if (!(cmd->t_task->t_mem_list))
-		return -ENOMEM;
-
 	/*
 	 * If the device uses memory mapping this is enough.
 	 */
 	if (cmd->se_dev->transport->do_se_mem_map)
 		return 0;
 
-	/*
-	 * Setup BIDI-COMMAND READ list of struct se_mem elements
-	 */
-	if (cmd->t_task->t_tasks_bidi) {
-		cmd->t_task->t_mem_bidi_list = transport_init_se_mem_list();
-		if (!(cmd->t_task->t_mem_bidi_list)) {
-			kfree(cmd->t_task->t_mem_list);
-			return -ENOMEM;
-		}
-	}
-
 	while (length) {
 		se_mem = kmem_cache_zalloc(se_mem_cache, GFP_KERNEL);
 		if (!(se_mem)) {
@@ -4263,8 +4158,8 @@ transport_generic_get_mem(struct se_cmd *cmd, u32 length, u32 dma_size)
 		memset(buf, 0, se_mem->se_len);
 		kunmap_atomic(buf, KM_IRQ0);
 
-		list_add_tail(&se_mem->se_list, cmd->t_task->t_mem_list);
-		cmd->t_task->t_tasks_se_num++;
+		list_add_tail(&se_mem->se_list, &cmd->t_task.t_mem_list);
+		cmd->t_task.t_tasks_se_num++;
 
 		DEBUG_MEM("Allocated struct se_mem page(%p) Length(%u)"
 			" Offset(%u)\n", se_mem->se_page, se_mem->se_len,
@@ -4274,7 +4169,7 @@ transport_generic_get_mem(struct se_cmd *cmd, u32 length, u32 dma_size)
 	}
 
 	DEBUG_MEM("Allocated total struct se_mem elements(%u)\n",
-			cmd->t_task->t_tasks_se_num);
+			cmd->t_task.t_tasks_se_num);
 
 	return 0;
 out:
@@ -4290,7 +4185,7 @@ int transport_init_task_sg(
 	u32 task_offset)
 {
 	struct se_cmd *se_cmd = task->task_se_cmd;
-	struct se_device *se_dev = se_cmd->se_lun->lun_se_dev;
+	struct se_device *se_dev = se_cmd->se_dev;
 	struct se_mem *se_mem = in_se_mem;
 	struct target_core_fabric_ops *tfo = se_cmd->se_tfo;
 	u32 sg_length, task_size = task->task_size, task_sg_num_padded;
@@ -4306,7 +4201,7 @@ int transport_init_task_sg(
 				sg_length = se_mem->se_len;
 
 				if (!(list_is_last(&se_mem->se_list,
-						se_cmd->t_task->t_mem_list)))
+						&se_cmd->t_task.t_mem_list)))
 					se_mem = list_entry(se_mem->se_list.next,
 							struct se_mem, se_list);
 			} else {
@@ -4326,7 +4221,7 @@ int transport_init_task_sg(
 				sg_length = (se_mem->se_len - task_offset);
 
 				if (!(list_is_last(&se_mem->se_list,
-						se_cmd->t_task->t_mem_list)))
+						&se_cmd->t_task.t_mem_list)))
 					se_mem = list_entry(se_mem->se_list.next,
 							struct se_mem, se_list);
 			}
@@ -4367,7 +4262,7 @@ next:
 	 * Setup task->task_sg_bidi for SCSI READ payload for
 	 * TCM/pSCSI passthrough if present for BIDI-COMMAND
 	 */
-	if ((se_cmd->t_task->t_mem_bidi_list != NULL) &&
+	if (!list_empty(&se_cmd->t_task.t_mem_bidi_list) &&
 	    (se_dev->transport->transport_type == TRANSPORT_PLUGIN_PHBA_PDEV)) {
 		task->task_sg_bidi = kzalloc(task_sg_num_padded *
 				sizeof(struct scatterlist), GFP_KERNEL);
@@ -4458,21 +4353,26 @@ static inline int transport_set_tasks_sectors(
 				max_sectors_set);
 }
 
+/*
+ * Convert a sgl into a linked list of se_mems.
+ */
 static int transport_map_sg_to_mem(
 	struct se_cmd *cmd,
 	struct list_head *se_mem_list,
-	void *in_mem,
-	u32 *se_mem_cnt)
+	struct scatterlist *sg,
+	u32 *sg_count)
 {
 	struct se_mem *se_mem;
-	struct scatterlist *sg;
-	u32 sg_count = 1, cmd_size = cmd->data_length;
+	u32 cmd_size = cmd->data_length;
 
-	WARN_ON(!in_mem);
-
-	sg = (struct scatterlist *)in_mem;
+	WARN_ON(!sg);
 
 	while (cmd_size) {
+		/*
+		 * NOTE: it is safe to return -ENOMEM at any time in creating this
+		 * list because transport_free_pages() will eventually be called, and is
+		 * smart enough to deallocate all list items for sg and sg_bidi lists.
+		 */
 		se_mem = kmem_cache_zalloc(se_mem_cache, GFP_KERNEL);
 		if (!(se_mem)) {
 			printk(KERN_ERR "Unable to allocate struct se_mem\n");
@@ -4489,26 +4389,21 @@ static int transport_map_sg_to_mem(
 		if (cmd_size > sg->length) {
 			se_mem->se_len = sg->length;
 			sg = sg_next(sg);
-			sg_count++;
 		} else
 			se_mem->se_len = cmd_size;
 
 		cmd_size -= se_mem->se_len;
+		(*sg_count)++;
 
-		DEBUG_MEM("sg_to_mem: *se_mem_cnt: %u cmd_size: %u\n",
-				*se_mem_cnt, cmd_size);
+		DEBUG_MEM("sg_to_mem: sg_count: %u cmd_size: %u\n",
+				sg_count, cmd_size);
 		DEBUG_MEM("sg_to_mem: Final se_page: %p se_off: %d se_len: %d\n",
 				se_mem->se_page, se_mem->se_off, se_mem->se_len);
 
 		list_add_tail(&se_mem->se_list, se_mem_list);
-		(*se_mem_cnt)++;
 	}
 
-	DEBUG_MEM("task[0] - Mapped(%u) struct scatterlist segments to(%u)"
-		" struct se_mem\n", sg_count, *se_mem_cnt);
-
-	if (sg_count != *se_mem_cnt)
-		BUG();
+	DEBUG_MEM("task[0] - Mapped(%u) struct scatterlist segments\n", sg_count);
 
 	return 0;
 }
@@ -4551,7 +4446,7 @@ int transport_map_mem_to_sg(
 				sg->length = se_mem->se_len;
 
 				if (!(list_is_last(&se_mem->se_list,
-						se_cmd->t_task->t_mem_list))) {
+						&se_cmd->t_task.t_mem_list))) {
 					se_mem = list_entry(se_mem->se_list.next,
 							struct se_mem, se_list);
 					(*se_mem_cnt)++;
@@ -4587,7 +4482,7 @@ int transport_map_mem_to_sg(
 				sg->length = (se_mem->se_len - *task_offset);
 
 				if (!(list_is_last(&se_mem->se_list,
-						se_cmd->t_task->t_mem_list))) {
+						&se_cmd->t_task.t_mem_list))) {
 					se_mem = list_entry(se_mem->se_list.next,
 							struct se_mem, se_list);
 					(*se_mem_cnt)++;
@@ -4645,7 +4540,7 @@ void transport_do_task_sg_chain(struct se_cmd *cmd)
 	 * Walk the struct se_task list and setup scatterlist chains
 	 * for each contiguosly allocated struct se_task->task_sg[].
 	 */
-	list_for_each_entry(task, &cmd->t_task->t_task_list, t_list) {
+	list_for_each_entry(task, &cmd->t_task.t_task_list, t_list) {
 		if (!(task->task_sg) || !(task->task_padded_sg))
 			continue;
 
@@ -4656,7 +4551,7 @@ void transport_do_task_sg_chain(struct se_cmd *cmd)
 			 * Either add chain or mark end of scatterlist
 			 */
 			if (!(list_is_last(&task->t_list,
-					&cmd->t_task->t_task_list))) {
+					&cmd->t_task.t_task_list))) {
 				/*
 				 * Clear existing SGL termination bit set in
 				 * transport_init_task_sg(), see sg_mark_end()
@@ -4682,7 +4577,7 @@ void transport_do_task_sg_chain(struct se_cmd *cmd)
 		/*
 		 * Check for single task..
 		 */
-		if (!(list_is_last(&task->t_list, &cmd->t_task->t_task_list))) {
+		if (!(list_is_last(&task->t_list, &cmd->t_task.t_task_list))) {
 			/*
 			 * Clear existing SGL termination bit set in
 			 * transport_init_task_sg(), see sg_mark_end()
@@ -4700,18 +4595,18 @@ void transport_do_task_sg_chain(struct se_cmd *cmd)
 	 * Setup the starting pointer and total t_tasks_sg_linked_no including
 	 * padding SGs for linking and to mark the end.
 	 */
-	cmd->t_task->t_tasks_sg_chained = sg_first;
-	cmd->t_task->t_tasks_sg_chained_no = sg_count;
+	cmd->t_task.t_tasks_sg_chained = sg_first;
+	cmd->t_task.t_tasks_sg_chained_no = sg_count;
 
-	DEBUG_CMD_M("Setup cmd: %p cmd->t_task->t_tasks_sg_chained: %p and"
-		" t_tasks_sg_chained_no: %u\n", cmd, cmd->t_task->t_tasks_sg_chained,
-		cmd->t_task->t_tasks_sg_chained_no);
+	DEBUG_CMD_M("Setup cmd: %p cmd->t_task.t_tasks_sg_chained: %p and"
+		" t_tasks_sg_chained_no: %u\n", cmd, cmd->t_task.t_tasks_sg_chained,
+		cmd->t_task.t_tasks_sg_chained_no);
 
-	for_each_sg(cmd->t_task->t_tasks_sg_chained, sg,
-			cmd->t_task->t_tasks_sg_chained_no, i) {
+	for_each_sg(cmd->t_task.t_tasks_sg_chained, sg,
+			cmd->t_task.t_tasks_sg_chained_no, i) {
 
-		DEBUG_CMD_M("SG[%d]: %p page: %p length: %d offset: %d, magic: 0x%08x\n",
-			i, sg, sg_page(sg), sg->length, sg->offset, sg->sg_magic);
+		DEBUG_CMD_M("SG[%d]: %p page: %p length: %d offset: %d\n",
+			i, sg, sg_page(sg), sg->length, sg->offset);
 		if (sg_is_chain(sg))
 			DEBUG_CMD_M("SG: %p sg_is_chain=1\n", sg);
 		if (sg_is_last(sg))
@@ -4741,7 +4636,7 @@ static int transport_do_se_mem_map(
 				in_mem, in_se_mem, out_se_mem, se_mem_cnt,
 				task_offset_in);
 		if (ret == 0)
-			task->task_se_cmd->t_task->t_tasks_se_num += *se_mem_cnt;
+			task->task_se_cmd->t_task.t_tasks_se_num += *se_mem_cnt;
 
 		return ret;
 	}
@@ -4791,7 +4686,7 @@ static u32 transport_generic_get_cdb_count(
 	struct se_task *task;
 	struct se_mem *se_mem = NULL, *se_mem_lout = NULL;
 	struct se_mem *se_mem_bidi = NULL, *se_mem_bidi_lout = NULL;
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	int max_sectors_set = 0, ret;
 	u32 task_offset_in = 0, se_mem_cnt = 0, se_mem_bidi_cnt = 0, task_cdbs = 0;
 
@@ -4805,15 +4700,14 @@ static u32 transport_generic_get_cdb_count(
 	 * mem_list will ever be empty at this point.
 	 */
 	if (!(list_empty(mem_list)))
-		se_mem = list_entry(mem_list->next, struct se_mem, se_list);
+		se_mem = list_first_entry(mem_list, struct se_mem, se_list);
 	/*
 	 * Check for extra se_mem_bidi mapping for BIDI-COMMANDs to
 	 * struct se_task->task_sg_bidi for TCM/pSCSI passthrough operation
 	 */
-	if ((cmd->t_task->t_mem_bidi_list != NULL) &&
-	    !(list_empty(cmd->t_task->t_mem_bidi_list)) &&
+	if (!list_empty(&cmd->t_task.t_mem_bidi_list) &&
 	    (dev->transport->transport_type == TRANSPORT_PLUGIN_PHBA_PDEV))
-		se_mem_bidi = list_entry(cmd->t_task->t_mem_bidi_list->next,
+		se_mem_bidi = list_first_entry(&cmd->t_task.t_mem_bidi_list,
 					struct se_mem, se_list);
 
 	while (sectors) {
@@ -4836,15 +4730,15 @@ static u32 transport_generic_get_cdb_count(
 
 		cdb = dev->transport->get_cdb(task);
 		if ((cdb)) {
-			memcpy(cdb, cmd->t_task->t_task_cdb,
-				scsi_command_size(cmd->t_task->t_task_cdb));
+			memcpy(cdb, cmd->t_task.t_task_cdb,
+				scsi_command_size(cmd->t_task.t_task_cdb));
 			cmd->transport_split_cdb(task->task_lba,
 					&task->task_sectors, cdb);
 		}
 
 		/*
 		 * Perform the SE OBJ plugin and/or Transport plugin specific
-		 * mapping for cmd->t_task->t_mem_list. And setup the
+		 * mapping for cmd->t_task.t_mem_list. And setup the
 		 * task->task_sg and if necessary task->task_sg_bidi
 		 */
 		ret = transport_do_se_mem_map(dev, task, mem_list,
@@ -4855,7 +4749,7 @@ static u32 transport_generic_get_cdb_count(
 
 		se_mem = se_mem_lout;
 		/*
-		 * Setup the cmd->t_task->t_mem_bidi_list -> task->task_sg_bidi
+		 * Setup the cmd->t_task.t_mem_bidi_list -> task->task_sg_bidi
 		 * mapping for SCSI READ for BIDI-COMMAND passthrough with TCM/pSCSI
 		 *
 		 * Note that the first call to transport_do_se_mem_map() above will
@@ -4865,7 +4759,7 @@ static u32 transport_generic_get_cdb_count(
 		 */
 		if (task->task_sg_bidi != NULL) {
 			ret = transport_do_se_mem_map(dev, task,
-				cmd->t_task->t_mem_bidi_list, NULL,
+				&cmd->t_task.t_mem_bidi_list, NULL,
 				se_mem_bidi, &se_mem_bidi_lout, &se_mem_bidi_cnt,
 				&task_offset_in);
 			if (ret < 0)
@@ -4888,8 +4782,8 @@ static u32 transport_generic_get_cdb_count(
 	}
 
 	if (set_counts) {
-		atomic_inc(&cmd->t_task->t_fe_count);
-		atomic_inc(&cmd->t_task->t_se_count);
+		atomic_inc(&cmd->t_task.t_fe_count);
+		atomic_inc(&cmd->t_task.t_se_count);
 	}
 
 	DEBUG_VOL("ITT[0x%08x] total %s cdbs(%u)\n",
@@ -4904,7 +4798,7 @@ out:
 static int
 transport_map_control_cmd_to_task(struct se_cmd *cmd)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	unsigned char *cdb;
 	struct se_task *task;
 	int ret;
@@ -4915,26 +4809,26 @@ transport_map_control_cmd_to_task(struct se_cmd *cmd)
 
 	cdb = dev->transport->get_cdb(task);
 	if (cdb)
-		memcpy(cdb, cmd->t_task->t_task_cdb,
-			scsi_command_size(cmd->t_task->t_task_cdb));
+		memcpy(cdb, cmd->t_task.t_task_cdb,
+			scsi_command_size(cmd->t_task.t_task_cdb));
 
 	task->task_size = cmd->data_length;
 	task->task_sg_num =
 		(cmd->se_cmd_flags & SCF_SCSI_CONTROL_SG_IO_CDB) ? 1 : 0;
 
-	atomic_inc(&cmd->t_task->t_fe_count);
-	atomic_inc(&cmd->t_task->t_se_count);
+	atomic_inc(&cmd->t_task.t_fe_count);
+	atomic_inc(&cmd->t_task.t_se_count);
 
 	if (cmd->se_cmd_flags & SCF_SCSI_CONTROL_SG_IO_CDB) {
 		struct se_mem *se_mem = NULL, *se_mem_lout = NULL;
 		u32 se_mem_cnt = 0, task_offset = 0;
 
-		if (!list_empty(cmd->t_task->t_mem_list))
-			se_mem = list_entry(cmd->t_task->t_mem_list->next,
+		if (!list_empty(&cmd->t_task.t_mem_list))
+			se_mem = list_first_entry(&cmd->t_task.t_mem_list,
 					struct se_mem, se_list);
 
 		ret = transport_do_se_mem_map(dev, task,
-				cmd->t_task->t_mem_list, NULL, se_mem,
+				&cmd->t_task.t_mem_list, NULL, se_mem,
 				&se_mem_lout, &se_mem_cnt, &task_offset);
 		if (ret < 0)
 			return PYX_TRANSPORT_OUT_OF_MEMORY_RESOURCES;
@@ -4969,14 +4863,14 @@ static int transport_generic_new_cmd(struct se_cmd *cmd)
 {
 	struct se_portal_group *se_tpg;
 	struct se_task *task;
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	int ret = 0;
 
 	/*
 	 * Determine is the TCM fabric module has already allocated physical
 	 * memory, and is directly calling transport_generic_map_mem_to_cmd()
 	 * to setup beforehand the linked list of physical memory at
-	 * cmd->t_task->t_mem_list of struct se_mem->se_page
+	 * cmd->t_task.t_mem_list of struct se_mem->se_page
 	 */
 	if (!(cmd->se_cmd_flags & SCF_PASSTHROUGH_SG_TO_MEM_NOALLOC)) {
 		ret = transport_allocate_resources(cmd);
@@ -5005,7 +4899,7 @@ static int transport_generic_new_cmd(struct se_cmd *cmd)
 	}
 
 	if (cmd->se_cmd_flags & SCF_SCSI_DATA_SG_IO_CDB) {
-		list_for_each_entry(task, &cmd->t_task->t_task_list, t_list) {
+		list_for_each_entry(task, &cmd->t_task.t_task_list, t_list) {
 			if (atomic_read(&task->task_sent))
 				continue;
 			if (!dev->transport->map_task_SG)
@@ -5052,9 +4946,9 @@ void transport_generic_process_write(struct se_cmd *cmd)
 	 * original EDTL
 	 */
 	if (cmd->se_cmd_flags & SCF_UNDERFLOW_BIT) {
-		if (!cmd->t_task->t_tasks_se_num) {
+		if (!cmd->t_task.t_tasks_se_num) {
 			unsigned char *dst, *buf =
-				(unsigned char *)cmd->t_task->t_task_buf;
+				(unsigned char *)cmd->t_task.t_task_buf;
 
 			dst = kzalloc(cmd->cmd_spdtl), GFP_KERNEL);
 			if (!(dst)) {
@@ -5066,15 +4960,15 @@ void transport_generic_process_write(struct se_cmd *cmd)
 			}
 			memcpy(dst, buf, cmd->cmd_spdtl);
 
-			kfree(cmd->t_task->t_task_buf);
-			cmd->t_task->t_task_buf = dst;
+			kfree(cmd->t_task.t_task_buf);
+			cmd->t_task.t_task_buf = dst;
 		} else {
 			struct scatterlist *sg =
-				(struct scatterlist *sg)cmd->t_task->t_task_buf;
+				(struct scatterlist *sg)cmd->t_task.t_task_buf;
 			struct scatterlist *orig_sg;
 
 			orig_sg = kzalloc(sizeof(struct scatterlist) *
-					cmd->t_task->t_tasks_se_num,
+					cmd->t_task.t_tasks_se_num,
 					GFP_KERNEL))) {
 			if (!(orig_sg)) {
 				printk(KERN_ERR "Unable to allocate memory"
@@ -5084,9 +4978,9 @@ void transport_generic_process_write(struct se_cmd *cmd)
 				return;
 			}
 
-			memcpy(orig_sg, cmd->t_task->t_task_buf,
+			memcpy(orig_sg, cmd->t_task.t_task_buf,
 					sizeof(struct scatterlist) *
-					cmd->t_task->t_tasks_se_num);
+					cmd->t_task.t_tasks_se_num);
 
 			cmd->data_length = cmd->cmd_spdtl;
 			/*
@@ -5117,22 +5011,22 @@ static int transport_generic_write_pending(struct se_cmd *cmd)
 	unsigned long flags;
 	int ret;
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 	cmd->t_state = TRANSPORT_WRITE_PENDING;
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 	/*
 	 * For the TCM control CDBs using a contiguous buffer, do the memcpy
 	 * from the passed Linux/SCSI struct scatterlist located at
-	 * se_cmd->t_task->t_task_pt_buf to the contiguous buffer at
-	 * se_cmd->t_task->t_task_buf.
+	 * se_cmd->t_task.t_task_pt_buf to the contiguous buffer at
+	 * se_cmd->t_task.t_task_buf.
 	 */
 	if (cmd->se_cmd_flags & SCF_PASSTHROUGH_CONTIG_TO_SG)
 		transport_memcpy_read_contig(cmd,
-				cmd->t_task->t_task_buf,
-				cmd->t_task->t_task_pt_sgl);
+				cmd->t_task.t_task_buf,
+				cmd->t_task.t_task_pt_sgl);
 	/*
 	 * Clear the se_cmd for WRITE_PENDING status in order to set
-	 * cmd->t_task->t_transport_active=0 so that transport_generic_handle_data
+	 * cmd->t_task.t_transport_active=0 so that transport_generic_handle_data
 	 * can be called from HW target mode interrupt code.  This is safe
 	 * to be called with transport_off=1 before the cmd->se_tfo->write_pending
 	 * because the se_cmd->se_lun pointer is not being cleared.
@@ -5156,7 +5050,6 @@ static int transport_generic_write_pending(struct se_cmd *cmd)
  */
 void transport_release_cmd_to_pool(struct se_cmd *cmd)
 {
-	BUG_ON(!cmd->t_task);
 	BUG_ON(!cmd->se_tfo);
 
 	transport_free_se_cmd(cmd);
@@ -5174,7 +5067,7 @@ void transport_generic_free_cmd(
 	int release_to_pool,
 	int session_reinstatement)
 {
-	if (!(cmd->se_cmd_flags & SCF_SE_LUN_CMD) || !cmd->t_task)
+	if (!(cmd->se_cmd_flags & SCF_SE_LUN_CMD))
 		transport_release_cmd_to_pool(cmd);
 	else {
 		core_dec_lacl_count(cmd->se_sess->se_node_acl, cmd);
@@ -5220,32 +5113,32 @@ static int transport_lun_wait_for_tasks(struct se_cmd *cmd, struct se_lun *lun)
 	 * If the frontend has already requested this struct se_cmd to
 	 * be stopped, we can safely ignore this struct se_cmd.
 	 */
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
-	if (atomic_read(&cmd->t_task->t_transport_stop)) {
-		atomic_set(&cmd->t_task->transport_lun_stop, 0);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
+	if (atomic_read(&cmd->t_task.t_transport_stop)) {
+		atomic_set(&cmd->t_task.transport_lun_stop, 0);
 		DEBUG_TRANSPORT_S("ConfigFS ITT[0x%08x] - t_transport_stop =="
 			" TRUE, skipping\n", cmd->se_tfo->get_task_tag(cmd));
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		transport_cmd_check_stop(cmd, 1, 0);
 		return -EPERM;
 	}
-	atomic_set(&cmd->t_task->transport_lun_fe_stop, 1);
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	atomic_set(&cmd->t_task.transport_lun_fe_stop, 1);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
-	wake_up_interruptible(&cmd->se_lun->lun_se_dev->dev_queue_obj.thread_wq);
+	wake_up_interruptible(&cmd->se_dev->dev_queue_obj.thread_wq);
 
 	ret = transport_stop_tasks_for_cmd(cmd);
 
 	DEBUG_TRANSPORT_S("ConfigFS: cmd: %p t_task_cdbs: %d stop tasks ret:"
-			" %d\n", cmd, cmd->t_task->t_task_cdbs, ret);
+			" %d\n", cmd, cmd->t_task.t_task_cdbs, ret);
 	if (!ret) {
 		DEBUG_TRANSPORT_S("ConfigFS: ITT[0x%08x] - stopping cmd....\n",
 				cmd->se_tfo->get_task_tag(cmd));
-		wait_for_completion(&cmd->t_task->transport_lun_stop_comp);
+		wait_for_completion(&cmd->t_task.transport_lun_stop_comp);
 		DEBUG_TRANSPORT_S("ConfigFS: ITT[0x%08x] - stopped cmd....\n",
 				cmd->se_tfo->get_task_tag(cmd));
 	}
-	transport_remove_cmd_from_queue(cmd, &cmd->se_lun->lun_se_dev->dev_queue_obj);
+	transport_remove_cmd_from_queue(cmd, &cmd->se_dev->dev_queue_obj);
 
 	return 0;
 }
@@ -5266,31 +5159,24 @@ static void __transport_clear_lun_from_sessions(struct se_lun *lun)
 	 * Initiator Port.
 	 */
 	spin_lock_irqsave(&lun->lun_cmd_lock, lun_flags);
-	while (!list_empty_careful(&lun->lun_cmd_list)) {
-		cmd = list_entry(lun->lun_cmd_list.next,
-			struct se_cmd, se_lun_list);
-		list_del(&cmd->se_lun_list);
-
-		if (!(cmd->t_task)) {
-			printk(KERN_ERR "ITT: 0x%08x, cmd->t_task = NULL"
-				"[i,t]_state: %u/%u\n",
-				cmd->se_tfo->get_task_tag(cmd),
-				cmd->se_tfo->get_cmd_state(cmd), cmd->t_state);
-			BUG();
-		}
-		atomic_set(&cmd->t_task->transport_lun_active, 0);
+	while (!list_empty(&lun->lun_cmd_list)) {
+		cmd = list_first_entry(&lun->lun_cmd_list,
+		       struct se_cmd, se_lun_node);
+		list_del(&cmd->se_lun_node);
+
+		atomic_set(&cmd->t_task.transport_lun_active, 0);
 		/*
 		 * This will notify iscsi_target_transport.c:
 		 * transport_cmd_check_stop() that a LUN shutdown is in
 		 * progress for the iscsi_cmd_t.
 		 */
-		spin_lock(&cmd->t_task->t_state_lock);
-		DEBUG_CLEAR_L("SE_LUN[%d] - Setting cmd->t_task->transport"
+		spin_lock(&cmd->t_task.t_state_lock);
+		DEBUG_CLEAR_L("SE_LUN[%d] - Setting cmd->t_task.transport"
 			"_lun_stop for  ITT: 0x%08x\n",
 			cmd->se_lun->unpacked_lun,
 			cmd->se_tfo->get_task_tag(cmd));
-		atomic_set(&cmd->t_task->transport_lun_stop, 1);
-		spin_unlock(&cmd->t_task->t_state_lock);
+		atomic_set(&cmd->t_task.transport_lun_stop, 1);
+		spin_unlock(&cmd->t_task.t_state_lock);
 
 		spin_unlock_irqrestore(&lun->lun_cmd_lock, lun_flags);
 
@@ -5318,14 +5204,14 @@ static void __transport_clear_lun_from_sessions(struct se_lun *lun)
 			cmd->se_lun->unpacked_lun,
 			cmd->se_tfo->get_task_tag(cmd));
 
-		spin_lock_irqsave(&cmd->t_task->t_state_lock, cmd_flags);
-		if (!(atomic_read(&cmd->t_task->transport_dev_active))) {
-			spin_unlock_irqrestore(&cmd->t_task->t_state_lock, cmd_flags);
+		spin_lock_irqsave(&cmd->t_task.t_state_lock, cmd_flags);
+		if (!(atomic_read(&cmd->t_task.transport_dev_active))) {
+			spin_unlock_irqrestore(&cmd->t_task.t_state_lock, cmd_flags);
 			goto check_cond;
 		}
-		atomic_set(&cmd->t_task->transport_dev_active, 0);
+		atomic_set(&cmd->t_task.transport_dev_active, 0);
 		transport_all_task_dev_remove_state(cmd);
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, cmd_flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, cmd_flags);
 
 		transport_free_dev_tasks(cmd);
 		/*
@@ -5342,24 +5228,24 @@ check_cond:
 		 * be released, notify the waiting thread now that LU has
 		 * finished accessing it.
 		 */
-		spin_lock_irqsave(&cmd->t_task->t_state_lock, cmd_flags);
-		if (atomic_read(&cmd->t_task->transport_lun_fe_stop)) {
+		spin_lock_irqsave(&cmd->t_task.t_state_lock, cmd_flags);
+		if (atomic_read(&cmd->t_task.transport_lun_fe_stop)) {
 			DEBUG_CLEAR_L("SE_LUN[%d] - Detected FE stop for"
 				" struct se_cmd: %p ITT: 0x%08x\n",
 				lun->unpacked_lun,
 				cmd, cmd->se_tfo->get_task_tag(cmd));
 
-			spin_unlock_irqrestore(&cmd->t_task->t_state_lock,
+			spin_unlock_irqrestore(&cmd->t_task.t_state_lock,
 					cmd_flags);
 			transport_cmd_check_stop(cmd, 1, 0);
-			complete(&cmd->t_task->transport_lun_fe_stop_comp);
+			complete(&cmd->t_task.transport_lun_fe_stop_comp);
 			spin_lock_irqsave(&lun->lun_cmd_lock, lun_flags);
 			continue;
 		}
 		DEBUG_CLEAR_L("SE_LUN[%d] - ITT: 0x%08x finished processing\n",
 			lun->unpacked_lun, cmd->se_tfo->get_task_tag(cmd));
 
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, cmd_flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, cmd_flags);
 		spin_lock_irqsave(&lun->lun_cmd_lock, lun_flags);
 	}
 	spin_unlock_irqrestore(&lun->lun_cmd_lock, lun_flags);
@@ -5379,7 +5265,7 @@ int transport_clear_lun_from_sessions(struct se_lun *lun)
 {
 	struct task_struct *kt;
 
-	kt = kthread_run(transport_clear_lun_thread, (void *)lun,
+	kt = kthread_run(transport_clear_lun_thread, lun,
 			"tcm_cl_%u", lun->unpacked_lun);
 	if (IS_ERR(kt)) {
 		printk(KERN_ERR "Unable to start clear_lun thread\n");
@@ -5405,15 +5291,15 @@ static void transport_generic_wait_for_tasks(
 	if (!(cmd->se_cmd_flags & SCF_SE_LUN_CMD) && !(cmd->se_tmr_req))
 		return;
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 	/*
 	 * If we are already stopped due to an external event (ie: LUN shutdown)
 	 * sleep until the connection can have the passed struct se_cmd back.
-	 * The cmd->t_task->transport_lun_stopped_sem will be upped by
+	 * The cmd->t_task.transport_lun_stopped_sem will be upped by
 	 * transport_clear_lun_from_sessions() once the ConfigFS context caller
 	 * has completed its operation on the struct se_cmd.
 	 */
-	if (atomic_read(&cmd->t_task->transport_lun_stop)) {
+	if (atomic_read(&cmd->t_task.transport_lun_stop)) {
 
 		DEBUG_TRANSPORT_S("wait_for_tasks: Stopping"
 			" wait_for_completion(&cmd->t_tasktransport_lun_fe"
@@ -5426,10 +5312,10 @@ static void transport_generic_wait_for_tasks(
 		 * We go ahead and up transport_lun_stop_comp just to be sure
 		 * here.
 		 */
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
-		complete(&cmd->t_task->transport_lun_stop_comp);
-		wait_for_completion(&cmd->t_task->transport_lun_fe_stop_comp);
-		spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
+		complete(&cmd->t_task.transport_lun_stop_comp);
+		wait_for_completion(&cmd->t_task.transport_lun_fe_stop_comp);
+		spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 
 		transport_all_task_dev_remove_state(cmd);
 		/*
@@ -5442,13 +5328,13 @@ static void transport_generic_wait_for_tasks(
 			"stop_comp); for ITT: 0x%08x\n",
 			cmd->se_tfo->get_task_tag(cmd));
 
-		atomic_set(&cmd->t_task->transport_lun_stop, 0);
+		atomic_set(&cmd->t_task.transport_lun_stop, 0);
 	}
-	if (!atomic_read(&cmd->t_task->t_transport_active) ||
-	     atomic_read(&cmd->t_task->t_transport_aborted))
+	if (!atomic_read(&cmd->t_task.t_transport_active) ||
+	     atomic_read(&cmd->t_task.t_transport_aborted))
 		goto remove;
 
-	atomic_set(&cmd->t_task->t_transport_stop, 1);
+	atomic_set(&cmd->t_task.t_transport_stop, 1);
 
 	DEBUG_TRANSPORT_S("wait_for_tasks: Stopping %p ITT: 0x%08x"
 		" i_state: %d, t_state/def_t_state: %d/%d, t_transport_stop"
@@ -5456,21 +5342,21 @@ static void transport_generic_wait_for_tasks(
 		cmd->se_tfo->get_cmd_state(cmd), cmd->t_state,
 		cmd->deferred_t_state);
 
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
-	wake_up_interruptible(&cmd->se_lun->lun_se_dev->dev_queue_obj.thread_wq);
+	wake_up_interruptible(&cmd->se_dev->dev_queue_obj.thread_wq);
 
-	wait_for_completion(&cmd->t_task->t_transport_stop_comp);
+	wait_for_completion(&cmd->t_task.t_transport_stop_comp);
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
-	atomic_set(&cmd->t_task->t_transport_active, 0);
-	atomic_set(&cmd->t_task->t_transport_stop, 0);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
+	atomic_set(&cmd->t_task.t_transport_active, 0);
+	atomic_set(&cmd->t_task.t_transport_stop, 0);
 
 	DEBUG_TRANSPORT_S("wait_for_tasks: Stopped wait_for_compltion("
-		"&cmd->t_task->t_transport_stop_comp) for ITT: 0x%08x\n",
+		"&cmd->t_task.t_transport_stop_comp) for ITT: 0x%08x\n",
 		cmd->se_tfo->get_task_tag(cmd));
 remove:
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 	if (!remove_cmd)
 		return;
 
@@ -5509,13 +5395,13 @@ int transport_send_check_condition_and_sense(
 	int offset;
 	u8 asc = 0, ascq = 0;
 
-	spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+	spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 	if (cmd->se_cmd_flags & SCF_SENT_CHECK_CONDITION) {
-		spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+		spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 		return 0;
 	}
 	cmd->se_cmd_flags |= SCF_SENT_CHECK_CONDITION;
-	spin_unlock_irqrestore(&cmd->t_task->t_state_lock, flags);
+	spin_unlock_irqrestore(&cmd->t_task.t_state_lock, flags);
 
 	if (!reason && from_transport)
 		goto after_reason;
@@ -5674,14 +5560,14 @@ int transport_check_aborted_status(struct se_cmd *cmd, int send_status)
 {
 	int ret = 0;
 
-	if (atomic_read(&cmd->t_task->t_transport_aborted) != 0) {
+	if (atomic_read(&cmd->t_task.t_transport_aborted) != 0) {
 		if (!(send_status) ||
 		     (cmd->se_cmd_flags & SCF_SENT_DELAYED_TAS))
 			return 1;
 #if 0
 		printk(KERN_INFO "Sending delayed SAM_STAT_TASK_ABORTED"
 			" status for CDB: 0x%02x ITT: 0x%08x\n",
-			cmd->t_task->t_task_cdb[0],
+			cmd->t_task.t_task_cdb[0],
 			cmd->se_tfo->get_task_tag(cmd));
 #endif
 		cmd->se_cmd_flags |= SCF_SENT_DELAYED_TAS;
@@ -5702,7 +5588,7 @@ void transport_send_task_abort(struct se_cmd *cmd)
 	 */
 	if (cmd->data_direction == DMA_TO_DEVICE) {
 		if (cmd->se_tfo->write_pending_status(cmd) != 0) {
-			atomic_inc(&cmd->t_task->t_transport_aborted);
+			atomic_inc(&cmd->t_task.t_transport_aborted);
 			smp_mb__after_atomic_inc();
 			cmd->scsi_status = SAM_STAT_TASK_ABORTED;
 			transport_new_cmd_failure(cmd);
@@ -5712,7 +5598,7 @@ void transport_send_task_abort(struct se_cmd *cmd)
 	cmd->scsi_status = SAM_STAT_TASK_ABORTED;
 #if 0
 	printk(KERN_INFO "Setting SAM_STAT_TASK_ABORTED status for CDB: 0x%02x,"
-		" ITT: 0x%08x\n", cmd->t_task->t_task_cdb[0],
+		" ITT: 0x%08x\n", cmd->t_task.t_task_cdb[0],
 		cmd->se_tfo->get_task_tag(cmd));
 #endif
 	cmd->se_tfo->queue_status(cmd);
@@ -5725,7 +5611,7 @@ void transport_send_task_abort(struct se_cmd *cmd)
 int transport_generic_do_tmr(struct se_cmd *cmd)
 {
 	struct se_cmd *ref_cmd;
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	struct se_tmr_req *tmr = cmd->se_tmr_req;
 	int ret;
 
@@ -5788,9 +5674,7 @@ transport_get_task_from_state_list(struct se_device *dev)
 static void transport_processing_shutdown(struct se_device *dev)
 {
 	struct se_cmd *cmd;
-	struct se_queue_req *qr;
 	struct se_task *task;
-	u8 state;
 	unsigned long flags;
 	/*
 	 * Empty the struct se_device's struct se_task state list.
@@ -5803,15 +5687,9 @@ static void transport_processing_shutdown(struct se_device *dev)
 		}
 		cmd = task->task_se_cmd;
 
-		if (!cmd->t_task) {
-			printk(KERN_ERR "cmd->t_task is NULL for task: %p cmd:"
-				" %p ITT: 0x%08x\n", task, cmd,
-				cmd->se_tfo->get_task_tag(cmd));
-			continue;
-		}
 		spin_unlock_irqrestore(&dev->execute_task_lock, flags);
 
-		spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
+		spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
 
 		DEBUG_DO("PT: cmd: %p task: %p ITT/CmdSN: 0x%08x/0x%08x,"
 			" i_state/def_i_state: %d/%d, t_state/def_t_state:"
@@ -5819,22 +5697,22 @@ static void transport_processing_shutdown(struct se_device *dev)
 			cmd->se_tfo->get_task_tag(cmd), cmd->cmd_sn,
 			cmd->se_tfo->get_cmd_state(cmd), cmd->deferred_i_state,
 			cmd->t_state, cmd->deferred_t_state,
-			cmd->t_task->t_task_cdb[0]);
+			cmd->t_task.t_task_cdb[0]);
 		DEBUG_DO("PT: ITT[0x%08x] - t_task_cdbs: %d t_task_cdbs_left:"
 			" %d t_task_cdbs_sent: %d -- t_transport_active: %d"
 			" t_transport_stop: %d t_transport_sent: %d\n",
 			cmd->se_tfo->get_task_tag(cmd),
-			cmd->t_task->t_task_cdbs,
-			atomic_read(&cmd->t_task->t_task_cdbs_left),
-			atomic_read(&cmd->t_task->t_task_cdbs_sent),
-			atomic_read(&cmd->t_task->t_transport_active),
-			atomic_read(&cmd->t_task->t_transport_stop),
-			atomic_read(&cmd->t_task->t_transport_sent));
+			cmd->t_task.t_task_cdbs,
+			atomic_read(&cmd->t_task.t_task_cdbs_left),
+			atomic_read(&cmd->t_task.t_task_cdbs_sent),
+			atomic_read(&cmd->t_task.t_transport_active),
+			atomic_read(&cmd->t_task.t_transport_stop),
+			atomic_read(&cmd->t_task.t_transport_sent));
 
 		if (atomic_read(&task->task_active)) {
 			atomic_set(&task->task_stop, 1);
 			spin_unlock_irqrestore(
-				&cmd->t_task->t_state_lock, flags);
+				&cmd->t_task.t_state_lock, flags);
 
 			DEBUG_DO("Waiting for task: %p to shutdown for dev:"
 				" %p\n", task, dev);
@@ -5842,8 +5720,8 @@ static void transport_processing_shutdown(struct se_device *dev)
 			DEBUG_DO("Completed task: %p shutdown for dev: %p\n",
 				task, dev);
 
-			spin_lock_irqsave(&cmd->t_task->t_state_lock, flags);
-			atomic_dec(&cmd->t_task->t_task_cdbs_left);
+			spin_lock_irqsave(&cmd->t_task.t_state_lock, flags);
+			atomic_dec(&cmd->t_task.t_task_cdbs_left);
 
 			atomic_set(&task->task_active, 0);
 			atomic_set(&task->task_stop, 0);
@@ -5853,39 +5731,39 @@ static void transport_processing_shutdown(struct se_device *dev)
 		}
 		__transport_stop_task_timer(task, &flags);
 
-		if (!(atomic_dec_and_test(&cmd->t_task->t_task_cdbs_ex_left))) {
+		if (!(atomic_dec_and_test(&cmd->t_task.t_task_cdbs_ex_left))) {
 			spin_unlock_irqrestore(
-					&cmd->t_task->t_state_lock, flags);
+					&cmd->t_task.t_state_lock, flags);
 
 			DEBUG_DO("Skipping task: %p, dev: %p for"
 				" t_task_cdbs_ex_left: %d\n", task, dev,
-				atomic_read(&cmd->t_task->t_task_cdbs_ex_left));
+				atomic_read(&cmd->t_task.t_task_cdbs_ex_left));
 
 			spin_lock_irqsave(&dev->execute_task_lock, flags);
 			continue;
 		}
 
-		if (atomic_read(&cmd->t_task->t_transport_active)) {
+		if (atomic_read(&cmd->t_task.t_transport_active)) {
 			DEBUG_DO("got t_transport_active = 1 for task: %p, dev:"
 					" %p\n", task, dev);
 
-			if (atomic_read(&cmd->t_task->t_fe_count)) {
+			if (atomic_read(&cmd->t_task.t_fe_count)) {
 				spin_unlock_irqrestore(
-					&cmd->t_task->t_state_lock, flags);
+					&cmd->t_task.t_state_lock, flags);
 				transport_send_check_condition_and_sense(
 					cmd, TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE,
 					0);
 				transport_remove_cmd_from_queue(cmd,
-					&cmd->se_lun->lun_se_dev->dev_queue_obj);
+					&cmd->se_dev->dev_queue_obj);
 
 				transport_lun_remove_cmd(cmd);
 				transport_cmd_check_stop(cmd, 1, 0);
 			} else {
 				spin_unlock_irqrestore(
-					&cmd->t_task->t_state_lock, flags);
+					&cmd->t_task.t_state_lock, flags);
 
 				transport_remove_cmd_from_queue(cmd,
-					&cmd->se_lun->lun_se_dev->dev_queue_obj);
+					&cmd->se_dev->dev_queue_obj);
 
 				transport_lun_remove_cmd(cmd);
 
@@ -5899,22 +5777,22 @@ static void transport_processing_shutdown(struct se_device *dev)
 		DEBUG_DO("Got t_transport_active = 0 for task: %p, dev: %p\n",
 				task, dev);
 
-		if (atomic_read(&cmd->t_task->t_fe_count)) {
+		if (atomic_read(&cmd->t_task.t_fe_count)) {
 			spin_unlock_irqrestore(
-				&cmd->t_task->t_state_lock, flags);
+				&cmd->t_task.t_state_lock, flags);
 			transport_send_check_condition_and_sense(cmd,
 				TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE, 0);
 			transport_remove_cmd_from_queue(cmd,
-				&cmd->se_lun->lun_se_dev->dev_queue_obj);
+				&cmd->se_dev->dev_queue_obj);
 
 			transport_lun_remove_cmd(cmd);
 			transport_cmd_check_stop(cmd, 1, 0);
 		} else {
 			spin_unlock_irqrestore(
-				&cmd->t_task->t_state_lock, flags);
+				&cmd->t_task.t_state_lock, flags);
 
 			transport_remove_cmd_from_queue(cmd,
-				&cmd->se_lun->lun_se_dev->dev_queue_obj);
+				&cmd->se_dev->dev_queue_obj);
 			transport_lun_remove_cmd(cmd);
 
 			if (transport_cmd_check_stop(cmd, 1, 0))
@@ -5927,15 +5805,12 @@ static void transport_processing_shutdown(struct se_device *dev)
 	/*
 	 * Empty the struct se_device's struct se_cmd list.
 	 */
-	while ((qr = transport_get_qr_from_queue(&dev->dev_queue_obj))) {
-		cmd = qr->cmd;
-		state = qr->state;
-		kfree(qr);
+	while ((cmd = transport_get_cmd_from_queue(&dev->dev_queue_obj))) {
 
 		DEBUG_DO("From Device Queue: cmd: %p t_state: %d\n",
-				cmd, state);
+				cmd, cmd->t_state);
 
-		if (atomic_read(&cmd->t_task->t_fe_count)) {
+		if (atomic_read(&cmd->t_task.t_fe_count)) {
 			transport_send_check_condition_and_sense(cmd,
 				TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE, 0);
 
@@ -5955,10 +5830,9 @@ static void transport_processing_shutdown(struct se_device *dev)
  */
 static int transport_processing_thread(void *param)
 {
-	int ret, t_state;
+	int ret;
 	struct se_cmd *cmd;
 	struct se_device *dev = (struct se_device *) param;
-	struct se_queue_req *qr;
 
 	set_user_nice(current, -20);
 
@@ -5980,15 +5854,11 @@ static int transport_processing_thread(void *param)
 get_cmd:
 		__transport_execute_tasks(dev);
 
-		qr = transport_get_qr_from_queue(&dev->dev_queue_obj);
-		if (!(qr))
+		cmd = transport_get_cmd_from_queue(&dev->dev_queue_obj);
+		if (!cmd)
 			continue;
 
-		cmd = qr->cmd;
-		t_state = qr->state;
-		kfree(qr);
-
-		switch (t_state) {
+		switch (cmd->t_state) {
 		case TRANSPORT_NEW_CMD_MAP:
 			if (!(cmd->se_tfo->new_cmd_map)) {
 				printk(KERN_ERR "cmd->se_tfo->new_cmd_map is"
@@ -6039,7 +5909,7 @@ get_cmd:
 		default:
 			printk(KERN_ERR "Unknown t_state: %d deferred_t_state:"
 				" %d for ITT: 0x%08x i_state: %d on SE LUN:"
-				" %u\n", t_state, cmd->deferred_t_state,
+				" %u\n", cmd->t_state, cmd->deferred_t_state,
 				cmd->se_tfo->get_task_tag(cmd),
 				cmd->se_tfo->get_cmd_state(cmd),
 				cmd->se_lun->unpacked_lun);
diff --git a/drivers/target/target_core_ua.c b/drivers/target/target_core_ua.c
index 16f41d1..3b8b02c 100644
--- a/drivers/target/target_core_ua.c
+++ b/drivers/target/target_core_ua.c
@@ -208,7 +208,7 @@ void core_scsi3_ua_for_check_condition(
 	u8 *asc,
 	u8 *ascq)
 {
-	struct se_device *dev = cmd->se_lun->lun_se_dev;
+	struct se_device *dev = cmd->se_dev;
 	struct se_dev_entry *deve;
 	struct se_session *sess = cmd->se_sess;
 	struct se_node_acl *nacl;
@@ -270,7 +270,7 @@ void core_scsi3_ua_for_check_condition(
 		nacl->se_tpg->se_tpg_tfo->get_fabric_name(),
 		(dev->se_sub_dev->se_dev_attrib.emulate_ua_intlck_ctrl != 0) ? "Reporting" :
 		"Releasing", dev->se_sub_dev->se_dev_attrib.emulate_ua_intlck_ctrl,
-		cmd->orig_fe_lun, cmd->t_task->t_task_cdb[0], *asc, *ascq);
+		cmd->orig_fe_lun, cmd->t_task.t_task_cdb[0], *asc, *ascq);
 }
 
 int core_scsi3_ua_clear_for_request_sense(
diff --git a/drivers/target/tcm_fc/tfc_cmd.c b/drivers/target/tcm_fc/tfc_cmd.c
index 19b2b99..6d9553b 100644
--- a/drivers/target/tcm_fc/tfc_cmd.c
+++ b/drivers/target/tcm_fc/tfc_cmd.c
@@ -72,16 +72,16 @@ void ft_dump_cmd(struct ft_cmd *cmd, const char *caller)
 		caller, cmd, cmd->cdb);
 	printk(KERN_INFO "%s: cmd %p lun %d\n", caller, cmd, cmd->lun);
 
-	task = se_cmd->t_task;
+	task = &se_cmd->t_task;
 	printk(KERN_INFO "%s: cmd %p task %p se_num %u buf %p len %u se_cmd_flags <0x%x>\n",
 	       caller, cmd, task, task->t_tasks_se_num,
 	       task->t_task_buf, se_cmd->data_length, se_cmd->se_cmd_flags);
-	if (task->t_mem_list)
-		list_for_each_entry(mem, task->t_mem_list, se_list)
-			printk(KERN_INFO "%s: cmd %p mem %p page %p "
-			       "len 0x%x off 0x%x\n",
-			       caller, cmd, mem,
-			       mem->se_page, mem->se_len, mem->se_off);
+
+	list_for_each_entry(mem, &task->t_mem_list, se_list)
+		printk(KERN_INFO "%s: cmd %p mem %p page %p "
+		       "len 0x%x off 0x%x\n",
+		       caller, cmd, mem,
+		       mem->se_page, mem->se_len, mem->se_off);
 	sp = cmd->seq;
 	if (sp) {
 		ep = fc_seq_exch(sp);
@@ -262,9 +262,9 @@ int ft_write_pending(struct se_cmd *se_cmd)
 				 * TCM/LIO target
 				 */
 				transport_do_task_sg_chain(se_cmd);
-				cmd->sg = se_cmd->t_task->t_tasks_sg_chained;
+				cmd->sg = se_cmd->t_task.t_tasks_sg_chained;
 				cmd->sg_cnt =
-					se_cmd->t_task->t_tasks_sg_chained_no;
+					se_cmd->t_task.t_tasks_sg_chained_no;
 			}
 			if (cmd->sg && lport->tt.ddp_setup(lport, ep->xid,
 						    cmd->sg, cmd->sg_cnt))
@@ -438,7 +438,7 @@ static void ft_send_tm(struct ft_cmd *cmd)
 	switch (fcp->fc_tm_flags) {
 	case FCP_TMF_LUN_RESET:
 		cmd->lun = scsilun_to_int((struct scsi_lun *)fcp->fc_lun);
-		if (transport_get_lun_for_tmr(&cmd->se_cmd, cmd->lun) < 0) {
+		if (transport_lookup_tmr_lun(&cmd->se_cmd, cmd->lun) < 0) {
 			/*
 			 * Make sure to clean up newly allocated TMR request
 			 * since "unable to  handle TMR request because failed
@@ -637,7 +637,7 @@ static void ft_send_cmd(struct ft_cmd *cmd)
 	fc_seq_exch(cmd->seq)->lp->tt.seq_set_resp(cmd->seq, ft_recv_seq, cmd);
 
 	cmd->lun = scsilun_to_int((struct scsi_lun *)fcp->fc_lun);
-	ret = transport_get_lun_for_cmd(&cmd->se_cmd, cmd->lun);
+	ret = transport_lookup_cmd_lun(&cmd->se_cmd, cmd->lun);
 	if (ret < 0) {
 		ft_dump_cmd(cmd, __func__);
 		transport_send_check_condition_and_sense(&cmd->se_cmd,
@@ -650,13 +650,13 @@ static void ft_send_cmd(struct ft_cmd *cmd)
 	FT_IO_DBG("r_ctl %x alloc task ret %d\n", fh->fh_r_ctl, ret);
 	ft_dump_cmd(cmd, __func__);
 
-	if (ret == -1) {
+	if (ret == -ENOMEM) {
 		transport_send_check_condition_and_sense(se_cmd,
 				TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE, 0);
 		transport_generic_free_cmd(se_cmd, 0, 1, 0);
 		return;
 	}
-	if (ret == -2) {
+	if (ret == -EINVAL) {
 		if (se_cmd->se_cmd_flags & SCF_SCSI_RESERVATION_CONFLICT)
 			ft_queue_status(se_cmd);
 		else
diff --git a/drivers/target/tcm_fc/tfc_conf.c b/drivers/target/tcm_fc/tfc_conf.c
index 8c5067c..58e4745 100644
--- a/drivers/target/tcm_fc/tfc_conf.c
+++ b/drivers/target/tcm_fc/tfc_conf.c
@@ -331,7 +331,7 @@ static struct se_portal_group *ft_add_tpg(
 	transport_init_queue_obj(&tpg->qobj);
 
 	ret = core_tpg_register(&ft_configfs->tf_ops, wwn, &tpg->se_tpg,
-				(void *)tpg, TRANSPORT_TPG_TYPE_NORMAL);
+				tpg, TRANSPORT_TPG_TYPE_NORMAL);
 	if (ret < 0) {
 		kfree(tpg);
 		return NULL;
diff --git a/drivers/target/tcm_fc/tfc_io.c b/drivers/target/tcm_fc/tfc_io.c
index 47efcfb..f18af6e 100644
--- a/drivers/target/tcm_fc/tfc_io.c
+++ b/drivers/target/tcm_fc/tfc_io.c
@@ -90,15 +90,14 @@ int ft_queue_data_in(struct se_cmd *se_cmd)
 	lport = ep->lp;
 	cmd->seq = lport->tt.seq_start_next(cmd->seq);
 
-	task = se_cmd->t_task;
-	BUG_ON(!task);
+	task = &se_cmd->t_task;
 	remaining = se_cmd->data_length;
 
 	/*
 	 * Setup to use first mem list entry if any.
 	 */
 	if (task->t_tasks_se_num) {
-		mem = list_first_entry(task->t_mem_list,
+		mem = list_first_entry(&task->t_mem_list,
 			 struct se_mem, se_list);
 		mem_len = mem->se_len;
 		mem_off = mem->se_off;
@@ -236,8 +235,7 @@ void ft_recv_write_data(struct ft_cmd *cmd, struct fc_frame *fp)
 	u32 f_ctl;
 	void *buf;
 
-	task = se_cmd->t_task;
-	BUG_ON(!task);
+	task = &se_cmd->t_task;
 
 	fh = fc_frame_header_get(fp);
 	if (!(ntoh24(fh->fh_f_ctl) & FC_FC_REL_OFF))
@@ -315,7 +313,7 @@ void ft_recv_write_data(struct ft_cmd *cmd, struct fc_frame *fp)
 	 * Setup to use first mem list entry if any.
 	 */
 	if (task->t_tasks_se_num) {
-		mem = list_first_entry(task->t_mem_list,
+		mem = list_first_entry(&task->t_mem_list,
 				       struct se_mem, se_list);
 		mem_len = mem->se_len;
 		mem_off = mem->se_off;
diff --git a/include/target/target_core_base.h b/include/target/target_core_base.h
index b0b83ed..94c838d 100644
--- a/include/target/target_core_base.h
+++ b/include/target/target_core_base.h
@@ -123,7 +123,7 @@ enum se_cmd_flags_table {
 	SCF_SENT_DELAYED_TAS		= 0x00020000,
 	SCF_ALUA_NON_OPTIMIZED		= 0x00040000,
 	SCF_DELAYED_CMD_FROM_SAM_ATTR	= 0x00080000,
-	SCF_PASSTHROUGH_SG_TO_MEM	= 0x00100000,
+	SCF_UNUSED			= 0x00100000,
 	SCF_PASSTHROUGH_CONTIG_TO_SG	= 0x00200000,
 	SCF_PASSTHROUGH_SG_TO_MEM_NOALLOC = 0x00400000,
 	SCF_EMULATE_SYNC_CACHE		= 0x00800000,
@@ -452,9 +452,9 @@ struct se_transport_task {
 	 * and other HW target mode fabric modules.
 	 */
 	struct scatterlist	*t_task_pt_sgl;
-	struct list_head	*t_mem_list;
+	struct list_head	t_mem_list;
 	/* Used for BIDI READ */
-	struct list_head	*t_mem_bidi_list;
+	struct list_head	t_mem_bidi_list;
 	struct list_head	t_task_list;
 } ____cacheline_aligned;
 
@@ -523,9 +523,9 @@ struct se_cmd {
 	atomic_t                transport_sent;
 	/* Used for sense data */
 	void			*sense_buffer;
-	struct list_head	se_delayed_list;
-	struct list_head	se_ordered_list;
-	struct list_head	se_lun_list;
+	struct list_head	se_delayed_node;
+	struct list_head	se_ordered_node;
+	struct list_head	se_lun_node;
 	struct se_device      *se_dev;
 	struct se_dev_entry   *se_deve;
 	struct se_device	*se_obj_ptr;
@@ -534,9 +534,8 @@ struct se_cmd {
 	/* Only used for internal passthrough and legacy TCM fabric modules */
 	struct se_session	*se_sess;
 	struct se_tmr_req	*se_tmr_req;
-	/* t_task is setup to t_task_backstore in transport_init_se_cmd() */
-	struct se_transport_task *t_task;
-	struct se_transport_task t_task_backstore;
+	struct se_transport_task t_task;
+	struct list_head	se_queue_node;
 	struct target_core_fabric_ops *se_tfo;
 	int (*transport_emulate_cdb)(struct se_cmd *);
 	void (*transport_split_cdb)(unsigned long long, u32 *, unsigned char *);
diff --git a/include/target/target_core_device.h b/include/target/target_core_device.h
index d9745bf..96586cc 100644
--- a/include/target/target_core_device.h
+++ b/include/target/target_core_device.h
@@ -1,8 +1,8 @@
 #ifndef TARGET_CORE_DEVICE_H
 #define TARGET_CORE_DEVICE_H
 
-extern int transport_get_lun_for_cmd(struct se_cmd *, u32);
-extern int transport_get_lun_for_tmr(struct se_cmd *, u32);
+extern int transport_lookup_cmd_lun(struct se_cmd *, u32);
+extern int transport_lookup_tmr_lun(struct se_cmd *, u32);
 extern struct se_dev_entry *core_get_se_deve_from_rtpi(
 					struct se_node_acl *, u16);
 extern int core_free_device_list_for_node(struct se_node_acl *,
diff --git a/include/target/target_core_transport.h b/include/target/target_core_transport.h
index 1dd4d18..acd5914 100644
--- a/include/target/target_core_transport.h
+++ b/include/target/target_core_transport.h
@@ -159,7 +159,6 @@ extern struct se_device *transport_add_device_to_core_hba(struct se_hba *,
 					struct se_subsystem_dev *, u32,
 					void *, struct se_dev_limits *,
 					const char *, const char *);
-extern void transport_device_setup_cmd(struct se_cmd *);
 extern void transport_init_se_cmd(struct se_cmd *,
 					struct target_core_fabric_ops *,
 					struct se_session *, u32, int, int,