Merge branch 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending

Pull SCSI target updates from Nicholas Bellinger:
 "Things were a lot more calm than previously expected. It's primarily
  fixes in various areas, with most of the new functionality centering
  around TCMU backend driver work that Xiubo Li has been driving.

  Here's the summary on the feature side:

   - Make T10-PI verify configurable for emulated (FILEIO + RD) backends
    (Dmitry Monakhov)
   - Allow target-core/TCMU pass-through to use in-kernel SPC-PR logic
    (Bryant Ly + MNC)
   - Add TCMU support for growing ring buffer size (Xiubo Li + MNC)
   - Add TCMU support for global block data pool (Xiubo Li + MNC)

  and on the bug-fix side:

   - Fix COMPARE_AND_WRITE non GOOD status handling for READ phase
    failures (Gary Guo + nab)
   - Fix iscsi-target hang with explicitly changing per NodeACL
    CmdSN number depth with concurrent login driven session
    reinstatement.  (Gary Guo + nab)
   - Fix ibmvscsis fabric driver ABORT task handling (Bryant Ly)
   - Fix target-core/FILEIO zero length handling (Bart Van Assche)

  Also, there was an OOPs introduced with the WRITE_VERIFY changes that
  I ended up reverting at the last minute, because as not unusual Bart
  and I could not agree on the fix in time for -rc1. Since it's specific
  to a conformance test, it's been reverted for now.

  There is a separate patch in the queue to address the underlying
  control CDB write overflow regression in >= v4.3 separate from the
  WRITE_VERIFY revert here, that will be pushed post -rc1"

* 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending: (30 commits)
  Revert "target: Fix VERIFY and WRITE VERIFY command parsing"
  IB/srpt: Avoid that aborting a command triggers a kernel warning
  IB/srpt: Fix abort handling
  target/fileio: Fix zero-length READ and WRITE handling
  ibmvscsis: Do not send aborted task response
  tcmu: fix module removal due to stuck thread
  target: Don't force session reset if queue_depth does not change
  iscsi-target: Set session_fall_back_to_erl0 when forcing reinstatement
  target: Fix compare_and_write_callback handling for non GOOD status
  tcmu: Recalculate the tcmu_cmd size to save cmd area memories
  tcmu: Add global data block pool support
  tcmu: Add dynamic growing data area feature support
  target: fixup error message in target_tg_pt_gp_tg_pt_gp_id_store()
  target: fixup error message in target_tg_pt_gp_alua_access_type_store()
  target/user: PGR Support
  target: Add WRITE_VERIFY_16
  Documentation/target: add an example script to configure an iSCSI target
  target: Use kmalloc_array() in transport_kmap_data_sg()
  target: Use kmalloc_array() in compare_and_write_callback()
  target: Improve size determinations in two functions
  ...

Documentation/target/target-export-device

+#!/bin/sh
+#
+# This script illustrates the sequence of operations in configfs to
+# create a very simple LIO iSCSI target with a file or block device
+# backstore.
+#
+# (C) Copyright 2014 Christophe Vu-Brugier <cvubrugier@fastmail.fm>
+#
+
+print_usage() {
+    cat <<EOF
+Usage: $(basename $0) [-p PORTAL] DEVICE|FILE
+Export a block device or a file as an iSCSI target with a single LUN
+EOF
+}
+
+die() {
+    echo $1
+    exit 1
+}
+
+while getopts "hp:" arg; do
+    case $arg in
+        h) print_usage; exit 0;;
+        p) PORTAL=${OPTARG};;
+    esac
+done
+shift $(($OPTIND - 1))
+
+DEVICE=$1
+[ -n "$DEVICE" ] || die "Missing device or file argument"
+[ -b $DEVICE -o -f $DEVICE ] || die "Invalid device or file: ${DEVICE}"
+IQN="iqn.2003-01.org.linux-iscsi.$(hostname):$(basename $DEVICE)"
+[ -n "$PORTAL" ] || PORTAL="0.0.0.0:3260"
+
+CONFIGFS=/sys/kernel/config
+CORE_DIR=$CONFIGFS/target/core
+ISCSI_DIR=$CONFIGFS/target/iscsi
+
+# Load the target modules and mount the config file system
+lsmod | grep -q configfs || modprobe configfs
+lsmod | grep -q target_core_mod || modprobe target_core_mod
+mount | grep -q ^configfs || mount -t configfs none $CONFIGFS
+mkdir -p $ISCSI_DIR
+
+# Create a backstore
+if [ -b $DEVICE ]; then
+    BACKSTORE_DIR=$CORE_DIR/iblock_0/data
+    mkdir -p $BACKSTORE_DIR
+    echo "udev_path=${DEVICE}" > $BACKSTORE_DIR/control
+else
+    BACKSTORE_DIR=$CORE_DIR/fileio_0/data
+    mkdir -p $BACKSTORE_DIR
+    DEVICE_SIZE=$(du -b $DEVICE | cut -f1)
+    echo "fd_dev_name=${DEVICE}" > $BACKSTORE_DIR/control
+    echo "fd_dev_size=${DEVICE_SIZE}" > $BACKSTORE_DIR/control
+    echo 1 > $BACKSTORE_DIR/attrib/emulate_write_cache
+fi
+echo 1 > $BACKSTORE_DIR/enable
+
+# Create an iSCSI target and a target portal group (TPG)
+mkdir $ISCSI_DIR/$IQN
+mkdir $ISCSI_DIR/$IQN/tpgt_1/
+
+# Create a LUN
+mkdir $ISCSI_DIR/$IQN/tpgt_1/lun/lun_0
+ln -s $BACKSTORE_DIR $ISCSI_DIR/$IQN/tpgt_1/lun/lun_0/data
+echo 1 > $ISCSI_DIR/$IQN/tpgt_1/enable
+
+# Create a network portal
+mkdir $ISCSI_DIR/$IQN/tpgt_1/np/$PORTAL
+
+# Disable authentication
+echo 0 > $ISCSI_DIR/$IQN/tpgt_1/attrib/authentication
+echo 1 > $ISCSI_DIR/$IQN/tpgt_1/attrib/generate_node_acls
+
+# Allow write access for non authenticated initiators
+echo 0 > $ISCSI_DIR/$IQN/tpgt_1/attrib/demo_mode_write_protect
+
+echo "Target ${IQN}, portal ${PORTAL} has been created"

drivers/infiniband/ulp/srpt/ib_srpt.c

@@ -2302,12 +2302,8 @@ static void srpt_queue_response(struct se_cmd *cmd)
 	}
 	spin_unlock_irqrestore(&ioctx->spinlock, flags);
 
-	if (unlikely(transport_check_aborted_status(&ioctx->cmd, false)
-		     || WARN_ON_ONCE(state == SRPT_STATE_CMD_RSP_SENT))) {
-		atomic_inc(&ch->req_lim_delta);
-		srpt_abort_cmd(ioctx);
+	if (unlikely(WARN_ON_ONCE(state == SRPT_STATE_CMD_RSP_SENT)))
 		return;
-	}
 
 	/* For read commands, transfer the data to the initiator. */
 	if (ioctx->cmd.data_direction == DMA_FROM_DEVICE &&
@@ -2689,7 +2685,8 @@ static void srpt_release_cmd(struct se_cmd *se_cmd)
 	struct srpt_rdma_ch *ch = ioctx->ch;
 	unsigned long flags;
 
-	WARN_ON(ioctx->state != SRPT_STATE_DONE);
+	WARN_ON_ONCE(ioctx->state != SRPT_STATE_DONE &&
+		     !(ioctx->cmd.transport_state & CMD_T_ABORTED));
 
 	if (ioctx->n_rw_ctx) {
 		srpt_free_rw_ctxs(ch, ioctx);

drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c

@@ -1170,6 +1170,7 @@ static struct ibmvscsis_cmd *ibmvscsis_get_free_cmd(struct scsi_info *vscsi)
 		cmd = list_first_entry_or_null(&vscsi->free_cmd,
 					       struct ibmvscsis_cmd, list);
 		if (cmd) {
+			cmd->flags &= ~(DELAY_SEND);
 			list_del(&cmd->list);
 			cmd->iue = iue;
 			cmd->type = UNSET_TYPE;
@@ -1749,16 +1750,45 @@ static void srp_snd_msg_failed(struct scsi_info *vscsi, long rc)
 static void ibmvscsis_send_messages(struct scsi_info *vscsi)
 {
 	u64 msg_hi = 0;
-	/* note do not attmempt to access the IU_data_ptr with this pointer
+	/* note do not attempt to access the IU_data_ptr with this pointer
 	 * it is not valid
 	 */
 	struct viosrp_crq *crq = (struct viosrp_crq *)&msg_hi;
 	struct ibmvscsis_cmd *cmd, *nxt;
 	struct iu_entry *iue;
 	long rc = ADAPT_SUCCESS;
+	bool retry = false;
 
 	if (!(vscsi->flags & RESPONSE_Q_DOWN)) {
-		list_for_each_entry_safe(cmd, nxt, &vscsi->waiting_rsp, list) {
+		do {
+			retry = false;
+			list_for_each_entry_safe(cmd, nxt, &vscsi->waiting_rsp,
+						 list) {
+				/*
+				 * Check to make sure abort cmd gets processed
+				 * prior to the abort tmr cmd
+				 */
+				if (cmd->flags & DELAY_SEND)
+					continue;
+
+				if (cmd->abort_cmd) {
+					retry = true;
+					cmd->abort_cmd->flags &= ~(DELAY_SEND);
+				}
+
+				/*
+				 * If CMD_T_ABORTED w/o CMD_T_TAS scenarios and
+				 * the case where LIO issued a
+				 * ABORT_TASK: Sending TMR_TASK_DOES_NOT_EXIST
+				 * case then we dont send a response, since it
+				 * was already done.
+				 */
+				if (cmd->se_cmd.transport_state & CMD_T_ABORTED &&
+				    !(cmd->se_cmd.transport_state & CMD_T_TAS)) {
+					list_del(&cmd->list);
+					ibmvscsis_free_cmd_resources(vscsi,
+								     cmd);
+				} else {
 					iue = cmd->iue;
 
 					crq->valid = VALID_CMD_RESP_EL;
@@ -1776,18 +1806,23 @@ static void ibmvscsis_send_messages(struct scsi_info *vscsi)
 					pr_debug("send_messages: cmd %p, tag 0x%llx, rc %ld\n",
 						 cmd, be64_to_cpu(cmd->rsp.tag), rc);
 
-			/* if all ok free up the command element resources */
+					/* if all ok free up the command
+					 * element resources
+					 */
 					if (rc == H_SUCCESS) {
 						/* some movement has occurred */
 						vscsi->rsp_q_timer.timer_pops = 0;
 						list_del(&cmd->list);
 
-				ibmvscsis_free_cmd_resources(vscsi, cmd);
+						ibmvscsis_free_cmd_resources(vscsi,
+									     cmd);
 					} else {
 						srp_snd_msg_failed(vscsi, rc);
 						break;
 					}
 				}
+			}
+		} while (retry);
 
 		if (!rc) {
 			/*
@@ -2708,6 +2743,7 @@ static int ibmvscsis_alloc_cmds(struct scsi_info *vscsi, int num)
 
 	for (i = 0, cmd = (struct ibmvscsis_cmd *)vscsi->cmd_pool; i < num;
 	     i++, cmd++) {
+		cmd->abort_cmd = NULL;
 		cmd->adapter = vscsi;
 		INIT_WORK(&cmd->work, ibmvscsis_scheduler);
 		list_add_tail(&cmd->list, &vscsi->free_cmd);
@@ -3579,9 +3615,20 @@ static int ibmvscsis_write_pending(struct se_cmd *se_cmd)
 {
 	struct ibmvscsis_cmd *cmd = container_of(se_cmd, struct ibmvscsis_cmd,
 						 se_cmd);
+	struct scsi_info *vscsi = cmd->adapter;
 	struct iu_entry *iue = cmd->iue;
 	int rc;
 
+	/*
+	 * If CLIENT_FAILED OR RESPONSE_Q_DOWN, then just return success
+	 * since LIO can't do anything about it, and we dont want to
+	 * attempt an srp_transfer_data.
+	 */
+	if ((vscsi->flags & (CLIENT_FAILED | RESPONSE_Q_DOWN))) {
+		pr_err("write_pending failed since: %d\n", vscsi->flags);
+		return 0;
+	}
+
 	rc = srp_transfer_data(cmd, &vio_iu(iue)->srp.cmd, ibmvscsis_rdma,
 			       1, 1);
 	if (rc) {
@@ -3660,11 +3707,28 @@ static void ibmvscsis_queue_tm_rsp(struct se_cmd *se_cmd)
 	struct ibmvscsis_cmd *cmd = container_of(se_cmd, struct ibmvscsis_cmd,
 						 se_cmd);
 	struct scsi_info *vscsi = cmd->adapter;
+	struct ibmvscsis_cmd *cmd_itr;
+	struct iu_entry *iue = iue = cmd->iue;
+	struct srp_tsk_mgmt *srp_tsk = &vio_iu(iue)->srp.tsk_mgmt;
+	u64 tag_to_abort = be64_to_cpu(srp_tsk->task_tag);
 	uint len;
 
 	pr_debug("queue_tm_rsp %p, status %d\n",
 		 se_cmd, (int)se_cmd->se_tmr_req->response);
 
+	if (srp_tsk->tsk_mgmt_func == SRP_TSK_ABORT_TASK &&
+	    cmd->se_cmd.se_tmr_req->response == TMR_TASK_DOES_NOT_EXIST) {
+		spin_lock_bh(&vscsi->intr_lock);
+		list_for_each_entry(cmd_itr, &vscsi->active_q, list) {
+			if (tag_to_abort == cmd_itr->se_cmd.tag) {
+				cmd_itr->abort_cmd = cmd;
+				cmd->flags |= DELAY_SEND;
+				break;
+			}
+		}
+		spin_unlock_bh(&vscsi->intr_lock);
+	}
+
 	srp_build_response(vscsi, cmd, &len);
 	cmd->rsp.format = SRP_FORMAT;
 	cmd->rsp.len = len;
@@ -3672,8 +3736,8 @@ static void ibmvscsis_queue_tm_rsp(struct se_cmd *se_cmd)
 
 static void ibmvscsis_aborted_task(struct se_cmd *se_cmd)
 {
-	/* TBD: What (if anything) should we do here? */
-	pr_debug("ibmvscsis_aborted_task %p\n", se_cmd);
+	pr_debug("ibmvscsis_aborted_task %p task_tag: %llu\n",
+		 se_cmd, se_cmd->tag);
 }
 
 static struct se_wwn *ibmvscsis_make_tport(struct target_fabric_configfs *tf,

drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h

@@ -168,10 +168,12 @@ struct ibmvscsis_cmd {
 	struct iu_rsp rsp;
 	struct work_struct work;
 	struct scsi_info *adapter;
+	struct ibmvscsis_cmd *abort_cmd;
 	/* Sense buffer that will be mapped into outgoing status */
 	unsigned char sense_buf[TRANSPORT_SENSE_BUFFER];
 	u64 init_time;
 #define CMD_FAST_FAIL	BIT(0)
+#define DELAY_SEND	BIT(1)
 	u32 flags;
 	char type;
 };

drivers/target/iscsi/iscsi_target.c

@@ -128,11 +128,9 @@ struct iscsi_tiqn *iscsit_add_tiqn(unsigned char *buf)
 		return ERR_PTR(-EINVAL);
 	}
 
-	tiqn = kzalloc(sizeof(struct iscsi_tiqn), GFP_KERNEL);
-	if (!tiqn) {
-		pr_err("Unable to allocate struct iscsi_tiqn\n");
+	tiqn = kzalloc(sizeof(*tiqn), GFP_KERNEL);
+	if (!tiqn)
 		return ERR_PTR(-ENOMEM);
-	}
 
 	sprintf(tiqn->tiqn, "%s", buf);
 	INIT_LIST_HEAD(&tiqn->tiqn_list);
@@ -362,9 +360,8 @@ struct iscsi_np *iscsit_add_np(
 		return np;
 	}
 
-	np = kzalloc(sizeof(struct iscsi_np), GFP_KERNEL);
+	np = kzalloc(sizeof(*np), GFP_KERNEL);
 	if (!np) {
-		pr_err("Unable to allocate memory for struct iscsi_np\n");
 		mutex_unlock(&np_lock);
 		return ERR_PTR(-ENOMEM);
 	}
@@ -696,12 +693,10 @@ static int __init iscsi_target_init_module(void)
 	int ret = 0, size;
 
 	pr_debug("iSCSI-Target "ISCSIT_VERSION"\n");
-
-	iscsit_global = kzalloc(sizeof(struct iscsit_global), GFP_KERNEL);
-	if (!iscsit_global) {
-		pr_err("Unable to allocate memory for iscsit_global\n");
+	iscsit_global = kzalloc(sizeof(*iscsit_global), GFP_KERNEL);
+	if (!iscsit_global)
 		return -1;
-	}
+
 	spin_lock_init(&iscsit_global->ts_bitmap_lock);
 	mutex_init(&auth_id_lock);
 	spin_lock_init(&sess_idr_lock);
@@ -714,10 +709,8 @@ static int __init iscsi_target_init_module(void)
 
 	size = BITS_TO_LONGS(ISCSIT_BITMAP_BITS) * sizeof(long);
 	iscsit_global->ts_bitmap = vzalloc(size);
-	if (!iscsit_global->ts_bitmap) {
-		pr_err("Unable to allocate iscsit_global->ts_bitmap\n");
+	if (!iscsit_global->ts_bitmap)
 		goto configfs_out;
-	}
 
 	lio_qr_cache = kmem_cache_create("lio_qr_cache",
 			sizeof(struct iscsi_queue_req),
@@ -984,12 +977,9 @@ static int iscsit_allocate_iovecs(struct iscsi_cmd *cmd)
 	u32 iov_count = max(1UL, DIV_ROUND_UP(cmd->se_cmd.data_length, PAGE_SIZE));
 
 	iov_count += ISCSI_IOV_DATA_BUFFER;
-
-	cmd->iov_data = kzalloc(iov_count * sizeof(struct kvec), GFP_KERNEL);
-	if (!cmd->iov_data) {
-		pr_err("Unable to allocate cmd->iov_data\n");
+	cmd->iov_data = kcalloc(iov_count, sizeof(*cmd->iov_data), GFP_KERNEL);
+	if (!cmd->iov_data)
 		return -ENOMEM;
-	}
 
 	cmd->orig_iov_data_count = iov_count;
 	return 0;
@@ -1850,8 +1840,6 @@ static int iscsit_handle_nop_out(struct iscsi_conn *conn, struct iscsi_cmd *cmd,
 
 		ping_data = kzalloc(payload_length + 1, GFP_KERNEL);
 		if (!ping_data) {
-			pr_err("Unable to allocate memory for"
-				" NOPOUT ping data.\n");
 			ret = -1;
 			goto out;
 		}
@@ -1997,15 +1985,11 @@ iscsit_handle_task_mgt_cmd(struct iscsi_conn *conn, struct iscsi_cmd *cmd,
 		hdr->refcmdsn = cpu_to_be32(ISCSI_RESERVED_TAG);
 
 	cmd->data_direction = DMA_NONE;
-
-	cmd->tmr_req = kzalloc(sizeof(struct iscsi_tmr_req), GFP_KERNEL);
-	if (!cmd->tmr_req) {
-		pr_err("Unable to allocate memory for"
-			" Task Management command!\n");
+	cmd->tmr_req = kzalloc(sizeof(*cmd->tmr_req), GFP_KERNEL);
+	if (!cmd->tmr_req)
 		return iscsit_add_reject_cmd(cmd,
 					     ISCSI_REASON_BOOKMARK_NO_RESOURCES,
 					     buf);
-	}
 
 	/*
 	 * TASK_REASSIGN for ERL=2 / connection stays inside of
@@ -2265,11 +2249,9 @@ iscsit_handle_text_cmd(struct iscsi_conn *conn, struct iscsi_cmd *cmd,
 		struct kvec iov[3];
 
 		text_in = kzalloc(payload_length, GFP_KERNEL);
-		if (!text_in) {
-			pr_err("Unable to allocate memory for"
-				" incoming text parameters\n");
+		if (!text_in)
 			goto reject;
-		}
+
 		cmd->text_in_ptr = text_in;
 
 		memset(iov, 0, 3 * sizeof(struct kvec));
@@ -3353,11 +3335,9 @@ iscsit_build_sendtargets_response(struct iscsi_cmd *cmd,
 			 SENDTARGETS_BUF_LIMIT);
 
 	payload = kzalloc(buffer_len, GFP_KERNEL);
-	if (!payload) {
-		pr_err("Unable to allocate memory for sendtargets"
-				" response.\n");
+	if (!payload)
 		return -ENOMEM;
-	}
+
 	/*
 	 * Locate pointer to iqn./eui. string for ICF_SENDTARGETS_SINGLE
 	 * explicit case..
@@ -4683,6 +4663,7 @@ int iscsit_release_sessions_for_tpg(struct iscsi_portal_group *tpg, int force)
 			continue;
 		}
 		atomic_set(&sess->session_reinstatement, 1);
+		atomic_set(&sess->session_fall_back_to_erl0, 1);
 		spin_unlock(&sess->conn_lock);
 
 		list_move_tail(&se_sess->sess_list, &free_list);

drivers/target/iscsi/iscsi_target_configfs.c

@@ -1506,6 +1506,7 @@ static void lio_tpg_close_session(struct se_session *se_sess)
 		return;
 	}
 	atomic_set(&sess->session_reinstatement, 1);
+	atomic_set(&sess->session_fall_back_to_erl0, 1);
 	spin_unlock(&sess->conn_lock);
 
 	iscsit_stop_time2retain_timer(sess);

drivers/target/iscsi/iscsi_target_login.c

@@ -208,6 +208,7 @@ int iscsi_check_for_session_reinstatement(struct iscsi_conn *conn)
 			    initiatorname_param->value) &&
 		   (sess_p->sess_ops->SessionType == sessiontype))) {
 			atomic_set(&sess_p->session_reinstatement, 1);
+			atomic_set(&sess_p->session_fall_back_to_erl0, 1);
 			spin_unlock(&sess_p->conn_lock);
 			iscsit_inc_session_usage_count(sess_p);
 			iscsit_stop_time2retain_timer(sess_p);

drivers/target/target_core_configfs.c

@@ -533,6 +533,7 @@ DEF_CONFIGFS_ATTRIB_SHOW(emulate_3pc);
 DEF_CONFIGFS_ATTRIB_SHOW(pi_prot_type);
 DEF_CONFIGFS_ATTRIB_SHOW(hw_pi_prot_type);
 DEF_CONFIGFS_ATTRIB_SHOW(pi_prot_format);
+DEF_CONFIGFS_ATTRIB_SHOW(pi_prot_verify);
 DEF_CONFIGFS_ATTRIB_SHOW(enforce_pr_isids);
 DEF_CONFIGFS_ATTRIB_SHOW(is_nonrot);
 DEF_CONFIGFS_ATTRIB_SHOW(emulate_rest_reord);
@@ -823,6 +824,7 @@ static ssize_t pi_prot_type_store(struct config_item *item,
 		ret = dev->transport->init_prot(dev);
 		if (ret) {
 			da->pi_prot_type = old_prot;
+			da->pi_prot_verify = (bool) da->pi_prot_type;
 			return ret;
 		}
 
@@ -830,6 +832,7 @@ static ssize_t pi_prot_type_store(struct config_item *item,
 		dev->transport->free_prot(dev);
 	}
 
+	da->pi_prot_verify = (bool) da->pi_prot_type;
 	pr_debug("dev[%p]: SE Device Protection Type: %d\n", dev, flag);
 	return count;
 }
@@ -872,6 +875,35 @@ static ssize_t pi_prot_format_store(struct config_item *item,
 	return count;
 }
 
+static ssize_t pi_prot_verify_store(struct config_item *item,
+		const char *page, size_t count)
+{
+	struct se_dev_attrib *da = to_attrib(item);
+	bool flag;
+	int ret;
+
+	ret = strtobool(page, &flag);
+	if (ret < 0)
+		return ret;
+
+	if (!flag) {
+		da->pi_prot_verify = flag;
+		return count;
+	}
+	if (da->hw_pi_prot_type) {
+		pr_warn("DIF protection enabled on underlying hardware,"
+			" ignoring\n");
+		return count;
+	}
+	if (!da->pi_prot_type) {
+		pr_warn("DIF protection not supported by backend, ignoring\n");
+		return count;
+	}
+	da->pi_prot_verify = flag;
+
+	return count;
+}
+
 static ssize_t force_pr_aptpl_store(struct config_item *item,
 		const char *page, size_t count)
 {
@@ -1067,6 +1099,7 @@ CONFIGFS_ATTR(, emulate_3pc);
 CONFIGFS_ATTR(, pi_prot_type);
 CONFIGFS_ATTR_RO(, hw_pi_prot_type);
 CONFIGFS_ATTR(, pi_prot_format);
+CONFIGFS_ATTR(, pi_prot_verify);
 CONFIGFS_ATTR(, enforce_pr_isids);
 CONFIGFS_ATTR(, is_nonrot);
 CONFIGFS_ATTR(, emulate_rest_reord);
@@ -1104,6 +1137,7 @@ struct configfs_attribute *sbc_attrib_attrs[] = {
 	&attr_pi_prot_type,
 	&attr_hw_pi_prot_type,
 	&attr_pi_prot_format,
+	&attr_pi_prot_verify,
 	&attr_enforce_pr_isids,
 	&attr_is_nonrot,
 	&attr_emulate_rest_reord,
@@ -1366,7 +1400,7 @@ static ssize_t target_pr_res_holder_show(struct config_item *item, char *page)
 	struct se_device *dev = pr_to_dev(item);
 	int ret;
 
-	if (dev->transport->transport_flags & TRANSPORT_FLAG_PASSTHROUGH)
+	if (dev->transport->transport_flags & TRANSPORT_FLAG_PASSTHROUGH_PGR)
 		return sprintf(page, "Passthrough\n");
 
 	spin_lock(&dev->dev_reservation_lock);
@@ -1506,7 +1540,7 @@ static ssize_t target_pr_res_type_show(struct config_item *item, char *page)
 {
 	struct se_device *dev = pr_to_dev(item);
 
-	if (dev->transport->transport_flags & TRANSPORT_FLAG_PASSTHROUGH)
+	if (dev->transport->transport_flags & TRANSPORT_FLAG_PASSTHROUGH_PGR)
 		return sprintf(page, "SPC_PASSTHROUGH\n");
 	else if (dev->dev_reservation_flags & DRF_SPC2_RESERVATIONS)
 		return sprintf(page, "SPC2_RESERVATIONS\n");
@@ -1519,7 +1553,7 @@ static ssize_t target_pr_res_aptpl_active_show(struct config_item *item,
 {
 	struct se_device *dev = pr_to_dev(item);
 
-	if (dev->transport->transport_flags & TRANSPORT_FLAG_PASSTHROUGH)
+	if (dev->transport->transport_flags & TRANSPORT_FLAG_PASSTHROUGH_PGR)
 		return 0;
 
 	return sprintf(page, "APTPL Bit Status: %s\n",
@@ -1531,7 +1565,7 @@ static ssize_t target_pr_res_aptpl_metadata_show(struct config_item *item,
 {
 	struct se_device *dev = pr_to_dev(item);
 
-	if (dev->transport->transport_flags & TRANSPORT_FLAG_PASSTHROUGH)
+	if (dev->transport->transport_flags & TRANSPORT_FLAG_PASSTHROUGH_PGR)
 		return 0;
 
 	return sprintf(page, "Ready to process PR APTPL metadata..\n");
@@ -1577,7 +1611,7 @@ static ssize_t target_pr_res_aptpl_metadata_store(struct config_item *item,
 	u16 tpgt = 0;
 	u8 type = 0;
 
-	if (dev->transport->transport_flags & TRANSPORT_FLAG_PASSTHROUGH)
+	if (dev->transport->transport_flags & TRANSPORT_FLAG_PASSTHROUGH_PGR)
 		return count;
 	if (dev->dev_reservation_flags & DRF_SPC2_RESERVATIONS)
 		return count;
@@ -2511,7 +2545,7 @@ static ssize_t target_tg_pt_gp_alua_support_##_name##_store(		\
 	int ret;							\
 									\
 	if (!t->tg_pt_gp_valid_id) {					\
-		pr_err("Unable to do set ##_name ALUA state on non"	\
+		pr_err("Unable to do set " #_name " ALUA state on non"	\
 		       " valid tg_pt_gp ID: %hu\n",			\
 		       t->tg_pt_gp_valid_id);				\
 		return -EINVAL;						\
@@ -2643,13 +2677,13 @@ static ssize_t target_tg_pt_gp_tg_pt_gp_id_store(struct config_item *item,
 
 	ret = kstrtoul(page, 0, &tg_pt_gp_id);
 	if (ret < 0) {
-		pr_err("kstrtoul() returned %d for"
-			" tg_pt_gp_id\n", ret);
+		pr_err("ALUA tg_pt_gp_id: invalid value '%s' for tg_pt_gp_id\n",
+		       page);
 		return ret;
 	}
 	if (tg_pt_gp_id > 0x0000ffff) {
-		pr_err("ALUA tg_pt_gp_id: %lu exceeds maximum:"
-			" 0x0000ffff\n", tg_pt_gp_id);
+		pr_err("ALUA tg_pt_gp_id: %lu exceeds maximum: 0x0000ffff\n",
+		       tg_pt_gp_id);
 		return -EINVAL;
 	}
 

drivers/target/target_core_device.c

@@ -1045,6 +1045,8 @@ passthrough_parse_cdb(struct se_cmd *cmd,
 	sense_reason_t (*exec_cmd)(struct se_cmd *cmd))
 {
 	unsigned char *cdb = cmd->t_task_cdb;
+	struct se_device *dev = cmd->se_dev;
+	unsigned int size;
 
 	/*
 	 * Clear a lun set in the cdb if the initiator talking to use spoke
@@ -1076,6 +1078,42 @@ passthrough_parse_cdb(struct se_cmd *cmd,
 		return TCM_NO_SENSE;
 	}
 
+	/*
+	 * For PERSISTENT RESERVE IN/OUT, RELEASE, and RESERVE we need to
+	 * emulate the response, since tcmu does not have the information
+	 * required to process these commands.
+	 */
+	if (!(dev->transport->transport_flags &
+	      TRANSPORT_FLAG_PASSTHROUGH_PGR)) {
+		if (cdb[0] == PERSISTENT_RESERVE_IN) {
+			cmd->execute_cmd = target_scsi3_emulate_pr_in;
+			size = (cdb[7] << 8) + cdb[8];
+			return target_cmd_size_check(cmd, size);
+		}
+		if (cdb[0] == PERSISTENT_RESERVE_OUT) {
+			cmd->execute_cmd = target_scsi3_emulate_pr_out;
+			size = (cdb[7] << 8) + cdb[8];
+			return target_cmd_size_check(cmd, size);
+		}
+
+		if (cdb[0] == RELEASE || cdb[0] == RELEASE_10) {
+			cmd->execute_cmd = target_scsi2_reservation_release;
+			if (cdb[0] == RELEASE_10)
+				size = (cdb[7] << 8) | cdb[8];
+			else
+				size = cmd->data_length;
+			return target_cmd_size_check(cmd, size);
+		}
+		if (cdb[0] == RESERVE || cdb[0] == RESERVE_10) {
+			cmd->execute_cmd = target_scsi2_reservation_reserve;
+			if (cdb[0] == RESERVE_10)
+				size = (cdb[7] << 8) | cdb[8];
+			else
+				size = cmd->data_length;
+			return target_cmd_size_check(cmd, size);
+		}
+	}
+
 	/* Set DATA_CDB flag for ops that should have it */
 	switch (cdb[0]) {
 	case READ_6:

drivers/target/target_core_file.c

@@ -277,12 +277,11 @@ static int fd_do_rw(struct se_cmd *cmd, struct file *fd,
 	else
 		ret = vfs_iter_read(fd, &iter, &pos);
 
-	kfree(bvec);
-
 	if (is_write) {
 		if (ret < 0 || ret != data_length) {
 			pr_err("%s() write returned %d\n", __func__, ret);
-			return (ret < 0 ? ret : -EINVAL);
+			if (ret >= 0)
+				ret = -EINVAL;
 		}
 	} else {
 		/*
@@ -295,17 +294,29 @@ static int fd_do_rw(struct se_cmd *cmd, struct file *fd,
 				pr_err("%s() returned %d, expecting %u for "
 						"S_ISBLK\n", __func__, ret,
 						data_length);
-				return (ret < 0 ? ret : -EINVAL);
+				if (ret >= 0)
+					ret = -EINVAL;
 			}
 		} else {
 			if (ret < 0) {
 				pr_err("%s() returned %d for non S_ISBLK\n",
 						__func__, ret);
-				return ret;
+			} else if (ret != data_length) {
+				/*
+				 * Short read case:
+				 * Probably some one truncate file under us.
+				 * We must explicitly zero sg-pages to prevent
+				 * expose uninizialized pages to userspace.
+				 */
+				if (ret < data_length)
+					ret += iov_iter_zero(data_length - ret, &iter);
+				else
+					ret = -EINVAL;
 			}
 		}
 	}
-	return 1;
+	kfree(bvec);
+	return ret;
 }
 
 static sense_reason_t
@@ -543,7 +554,8 @@ fd_execute_rw(struct se_cmd *cmd, struct scatterlist *sgl, u32 sgl_nents,
 		ret = fd_do_rw(cmd, file, dev->dev_attrib.block_size,
 			       sgl, sgl_nents, cmd->data_length, 0);
 
-		if (ret > 0 && cmd->prot_type && dev->dev_attrib.pi_prot_type) {
+		if (ret > 0 && cmd->prot_type && dev->dev_attrib.pi_prot_type &&
+		    dev->dev_attrib.pi_prot_verify) {
 			u32 sectors = cmd->data_length >>
 					ilog2(dev->dev_attrib.block_size);
 
@@ -553,7 +565,8 @@ fd_execute_rw(struct se_cmd *cmd, struct scatterlist *sgl, u32 sgl_nents,
 				return rc;
 		}
 	} else {
-		if (cmd->prot_type && dev->dev_attrib.pi_prot_type) {
+		if (cmd->prot_type && dev->dev_attrib.pi_prot_type &&
+		    dev->dev_attrib.pi_prot_verify) {
 			u32 sectors = cmd->data_length >>
 					ilog2(dev->dev_attrib.block_size);
 
@@ -595,7 +608,6 @@ fd_execute_rw(struct se_cmd *cmd, struct scatterlist *sgl, u32 sgl_nents,
 	if (ret < 0)
 		return TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
 
-	if (ret)
 	target_complete_cmd(cmd, SAM_STAT_GOOD);
 	return 0;
 }

drivers/target/target_core_iblock.c

@@ -279,7 +279,7 @@ static void iblock_complete_cmd(struct se_cmd *cmd)
 	struct iblock_req *ibr = cmd->priv;
 	u8 status;
 
-	if (!atomic_dec_and_test(&ibr->pending))
+	if (!refcount_dec_and_test(&ibr->pending))
 		return;
 
 	if (atomic_read(&ibr->ib_bio_err_cnt))
@@ -487,7 +487,7 @@ iblock_execute_write_same(struct se_cmd *cmd)
 	bio_list_init(&list);
 	bio_list_add(&list, bio);
 
-	atomic_set(&ibr->pending, 1);
+	refcount_set(&ibr->pending, 1);
 
 	while (sectors) {
 		while (bio_add_page(bio, sg_page(sg), sg->length, sg->offset)
@@ -498,7 +498,7 @@ iblock_execute_write_same(struct se_cmd *cmd)
 			if (!bio)
 				goto fail_put_bios;
 
-			atomic_inc(&ibr->pending);
+			refcount_inc(&ibr->pending);
 			bio_list_add(&list, bio);
 		}
 
@@ -706,7 +706,7 @@ iblock_execute_rw(struct se_cmd *cmd, struct scatterlist *sgl, u32 sgl_nents,
 	cmd->priv = ibr;
 
 	if (!sgl_nents) {
-		atomic_set(&ibr->pending, 1);
+		refcount_set(&ibr->pending, 1);
 		iblock_complete_cmd(cmd);
 		return 0;
 	}
@@ -719,7 +719,7 @@ iblock_execute_rw(struct se_cmd *cmd, struct scatterlist *sgl, u32 sgl_nents,
 	bio_list_init(&list);
 	bio_list_add(&list, bio);
 
-	atomic_set(&ibr->pending, 2);
+	refcount_set(&ibr->pending, 2);
 	bio_cnt = 1;
 
 	for_each_sg(sgl, sg, sgl_nents, i) {
@@ -740,7 +740,7 @@ iblock_execute_rw(struct se_cmd *cmd, struct scatterlist *sgl, u32 sgl_nents,
 			if (!bio)
 				goto fail_put_bios;
 
-			atomic_inc(&ibr->pending);
+			refcount_inc(&ibr->pending);
 			bio_list_add(&list, bio);
 			bio_cnt++;
 		}

drivers/target/target_core_iblock.h

@@ -2,6 +2,7 @@
 #define TARGET_CORE_IBLOCK_H
 
 #include <linux/atomic.h>
+#include <linux/refcount.h>
 #include <target/target_core_base.h>
 
 #define IBLOCK_VERSION		"4.0"
@@ -10,7 +11,7 @@
 #define IBLOCK_LBA_SHIFT	9
 
 struct iblock_req {
-	atomic_t pending;
+	refcount_t pending;
 	atomic_t ib_bio_err_cnt;
 } ____cacheline_aligned;
 

drivers/target/target_core_pr.c

@@ -4147,7 +4147,7 @@ target_check_reservation(struct se_cmd *cmd)
 		return 0;
 	if (dev->se_hba->hba_flags & HBA_FLAGS_INTERNAL_USE)
 		return 0;
-	if (dev->transport->transport_flags & TRANSPORT_FLAG_PASSTHROUGH)
+	if (dev->transport->transport_flags & TRANSPORT_FLAG_PASSTHROUGH_PGR)
 		return 0;
 
 	spin_lock(&dev->dev_reservation_lock);

drivers/target/target_core_pr.h

@@ -7,7 +7,7 @@
 /*
  * PERSISTENT_RESERVE_OUT service action codes
  *
- * spc4r17 section 6.14.2 Table 171
+ * spc5r04b section 6.15.2 Table 174
  */
 #define PRO_REGISTER				0x00
 #define PRO_RESERVE				0x01
@@ -17,10 +17,11 @@
 #define PRO_PREEMPT_AND_ABORT			0x05
 #define PRO_REGISTER_AND_IGNORE_EXISTING_KEY	0x06
 #define PRO_REGISTER_AND_MOVE			0x07
+#define PRO_REPLACE_LOST_RESERVATION	0x08
 /*
  * PERSISTENT_RESERVE_IN service action codes
  *
- * spc4r17 section 6.13.1 Table 159
+ * spc5r04b section 6.14.1 Table 162
  */
 #define PRI_READ_KEYS				0x00
 #define PRI_READ_RESERVATION			0x01
@@ -29,13 +30,13 @@
 /*
  * PERSISTENT_RESERVE_ SCOPE field
  *
- * spc4r17 section 6.13.3.3 Table 163
+ * spc5r04b section 6.14.3.2 Table 166
  */
 #define PR_SCOPE_LU_SCOPE			0x00
 /*
  * PERSISTENT_RESERVE_* TYPE field
  *
- * spc4r17 section 6.13.3.4 Table 164
+ * spc5r04b section 6.14.3.3 Table 167
  */
 #define PR_TYPE_WRITE_EXCLUSIVE			0x01
 #define PR_TYPE_EXCLUSIVE_ACCESS		0x03

drivers/target/target_core_pscsi.c

@@ -1081,7 +1081,8 @@ static const struct target_backend_ops pscsi_ops = {
 	.name			= "pscsi",
 	.owner			= THIS_MODULE,
 	.transport_flags	= TRANSPORT_FLAG_PASSTHROUGH |
-				  TRANSPORT_FLAG_PASSTHROUGH_ALUA,
+				  TRANSPORT_FLAG_PASSTHROUGH_ALUA |
+				  TRANSPORT_FLAG_PASSTHROUGH_PGR,
 	.attach_hba		= pscsi_attach_hba,
 	.detach_hba		= pscsi_detach_hba,
 	.pmode_enable_hba	= pscsi_pmode_enable_hba,

drivers/target/target_core_rd.c

@@ -47,11 +47,9 @@ static int rd_attach_hba(struct se_hba *hba, u32 host_id)
 {
 	struct rd_host *rd_host;
 
-	rd_host = kzalloc(sizeof(struct rd_host), GFP_KERNEL);
-	if (!rd_host) {
-		pr_err("Unable to allocate memory for struct rd_host\n");
+	rd_host = kzalloc(sizeof(*rd_host), GFP_KERNEL);
+	if (!rd_host)
 		return -ENOMEM;
-	}
 
 	rd_host->rd_host_id = host_id;
 
@@ -148,11 +146,8 @@ static int rd_allocate_sgl_table(struct rd_dev *rd_dev, struct rd_dev_sg_table *
 
 		sg = kcalloc(sg_per_table + chain_entry, sizeof(*sg),
 				GFP_KERNEL);
-		if (!sg) {
-			pr_err("Unable to allocate scatterlist array"
-				" for struct rd_dev\n");
+		if (!sg)
 			return -ENOMEM;
-		}
 
 		sg_init_table(sg, sg_per_table + chain_entry);
 
@@ -210,13 +205,9 @@ static int rd_build_device_space(struct rd_dev *rd_dev)
 	total_sg_needed = rd_dev->rd_page_count;
 
 	sg_tables = (total_sg_needed / max_sg_per_table) + 1;
-
-	sg_table = kzalloc(sg_tables * sizeof(struct rd_dev_sg_table), GFP_KERNEL);
-	if (!sg_table) {
-		pr_err("Unable to allocate memory for Ramdisk"
-		       " scatterlist tables\n");
+	sg_table = kcalloc(sg_tables, sizeof(*sg_table), GFP_KERNEL);
+	if (!sg_table)
 		return -ENOMEM;
-	}
 
 	rd_dev->sg_table_array = sg_table;
 	rd_dev->sg_table_count = sg_tables;
@@ -271,13 +262,9 @@ static int rd_build_prot_space(struct rd_dev *rd_dev, int prot_length, int block
 	total_sg_needed = (rd_dev->rd_page_count * prot_length / block_size) + 1;
 
 	sg_tables = (total_sg_needed / max_sg_per_table) + 1;
-
-	sg_table = kzalloc(sg_tables * sizeof(struct rd_dev_sg_table), GFP_KERNEL);
-	if (!sg_table) {
-		pr_err("Unable to allocate memory for Ramdisk protection"
-		       " scatterlist tables\n");
+	sg_table = kcalloc(sg_tables, sizeof(*sg_table), GFP_KERNEL);
+	if (!sg_table)
 		return -ENOMEM;
-	}
 
 	rd_dev->sg_prot_array = sg_table;
 	rd_dev->sg_prot_count = sg_tables;
@@ -298,11 +285,9 @@ static struct se_device *rd_alloc_device(struct se_hba *hba, const char *name)
 	struct rd_dev *rd_dev;
 	struct rd_host *rd_host = hba->hba_ptr;
 
-	rd_dev = kzalloc(sizeof(struct rd_dev), GFP_KERNEL);
-	if (!rd_dev) {
-		pr_err("Unable to allocate memory for struct rd_dev\n");
+	rd_dev = kzalloc(sizeof(*rd_dev), GFP_KERNEL);
+	if (!rd_dev)
 		return NULL;
-	}
 
 	rd_dev->rd_host = rd_host;
 
@@ -410,7 +395,7 @@ static sense_reason_t rd_do_prot_rw(struct se_cmd *cmd, bool is_read)
 	u32 prot_offset, prot_page;
 	u32 prot_npages __maybe_unused;
 	u64 tmp;
-	sense_reason_t rc = TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
+	sense_reason_t rc = 0;
 
 	tmp = cmd->t_task_lba * se_dev->prot_length;
 	prot_offset = do_div(tmp, PAGE_SIZE);
@@ -423,13 +408,14 @@ static sense_reason_t rd_do_prot_rw(struct se_cmd *cmd, bool is_read)
 	prot_sg = &prot_table->sg_table[prot_page -
 					prot_table->page_start_offset];
 
+	if (se_dev->dev_attrib.pi_prot_verify) {
 		if (is_read)
 			rc = sbc_dif_verify(cmd, cmd->t_task_lba, sectors, 0,
 					    prot_sg, prot_offset);
 		else
 			rc = sbc_dif_verify(cmd, cmd->t_task_lba, sectors, 0,
 					    cmd->t_prot_sg, 0);
-
+	}
 	if (!rc)
 		sbc_dif_copy_prot(cmd, sectors, is_read, prot_sg, prot_offset);
 

drivers/target/target_core_sbc.c

@@ -507,8 +507,11 @@ static sense_reason_t compare_and_write_callback(struct se_cmd *cmd, bool succes
 	 * been failed with a non-zero SCSI status.
 	 */
 	if (cmd->scsi_status) {
-		pr_err("compare_and_write_callback: non zero scsi_status:"
+		pr_debug("compare_and_write_callback: non zero scsi_status:"
 			" 0x%02x\n", cmd->scsi_status);
+		*post_ret = 1;
+		if (cmd->scsi_status == SAM_STAT_CHECK_CONDITION)
+			ret = TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
 		goto out;
 	}
 
@@ -519,7 +522,7 @@ static sense_reason_t compare_and_write_callback(struct se_cmd *cmd, bool succes
 		goto out;
 	}
 
-	write_sg = kmalloc(sizeof(struct scatterlist) * cmd->t_data_nents,
+	write_sg = kmalloc_array(cmd->t_data_nents, sizeof(*write_sg),
 				 GFP_KERNEL);
 	if (!write_sg) {
 		pr_err("Unable to allocate compare_and_write sg\n");
@@ -924,6 +927,7 @@ sbc_parse_cdb(struct se_cmd *cmd, struct sbc_ops *ops)
 		cmd->execute_cmd = sbc_execute_rw;
 		break;
 	case WRITE_16:
+	case WRITE_VERIFY_16:
 		sectors = transport_get_sectors_16(cdb);
 		cmd->t_task_lba = transport_lba_64(cdb);
 

drivers/target/target_core_tpg.c

@@ -397,6 +397,13 @@ int core_tpg_set_initiator_node_queue_depth(
 {
 	struct se_portal_group *tpg = acl->se_tpg;
 
+	/*
+	 * Allow the setting of se_node_acl queue_depth to be idempotent,
+	 * and not force a session shutdown event if the value is not
+	 * changing.
+	 */
+	if (acl->queue_depth == queue_depth)
+		return 0;
 	/*
 	 * User has requested to change the queue depth for a Initiator Node.
 	 * Change the value in the Node's struct se_node_acl, and call

drivers/target/target_core_transport.c

@@ -2311,7 +2311,7 @@ void *transport_kmap_data_sg(struct se_cmd *cmd)
 		return kmap(sg_page(sg)) + sg->offset;
 
 	/* >1 page. use vmap */
-	pages = kmalloc(sizeof(*pages) * cmd->t_data_nents, GFP_KERNEL);
+	pages = kmalloc_array(cmd->t_data_nents, sizeof(*pages), GFP_KERNEL);
 	if (!pages)
 		return NULL;
 

drivers/target/target_core_user.c

@@ -2,6 +2,7 @@
  * Copyright (C) 2013 Shaohua Li <shli@kernel.org>
  * Copyright (C) 2014 Red Hat, Inc.
  * Copyright (C) 2015 Arrikto, Inc.
+ * Copyright (C) 2017 Chinamobile, Inc.
  *
  * This program is free software; you can redistribute it and/or modify it
  * under the terms and conditions of the GNU General Public License,
@@ -25,10 +26,13 @@
 #include <linux/parser.h>
 #include <linux/vmalloc.h>
 #include <linux/uio_driver.h>
+#include <linux/radix-tree.h>
 #include <linux/stringify.h>
 #include <linux/bitops.h>
 #include <linux/highmem.h>
 #include <linux/configfs.h>
+#include <linux/mutex.h>
+#include <linux/kthread.h>
 #include <net/genetlink.h>
 #include <scsi/scsi_common.h>
 #include <scsi/scsi_proto.h>
@@ -63,17 +67,26 @@
  * this may have a 'UAM' comment.
  */
 
-
 #define TCMU_TIME_OUT (30 * MSEC_PER_SEC)
 
-#define DATA_BLOCK_BITS 256
-#define DATA_BLOCK_SIZE 4096
+/* For cmd area, the size is fixed 8MB */
+#define CMDR_SIZE (8 * 1024 * 1024)
 
-#define CMDR_SIZE (16 * 4096)
+/*
+ * For data area, the block size is PAGE_SIZE and
+ * the total size is 256K * PAGE_SIZE.
+ */
+#define DATA_BLOCK_SIZE PAGE_SIZE
+#define DATA_BLOCK_BITS (256 * 1024)
 #define DATA_SIZE (DATA_BLOCK_BITS * DATA_BLOCK_SIZE)
+#define DATA_BLOCK_INIT_BITS 128
 
+/* The total size of the ring is 8M + 256K * PAGE_SIZE */
 #define TCMU_RING_SIZE (CMDR_SIZE + DATA_SIZE)
 
+/* Default maximum of the global data blocks(512K * PAGE_SIZE) */
+#define TCMU_GLOBAL_MAX_BLOCKS (512 * 1024)
+
 static struct device *tcmu_root_device;
 
 struct tcmu_hba {
@@ -83,6 +96,8 @@ struct tcmu_hba {
 #define TCMU_CONFIG_LEN 256
 
 struct tcmu_dev {
+	struct list_head node;
+
 	struct se_device se_dev;
 
 	char *name;
@@ -94,6 +109,8 @@ struct tcmu_dev {
 
 	struct uio_info uio_info;
 
+	struct inode *inode;
+
 	struct tcmu_mailbox *mb_addr;
 	size_t dev_size;
 	u32 cmdr_size;
@@ -103,11 +120,14 @@ struct tcmu_dev {
 	size_t data_off;
 	size_t data_size;
 
-	DECLARE_BITMAP(data_bitmap, DATA_BLOCK_BITS);
-
 	wait_queue_head_t wait_cmdr;
-	/* TODO should this be a mutex? */
-	spinlock_t cmdr_lock;
+	struct mutex cmdr_lock;
+
+	bool waiting_global;
+	uint32_t dbi_max;
+	uint32_t dbi_thresh;
+	DECLARE_BITMAP(data_bitmap, DATA_BLOCK_BITS);
+	struct radix_tree_root data_blocks;
 
 	struct idr commands;
 	spinlock_t commands_lock;
@@ -130,7 +150,9 @@ struct tcmu_cmd {
 
 	/* Can't use se_cmd when cleaning up expired cmds, because if
 	   cmd has been completed then accessing se_cmd is off limits */
-	DECLARE_BITMAP(data_bitmap, DATA_BLOCK_BITS);
+	uint32_t dbi_cnt;
+	uint32_t dbi_cur;
+	uint32_t *dbi;
 
 	unsigned long deadline;
 
@@ -138,6 +160,13 @@ struct tcmu_cmd {
 	unsigned long flags;
 };
 
+static struct task_struct *unmap_thread;
+static wait_queue_head_t unmap_wait;
+static DEFINE_MUTEX(root_udev_mutex);
+static LIST_HEAD(root_udev);
+
+static atomic_t global_db_count = ATOMIC_INIT(0);
+
 static struct kmem_cache *tcmu_cmd_cache;
 
 /* multicast group */
@@ -161,6 +190,114 @@ static struct genl_family tcmu_genl_family __ro_after_init = {
 	.netnsok = true,
 };
 
+#define tcmu_cmd_set_dbi_cur(cmd, index) ((cmd)->dbi_cur = (index))
+#define tcmu_cmd_reset_dbi_cur(cmd) tcmu_cmd_set_dbi_cur(cmd, 0)
+#define tcmu_cmd_set_dbi(cmd, index) ((cmd)->dbi[(cmd)->dbi_cur++] = (index))
+#define tcmu_cmd_get_dbi(cmd) ((cmd)->dbi[(cmd)->dbi_cur++])
+
+static void tcmu_cmd_free_data(struct tcmu_cmd *tcmu_cmd, uint32_t len)
+{
+	struct tcmu_dev *udev = tcmu_cmd->tcmu_dev;
+	uint32_t i;
+
+	for (i = 0; i < len; i++)
+		clear_bit(tcmu_cmd->dbi[i], udev->data_bitmap);
+}
+
+static inline bool tcmu_get_empty_block(struct tcmu_dev *udev,
+					struct tcmu_cmd *tcmu_cmd)
+{
+	struct page *page;
+	int ret, dbi;
+
+	dbi = find_first_zero_bit(udev->data_bitmap, udev->dbi_thresh);
+	if (dbi == udev->dbi_thresh)
+		return false;
+
+	page = radix_tree_lookup(&udev->data_blocks, dbi);
+	if (!page) {
+
+		if (atomic_add_return(1, &global_db_count) >
+					TCMU_GLOBAL_MAX_BLOCKS) {
+			atomic_dec(&global_db_count);
+			return false;
+		}
+
+		/* try to get new page from the mm */
+		page = alloc_page(GFP_KERNEL);
+		if (!page)
+			return false;
+
+		ret = radix_tree_insert(&udev->data_blocks, dbi, page);
+		if (ret) {
+			__free_page(page);
+			return false;
+		}
+
+	}
+
+	if (dbi > udev->dbi_max)
+		udev->dbi_max = dbi;
+
+	set_bit(dbi, udev->data_bitmap);
+	tcmu_cmd_set_dbi(tcmu_cmd, dbi);
+
+	return true;
+}
+
+static bool tcmu_get_empty_blocks(struct tcmu_dev *udev,
+				  struct tcmu_cmd *tcmu_cmd)
+{
+	int i;
+
+	udev->waiting_global = false;
+
+	for (i = tcmu_cmd->dbi_cur; i < tcmu_cmd->dbi_cnt; i++) {
+		if (!tcmu_get_empty_block(udev, tcmu_cmd))
+			goto err;
+	}
+	return true;
+
+err:
+	udev->waiting_global = true;
+	/* Try to wake up the unmap thread */
+	wake_up(&unmap_wait);
+	return false;
+}
+
+static inline struct page *
+tcmu_get_block_page(struct tcmu_dev *udev, uint32_t dbi)
+{
+	return radix_tree_lookup(&udev->data_blocks, dbi);
+}
+
+static inline void tcmu_free_cmd(struct tcmu_cmd *tcmu_cmd)
+{
+	kfree(tcmu_cmd->dbi);
+	kmem_cache_free(tcmu_cmd_cache, tcmu_cmd);
+}
+
+static inline size_t tcmu_cmd_get_data_length(struct tcmu_cmd *tcmu_cmd)
+{
+	struct se_cmd *se_cmd = tcmu_cmd->se_cmd;
+	size_t data_length = round_up(se_cmd->data_length, DATA_BLOCK_SIZE);
+
+	if (se_cmd->se_cmd_flags & SCF_BIDI) {
+		BUG_ON(!(se_cmd->t_bidi_data_sg && se_cmd->t_bidi_data_nents));
+		data_length += round_up(se_cmd->t_bidi_data_sg->length,
+				DATA_BLOCK_SIZE);
+	}
+
+	return data_length;
+}
+
+static inline uint32_t tcmu_cmd_get_block_cnt(struct tcmu_cmd *tcmu_cmd)
+{
+	size_t data_length = tcmu_cmd_get_data_length(tcmu_cmd);
+
+	return data_length / DATA_BLOCK_SIZE;
+}
+
 static struct tcmu_cmd *tcmu_alloc_cmd(struct se_cmd *se_cmd)
 {
 	struct se_device *se_dev = se_cmd->se_dev;
@@ -178,6 +315,15 @@ static struct tcmu_cmd *tcmu_alloc_cmd(struct se_cmd *se_cmd)
 		tcmu_cmd->deadline = jiffies +
 					msecs_to_jiffies(udev->cmd_time_out);
 
+	tcmu_cmd_reset_dbi_cur(tcmu_cmd);
+	tcmu_cmd->dbi_cnt = tcmu_cmd_get_block_cnt(tcmu_cmd);
+	tcmu_cmd->dbi = kcalloc(tcmu_cmd->dbi_cnt, sizeof(uint32_t),
+				GFP_KERNEL);
+	if (!tcmu_cmd->dbi) {
+		kmem_cache_free(tcmu_cmd_cache, tcmu_cmd);
+		return NULL;
+	}
+
 	idr_preload(GFP_KERNEL);
 	spin_lock_irq(&udev->commands_lock);
 	cmd_id = idr_alloc(&udev->commands, tcmu_cmd, 0,
@@ -186,7 +332,7 @@ static struct tcmu_cmd *tcmu_alloc_cmd(struct se_cmd *se_cmd)
 	idr_preload_end();
 
 	if (cmd_id < 0) {
-		kmem_cache_free(tcmu_cmd_cache, tcmu_cmd);
+		tcmu_free_cmd(tcmu_cmd);
 		return NULL;
 	}
 	tcmu_cmd->cmd_id = cmd_id;
@@ -248,10 +394,10 @@ static inline void new_iov(struct iovec **iov, int *iov_cnt,
 #define UPDATE_HEAD(head, used, size) smp_store_release(&head, ((head % size) + used) % size)
 
 /* offset is relative to mb_addr */
-static inline size_t get_block_offset(struct tcmu_dev *dev,
-		int block, int remaining)
+static inline size_t get_block_offset_user(struct tcmu_dev *dev,
+		int dbi, int remaining)
 {
-	return dev->data_off + block * DATA_BLOCK_SIZE +
+	return dev->data_off + dbi * DATA_BLOCK_SIZE +
 		DATA_BLOCK_SIZE - remaining;
 }
 
@@ -260,37 +406,45 @@ static inline size_t iov_tail(struct tcmu_dev *udev, struct iovec *iov)
 	return (size_t)iov->iov_base + iov->iov_len;
 }
 
-static void alloc_and_scatter_data_area(struct tcmu_dev *udev,
-	struct scatterlist *data_sg, unsigned int data_nents,
-	struct iovec **iov, int *iov_cnt, bool copy_data)
+static int scatter_data_area(struct tcmu_dev *udev,
+	struct tcmu_cmd *tcmu_cmd, struct scatterlist *data_sg,
+	unsigned int data_nents, struct iovec **iov,
+	int *iov_cnt, bool copy_data)
 {
-	int i, block;
+	int i, dbi;
 	int block_remaining = 0;
-	void *from, *to;
-	size_t copy_bytes, to_offset;
+	void *from, *to = NULL;
+	size_t copy_bytes, to_offset, offset;
 	struct scatterlist *sg;
+	struct page *page;
 
 	for_each_sg(data_sg, sg, data_nents, i) {
 		int sg_remaining = sg->length;
 		from = kmap_atomic(sg_page(sg)) + sg->offset;
 		while (sg_remaining > 0) {
 			if (block_remaining == 0) {
-				block = find_first_zero_bit(udev->data_bitmap,
-						DATA_BLOCK_BITS);
+				if (to)
+					kunmap_atomic(to);
+
 				block_remaining = DATA_BLOCK_SIZE;
-				set_bit(block, udev->data_bitmap);
+				dbi = tcmu_cmd_get_dbi(tcmu_cmd);
+				page = tcmu_get_block_page(udev, dbi);
+				to = kmap_atomic(page);
 			}
+
 			copy_bytes = min_t(size_t, sg_remaining,
 					block_remaining);
-			to_offset = get_block_offset(udev, block,
+			to_offset = get_block_offset_user(udev, dbi,
 					block_remaining);
-			to = (void *)udev->mb_addr + to_offset;
+			offset = DATA_BLOCK_SIZE - block_remaining;
+			to = (void *)(unsigned long)to + offset;
+
 			if (*iov_cnt != 0 &&
 			    to_offset == iov_tail(udev, *iov)) {
 				(*iov)->iov_len += copy_bytes;
 			} else {
 				new_iov(iov, iov_cnt, udev);
-				(*iov)->iov_base = (void __user *) to_offset;
+				(*iov)->iov_base = (void __user *)to_offset;
 				(*iov)->iov_len = copy_bytes;
 			}
 			if (copy_data) {
@@ -303,33 +457,29 @@ static void alloc_and_scatter_data_area(struct tcmu_dev *udev,
 		}
 		kunmap_atomic(from - sg->offset);
 	}
-}
+	if (to)
+		kunmap_atomic(to);
 
-static void free_data_area(struct tcmu_dev *udev, struct tcmu_cmd *cmd)
-{
-	bitmap_xor(udev->data_bitmap, udev->data_bitmap, cmd->data_bitmap,
-		   DATA_BLOCK_BITS);
+	return 0;
 }
 
 static void gather_data_area(struct tcmu_dev *udev, struct tcmu_cmd *cmd,
 			     bool bidi)
 {
 	struct se_cmd *se_cmd = cmd->se_cmd;
-	int i, block;
+	int i, dbi;
 	int block_remaining = 0;
-	void *from, *to;
-	size_t copy_bytes, from_offset;
+	void *from = NULL, *to;
+	size_t copy_bytes, offset;
 	struct scatterlist *sg, *data_sg;
+	struct page *page;
 	unsigned int data_nents;
-	DECLARE_BITMAP(bitmap, DATA_BLOCK_BITS);
-
-	bitmap_copy(bitmap, cmd->data_bitmap, DATA_BLOCK_BITS);
+	uint32_t count = 0;
 
 	if (!bidi) {
 		data_sg = se_cmd->t_data_sg;
 		data_nents = se_cmd->t_data_nents;
 	} else {
-		uint32_t count;
 
 		/*
 		 * For bidi case, the first count blocks are for Data-Out
@@ -337,30 +487,30 @@ static void gather_data_area(struct tcmu_dev *udev, struct tcmu_cmd *cmd,
 		 * the Data-Out buffer blocks should be discarded.
 		 */
 		count = DIV_ROUND_UP(se_cmd->data_length, DATA_BLOCK_SIZE);
-		while (count--) {
-			block = find_first_bit(bitmap, DATA_BLOCK_BITS);
-			clear_bit(block, bitmap);
-		}
 
 		data_sg = se_cmd->t_bidi_data_sg;
 		data_nents = se_cmd->t_bidi_data_nents;
 	}
 
+	tcmu_cmd_set_dbi_cur(cmd, count);
+
 	for_each_sg(data_sg, sg, data_nents, i) {
 		int sg_remaining = sg->length;
 		to = kmap_atomic(sg_page(sg)) + sg->offset;
 		while (sg_remaining > 0) {
 			if (block_remaining == 0) {
-				block = find_first_bit(bitmap,
-						DATA_BLOCK_BITS);
+				if (from)
+					kunmap_atomic(from);
+
 				block_remaining = DATA_BLOCK_SIZE;
-				clear_bit(block, bitmap);
+				dbi = tcmu_cmd_get_dbi(cmd);
+				page = tcmu_get_block_page(udev, dbi);
+				from = kmap_atomic(page);
 			}
 			copy_bytes = min_t(size_t, sg_remaining,
 					block_remaining);
-			from_offset = get_block_offset(udev, block,
-					block_remaining);
-			from = (void *) udev->mb_addr + from_offset;
+			offset = DATA_BLOCK_SIZE - block_remaining;
+			from = (void *)(unsigned long)from + offset;
 			tcmu_flush_dcache_range(from, copy_bytes);
 			memcpy(to + sg->length - sg_remaining, from,
 					copy_bytes);
@@ -370,12 +520,13 @@ static void gather_data_area(struct tcmu_dev *udev, struct tcmu_cmd *cmd,
 		}
 		kunmap_atomic(to - sg->offset);
 	}
+	if (from)
+		kunmap_atomic(from);
 }
 
-static inline size_t spc_bitmap_free(unsigned long *bitmap)
+static inline size_t spc_bitmap_free(unsigned long *bitmap, uint32_t thresh)
 {
-	return DATA_BLOCK_SIZE * (DATA_BLOCK_BITS -
-			bitmap_weight(bitmap, DATA_BLOCK_BITS));
+	return DATA_BLOCK_SIZE * (thresh - bitmap_weight(bitmap, thresh));
 }
 
 /*
@@ -384,9 +535,12 @@ static inline size_t spc_bitmap_free(unsigned long *bitmap)
  *
  * Called with ring lock held.
  */
-static bool is_ring_space_avail(struct tcmu_dev *udev, size_t cmd_size, size_t data_needed)
+static bool is_ring_space_avail(struct tcmu_dev *udev, struct tcmu_cmd *cmd,
+		size_t cmd_size, size_t data_needed)
 {
 	struct tcmu_mailbox *mb = udev->mb_addr;
+	uint32_t blocks_needed = (data_needed + DATA_BLOCK_SIZE - 1)
+				/ DATA_BLOCK_SIZE;
 	size_t space, cmd_needed;
 	u32 cmd_head;
 
@@ -410,35 +564,63 @@ static bool is_ring_space_avail(struct tcmu_dev *udev, size_t cmd_size, size_t d
 		return false;
 	}
 
-	space = spc_bitmap_free(udev->data_bitmap);
+	/* try to check and get the data blocks as needed */
+	space = spc_bitmap_free(udev->data_bitmap, udev->dbi_thresh);
 	if (space < data_needed) {
-		pr_debug("no data space: only %zu available, but ask for %zu\n",
-				space, data_needed);
+		unsigned long blocks_left = DATA_BLOCK_BITS - udev->dbi_thresh;
+		unsigned long grow;
+
+		if (blocks_left < blocks_needed) {
+			pr_debug("no data space: only %lu available, but ask for %zu\n",
+					blocks_left * DATA_BLOCK_SIZE,
+					data_needed);
 			return false;
 		}
 
+		/* Try to expand the thresh */
+		if (!udev->dbi_thresh) {
+			/* From idle state */
+			uint32_t init_thresh = DATA_BLOCK_INIT_BITS;
+
+			udev->dbi_thresh = max(blocks_needed, init_thresh);
+		} else {
+			/*
+			 * Grow the data area by max(blocks needed,
+			 * dbi_thresh / 2), but limited to the max
+			 * DATA_BLOCK_BITS size.
+			 */
+			grow = max(blocks_needed, udev->dbi_thresh / 2);
+			udev->dbi_thresh += grow;
+			if (udev->dbi_thresh > DATA_BLOCK_BITS)
+				udev->dbi_thresh = DATA_BLOCK_BITS;
+		}
+	}
+
+	if (!tcmu_get_empty_blocks(udev, cmd))
+		return false;
+
 	return true;
 }
 
-static inline size_t tcmu_cmd_get_data_length(struct tcmu_cmd *tcmu_cmd)
+static inline size_t tcmu_cmd_get_base_cmd_size(size_t iov_cnt)
 {
-	struct se_cmd *se_cmd = tcmu_cmd->se_cmd;
-	size_t data_length = round_up(se_cmd->data_length, DATA_BLOCK_SIZE);
-
-	if (se_cmd->se_cmd_flags & SCF_BIDI) {
-		BUG_ON(!(se_cmd->t_bidi_data_sg && se_cmd->t_bidi_data_nents));
-		data_length += round_up(se_cmd->t_bidi_data_sg->length,
-				DATA_BLOCK_SIZE);
-	}
-
-	return data_length;
+	return max(offsetof(struct tcmu_cmd_entry, req.iov[iov_cnt]),
+			sizeof(struct tcmu_cmd_entry));
 }
 
-static inline uint32_t tcmu_cmd_get_block_cnt(struct tcmu_cmd *tcmu_cmd)
+static inline size_t tcmu_cmd_get_cmd_size(struct tcmu_cmd *tcmu_cmd,
+					   size_t base_command_size)
 {
-	size_t data_length = tcmu_cmd_get_data_length(tcmu_cmd);
+	struct se_cmd *se_cmd = tcmu_cmd->se_cmd;
+	size_t command_size;
 
-	return data_length / DATA_BLOCK_SIZE;
+	command_size = base_command_size +
+		round_up(scsi_command_size(se_cmd->t_task_cdb),
+				TCMU_OP_ALIGN_SIZE);
+
+	WARN_ON(command_size & (TCMU_OP_ALIGN_SIZE-1));
+
+	return command_size;
 }
 
 static sense_reason_t
@@ -450,12 +632,11 @@ tcmu_queue_cmd_ring(struct tcmu_cmd *tcmu_cmd)
 	struct tcmu_mailbox *mb;
 	struct tcmu_cmd_entry *entry;
 	struct iovec *iov;
-	int iov_cnt;
+	int iov_cnt, ret;
 	uint32_t cmd_head;
 	uint64_t cdb_off;
 	bool copy_to_data_area;
 	size_t data_length = tcmu_cmd_get_data_length(tcmu_cmd);
-	DECLARE_BITMAP(old_bitmap, DATA_BLOCK_BITS);
 
 	if (test_bit(TCMU_DEV_BIT_BROKEN, &udev->flags))
 		return TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
@@ -464,18 +645,18 @@ tcmu_queue_cmd_ring(struct tcmu_cmd *tcmu_cmd)
 	 * Must be a certain minimum size for response sense info, but
 	 * also may be larger if the iov array is large.
 	 *
-	 * We prepare way too many iovs for potential uses here, because it's
-	 * expensive to tell how many regions are freed in the bitmap
+	 * We prepare as many iovs as possbile for potential uses here,
+	 * because it's expensive to tell how many regions are freed in
+	 * the bitmap & global data pool, as the size calculated here
+	 * will only be used to do the checks.
+	 *
+	 * The size will be recalculated later as actually needed to save
+	 * cmd area memories.
 	 */
-	base_command_size = max(offsetof(struct tcmu_cmd_entry,
-				req.iov[tcmu_cmd_get_block_cnt(tcmu_cmd)]),
-				sizeof(struct tcmu_cmd_entry));
-	command_size = base_command_size
-		+ round_up(scsi_command_size(se_cmd->t_task_cdb), TCMU_OP_ALIGN_SIZE);
-
-	WARN_ON(command_size & (TCMU_OP_ALIGN_SIZE-1));
+	base_command_size = tcmu_cmd_get_base_cmd_size(tcmu_cmd->dbi_cnt);
+	command_size = tcmu_cmd_get_cmd_size(tcmu_cmd, base_command_size);
 
-	spin_lock_irq(&udev->cmdr_lock);
+	mutex_lock(&udev->cmdr_lock);
 
 	mb = udev->mb_addr;
 	cmd_head = mb->cmd_head % udev->cmdr_size; /* UAM */
@@ -484,18 +665,18 @@ tcmu_queue_cmd_ring(struct tcmu_cmd *tcmu_cmd)
 		pr_warn("TCMU: Request of size %zu/%zu is too big for %u/%zu "
 			"cmd ring/data area\n", command_size, data_length,
 			udev->cmdr_size, udev->data_size);
-		spin_unlock_irq(&udev->cmdr_lock);
+		mutex_unlock(&udev->cmdr_lock);
 		return TCM_INVALID_CDB_FIELD;
 	}
 
-	while (!is_ring_space_avail(udev, command_size, data_length)) {
+	while (!is_ring_space_avail(udev, tcmu_cmd, command_size, data_length)) {
 		int ret;
 		DEFINE_WAIT(__wait);
 
 		prepare_to_wait(&udev->wait_cmdr, &__wait, TASK_INTERRUPTIBLE);
 
 		pr_debug("sleeping for ring space\n");
-		spin_unlock_irq(&udev->cmdr_lock);
+		mutex_unlock(&udev->cmdr_lock);
 		if (udev->cmd_time_out)
 			ret = schedule_timeout(
 					msecs_to_jiffies(udev->cmd_time_out));
@@ -507,7 +688,7 @@ tcmu_queue_cmd_ring(struct tcmu_cmd *tcmu_cmd)
 			return TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
 		}
 
-		spin_lock_irq(&udev->cmdr_lock);
+		mutex_lock(&udev->cmdr_lock);
 
 		/* We dropped cmdr_lock, cmd_head is stale */
 		cmd_head = mb->cmd_head % udev->cmdr_size; /* UAM */
@@ -534,20 +715,26 @@ tcmu_queue_cmd_ring(struct tcmu_cmd *tcmu_cmd)
 	entry = (void *) mb + CMDR_OFF + cmd_head;
 	tcmu_flush_dcache_range(entry, sizeof(*entry));
 	tcmu_hdr_set_op(&entry->hdr.len_op, TCMU_OP_CMD);
-	tcmu_hdr_set_len(&entry->hdr.len_op, command_size);
 	entry->hdr.cmd_id = tcmu_cmd->cmd_id;
 	entry->hdr.kflags = 0;
 	entry->hdr.uflags = 0;
 
-	bitmap_copy(old_bitmap, udev->data_bitmap, DATA_BLOCK_BITS);
-
 	/* Handle allocating space from the data area */
+	tcmu_cmd_reset_dbi_cur(tcmu_cmd);
 	iov = &entry->req.iov[0];
 	iov_cnt = 0;
 	copy_to_data_area = (se_cmd->data_direction == DMA_TO_DEVICE
 		|| se_cmd->se_cmd_flags & SCF_BIDI);
-	alloc_and_scatter_data_area(udev, se_cmd->t_data_sg,
-		se_cmd->t_data_nents, &iov, &iov_cnt, copy_to_data_area);
+	ret = scatter_data_area(udev, tcmu_cmd, se_cmd->t_data_sg,
+				se_cmd->t_data_nents, &iov, &iov_cnt,
+				copy_to_data_area);
+	if (ret) {
+		tcmu_cmd_free_data(tcmu_cmd, tcmu_cmd->dbi_cnt);
+		mutex_unlock(&udev->cmdr_lock);
+
+		pr_err("tcmu: alloc and scatter data failed\n");
+		return TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
+	}
 	entry->req.iov_cnt = iov_cnt;
 	entry->req.iov_dif_cnt = 0;
 
@@ -555,14 +742,29 @@ tcmu_queue_cmd_ring(struct tcmu_cmd *tcmu_cmd)
 	if (se_cmd->se_cmd_flags & SCF_BIDI) {
 		iov_cnt = 0;
 		iov++;
-		alloc_and_scatter_data_area(udev, se_cmd->t_bidi_data_sg,
-				se_cmd->t_bidi_data_nents, &iov, &iov_cnt,
-				false);
+		ret = scatter_data_area(udev, tcmu_cmd,
+					se_cmd->t_bidi_data_sg,
+					se_cmd->t_bidi_data_nents,
+					&iov, &iov_cnt, false);
+		if (ret) {
+			tcmu_cmd_free_data(tcmu_cmd, tcmu_cmd->dbi_cnt);
+			mutex_unlock(&udev->cmdr_lock);
+
+			pr_err("tcmu: alloc and scatter bidi data failed\n");
+			return TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
+		}
 		entry->req.iov_bidi_cnt = iov_cnt;
 	}
-	/* cmd's data_bitmap is what changed in process */
-	bitmap_xor(tcmu_cmd->data_bitmap, old_bitmap, udev->data_bitmap,
-			DATA_BLOCK_BITS);
+
+	/*
+	 * Recalaulate the command's base size and size according
+	 * to the actual needs
+	 */
+	base_command_size = tcmu_cmd_get_base_cmd_size(entry->req.iov_cnt +
+						       entry->req.iov_bidi_cnt);
+	command_size = tcmu_cmd_get_cmd_size(tcmu_cmd, base_command_size);
+
+	tcmu_hdr_set_len(&entry->hdr.len_op, command_size);
 
 	/* All offsets relative to mb_addr, not start of entry! */
 	cdb_off = CMDR_OFF + cmd_head + base_command_size;
@@ -572,8 +774,7 @@ tcmu_queue_cmd_ring(struct tcmu_cmd *tcmu_cmd)
 
 	UPDATE_HEAD(mb->cmd_head, command_size, udev->cmdr_size);
 	tcmu_flush_dcache_range(mb, sizeof(*mb));
-
-	spin_unlock_irq(&udev->cmdr_lock);
+	mutex_unlock(&udev->cmdr_lock);
 
 	/* TODO: only if FLUSH and FUA? */
 	uio_event_notify(&udev->uio_info);
@@ -604,7 +805,7 @@ tcmu_queue_cmd(struct se_cmd *se_cmd)
 		idr_remove(&udev->commands, tcmu_cmd->cmd_id);
 		spin_unlock_irq(&udev->commands_lock);
 
-		kmem_cache_free(tcmu_cmd_cache, tcmu_cmd);
+		tcmu_free_cmd(tcmu_cmd);
 	}
 
 	return ret;
@@ -615,50 +816,45 @@ static void tcmu_handle_completion(struct tcmu_cmd *cmd, struct tcmu_cmd_entry *
 	struct se_cmd *se_cmd = cmd->se_cmd;
 	struct tcmu_dev *udev = cmd->tcmu_dev;
 
-	if (test_bit(TCMU_CMD_BIT_EXPIRED, &cmd->flags)) {
 	/*
 	 * cmd has been completed already from timeout, just reclaim
 	 * data area space and free cmd
 	 */
-		free_data_area(udev, cmd);
+	if (test_bit(TCMU_CMD_BIT_EXPIRED, &cmd->flags))
+		goto out;
 
-		kmem_cache_free(tcmu_cmd_cache, cmd);
-		return;
-	}
+	tcmu_cmd_reset_dbi_cur(cmd);
 
 	if (entry->hdr.uflags & TCMU_UFLAG_UNKNOWN_OP) {
-		free_data_area(udev, cmd);
 		pr_warn("TCMU: Userspace set UNKNOWN_OP flag on se_cmd %p\n",
 			cmd->se_cmd);
 		entry->rsp.scsi_status = SAM_STAT_CHECK_CONDITION;
 	} else if (entry->rsp.scsi_status == SAM_STAT_CHECK_CONDITION) {
 		memcpy(se_cmd->sense_buffer, entry->rsp.sense_buffer,
 			       se_cmd->scsi_sense_length);
-		free_data_area(udev, cmd);
 	} else if (se_cmd->se_cmd_flags & SCF_BIDI) {
 		/* Get Data-In buffer before clean up */
 		gather_data_area(udev, cmd, true);
-		free_data_area(udev, cmd);
 	} else if (se_cmd->data_direction == DMA_FROM_DEVICE) {
 		gather_data_area(udev, cmd, false);
-		free_data_area(udev, cmd);
 	} else if (se_cmd->data_direction == DMA_TO_DEVICE) {
-		free_data_area(udev, cmd);
+		/* TODO: */
 	} else if (se_cmd->data_direction != DMA_NONE) {
 		pr_warn("TCMU: data direction was %d!\n",
 			se_cmd->data_direction);
 	}
 
 	target_complete_cmd(cmd->se_cmd, entry->rsp.scsi_status);
-	cmd->se_cmd = NULL;
 
-	kmem_cache_free(tcmu_cmd_cache, cmd);
+out:
+	cmd->se_cmd = NULL;
+	tcmu_cmd_free_data(cmd, cmd->dbi_cnt);
+	tcmu_free_cmd(cmd);
 }
 
 static unsigned int tcmu_handle_completions(struct tcmu_dev *udev)
 {
 	struct tcmu_mailbox *mb;
-	unsigned long flags;
 	int handled = 0;
 
 	if (test_bit(TCMU_DEV_BIT_BROKEN, &udev->flags)) {
@@ -666,8 +862,6 @@ static unsigned int tcmu_handle_completions(struct tcmu_dev *udev)
 		return 0;
 	}
 
-	spin_lock_irqsave(&udev->cmdr_lock, flags);
-
 	mb = udev->mb_addr;
 	tcmu_flush_dcache_range(mb, sizeof(*mb));
 
@@ -708,8 +902,6 @@ static unsigned int tcmu_handle_completions(struct tcmu_dev *udev)
 	if (mb->cmd_tail == mb->cmd_head)
 		del_timer(&udev->timeout); /* no more pending cmds */
 
-	spin_unlock_irqrestore(&udev->cmdr_lock, flags);
-
 	wake_up(&udev->wait_cmdr);
 
 	return handled;
@@ -736,16 +928,14 @@ static void tcmu_device_timedout(unsigned long data)
 {
 	struct tcmu_dev *udev = (struct tcmu_dev *)data;
 	unsigned long flags;
-	int handled;
-
-	handled = tcmu_handle_completions(udev);
-
-	pr_warn("%d completions handled from timeout\n", handled);
 
 	spin_lock_irqsave(&udev->commands_lock, flags);
 	idr_for_each(&udev->commands, tcmu_check_expired_cmd, NULL);
 	spin_unlock_irqrestore(&udev->commands_lock, flags);
 
+	/* Try to wake up the ummap thread */
+	wake_up(&unmap_wait);
+
 	/*
 	 * We don't need to wakeup threads on wait_cmdr since they have their
 	 * own timeout.
@@ -790,7 +980,7 @@ static struct se_device *tcmu_alloc_device(struct se_hba *hba, const char *name)
 	udev->cmd_time_out = TCMU_TIME_OUT;
 
 	init_waitqueue_head(&udev->wait_cmdr);
-	spin_lock_init(&udev->cmdr_lock);
+	mutex_init(&udev->cmdr_lock);
 
 	idr_init(&udev->commands);
 	spin_lock_init(&udev->commands_lock);
@@ -805,7 +995,9 @@ static int tcmu_irqcontrol(struct uio_info *info, s32 irq_on)
 {
 	struct tcmu_dev *tcmu_dev = container_of(info, struct tcmu_dev, uio_info);
 
+	mutex_lock(&tcmu_dev->cmdr_lock);
 	tcmu_handle_completions(tcmu_dev);
+	mutex_unlock(&tcmu_dev->cmdr_lock);
 
 	return 0;
 }
@@ -827,6 +1019,60 @@ static int tcmu_find_mem_index(struct vm_area_struct *vma)
 	return -1;
 }
 
+static struct page *tcmu_try_get_block_page(struct tcmu_dev *udev, uint32_t dbi)
+{
+	struct page *page;
+	int ret;
+
+	mutex_lock(&udev->cmdr_lock);
+	page = tcmu_get_block_page(udev, dbi);
+	if (likely(page)) {
+		mutex_unlock(&udev->cmdr_lock);
+		return page;
+	}
+
+	/*
+	 * Normally it shouldn't be here:
+	 * Only when the userspace has touched the blocks which
+	 * are out of the tcmu_cmd's data iov[], and will return
+	 * one zeroed page.
+	 */
+	pr_warn("Block(%u) out of cmd's iov[] has been touched!\n", dbi);
+	pr_warn("Mostly it will be a bug of userspace, please have a check!\n");
+
+	if (dbi >= udev->dbi_thresh) {
+		/* Extern the udev->dbi_thresh to dbi + 1 */
+		udev->dbi_thresh = dbi + 1;
+		udev->dbi_max = dbi;
+	}
+
+	page = radix_tree_lookup(&udev->data_blocks, dbi);
+	if (!page) {
+		page = alloc_page(GFP_KERNEL | __GFP_ZERO);
+		if (!page) {
+			mutex_unlock(&udev->cmdr_lock);
+			return NULL;
+		}
+
+		ret = radix_tree_insert(&udev->data_blocks, dbi, page);
+		if (ret) {
+			mutex_unlock(&udev->cmdr_lock);
+			__free_page(page);
+			return NULL;
+		}
+
+		/*
+		 * Since this case is rare in page fault routine, here we
+		 * will allow the global_db_count >= TCMU_GLOBAL_MAX_BLOCKS
+		 * to reduce possible page fault call trace.
+		 */
+		atomic_inc(&global_db_count);
+	}
+	mutex_unlock(&udev->cmdr_lock);
+
+	return page;
+}
+
 static int tcmu_vma_fault(struct vm_fault *vmf)
 {
 	struct tcmu_dev *udev = vmf->vma->vm_private_data;
@@ -845,11 +1091,20 @@ static int tcmu_vma_fault(struct vm_fault *vmf)
 	 */
 	offset = (vmf->pgoff - mi) << PAGE_SHIFT;
 
+	if (offset < udev->data_off) {
+		/* For the vmalloc()ed cmd area pages */
 		addr = (void *)(unsigned long)info->mem[mi].addr + offset;
-	if (info->mem[mi].memtype == UIO_MEM_LOGICAL)
-		page = virt_to_page(addr);
-	else
 		page = vmalloc_to_page(addr);
+	} else {
+		uint32_t dbi;
+
+		/* For the dynamically growing data area pages */
+		dbi = (offset - udev->data_off) / DATA_BLOCK_SIZE;
+		page = tcmu_try_get_block_page(udev, dbi);
+		if (!page)
+			return VM_FAULT_NOPAGE;
+	}
+
 	get_page(page);
 	vmf->page = page;
 	return 0;
@@ -883,6 +1138,8 @@ static int tcmu_open(struct uio_info *info, struct inode *inode)
 	if (test_and_set_bit(TCMU_DEV_BIT_OPEN, &udev->flags))
 		return -EBUSY;
 
+	udev->inode = inode;
+
 	pr_debug("open\n");
 
 	return 0;
@@ -963,7 +1220,7 @@ static int tcmu_configure_device(struct se_device *dev)
 
 	info->name = str;
 
-	udev->mb_addr = vzalloc(TCMU_RING_SIZE);
+	udev->mb_addr = vzalloc(CMDR_SIZE);
 	if (!udev->mb_addr) {
 		ret = -ENOMEM;
 		goto err_vzalloc;
@@ -972,8 +1229,11 @@ static int tcmu_configure_device(struct se_device *dev)
 	/* mailbox fits in first part of CMDR space */
 	udev->cmdr_size = CMDR_SIZE - CMDR_OFF;
 	udev->data_off = CMDR_SIZE;
-	udev->data_size = TCMU_RING_SIZE - CMDR_SIZE;
+	udev->data_size = DATA_SIZE;
+	udev->dbi_thresh = 0; /* Default in Idle state */
+	udev->waiting_global = false;
 
+	/* Initialise the mailbox of the ring buffer */
 	mb = udev->mb_addr;
 	mb->version = TCMU_MAILBOX_VERSION;
 	mb->flags = TCMU_MAILBOX_FLAG_CAP_OOOC;
@@ -984,12 +1244,14 @@ static int tcmu_configure_device(struct se_device *dev)
 	WARN_ON(udev->data_size % PAGE_SIZE);
 	WARN_ON(udev->data_size % DATA_BLOCK_SIZE);
 
+	INIT_RADIX_TREE(&udev->data_blocks, GFP_KERNEL);
+
 	info->version = __stringify(TCMU_MAILBOX_VERSION);
 
 	info->mem[0].name = "tcm-user command & data buffer";
 	info->mem[0].addr = (phys_addr_t)(uintptr_t)udev->mb_addr;
 	info->mem[0].size = TCMU_RING_SIZE;
-	info->mem[0].memtype = UIO_MEM_VIRTUAL;
+	info->mem[0].memtype = UIO_MEM_NONE;
 
 	info->irqcontrol = tcmu_irqcontrol;
 	info->irq = UIO_IRQ_CUSTOM;
@@ -1015,6 +1277,10 @@ static int tcmu_configure_device(struct se_device *dev)
 	if (ret)
 		goto err_netlink;
 
+	mutex_lock(&root_udev_mutex);
+	list_add(&udev->node, &root_udev);
+	mutex_unlock(&root_udev_mutex);
+
 	return 0;
 
 err_netlink:
@@ -1049,6 +1315,23 @@ static bool tcmu_dev_configured(struct tcmu_dev *udev)
 	return udev->uio_info.uio_dev ? true : false;
 }
 
+static void tcmu_blocks_release(struct tcmu_dev *udev)
+{
+	int i;
+	struct page *page;
+
+	/* Try to release all block pages */
+	mutex_lock(&udev->cmdr_lock);
+	for (i = 0; i <= udev->dbi_max; i++) {
+		page = radix_tree_delete(&udev->data_blocks, i);
+		if (page) {
+			__free_page(page);
+			atomic_dec(&global_db_count);
+		}
+	}
+	mutex_unlock(&udev->cmdr_lock);
+}
+
 static void tcmu_free_device(struct se_device *dev)
 {
 	struct tcmu_dev *udev = TCMU_DEV(dev);
@@ -1058,6 +1341,10 @@ static void tcmu_free_device(struct se_device *dev)
 
 	del_timer_sync(&udev->timeout);
 
+	mutex_lock(&root_udev_mutex);
+	list_del(&udev->node);
+	mutex_unlock(&root_udev_mutex);
+
 	vfree(udev->mb_addr);
 
 	/* Upper layer should drain all requests before calling this */
@@ -1070,6 +1357,8 @@ static void tcmu_free_device(struct se_device *dev)
 	spin_unlock_irq(&udev->commands_lock);
 	WARN_ON(!all_expired);
 
+	tcmu_blocks_release(udev);
+
 	if (tcmu_dev_configured(udev)) {
 		tcmu_netlink_event(TCMU_CMD_REMOVED_DEVICE, udev->uio_info.name,
 				   udev->uio_info.uio_dev->minor);
@@ -1256,6 +1545,84 @@ static struct target_backend_ops tcmu_ops = {
 	.tb_dev_attrib_attrs	= NULL,
 };
 
+static int unmap_thread_fn(void *data)
+{
+	struct tcmu_dev *udev;
+	loff_t off;
+	uint32_t start, end, block;
+	struct page *page;
+	int i;
+
+	while (1) {
+		DEFINE_WAIT(__wait);
+
+		prepare_to_wait(&unmap_wait, &__wait, TASK_INTERRUPTIBLE);
+		schedule();
+		finish_wait(&unmap_wait, &__wait);
+
+		if (kthread_should_stop())
+			break;
+
+		mutex_lock(&root_udev_mutex);
+		list_for_each_entry(udev, &root_udev, node) {
+			mutex_lock(&udev->cmdr_lock);
+
+			/* Try to complete the finished commands first */
+			tcmu_handle_completions(udev);
+
+			/* Skip the udevs waiting the global pool or in idle */
+			if (udev->waiting_global || !udev->dbi_thresh) {
+				mutex_unlock(&udev->cmdr_lock);
+				continue;
+			}
+
+			end = udev->dbi_max + 1;
+			block = find_last_bit(udev->data_bitmap, end);
+			if (block == udev->dbi_max) {
+				/*
+				 * The last bit is dbi_max, so there is
+				 * no need to shrink any blocks.
+				 */
+				mutex_unlock(&udev->cmdr_lock);
+				continue;
+			} else if (block == end) {
+				/* The current udev will goto idle state */
+				udev->dbi_thresh = start = 0;
+				udev->dbi_max = 0;
+			} else {
+				udev->dbi_thresh = start = block + 1;
+				udev->dbi_max = block;
+			}
+
+			/* Here will truncate the data area from off */
+			off = udev->data_off + start * DATA_BLOCK_SIZE;
+			unmap_mapping_range(udev->inode->i_mapping, off, 0, 1);
+
+			/* Release the block pages */
+			for (i = start; i < end; i++) {
+				page = radix_tree_delete(&udev->data_blocks, i);
+				if (page) {
+					__free_page(page);
+					atomic_dec(&global_db_count);
+				}
+			}
+			mutex_unlock(&udev->cmdr_lock);
+		}
+
+		/*
+		 * Try to wake up the udevs who are waiting
+		 * for the global data pool.
+		 */
+		list_for_each_entry(udev, &root_udev, node) {
+			if (udev->waiting_global)
+				wake_up(&udev->wait_cmdr);
+		}
+		mutex_unlock(&root_udev_mutex);
+	}
+
+	return 0;
+}
+
 static int __init tcmu_module_init(void)
 {
 	int ret, i, len = 0;
@@ -1301,8 +1668,17 @@ static int __init tcmu_module_init(void)
 	if (ret)
 		goto out_attrs;
 
+	init_waitqueue_head(&unmap_wait);
+	unmap_thread = kthread_run(unmap_thread_fn, NULL, "tcmu_unmap");
+	if (IS_ERR(unmap_thread)) {
+		ret = PTR_ERR(unmap_thread);
+		goto out_unreg_transport;
+	}
+
 	return 0;
 
+out_unreg_transport:
+	target_backend_unregister(&tcmu_ops);
 out_attrs:
 	kfree(tcmu_attrs);
 out_unreg_genl:
@@ -1317,6 +1693,7 @@ static int __init tcmu_module_init(void)
 
 static void __exit tcmu_module_exit(void)
 {
+	kthread_stop(unmap_thread);
 	target_backend_unregister(&tcmu_ops);
 	kfree(tcmu_attrs);
 	genl_unregister_family(&tcmu_genl_family);

include/scsi/scsi_proto.h

@@ -112,6 +112,7 @@
 #define WRITE_16              0x8a
 #define READ_ATTRIBUTE        0x8c
 #define WRITE_ATTRIBUTE	      0x8d
+#define WRITE_VERIFY_16	      0x8e
 #define VERIFY_16	      0x8f
 #define SYNCHRONIZE_CACHE_16  0x91
 #define WRITE_SAME_16	      0x93

include/target/target_core_backend.h

@@ -10,6 +10,7 @@
  * backend module.
  */
 #define TRANSPORT_FLAG_PASSTHROUGH_ALUA		0x2
+#define TRANSPORT_FLAG_PASSTHROUGH_PGR          0x4
 
 struct request_queue;
 struct scatterlist;

include/target/target_core_base.h

@@ -664,6 +664,7 @@ struct se_dev_attrib {
 	int		pi_prot_format;
 	enum target_prot_type pi_prot_type;
 	enum target_prot_type hw_pi_prot_type;
+	int		pi_prot_verify;
 	int		enforce_pr_isids;
 	int		force_pr_aptpl;
 	int		is_nonrot;