diff --git a/drivers/scsi/iscsi_tcp.c b/drivers/scsi/iscsi_tcp.c
index 55181d28291e7daab0e642a12a88d7f5cac01e40..02621588529eea3f7dd5bc82c84982994f6f14b4 100644
--- a/drivers/scsi/iscsi_tcp.c
+++ b/drivers/scsi/iscsi_tcp.c
@@ -132,6 +132,7 @@ static void iscsi_sw_tcp_data_ready(struct sock *sk)
 	struct iscsi_conn *conn;
 	struct iscsi_tcp_conn *tcp_conn;
 	read_descriptor_t rd_desc;
+	int current_cpu;
 
 	read_lock_bh(&sk->sk_callback_lock);
 	conn = sk->sk_user_data;
@@ -141,6 +142,13 @@ static void iscsi_sw_tcp_data_ready(struct sock *sk)
 	}
 	tcp_conn = conn->dd_data;
 
+	/* save intimate cpu when in softirq */
+	if (!sock_owned_by_user_nocheck(sk)) {
+		current_cpu = smp_processor_id();
+		if (conn->intimate_cpu != current_cpu)
+			conn->intimate_cpu = current_cpu;
+	}
+
 	/*
 	 * Use rd_desc to pass 'conn' to iscsi_tcp_recv.
 	 * We set count to 1 because we want the network layer to
diff --git a/drivers/scsi/libiscsi.c b/drivers/scsi/libiscsi.c
index 214fe77eabc0d74bf0360b74ce3a874027935d90..ec356e035c5f0d47ccb2b5ae3be882cab1aaba89 100644
--- a/drivers/scsi/libiscsi.c
+++ b/drivers/scsi/libiscsi.c
@@ -90,9 +90,15 @@ inline void iscsi_conn_queue_work(struct iscsi_conn *conn)
 {
 	struct Scsi_Host *shost = conn->session->host;
 	struct iscsi_host *ihost = shost_priv(shost);
+	int intimate_cpu = conn->intimate_cpu;
 
-	if (ihost->workq)
-		queue_work(ihost->workq, &conn->xmitwork);
+	if (ihost->workq) {
+		/* we expect it to be excuted on the same numa of the intimate cpu */
+		if ((intimate_cpu >= 0) && cpu_possible(intimate_cpu))
+			queue_work_on(intimate_cpu, ihost->workq, &conn->xmitwork);
+		else
+			queue_work(ihost->workq, &conn->xmitwork);
+	}
 }
 EXPORT_SYMBOL_GPL(iscsi_conn_queue_work);
 
@@ -2682,7 +2688,9 @@ struct Scsi_Host *iscsi_host_alloc(struct scsi_host_template *sht,
 	if (xmit_can_sleep) {
 		snprintf(ihost->workq_name, sizeof(ihost->workq_name),
 			"iscsi_q_%d", shost->host_no);
-		ihost->workq = create_singlethread_workqueue(ihost->workq_name);
+		/* this kind of workqueue only support single work */
+		ihost->workq = alloc_ordered_workqueue("%s", __WQ_LEGACY | WQ_MEM_RECLAIM |
+						__WQ_DYNAMIC, ihost->workq_name);
 		if (!ihost->workq)
 			goto free_host;
 	}
@@ -2959,6 +2967,7 @@ iscsi_conn_setup(struct iscsi_cls_session *cls_session, int dd_size,
 	conn->id = conn_idx;
 	conn->exp_statsn = 0;
 	conn->tmf_state = TMF_INITIAL;
+	conn->intimate_cpu = -1;
 
 	timer_setup(&conn->transport_timer, iscsi_check_transport_timeouts, 0);