diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index 78d776b..4c5bbb3 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -974,6 +974,21 @@
 	return 0;
 }
 
+static int mlx4_en_set_features(struct net_device *netdev,
+		netdev_features_t features)
+{
+	struct mlx4_en_priv *priv = netdev_priv(netdev);
+
+	if (features & NETIF_F_LOOPBACK)
+		priv->ctrl_flags |= cpu_to_be32(MLX4_WQE_CTRL_FORCE_LOOPBACK);
+	else
+		priv->ctrl_flags &=
+			cpu_to_be32(~MLX4_WQE_CTRL_FORCE_LOOPBACK);
+
+	return 0;
+
+}
+
 static const struct net_device_ops mlx4_netdev_ops = {
 	.ndo_open		= mlx4_en_open,
 	.ndo_stop		= mlx4_en_close,
@@ -990,6 +1005,7 @@
 #ifdef CONFIG_NET_POLL_CONTROLLER
 	.ndo_poll_controller	= mlx4_en_netpoll,
 #endif
+	.ndo_set_features	= mlx4_en_set_features,
 };
 
 int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
@@ -1022,6 +1038,8 @@
 	priv->port = port;
 	priv->port_up = false;
 	priv->flags = prof->flags;
+	priv->ctrl_flags = cpu_to_be32(MLX4_WQE_CTRL_CQ_UPDATE |
+			MLX4_WQE_CTRL_SOLICITED);
 	priv->tx_ring_num = prof->tx_ring_num;
 	priv->rx_ring_num = prof->rx_ring_num;
 	priv->mac_index = -1;
@@ -1088,6 +1106,7 @@
 	dev->features = dev->hw_features | NETIF_F_HIGHDMA |
 			NETIF_F_HW_VLAN_TX | NETIF_F_HW_VLAN_RX |
 			NETIF_F_HW_VLAN_FILTER;
+	dev->hw_features |= NETIF_F_LOOPBACK;
 
 	mdev->pndev[port] = dev;
 
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_tx.c b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
index 3094f940..807c218 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
@@ -679,8 +679,7 @@
 	tx_desc->ctrl.vlan_tag = cpu_to_be16(vlan_tag);
 	tx_desc->ctrl.ins_vlan = MLX4_WQE_CTRL_INS_VLAN * !!vlan_tag;
 	tx_desc->ctrl.fence_size = (real_size / 16) & 0x3f;
-	tx_desc->ctrl.srcrb_flags = cpu_to_be32(MLX4_WQE_CTRL_CQ_UPDATE |
-						MLX4_WQE_CTRL_SOLICITED);
+	tx_desc->ctrl.srcrb_flags = priv->ctrl_flags;
 	if (likely(skb->ip_summed == CHECKSUM_PARTIAL)) {
 		tx_desc->ctrl.srcrb_flags |= cpu_to_be32(MLX4_WQE_CTRL_IP_CSUM |
 							 MLX4_WQE_CTRL_TCP_UDP_CSUM);
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
index d2dd97f..ea2ba68 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
@@ -453,6 +453,7 @@
 	int base_qpn;
 
 	struct mlx4_en_rss_map rss_map;
+	u32 ctrl_flags;
 	u32 flags;
 #define MLX4_EN_FLAG_PROMISC	0x1
 #define MLX4_EN_FLAG_MC_PROMISC	0x2
diff --git a/include/linux/mlx4/qp.h b/include/linux/mlx4/qp.h
index 6562ff6..bee8fa2 100644
--- a/include/linux/mlx4/qp.h
+++ b/include/linux/mlx4/qp.h
@@ -210,6 +210,7 @@
 	 * [4]   IP checksum
 	 * [3:2] C (generate completion queue entry)
 	 * [1]   SE (solicited event)
+	 * [0]   FL (force loopback)
 	 */
 	__be32			srcrb_flags;
 	/*