[RFC PATCH 0/7] tc cls_u32 hardware interface

@@ -4268,6 +4268,13 @@ int bnx2x_setup_tc(struct net_device *dev, u8 num_tc)
 	return 0;
 }
 
+int __bnx2x_setup_tc(struct net_device *dev, u32 handle, u8 num_tc)
+{
+	if (handle != TC_H_ROOT)
+		return -EINVAL;
+	return bnx2x_setup_tc(dev, num_tc);
+}
+
 /* called with rtnl_lock */
 int bnx2x_change_mac_addr(struct net_device *dev, void *p)
 {

@@ -486,6 +486,7 @@ netdev_tx_t bnx2x_start_xmit(struct sk_buff *skb, struct net_device *dev);
 
 /* setup_tc callback */
 int bnx2x_setup_tc(struct net_device *dev, u8 num_tc);
+int __bnx2x_setup_tc(struct net_device *dev, u32 handle, u8 num_tc);
 
 int bnx2x_get_vf_config(struct net_device *dev, int vf,
 			struct ifla_vf_info *ivi);

@@ -12994,7 +12994,7 @@ static const struct net_device_ops bnx2x_netdev_ops = {
 #ifdef CONFIG_NET_POLL_CONTROLLER
 	.ndo_poll_controller	= poll_bnx2x,
 #endif
-	.ndo_setup_tc		= bnx2x_setup_tc,
+	.ndo_setup_tc		= __bnx2x_setup_tc,
 #ifdef CONFIG_BNX2X_SRIOV
 	.ndo_set_vf_mac		= bnx2x_set_vf_mac,
 	.ndo_set_vf_vlan	= bnx2x_set_vf_vlan,

@@ -5260,10 +5260,13 @@ static int bnxt_change_mtu(struct net_device *dev, int new_mtu)
 	return 0;
 }
 
-static int bnxt_setup_tc(struct net_device *dev, u8 tc)
+static int bnxt_setup_tc(struct net_device *dev, u32 handle, u8 tc)
 {
 	struct bnxt *bp = netdev_priv(dev);
 
+	if (handle != TC_H_ROOT)
+		return -EINVAL;
+
 	if (tc > bp->max_tc) {
 		netdev_err(dev, "too many traffic classes requested: %d Max supported is %d\n",
 			   tc, bp->max_tc);

@@ -1190,6 +1190,14 @@ int fm10k_setup_tc(struct net_device *dev, u8 tc)
 	return 0;
 }
 
+static int __fm10k_setup_tc(struct net_device *dev, u32 handle, u8 tc)
+{
+	if (handle != TC_H_ROOT)
+		return -EINVAL;
+
+	return fm10k_setup_tc(dev, tc);
+}
+
 static int fm10k_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd)
 {
 	switch (cmd) {

@@ -1372,7 +1380,7 @@ static const struct net_device_ops fm10k_netdev_ops = {
 	.ndo_vlan_rx_kill_vid	= fm10k_vlan_rx_kill_vid,
 	.ndo_set_rx_mode	= fm10k_set_rx_mode,
 	.ndo_get_stats64	= fm10k_get_stats64,
-	.ndo_setup_tc		= fm10k_setup_tc,
+	.ndo_setup_tc		= __fm10k_setup_tc,
 	.ndo_set_vf_mac		= fm10k_ndo_set_vf_mac,
 	.ndo_set_vf_vlan	= fm10k_ndo_set_vf_vlan,
 	.ndo_set_vf_rate	= fm10k_ndo_set_vf_bw,

@@ -5298,6 +5298,13 @@ exit:
 	return ret;
 }
 
+static int __i40e_setup_tc(struct net_device *netdev, u32 handle, u8 tc)
+{
+	if (handle != TC_H_ROOT)
+		return -EINVAL;
+	return i40e_setup_tc(netdev, tc);
+}
+
 /**
  * i40e_open - Called when a network interface is made active
  * @netdev: network interface device structure

@@ -8887,7 +8894,7 @@ static const struct net_device_ops i40e_netdev_ops = {
 #ifdef CONFIG_NET_POLL_CONTROLLER
 	.ndo_poll_controller	= i40e_netpoll,
 #endif
-	.ndo_setup_tc		= i40e_setup_tc,
+	.ndo_setup_tc		= __i40e_setup_tc,
 #ifdef I40E_FCOE
 	.ndo_fcoe_enable	= i40e_fcoe_enable,
 	.ndo_fcoe_disable	= i40e_fcoe_disable,

@@ -8167,6 +8167,15 @@ int ixgbe_setup_tc(struct net_device *dev, u8 tc)
 	return 0;
 }
 
+int __ixgbe_setup_tc(struct net_device *dev, u32 handle, u8 tc)
+{
+	/* Only support egress tc setup for now */
+	if (handle != TC_H_ROOT)
+		return -EINVAL;
+
+	return ixgbe_setup_tc(dev, tc);
+}
+
 #ifdef CONFIG_PCI_IOV
 void ixgbe_sriov_reinit(struct ixgbe_adapter *adapter)
 {

@@ -8625,7 +8634,7 @@ static const struct net_device_ops ixgbe_netdev_ops = {
 	.ndo_get_vf_config	= ixgbe_ndo_get_vf_config,
 	.ndo_get_stats64	= ixgbe_get_stats64,
 #ifdef CONFIG_IXGBE_DCB
-	.ndo_setup_tc		= ixgbe_setup_tc,
+	.ndo_setup_tc		= __ixgbe_setup_tc,
 #endif
 #ifdef CONFIG_NET_POLL_CONTROLLER
 	.ndo_poll_controller	= ixgbe_netpoll,

@@ -69,6 +69,14 @@ int mlx4_en_setup_tc(struct net_device *dev, u8 up)
 	return 0;
 }
 
+static int __mlx4_en_setup_tc(struct net_device *dev, u32 handle, u8 up)
+{
+	if (handle != TC_H_ROOT)
+		return -EINVAL;
+
+	return mlx4_en_setup_tc(dev, up);
+}
+
 #ifdef CONFIG_RFS_ACCEL
 
 struct mlx4_en_filter {

@@ -2463,7 +2471,7 @@ static const struct net_device_ops mlx4_netdev_ops = {
 #endif
 	.ndo_set_features	= mlx4_en_set_features,
 	.ndo_fix_features	= mlx4_en_fix_features,
-	.ndo_setup_tc		= mlx4_en_setup_tc,
+	.ndo_setup_tc		= __mlx4_en_setup_tc,
 #ifdef CONFIG_RFS_ACCEL
 	.ndo_rx_flow_steer	= mlx4_en_filter_rfs,
 #endif

@@ -2501,7 +2509,7 @@ static const struct net_device_ops mlx4_netdev_ops_master = {
 #endif
 	.ndo_set_features	= mlx4_en_set_features,
 	.ndo_fix_features	= mlx4_en_fix_features,
-	.ndo_setup_tc		= mlx4_en_setup_tc,
+	.ndo_setup_tc		= __mlx4_en_setup_tc,
 #ifdef CONFIG_RFS_ACCEL
 	.ndo_rx_flow_steer	= mlx4_en_filter_rfs,
 #endif

@@ -32,7 +32,7 @@ netdev_tx_t efx_hard_start_xmit(struct sk_buff *skb,
 				struct net_device *net_dev);
 netdev_tx_t efx_enqueue_skb(struct efx_tx_queue *tx_queue, struct sk_buff *skb);
 void efx_xmit_done(struct efx_tx_queue *tx_queue, unsigned int index);
-int efx_setup_tc(struct net_device *net_dev, u8 num_tc);
+int efx_setup_tc(struct net_device *net_dev, u32 handle, u8 num_tc);
 unsigned int efx_tx_max_skb_descs(struct efx_nic *efx);
 extern unsigned int efx_piobuf_size;
 extern bool efx_separate_tx_channels;

@@ -562,7 +562,7 @@ void efx_init_tx_queue_core_txq(struct efx_tx_queue *tx_queue)
 				     efx->n_tx_channels : 0));
 }
 
-int efx_setup_tc(struct net_device *net_dev, u8 num_tc)
+int efx_setup_tc(struct net_device *net_dev, u32 handle, u8 num_tc)
 {
 	struct efx_nic *efx = netdev_priv(net_dev);
 	struct efx_channel *channel;

@@ -570,6 +570,9 @@ int efx_setup_tc(struct net_device *net_dev, u8 num_tc)
 	unsigned tc;
 	int rc;
 
+	if (handle != TC_H_ROOT)
+		return -EINVAL;
+
 	if (efx_nic_rev(efx) < EFX_REV_FALCON_B0 || num_tc > EFX_MAX_TX_TC)
 		return -EINVAL;
 

@@ -1831,13 +1831,16 @@ static u16 netcp_select_queue(struct net_device *dev, struct sk_buff *skb,
 	return 0;
 }
 
-static int netcp_setup_tc(struct net_device *dev, u8 num_tc)
+static int netcp_setup_tc(struct net_device *dev, u32 handle, u8 num_tc)
 {
 	int i;
 
 	/* setup tc must be called under rtnl lock */
 	ASSERT_RTNL();
 
+	if (handle != TC_H_ROOT)
+		return -EINVAL;
+
 	/* Sanity-check the number of traffic classes requested */
 	if ((dev->real_num_tx_queues <= 1) ||
 	    (dev->real_num_tx_queues < num_tc))

@@ -1148,7 +1148,7 @@ struct net_device_ops {
 	int			(*ndo_set_vf_rss_query_en)(
 						   struct net_device *dev,
 						   int vf, bool setting);
-	int			(*ndo_setup_tc)(struct net_device *dev, u8 tc);
+	int			(*ndo_setup_tc)(struct net_device *dev, u32 handle, u8 tc);
 #if IS_ENABLED(CONFIG_FCOE)
 	int			(*ndo_fcoe_enable)(struct net_device *dev);
 	int			(*ndo_fcoe_disable)(struct net_device *dev);

@@ -39,7 +39,7 @@ static void mqprio_destroy(struct Qdisc *sch)
 	}
 
 	if (priv->hw_owned && dev->netdev_ops->ndo_setup_tc)
-		dev->netdev_ops->ndo_setup_tc(dev, 0);
+		dev->netdev_ops->ndo_setup_tc(dev, sch->handle, 0);
 	else
 		netdev_set_num_tc(dev, 0);
 }

@@ -141,7 +141,8 @@ static int mqprio_init(struct Qdisc *sch, struct nlattr *opt)
 	 */
 	if (qopt->hw) {
 		priv->hw_owned = 1;
-		err = dev->netdev_ops->ndo_setup_tc(dev, qopt->num_tc);
+		err = dev->netdev_ops->ndo_setup_tc(dev, sch->handle,
+						    qopt->num_tc);
 		if (err)
 			goto err;
 	} else {

@@ -4268,11 +4268,12 @@ int bnx2x_setup_tc(struct net_device *dev, u8 num_tc)
 	return 0;
 }
 
-int __bnx2x_setup_tc(struct net_device *dev, u32 handle, u8 num_tc)
+int __bnx2x_setup_tc(struct net_device *dev, u32 handle, __be16 proto,
+		     struct tc_to_netdev *tc)
 {
-	if (handle != TC_H_ROOT)
+	if (handle != TC_H_ROOT || tc->type != TC_SETUP_MQPRIO)
 		return -EINVAL;
-	return bnx2x_setup_tc(dev, num_tc);
+	return bnx2x_setup_tc(dev, tc->tc);
 }
 
 /* called with rtnl_lock */

@@ -486,7 +486,8 @@ netdev_tx_t bnx2x_start_xmit(struct sk_buff *skb, struct net_device *dev);
 
 /* setup_tc callback */
 int bnx2x_setup_tc(struct net_device *dev, u8 num_tc);
-int __bnx2x_setup_tc(struct net_device *dev, u32 handle, u8 num_tc);
+int __bnx2x_setup_tc(struct net_device *dev, u32 handle, __be16 proto,
+		     struct tc_to_netdev *tc);
 
 int bnx2x_get_vf_config(struct net_device *dev, int vf,
 			struct ifla_vf_info *ivi);

@@ -5260,13 +5260,17 @@ static int bnxt_change_mtu(struct net_device *dev, int new_mtu)
 	return 0;
 }
 
-static int bnxt_setup_tc(struct net_device *dev, u32 handle, u8 tc)
+static int bnxt_setup_tc(struct net_device *dev, u32 handle, __be16 proto,
+			 struct tc_to_netdev *ntc)
 {
 	struct bnxt *bp = netdev_priv(dev);
+	u8 tc;
 
-	if (handle != TC_H_ROOT)
+	if (handle != TC_H_ROOT || ntc->type != TC_SETUP_MQPRIO)
 		return -EINVAL;
 
+	tc = ntc->tc;
+
 	if (tc > bp->max_tc) {
 		netdev_err(dev, "too many traffic classes requested: %d Max supported is %d\n",
 			   tc, bp->max_tc);

@@ -1190,12 +1190,13 @@ int fm10k_setup_tc(struct net_device *dev, u8 tc)
 	return 0;
 }
 
-static int __fm10k_setup_tc(struct net_device *dev, u32 handle, u8 tc)
+static int __fm10k_setup_tc(struct net_device *dev, u32 handle, __be16 proto,
+			    struct tc_to_netdev *tc)
 {
-	if (handle != TC_H_ROOT)
+	if (handle != TC_H_ROOT || tc->type != TC_SETUP_MQPRIO)
 		return -EINVAL;
 
-	return fm10k_setup_tc(dev, tc);
+	return fm10k_setup_tc(dev, tc->tc);
 }
 
 static int fm10k_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd)

@@ -5298,11 +5298,12 @@ exit:
 	return ret;
 }
 
-static int __i40e_setup_tc(struct net_device *netdev, u32 handle, u8 tc)
+static int __i40e_setup_tc(struct net_device *netdev, u32 handle, __be16 proto,
+			   struct tc_to_netdev *tc)
 {
-	if (handle != TC_H_ROOT)
+	if (handle != TC_H_ROOT || tc->type != TC_SETUP_MQPRIO)
 		return -EINVAL;
-	return i40e_setup_tc(netdev, tc);
+	return i40e_setup_tc(netdev, tc->tc);
 }
 
 /**

@@ -8167,13 +8167,14 @@ int ixgbe_setup_tc(struct net_device *dev, u8 tc)
 	return 0;
 }
 
-int __ixgbe_setup_tc(struct net_device *dev, u32 handle, u8 tc)
+int __ixgbe_setup_tc(struct net_device *dev, u32 handle, __be16 proto,
+		     struct tc_to_netdev *tc)
 {
 	/* Only support egress tc setup for now */
-	if (handle != TC_H_ROOT)
+	if (handle != TC_H_ROOT || tc->type != TC_SETUP_MQPRIO)
 		return -EINVAL;
 
-	return ixgbe_setup_tc(dev, tc);
+	return ixgbe_setup_tc(dev, tc->tc);
 }
 
 #ifdef CONFIG_PCI_IOV

@@ -69,12 +69,13 @@ int mlx4_en_setup_tc(struct net_device *dev, u8 up)
 	return 0;
 }
 
-static int __mlx4_en_setup_tc(struct net_device *dev, u32 handle, u8 up)
+static int __mlx4_en_setup_tc(struct net_device *dev, u32 handle, __be16 proto,
+			      struct tc_to_netdev *tc)
 {
-	if (handle != TC_H_ROOT)
+	if (handle != TC_H_ROOT || tc->type != TC_SETUP_MQPRIO)
 		return -EINVAL;
 
-	return mlx4_en_setup_tc(dev, up);
+	return mlx4_en_setup_tc(dev, tc->tc);
 }
 
 #ifdef CONFIG_RFS_ACCEL

@@ -32,7 +32,8 @@ netdev_tx_t efx_hard_start_xmit(struct sk_buff *skb,
 				struct net_device *net_dev);
 netdev_tx_t efx_enqueue_skb(struct efx_tx_queue *tx_queue, struct sk_buff *skb);
 void efx_xmit_done(struct efx_tx_queue *tx_queue, unsigned int index);
-int efx_setup_tc(struct net_device *net_dev, u32 handle, u8 num_tc);
+int efx_setup_tc(struct net_device *net_dev, u32 handle, __be16 proto,
+		 struct tc_to_netdev *tc);
 unsigned int efx_tx_max_skb_descs(struct efx_nic *efx);
 extern unsigned int efx_piobuf_size;
 extern bool efx_separate_tx_channels;

@@ -562,17 +562,20 @@ void efx_init_tx_queue_core_txq(struct efx_tx_queue *tx_queue)
 				     efx->n_tx_channels : 0));
 }
 
-int efx_setup_tc(struct net_device *net_dev, u32 handle, u8 num_tc)
+int efx_setup_tc(struct net_device *net_dev, u32 handle, __be16 proto,
+		 struct tc_to_netdev *ntc)
 {
 	struct efx_nic *efx = netdev_priv(net_dev);
 	struct efx_channel *channel;
 	struct efx_tx_queue *tx_queue;
-	unsigned tc;
+	unsigned tc, num_tc;
 	int rc;
 
-	if (handle != TC_H_ROOT)
+	if (handle != TC_H_ROOT || ntc->type != TC_SETUP_MQPRIO)
 		return -EINVAL;
 
+	num_tc = ntc->tc;
+
 	if (efx_nic_rev(efx) < EFX_REV_FALCON_B0 || num_tc > EFX_MAX_TX_TC)
 		return -EINVAL;
 

@@ -1831,25 +1831,26 @@ static u16 netcp_select_queue(struct net_device *dev, struct sk_buff *skb,
 	return 0;
 }
 
-static int netcp_setup_tc(struct net_device *dev, u32 handle, u8 num_tc)
+static int netcp_setup_tc(struct net_device *dev, u32 handle, __be16 proto,
+			  struct tc_to_netdev tc)
 {
 	int i;
 
 	/* setup tc must be called under rtnl lock */
 	ASSERT_RTNL();
 
-	if (handle != TC_H_ROOT)
+	if (handle != TC_H_ROOT || tc->type != TC_SETUP_MQPRIO)
 		return -EINVAL;
 
 	/* Sanity-check the number of traffic classes requested */
 	if ((dev->real_num_tx_queues <= 1) ||
-	    (dev->real_num_tx_queues < num_tc))
+	    (dev->real_num_tx_queues < tc->tc))
 		return -EINVAL;
 
 	/* Configure traffic class to queue mappings */
-	if (num_tc) {
-		netdev_set_num_tc(dev, num_tc);
-		for (i = 0; i < num_tc; i++)
+	if (tc->tc) {
+		netdev_set_num_tc(dev, tc->tc);
+		for (i = 0; i < tc->tc; i++)
 			netdev_set_tc_queue(dev, i, 1, i);
 	} else {
 		netdev_reset_tc(dev);

@@ -779,6 +779,21 @@ static inline bool netdev_phys_item_id_same(struct netdev_phys_item_id *a,
 typedef u16 (*select_queue_fallback_t)(struct net_device *dev,
 				       struct sk_buff *skb);
 
+/* This structure holds attributes of qdisc and classifiers
+ * that are being passed to the netdevice through the setup_tc op.
+ */
+enum {
+	TC_SETUP_MQPRIO,
+};
+
+struct tc_to_netdev {
+	unsigned int type;
+	union {
+		u8 tc;
+	};
+};
+
+
 /*
  * This structure defines the management hooks for network devices.
  * The following hooks can be defined; unless noted otherwise, they are

@@ -1148,7 +1163,10 @@ struct net_device_ops {
 	int			(*ndo_set_vf_rss_query_en)(
 						   struct net_device *dev,
 						   int vf, bool setting);
-	int			(*ndo_setup_tc)(struct net_device *dev, u32 handle, u8 tc);
+	int			(*ndo_setup_tc)(struct net_device *dev,
+						u32 handle,
+						__be16 protocol,
+						struct tc_to_netdev *tc);
 #if IS_ENABLED(CONFIG_FCOE)
 	int			(*ndo_fcoe_enable)(struct net_device *dev);
 	int			(*ndo_fcoe_disable)(struct net_device *dev);

@@ -28,6 +28,7 @@ static void mqprio_destroy(struct Qdisc *sch)
 {
 	struct net_device *dev = qdisc_dev(sch);
 	struct mqprio_sched *priv = qdisc_priv(sch);
+	struct tc_to_netdev tc = {.type = TC_SETUP_MQPRIO};
 	unsigned int ntx;
 
 	if (priv->qdiscs) {

@@ -39,7 +40,7 @@ static void mqprio_destroy(struct Qdisc *sch)
 	}
 
 	if (priv->hw_owned && dev->netdev_ops->ndo_setup_tc)
-		dev->netdev_ops->ndo_setup_tc(dev, sch->handle, 0);
+		dev->netdev_ops->ndo_setup_tc(dev, sch->handle, 0, &tc);
 	else
 		netdev_set_num_tc(dev, 0);
 }

@@ -140,9 +141,11 @@ static int mqprio_init(struct Qdisc *sch, struct nlattr *opt)
 	 * supplied and verified mapping
 	 */
 	if (qopt->hw) {
+		struct tc_to_netdev tc = {.type = TC_SETUP_MQPRIO,
+					  .tc = qopt->num_tc};
+
 		priv->hw_owned = 1;
-		err = dev->netdev_ops->ndo_setup_tc(dev, sch->handle,
-						    qopt->num_tc);
+		err = dev->netdev_ops->ndo_setup_tc(dev, sch->handle, 0, &tc);
 		if (err)
 			goto err;
 	} else {

@@ -779,17 +779,21 @@ static inline bool netdev_phys_item_id_same(struct netdev_phys_item_id *a,
 typedef u16 (*select_queue_fallback_t)(struct net_device *dev,
 				       struct sk_buff *skb);
 
-/* This structure holds attributes of qdisc and classifiers
+/* These structures hold the attributes of qdisc and classifiers
  * that are being passed to the netdevice through the setup_tc op.
  */
 enum {
 	TC_SETUP_MQPRIO,
+	TC_SETUP_CLSU32,
 };
 
+struct tc_cls_u32_offload;
+
 struct tc_to_netdev {
 	unsigned int type;
 	union {
 		u8 tc;
+		struct tc_cls_u32_offload *cls_u32;
 	};
 };
 

@@ -358,4 +358,37 @@ tcf_match_indev(struct sk_buff *skb, int ifindex)
 }
 #endif /* CONFIG_NET_CLS_IND */
 
+struct tc_cls_u32_knode {
+	struct tcf_exts *exts;
+	u8 fshift;
+	u32 handle;
+	u32 val;
+	u32 mask;
+	u32 link_handle;
+	struct tc_u32_sel *sel;
+};
+
+struct tc_cls_u32_hnode {
+	u32 handle;
+	u32 prio;
+	unsigned int divisor;
+};
+
+enum {
+	TC_CLSU32_NEW_KNODE,
+	TC_CLSU32_REPLACE_KNODE,
+	TC_CLSU32_DELETE_KNODE,
+	TC_CLSU32_NEW_HNODE,
+	TC_CLSU32_REPLACE_HNODE,
+};
+
+struct tc_cls_u32_offload {
+	/* knode values */
+	int command;
+	union {
+		struct tc_cls_u32_knode knode;
+		struct tc_cls_u32_hnode hnode;
+	};
+};
+
 #endif

@@ -43,6 +43,7 @@
 #include <net/netlink.h>
 #include <net/act_api.h>
 #include <net/pkt_cls.h>
+#include <linux/netdevice.h>
 
 struct tc_u_knode {
 	struct tc_u_knode __rcu	*next;

@@ -424,6 +425,68 @@ static int u32_delete_key(struct tcf_proto *tp, struct tc_u_knode *key)
 	return 0;
 }
 
+static void u32_remove_hw_knode(struct tcf_proto *tp, u32 handle)
+{
+	struct net_device *dev = tp->q->dev_queue->dev;
+	struct tc_cls_u32_offload u32_offload = {0};
+	struct tc_to_netdev offload;
+
+	offload.type = TC_SETUP_CLSU32;
+	offload.cls_u32 = &u32_offload;
+
+	if (dev->netdev_ops->ndo_setup_tc) {
+		offload.cls_u32->command = TC_CLSU32_DELETE_KNODE;
+		offload.cls_u32->knode.handle = handle;
+		dev->netdev_ops->ndo_setup_tc(dev, tp->q->handle,
+					      tp->protocol, &offload);
+	}
+}
+
+static void u32_replace_hw_hnode(struct tcf_proto *tp, struct tc_u_hnode *h)
+{
+	struct net_device *dev = tp->q->dev_queue->dev;
+	struct tc_cls_u32_offload u32_offload = {0};
+	struct tc_to_netdev offload;
+
+	offload.type = TC_SETUP_CLSU32;
+	offload.cls_u32 = &u32_offload;
+
+	if (dev->netdev_ops->ndo_setup_tc) {
+		offload.cls_u32->command = TC_CLSU32_NEW_HNODE;
+		offload.cls_u32->hnode.divisor = h->divisor;
+		offload.cls_u32->hnode.handle = h->handle;
+		offload.cls_u32->hnode.prio = h->prio;
+
+		dev->netdev_ops->ndo_setup_tc(dev, tp->q->handle,
+					      tp->protocol, &offload);
+	}
+}
+
+static void u32_replace_hw_knode(struct tcf_proto *tp, struct tc_u_knode *n)
+{
+	struct net_device *dev = tp->q->dev_queue->dev;
+	struct tc_cls_u32_offload u32_offload = {0};
+	struct tc_to_netdev offload;
+
+	offload.type = TC_SETUP_CLSU32;
+	offload.cls_u32 = &u32_offload;
+
+	if (dev->netdev_ops->ndo_setup_tc) {
+		offload.cls_u32->command = TC_CLSU32_REPLACE_KNODE;
+		offload.cls_u32->knode.handle = n->handle;
+		offload.cls_u32->knode.fshift = n->fshift;
+		offload.cls_u32->knode.val = n->val;
+		offload.cls_u32->knode.mask = n->mask;
+		offload.cls_u32->knode.sel = &n->sel;
+		offload.cls_u32->knode.exts = &n->exts;
+		if (n->ht_down)
+			offload.cls_u32->knode.link_handle = n->ht_down->handle;
+
+		dev->netdev_ops->ndo_setup_tc(dev, tp->q->handle,
+					      tp->protocol, &offload);
+	}
+}
+
 static void u32_clear_hnode(struct tcf_proto *tp, struct tc_u_hnode *ht)
 {
 	struct tc_u_knode *n;

@@ -434,6 +497,7 @@ static void u32_clear_hnode(struct tcf_proto *tp, struct tc_u_hnode *ht)
 			RCU_INIT_POINTER(ht->ht[h],
 					 rtnl_dereference(n->next));
 			tcf_unbind_filter(tp, &n->res);
+			u32_remove_hw_knode(tp, n->handle);
 			call_rcu(&n->rcu, u32_delete_key_freepf_rcu);
 		}
 	}

@@ -540,8 +604,10 @@ static int u32_delete(struct tcf_proto *tp, unsigned long arg)
 	if (ht == NULL)
 		return 0;
 
-	if (TC_U32_KEY(ht->handle))
+	if (TC_U32_KEY(ht->handle)) {
+		u32_remove_hw_knode(tp, ht->handle);
 		return u32_delete_key(tp, (struct tc_u_knode *)ht);
+	}
 
 	if (root_ht == ht)
 		return -EINVAL;

@@ -769,6 +835,7 @@ static int u32_change(struct net *net, struct sk_buff *in_skb,
 		u32_replace_knode(tp, tp_c, new);
 		tcf_unbind_filter(tp, &n->res);
 		call_rcu(&n->rcu, u32_delete_key_rcu);
+		u32_replace_hw_knode(tp, new);
 		return 0;
 	}
 

@@ -795,6 +862,8 @@ static int u32_change(struct net *net, struct sk_buff *in_skb,
 		RCU_INIT_POINTER(ht->next, tp_c->hlist);
 		rcu_assign_pointer(tp_c->hlist, ht);
 		*arg = (unsigned long)ht;
+
+		u32_replace_hw_hnode(tp, ht);
 		return 0;
 	}
 

@@ -877,7 +946,7 @@ static int u32_change(struct net *net, struct sk_buff *in_skb,
 
 		RCU_INIT_POINTER(n->next, pins);
 		rcu_assign_pointer(*ins, n);
-
+		u32_replace_hw_knode(tp, n);
 		*arg = (unsigned long)n;
 		return 0;
 	}

@@ -67,6 +67,8 @@ enum {
 	NETIF_F_HW_L2FW_DOFFLOAD_BIT,	/* Allow L2 Forwarding in Hardware */
 	NETIF_F_BUSY_POLL_BIT,		/* Busy poll */
 
+	NETIF_F_HW_TC_BIT,		/* Offload TC infrastructure */
+
 	/*
 	 * Add your fresh new feature above and remember to update
 	 * netdev_features_strings[] in net/core/ethtool.c and maybe

@@ -124,6 +126,7 @@ enum {
 #define NETIF_F_HW_VLAN_STAG_TX	__NETIF_F(HW_VLAN_STAG_TX)
 #define NETIF_F_HW_L2FW_DOFFLOAD	__NETIF_F(HW_L2FW_DOFFLOAD)
 #define NETIF_F_BUSY_POLL	__NETIF_F(BUSY_POLL)
+#define NETIF_F_HW_TC		__NETIF_F(HW_TC)
 
 #define for_each_netdev_feature(mask_addr, bit)	\
 	for_each_set_bit(bit, (unsigned long *)mask_addr, NETDEV_FEATURE_COUNT)

@@ -98,6 +98,7 @@ static const char netdev_features_strings[NETDEV_FEATURE_COUNT][ETH_GSTRING_LEN]
 	[NETIF_F_RXALL_BIT] =            "rx-all",
 	[NETIF_F_HW_L2FW_DOFFLOAD_BIT] = "l2-fwd-offload",
 	[NETIF_F_BUSY_POLL_BIT] =        "busy-poll",
+	[NETIF_F_HW_TC_BIT] =		 "hw-tc-offload",
 };
 
 static const char

@@ -2,6 +2,7 @@
 #define __NET_TC_GACT_H
 
 #include <net/act_api.h>
+#include <linux/tc_act/tc_gact.h>
 
 struct tcf_gact {
 	struct tcf_common	common;

@@ -15,4 +16,17 @@ struct tcf_gact {
 #define to_gact(a) \
 	container_of(a->priv, struct tcf_gact, common)
 
+static inline bool is_tcf_gact_dropped(const struct tc_action *a)
+{
+	struct tcf_gact *gact;
+
+	if (a->ops && a->ops->type != TCA_ACT_GACT)
+		return false;
+
+	gact = a->priv;
+	if (gact->tcf_action == TC_ACT_SHOT)
+		return true;
+
+	return false;
+}
 #endif /* __NET_TC_GACT_H */

@@ -0,0 +1,110 @@
+/*******************************************************************************
+
+  Intel 10 Gigabit PCI Express Linux driver
+  Copyright(c) 2016 Intel Corporation.
+
+  This program is free software; you can redistribute it and/or modify it
+  under the terms and conditions of the GNU General Public License,
+  version 2, as published by the Free Software Foundation.
+
+  This program is distributed in the hope it will be useful, but WITHOUT
+  ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+  FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+  more details.
+
+  You should have received a copy of the GNU General Public License along with
+  this program; if not, write to the Free Software Foundation, Inc.,
+  51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+
+  The full GNU General Public License is included in this distribution in
+  the file called "COPYING".
+
+  Contact Information:
+  Linux NICS <linux.nics@intel.com>
+  e1000-devel Mailing List <e1000-devel@lists.sourceforge.net>
+  Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497
+
+*******************************************************************************/
+
+#ifndef _IXGBE_MODEL_H_
+#define _IXGBE_MODEL_H_
+
+#include "ixgbe.h"
+#include "ixgbe_type.h"
+
+struct ixgbe_mat_field {
+	unsigned int off;
+	unsigned int mask;
+	int (*val)(struct ixgbe_fdir_filter *input,
+		   union ixgbe_atr_input *mask,
+		   __u32 val, __u32 m);
+	int link;
+};
+
+static inline int ixgbe_mat_prgm_sip(struct ixgbe_fdir_filter *input,
+				     union ixgbe_atr_input *mask,
+				     __u32 val, __u32 m)
+{
+	input->filter.formatted.src_ip[0] = val;
+	mask->formatted.src_ip[0] = m;
+	return 0;
+}
+
+static inline int ixgbe_mat_prgm_dip(struct ixgbe_fdir_filter *input,
+				     union ixgbe_atr_input *mask,
+				     __u32 val, __u32 m)
+{
+	input->filter.formatted.dst_ip[0] = val;
+	mask->formatted.dst_ip[0] = m;
+	return 0;
+}
+
+static struct ixgbe_mat_field ixgbe_ipv4_fields[] = {
+	{ .off = 12, .mask = -1, .val = ixgbe_mat_prgm_sip, .link = 0},
+	{ .off = 16, .mask = -1, .val = ixgbe_mat_prgm_dip, .link = 0},
+	{ .val = NULL } /* terminal node */
+};
+
+static inline int ixgbe_mat_prgm_sport(struct ixgbe_fdir_filter *input,
+				       union ixgbe_atr_input *mask,
+				       __u32 val, __u32 m)
+{
+	input->filter.formatted.src_port = val & 0xffff;
+	mask->formatted.src_port = m & 0xffff;
+	return 0;
+};
+
+static inline int ixgbe_mat_prgm_dport(struct ixgbe_fdir_filter *input,
+				       union ixgbe_atr_input *mask,
+				       __u32 val, __u32 m)
+{
+	input->filter.formatted.dst_port = val & 0xffff;
+	mask->formatted.dst_port = m & 0xffff;
+	return 0;
+};
+
+static struct ixgbe_mat_field ixgbe_tcp_fields[] = {
+	{.off = 0, .mask = 0xffff, .val = ixgbe_mat_prgm_sport, .link = -1},
+	{.off = 2, .mask = 0xffff, .val = ixgbe_mat_prgm_dport, .link = -1},
+	{ .val = NULL } /* terminal node */
+};
+
+struct ixgbe_nexthdr {
+	/* offset, shift, and mask of position to next header */
+	unsigned int o;
+	__u32 s;
+	__u32 m;
+	/* match criteria to make this jump*/
+	unsigned int off;
+	__u32 val;
+	__u32 mask;
+	/* location of jump to make */
+	struct ixgbe_mat_field *jump;
+};
+
+static struct ixgbe_nexthdr ixgbe_ipv4_jumps[] = {
+	{ .o = 0, .s = 6, .m = 0xf,
+	  .off = 8, .val = 0x600, .mask = 0xff00, .jump = ixgbe_tcp_fields},
+	{ .jump = NULL } /* terminal node */
+};
+#endif /* _IXGBE_MODEL_H_ */

@@ -925,6 +925,9 @@ s32 ixgbe_fdir_erase_perfect_filter_82599(struct ixgbe_hw *hw,
 					  u16 soft_id);
 void ixgbe_atr_compute_perfect_hash_82599(union ixgbe_atr_input *input,
 					  union ixgbe_atr_input *mask);
+int ixgbe_update_ethtool_fdir_entry(struct ixgbe_adapter *adapter,
+				    struct ixgbe_fdir_filter *input,
+				    u16 sw_idx);
 void ixgbe_set_rx_mode(struct net_device *netdev);
 #ifdef CONFIG_IXGBE_DCB
 void ixgbe_set_rx_drop_en(struct ixgbe_adapter *adapter);

@@ -2490,9 +2490,9 @@ static int ixgbe_get_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd,
 	return ret;
 }
 
-static int ixgbe_update_ethtool_fdir_entry(struct ixgbe_adapter *adapter,
-					   struct ixgbe_fdir_filter *input,
-					   u16 sw_idx)
+int ixgbe_update_ethtool_fdir_entry(struct ixgbe_adapter *adapter,
+				    struct ixgbe_fdir_filter *input,
+				    u16 sw_idx)
 {
 	struct ixgbe_hw *hw = &adapter->hw;
 	struct hlist_node *node2;

@@ -51,6 +51,7 @@
 #include <linux/prefetch.h>
 #include <scsi/fc/fc_fcoe.h>
 #include <net/vxlan.h>
+#include <net/pkt_cls.h>
 
 #ifdef CONFIG_OF
 #include <linux/of_net.h>

@@ -8167,10 +8168,189 @@ int ixgbe_setup_tc(struct net_device *dev, u8 tc)
 	return 0;
 }
 
+#include <net/tc_act/tc_gact.h>
+#include "ixgbe_model.h"
+static int ixgbe_delete_clsu32(struct ixgbe_adapter *adapter,
+			       struct tc_cls_u32_offload *cls)
+{
+	int err;
+
+	spin_lock(&adapter->fdir_perfect_lock);
+	err = ixgbe_update_ethtool_fdir_entry(adapter, NULL, cls->knode.handle);
+	spin_unlock(&adapter->fdir_perfect_lock);
+	return err;
+}
+
+#define IXGBE_MAX_LINK_HANDLE 10
+static struct ixgbe_mat_fieldi *
+ixgbe_jump_tables[IXGBE_MAX_LINK_HANDLE] = {ixgbe_ipv4_fields,};
+
+static int ixgbe_configure_clsu32(struct ixgbe_adapter *adapter,
+				  __be16 protocol,
+				  struct tc_cls_u32_offload *cls)
+{
+	u32 loc = cls->knode.handle & 0xfffff;
+	struct ixgbe_hw *hw = &adapter->hw;
+	struct ixgbe_mat_field *field_ptr;
+	struct ixgbe_fdir_filter *input;
+	union ixgbe_atr_input mask;
+	const struct tc_action *a;
+	int i, err = 0;
+	u8 queue;
+	u32 handle;
+
+	memset(&mask, 0, sizeof(union ixgbe_atr_input));
+	handle = cls->knode.handle;
+
+	/* At the moment cls_u32 jumps to transport layer and skips past
+	 * L2 headers. The canonical method to match L2 frames is to use
+	 * negative values. However this is error prone at best but really
+	 * just broken because there is no way to "know" what sort of hdr
+	 * is in front of the transport layer. Fix cls_u32 to support L2
+	 * headers when needed.
+	 */
+	if (protocol != htons(ETH_P_IP))
+		return -EINVAL;
+
+	if (cls->knode.link_handle ||
+	    cls->knode.link_handle >= IXGBE_MAX_LINK_HANDLE) {
+		struct ixgbe_nexthdr *nexthdr = ixgbe_ipv4_jumps;
+		u32 uhtid = TC_U32_USERHTID(cls->knode.link_handle);
+
+		for (i = 0; nexthdr[i].jump; i++) {
+			if (nexthdr->o != cls->knode.sel->offoff ||
+			    nexthdr->s != cls->knode.sel->offshift ||
+			    nexthdr->m != cls->knode.sel->offmask ||
+			    /* do not support multiple key jumps its just mad */
+			    cls->knode.sel->nkeys > 1)
+				return -EINVAL;
+
+			if (nexthdr->off != cls->knode.sel->keys[0].off ||
+			    nexthdr->val != cls->knode.sel->keys[0].val ||
+			    nexthdr->mask != cls->knode.sel->keys[0].mask)
+				return -EINVAL;
+
+			if (uhtid >= IXGBE_MAX_LINK_HANDLE)
+				return -EINVAL;
+
+			ixgbe_jump_tables[uhtid] = nexthdr->jump;
+		}
+		return 0;
+	}
+
+	if (loc >= ((1024 << adapter->fdir_pballoc) - 2)) {
+		e_err(drv, "Location out of range\n");
+		return -EINVAL;
+	}
+
+	/* cls u32 is a graph starting at root node 0x800. The driver tracks
+	 * links and also the fields used to advance the parser across each
+	 * link (e.g. nexthdr/eat parameters from 'tc'). This way we can map
+	 * the u32 graph onto the hardware parse graph denoted in ixgbe_model.h
+	 * To add support for new nodes update ixgbe_model.h parse structures
+	 * this function _should_ be generic try not to hardcode values here.
+	 */
+	if (TC_U32_USERHTID(handle) == 0x800) {
+		field_ptr = ixgbe_jump_tables[0];
+	} else {
+		if (TC_U32_USERHTID(handle) >= ARRAY_SIZE(ixgbe_jump_tables))
+			return -EINVAL;
+
+		field_ptr = ixgbe_jump_tables[TC_U32_USERHTID(handle)];
+	}
+
+	if (!field_ptr)
+		return -EINVAL;
+
+	input = kzalloc(sizeof(*input), GFP_ATOMIC);
+	if (!input)
+		return -ENOMEM;
+
+	for (i = 0; i < cls->knode.sel->nkeys; i++) {
+		int off = cls->knode.sel->keys[i].off;
+		__be32 val = cls->knode.sel->keys[i].val;
+		__be32 m = cls->knode.sel->keys[i].mask;
+		bool found_entry = false;
+
+		mask.formatted.flow_type = IXGBE_ATR_L4TYPE_IPV6_MASK |
+					   IXGBE_ATR_L4TYPE_MASK;
+
+		for (i = 0; field_ptr[i].val; i++) {
+			if (field_ptr[i].off == off &&
+			    field_ptr[i].mask == m) {
+				field_ptr[i].val(input, &mask, val, m);
+				found_entry = true;
+				break;
+			}
+		}
+
+		if (!found_entry)
+			goto err_out;
+	}
+
+	if (list_empty(&cls->knode.exts->actions))
+		goto err_out;
+
+	list_for_each_entry(a, &cls->knode.exts->actions, list) {
+		if (!is_tcf_gact_dropped(a))
+			goto err_out;
+	}
+
+	input->action = IXGBE_FDIR_DROP_QUEUE;
+	queue = IXGBE_FDIR_DROP_QUEUE;
+	input->sw_idx = loc;
+
+	spin_lock(&adapter->fdir_perfect_lock);
+
+	if (hlist_empty(&adapter->fdir_filter_list)) {
+		memcpy(&adapter->fdir_mask, &mask, sizeof(mask));
+		err = ixgbe_fdir_set_input_mask_82599(hw, &mask);
+		if (err)
+			goto err_out_w_lock;
+	} else if (memcmp(&adapter->fdir_mask, &mask, sizeof(mask))) {
+		err = -EINVAL;
+		goto err_out_w_lock;
+	}
+
+	ixgbe_atr_compute_perfect_hash_82599(&input->filter, &mask);
+	err = ixgbe_fdir_write_perfect_filter_82599(hw, &input->filter,
+						    input->sw_idx, queue);
+	if (!err)
+		ixgbe_update_ethtool_fdir_entry(adapter, input, input->sw_idx);
+	spin_unlock(&adapter->fdir_perfect_lock);
+
+	return err;
+err_out_w_lock:
+	spin_unlock(&adapter->fdir_perfect_lock);
+err_out:
+	kfree(input);
+	return -EINVAL;
+}
+
 int __ixgbe_setup_tc(struct net_device *dev, u32 handle, __be16 proto,
 		     struct tc_to_netdev *tc)
 {
-	/* Only support egress tc setup for now */
+	struct ixgbe_adapter *adapter = netdev_priv(dev);
+
+	if (TC_H_MAJ(handle) == TC_H_MAJ(TC_H_INGRESS) &&
+	    tc->type == TC_SETUP_CLSU32) {
+		struct tc_cls_u32_offload *cls = tc->cls_u32;
+
+		if (!(dev->hw_features & NETIF_F_HW_TC))
+			return -EINVAL;
+
+		switch (tc->cls_u32->command) {
+		case TC_CLSU32_NEW_KNODE:
+		case TC_CLSU32_REPLACE_KNODE:
+			return ixgbe_configure_clsu32(adapter,
+						      proto, tc->cls_u32);
+		case TC_CLSU32_DELETE_KNODE:
+			return ixgbe_delete_clsu32(adapter, tc->cls_u32);
+		default:
+			return -EINVAL;
+		}
+	}
+
 	if (handle != TC_H_ROOT || tc->type != TC_SETUP_MQPRIO)
 		return -EINVAL;