dpaa2-switch: add tc flower hardware offload on ingress traffic

This patch adds support for tc flower hardware offload on the ingress
path. Shared filter blocks are supported by sharing a single ACL table
between multiple ports.

The following flow keys are supported:
 - Ethernet: dst_mac/src_mac
 - IPv4: dst_ip/src_ip/ip_proto/tos
 - VLAN: vlan_id/vlan_prio/vlan_tpid/vlan_dei
 - L4: dst_port/src_port

As per flow actions, the following are supported:
 - drop
 - mirred egress redirect
 - trap
Each ACL entry (filter) can be setup with only one of the listed
actions.

A sorted single linked list is used to keep the ACL entries by their
order of priority. When adding a new filter, this enables us to quickly
ascertain if the new entry has the highest priority of the entire block
or if we should make some space in the ACL table by increasing the
priority of the filters already in the table.

Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 248 insertions, 13 deletions

diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-switch.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-switch.c
index fc9e2eb0ad11..5080788c692b 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-switch.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-switch.c
@@ -14,6 +14,7 @@
 #include <linux/kthread.h>
 #include <linux/workqueue.h>
 #include <linux/iommu.h>
+#include <net/pkt_cls.h>
 
 #include <linux/fsl/mc.h>
 
@@ -1125,6 +1126,243 @@ err_exit:
 	return NETDEV_TX_OK;
 }
 
+static int
+dpaa2_switch_setup_tc_cls_flower(struct dpaa2_switch_acl_tbl *acl_tbl,
+				 struct flow_cls_offload *f)
+{
+	switch (f->command) {
+	case FLOW_CLS_REPLACE:
+		return dpaa2_switch_cls_flower_replace(acl_tbl, f);
+	case FLOW_CLS_DESTROY:
+		return dpaa2_switch_cls_flower_destroy(acl_tbl, f);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static int dpaa2_switch_port_setup_tc_block_cb_ig(enum tc_setup_type type,
+						  void *type_data,
+						  void *cb_priv)
+{
+	switch (type) {
+	case TC_SETUP_CLSFLOWER:
+		return dpaa2_switch_setup_tc_cls_flower(cb_priv, type_data);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static LIST_HEAD(dpaa2_switch_block_cb_list);
+
+static int dpaa2_switch_port_acl_tbl_bind(struct ethsw_port_priv *port_priv,
+					  struct dpaa2_switch_acl_tbl *acl_tbl)
+{
+	struct ethsw_core *ethsw = port_priv->ethsw_data;
+	struct net_device *netdev = port_priv->netdev;
+	struct dpsw_acl_if_cfg acl_if_cfg;
+	int err;
+
+	if (port_priv->acl_tbl)
+		return -EINVAL;
+
+	acl_if_cfg.if_id[0] = port_priv->idx;
+	acl_if_cfg.num_ifs = 1;
+	err = dpsw_acl_add_if(ethsw->mc_io, 0, ethsw->dpsw_handle,
+			      acl_tbl->id, &acl_if_cfg);
+	if (err) {
+		netdev_err(netdev, "dpsw_acl_add_if err %d\n", err);
+		return err;
+	}
+
+	acl_tbl->ports |= BIT(port_priv->idx);
+	port_priv->acl_tbl = acl_tbl;
+
+	return 0;
+}
+
+static int
+dpaa2_switch_port_acl_tbl_unbind(struct ethsw_port_priv *port_priv,
+				 struct dpaa2_switch_acl_tbl *acl_tbl)
+{
+	struct ethsw_core *ethsw = port_priv->ethsw_data;
+	struct net_device *netdev = port_priv->netdev;
+	struct dpsw_acl_if_cfg acl_if_cfg;
+	int err;
+
+	if (port_priv->acl_tbl != acl_tbl)
+		return -EINVAL;
+
+	acl_if_cfg.if_id[0] = port_priv->idx;
+	acl_if_cfg.num_ifs = 1;
+	err = dpsw_acl_remove_if(ethsw->mc_io, 0, ethsw->dpsw_handle,
+				 acl_tbl->id, &acl_if_cfg);
+	if (err) {
+		netdev_err(netdev, "dpsw_acl_add_if err %d\n", err);
+		return err;
+	}
+
+	acl_tbl->ports &= ~BIT(port_priv->idx);
+	port_priv->acl_tbl = NULL;
+	return 0;
+}
+
+static int dpaa2_switch_port_block_bind(struct ethsw_port_priv *port_priv,
+					struct dpaa2_switch_acl_tbl *acl_tbl)
+{
+	struct dpaa2_switch_acl_tbl *old_acl_tbl = port_priv->acl_tbl;
+	int err;
+
+	/* If the port is already bound to this ACL table then do nothing. This
+	 * can happen when this port is the first one to join a tc block
+	 */
+	if (port_priv->acl_tbl == acl_tbl)
+		return 0;
+
+	err = dpaa2_switch_port_acl_tbl_unbind(port_priv, old_acl_tbl);
+	if (err)
+		return err;
+
+	/* Mark the previous ACL table as being unused if this was the last
+	 * port that was using it.
+	 */
+	if (old_acl_tbl->ports == 0)
+		old_acl_tbl->in_use = false;
+
+	return dpaa2_switch_port_acl_tbl_bind(port_priv, acl_tbl);
+}
+
+static int dpaa2_switch_port_block_unbind(struct ethsw_port_priv *port_priv,
+					  struct dpaa2_switch_acl_tbl *acl_tbl)
+{
+	struct ethsw_core *ethsw = port_priv->ethsw_data;
+	struct dpaa2_switch_acl_tbl *new_acl_tbl;
+	int err;
+
+	/* We are the last port that leaves a block (an ACL table).
+	 * We'll continue to use this table.
+	 */
+	if (acl_tbl->ports == BIT(port_priv->idx))
+		return 0;
+
+	err = dpaa2_switch_port_acl_tbl_unbind(port_priv, acl_tbl);
+	if (err)
+		return err;
+
+	if (acl_tbl->ports == 0)
+		acl_tbl->in_use = false;
+
+	new_acl_tbl = dpaa2_switch_acl_tbl_get_unused(ethsw);
+	new_acl_tbl->in_use = true;
+	return dpaa2_switch_port_acl_tbl_bind(port_priv, new_acl_tbl);
+}
+
+static int dpaa2_switch_setup_tc_block_bind(struct net_device *netdev,
+					    struct flow_block_offload *f)
+{
+	struct ethsw_port_priv *port_priv = netdev_priv(netdev);
+	struct ethsw_core *ethsw = port_priv->ethsw_data;
+	struct dpaa2_switch_acl_tbl *acl_tbl;
+	struct flow_block_cb *block_cb;
+	bool register_block = false;
+	int err;
+
+	block_cb = flow_block_cb_lookup(f->block,
+					dpaa2_switch_port_setup_tc_block_cb_ig,
+					ethsw);
+
+	if (!block_cb) {
+		/* If the ACL table is not already known, then this port must
+		 * be the first to join it. In this case, we can just continue
+		 * to use our private table
+		 */
+		acl_tbl = port_priv->acl_tbl;
+
+		block_cb = flow_block_cb_alloc(dpaa2_switch_port_setup_tc_block_cb_ig,
+					       ethsw, acl_tbl, NULL);
+		if (IS_ERR(block_cb))
+			return PTR_ERR(block_cb);
+
+		register_block = true;
+	} else {
+		acl_tbl = flow_block_cb_priv(block_cb);
+	}
+
+	flow_block_cb_incref(block_cb);
+	err = dpaa2_switch_port_block_bind(port_priv, acl_tbl);
+	if (err)
+		goto err_block_bind;
+
+	if (register_block) {
+		flow_block_cb_add(block_cb, f);
+		list_add_tail(&block_cb->driver_list,
+			      &dpaa2_switch_block_cb_list);
+	}
+
+	return 0;
+
+err_block_bind:
+	if (!flow_block_cb_decref(block_cb))
+		flow_block_cb_free(block_cb);
+	return err;
+}
+
+static void dpaa2_switch_setup_tc_block_unbind(struct net_device *netdev,
+					       struct flow_block_offload *f)
+{
+	struct ethsw_port_priv *port_priv = netdev_priv(netdev);
+	struct ethsw_core *ethsw = port_priv->ethsw_data;
+	struct dpaa2_switch_acl_tbl *acl_tbl;
+	struct flow_block_cb *block_cb;
+	int err;
+
+	block_cb = flow_block_cb_lookup(f->block,
+					dpaa2_switch_port_setup_tc_block_cb_ig,
+					ethsw);
+	if (!block_cb)
+		return;
+
+	acl_tbl = flow_block_cb_priv(block_cb);
+	err = dpaa2_switch_port_block_unbind(port_priv, acl_tbl);
+	if (!err && !flow_block_cb_decref(block_cb)) {
+		flow_block_cb_remove(block_cb, f);
+		list_del(&block_cb->driver_list);
+	}
+}
+
+static int dpaa2_switch_setup_tc_block(struct net_device *netdev,
+				       struct flow_block_offload *f)
+{
+	if (f->binder_type != FLOW_BLOCK_BINDER_TYPE_CLSACT_INGRESS)
+		return -EOPNOTSUPP;
+
+	f->driver_block_list = &dpaa2_switch_block_cb_list;
+
+	switch (f->command) {
+	case FLOW_BLOCK_BIND:
+		return dpaa2_switch_setup_tc_block_bind(netdev, f);
+	case FLOW_BLOCK_UNBIND:
+		dpaa2_switch_setup_tc_block_unbind(netdev, f);
+		return 0;
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static int dpaa2_switch_port_setup_tc(struct net_device *netdev,
+				      enum tc_setup_type type,
+				      void *type_data)
+{
+	switch (type) {
+	case TC_SETUP_BLOCK: {
+		return dpaa2_switch_setup_tc_block(netdev, type_data);
+	}
+	default:
+		return -EOPNOTSUPP;
+	}
+
+	return 0;
+}
+
 static const struct net_device_ops dpaa2_switch_port_ops = {
 	.ndo_open		= dpaa2_switch_port_open,
 	.ndo_stop		= dpaa2_switch_port_stop,
@@ -1141,6 +1379,7 @@ static const struct net_device_ops dpaa2_switch_port_ops = {
 	.ndo_start_xmit		= dpaa2_switch_port_tx,
 	.ndo_get_port_parent_id	= dpaa2_switch_port_parent_id,
 	.ndo_get_phys_port_name = dpaa2_switch_port_get_phys_name,
+	.ndo_setup_tc		= dpaa2_switch_port_setup_tc,
 };
 
 bool dpaa2_switch_port_dev_check(const struct net_device *netdev)
@@ -2749,7 +2988,6 @@ static int dpaa2_switch_port_init(struct ethsw_port_priv *port_priv, u16 port)
 	struct ethsw_core *ethsw = port_priv->ethsw_data;
 	struct dpaa2_switch_acl_tbl *acl_tbl;
 	struct dpsw_fdb_cfg fdb_cfg = {0};
-	struct dpsw_acl_if_cfg acl_if_cfg;
 	struct dpsw_if_attr dpsw_if_attr;
 	struct dpaa2_switch_fdb *fdb;
 	struct dpsw_acl_cfg acl_cfg;
@@ -2803,21 +3041,16 @@ static int dpaa2_switch_port_init(struct ethsw_port_priv *port_priv, u16 port)
 		return err;
 	}
 
-	acl_if_cfg.if_id[0] = port_priv->idx;
-	acl_if_cfg.num_ifs = 1;
-	err = dpsw_acl_add_if(ethsw->mc_io, 0, ethsw->dpsw_handle,
-			      acl_tbl_id, &acl_if_cfg);
-	if (err) {
-		netdev_err(netdev, "dpsw_acl_add_if err %d\n", err);
-		dpsw_acl_remove(ethsw->mc_io, 0, ethsw->dpsw_handle,
-				acl_tbl_id);
-	}
-
 	acl_tbl = dpaa2_switch_acl_tbl_get_unused(ethsw);
+	acl_tbl->ethsw = ethsw;
 	acl_tbl->id = acl_tbl_id;
 	acl_tbl->in_use = true;
 	acl_tbl->num_rules = 0;
-	port_priv->acl_tbl = acl_tbl;
+	INIT_LIST_HEAD(&acl_tbl->entries);
+
+	err = dpaa2_switch_port_acl_tbl_bind(port_priv, acl_tbl);
+	if (err)
+		return err;
 
 	err = dpaa2_switch_port_trap_mac_addr(port_priv, stpa);
 	if (err)
@@ -2927,7 +3160,9 @@ static int dpaa2_switch_probe_port(struct ethsw_core *ethsw,
 	/* The DPAA2 switch's ingress path depends on the VLAN table,
 	 * thus we are not able to disable VLAN filtering.
 	 */
-	port_netdev->features = NETIF_F_HW_VLAN_CTAG_FILTER | NETIF_F_HW_VLAN_STAG_FILTER;
+	port_netdev->features = NETIF_F_HW_VLAN_CTAG_FILTER |
+				NETIF_F_HW_VLAN_STAG_FILTER |
+				NETIF_F_HW_TC;
 
 	err = dpaa2_switch_port_init(port_priv, port_idx);
 	if (err)