dpaa2-eth: Add flow steering support without masking

On platforms that lack a TCAM (like LS1088A), masking of
flow steering keys is not supported. Until now we didn't
offer flow steering capabilities at all on these platforms,
since our driver implementation configured a "comprehensive"
FS key (containing all supported header fields), with masks
used to ignore the fields not present in the rules provided
by the user.

We now allow ethtool rules that share a common key (i.e. have
the same header fields). The FS key is now kept in the driver
private data and initialized when the first rule is added to
an empty table, rather than at probe time. If a rule with a new
composition key is wanted, the user must first manually delete
all previous rules.

When building a FS table entry to pass to firmware, we still use
the old building algorithm, which assumes an all-supported-fields
key, and later collapse the fields which aren't actually needed.

Masked rules are not supported; if provided, the mask value
will be ignored. For firmware versions older than MC10.7.0
(that only offer the legacy ABIs for configuring distribution
keys) flow steering without masking support remains unavailable.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c

@@ -2692,12 +2692,15 @@ static int config_cls_key(struct dpaa2_eth_priv *priv, dma_addr_t key)
 }
 
 /* Size of the Rx flow classification key */
-int dpaa2_eth_cls_key_size(void)
+int dpaa2_eth_cls_key_size(u64 fields)
 {
 	int i, size = 0;
 
-	for (i = 0; i < ARRAY_SIZE(dist_fields); i++)
+	for (i = 0; i < ARRAY_SIZE(dist_fields); i++) {
+		if (!(fields & dist_fields[i].id))
+			continue;
 		size += dist_fields[i].size;
+	}
 
 	return size;
 }
@@ -2718,6 +2721,24 @@ int dpaa2_eth_cls_fld_off(int prot, int field)
 	return 0;
 }
 
+/* Prune unused fields from the classification rule.
+ * Used when masking is not supported
+ */
+void dpaa2_eth_cls_trim_rule(void *key_mem, u64 fields)
+{
+	int off = 0, new_off = 0;
+	int i, size;
+
+	for (i = 0; i < ARRAY_SIZE(dist_fields); i++) {
+		size = dist_fields[i].size;
+		if (dist_fields[i].id & fields) {
+			memcpy(key_mem + new_off, key_mem + off, size);
+			new_off += size;
+		}
+		off += size;
+	}
+}
+
 /* Set Rx distribution (hash or flow classification) key
  * flags is a combination of RXH_ bits
  */
@@ -2739,14 +2760,13 @@ static int dpaa2_eth_set_dist_key(struct net_device *net_dev,
 		struct dpkg_extract *key =
 			&cls_cfg.extracts[cls_cfg.num_extracts];
 
-		/* For Rx hashing key we set only the selected fields.
-		 * For Rx flow classification key we set all supported fields
+		/* For both Rx hashing and classification keys
+		 * we set only the selected fields.
 		 */
-		if (type == DPAA2_ETH_RX_DIST_HASH) {
-			if (!(flags & dist_fields[i].id))
-				continue;
+		if (!(flags & dist_fields[i].id))
+			continue;
+		if (type == DPAA2_ETH_RX_DIST_HASH)
 			rx_hash_fields |= dist_fields[i].rxnfc_field;
-		}
 
 		if (cls_cfg.num_extracts >= DPKG_MAX_NUM_OF_EXTRACTS) {
 			dev_err(dev, "error adding key extraction rule, too many rules?\n");
@@ -2814,7 +2834,12 @@ int dpaa2_eth_set_hash(struct net_device *net_dev, u64 flags)
 	return dpaa2_eth_set_dist_key(net_dev, DPAA2_ETH_RX_DIST_HASH, key);
 }
 
-static int dpaa2_eth_set_cls(struct dpaa2_eth_priv *priv)
+int dpaa2_eth_set_cls(struct net_device *net_dev, u64 flags)
+{
+	return dpaa2_eth_set_dist_key(net_dev, DPAA2_ETH_RX_DIST_CLS, flags);
+}
+
+static int dpaa2_eth_set_default_cls(struct dpaa2_eth_priv *priv)
 {
 	struct device *dev = priv->net_dev->dev.parent;
 	int err;
@@ -2825,7 +2850,7 @@ static int dpaa2_eth_set_cls(struct dpaa2_eth_priv *priv)
 		return -EOPNOTSUPP;
 	}
 
-	if (!dpaa2_eth_fs_enabled(priv) || !dpaa2_eth_fs_mask_enabled(priv)) {
+	if (!dpaa2_eth_fs_enabled(priv)) {
 		dev_dbg(dev, "Rx cls disabled in DPNI options\n");
 		return -EOPNOTSUPP;
 	}
@@ -2835,10 +2860,18 @@ static int dpaa2_eth_set_cls(struct dpaa2_eth_priv *priv)
 		return -EOPNOTSUPP;
 	}
 
-	err = dpaa2_eth_set_dist_key(priv->net_dev, DPAA2_ETH_RX_DIST_CLS, 0);
+	/* If there is no support for masking in the classification table,
+	 * we don't set a default key, as it will depend on the rules
+	 * added by the user at runtime.
+	 */
+	if (!dpaa2_eth_fs_mask_enabled(priv))
+		goto out;
+
+	err = dpaa2_eth_set_cls(priv->net_dev, DPAA2_ETH_DIST_ALL);
 	if (err)
 		return err;
 
+out:
 	priv->rx_cls_enabled = 1;
 
 	return 0;
@@ -2876,7 +2909,7 @@ static int bind_dpni(struct dpaa2_eth_priv *priv)
 	/* Configure the flow classification key; it includes all
 	 * supported header fields and cannot be modified at runtime
 	 */
-	err = dpaa2_eth_set_cls(priv);
+	err = dpaa2_eth_set_default_cls(priv);
 	if (err && err != -EOPNOTSUPP)
 		dev_err(dev, "Failed to configure Rx classification key\n");
 

drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h

@@ -395,6 +395,7 @@ struct dpaa2_eth_priv {
 
 	/* enabled ethtool hashing bits */
 	u64 rx_hash_fields;
+	u64 rx_cls_fields;
 	struct dpaa2_eth_cls_rule *cls_rules;
 	u8 rx_cls_enabled;
 	struct bpf_prog *xdp_prog;
@@ -502,7 +503,9 @@ static inline unsigned int dpaa2_eth_rx_head_room(struct dpaa2_eth_priv *priv)
 }
 
 int dpaa2_eth_set_hash(struct net_device *net_dev, u64 flags);
-int dpaa2_eth_cls_key_size(void);
+int dpaa2_eth_set_cls(struct net_device *net_dev, u64 key);
+int dpaa2_eth_cls_key_size(u64 key);
 int dpaa2_eth_cls_fld_off(int prot, int field);
+void dpaa2_eth_cls_trim_rule(void *key_mem, u64 fields);
 
 #endif	/* __DPAA2_H */