patch 'net/mlx5: fix RSS expansion with explicit next protocol' has been queued to stable release 21.11.1

[Kevin Traynor <ktraynor@redhat.com>]

Hi,

FYI, your patch has been queued to stable release 21.11.1

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 02/26/22. So please
shout if anyone has objections.

Also note that after the patch there's a diff of the upstream commit vs the
patch applied to the branch. This will indicate if there was any rebasing
needed to apply to the stable branch. If there were code changes for rebasing
(ie: not only metadata diffs), please double check that the rebase was
correctly done.

Queued patches are on a temporary branch at:
https://github.com/kevintraynor/dpdk-stable

This queued commit can be viewed at:
https://github.com/kevintraynor/dpdk-stable/commit/d8d54171bd5e35b4a471a9be2f4a85a391f616c6

Thanks.

Kevin

---
From d8d54171bd5e35b4a471a9be2f4a85a391f616c6 Mon Sep 17 00:00:00 2001
From: Gregory Etelson <getelson@nvidia.com>
Date: Thu, 23 Dec 2021 15:08:17 +0200
Subject: [PATCH] net/mlx5: fix RSS expansion with explicit next protocol

[ upstream commit f3f1f576f43804b30f81deae967bc09aff7000df ]

The PMD RSS expansion scheme by default compiles flow rules for all
flow item types that may branch out from a stub supplied
by application.
For example,
ETH can lead to VLAN, IPv4 or IPv6.
IPv4 can lead to UDP, TCP, IPv4 or IPv6.

If application explicitly specified next protocol type, expansion must
use that option only and not create flows with other protocol types.

The PMD ignored explicit next protocol values in GRE and VXLAN-GPE.

The patch updates RSS expansion for GRE and VXLAN-GPE with explicit
next protocol settings.

Fixes: c7870bfe09dc ("ethdev: move RSS expansion code to mlx5 driver")

Signed-off-by: Gregory Etelson <getelson@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
---
 drivers/net/mlx5/mlx5_flow.c | 217 +++++++++++++++++++----------------
 1 file changed, 116 insertions(+), 101 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index b7cf4143d5..2cadf615ec 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -165,120 +165,134 @@ mlx5_flow_is_rss_expandable_item(const struct rte_flow_item *item)
 }
 
+/**
+ * Network Service Header (NSH) and its next protocol values
+ * are described in RFC-8393.
+ */
+static enum rte_flow_item_type
+mlx5_nsh_proto_to_item_type(uint8_t proto_spec, uint8_t proto_mask)
+{
+	enum rte_flow_item_type type;
+
+	switch (proto_mask & proto_spec) {
+	case RTE_VXLAN_GPE_TYPE_IPV4:
+		type = RTE_FLOW_ITEM_TYPE_IPV4;
+		break;
+	case RTE_VXLAN_GPE_TYPE_IPV6:
+		type = RTE_VXLAN_GPE_TYPE_IPV6;
+		break;
+	case RTE_VXLAN_GPE_TYPE_ETH:
+		type = RTE_FLOW_ITEM_TYPE_ETH;
+		break;
+	default:
+		type = RTE_FLOW_ITEM_TYPE_END;
+	}
+	return type;
+}
+
+static enum rte_flow_item_type
+mlx5_inet_proto_to_item_type(uint8_t proto_spec, uint8_t proto_mask)
+{
+	enum rte_flow_item_type type;
+
+	switch (proto_mask & proto_spec) {
+	case IPPROTO_UDP:
+		type = RTE_FLOW_ITEM_TYPE_UDP;
+		break;
+	case IPPROTO_TCP:
+		type = RTE_FLOW_ITEM_TYPE_TCP;
+		break;
+	case IPPROTO_IP:
+		type = RTE_FLOW_ITEM_TYPE_IPV4;
+		break;
+	case IPPROTO_IPV6:
+		type = RTE_FLOW_ITEM_TYPE_IPV6;
+		break;
+	default:
+		type = RTE_FLOW_ITEM_TYPE_END;
+	}
+	return type;
+}
+
+static enum rte_flow_item_type
+mlx5_ethertype_to_item_type(rte_be16_t type_spec,
+			    rte_be16_t type_mask, bool is_tunnel)
+{
+	enum rte_flow_item_type type;
+
+	switch (rte_be_to_cpu_16(type_spec & type_mask)) {
+	case RTE_ETHER_TYPE_TEB:
+		type = is_tunnel ?
+		       RTE_FLOW_ITEM_TYPE_ETH : RTE_FLOW_ITEM_TYPE_END;
+		break;
+	case RTE_ETHER_TYPE_VLAN:
+		type = !is_tunnel ?
+		       RTE_FLOW_ITEM_TYPE_VLAN : RTE_FLOW_ITEM_TYPE_END;
+		break;
+	case RTE_ETHER_TYPE_IPV4:
+		type = RTE_FLOW_ITEM_TYPE_IPV4;
+		break;
+	case RTE_ETHER_TYPE_IPV6:
+		type = RTE_FLOW_ITEM_TYPE_IPV6;
+		break;
+	default:
+		type = RTE_FLOW_ITEM_TYPE_END;
+	}
+	return type;
+}
+
 static enum rte_flow_item_type
 mlx5_flow_expand_rss_item_complete(const struct rte_flow_item *item)
 {
-	enum rte_flow_item_type ret = RTE_FLOW_ITEM_TYPE_VOID;
-	uint16_t ether_type = 0;
-	uint16_t ether_type_m;
-	uint8_t ip_next_proto = 0;
-	uint8_t ip_next_proto_m;
+#define MLX5_XSET_ITEM_MASK_SPEC(type, fld)                              \
+	do {                                                             \
+		const void *m = item->mask;                              \
+		const void *s = item->spec;                              \
+		mask = m ?                                               \
+			((const struct rte_flow_item_##type *)m)->fld :  \
+			rte_flow_item_##type##_mask.fld;                 \
+		spec = ((const struct rte_flow_item_##type *)s)->fld;    \
+	} while (0)
+
+	enum rte_flow_item_type ret;
+	uint16_t spec, mask;
 
 	if (item == NULL || item->spec == NULL)
-		return ret;
+		return RTE_FLOW_ITEM_TYPE_VOID;
 	switch (item->type) {
 	case RTE_FLOW_ITEM_TYPE_ETH:
-		if (item->mask)
-			ether_type_m = ((const struct rte_flow_item_eth *)
-						(item->mask))->type;
-		else
-			ether_type_m = rte_flow_item_eth_mask.type;
-		if (ether_type_m != RTE_BE16(0xFFFF))
-			break;
-		ether_type = ((const struct rte_flow_item_eth *)
-				(item->spec))->type;
-		if (rte_be_to_cpu_16(ether_type) == RTE_ETHER_TYPE_IPV4)
-			ret = RTE_FLOW_ITEM_TYPE_IPV4;
-		else if (rte_be_to_cpu_16(ether_type) == RTE_ETHER_TYPE_IPV6)
-			ret = RTE_FLOW_ITEM_TYPE_IPV6;
-		else if (rte_be_to_cpu_16(ether_type) == RTE_ETHER_TYPE_VLAN)
-			ret = RTE_FLOW_ITEM_TYPE_VLAN;
-		else
-			ret = RTE_FLOW_ITEM_TYPE_END;
+		MLX5_XSET_ITEM_MASK_SPEC(eth, type);
+		if (!mask)
+			return RTE_FLOW_ITEM_TYPE_VOID;
+		ret = mlx5_ethertype_to_item_type(spec, mask, false);
 		break;
 	case RTE_FLOW_ITEM_TYPE_VLAN:
-		if (item->mask)
-			ether_type_m = ((const struct rte_flow_item_vlan *)
-						(item->mask))->inner_type;
-		else
-			ether_type_m = rte_flow_item_vlan_mask.inner_type;
-		if (ether_type_m != RTE_BE16(0xFFFF))
-			break;
-		ether_type = ((const struct rte_flow_item_vlan *)
-				(item->spec))->inner_type;
-		if (rte_be_to_cpu_16(ether_type) == RTE_ETHER_TYPE_IPV4)
-			ret = RTE_FLOW_ITEM_TYPE_IPV4;
-		else if (rte_be_to_cpu_16(ether_type) == RTE_ETHER_TYPE_IPV6)
-			ret = RTE_FLOW_ITEM_TYPE_IPV6;
-		else if (rte_be_to_cpu_16(ether_type) == RTE_ETHER_TYPE_VLAN)
-			ret = RTE_FLOW_ITEM_TYPE_VLAN;
-		else
-			ret = RTE_FLOW_ITEM_TYPE_END;
+		MLX5_XSET_ITEM_MASK_SPEC(vlan, inner_type);
+		if (!mask)
+			return RTE_FLOW_ITEM_TYPE_VOID;
+		ret = mlx5_ethertype_to_item_type(spec, mask, false);
 		break;
 	case RTE_FLOW_ITEM_TYPE_IPV4:
-		if (item->mask)
-			ip_next_proto_m = ((const struct rte_flow_item_ipv4 *)
-					(item->mask))->hdr.next_proto_id;
-		else
-			ip_next_proto_m =
-				rte_flow_item_ipv4_mask.hdr.next_proto_id;
-		if (ip_next_proto_m != 0xFF)
-			break;
-		ip_next_proto = ((const struct rte_flow_item_ipv4 *)
-				(item->spec))->hdr.next_proto_id;
-		if (ip_next_proto == IPPROTO_UDP)
-			ret = RTE_FLOW_ITEM_TYPE_UDP;
-		else if (ip_next_proto == IPPROTO_TCP)
-			ret = RTE_FLOW_ITEM_TYPE_TCP;
-		else if (ip_next_proto == IPPROTO_IP)
-			ret = RTE_FLOW_ITEM_TYPE_IPV4;
-		else if (ip_next_proto == IPPROTO_IPV6)
-			ret = RTE_FLOW_ITEM_TYPE_IPV6;
-		else
-			ret = RTE_FLOW_ITEM_TYPE_END;
+		MLX5_XSET_ITEM_MASK_SPEC(ipv4, hdr.next_proto_id);
+		if (!mask)
+			return RTE_FLOW_ITEM_TYPE_VOID;
+		ret = mlx5_inet_proto_to_item_type(spec, mask);
 		break;
 	case RTE_FLOW_ITEM_TYPE_IPV6:
-		if (item->mask)
-			ip_next_proto_m = ((const struct rte_flow_item_ipv6 *)
-						(item->mask))->hdr.proto;
-		else
-			ip_next_proto_m =
-				rte_flow_item_ipv6_mask.hdr.proto;
-		if (ip_next_proto_m != 0xFF)
-			break;
-		ip_next_proto = ((const struct rte_flow_item_ipv6 *)
-				(item->spec))->hdr.proto;
-		if (ip_next_proto == IPPROTO_UDP)
-			ret = RTE_FLOW_ITEM_TYPE_UDP;
-		else if (ip_next_proto == IPPROTO_TCP)
-			ret = RTE_FLOW_ITEM_TYPE_TCP;
-		else if (ip_next_proto == IPPROTO_IP)
-			ret = RTE_FLOW_ITEM_TYPE_IPV4;
-		else if (ip_next_proto == IPPROTO_IPV6)
-			ret = RTE_FLOW_ITEM_TYPE_IPV6;
-		else
-			ret = RTE_FLOW_ITEM_TYPE_END;
+		MLX5_XSET_ITEM_MASK_SPEC(ipv6, hdr.proto);
+		if (!mask)
+			return RTE_FLOW_ITEM_TYPE_VOID;
+		ret = mlx5_inet_proto_to_item_type(spec, mask);
 		break;
 	case RTE_FLOW_ITEM_TYPE_GENEVE:
-		ether_type_m = item->mask ?
-			       ((const struct rte_flow_item_geneve *)
-			       (item->mask))->protocol :
-			       rte_flow_item_geneve_mask.protocol;
-		ether_type = ((const struct rte_flow_item_geneve *)
-			     (item->spec))->protocol;
-		ether_type_m = rte_be_to_cpu_16(ether_type_m);
-		ether_type = rte_be_to_cpu_16(ether_type);
-		switch (ether_type_m & ether_type) {
-		case RTE_ETHER_TYPE_TEB:
-			ret = RTE_FLOW_ITEM_TYPE_ETH;
-			break;
-		case RTE_ETHER_TYPE_IPV4:
-			ret = RTE_FLOW_ITEM_TYPE_IPV4;
-			break;
-		case RTE_ETHER_TYPE_IPV6:
-			ret = RTE_FLOW_ITEM_TYPE_IPV6;
-			break;
-		default:
-			ret = RTE_FLOW_ITEM_TYPE_END;
-		}
+		MLX5_XSET_ITEM_MASK_SPEC(geneve, protocol);
+		ret = mlx5_ethertype_to_item_type(spec, mask, true);
+		break;
+	case RTE_FLOW_ITEM_TYPE_GRE:
+		MLX5_XSET_ITEM_MASK_SPEC(gre, protocol);
+		ret = mlx5_ethertype_to_item_type(spec, mask, true);
+		break;
+	case RTE_FLOW_ITEM_TYPE_VXLAN_GPE:
+		MLX5_XSET_ITEM_MASK_SPEC(vxlan_gpe, protocol);
+		ret = mlx5_nsh_proto_to_item_type(spec, mask);
 		break;
 	default:
@@ -287,4 +301,5 @@ mlx5_flow_expand_rss_item_complete(const struct rte_flow_item *item)
 	}
 	return ret;
+#undef MLX5_XSET_ITEM_MASK_SPEC
 }
 
-- 
2.34.1

---
  Diff of the applied patch vs upstream commit (please double-check if non-empty:
---
--- -	2022-02-21 15:22:45.616795045 +0000
+++ 0045-net-mlx5-fix-RSS-expansion-with-explicit-next-protoc.patch	2022-02-21 15:22:44.094704132 +0000
@@ -1 +1 @@
-From f3f1f576f43804b30f81deae967bc09aff7000df Mon Sep 17 00:00:00 2001
+From d8d54171bd5e35b4a471a9be2f4a85a391f616c6 Mon Sep 17 00:00:00 2001
@@ -5,0 +6,2 @@
+[ upstream commit f3f1f576f43804b30f81deae967bc09aff7000df ]
+
@@ -22 +23,0 @@
-Cc: stable@dpdk.org