[dpdk-dev] net/mlx5: fix tunnel offloads cap query

Message ID 20180225072837.85663-1-shahafs@mellanox.com
State Accepted, archived
Delegated to: Ferruh Yigit
Headers show

Checks

Context Check Description
ci/Intel-compilation success Compilation OK
ci/checkpatch success coding style OK

Commit Message

Shahaf Shuler Feb. 25, 2018, 7:28 a.m.
The query for the tunnel stateless offloads is wrongly implemented
because of:

1. It was using the device id to query for the offloads.
2. It was using a compilation flag for Verbs which no longer exits.

The main reason was lack of proper API from Verbs.

Fixing the query to use rdma-core API. The capability returned from
rdma-core refer to both Tx and Rx sides.
Eventhough there is a separate cap for GRE and VXLAN, implementation merge
them into a single flag in order to simplify the checks on the data
path.

Fixes: 43e9d9794cde ("net/mlx5: support upstream rdma-core")
Fixes: f5fde5205101 ("net/mlx5: add hardware checksum offload for tunnel packets")
Cc: stable@dpdk.org

Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
Acked-by: Xueming Li <xuemingl@mellanox.com>
---
 drivers/net/mlx5/Makefile   |  6 +++---
 drivers/net/mlx5/mlx5.c     | 39 +++++++++++++++------------------------
 drivers/net/mlx5/mlx5.h     |  4 ++--
 drivers/net/mlx5/mlx5_rxq.c |  2 +-
 4 files changed, 21 insertions(+), 30 deletions(-)

Comments

Nélio Laranjeiro Feb. 26, 2018, 8:12 a.m. | #1
On Sun, Feb 25, 2018 at 09:28:37AM +0200, Shahaf Shuler wrote:
> The query for the tunnel stateless offloads is wrongly implemented
> because of:
> 
> 1. It was using the device id to query for the offloads.
> 2. It was using a compilation flag for Verbs which no longer exits.
> 
> The main reason was lack of proper API from Verbs.
> 
> Fixing the query to use rdma-core API. The capability returned from
> rdma-core refer to both Tx and Rx sides.
> Eventhough there is a separate cap for GRE and VXLAN, implementation merge
> them into a single flag in order to simplify the checks on the data
> path.
> 
> Fixes: 43e9d9794cde ("net/mlx5: support upstream rdma-core")
> Fixes: f5fde5205101 ("net/mlx5: add hardware checksum offload for tunnel packets")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
> Acked-by: Yongseok Koh <yskoh@mellanox.com>
> Acked-by: Xueming Li <xuemingl@mellanox.com>
>[...]
Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Shahaf Shuler Feb. 28, 2018, 8:22 a.m. | #2
Monday, February 26, 2018 10:13 AM, Nélio Laranjeiro:
> On Sun, Feb 25, 2018 at 09:28:37AM +0200, Shahaf Shuler wrote:
> > The query for the tunnel stateless offloads is wrongly implemented
> > because of:
> >
> > 1. It was using the device id to query for the offloads.
> > 2. It was using a compilation flag for Verbs which no longer exits.
> >
> > The main reason was lack of proper API from Verbs.
> >
> > Fixing the query to use rdma-core API. The capability returned from
> > rdma-core refer to both Tx and Rx sides.
> > Eventhough there is a separate cap for GRE and VXLAN, implementation
> > merge them into a single flag in order to simplify the checks on the
> > data path.
> >
> > Fixes: 43e9d9794cde ("net/mlx5: support upstream rdma-core")
> > Fixes: f5fde5205101 ("net/mlx5: add hardware checksum offload for
> > tunnel packets")
> > Cc: stable@dpdk.org
> >
> > Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
> > Acked-by: Yongseok Koh <yskoh@mellanox.com>
> > Acked-by: Xueming Li <xuemingl@mellanox.com> [...]
> Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>

Applied to next-net-mlx, thanks. 

> 
> --
> Nélio Laranjeiro
> 6WIND

Patch

diff --git a/drivers/net/mlx5/Makefile b/drivers/net/mlx5/Makefile
index 3bc9736c9..afda4118f 100644
--- a/drivers/net/mlx5/Makefile
+++ b/drivers/net/mlx5/Makefile
@@ -125,9 +125,9 @@  mlx5_autoconf.h.new: FORCE
 mlx5_autoconf.h.new: $(RTE_SDK)/buildtools/auto-config-h.sh
 	$Q $(RM) -f -- '$@'
 	$Q sh -- '$<' '$@' \
-		HAVE_IBV_DEVICE_VXLAN_SUPPORT \
-		infiniband/verbs.h \
-		enum IBV_DEVICE_VXLAN_SUPPORT \
+		HAVE_IBV_DEVICE_TUNNEL_SUPPORT \
+		infiniband/mlx5dv.h \
+		enum MLX5DV_CONTEXT_MASK_TUNNEL_OFFLOADS \
 		$(AUTOCONF_OUTPUT)
 	$Q sh -- '$<' '$@' \
 		HAVE_IBV_WQ_FLAG_RX_END_PADDING \
diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 6c0985bd6..61cb93101 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -584,7 +584,7 @@  mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 	unsigned int tunnel_en = 0;
 	int idx;
 	int i;
-	struct mlx5dv_context attrs_out;
+	struct mlx5dv_context attrs_out = {0};
 #ifdef HAVE_IBV_DEVICE_COUNTERS_SET_SUPPORT
 	struct ibv_counter_set_description cs_desc;
 #endif
@@ -633,20 +633,6 @@  mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 		       PCI_DEVICE_ID_MELLANOX_CONNECTX5VF) ||
 		      (pci_dev->id.device_id ==
 		       PCI_DEVICE_ID_MELLANOX_CONNECTX5EXVF));
-		switch (pci_dev->id.device_id) {
-		case PCI_DEVICE_ID_MELLANOX_CONNECTX4:
-			tunnel_en = 1;
-			break;
-		case PCI_DEVICE_ID_MELLANOX_CONNECTX4LX:
-		case PCI_DEVICE_ID_MELLANOX_CONNECTX5:
-		case PCI_DEVICE_ID_MELLANOX_CONNECTX5VF:
-		case PCI_DEVICE_ID_MELLANOX_CONNECTX5EX:
-		case PCI_DEVICE_ID_MELLANOX_CONNECTX5EXVF:
-			tunnel_en = 1;
-			break;
-		default:
-			break;
-		}
 		INFO("PCI information matches, using device \"%s\""
 		     " (SR-IOV: %s)",
 		     list[i]->name,
@@ -675,6 +661,9 @@  mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 	 * Multi-packet send is supported by ConnectX-4 Lx PF as well
 	 * as all ConnectX-5 devices.
 	 */
+#ifdef HAVE_IBV_DEVICE_TUNNEL_SUPPORT
+	attrs_out.comp_mask |= MLX5DV_CONTEXT_MASK_TUNNEL_OFFLOADS;
+#endif
 	mlx5_glue->dv_query_device(attr_ctx, &attrs_out);
 	if (attrs_out.flags & MLX5DV_CONTEXT_FLAGS_MPW_ALLOWED) {
 		if (attrs_out.flags & MLX5DV_CONTEXT_FLAGS_ENHANCED_MPW) {
@@ -693,6 +682,17 @@  mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 		cqe_comp = 0;
 	else
 		cqe_comp = 1;
+#ifdef HAVE_IBV_DEVICE_TUNNEL_SUPPORT
+	if (attrs_out.comp_mask & MLX5DV_CONTEXT_MASK_TUNNEL_OFFLOADS) {
+		tunnel_en = ((attrs_out.tunnel_offloads_caps &
+			      MLX5DV_RAW_PACKET_CAP_TUNNELED_OFFLOAD_VXLAN) &&
+			     (attrs_out.tunnel_offloads_caps &
+			      MLX5DV_RAW_PACKET_CAP_TUNNELED_OFFLOAD_GRE));
+	}
+	DEBUG("Tunnel offloading is %ssupported", tunnel_en ? "" : "not ");
+#else
+	WARN("Tunnel offloading disabled due to old OFED/rdma-core version");
+#endif
 	if (mlx5_glue->query_device_ex(attr_ctx, NULL, &device_attr))
 		goto error;
 	INFO("%u port(s) detected", device_attr.orig_attr.phys_port_cnt);
@@ -838,15 +838,6 @@  mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 				    IBV_DEVICE_RAW_IP_CSUM);
 		DEBUG("checksum offloading is %ssupported",
 		      (config.hw_csum ? "" : "not "));
-
-#ifdef HAVE_IBV_DEVICE_VXLAN_SUPPORT
-		config.hw_csum_l2tun =
-				!!(exp_device_attr.exp_device_cap_flags &
-				   IBV_DEVICE_VXLAN_SUPPORT);
-#endif
-		DEBUG("Rx L2 tunnel checksum offloads are %ssupported",
-		      (config.hw_csum_l2tun ? "" : "not "));
-
 #ifdef HAVE_IBV_DEVICE_COUNTERS_SET_SUPPORT
 		config.flow_counter_en = !!(device_attr.max_counter_sets);
 		mlx5_glue->describe_counter_set(ctx, 0, &cs_desc);
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 965c19f21..3e2d96abc 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -75,13 +75,13 @@  TAILQ_HEAD(mlx5_flows, rte_flow);
  */
 struct mlx5_dev_config {
 	unsigned int hw_csum:1; /* Checksum offload is supported. */
-	unsigned int hw_csum_l2tun:1; /* Same for L2 tunnels. */
 	unsigned int hw_vlan_strip:1; /* VLAN stripping is supported. */
 	unsigned int hw_fcs_strip:1; /* FCS stripping is supported. */
 	unsigned int hw_padding:1; /* End alignment padding is supported. */
 	unsigned int sriov:1; /* This is a VF or PF with VF devices. */
 	unsigned int mps:2; /* Multi-packet send supported mode. */
-	unsigned int tunnel_en:1; /* Whether tunnel is supported. */
+	unsigned int tunnel_en:1;
+	/* Whether tunnel stateless offloads are supported. */
 	unsigned int flow_counter_en:1; /* Whether flow counter is supported. */
 	unsigned int cqe_comp:1; /* CQE compression is enabled. */
 	unsigned int tso:1; /* Whether TSO is supported. */
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index ff58c4921..238fa7e56 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -1006,7 +1006,7 @@  mlx5_priv_rxq_new(struct priv *priv, uint16_t idx, uint16_t desc,
 	/* Toggle RX checksum offload if hardware supports it. */
 	tmpl->rxq.csum = !!(conf->offloads & DEV_RX_OFFLOAD_CHECKSUM);
 	tmpl->rxq.csum_l2tun = (!!(conf->offloads & DEV_RX_OFFLOAD_CHECKSUM) &&
-				priv->config.hw_csum_l2tun);
+				priv->config.tunnel_en);
 	tmpl->rxq.hw_timestamp = !!(conf->offloads & DEV_RX_OFFLOAD_TIMESTAMP);
 	/* Configure VLAN stripping. */
 	tmpl->rxq.vlan_strip = !!(conf->offloads & DEV_RX_OFFLOAD_VLAN_STRIP);