[PATCH net-next V2 0/2] net/mlx4: Add support for TCP/IP offloads of vxlan tunneling

[PATCH net-next V2 0/2] net/mlx4: Add support for TCP/IP offloads of vxlan tunneling

[Or Gerlitz]

Hi Dave,

These two patches add mlx4 support for TCP/IP offloads under vxlan tunneling.

V0 was submitted rebased against the other inflight mlx4 series we sent, 
but it turns out that there was another mlx4 patch in the air, its now merged 
as 6917441 "net: mlx4 calls skb_set_hash" which has tiny collision with V0 
as of content changes.


changes from V1:
 - applied feedback from Joe Perches: fixed an overly indented line, made a variable bool
   instead of int, initialize dev->hw_enc_features with |= to take into account
   setting done by the networking core during register_netdevice

changes from V0:
 - rebased against net-next commit 289dccb "net: use kfree_skb_list() helper",
   to avoid git am failure as of content change

Or.

Or Gerlitz (2):
  net/mlx4_core: Add basic support for TCP/IP offloads under tunneling
  net/mlx4_en: Add netdev support for TCP/IP offloads of vxlan
    tunneling

 drivers/net/ethernet/mellanox/mlx4/en_netdev.c    |   94 +++++++++++++++++++++
 drivers/net/ethernet/mellanox/mlx4/en_resources.c |    6 ++
 drivers/net/ethernet/mellanox/mlx4/en_rx.c        |   14 +++
 drivers/net/ethernet/mellanox/mlx4/en_tx.c        |   14 +++-
 drivers/net/ethernet/mellanox/mlx4/fw.c           |   19 ++++-
 drivers/net/ethernet/mellanox/mlx4/main.c         |   14 +++
 drivers/net/ethernet/mellanox/mlx4/mcg.c          |   14 +++-
 drivers/net/ethernet/mellanox/mlx4/mlx4_en.h      |    3 +-
 drivers/net/ethernet/mellanox/mlx4/port.c         |   41 +++++++++
 include/linux/mlx4/cmd.h                          |    1 +
 include/linux/mlx4/cq.h                           |    5 +
 include/linux/mlx4/device.h                       |   37 ++++++++-
 include/linux/mlx4/qp.h                           |    6 ++
 13 files changed, 262 insertions(+), 6 deletions(-)

[PATCH net-next V2 1/2] net/mlx4_core: Add basic support for TCP/IP offloads under tunneling

[Or Gerlitz]

Add the low-level device commands and definitions used for TCP/IP HW offloads
of tunneled/vxlan traffic which are supported by the ConnectX3-pro NIC.

This is done through the following elements:

 - read tunneling device caps in QUERY_DEV_CAP
 - add helper function to do SET_PORT for tunneling
 - add DMFS VXLAN steering rule definitions
 - add CQE and WQE checksum offload field definitions

Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/fw.c   |   19 ++++++++++++-
 drivers/net/ethernet/mellanox/mlx4/main.c |   14 ++++++++++
 drivers/net/ethernet/mellanox/mlx4/mcg.c  |   14 ++++++++-
 drivers/net/ethernet/mellanox/mlx4/port.c |   41 +++++++++++++++++++++++++++++
 include/linux/mlx4/cmd.h                  |    1 +
 include/linux/mlx4/cq.h                   |    5 +++
 include/linux/mlx4/device.h               |   37 +++++++++++++++++++++++++-
 include/linux/mlx4/qp.h                   |    6 ++++
 8 files changed, 133 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c
index 27a8434..55c4ea7 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw.c
+++ b/drivers/net/ethernet/mellanox/mlx4/fw.c
@@ -134,7 +134,8 @@ static void dump_dev_cap_flags2(struct mlx4_dev *dev, u64 flags)
 		[5] = "Time stamping support",
 		[6] = "VST (control vlan insertion/stripping) support",
 		[7] = "FSM (MAC anti-spoofing) support",
-		[8] = "Dynamic QP updates support"
+		[8] = "Dynamic QP updates support",
+		[9] = "TCP/IP offloads/flow-steering for VXLAN support"
 	};
 	int i;
 
@@ -536,6 +537,7 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 #define QUERY_DEV_CAP_RSVD_LKEY_OFFSET		0x98
 #define QUERY_DEV_CAP_MAX_ICM_SZ_OFFSET		0xa0
 #define QUERY_DEV_CAP_FW_REASSIGN_MAC		0x9d
+#define QUERY_DEV_CAP_VXLAN			0x9e
 
 	dev_cap->flags2 = 0;
 	mailbox = mlx4_alloc_cmd_mailbox(dev);
@@ -701,6 +703,9 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 	MLX4_GET(field, outbox, QUERY_DEV_CAP_FW_REASSIGN_MAC);
 	if (field & 1<<6)
 		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_REASSIGN_MAC_EN;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_VXLAN);
+	if (field & 1<<3)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_VXLAN_OFFLOADS;
 	MLX4_GET(dev_cap->max_icm_sz, outbox,
 		 QUERY_DEV_CAP_MAX_ICM_SZ_OFFSET);
 	if (dev_cap->flags & MLX4_DEV_CAP_FLAG_COUNTERS)
@@ -849,6 +854,11 @@ int mlx4_QUERY_DEV_CAP_wrapper(struct mlx4_dev *dev, int slave,
 	field &= 0x7f;
 	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_CQ_TS_SUPPORT_OFFSET);
 
+	/* For guests, disable vxlan tunneling */
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_VXLAN);
+	field &= 0xf7;
+	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_VXLAN);
+
 	/* For guests, report Blueflame disabled */
 	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_BF_OFFSET);
 	field &= 0x7f;
@@ -1274,6 +1284,7 @@ int mlx4_INIT_HCA(struct mlx4_dev *dev, struct mlx4_init_hca_param *param)
 #define INIT_HCA_IN_SIZE		 0x200
 #define INIT_HCA_VERSION_OFFSET		 0x000
 #define	 INIT_HCA_VERSION		 2
+#define INIT_HCA_VXLAN_OFFSET		 0x0c
 #define INIT_HCA_CACHELINE_SZ_OFFSET	 0x0e
 #define INIT_HCA_FLAGS_OFFSET		 0x014
 #define INIT_HCA_QPC_OFFSET		 0x020
@@ -1432,6 +1443,12 @@ int mlx4_INIT_HCA(struct mlx4_dev *dev, struct mlx4_init_hca_param *param)
 	MLX4_PUT(inbox, param->uar_page_sz,	INIT_HCA_UAR_PAGE_SZ_OFFSET);
 	MLX4_PUT(inbox, param->log_uar_sz,      INIT_HCA_LOG_UAR_SZ_OFFSET);
 
+	/* set parser VXLAN attributes */
+	if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_VXLAN_OFFLOADS) {
+		u8 parser_params = 0;
+		MLX4_PUT(inbox, parser_params,	INIT_HCA_VXLAN_OFFSET);
+	}
+
 	err = mlx4_cmd(dev, mailbox->dma, 0, 0, MLX4_CMD_INIT_HCA, 10000,
 		       MLX4_CMD_NATIVE);
 
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index fbebfb0..d2b8b39 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -1444,6 +1444,19 @@ static void choose_steering_mode(struct mlx4_dev *dev,
 		 mlx4_log_num_mgm_entry_size);
 }
 
+static void choose_tunnel_offload_mode(struct mlx4_dev *dev,
+				       struct mlx4_dev_cap *dev_cap)
+{
+	if (dev->caps.steering_mode == MLX4_STEERING_MODE_DEVICE_MANAGED &&
+	    dev_cap->flags2 & MLX4_DEV_CAP_FLAG2_VXLAN_OFFLOADS)
+		dev->caps.tunnel_offload_mode = MLX4_TUNNEL_OFFLOAD_MODE_VXLAN;
+	else
+		dev->caps.tunnel_offload_mode = MLX4_TUNNEL_OFFLOAD_MODE_NONE;
+
+	mlx4_dbg(dev, "Tunneling offload mode is: %s\n",  (dev->caps.tunnel_offload_mode
+		 == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN) ? "vxlan" : "none");
+}
+
 static int mlx4_init_hca(struct mlx4_dev *dev)
 {
 	struct mlx4_priv	  *priv = mlx4_priv(dev);
@@ -1484,6 +1497,7 @@ static int mlx4_init_hca(struct mlx4_dev *dev)
 		}
 
 		choose_steering_mode(dev, &dev_cap);
+		choose_tunnel_offload_mode(dev, &dev_cap);
 
 		err = mlx4_get_phys_port_id(dev);
 		if (err)
diff --git a/drivers/net/ethernet/mellanox/mlx4/mcg.c b/drivers/net/ethernet/mellanox/mlx4/mcg.c
index 7c83e6c..bfe65f7 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mcg.c
+++ b/drivers/net/ethernet/mellanox/mlx4/mcg.c
@@ -697,7 +697,8 @@ const u16 __sw_id_hw[] = {
 	[MLX4_NET_TRANS_RULE_ID_IPV6]    = 0xE003,
 	[MLX4_NET_TRANS_RULE_ID_IPV4]    = 0xE002,
 	[MLX4_NET_TRANS_RULE_ID_TCP]     = 0xE004,
-	[MLX4_NET_TRANS_RULE_ID_UDP]     = 0xE006
+	[MLX4_NET_TRANS_RULE_ID_UDP]     = 0xE006,
+	[MLX4_NET_TRANS_RULE_ID_VXLAN]	 = 0xE008
 };
 
 int mlx4_map_sw_to_hw_steering_id(struct mlx4_dev *dev,
@@ -722,7 +723,9 @@ static const int __rule_hw_sz[] = {
 	[MLX4_NET_TRANS_RULE_ID_TCP] =
 		sizeof(struct mlx4_net_trans_rule_hw_tcp_udp),
 	[MLX4_NET_TRANS_RULE_ID_UDP] =
-		sizeof(struct mlx4_net_trans_rule_hw_tcp_udp)
+		sizeof(struct mlx4_net_trans_rule_hw_tcp_udp),
+	[MLX4_NET_TRANS_RULE_ID_VXLAN] =
+		sizeof(struct mlx4_net_trans_rule_hw_vxlan)
 };
 
 int mlx4_hw_rule_sz(struct mlx4_dev *dev,
@@ -787,6 +790,13 @@ static int parse_trans_rule(struct mlx4_dev *dev, struct mlx4_spec_list *spec,
 		rule_hw->tcp_udp.src_port_msk = spec->tcp_udp.src_port_msk;
 		break;
 
+	case MLX4_NET_TRANS_RULE_ID_VXLAN:
+		rule_hw->vxlan.vni =
+			cpu_to_be32(be32_to_cpu(spec->vxlan.vni) << 8);
+		rule_hw->vxlan.vni_mask =
+			cpu_to_be32(be32_to_cpu(spec->vxlan.vni_mask) << 8);
+		break;
+
 	default:
 		return -EINVAL;
 	}
diff --git a/drivers/net/ethernet/mellanox/mlx4/port.c b/drivers/net/ethernet/mellanox/mlx4/port.c
index 97d342f..93f75ec 100644
--- a/drivers/net/ethernet/mellanox/mlx4/port.c
+++ b/drivers/net/ethernet/mellanox/mlx4/port.c
@@ -800,6 +800,47 @@ int mlx4_SET_PORT_SCHEDULER(struct mlx4_dev *dev, u8 port, u8 *tc_tx_bw,
 }
 EXPORT_SYMBOL(mlx4_SET_PORT_SCHEDULER);
 
+enum {
+	VXLAN_ENABLE_MODIFY	= 1 << 7,
+	VXLAN_STEERING_MODIFY	= 1 << 6,
+
+	VXLAN_ENABLE		= 1 << 7,
+};
+
+struct mlx4_set_port_vxlan_context {
+	u32	reserved1;
+	u8	modify_flags;
+	u8	reserved2;
+	u8	enable_flags;
+	u8	steering;
+};
+
+int mlx4_SET_PORT_VXLAN(struct mlx4_dev *dev, u8 port, u8 steering)
+{
+	int err;
+	u32 in_mod;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_vxlan_context  *context;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	context = mailbox->buf;
+	memset(context, 0, sizeof(*context));
+
+	context->modify_flags = VXLAN_ENABLE_MODIFY | VXLAN_STEERING_MODIFY;
+	context->enable_flags = VXLAN_ENABLE;
+	context->steering  = steering;
+
+	in_mod = MLX4_SET_PORT_VXLAN << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_VXLAN);
+
 int mlx4_SET_MCAST_FLTR_wrapper(struct mlx4_dev *dev, int slave,
 				struct mlx4_vhcr *vhcr,
 				struct mlx4_cmd_mailbox *inbox,
diff --git a/include/linux/mlx4/cmd.h b/include/linux/mlx4/cmd.h
index 8df61bc..b0ec0fe 100644
--- a/include/linux/mlx4/cmd.h
+++ b/include/linux/mlx4/cmd.h
@@ -180,6 +180,7 @@ enum {
 	MLX4_SET_PORT_GID_TABLE = 0x5,
 	MLX4_SET_PORT_PRIO2TC	= 0x8,
 	MLX4_SET_PORT_SCHEDULER = 0x9,
+	MLX4_SET_PORT_VXLAN	= 0xB
 };
 
 enum {
diff --git a/include/linux/mlx4/cq.h b/include/linux/mlx4/cq.h
index 98fa492..5dd0d70 100644
--- a/include/linux/mlx4/cq.h
+++ b/include/linux/mlx4/cq.h
@@ -81,7 +81,12 @@ struct mlx4_ts_cqe {
 } __packed;
 
 enum {
+	MLX4_CQE_L2_TUNNEL_IPOK		= 1 << 31,
 	MLX4_CQE_VLAN_PRESENT_MASK	= 1 << 29,
+	MLX4_CQE_L2_TUNNEL		= 1 << 27,
+	MLX4_CQE_L2_TUNNEL_CSUM		= 1 << 26,
+	MLX4_CQE_L2_TUNNEL_IPV4		= 1 << 25,
+
 	MLX4_CQE_QPN_MASK		= 0xffffff,
 };
 
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index 294b7c5..c99ecf6 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -119,6 +119,11 @@ static inline const char *mlx4_steering_mode_str(int steering_mode)
 }
 
 enum {
+	MLX4_TUNNEL_OFFLOAD_MODE_NONE,
+	MLX4_TUNNEL_OFFLOAD_MODE_VXLAN
+};
+
+enum {
 	MLX4_DEV_CAP_FLAG_RC		= 1LL <<  0,
 	MLX4_DEV_CAP_FLAG_UC		= 1LL <<  1,
 	MLX4_DEV_CAP_FLAG_UD		= 1LL <<  2,
@@ -160,7 +165,8 @@ enum {
 	MLX4_DEV_CAP_FLAG2_TS			= 1LL <<  5,
 	MLX4_DEV_CAP_FLAG2_VLAN_CONTROL		= 1LL <<  6,
 	MLX4_DEV_CAP_FLAG2_FSM			= 1LL <<  7,
-	MLX4_DEV_CAP_FLAG2_UPDATE_QP		= 1LL <<  8
+	MLX4_DEV_CAP_FLAG2_UPDATE_QP		= 1LL <<  8,
+	MLX4_DEV_CAP_FLAG2_VXLAN_OFFLOADS	= 1LL <<  9
 };
 
 enum {
@@ -455,6 +461,7 @@ struct mlx4_caps {
 	u32			function_caps;  /* VFs must be aware of these */
 	u16			hca_core_clock;
 	u64			phys_port_id[MLX4_MAX_PORTS + 1];
+	int			tunnel_offload_mode;
 };
 
 struct mlx4_buf_list {
@@ -909,6 +916,7 @@ enum mlx4_net_trans_rule_id {
 	MLX4_NET_TRANS_RULE_ID_IPV4,
 	MLX4_NET_TRANS_RULE_ID_TCP,
 	MLX4_NET_TRANS_RULE_ID_UDP,
+	MLX4_NET_TRANS_RULE_ID_VXLAN,
 	MLX4_NET_TRANS_RULE_NUM, /* should be last */
 };
 
@@ -966,6 +974,12 @@ struct mlx4_spec_ib {
 	u8	dst_gid_msk[16];
 };
 
+struct mlx4_spec_vxlan {
+	__be32 vni;
+	__be32 vni_mask;
+
+};
+
 struct mlx4_spec_list {
 	struct	list_head list;
 	enum	mlx4_net_trans_rule_id id;
@@ -974,6 +988,7 @@ struct mlx4_spec_list {
 		struct mlx4_spec_ib ib;
 		struct mlx4_spec_ipv4 ipv4;
 		struct mlx4_spec_tcp_udp tcp_udp;
+		struct mlx4_spec_vxlan vxlan;
 	};
 };
 
@@ -1060,6 +1075,15 @@ struct mlx4_net_trans_rule_hw_ipv4 {
 	__be32	src_ip_msk;
 } __packed;
 
+struct mlx4_net_trans_rule_hw_vxlan {
+	u8	size;
+	u8	rsvd;
+	__be16	id;
+	__be32	rsvd1;
+	__be32	vni;
+	__be32	vni_mask;
+} __packed;
+
 struct _rule_hw {
 	union {
 		struct {
@@ -1071,9 +1095,19 @@ struct _rule_hw {
 		struct mlx4_net_trans_rule_hw_ib ib;
 		struct mlx4_net_trans_rule_hw_ipv4 ipv4;
 		struct mlx4_net_trans_rule_hw_tcp_udp tcp_udp;
+		struct mlx4_net_trans_rule_hw_vxlan vxlan;
 	};
 };
 
+enum {
+	VXLAN_STEER_BY_OUTER_MAC	= 1 << 0,
+	VXLAN_STEER_BY_OUTER_VLAN	= 1 << 1,
+	VXLAN_STEER_BY_VSID_VNI		= 1 << 2,
+	VXLAN_STEER_BY_INNER_MAC	= 1 << 3,
+	VXLAN_STEER_BY_INNER_VLAN	= 1 << 4,
+};
+
+
 int mlx4_flow_steer_promisc_add(struct mlx4_dev *dev, u8 port, u32 qpn,
 				enum mlx4_net_trans_promisc_mode mode);
 int mlx4_flow_steer_promisc_remove(struct mlx4_dev *dev, u8 port,
@@ -1096,6 +1130,7 @@ int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn,
 int mlx4_SET_PORT_PRIO2TC(struct mlx4_dev *dev, u8 port, u8 *prio2tc);
 int mlx4_SET_PORT_SCHEDULER(struct mlx4_dev *dev, u8 port, u8 *tc_tx_bw,
 		u8 *pg, u16 *ratelimit);
+int mlx4_SET_PORT_VXLAN(struct mlx4_dev *dev, u8 port, u8 steering);
 int mlx4_find_cached_vlan(struct mlx4_dev *dev, u8 port, u16 vid, int *idx);
 int mlx4_register_vlan(struct mlx4_dev *dev, u8 port, u16 vlan, int *index);
 void mlx4_unregister_vlan(struct mlx4_dev *dev, u8 port, u16 vlan);
diff --git a/include/linux/mlx4/qp.h b/include/linux/mlx4/qp.h
index 6d35147..59f8ba8 100644
--- a/include/linux/mlx4/qp.h
+++ b/include/linux/mlx4/qp.h
@@ -109,6 +109,10 @@ enum {
 	MLX4_RSS_TCP_IPV4			= 1 << 4,
 	MLX4_RSS_IPV4				= 1 << 5,
 
+	MLX4_RSS_BY_OUTER_HEADERS		= 0 << 6,
+	MLX4_RSS_BY_INNER_HEADERS		= 2 << 6,
+	MLX4_RSS_BY_INNER_HEADERS_IPONLY	= 3 << 6,
+
 	/* offset of mlx4_rss_context within mlx4_qp_context.pri_path */
 	MLX4_RSS_OFFSET_IN_QPC_PRI_PATH		= 0x24,
 	/* offset of being RSS indirection QP within mlx4_qp_context.flags */
@@ -252,6 +256,8 @@ enum { /* param3 */
 
 enum {
 	MLX4_WQE_CTRL_NEC		= 1 << 29,
+	MLX4_WQE_CTRL_IIP		= 1 << 28,
+	MLX4_WQE_CTRL_ILP		= 1 << 27,
 	MLX4_WQE_CTRL_FENCE		= 1 << 6,
 	MLX4_WQE_CTRL_CQ_UPDATE		= 3 << 2,
 	MLX4_WQE_CTRL_SOLICITED		= 1 << 1,
-- 
1.7.1

[PATCH net-next V2 2/2] net/mlx4_en: Add netdev support for TCP/IP offloads of vxlan tunneling

[Or Gerlitz]

When the device tunneling offloads mode is vxlan do the following

 - call SET_PORT with the relevant setting

 - add DMFS steering vxlan rule for the device self and multicast mac addresses
   of the form: {<ETH, outer-mac> <VXLAN, ANY vnid> <ETH, ANY mac>} --> RSS QP

 - set relevant QPC fields in RSS context and RX ring QPs

 - in TX flow, set WQE fields to generate HW checksum, and handle gso skbs
   which are marked for encapsulation such that the HW will segment them properly.

 - in RX flow, read HW offloaded checksum for encapsulated packets from the CQE

 - advertize hw_enc_features and NETIF_F_GSO_UDP_TUNNEL to the networking stack

Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_netdev.c    |   94 +++++++++++++++++++++
 drivers/net/ethernet/mellanox/mlx4/en_resources.c |    6 ++
 drivers/net/ethernet/mellanox/mlx4/en_rx.c        |   14 +++
 drivers/net/ethernet/mellanox/mlx4/en_tx.c        |   14 +++-
 drivers/net/ethernet/mellanox/mlx4/mlx4_en.h      |    3 +-
 5 files changed, 129 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index 6f92090..0a19c1d 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -468,6 +468,53 @@ static void mlx4_en_u64_to_mac(unsigned char dst_mac[ETH_ALEN + 2], u64 src_mac)
 	memset(&dst_mac[ETH_ALEN], 0, 2);
 }
 
+
+static int mlx4_en_tunnel_steer_add(struct mlx4_en_priv *priv, unsigned char *addr,
+				    int qpn, u64 *reg_id)
+{
+	int err;
+	struct mlx4_spec_list spec_eth_outer = { {NULL} };
+	struct mlx4_spec_list spec_vxlan     = { {NULL} };
+	struct mlx4_spec_list spec_eth_inner = { {NULL} };
+
+	struct mlx4_net_trans_rule rule = {
+		.queue_mode = MLX4_NET_TRANS_Q_FIFO,
+		.exclusive = 0,
+		.allow_loopback = 1,
+		.promisc_mode = MLX4_FS_REGULAR,
+		.priority = MLX4_DOMAIN_NIC,
+	};
+
+	__be64 mac_mask = cpu_to_be64(MLX4_MAC_MASK << 16);
+
+	if ((priv->mdev->dev->caps.tunnel_offload_mode != MLX4_TUNNEL_OFFLOAD_MODE_VXLAN))
+		return 0; /* do nothing */
+
+	rule.port = priv->port;
+	rule.qpn = qpn;
+	INIT_LIST_HEAD(&rule.list);
+
+	spec_eth_outer.id = MLX4_NET_TRANS_RULE_ID_ETH;
+	memcpy(spec_eth_outer.eth.dst_mac, addr, ETH_ALEN);
+	memcpy(spec_eth_outer.eth.dst_mac_msk, &mac_mask, ETH_ALEN);
+
+	spec_vxlan.id = MLX4_NET_TRANS_RULE_ID_VXLAN;    /* any vxlan header */
+	spec_eth_inner.id = MLX4_NET_TRANS_RULE_ID_ETH;	 /* any inner eth header */
+
+	list_add_tail(&spec_eth_outer.list, &rule.list);
+	list_add_tail(&spec_vxlan.list,     &rule.list);
+	list_add_tail(&spec_eth_inner.list, &rule.list);
+
+	err = mlx4_flow_attach(priv->mdev->dev, &rule, reg_id);
+	if (err) {
+		en_err(priv, "failed to add vxlan steering rule, err %d\n", err);
+		return err;
+	}
+	en_dbg(DRV, priv, "added vxlan steering rule, mac %pM reg_id %llx\n", addr, *reg_id);
+	return 0;
+}
+
+
 static int mlx4_en_uc_steer_add(struct mlx4_en_priv *priv,
 				unsigned char *mac, int *qpn, u64 *reg_id)
 {
@@ -585,6 +632,10 @@ static int mlx4_en_get_qp(struct mlx4_en_priv *priv)
 	if (err)
 		goto steer_err;
 
+	if (mlx4_en_tunnel_steer_add(priv, priv->dev->dev_addr, *qpn,
+				     &priv->tunnel_reg_id))
+		goto tunnel_err;
+
 	entry = kmalloc(sizeof(*entry), GFP_KERNEL);
 	if (!entry) {
 		err = -ENOMEM;
@@ -599,6 +650,9 @@ static int mlx4_en_get_qp(struct mlx4_en_priv *priv)
 	return 0;
 
 alloc_err:
+	if (priv->tunnel_reg_id)
+		mlx4_flow_detach(priv->mdev->dev, priv->tunnel_reg_id);
+tunnel_err:
 	mlx4_en_uc_steer_release(priv, priv->dev->dev_addr, *qpn, reg_id);
 
 steer_err:
@@ -642,6 +696,11 @@ static void mlx4_en_put_qp(struct mlx4_en_priv *priv)
 			}
 		}
 
+		if (priv->tunnel_reg_id) {
+			mlx4_flow_detach(priv->mdev->dev, priv->tunnel_reg_id);
+			priv->tunnel_reg_id = 0;
+		}
+
 		en_dbg(DRV, priv, "Releasing qp: port %d, qpn %d\n",
 		       priv->port, qpn);
 		mlx4_qp_release_range(dev, qpn, 1);
@@ -1044,6 +1103,12 @@ static void mlx4_en_do_multicast(struct mlx4_en_priv *priv,
 				if (err)
 					en_err(priv, "Fail to detach multicast address\n");
 
+				if (mclist->tunnel_reg_id) {
+					err = mlx4_flow_detach(priv->mdev->dev, mclist->tunnel_reg_id);
+					if (err)
+						en_err(priv, "Failed to detach multicast address\n");
+				}
+
 				/* remove from list */
 				list_del(&mclist->list);
 				kfree(mclist);
@@ -1061,6 +1126,10 @@ static void mlx4_en_do_multicast(struct mlx4_en_priv *priv,
 				if (err)
 					en_err(priv, "Fail to attach multicast address\n");
 
+				err = mlx4_en_tunnel_steer_add(priv, &mc_list[10], priv->base_qpn,
+							       &mclist->tunnel_reg_id);
+				if (err)
+					en_err(priv, "Failed to attach multicast address\n");
 			}
 		}
 	}
@@ -1598,6 +1667,15 @@ int mlx4_en_start_port(struct net_device *dev)
 		goto tx_err;
 	}
 
+	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
+		err = mlx4_SET_PORT_VXLAN(mdev->dev, priv->port, VXLAN_STEER_BY_OUTER_MAC);
+		if (err) {
+			en_err(priv, "Failed setting port L2 tunnel configuration, err %d\n",
+			       err);
+			goto tx_err;
+		}
+	}
+
 	/* Init port */
 	en_dbg(HW, priv, "Initializing port\n");
 	err = mlx4_INIT_PORT(mdev->dev, priv->port);
@@ -2400,6 +2478,13 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 	if (mdev->dev->caps.steering_mode != MLX4_STEERING_MODE_A0)
 		dev->priv_flags |= IFF_UNICAST_FLT;
 
+	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
+		dev->hw_enc_features |= NETIF_F_IP_CSUM | NETIF_F_RXCSUM |
+					NETIF_F_TSO | NETIF_F_GSO_UDP_TUNNEL;
+		dev->hw_features |= NETIF_F_GSO_UDP_TUNNEL;
+		dev->features    |= NETIF_F_GSO_UDP_TUNNEL;
+	}
+
 	mdev->pndev[port] = dev;
 
 	netif_carrier_off(dev);
@@ -2429,6 +2514,15 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 		goto out;
 	}
 
+	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
+		err = mlx4_SET_PORT_VXLAN(mdev->dev, priv->port, VXLAN_STEER_BY_OUTER_MAC);
+		if (err) {
+			en_err(priv, "Failed setting port L2 tunnel configuration, err %d\n",
+			       err);
+			goto out;
+		}
+	}
+
 	/* Init port */
 	en_warn(priv, "Initializing port\n");
 	err = mlx4_INIT_PORT(mdev->dev, priv->port);
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_resources.c b/drivers/net/ethernet/mellanox/mlx4/en_resources.c
index d3f5086..f1a5500 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_resources.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_resources.c
@@ -68,6 +68,12 @@ void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
 	context->db_rec_addr = cpu_to_be64(priv->res.db.dma << 2);
 	if (!(dev->features & NETIF_F_HW_VLAN_CTAG_RX))
 		context->param3 |= cpu_to_be32(1 << 30);
+
+	if (!is_tx && !rss &&
+	    (mdev->dev->caps.tunnel_offload_mode ==  MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
+		en_dbg(HW, priv, "Setting RX qp %x tunnel mode to RX tunneled & non-tunneled\n", qpn);
+		context->srqn = cpu_to_be32(7 << 28); /* this fills bits 30:28 */
+	}
 }
 
 
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_rx.c b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
index 2aa0ae1..9e67222 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
@@ -631,6 +631,7 @@ int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int bud
 	int ip_summed;
 	int factor = priv->cqe_factor;
 	u64 timestamp;
+	bool l2_tunnel;
 
 	if (!priv->port_up)
 		return 0;
@@ -709,6 +710,8 @@ int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int bud
 		length -= ring->fcs_del;
 		ring->bytes += length;
 		ring->packets++;
+		l2_tunnel = (dev->hw_enc_features & NETIF_F_RXCSUM) &&
+			(cqe->vlan_my_qpn & cpu_to_be32(MLX4_CQE_L2_TUNNEL));
 
 		if (likely(dev->features & NETIF_F_RXCSUM)) {
 			if ((cqe->status & cpu_to_be16(MLX4_CQE_STATUS_IPOK)) &&
@@ -738,6 +741,8 @@ int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int bud
 					gro_skb->data_len = length;
 					gro_skb->ip_summed = CHECKSUM_UNNECESSARY;
 
+					if (l2_tunnel)
+						gro_skb->encapsulation = 1;
 					if ((cqe->vlan_my_qpn &
 					    cpu_to_be32(MLX4_CQE_VLAN_PRESENT_MASK)) &&
 					    (dev->features & NETIF_F_HW_VLAN_CTAG_RX)) {
@@ -790,6 +795,9 @@ int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int bud
 		skb->protocol = eth_type_trans(skb, dev);
 		skb_record_rx_queue(skb, cq->ring);
 
+		if (l2_tunnel)
+			skb->encapsulation = 1;
+
 		if (dev->features & NETIF_F_RXHASH)
 			skb_set_hash(skb,
 				     be32_to_cpu(cqe->immed_rss_invalid),
@@ -1057,6 +1065,12 @@ int mlx4_en_config_rss_steer(struct mlx4_en_priv *priv)
 		rss_mask |=  MLX4_RSS_UDP_IPV4 | MLX4_RSS_UDP_IPV6;
 		rss_context->base_qpn_udp = rss_context->default_qpn;
 	}
+
+	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
+		en_info(priv, "Setting RSS context tunnel type to RSS on inner headers\n");
+		rss_mask |= MLX4_RSS_BY_INNER_HEADERS;
+	}
+
 	rss_context->flags = rss_mask;
 	rss_context->hash_fn = MLX4_RSS_HASH_TOP;
 	for (i = 0; i < 10; i++)
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_tx.c b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
index e3adceb..160e86d 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
@@ -39,6 +39,7 @@
 #include <linux/if_vlan.h>
 #include <linux/vmalloc.h>
 #include <linux/tcp.h>
+#include <linux/ip.h>
 #include <linux/moduleparam.h>
 
 #include "mlx4_en.h"
@@ -560,7 +561,10 @@ static int get_real_size(struct sk_buff *skb, struct net_device *dev,
 	int real_size;
 
 	if (skb_is_gso(skb)) {
-		*lso_header_size = skb_transport_offset(skb) + tcp_hdrlen(skb);
+		if (skb->encapsulation)
+			*lso_header_size = (skb_inner_transport_header(skb) - skb->data) + inner_tcp_hdrlen(skb);
+		else
+			*lso_header_size = skb_transport_offset(skb) + tcp_hdrlen(skb);
 		real_size = CTRL_SIZE + skb_shinfo(skb)->nr_frags * DS_SIZE +
 			ALIGN(*lso_header_size + 4, DS_SIZE);
 		if (unlikely(*lso_header_size != skb_headlen(skb))) {
@@ -859,6 +863,14 @@ netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
 		tx_info->inl = 1;
 	}
 
+	if (skb->encapsulation) {
+		struct iphdr *ipv4 = (struct iphdr *)skb_inner_network_header(skb);
+		if (ipv4->protocol == IPPROTO_TCP || ipv4->protocol == IPPROTO_UDP)
+			op_own |= cpu_to_be32(MLX4_WQE_CTRL_IIP | MLX4_WQE_CTRL_ILP);
+		else
+			op_own |= cpu_to_be32(MLX4_WQE_CTRL_IIP);
+	}
+
 	ring->prod += nr_txbb;
 
 	/* If we used a bounce buffer then copy descriptor back into place */
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
index 123714c..766691c 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
@@ -436,6 +436,7 @@ struct mlx4_en_mc_list {
 	enum mlx4_en_mclist_act	action;
 	u8			addr[ETH_ALEN];
 	u64			reg_id;
+	u64			tunnel_reg_id;
 };
 
 struct mlx4_en_frag_info {
@@ -567,7 +568,7 @@ struct mlx4_en_priv {
 	struct list_head filters;
 	struct hlist_head filter_hash[1 << MLX4_EN_FILTER_HASH_SHIFT];
 #endif
-
+	u64 tunnel_reg_id;
 };
 
 enum mlx4_en_wol {
-- 
1.7.1

Re: [PATCH net-next V2 2/2] net/mlx4_en: Add netdev support for TCP/IP offloads of vxlan tunneling

[David Miller]

From: Or Gerlitz <ogerlitz@mellanox.com>
Date: Sun, 22 Dec 2013 17:26:12 +0200

> +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
> +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
> +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
> +	    (mdev->dev->caps.tunnel_offload_mode ==  MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
> +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {

Way too many parenthesis, this isn't LISP :-)

Re: [PATCH net-next V2 2/2] net/mlx4_en: Add netdev support for TCP/IP offloads of vxlan tunneling

[Joe Perches]

On Sun, 2013-12-22 at 18:12 -0500, David Miller wrote:
> From: Or Gerlitz <ogerlitz@mellanox.com>
> Date: Sun, 22 Dec 2013 17:26:12 +0200
> 
> > +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
> > +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
> > +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
> > +	    (mdev->dev->caps.tunnel_offload_mode ==  MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
> > +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
> 
> Way too many parenthesis, this isn't LISP :-)

Is patchwork stuttering?
I don't see that in the patch I see.

@@ -2400,6 +2478,13 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
        if (mdev->dev->caps.steering_mode != MLX4_STEERING_MODE_A0)
                dev->priv_flags |= IFF_UNICAST_FLT;
 
+       if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
+               dev->hw_enc_features |= NETIF_F_IP_CSUM | NETIF_F_RXCSUM |
+                                       NETIF_F_TSO | NETIF_F_GSO_UDP_TUNNEL;
+               dev->hw_features |= NETIF_F_GSO_UDP_TUNNEL;
+               dev->features    |= NETIF_F_GSO_UDP_TUNNEL;
+       }

Re: [PATCH net-next V2 2/2] net/mlx4_en: Add netdev support for TCP/IP offloads of vxlan tunneling

[David Miller]

From: Joe Perches <joe@perches.com>
Date: Sun, 22 Dec 2013 15:32:39 -0800

> On Sun, 2013-12-22 at 18:12 -0500, David Miller wrote:
>> From: Or Gerlitz <ogerlitz@mellanox.com>
>> Date: Sun, 22 Dec 2013 17:26:12 +0200
>> 
>> > +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
>> > +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
>> > +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
>> > +	    (mdev->dev->caps.tunnel_offload_mode ==  MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
>> > +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
>> 
>> Way too many parenthesis, this isn't LISP :-)
> 
> Is patchwork stuttering?

I edited it, I'm quoting all of the lines in your patch that add the
"too many parenthsis" problem.

[PATCH] checkpatch: Add warning with unnecessary parentheses in if statements

[Joe Perches]

Using doubled parentheses in statements like "if ((foo == bar))"
is unnecessary or may be a sign of an intended single assignment
instead of a comparison.

Bleat a warning message on those uses.

Signed-off-by: Joe Perches <joe@perches.com>
---
> > On Sun, 2013-12-22 at 18:12 -0500, David Miller wrote:
> >> > +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
> >> > +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
> >> > +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
> >> > +	    (mdev->dev->caps.tunnel_offload_mode ==  MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
> >> > +	if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
> >> 
> >> Way too many parenthesis, this isn't LISP :-)
> > 
> > Is patchwork stuttering?
> 
> I edited it, I'm quoting all of the lines in your patch that add the
> "too many parenthsis" problem.

OK then.  Or's patch, not mine, but maybe this is useful...

(this also happened today with another nominal cleanup patch)

 scripts/checkpatch.pl | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl
index 8f3aecd..29af02a 100755
--- a/scripts/checkpatch.pl
+++ b/scripts/checkpatch.pl
@@ -3251,6 +3251,13 @@ sub process {
 			}
 		}
 
+# if statements using unnecessary parentheses - ie: if ((foo == bar))
+		if ($^V && $^V ge 5.10.0 &&
+		    $line =~ /\bif\s*\(\s*\(\s*$LvalOrFunc\s*$Compare\s*$LvalOrFunc\s*\)\s*\)/) {
+			WARN("UNNECESSARY_PARENTHESES",
+			     "Unnecessary parentheses\n" . $herecurr);
+		}
+
 # Return of what appears to be an errno should normally be -'ve
 		if ($line =~ /^.\s*return\s*(E[A-Z]*)\s*;/) {
 			my $name = $1;

Re: [PATCH net-next V2 2/2] net/mlx4_en: Add netdev support for TCP/IP offloads of vxlan tunneling

[Or Gerlitz]

On Mon, Dec 23, 2013 at 1:12 AM, David Miller <davem@davemloft.net> wrote:
> From: Or Gerlitz <ogerlitz@mellanox.com>
> Date: Sun, 22 Dec 2013 17:26:12 +0200
>
>> +     if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
>> +     if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
>> +     if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
>> +         (mdev->dev->caps.tunnel_offload_mode ==  MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
>> +     if ((mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
>
> Way too many parenthesis, this isn't LISP :-)

mmm, not sure how this happened... anyway, will fix and submit V3