Merge tag 'mlx5e-updates-2018-12-14' of git://git.kernel.org/pub/scm/…

…linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5e-updates-2018-12-14 (VF Lag)

From Aviv Heller,

Subsequent patches introduce VF LAG, which provdies load-balancing and
high-availability capabilities for VFs associated with different
physical ports of the same Connect-X card.

This series consists of the following:
 - mlx5 devcom, driver infrastructure that facilitates operations that involve
   both core devices (physical functions) of the same card, to synchronize and
   communicate between two driver instances of the same card.
 - Infrastructure for TC rule duplication.
 - Changes to LAG logic to enable its use when SR-IOV is enabled
 - PFs in switchdev mode is the only mode currently supported.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/infiniband/hw/mlx5/ib_rep.c

@@ -48,13 +48,21 @@ static const struct mlx5_ib_profile rep_profile = {
 static int
 mlx5_ib_nic_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
 {
+	struct mlx5_ib_dev *ibdev;
+
+	ibdev = mlx5_ib_rep_to_dev(rep);
+	if (!__mlx5_ib_add(ibdev, ibdev->profile))
+		return -EINVAL;
 	return 0;
 }
 
 static void
 mlx5_ib_nic_rep_unload(struct mlx5_eswitch_rep *rep)
 {
-	rep->rep_if[REP_IB].priv = NULL;
+	struct mlx5_ib_dev *ibdev;
+
+	ibdev = mlx5_ib_rep_to_dev(rep);
+	__mlx5_ib_remove(ibdev, ibdev->profile, MLX5_IB_STAGE_MAX);
 }
 
 static int
@@ -89,6 +97,7 @@ mlx5_ib_vport_rep_unload(struct mlx5_eswitch_rep *rep)
 	dev = mlx5_ib_rep_to_dev(rep);
 	__mlx5_ib_remove(dev, dev->profile, MLX5_IB_STAGE_MAX);
 	rep->rep_if[REP_IB].priv = NULL;
+	ib_dealloc_device(&dev->ib_dev);
 }
 
 static void *mlx5_ib_vport_get_proto_dev(struct mlx5_eswitch_rep *rep)

drivers/infiniband/hw/mlx5/main.c

@@ -445,7 +445,7 @@ static int mlx5_query_port_roce(struct ib_device *device, u8 port_num,
 	if (!ndev)
 		goto out;
 
-	if (mlx5_lag_is_active(dev->mdev)) {
+	if (dev->lag_active) {
 		rcu_read_lock();
 		upper = netdev_master_upper_dev_get_rcu(ndev);
 		if (upper) {
@@ -1848,7 +1848,7 @@ static struct ib_ucontext *mlx5_ib_alloc_ucontext(struct ib_device *ibdev,
 	context->lib_caps = req.lib_caps;
 	print_lib_caps(dev, context->lib_caps);
 
-	if (mlx5_lag_is_active(dev->mdev)) {
+	if (dev->lag_active) {
 		u8 port = mlx5_core_native_port_num(dev->mdev);
 
 		atomic_set(&context->tx_port_affinity,
@@ -4841,7 +4841,7 @@ static int mlx5_eth_lag_init(struct mlx5_ib_dev *dev)
 	struct mlx5_flow_table *ft;
 	int err;
 
-	if (!ns || !mlx5_lag_is_active(mdev))
+	if (!ns || !mlx5_lag_is_roce(mdev))
 		return 0;
 
 	err = mlx5_cmd_create_vport_lag(mdev);
@@ -4855,6 +4855,7 @@ static int mlx5_eth_lag_init(struct mlx5_ib_dev *dev)
 	}
 
 	dev->flow_db->lag_demux_ft = ft;
+	dev->lag_active = true;
 	return 0;
 
 err_destroy_vport_lag:
@@ -4866,7 +4867,9 @@ static void mlx5_eth_lag_cleanup(struct mlx5_ib_dev *dev)
 {
 	struct mlx5_core_dev *mdev = dev->mdev;
 
-	if (dev->flow_db->lag_demux_ft) {
+	if (dev->lag_active) {
+		dev->lag_active = false;
+
 		mlx5_destroy_flow_table(dev->flow_db->lag_demux_ft);
 		dev->flow_db->lag_demux_ft = NULL;
 
@@ -6173,7 +6176,7 @@ int mlx5_ib_stage_ib_reg_init(struct mlx5_ib_dev *dev)
 	const char *name;
 
 	rdma_set_device_sysfs_group(&dev->ib_dev, &mlx5_attr_group);
-	if (!mlx5_lag_is_active(dev->mdev))
+	if (!mlx5_lag_is_roce(dev->mdev))
 		name = "mlx5_%d";
 	else
 		name = "mlx5_bond_%d";
@@ -6207,18 +6210,6 @@ static void mlx5_ib_stage_delay_drop_cleanup(struct mlx5_ib_dev *dev)
 	cancel_delay_drop(dev);
 }
 
-static int mlx5_ib_stage_rep_reg_init(struct mlx5_ib_dev *dev)
-{
-	mlx5_ib_register_vport_reps(dev);
-
-	return 0;
-}
-
-static void mlx5_ib_stage_rep_reg_cleanup(struct mlx5_ib_dev *dev)
-{
-	mlx5_ib_unregister_vport_reps(dev);
-}
-
 static int mlx5_ib_stage_dev_notifier_init(struct mlx5_ib_dev *dev)
 {
 	dev->mdev_events.notifier_call = mlx5_ib_event;
@@ -6257,8 +6248,6 @@ void __mlx5_ib_remove(struct mlx5_ib_dev *dev,
 		if (profile->stage[stage].cleanup)
 			profile->stage[stage].cleanup(dev);
 	}
-
-	ib_dealloc_device((struct ib_device *)dev);
 }
 
 void *__mlx5_ib_add(struct mlx5_ib_dev *dev,
@@ -6392,9 +6381,6 @@ static const struct mlx5_ib_profile nic_rep_profile = {
 	STAGE_CREATE(MLX5_IB_STAGE_POST_IB_REG_UMR,
 		     mlx5_ib_stage_post_ib_reg_umr_init,
 		     NULL),
-	STAGE_CREATE(MLX5_IB_STAGE_REP_REG,
-		     mlx5_ib_stage_rep_reg_init,
-		     mlx5_ib_stage_rep_reg_cleanup),
 };
 
 static void *mlx5_ib_add_slave_port(struct mlx5_core_dev *mdev)
@@ -6462,8 +6448,9 @@ static void *mlx5_ib_add(struct mlx5_core_dev *mdev)
 	if (MLX5_ESWITCH_MANAGER(mdev) &&
 	    mlx5_ib_eswitch_mode(mdev->priv.eswitch) == SRIOV_OFFLOADS) {
 		dev->rep = mlx5_ib_vport_rep(mdev->priv.eswitch, 0);
-
-		return __mlx5_ib_add(dev, &nic_rep_profile);
+		dev->profile = &nic_rep_profile;
+		mlx5_ib_register_vport_reps(dev);
+		return dev;
 	}
 
 	return __mlx5_ib_add(dev, &pf_profile);
@@ -6485,7 +6472,12 @@ static void mlx5_ib_remove(struct mlx5_core_dev *mdev, void *context)
 	}
 
 	dev = context;
-	__mlx5_ib_remove(dev, dev->profile, MLX5_IB_STAGE_MAX);
+	if (dev->profile == &nic_rep_profile)
+		mlx5_ib_unregister_vport_reps(dev);
+	else
+		__mlx5_ib_remove(dev, dev->profile, MLX5_IB_STAGE_MAX);
+
+	ib_dealloc_device((struct ib_device *)dev);
 }
 
 static struct mlx5_interface mlx5_ib_interface = {

drivers/infiniband/hw/mlx5/mlx5_ib.h

@@ -790,7 +790,6 @@ enum mlx5_ib_stages {
 	MLX5_IB_STAGE_POST_IB_REG_UMR,
 	MLX5_IB_STAGE_DELAY_DROP,
 	MLX5_IB_STAGE_CLASS_ATTR,
-	MLX5_IB_STAGE_REP_REG,
 	MLX5_IB_STAGE_MAX,
 };
 
@@ -937,6 +936,7 @@ struct mlx5_ib_dev {
 	struct mlx5_ib_delay_drop	delay_drop;
 	const struct mlx5_ib_profile	*profile;
 	struct mlx5_eswitch_rep		*rep;
+	int				lag_active;
 
 	struct mlx5_ib_lb_state		lb;
 	u8			umr_fence;

drivers/infiniband/hw/mlx5/qp.c

@@ -3258,7 +3258,7 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
 		    (ibqp->qp_type == IB_QPT_RAW_PACKET) ||
 		    (ibqp->qp_type == IB_QPT_XRC_INI) ||
 		    (ibqp->qp_type == IB_QPT_XRC_TGT)) {
-			if (mlx5_lag_is_active(dev->mdev)) {
+			if (dev->lag_active) {
 				u8 p = mlx5_core_native_port_num(dev->mdev);
 				tx_affinity = get_tx_affinity(dev, pd, base, p);
 				context->flags |= cpu_to_be32(tx_affinity << 24);

drivers/net/ethernet/mellanox/mlx5/core/Makefile

@@ -15,7 +15,7 @@ mlx5_core-y :=	main.o cmd.o debugfs.o fw.o eq.o uar.o pagealloc.o \
 		health.o mcg.o cq.o alloc.o qp.o port.o mr.o pd.o \
 		mad.o transobj.o vport.o sriov.o fs_cmd.o fs_core.o \
 		fs_counters.o rl.o lag.o dev.o events.o wq.o lib/gid.o \
-		diag/fs_tracepoint.o diag/fw_tracer.o
+		lib/devcom.o diag/fs_tracepoint.o diag/fw_tracer.o
 
 #
 # Netdev basic

drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c

@@ -14,7 +14,8 @@ static int mlx5e_route_lookup_ipv4(struct mlx5e_priv *priv,
 				   u8 *out_ttl)
 {
 	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
-	struct mlx5e_rep_priv *uplink_rpriv;
+	struct net_device *uplink_dev, *uplink_upper;
+	bool dst_is_lag_dev;
 	struct rtable *rt;
 	struct neighbour *n = NULL;
 
@@ -28,10 +29,20 @@ static int mlx5e_route_lookup_ipv4(struct mlx5e_priv *priv,
 #else
 	return -EOPNOTSUPP;
 #endif
-	uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
-	/* if the egress device isn't on the same HW e-switch, we use the uplink */
-	if (!switchdev_port_same_parent_id(priv->netdev, rt->dst.dev))
-		*out_dev = uplink_rpriv->netdev;
+
+	uplink_dev = mlx5_eswitch_uplink_get_proto_dev(esw, REP_ETH);
+	uplink_upper = netdev_master_upper_dev_get(uplink_dev);
+	dst_is_lag_dev = (uplink_upper &&
+			  netif_is_lag_master(uplink_upper) &&
+			  rt->dst.dev == uplink_upper &&
+			  mlx5_lag_is_sriov(priv->mdev));
+
+	/* if the egress device isn't on the same HW e-switch or
+	 * it's a LAG device, use the uplink
+	 */
+	if (!switchdev_port_same_parent_id(priv->netdev, rt->dst.dev) ||
+	    dst_is_lag_dev)
+		*out_dev = uplink_dev;
 	else
 		*out_dev = rt->dst.dev;
 
@@ -65,8 +76,9 @@ static int mlx5e_route_lookup_ipv6(struct mlx5e_priv *priv,
 	struct dst_entry *dst;
 
 #if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6)
-	struct mlx5e_rep_priv *uplink_rpriv;
 	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	struct net_device *uplink_dev, *uplink_upper;
+	bool dst_is_lag_dev;
 	int ret;
 
 	ret = ipv6_stub->ipv6_dst_lookup(dev_net(mirred_dev), NULL, &dst,
@@ -77,10 +89,19 @@ static int mlx5e_route_lookup_ipv6(struct mlx5e_priv *priv,
 	if (!(*out_ttl))
 		*out_ttl = ip6_dst_hoplimit(dst);
 
-	uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
-	/* if the egress device isn't on the same HW e-switch, we use the uplink */
-	if (!switchdev_port_same_parent_id(priv->netdev, dst->dev))
-		*out_dev = uplink_rpriv->netdev;
+	uplink_dev = mlx5_eswitch_uplink_get_proto_dev(esw, REP_ETH);
+	uplink_upper = netdev_master_upper_dev_get(uplink_dev);
+	dst_is_lag_dev = (uplink_upper &&
+			  netif_is_lag_master(uplink_upper) &&
+			  dst->dev == uplink_upper &&
+			  mlx5_lag_is_sriov(priv->mdev));
+
+	/* if the egress device isn't on the same HW e-switch or
+	 * it's a LAG device, use the uplink
+	 */
+	if (!switchdev_port_same_parent_id(priv->netdev, dst->dev) ||
+	    dst_is_lag_dev)
+		*out_dev = uplink_dev;
 	else
 		*out_dev = dst->dev;
 #else

drivers/net/ethernet/mellanox/mlx5/core/en_rep.c

@@ -297,17 +297,31 @@ static const struct ethtool_ops mlx5e_rep_ethtool_ops = {
 int mlx5e_attr_get(struct net_device *dev, struct switchdev_attr *attr)
 {
 	struct mlx5e_priv *priv = netdev_priv(dev);
-	struct mlx5e_rep_priv *rpriv = priv->ppriv;
-	struct mlx5_eswitch_rep *rep = rpriv->rep;
 	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	struct net_device *uplink_upper = NULL;
+	struct mlx5e_priv *uplink_priv = NULL;
+	struct net_device *uplink_dev;
 
 	if (esw->mode == SRIOV_NONE)
 		return -EOPNOTSUPP;
 
+	uplink_dev = mlx5_eswitch_uplink_get_proto_dev(esw, REP_ETH);
+	if (uplink_dev) {
+		uplink_upper = netdev_master_upper_dev_get(uplink_dev);
+		uplink_priv = netdev_priv(uplink_dev);
+	}
+
 	switch (attr->id) {
 	case SWITCHDEV_ATTR_ID_PORT_PARENT_ID:
 		attr->u.ppid.id_len = ETH_ALEN;
-		ether_addr_copy(attr->u.ppid.id, rep->hw_id);
+		if (uplink_upper && mlx5_lag_is_sriov(uplink_priv->mdev)) {
+			ether_addr_copy(attr->u.ppid.id, uplink_upper->dev_addr);
+		} else {
+			struct mlx5e_rep_priv *rpriv = priv->ppriv;
+			struct mlx5_eswitch_rep *rep = rpriv->rep;
+
+			ether_addr_copy(attr->u.ppid.id, rep->hw_id);
+		}
 		break;
 	default:
 		return -EOPNOTSUPP;