diff options
| author | Saeed Mahameed <saeedm@mellanox.com> | 2019-05-15 02:21:35 -0700 |
|---|---|---|
| committer | Saeed Mahameed <saeedm@mellanox.com> | 2020-03-09 16:58:44 -0700 |
| commit | 20f7b37ffc7da5beb9c98382ca70b918f2282060 (patch) | |
| tree | d5fb14992a8e2792045770cdf31c5f13e2b245c2 /drivers/net/ethernet/mellanox/mlx5/core/en_rep.h | |
| parent | cc617ceda0ebad888d592c6789276136d61d4a88 (diff) | |
net/mlx5e: Introduce root ft concept for representors netdevs
Uplink representor traffic will be redirected to an empty root ft rather
than directly to a direct tir or ttc table, this root ft will be empty and
will be used as a link for auto-chaining with ttc table or ethtool tables
in downstream patches.
On load, fs core will connect uplink rep root_ft with ttc table. In case
ethtool steering will be used, fs core will auto connect root_ft with
the ethtool bypass tables, which will be connected with the ttc table.
vport_rx_rule[uplink_rep]->root_ft->ethtool->ttc.
For non-uplink representors, for simplicity root_ft will always point at
ttc table, hence the replace vport_rx rule logic is removed.
vport_rx_rule[non_uplink_rep]->root_ft(ttc).
For now ethtool steering support can only be available on uplink rep.
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Diffstat (limited to 'drivers/net/ethernet/mellanox/mlx5/core/en_rep.h')
| -rw-r--r-- | drivers/net/ethernet/mellanox/mlx5/core/en_rep.h | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h index 8336301476a9..3d9c72eee9fa 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h @@ -87,6 +87,7 @@ struct mlx5e_rep_priv { struct mlx5_eswitch_rep *rep; struct mlx5e_neigh_update_table neigh_update; struct net_device *netdev; + struct mlx5_flow_table *root_ft; struct mlx5_flow_handle *vport_rx_rule; struct list_head vport_sqs_list; struct mlx5_rep_uplink_priv uplink_priv; /* valid for uplink rep */ |