net/mlx5e: For TC offloads, always add new flow instead of appending the actions
authorPaul Blakey <paulb@mellanox.com>
Sun, 2 Sep 2018 06:23:24 +0000 (09:23 +0300)
committerSaeed Mahameed <saeedm@mellanox.com>
Wed, 17 Oct 2018 21:18:50 +0000 (14:18 -0700)
commit42f7ad6760077e125eb7849c5499b7b77a54c2f3
tree5cb5c377a2a11ae80a2b6a66b0fabd1d99efdd1a
parentd5634fee245f9e92787e3a34ef621fc12b2cbf16
net/mlx5e: For TC offloads, always add new flow instead of appending the actions

When replacing a tc flower rule, flower first requests to add the
new rule (new action), then deletes the old one.
But currently when asked to add a new tc flower flow, we append the
actions (and counters to it).

This can result in a fte with two flow counters or conflicting
actions (drop and encap action) which firmware complains/errs
about and isn't achieving what the user aimed for.

Instead, insert the flow using the new no-append flag which will add a
new HW rule, the old flow and rule will be deleted later by flower

Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanmox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c