net: core: rename indirect block ingress cb function
authorJohn Hurley <john.hurley@netronome.com>
Thu, 5 Dec 2019 17:03:34 +0000 (17:03 +0000)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 18 Dec 2019 15:08:47 +0000 (16:08 +0100)
commit1b511a9d2c09bc7a0b0ea3d2b4538547b7615284
tree9667e5f2bb32911d25568137101c6f1a7c322360
parentee0dc0c3f371197ff8dbaf4ce874bef2e33674ea
net: core: rename indirect block ingress cb function

[ Upstream commit dbad3408896c3c5722ec9cda065468b3df16c5bf ]

With indirect blocks, a driver can register for callbacks from a device
that is does not 'own', for example, a tunnel device. When registering to
or unregistering from a new device, a callback is triggered to generate
a bind/unbind event. This, in turn, allows the driver to receive any
existing rules or to properly clean up installed rules.

When first added, it was assumed that all indirect block registrations
would be for ingress offloads. However, the NFP driver can, in some
instances, support clsact qdisc binds for egress offload.

Change the name of the indirect block callback command in flow_offload to
remove the 'ingress' identifier from it. While this does not change
functionality, a follow up patch will implement a more more generic
callback than just those currently just supporting ingress offload.

Fixes: 4d12ba42787b ("nfp: flower: allow offloading of matches on 'internal' ports")
Signed-off-by: John Hurley <john.hurley@netronome.com>
Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
include/net/flow_offload.h
net/core/flow_offload.c
net/netfilter/nf_tables_offload.c
net/sched/cls_api.c