On Thu, 25 Jan 2018 09:03:48 +0100, Jiri Pirko wrote: > >@@ -144,9 +144,6 @@ int nsim_bpf_setup_tc_block_cb(enum tc_setup_type type, > > return -EOPNOTSUPP; > > } > > > >- if (cls_bpf->common.chain_index) > >- return -EOPNOTSUPP; > >- > > For nfp and netdevsim you do this in a patch that introduces the helper, > for the rest you have a separate patch? Why this inconsistency? > Again, from my perspective, this can be done in a single patch for all > drivers. All the same hunks.
Ah, indeed, slightly sloppy. This patch used to remove the tc_can_offload_extack() helper hence the conversion here. Let me split them out.
