On Tue, 06 Mar 2018 11:08:26 -0800, Vinicius Costa Gomes wrote:
> >> +static int igb_parse_cls_flower(struct igb_adapter *adapter,
> >> + struct tc_cls_flower_offload *f,
> >> + int traffic_class,
> >> + struct igb_nfc_filter *input)
> >> +{
> >> + if (f->dissector->used_keys &
> >> + ~(BIT(FLOW_DISSECTOR_KEY_BASIC) |
> >> + BIT(FLOW_DISSECTOR_KEY_CONTROL) |
> >> + BIT(FLOW_DISSECTOR_KEY_ETH_ADDRS) |
> >> + BIT(FLOW_DISSECTOR_KEY_VLAN))) {
> >> + dev_err(&adapter->pdev->dev, "Unsupported key used: 0x%x\n",
> >> + f->dissector->used_keys);
> >
> > This will probably trigger for opportunistic offload (non-skip_sw) and
> > confuse users.
>
> I see your point. I will change to 'dev_warn()', it should not surprise
> users too much, right?
Yes, I think that would be fine, other drivers are doing that already.
IMHO best approach is to not print anything unless skip-sw is set. If
you used netlink's extack capability exclusively it would "just work".
Extack will only carry the error in case offload is requested. Could
you consider using extack or do you have a preference to print into the
logs? You could do both, too.
You can get to extack via f->common->extack.
