call for review: 802.11q QinQ netgraph support

[Adrian Chadd]

2012/2/8 Gleb Smirnoff <glebius at freebsd.org>:

> A> I would appreciate further review from network/netgraph related
> A> people. I'm going to borrow a term from gnn and say "Silence implies
> A> consent." :-)
>
> I have only minor comments:

Thanks for your feedback!

Ivan, can you please review these?


Adrian

>
> 1.1) According to style(9) new code should use uintXX_t instead of u_intXX_t.
> 1.2) Some lines are really loooong, they need to be broken into shorter ones
>     accoording to style(9).
> 1.3) Operators at beginning of a line - also style(9) violation.
>
> 2) NETGRAPH_DEBUG wasn't designed for the things the patch is doing. But
>   the KASSERT was. So it'll be better to change the code under NETGRAPH_DEBUG
>   to KASSERTs. For example in the chunk '@@ -262,35 +322,143 @@':
>
>        KASSERT(priv->vlan_hook[EVL_VLANOFTAG(hook_data)] == hook,
>            ("%s: NGM_VLAN_DEL_FILTER: Invalid VID for Hook = %s\n",
>            __func__, (char *)msg->data));
>
>        and
>
>        KASSERT(EVL_VLANOFTAG(hook_data) == vid,
>            ("%s: NGM_VLAN_DEL_VID_FLT: Invalid VID Hook = %us, must be: %us\n",
>            __func__, (uint16_t )EVL_VLANOFTAG(hook_data), vid));
>
> --
> Totus tuus, Glebius.