NETGRAPH- bridge vlans using netgraph help

[Abdullah Tariq]

>
> You are still ignoring requests to not top-post and I'm getting tired to
> fix it for you.
> I may just stop responding.
>

Apologies again, last i thought i had replied to all after your first
pointing out😷

If you really want to bridge 802.1q trunks, you should be stating so from
> the beginning.
>
Sorry if i couldnt make you understand properly.

  However, we still have several ways to bridge tagged traffic
> by means of creation multiple bridges (one per vlan) or using
> ng_vlan+ng_bridge to do the same.



bridge1 will contain vlan 1
bridge2 will containn vlan2

and bridge3 should have bridge1 and bridge2?

If not, can you please give some instructions/guidelines for
ng_vlan+ng_bridge?
Thanks


On Sat, Apr 28, 2018 at 3:20 PM, Eugene Grosbein <eugen at grosbein.net> wrote:

> 28.04.2018 16:26, Abdullah wrote:
>
> You are still ignoring requests to not top-post and I'm getting tired to
> fix it for you.
> I may just stop responding.
>
> >>     You still top-posting. Please don't.
> >>     It is possible to replicate *behaviour* but not configuration style.
> >>     Unless you have real embedded switching chip in your system like
> many ARM or MIPS-based routers do,
> >>     in which case you can use etherswitch(4) and etherswitchcfg(8) to
> create vlans and configure
> >>     ports of such chip similar to L2 switches.
>
> > Configuration style doesn't really matter.
> > I am trying to figure out how to make tagged traffic talk in  a bridge.
> > I have always had good support from bsd people over the last 4 5 years
> since i started using it, maybe it's time to add something to it :)
> > What gives me hope is that linux (cumulus linux, openswitch, ONL) are
> doing so, i dont see why BSD can't do it. I haven't searched much other bsd
> flavors though.
>
> Traffic of virtual lan (vlan) is not the same as tagged traffic, you know.
> You described a task where all vlan's ports are untagged and this is
> easily doable
> by bridging NIC interfaces without creation of syntetic vlan interface
> that is useless in case of untagged-only ports.
>
> If you really want to bridge 802.1q trunks, you should be stating so from
> the beginning.
> if_bridge(8) does not support tagged frames currently.
>
> It is generally bad idea to build full-blown switch out of FreeBSD unless
> you have mentioned
> companion switching chip. However, we still have several ways to bridge
> tagged traffic
> by means of creation multiple bridges (one per vlan) or using
> ng_vlan+ng_bridge to do the same.
>
>