Fwd: ipfw, pipes, queues, and IPv6

[Luigi Rizzo]

ipfw/ipfw2 will only see non-ipv4 packets at the layer2 level, i.e.
if you have net.link.ether.bridge.ipfw=1 or net.link.ether.ipfw=1

So you cannot do much classification except for the protocol type.

Re. ipv6 and pipes, i do not see any dummynet hooks in netinet6/
so almost surely it won't work.

The long term plan is to add ipv6-specific instructions to ipfw2
so you can replace ip6_fw with ipfw2, and then you "only" need
to add the post-ipfw processing from netinet/ip_{input,output}.c
in the corresponding netinet6/ files.

	cheers
	luigi


On Sun, Sep 21, 2003 at 08:10:29PM -0400, Michael W. Oliver wrote:
Content-Description: signed data
> [Folks, I posted this to ipfw@ last week, but had no takers over there.  
> Anyone here want to take a stab at it?  Thanks in advance!]
> 
> I am looking for a little sanity check here.  I have a gif tunnel set up for 
> IPv6, and I am passing protocol 41 through the ipfw pipe/queue mechanism 
> to/from my tunnel provider (part of a grander pipe/queue scheme due to my 
> new Vonage service :).  A couple of questions:
> 
> 1) Am I correct in assuming that all IPv6 traffic that I am permitting via
> ip6fw is governed by the pipe/queue rules that I have configured for 
> protocol 41 in ipfw?
> 
> 2) If #1 is true, is there a way for me to use pipes/queues within ip6fw
> itself to discriminate by type of IPv6 traffic?
> 
> Thanks for the info!
> 
> -- 
> Mike
> perl -e 'print unpack("u","88V]N=&%C=\"!I;F9O(&EN(&AE861E<G,*");'
> 
>