if_stf bug/feature

[Hajimu UMEMOTO]

Hi,

>>>>> On Thu, 6 May 2004 10:21:13 +0200
>>>>> Lukasz Stelmach <Lukasz.Stelmach at telmark.waw.pl> said:

Lukasz> Well i *have*got* one v4ADDR that is assigned to my nat/router-box.  I
Lukasz> have also configured that it should pass all packets that are not part
Lukasz> of some known connections (from M1 M2 .. Mn) (including but not limited
Lukasz> to proto 41) to one specified machine (name it TIGGER) that acts as the
Lukasz> end of 6to4 tunnel for all other computers in the LAN.  Now, for i
Lukasz> controll both the nat and TIGGER i can do such manglig without any
Lukasz> harm. Let's say taht to the rest of the world the nat+TIGGER act like
Lukasz> a single machine.

Yes, current if_stf is too restrictive against NAT, and skipping
certain checks enablea us to use 6to4 even behind NAT.  I believe it
doesn't break RFC3056.
Once, I made a patch to do so for a friend of mine.  But, it was based
on old source and somewhat redundant.  I've just made a patch against
recent 5-CURRENT.  But, I've not estimated if there are side effects.
I don't have testing environment for 6to4, now.  Could you test it?

Sincerely,

--
Hajimu UMEMOTO @ Internet Mutual Aid Society Yokohama, Japan
ume at mahoroba.org  ume@{,jp.}FreeBSD.org
http://www.imasy.org/~ume/