Tor Ipv6-Patch

Karsten Loesing karsten.loesing at
Tue Dec 4 13:42:40 UTC 2007

Hash: SHA1

Hi Marcus,

> I wrote an extensive patch to make tor support
> * hidden services on ipv6-addresses

I've had a quick look on your changes and would like to point out some
issues related to hidden services:

- - You changed encoding of v2 hidden service descriptors, so that
introduction points can be addressed via IPv6 in
rend_encode_v2_intro_points(), but I think you did not change parsing of
those descriptors, so that descriptors containing introduction points
with IPv6 addresses would be rejected; see

- - Even without your changes, the 0.2.0.x code should be more tolerant
with hidden service descriptors containing IPv6 addresses for
introduction points. Currently, the whole descriptor would be rejected.
But if newer Tor versions could encode both introduction points with
IPv4 and IPv6 addresses, current versions could at least keep the IPv4
introduction points and discard only the IPv6 ones. Otherwise we might
need a new descriptor version only for understanding IPv6 addresses,
which does not really make sense.

- - When supporting IPv6 for introduction points and the actual service,
rendezvous points should also be addressable using IPv6. This requires
using v3 INTRODUCE2 cells or something similar (see rend-spec.txt
section 1.8.1). Maybe that cell format should be implemented in this
context, too. However, it still requires some thinking about proposal
121 on hidden service authentication in order to incorporate all
necessary changes in one step. For 0.2.1.x?

> The code may not use the prefered naming-conventions [...]

Hmm, maybe you also want to run "make check-spaces". At least that's
what I was told when starting to write code for Tor. :)

- --Karsten
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla -


More information about the tor-dev mailing list