ocserv-fw script design

Lance LeFlore lance at 3t218.org
Mon May 9 07:34:58 PDT 2016

I can do that.

Would it be possible to get export_fw_info to also export ipv4-network
so that ocserv-fw can configure SNATing on the POSTROUTING chain?

On Mon, May 9, 2016 at 7:45 AM, Nikos Mavrogiannopoulos
<n.mavrogiannopoulos at gmail.com> wrote:
> On Fri, May 6, 2016 at 2:36 PM, Lance LeFlore <lance at 3t218.org> wrote:
>> Hi,
>> I'm trying to understand why the ocserv-fw script makes use of the
>> iptables INPUT chain. One would think that the FORWARD chain would be
>> where the script would put all of the allow/deny rules since traffic
>> will need to be forwarded from the tun interface to eth0 if it wants
>> to get out to other networks. I know that $INPUT_CHAIN can be set to
>> whatever I'd like, but there are several places where the INPUT chain
>> is hardcoded in the script.
>> I'd also like to understand why ocserv-fw appears to create empty
>> INPUT-ocserv-fw-vpns* chains which cause the rules in the INPUT chain
>> which reference said chains to go nowhere useful. It seems to me that
>> the INPUT-ocserv-fw-vpns* chains should each have a blanket allow rule
>> so that the rules in the INPUT (should be FORWARD?) chain which
>> reference them actually work.
>> Am I thinking about this all wrong?
> You are right, it seems that the current script doesn't fulfill its
> purpose; I should have tagged this as an experimental feature. It
> seems that both INPUT/FORWARD and OUTPUT/FORWARD should have been
> specified. Would you be interested in providing a fixed ocserv-fw
> script?
> regards,
> Nikos

More information about the openconnect-devel mailing list