Re: [Openvpn-devel] [HACK/PATCH] Prioritize TCP ACKs

2005-11-29 Thread Nate Nielsen
Florian Weimer wrote: >>Here's a hack of a patch which puts nice TOS bits on TCP ACK packets. >>This allows you to use pf or another firewall to prioritize the >>resulting openvpn traffic in the way you normally do over asynchronous >>connections. > > I think you have to make those TOS values conf

Re: [Openvpn-devel] [PATCH] Allow dynamic choice of source port

2005-11-28 Thread Nate Nielsen
James Yonan wrote: > On Fri, 25 Nov 2005, Nielsen wrote: >>As it is openvpn has no way to allow the kernel to dynamically choose >>a source port for it's connections. If you have two openvpn client >>processes on the same machine both you have to configure lport to >&g

[Openvpn-devel] [HACK/PATCH] Prioritize TCP ACKs

2005-11-24 Thread Nielsen
this in openvpn. Just figured someone else out there might have run into the same problems. Cheers, Nate Nielsen -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.1 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFDhqUue/sRCNknZa8RAoD

[Openvpn-devel] [PATCH] Allow dynamic choice of source port

2005-11-24 Thread Nielsen
s you to specify '--lport 0' and get the behaviour of normal programs, where the source port is dynamically chosen. Not sure if this is an issue for others, but here's the patch in case anyone has run into the same issue. Cheers, Nate Nielsen -BEGIN PGP SIGNATURE- Versio

[Openvpn-devel] [PATCH] Openvpn --pass-tos on FreeBSD

2005-11-24 Thread Nielsen
The attached patch makes the --pass-tos option work on FreeBSD. If included in OpenVPN a configure test would probably be necessary. This is one of several things we modified in OpenVPN while evaluating it for deployment. Thought I'd share the fixes. Cheers, Nate Nielsen diff -U3 ../openvp

Re: [Openvpn-devel] new tap device for each connection

2005-11-18 Thread Nate Nielsen
Ian Berry wrote: > I ran into this exact situation. Based on my knowledge of how OpenVPN > operates in "server mode", I figured that it would require a significant > change to allow this. As a workaround, I run OpenVPN in "inetd mode", > which forces a new openvpn process for each client and theref