Description:

After a system upgrade and a reboot my VPN setup stopped working.
The full description with a possible workaround follows, but action should be taken to fix the bug (possibly in netcfg directly)

I have a tap interfaces my openvpn instance relies upon.
It's defined as follow in /etc/network.d/tap0:
INTERFACE='tap0'
CONNECTION='tuntap'
MODE='tap'
USER='nobody'
GROUP='nobody'

and correctly configured to be started up at boot in rc.conf (amongst other):
NETWORKS=(tap0 intra extra)

I upgraded the system today, current netcfg version 2.8.5-3 and kernel is:
$uname -a
Linux bastione 3.4.4-2-ARCH #1 SMP PREEMPT Sun Jun 24 17:28:37 UTC 2012 i686 GNU/Linux

after a reboot my VPN stopped properly working, the client could get an IP address (strangely enough, and via a DHCP server listening on the bridge interface tap0 is part of) BUT COULD NOT work properly on the local lan.

To fix the problem I had to manually issue (on the server of course)
sudo ifconfig tap0 up

Right after launching the command my client started working in the LAN, without the need to restart openvpn neither on the client nor on the server.

tcpdump was also refusing to dump traffic on that interface before enabling it.

Steps to reproduce:
My system is acting as a firewall, clients connect to the VPN as part of

set up openvpn to use tap0, which is created automatically at boot via netcfg
add tap0 as a brige with eth0 using netcfg
restart the system

confirm that before manually forcing tap0 up the client cannot connect properly to services hosted at eth0