Linux Ethernet Bridging - Oct 2011 - [Bridge] Bridge down when all ports are down

I have problems much like earlier postings:
 - Bridge interface always down since Linux >= 2.6.39
   http://lists.linuxfoundation.org/pipermail/bridge/2011-June/007711.html
 - possible bridge regression in "bridge: implement [add/del]_slave
ops"?
   http://lists.linuxfoundation.org/pipermail/bridge/2011-June/007709.html

In my case, an app sends (multicast) traffic to br0 with eth1 and wlan0 ports.
There are conditions with no wlan0 HW on subset of my gear and eth0 down because
no connected workstation.

Problems:
1) Multicast between apps doesn't work when bridge is down. To send
multicast _also_ on virtual interface needs update in app and is a bit of a
hack, I think.
2) Troubleshoot the multicast packets with tcpdump on br0 doesn't work.

To send multicast _also_ on virtual interface needs update in app. It is also a
bit of a hack, I think.

I tried to add interface lo or tap0 in the bridge. This didn't work.

Is there a trick to always bring up the bridge interface?
Add a config option? Such as brctl autostate <bridge> {on|off}
I don't mind having the default autostate set to on. But is was off before
2.6.39.

Thanks, Teco