samba - May 2009 - broken swat when using socket address=xxx.xxx.xxx.xxx, v3.2.8

We are trying out samba 3.2.8.  All of our samba servers have multiple network
interfaces.  With version 3.0.31 we were able to use
                Socket address = xxx.xxx.xxx.xxx, 127.0.0.1
and nmbd worked fine as did swat.

With 3.2.8 that syntax causes nmbd to crash and complain about an invalid
address.
If we adjust the syntax to use
Socket address = xxx.xxx.xxx.xxx
Or           socket address = xxx.xxx.xxx.xxx 127.0.0.1
Nmbd starts and continues to run.

However, the swat status tab shows nmbd as "not running" when it IS
running.

Is this a known issue or can we "configure" around it.


Our v3.2.8 globals:
[global]
        workgroup = OUR_WORKGROUP
        realm = SOMEWHERE.EDU
        netbios aliases = ALIAS3
        server string = Our File Server
        interfaces = lo, 127.0.0.1, eth0:2
        bind interfaces only = Yes
        security = DOMAIN
        password server = xxx.xxx.xxx.xxx, xxx.xxx.xxx.xxx
        log level = 0
        syslog = 0
        syslog only = Yes
        max log size = 50000
        name resolve order = host
        socket address = xxx.xxx.xxx.xxx 127.0.0.1
        socket options = TCP_NODELAY IPTOS_LOWDELAY SO_RCVBUF=20480
SO_SNDBUF=20480
        load printers = No
        printcap name = /dev/null
        usershare path = /usr/lib/samba/usershares
        template homedir = /home/%U
        hosts deny = all
        block size = 4096
        veto oplock files = /*.*db/
        strict locking = No



Helen