browsing across subnets

[Andy Worthington]

Ok, I now have the machines visible on each subnet in network neighborhood.
Here is the setup of the network:

Workgroup:MSP - 192.168.4.0
2 Windows ME clients
            |
            |   192.168.4.1
 Linux samba, firewall, ipsec machine
           |  64.218.*.*
           |
Internet
           |
           |  64.217.*.*
Linux samba, firewall, ipsec machine
           | 192.168.3.1
           |
6 Windows 98, ME clients
Workgroup: Irish 192.168.3.0

Both subnets can see all clients but when Irish workgroup clients try to
connect to either of the clients on the other subnet they get an error
message about the client not being found.  Same with clients on the MSP
subnet.  If they try to communicate with the Linux server on the other
subnet that works fine.  Here is what I get if I try to connect from the
Irish linux server to one of the MSP clients:

[root at fw samba]# smbclient -L "Mindy"
doing parameter domain logons = yes
doing parameter logon script = netlogon.bat
doing parameter printcap name = /etc/printcap
doing parameter load printers = yes
doing parameter printing = lprng
doing parameter print command = /usr/bin/lpr -b -h -r -P%p %s
pm_process() returned Yes
added interface ip=127.0.0.1 bcast=127.255.255.255 nmask=255.0.0.0
added interface ip=192.168.3.1 bcast=192.168.3.255 nmask=255.255.255.0
added interface ip=64.217.29.97 bcast=64.217.29.103 nmask=255.255.255.248
Client started (version 2.0.7).
resolve_wins: Attempting wins lookup for name Mindy<0x20>
bind succeeded on port 0
Got a positive name query response from 127.0.0.1 ( 192.168.4.66 )
Connecting to 192.168.4.66 at port 139
timeout connecting to 192.168.4.66:139
Connection to Mindy failed

By that I am assuming it is either a routing or firewall problem keeping the
connection from working.  Pinging the client works fine though.  How can I
find where the problem is to solve it?  The connection between 192.168.3.0
and 192.168.4.0 is via a freeswan ipsec tunnel.

Thanks
Andy