DHCP issues in v3

[Arran Cudbard-Bell]

On 11 Oct 2013, at 22:23, Matthew Newton <mcn4 at LEICESTER.AC.UK> wrote:

> On Fri, Oct 11, 2013 at 08:31:50AM -0400, Alan DeKok wrote:
>>> Next - with as above, and also "interface = br0". This time, no warning about
>>> unicast only, but it still didn't see the broadcast packets.  This doesn't
>>> surprise me entirely, as I've specifically set an IP address, but there is no
>>> warning.
>> 
>>  IIRC, it's an issue with bridges.  As always, the server is at the
>> mercy of the OS.  It asks the OS for broadcast packets.  If the OS
>> doesn't send them...
> 
> Agreed. Seems a bit surprising though. I'll test with plain eth
> device (bridges are common here, VMs all over the place...)
> 
>>> Next I have "ipaddr = *" and again commented out the interface option (not
>>> expecting it necessarily to work - the warning appears again). This works fine
>>> until FR tries to send a reply, and we get a segfault.
>> 
>>  Arran just committed a fix for that.
> 
> Didn't get a segfault in testing just now, but it still didn't
> send any data in the outgoing packets.
> 
> Will report back when I've done some more testing.
> 

Ok, well, I just made copying the interface name conditional, hopefully the ARP subsystem will be smart enough to figure out interfaces based on the routing table.

Arran Cudbard-Bell <a.cudbardb at freeradius.org>
FreeRADIUS Development Team