- completely rewrote the wins_srv.c code. It is now much simpler, and

gives us a good grounding to properly support multiple wins servers
for different interfaces (which will be coming soon ...)

- fixed our wins registration failover code to actually do failover!
We were not trying to register with a secondary wins server at all
when the primary was down. We now fallback correctly.

- fixed the multi-homed name registration packets so that they work
even in a non-connected network (ie. when one of our interfaces is not
routable from the wins server. Yes, this really happens in the real
world).
(This used to be commit a049360d5b0d95a935b06aad43efc17d34de46dc)

1 files changed, 12 insertions, 4 deletions

diff --git a/source3/nmbd/nmbd_packets.c b/source3/nmbd/nmbd_packets.c
index a11b30b1dc..b5741caae0 100644
--- a/source3/nmbd/nmbd_packets.c
+++ b/source3/nmbd/nmbd_packets.c
@@ -264,11 +264,19 @@ static BOOL create_and_init_additional_record(struct packet_struct *packet,
   /* Set the address for the name we are registering. */
   putip(&nmb->additional->rdata[2], register_ip);
 
-  /* Ensure that we send out the file descriptor to give us the
-     the specific source address we are registering as our
-     IP source address. */
-
+#if 0
+  /* I removed this forced source IP as it breaks wins failover. The
+     problem is that our 2nd interface IP may not be routable to the
+     wins server, in which case all these nice packets we are senidng
+     out will never get a response and we end up marking a perfectly good wins server dead. 
+
+     In general I can't see any reason why we should force the source
+     ip on a packet anyway. We should just let the kernels routin
+     table take care of it, as that is the only place which really
+     knows what is routable and what isn't. (tridge)
+  */
   packet->fd = find_subnet_fd_for_address( *register_ip );
+#endif
 
   return True;
 }