This should solve problems people are seeing when a file changes
twice within a second and thus is missed for polling.

While testing https://github.com/sbyx/odhcp6c client I have noticed it
permanently crashes after startup.

The reason was it (odhcp6c) doesn't expect empty IA options in ADVERTISE
message without any suboptions.

Despite this validation bug of odhcp6c, dnsmasq should not generate
ADVERTISE messages with IA if there's nothing to advert per  RFC 3315
17.2.2:

   If the server will not assign any addresses to any IAs in a

   subsequent Request from the client, the server MUST send an Advertise

   message to the client that includes only a Status Code option with

   code NoAddrsAvail and a status message for the user, a Server

   Identifier option with the server's DUID, and a Client Identifier

   option with the client's DUID.

Meanwhile it's need to add status code for every IA in REPLY message per
RFC3315 18.2.1:

   If the server cannot assign any addresses to an IA in the message
   from the client, the server MUST include the IA in the Reply message
   with no addresses in the IA and a Status Code option in the IA
   containing status code NoAddrsAvail.

So, I've changed the logic to skip IA completely from ADVERTISE messages and
to add NoAddrsAvail subcode into IA of REPLY messages.

As for overhead, yes, I believe it's ok to return NoAddrsAvail twice in IA
and in global section for compatibility with all old and new clients.

Simon's fault.

SO_MARK has to be done before issuing the connect() call on the
TCP socket.

check_for_local_domain() was broken due to new code matching F_*
bits in cache entries for DNSSEC. Because F_DNSKEY | F_DS is
used to match RRSIG entries, cache_find_by_name() insists on an exact match
of those bits. So adding F_DS to the bits that check_for_local_domain()
sends to cache_find_by_name() won't result in DS records as well
as the others, it results in only DS records. Add a new bit, F_NSIGMATCH
which suitably changes the behaviour of cache_find_by_name().

http://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2014q3/008796.html

This handles the case that more than one interface contains
the network the lease address is on, but the interfaces have different
prefix lengths. Use the longest prefix length.

- With nested prefixes reside on different interfaces of single host

  (e.g., in 6to4, 2002::/16 on WAN and 2002:<IPv4>:<subnet>::/64 on LAN),

  current matching mechanism might return the interface with shorter prefix

  length instead of the longer one, if it appears later in the netlink message.
Signed-off-by: Lung-Pin Chang <changlp@cs.nctu.edu.tw>