Fix problems with ipset or nftset and TCP DNS transport.

If DNS is happening over TCP, the query is handled by a forked
process. Of ipset ot nftset is configured, this might include
inserting addresses in the *sets. Before this update, that
was done by the forked process using handles inherited from the
parent "master" process.

This is inherently racy. If the master process or another
child process tries to do updates at the same time, the
updates can clash and fail.

To see this, you need a busy server doing lots of DNS
queries over TCP, and ipset or nftset configured.

Going forward, we use the already established pipe to send the
updates from the child back to the master process, which
serialises them.

src/dnsmasq.h

@@ -539,8 +539,10 @@ struct crec {
 
 #define PIPE_OP_RR      1  /* Resource record */
 #define PIPE_OP_END     2  /* Cache entry complete: commit */
-#define PIPE_OP_RESULT  3  /* validation result. */
+#define PIPE_OP_RESULT  3  /* Validation result */
 #define PIPE_OP_STATS   4  /* Update parent's stats */
+#define PIPE_OP_IPSET   5  /* Update IPset */
+#define PIPE_OP_NFTSET  6  /* Update NFTset */
 
 /* struct sockaddr is not large enough to hold any address,
    and specifically not big enough to hold an IPv6 address.
@@ -1361,6 +1363,10 @@ int cache_recv_insert(time_t now, int fd);
 #ifdef HAVE_DNSSEC
 void cache_update_hwm(void);
 #endif
+#if defined(HAVE_IPSET) || defined(HAVE_NFTSET)
+void cache_send_ipset(unsigned char op, struct ipsets *sets,
+		      int flags, union all_addr *addr);
+#endif
 struct crec *cache_insert(char *name, union all_addr *addr, unsigned short class, 
 			  time_t now, unsigned long ttl, unsigned int flags);
 void cache_reload(void);