Fix remote buffer overflow CERT VU#434904

The problem is in the sort_rrset() function and allows a remote
attacker to overwrite memory. Any dnsmasq instance with DNSSEC
enabled is vulnerable.

CHANGELOG

@@ -1,12 +1,39 @@
+version 2.83
+	Fix a remote buffer overflow problem in the DNSSEC code. Any
+	dnsmasq with DNSSEC compiled in and enabled is vulnerable to this,
+	referenced by CERT VU#434904.
+	
+
+version 2.82
+	Improve behaviour in the face of network interfaces which come
+	and go and change index. Thanks to Petr Mensik for the patch.
+
+	Convert hard startup failure on NETLINK_NO_ENOBUFS under qemu-user
+	to a warning.
+
+	Allow IPv6 addresses ofthe form [::ffff:1.2.3.4] in --dhcp-option.
+
+	Fix crash under heavy TCP connection load introduced in 2.81.
+	Thanks to Frank for good work chasing this down.
+
+	Change default lease time for DHCPv6 to one day.
+
+	Alter calculation of preferred and valid times in router
+	advertisements, so that these do not have a floor applied
+	of the lease time in the dhcp-range if this is not explicitly
+	specified and is merely the default.
+	Thanks to Martin-Éric Racine for suggestions on this.
+
+	
 version 2.81
-	Impove cache behaviour for TCP connections. For ease of
-	implementaion, dnsmasq has always forked a new process to handle
+	Improve cache behaviour for TCP connections. For ease of
+	implementation, dnsmasq has always forked a new process to handle
 	each incoming TCP connection. A side-effect of this is that
 	any DNS queries answered from TCP connections are not cached:
 	when TCP connections were rare, this was not a problem.
-	With the coming of DNSSEC, it's now the case that some
+	With the coming of DNSSEC, it is now the case that some
 	DNSSEC queries have answers which spill to TCP, and if,
-	for instance, this applies to the keys for the root then
+	for instance, this applies to the keys for the root, then
 	those never get cached, and performance is very bad.
 	This fix passes cache entries back from the TCP child process to
 	the main server process, and fixes the problem.
@@ -35,7 +62,7 @@ version 2.81
 	configuration.
 
 	Add --shared-network config. This enables allocation of addresses
-	the DHCP server in subnets where the server (or relay) doesn't
+	by the DHCP server in subnets where the server (or relay) does not
 	have an interface on the network in that subnet. Many thanks to
 	kamp.de for sponsoring this feature.
 	
@@ -66,7 +93,7 @@ version 2.81
 	Evgenii Seliavka for the suggestion and initial patch.
 
 	In the router advert code, handle case where we have two
-	different interfaces on the same IPv6 net, and we're doing
+	different interfaces on the same IPv6 net, and we are doing
 	RA/DHCP service on only one of them. Thanks to NIIBE Yutaka
 	for spotting this case and making the initial patch.
 

debian/changelog

@@ -1,3 +1,27 @@
+dnsmasq (2.82-1) unstable; urgency=low
+
+   * New upstream.
+
+ -- Simon Kelley <simon@thekelleys.org.uk>  Fri, 26 Jun 2020 22:22:41 +0000
+
+dnsmasq (2.81-4) unstable; urgency=low
+
+   * Remove runit support when building for Ubuntu. (closes: #960401)
+
+ -- Simon Kelley <simon@thekelleys.org.uk>  Fri, 26 Jun 2020 21:52:44 +0000
+
+dnsmasq (2.81-3) unstable; urgency=low
+
+   * Fixes to control file for bug 958100
+
+ -- Simon Kelley <simon@thekelleys.org.uk>  Sun, 19 Apr 2020 21:44:12 +0000
+
+dnsmasq (2.81-2) unstable; urgency=low
+
+   * Fix FTBFS on kFreeBSD. (closes: #958100)
+	
+ -- Simon Kelley <simon@thekelleys.org.uk>  Sat, 18 Apr 2020 18:34:15 +0000
+
 dnsmasq (2.81-1) unstable; urgency=low
 
    * New upstream.
@@ -12,8 +36,9 @@ dnsmasq (2.81-1) unstable; urgency=low
    * Fix breakage of dig +trace. (closes: #942363)
    * Fix build faliure with newer Nettle libraries. (closes: #940985)
    * Support runscript init-system (closes: #929884)
+   * Security fix for CVE-2019-14834 (closes: #948373)
 	
- -- Simon Kelley <simon@thekelleys.org.uk>  Mon, 27 Jan 2020 22:31:15 +0000
+ -- Simon Kelley <simon@thekelleys.org.uk>  Wed, 8 Apr 2020 17:33:15 +0000
 
 dnsmasq (2.80-1) unstable; urgency=low
 

debian/control

@@ -3,8 +3,9 @@ Section: net
 Priority: optional
 Build-depends: gettext, libnetfilter-conntrack-dev [linux-any],
                libidn2-dev, libdbus-1-dev (>=0.61), libgmp-dev, 
-               nettle-dev (>=2.4-3), libbsd-dev [!linux-any],
-	       liblua5.2-dev, dh-runit, debhelper-compat (= 10)
+               nettle-dev (>=2.4-3), libbsd-dev [kfreebsd-any],
+	       liblua5.2-dev, dh-runit, debhelper-compat (= 10),
+               pkg-config
 Maintainer: Simon Kelley <simon@thekelleys.org.uk>
 Homepage: http://www.thekelleys.org.uk/dnsmasq/doc.html
 Vcs-Git: http://thekelleys.org.uk/git/dnsmasq.git

debian/rules

@@ -98,9 +98,10 @@ ifeq (,$(filter nodnssec,$(DEB_BUILD_OPTIONS)))
      DEB_COPTS += -DHAVE_DNSSEC
 endif
 
-ifneq ($(DEB_HOST_ARCH_OS),linux)
+ifeq ($(DEB_HOST_ARCH_OS),kfreebsd)
      # For strlcpy in FreeBSD
-     LDFLAGS += -lbsd
+     LIBS += $(shell ${PKG_CONFIG} --libs libbsd-overlay)
+     CFLAGS += $(shell ${PKG_CONFIG} --cflags libbsd-overlay)
 endif
 
 define build_tree
@@ -169,16 +170,20 @@ binary-indep:	checkroot
 	        -d debian/trees/daemon/etc/resolvconf/update.d \
 		-d debian/trees/daemon/usr/lib/resolvconf/dpkg-event.d \
 		-d debian/trees/daemon/usr/share/dnsmasq \
+		-d debian/trees/daemon/usr/share/doc/dnsmasq \
 	        -d debian/trees/daemon/etc/default \
 		-d debian/trees/daemon/lib/systemd/system \
 		-d debian/trees/daemon/usr/lib/tmpfiles.d \
                 -d debian/trees/daemon/etc/insserv.conf.d
 	install -m 644 debian/conffiles debian/trees/daemon/DEBIAN
 	install -m 755 debian/postinst debian/postrm debian/prerm debian/trees/daemon/DEBIAN
-	rm -f debian/dnsmasq.postinst.debhelper debian/dnsmasq.postrm.debhelper
-	dh_runit -pdnsmasq -Pdebian/trees/daemon
-	cat debian/dnsmasq.postinst.debhelper >> debian/trees/daemon/DEBIAN/postinst
-	cat debian/dnsmasq.postrm.debhelper   >> debian/trees/daemon/DEBIAN/postrm
+	if ! dpkg-vendor --derives-from Ubuntu; then \
+                rm -f debian/dnsmasq.postinst.debhelper debian/dnsmasq.postrm.debhelper; \
+	      	dh_runit -pdnsmasq -Pdebian/trees/daemon; \
+		cat debian/dnsmasq.postinst.debhelper >> debian/trees/daemon/DEBIAN/postinst; \
+		cat debian/dnsmasq.postrm.debhelper   >> debian/trees/daemon/DEBIAN/postrm; \
+		cd debian/trees/daemon && find etc/sv -type f -printf '/%p\n' >>DEBIAN/conffiles; \
+	fi
 	install -m 755 debian/init debian/trees/daemon/etc/init.d/dnsmasq
 	install -m 755 debian/resolvconf debian/trees/daemon/etc/resolvconf/update.d/dnsmasq
 	install -m 755 debian/resolvconf-package debian/trees/daemon/usr/lib/resolvconf/dpkg-event.d/dnsmasq
@@ -190,7 +195,9 @@ binary-indep:	checkroot
 	install -m 644 debian/systemd@.service debian/trees/daemon/lib/systemd/system/dnsmasq@.service
 	install -m 644 debian/tmpfiles.conf debian/trees/daemon/usr/lib/tmpfiles.d/dnsmasq.conf
 	install -m 644 debian/insserv debian/trees/daemon/etc/insserv.conf.d/dnsmasq
-	ln -s $(package) debian/trees/daemon/usr/share/doc/dnsmasq
+	install -m 644 debian/copyright debian/trees/daemon/usr/share/doc/dnsmasq/copyright
+	install -m 644 debian/changelog debian/trees/daemon/usr/share/doc/dnsmasq/changelog.Debian
+	gzip -9n debian/trees/daemon/usr/share/doc/dnsmasq/changelog.Debian
 	cd debian/trees/daemon && find . -type f ! -regex '.*DEBIAN/.*' -printf '%P\0' | LC_ALL=C sort -z | xargs -r0 md5sum > DEBIAN/md5sums	
 	dpkg-gencontrol $(PACKAGE_VERSION) -Tdebian/dnsmasq.substvars -pdnsmasq -Pdebian/trees/daemon
 	find debian/trees/daemon -depth -newermt '$(BUILD_DATE)' -print0 | xargs -0r touch --no-dereference --date='$(BUILD_DATE)'
@@ -200,7 +207,7 @@ binary-indep:	checkroot
 
 binary-arch:	checkroot 
 	$(call build_tree,debian/trees/base)
-	make $(TARGET) BUILDDIR=debian/build/no-lua PREFIX=/usr DESTDIR=`pwd`/debian/trees/base CFLAGS="$(CFLAGS)" LDFLAGS="$(LDFLAGS)" COPTS="$(DEB_COPTS)" CC=$(CC) PKG_CONFIG=$(PKG_CONFIG)
+	make $(TARGET) BUILDDIR=debian/build/no-lua PREFIX=/usr DESTDIR=`pwd`/debian/trees/base CFLAGS="$(CFLAGS)" LDFLAGS="$(LDFLAGS)" COPTS="$(DEB_COPTS)" CC=$(CC) PKG_CONFIG=$(PKG_CONFIG) LIBS="$(LIBS)"
 ifeq (,$(findstring nodoc,$(DEB_BUILD_OPTIONS)))
 	$(call add_docs,debian/trees/base)
 else	
@@ -219,7 +226,7 @@ endif
 	dpkg --build debian/trees/base ..
 
 	$(call build_tree,debian/trees/lua-base)
-	make $(TARGET) BUILDDIR=debian/build/lua PREFIX=/usr DESTDIR=`pwd`/debian/trees/lua-base CFLAGS="$(CFLAGS)" LDFLAGS="$(LDFLAGS)" COPTS="-DHAVE_LUASCRIPT $(DEB_COPTS)" CC=$(CC) PKG_CONFIG=$(PKG_CONFIG)
+	make $(TARGET) BUILDDIR=debian/build/lua PREFIX=/usr DESTDIR=`pwd`/debian/trees/lua-base CFLAGS="$(CFLAGS)" LDFLAGS="$(LDFLAGS)" COPTS="-DHAVE_LUASCRIPT $(DEB_COPTS)" CC=$(CC) PKG_CONFIG=$(PKG_CONFIG) LIBS="$(LIBS)"
 ifeq (,$(findstring nodoc,$(DEB_BUILD_OPTIONS)))
 	$(call add_docs,debian/trees/lua-base)
 else	
@@ -249,7 +256,7 @@ ifeq ($(DEB_HOST_ARCH_OS),linux)
 ifeq (,$(findstring nodoc,$(DEB_BUILD_OPTIONS)))
 	install -m 755 -d debian/trees/utils/usr/share/man/man1
 endif
-	make -C contrib/lease-tools PREFIX=/usr DESTDIR=`pwd`/debian/trees/utils CFLAGS="$(CFLAGS)" LDFLAGS="$(LDFLAGS)" COPTS="$(DEB_COPTS)" CC=$(CC) PKG_CONFIG=$(PKG_CONFIG)
+	make -C contrib/lease-tools PREFIX=/usr DESTDIR=`pwd`/debian/trees/utils CFLAGS="$(CFLAGS)" LDFLAGS="$(LDFLAGS)" COPTS="$(DEB_COPTS)" CC=$(CC) PKG_CONFIG=$(PKG_CONFIG) LIBS="$(LIBS)"
 	install -m 755 contrib/lease-tools/dhcp_release debian/trees/utils/usr/bin/dhcp_release
 	install -m 755 contrib/lease-tools/dhcp_release6 debian/trees/utils/usr/bin/dhcp_release6
 	install -m 755 contrib/lease-tools/dhcp_lease_time debian/trees/utils/usr/bin/dhcp_lease_time

man/dnsmasq.8

@@ -1,4 +1,4 @@
-.TH DNSMASQ 8
+.TH DNSMASQ 8 2020-04-05
 .SH NAME
 dnsmasq \- A lightweight DHCP and caching DNS server.
 .SH SYNOPSIS
@@ -428,7 +428,7 @@ Tells dnsmasq to never forward A or AAAA queries for plain names, without dots
 or domain parts, to upstream nameservers. If the name is not known
 from /etc/hosts or DHCP then a "not found" answer is returned.
 .TP
-.B \-S, --local, --server=[/[<domain>]/[domain/]][<ipaddr>[#<port>][@<source-ip>|<interface>[#<port>]]
+.B \-S, --local, --server=[/[<domain>]/[domain/]][<ipaddr>[#<port>]][@<source-ip>|<interface>[#<port>]]
 Specify IP address of upstream servers directly. Setting this flag does
 not suppress reading of /etc/resolv.conf, use \fB--no-resolv\fP to do that. If one or more
 optional domains are given, that server is used only for those domains
@@ -861,7 +861,7 @@ in
 options. If the lease time is given, then leases
 will be given for that length of time. The lease time is in seconds,
 or minutes (eg 45m) or hours (eg 1h) or "infinite". If not given,
-the default lease time is one hour. The
+the default lease time is one hour for IPv4 and one day for IPv6. The
 minimum lease time is two minutes. For IPv6 ranges, the lease time
 maybe "deprecated"; this sets the preferred lifetime sent in a DHCP
 lease or router advertisement to zero, which causes clients to use
@@ -909,7 +909,7 @@ then the address can be simply ::
 The optional 
 .B set:<tag> 
 sets an alphanumeric label which marks this network so that
-dhcp options may be specified on a per-network basis. 
+DHCP options may be specified on a per-network basis.
 When it is prefixed with 'tag:' instead, then its meaning changes from setting
 a tag to matching it. Only one tag may be set, but more than one tag
 may be matched.
@@ -1017,7 +1017,7 @@ may contain an IPv4 address or one or more IPv6 addresses, or both. IPv6 address
 .B --dhcp-host=laptop,[1234::56]
 IPv6 addresses may contain only the host-identifier part:
 .B --dhcp-host=laptop,[::56]
-in which case they act as wildcards in constructed dhcp ranges, with
+in which case they act as wildcards in constructed DHCP ranges, with
 the appropriate network part inserted. For IPv6, an address may include a prefix length:
 .B --dhcp-host=laptop,[1234:50/126]
 which (in this case) specifies four addresses, 1234::50 to 1234::53. This (an the ability
@@ -1354,7 +1354,7 @@ vendor-identifying vendor classes for the specified enterprise. Please
 see RFC 3925 for more details of these rare and interesting beasts.
 .TP
 .B --dhcp-name-match=set:<tag>,<name>[*]
-Set the tag if the given name is supplied by a dhcp client. There may be a single trailing wildcard *, which has the usual meaning. Combined with dhcp-ignore or dhcp-ignore-names this gives the ability to ignore certain clients by name, or disallow certain hostnames from being claimed by a client.
+Set the tag if the given name is supplied by a DHCP client. There may be a single trailing wildcard *, which has the usual meaning. Combined with dhcp-ignore or dhcp-ignore-names this gives the ability to ignore certain clients by name, or disallow certain hostnames from being claimed by a client.
 .TP
 .B --tag-if=set:<tag>[,set:<tag>[,tag:<tag>[,tag:<tag>]]]
 Perform boolean operations on tags. Any tag appearing as set:<tag> is set if
@@ -1751,7 +1751,7 @@ option also forces the leasechange script to be called on changes
 to the client-id and lease length and expiry time.
 .TP
 .B --script-on-renewal
-Call the dhcp script when the lease expiry time changes, for instance when the
+Call the DHCP script when the lease expiry time changes, for instance when the
 lease is renewed.
 .TP
 .B --bridge-interface=<interface>,<alias>[,<alias>]
@@ -1767,25 +1767,34 @@ It is permissible to add more than one alias using more than one \fB--bridge-int
 \fB--bridge-interface=int1,alias1,alias2\fP is exactly equivalent to
 \fB--bridge-interface=int1,alias1 --bridge-interface=int1,alias2\fP
 .TP
-.B --shared-network=<interface>|<addr>,<addr>
-The DHCP server determines which dhcp ranges are useable for allocating and
+.B --shared-network=<interface>,<addr>
+.PD 0
+.TP
+.B --shared-network=<addr>,<addr>
+.PD 1v
+The DHCP server determines which DHCP ranges are useable for allocating an
 address to a DHCP client based on the network from which the DHCP request arrives,
 and the IP configuration of the server's interface on that network. The shared-network
-option extends the available subnets (and therefore dhcp ranges) beyond the
-subnets configured on the arrival interface. The first argument is either the
-name of an interface or an address which is configured on a local interface, and the
+option extends the available subnets (and therefore DHCP ranges) beyond the
+subnets configured on the arrival interface.
+
+The first argument is either the
+name of an interface, or an address that is configured on a local interface, and the
 second argument is an address which defines another subnet on which addresses can be allocated.
-To be useful, there must be suitable dhcp-range which allows address allocation on this subnet
-and this dhcp-range MUST include the netmask. Use shared-network also needs extra
-consideration of routing. Dnsmasq doesn't have the usual information which it uses to
+
+To be useful, there must be a suitable dhcp-range which allows address allocation on this subnet
+and this dhcp-range MUST include the netmask.
+
+Using shared-network also needs extra
+consideration of routing. Dnsmasq does not have the usual information that it uses to
 determine the default route, so the default route option (or other routing) MUST be
-manually configured. The client must have a route to the server: if the two-address form
-of shared-network is used, this will be to the first specified address. If the interface,address
+configured manually. The client must have a route to the server: if the two-address form
+of shared-network is used, this needs to be to the first specified address. If the interface,address
 form is used, there must be a route to all of the addresses configured on the interface.
 
 The two-address form of shared-network is also usable with a DHCP relay: the first address
 is the address of the relay and the second, as before, specifies an extra subnet which
-may be allocated.
+addresses may be allocated from.
 
 .TP
 .B \-s, --domain=<domain>[,<address range>[,local]]
@@ -1795,7 +1804,7 @@ firstly it causes the DHCP server to return the domain to any hosts
 which request it, and secondly it sets the domain which it is legal
 for DHCP-configured hosts to claim. The intention is to constrain
 hostnames so that an untrusted host on the LAN cannot advertise 
-its name via dhcp as e.g. "microsoft.com" and capture traffic not 
+its name via DHCP as e.g. "microsoft.com" and capture traffic not 
 meant for it. If no domain suffix is specified, then any DHCP
 hostname with a domain part (ie with a period) will be disallowed 
 and logged. If suffix is specified, then hostnames with a domain 
@@ -1886,7 +1895,7 @@ The mtu: parameter may be an arbitrary interface name, in which case the MTU val
 for (eg) advertising the MTU of a WAN interface on the other interfaces of a router.
 .TP
 .B --dhcp-reply-delay=[tag:<tag>,]<integer>
-Delays sending DHCPOFFER and proxydhcp replies for at least the specified number of seconds.
+Delays sending DHCPOFFER and PROXYDHCP replies for at least the specified number of seconds.
 This can be used as workaround for bugs in PXE boot firmware that does not function properly when
 receiving an instant reply.
 This option takes into account the time already spent waiting (e.g. performing ping check) if any.
@@ -2022,7 +2031,7 @@ and
 .I /etc/ethers 
 and any file given by \fB--dhcp-hostsfile\fP, \fB--dhcp-hostsdir\fP, \fB--dhcp-optsfile\fP,
 \fB--dhcp-optsdir\fP, \fB--addn-hosts\fP or \fB--hostsdir\fP.
-The dhcp lease change script is called for all
+The DHCP lease change script is called for all
 existing DHCP leases. If 
 .B
 --no-poll
@@ -2200,7 +2209,6 @@ as is the tag "bootp", allowing some control over the options returned to
 different classes of hosts.
 
 .SH AUTHORITATIVE CONFIGURATION
-.PP 
 Configuring dnsmasq to act as an authoritative DNS server is
 complicated by the fact that it involves configuration of external DNS
 servers to provide delegation. We will walk through three scenarios of
@@ -2370,7 +2378,6 @@ used, and must match the zone's domain.
 
 
 .SH EXIT CODES
-.PP
 0 - Dnsmasq successfully forked into the background, or terminated
 normally if backgrounding is not enabled.
 .PP

src/cache.c

@@ -472,16 +472,29 @@ void cache_start_insert(void)
 struct crec *cache_insert(char *name, union all_addr *addr, unsigned short class,
 			  time_t now,  unsigned long ttl, unsigned int flags)
 {
-  /* Don't log DNSSEC records here, done elsewhere */
-  if (flags & (F_IPV4 | F_IPV6 | F_CNAME | F_SRV))
+#ifdef HAVE_DNSSEC
+  if (flags & (F_DNSKEY | F_DS)) 
     {
+      /* The DNSSEC validation process works by getting needed records into the
+	 cache, then retrying the validation until they are all in place.
+	 This can be messed up by very short TTLs, and _really_ messed up by
+	 zero TTLs, so we force the TTL to be at least long enough to do a validation.
+	 Ideally, we should use some kind of reference counting so that records are
+	 locked until the validation that asked for them is complete, but this
+	 is much easier, and just as effective. */
+      if (ttl < DNSSEC_MIN_TTL)
+	ttl = DNSSEC_MIN_TTL;
+    }
+  else
+#endif
+    {
+      /* Don't log DNSSEC records here, done elsewhere */
       log_query(flags | F_UPSTREAM, name, addr, NULL);
-      /* Don't mess with TTL for DNSSEC records. */
       if (daemon->max_cache_ttl != 0 && daemon->max_cache_ttl < ttl)
 	ttl = daemon->max_cache_ttl;
       if (daemon->min_cache_ttl != 0 && daemon->min_cache_ttl > ttl)
 	ttl = daemon->min_cache_ttl;
-    }
+    }	
   
   return really_insert(name, addr, class, now, ttl, flags);
 }

src/config.h

@@ -40,9 +40,11 @@
 #define DHCP_PACKET_MAX 16384 /* hard limit on DHCP packet size */
 #define SMALLDNAME 50 /* most domain names are smaller than this */
 #define CNAME_CHAIN 10 /* chains longer than this atr dropped for loop protection */
+#define DNSSEC_MIN_TTL 60 /* DNSKEY and DS records in cache last at least this long */
 #define HOSTSFILE "/etc/hosts"
 #define ETHERSFILE "/etc/ethers"
-#define DEFLEASE 3600 /* default lease time, 1 hour */
+#define DEFLEASE 3600 /* default DHCPv4 lease time, one hour */
+#define DEFLEASE6 (3600*24) /* default lease time for DHCPv6. One day. */
 #define CHUSER "nobody"
 #define CHGRP "dip"
 #define TFTP_MAX_CONNECTIONS 50 /* max simultaneous connections */

src/dbus.c

@@ -237,7 +237,7 @@ static DBusMessage *dbus_reply_server_loop(DBusMessage *message)
   for (serv = daemon->servers; serv; serv = serv->next)
     if (serv->flags & SERV_LOOP)
       {
-	prettyprint_addr(&serv->addr, daemon->addrbuff);
+	(void)prettyprint_addr(&serv->addr, daemon->addrbuff);
 	dbus_message_iter_append_basic (&args_iter, DBUS_TYPE_STRING, &daemon->addrbuff);
       }
   

src/dhcp-common.c

@@ -38,7 +38,7 @@ void dhcp_common_init(void)
 
 ssize_t recv_dhcp_packet(int fd, struct msghdr *msg)
 {  
-  ssize_t sz;
+  ssize_t sz, new_sz;
  
   while (1)
     {
@@ -65,9 +65,18 @@ ssize_t recv_dhcp_packet(int fd, struct msghdr *msg)
 	}
     }
   
-  while ((sz = recvmsg(fd, msg, 0)) == -1 && errno == EINTR);
+  while ((new_sz = recvmsg(fd, msg, 0)) == -1 && errno == EINTR);
+
+  /* Some kernels seem to ignore MSG_PEEK, and dequeue the packet anyway. 
+     If that happens we get EAGAIN here because the socket is non-blocking.
+     Use the result of the original testing recvmsg as long as the buffer
+     was big enough. There's a small race here that may lose the odd packet,
+     but it's UDP anyway. */
   
-  return (msg->msg_flags & MSG_TRUNC) ? -1 : sz;
+  if (new_sz == -1 && (errno == EWOULDBLOCK || errno == EAGAIN))
+    new_sz = sz;
+  
+  return (msg->msg_flags & MSG_TRUNC) ? -1 : new_sz;
 }
 
 struct dhcp_netid *run_tag_if(struct dhcp_netid *tags)

src/dnsmasq.c

@@ -58,6 +58,7 @@ int main (int argc, char **argv)
   char *bound_device = NULL;
   int did_bind = 0;
   struct server *serv;
+  char *netlink_warn;
 #endif 
 #if defined(HAVE_DHCP) || defined(HAVE_DHCP6)
   struct dhcp_context *context;
@@ -327,7 +328,7 @@ int main (int argc, char **argv)
 #endif
 
 #if  defined(HAVE_LINUX_NETWORK)
-  netlink_init();
+  netlink_warn = netlink_init();
 #elif defined(HAVE_BSD_NETWORK)
   route_init();
 #endif
@@ -946,6 +947,9 @@ int main (int argc, char **argv)
 #  ifdef HAVE_LINUX_NETWORK
   if (did_bind)
     my_syslog(MS_DHCP | LOG_INFO, _("DHCP, sockets bound exclusively to interface %s"), bound_device);
+
+  if (netlink_warn)
+    my_syslog(LOG_WARNING, netlink_warn);
 #  endif
 
   /* after dhcp_construct_contexts */
@@ -1820,7 +1824,8 @@ static void check_dns_listeners(time_t now)
 		    addr.addr4 = tcp_addr.in.sin_addr;
 		  
 		  for (iface = daemon->interfaces; iface; iface = iface->next)
-		    if (iface->index == if_index)
+		    if (iface->index == if_index &&
+		        iface->addr.sa.sa_family == tcp_addr.sa.sa_family)
 		      break;
 		  
 		  if (!iface && !loopback_exception(listener->tcpfd, tcp_addr.sa.sa_family, &addr, intr_name))
@@ -1859,31 +1864,30 @@ static void check_dns_listeners(time_t now)
 	      else
 		{
 		  int i;
+#ifdef HAVE_LINUX_NETWORK
+		  /* The child process inherits the netlink socket, 
+		     which it never uses, but when the parent (us) 
+		     uses it in the future, the answer may go to the 
+		     child, resulting in the parent blocking
+		     forever awaiting the result. To avoid this
+		     the child closes the netlink socket, but there's
+		     a nasty race, since the parent may use netlink
+		     before the child has done the close.
+		     
+		     To avoid this, the parent blocks here until a 
+		     single byte comes back up the pipe, which
+		     is sent by the child after it has closed the
+		     netlink socket. */
+		  
+		  unsigned char a;
+		  read_write(pipefd[0], &a, 1, 1);
+#endif
 
 		  for (i = 0; i < MAX_PROCS; i++)
 		    if (daemon->tcp_pids[i] == 0 && daemon->tcp_pipes[i] == -1)
 		      {
-			char a;
-			(void)a; /* suppress potential unused warning */
-
 			daemon->tcp_pids[i] = p;
 			daemon->tcp_pipes[i] = pipefd[0];
-#ifdef HAVE_LINUX_NETWORK
-			/* The child process inherits the netlink socket, 
-			   which it never uses, but when the parent (us) 
-			   uses it in the future, the answer may go to the 
-			   child, resulting in the parent blocking
-			   forever awaiting the result. To avoid this
-			   the child closes the netlink socket, but there's
-			   a nasty race, since the parent may use netlink
-			   before the child has done the close.
-
-			   To avoid this, the parent blocks here until a 
-			   single byte comes back up the pipe, which
-			   is sent by the child after it has closed the
-			   netlink socket. */
-			retry_send(read(pipefd[0], &a, 1));
-#endif
 			break;
 		      }
 		}
@@ -1915,16 +1919,16 @@ static void check_dns_listeners(time_t now)
 		 terminate the process. */
 	      if (!option_bool(OPT_DEBUG))
 		{
-		  char a = 0;
-		  (void)a; /* suppress potential unused warning */
+#ifdef HAVE_LINUX_NETWORK
+		  /* See comment above re: netlink socket. */
+		  unsigned char a = 0;
+
+		  close(daemon->netlinkfd);
+		  read_write(pipefd[1], &a, 1, 0);
+#endif		  
 		  alarm(CHILD_LIFETIME);
 		  close(pipefd[0]); /* close read end in child. */
 		  daemon->pipe_to_parent = pipefd[1];
-#ifdef HAVE_LINUX_NETWORK
-		  /* See comment above re netlink socket. */
-		  close(daemon->netlinkfd);
-		  retry_send(write(pipefd[1], &a, 1));
-#endif
 		}
 
 	      /* start with no upstream connections. */
@@ -1951,8 +1955,10 @@ static void check_dns_listeners(time_t now)
 		    shutdown(s->tcpfd, SHUT_RDWR);
 		    close(s->tcpfd);
 		  }
+	      
 	      if (!option_bool(OPT_DEBUG))
 		{
+		  close(daemon->pipe_to_parent);
 		  flush_log();
 		  _exit(0);
 		}
@@ -2112,6 +2118,4 @@ int delay_dhcp(time_t start, int sec, int fd, uint32_t addr, unsigned short id)
 
   return 0;
 }
-#endif
-
- 
+#endif /* HAVE_DHCP */

src/dnsmasq.h

@@ -572,7 +572,8 @@ struct irec {
 };
 
 struct listener {
-  int fd, tcpfd, tftpfd, family;
+  int fd, tcpfd, tftpfd, used;
+  union mysockaddr addr;
   struct irec *iface; /* only sometimes valid for non-wildcard */
   struct listener *next;
 };
@@ -941,6 +942,7 @@ struct shared_network {
 #define CONTEXT_OLD            (1u<<16)
 #define CONTEXT_V6             (1u<<17)
 #define CONTEXT_RA_OFF_LINK    (1u<<18)
+#define CONTEXT_SETLEASE       (1u<<19)
 
 struct ping_result {
   struct in_addr addr;
@@ -1450,7 +1452,7 @@ void clear_cache_and_reload(time_t now);
 
 /* netlink.c */
 #ifdef HAVE_LINUX_NETWORK
-void netlink_init(void);
+char *netlink_init(void);
 void netlink_multicast(void);
 #endif
 

src/dnssec.c

@@ -223,138 +223,147 @@ static int check_date_range(unsigned long curtime, u32 date_start, u32 date_end)
     && serial_compare_32(curtime, date_end) == SERIAL_LT;
 }
 
-/* Return bytes of canonicalised rdata, when the return value is zero, the remaining 
-   data, pointed to by *p, should be used raw. */
-static int get_rdata(struct dns_header *header, size_t plen, unsigned char *end, char *buff, int bufflen,
-		     unsigned char **p, u16 **desc)
+/* Return bytes of canonicalised rrdata one by one.
+   Init state->ip with the RR, and state->end with the end of same.
+   Init state->op to NULL.
+   Init state->desc to RR descriptor.
+   Init state->buff with a MAXDNAME * 2 buffer.
+   
+   After each call which returns 1, state->op points to the next byte of data.
+   On returning 0, the end has been reached.
+*/
+struct rdata_state {
+  u16 *desc;
+  size_t c;
+  unsigned char *end, *ip, *op;
+  char *buff;
+};
+
+static int get_rdata(struct dns_header *header, size_t plen, struct rdata_state *state)
 {
-  int d = **desc;
+  int d;
   
-  /* No more data needs mangling */
-  if (d == (u16)-1)
+  if (state->op && state->c != 1)
     {
-      /* If there's more data than we have space for, just return what fits,
-	 we'll get called again for more chunks */
-      if (end - *p > bufflen)
-	{
-	  memcpy(buff, *p, bufflen);
-	  *p += bufflen;
-	  return bufflen;
-	}
-      
-      return 0;
+      state->op++;
+      state->c--;
+      return 1;
     }
- 
-  (*desc)++;
-  
-  if (d == 0 && extract_name(header, plen, p, buff, 1, 0))
-    /* domain-name, canonicalise */
-    return to_wire(buff);
-  else
-    { 
-      /* plain data preceding a domain-name, don't run off the end of the data */
-      if ((end - *p) < d)
-	d = end - *p;
+
+  while (1)
+    {
+      d = *(state->desc);
       
-      if (d != 0)
+      if (d == (u16)-1)
 	{
-	  memcpy(buff, *p, d);
-	  *p += d;
+	  /* all the bytes to the end. */
+	  if ((state->c = state->end - state->ip) != 0)
+	    {
+	      state->op = state->ip;
+	      state->ip = state->end;;
+	    }
+	  else
+	    return 0;
+	}
+      else
+	{
+	  state->desc++;
+	  
+	  if (d == (u16)0)
+	    {
+	      /* domain-name, canonicalise */
+	      int len;
+	      
+	      if (!extract_name(header, plen, &state->ip, state->buff, 1, 0) ||
+		  (len = to_wire(state->buff)) == 0)
+		continue;
+	      
+	      state->c = len;
+	      state->op = (unsigned char *)state->buff;
+	    }
+	  else
+	    {
+	      /* plain data preceding a domain-name, don't run off the end of the data */
+	      if ((state->end - state->ip) < d)
+		d = state->end - state->ip;
+	      
+	      if (d == 0)
+		continue;
+		  
+	      state->op = state->ip;
+	      state->c = d;
+	      state->ip += d;
+	    }
 	}
       
-      return d;
+      return 1;
     }
 }
 
-/* Bubble sort the RRset into the canonical order. 
-   Note that the byte-streams from two RRs may get unsynced: consider 
-   RRs which have two domain-names at the start and then other data.
-   The domain-names may have different lengths in each RR, but sort equal
-
-   ------------
-   |abcde|fghi|
-   ------------
-   |abcd|efghi|
-   ------------
-
-   leaving the following bytes as deciding the order. Hence the nasty left1 and left2 variables.
-*/
+/* Bubble sort the RRset into the canonical order. */
 
 static int sort_rrset(struct dns_header *header, size_t plen, u16 *rr_desc, int rrsetidx, 
 		      unsigned char **rrset, char *buff1, char *buff2)
 {
-  int swap, quit, i, j;
+  int swap, i, j;
   
   do
     {
       for (swap = 0, i = 0; i < rrsetidx-1; i++)
 	{
-	  int rdlen1, rdlen2, left1, left2, len1, len2, len, rc;
-	  u16 *dp1, *dp2;
-	  unsigned char *end1, *end2;
+	  int rdlen1, rdlen2;
+	  struct rdata_state state1, state2;
+	  
 	  /* Note that these have been determined to be OK previously,
 	     so we don't need to check for NULL return here. */
-	  unsigned char *p1 = skip_name(rrset[i], header, plen, 10);
-	  unsigned char *p2 = skip_name(rrset[i+1], header, plen, 10);
+	  state1.ip = skip_name(rrset[i], header, plen, 10);
+	  state2.ip = skip_name(rrset[i+1], header, plen, 10);
+	  state1.op = state2.op = NULL;
+	  state1.buff = buff1;
+	  state2.buff = buff2;
+	  state1.desc = state2.desc = rr_desc;
 	  
-	  p1 += 8; /* skip class, type, ttl */
-	  GETSHORT(rdlen1, p1);
-	  end1 = p1 + rdlen1;
+	  state1.ip += 8; /* skip class, type, ttl */
+	  GETSHORT(rdlen1, state1.ip);
+	  if (!CHECK_LEN(header, state1.ip, plen, rdlen1))
+	    return rrsetidx; /* short packet */
+	  state1.end = state1.ip + rdlen1;
 	  
-	  p2 += 8; /* skip class, type, ttl */
-	  GETSHORT(rdlen2, p2);
-	  end2 = p2 + rdlen2; 
-	  
-	  dp1 = dp2 = rr_desc;
-	  
-	  for (quit = 0, left1 = 0, left2 = 0, len1 = 0, len2 = 0; !quit;)
+	  state2.ip += 8; /* skip class, type, ttl */
+	  GETSHORT(rdlen2, state2.ip);
+	  if (!CHECK_LEN(header, state2.ip, plen, rdlen2))
+	    return rrsetidx; /* short packet */
+	  state2.end = state2.ip + rdlen2; 
+	  	  
+	  while (1)
 	    {
-	      if (left1 != 0)
-		memmove(buff1, buff1 + len1 - left1, left1);
+	      int ok1, ok2;
 	      
-	      if ((len1 = get_rdata(header, plen, end1, buff1 + left1, (MAXDNAME * 2) - left1, &p1, &dp1)) == 0)
-		{
-		  quit = 1;
-		  len1 = end1 - p1;
-		  memcpy(buff1 + left1, p1, len1);
-		}
-	      len1 += left1;
-	      
-	      if (left2 != 0)
-		memmove(buff2, buff2 + len2 - left2, left2);
-	      
-	      if ((len2 = get_rdata(header, plen, end2, buff2 + left2, (MAXDNAME *2) - left2, &p2, &dp2)) == 0)
-		{
-		  quit = 1;
-		  len2 = end2 - p2;
-		  memcpy(buff2 + left2, p2, len2);
-		}
-	      len2 += left2;
-	       
-	      if (len1 > len2)
-		left1 = len1 - len2, left2 = 0, len = len2;
-	      else
-		left2 = len2 - len1, left1 = 0, len = len1;
-	      
-	      rc = (len == 0) ? 0 : memcmp(buff1, buff2, len);
-	      
-	      if (rc > 0 || (rc == 0 && quit && len1 > len2))
-		{
-		  unsigned char *tmp = rrset[i+1];
-		  rrset[i+1] = rrset[i];
-		  rrset[i] = tmp;
-		  swap = quit = 1;
-		}
-	      else if (rc == 0 && quit && len1 == len2)
+	      ok1 = get_rdata(header, plen, &state1);
+	      ok2 = get_rdata(header, plen, &state2);
+
+	      if (!ok1 && !ok2)
 		{
 		  /* Two RRs are equal, remove one copy. RFC 4034, para 6.3 */
 		  for (j = i+1; j < rrsetidx-1; j++)
 		    rrset[j] = rrset[j+1];
 		  rrsetidx--;
 		  i--;
+		  break;
 		}
-	      else if (rc < 0)
-		quit = 1;
+	      else if (ok1 && (!ok2 || *state1.op > *state2.op)) 
+		{
+		  unsigned char *tmp = rrset[i+1];
+		  rrset[i+1] = rrset[i];
+		  rrset[i] = tmp;
+		  swap = 1;
+		  break;
+		}
+	      else if (ok2 && (!ok1 || *state2.op > *state1.op))
+		break;
+	      
+	      /* arrive here when bytes are equal, go round the loop again
+		 and compare the next ones. */
 	    }
 	}
     } while (swap);
@@ -569,15 +578,18 @@ static int validate_rrset(time_t now, struct dns_header *header, size_t plen, in
       wire_len = to_wire(keyname);
       hash->update(ctx, (unsigned int)wire_len, (unsigned char*)keyname);
       from_wire(keyname);
+
+#define RRBUFLEN 300 /* Most RRs are smaller than this. */
       
       for (i = 0; i < rrsetidx; ++i)
 	{
-	  int seg;
-	  unsigned char *end, *cp;
-	  u16 len, *dp;
+	  int j;
+	  struct rdata_state state;
+	  u16 len;
+	  unsigned char rrbuf[RRBUFLEN];
 	  
 	  p = rrset[i];
-	 	  
+	  
 	  if (!extract_name(header, plen, &p, name, 1, 10)) 
 	    return STAT_BOGUS;
 
@@ -586,12 +598,11 @@ static int validate_rrset(time_t now, struct dns_header *header, size_t plen, in
 	  /* if more labels than in RRsig name, hash *.<no labels in rrsig labels field>  4035 5.3.2 */
 	  if (labels < name_labels)
 	    {
-	      int k;
-	      for (k = name_labels - labels; k != 0; k--)
+	      for (j = name_labels - labels; j != 0; j--)
 		{
 		  while (*name_start != '.' && *name_start != 0)
 		    name_start++;
-		  if (k != 1 && *name_start == '.')
+		  if (j != 1 && *name_start == '.')
 		    name_start++;
 		}
 	      
@@ -612,24 +623,44 @@ static int validate_rrset(time_t now, struct dns_header *header, size_t plen, in
 	  if (!CHECK_LEN(header, p, plen, rdlen))
 	    return STAT_BOGUS; 
 	  
-	  end = p + rdlen;
+	  /* canonicalise rdata and calculate length of same, use 
+	     name buffer as workspace for get_rdata. */
+	  state.ip = p;
+	  state.op = NULL;
+	  state.desc = rr_desc;
+	  state.buff = name;
+	  state.end = p + rdlen;
 	  
-	  /* canonicalise rdata and calculate length of same, use name buffer as workspace.
-	     Note that name buffer is twice MAXDNAME long in DNSSEC mode. */
-	  cp = p;
-	  dp = rr_desc;
-	  for (len = 0; (seg = get_rdata(header, plen, end, name, MAXDNAME * 2, &cp, &dp)) != 0; len += seg);
-	  len += end - cp;
-	  len = htons(len);
+	  for (j = 0; get_rdata(header, plen, &state); j++)
+	    if (j < RRBUFLEN)
+	      rrbuf[j] = *state.op;
+
+	  len = htons((u16)j);
 	  hash->update(ctx, 2, (unsigned char *)&len); 
+
+	  /* If the RR is shorter than RRBUFLEN (most of them, in practice)
+	     then we can just digest it now. If it exceeds RRBUFLEN we have to
+	     go back to the start and do it in chunks. */
+	  if (j >= RRBUFLEN)
+	    {
+	      state.ip = p;
+	      state.op = NULL;
+	      state.desc = rr_desc;
+
+	      for (j = 0; get_rdata(header, plen, &state); j++)
+		{
+		   rrbuf[j] = *state.op;
+
+		   if (j == RRBUFLEN - 1)
+		     {
+		       hash->update(ctx, RRBUFLEN, rrbuf);
+		       j = -1;
+		     }
+		}
+	    }
 	  
-	  /* Now canonicalise again and digest. */
-	  cp = p;
-	  dp = rr_desc;
-	  while ((seg = get_rdata(header, plen, end, name, MAXDNAME * 2, &cp, &dp)))
-	    hash->update(ctx, seg, (unsigned char *)name);
-	  if (cp != end)
-	    hash->update(ctx, end - cp, cp);
+	  if (j != 0)
+	    hash->update(ctx, j, rrbuf);
 	}
      
       hash->digest(ctx, hash->digest_size, digest);

src/forward.c

@@ -676,7 +676,7 @@ static size_t process_reply(struct dns_header *header, time_t now, struct server
   if (!(header->hb4 & HB4_RA) && rcode == NOERROR &&
       server && !(server->flags & SERV_WARNED_RECURSIVE))
     {
-      prettyprint_addr(&server->addr, daemon->namebuff);
+      (void)prettyprint_addr(&server->addr, daemon->namebuff);
       my_syslog(LOG_WARNING, _("nameserver %s refused to do a recursive query"), daemon->namebuff);
       if (!option_bool(OPT_LOG))
 	server->flags |= SERV_WARNED_RECURSIVE;
@@ -960,7 +960,7 @@ void reply_query(int fd, int family, time_t now)
     {
       forward->sentto->edns_pktsz = SAFE_PKTSZ;
       forward->sentto->pktsz_reduced = now;
-      prettyprint_addr(&forward->sentto->addr, daemon->addrbuff);
+      (void)prettyprint_addr(&forward->sentto->addr, daemon->addrbuff);
       my_syslog(LOG_WARNING, _("reducing DNS packet size for nameserver %s to %d"), daemon->addrbuff, SAFE_PKTSZ);
     }
 
@@ -1284,8 +1284,9 @@ void receive_query(struct listener *listen, time_t now)
 		 CMSG_SPACE(sizeof(struct sockaddr_dl))];
 #endif
   } control_u;
+  int family = listen->addr.sa.sa_family;
    /* Can always get recvd interface for IPv6 */
-  int check_dst = !option_bool(OPT_NOWILD) || listen->family == AF_INET6;
+  int check_dst = !option_bool(OPT_NOWILD) || family == AF_INET6;
 
   /* packet buffer overwritten */
   daemon->srv_save = NULL;
@@ -1297,7 +1298,7 @@ void receive_query(struct listener *listen, time_t now)
     {
       auth_dns = listen->iface->dns_auth;
      
-      if (listen->family == AF_INET)
+      if (family == AF_INET)
 	{
 	  dst_addr_4 = dst_addr.addr4 = listen->iface->addr.in.sin_addr;
 	  netmask = listen->iface->netmask;
@@ -1327,9 +1328,9 @@ void receive_query(struct listener *listen, time_t now)
      information disclosure. */
   memset(daemon->packet + n, 0, daemon->edns_pktsz - n);
   
-  source_addr.sa.sa_family = listen->family;
+  source_addr.sa.sa_family = family;
   
-  if (listen->family == AF_INET)
+  if (family == AF_INET)
     {
        /* Source-port == 0 is an error, we can't send back to that. 
 	  http://www.ietf.org/mail-archive/web/dnsop/current/msg11441.html */
@@ -1349,7 +1350,7 @@ void receive_query(struct listener *listen, time_t now)
     {
       struct addrlist *addr;
 
-      if (listen->family == AF_INET6) 
+      if (family == AF_INET6) 
 	{
 	  for (addr = daemon->interface_addrs; addr; addr = addr->next)
 	    if ((addr->flags & ADDRLIST_IPV6) &&
@@ -1387,7 +1388,7 @@ void receive_query(struct listener *listen, time_t now)
 	return;
 
 #if defined(HAVE_LINUX_NETWORK)
-      if (listen->family == AF_INET)
+      if (family == AF_INET)
 	for (cmptr = CMSG_FIRSTHDR(&msg); cmptr; cmptr = CMSG_NXTHDR(&msg, cmptr))
 	  if (cmptr->cmsg_level == IPPROTO_IP && cmptr->cmsg_type == IP_PKTINFO)
 	    {
@@ -1400,7 +1401,7 @@ void receive_query(struct listener *listen, time_t now)
 	      if_index = p.p->ipi_ifindex;
 	    }
 #elif defined(IP_RECVDSTADDR) && defined(IP_RECVIF)
-      if (listen->family == AF_INET)
+      if (family == AF_INET)
 	{
 	  for (cmptr = CMSG_FIRSTHDR(&msg); cmptr; cmptr = CMSG_NXTHDR(&msg, cmptr))
 	    {
@@ -1425,7 +1426,7 @@ void receive_query(struct listener *listen, time_t now)
 	}
 #endif
       
-      if (listen->family == AF_INET6)
+      if (family == AF_INET6)
 	{
 	  for (cmptr = CMSG_FIRSTHDR(&msg); cmptr; cmptr = CMSG_NXTHDR(&msg, cmptr))
 	    if (cmptr->cmsg_level == IPPROTO_IPV6 && cmptr->cmsg_type == daemon->v6pktinfo)
@@ -1446,16 +1447,16 @@ void receive_query(struct listener *listen, time_t now)
       if (!indextoname(listen->fd, if_index, ifr.ifr_name))
 	return;
       
-      if (!iface_check(listen->family, &dst_addr, ifr.ifr_name, &auth_dns))
+      if (!iface_check(family, &dst_addr, ifr.ifr_name, &auth_dns))
 	{
 	   if (!option_bool(OPT_CLEVERBIND))
 	     enumerate_interfaces(0); 
-	   if (!loopback_exception(listen->fd, listen->family, &dst_addr, ifr.ifr_name) &&
-	       !label_exception(if_index, listen->family, &dst_addr))
+	   if (!loopback_exception(listen->fd, family, &dst_addr, ifr.ifr_name) &&
+	       !label_exception(if_index, family, &dst_addr))
 	     return;
 	}
 
-      if (listen->family == AF_INET && option_bool(OPT_LOCALISE))
+      if (family == AF_INET && option_bool(OPT_LOCALISE))
 	{
 	  struct irec *iface;
 	  
@@ -1500,7 +1501,7 @@ void receive_query(struct listener *listen, time_t now)
 #endif
       char *types = querystr(auth_dns ? "auth" : "query", type);
       
-      if (listen->family == AF_INET) 
+      if (family == AF_INET) 
 	log_query(F_QUERY | F_IPV4 | F_FORWARD, daemon->namebuff, 
 		  (union all_addr *)&source_addr.in.sin_addr, types);
       else

src/netlink.c

@@ -49,7 +49,7 @@ static u32 netlink_pid;
 
 static void nl_async(struct nlmsghdr *h);
 
-void netlink_init(void)
+char *netlink_init(void)
 {
   struct sockaddr_nl addr;
   socklen_t slen = sizeof(addr);
@@ -82,16 +82,21 @@ void netlink_init(void)
     }
   
   if (daemon->netlinkfd == -1 || 
-      (daemon->kernel_version >= KERNEL_VERSION(2,6,30) &&
-       setsockopt(daemon->netlinkfd, SOL_NETLINK, NETLINK_NO_ENOBUFS, &opt, sizeof(opt)) == -1) ||
       getsockname(daemon->netlinkfd, (struct sockaddr *)&addr, &slen) == -1)
     die(_("cannot create netlink socket: %s"), NULL, EC_MISC);
   
+  
   /* save pid assigned by bind() and retrieved by getsockname() */ 
   netlink_pid = addr.nl_pid;
   
   iov.iov_len = 100;
   iov.iov_base = safe_malloc(iov.iov_len);
+  
+  if (daemon->kernel_version >= KERNEL_VERSION(2,6,30) &&
+      setsockopt(daemon->netlinkfd, SOL_NETLINK, NETLINK_NO_ENOBUFS, &opt, sizeof(opt)) == -1)
+    return _("warning: failed to set NETLINK_NO_ENOBUFS on netlink socket");
+  
+  return NULL;
 }
 
 static ssize_t netlink_recv(void)

src/network.c

@@ -388,10 +388,11 @@ static int iface_allowed(struct iface_param *param, int if_index, char *label,
   /* check whether the interface IP has been added already 
      we call this routine multiple times. */
   for (iface = daemon->interfaces; iface; iface = iface->next) 
-    if (sockaddr_isequal(&iface->addr, addr))
+    if (sockaddr_isequal(&iface->addr, addr) && iface->index == if_index)
       {
 	iface->dad = !!(iface_flags & IFACE_TENTATIVE);
 	iface->found = 1; /* for garbage collection */
+	iface->netmask = netmask;
 	return 1;
       }
 
@@ -532,7 +533,82 @@ static int iface_allowed_v4(struct in_addr local, int if_index, char *label,
 
   return iface_allowed((struct iface_param *)vparam, if_index, label, &addr, netmask, prefix, 0);
 }
-   
+
+/*
+ * Clean old interfaces no longer found.
+ */
+static void clean_interfaces()
+{
+  struct irec *iface;
+  struct irec **up = &daemon->interfaces;
+
+  for (iface = *up; iface; iface = *up)
+  {
+    if (!iface->found && !iface->done)
+      {
+        *up = iface->next;
+        free(iface->name);
+        free(iface);
+      }
+    else
+      {
+        up = &iface->next;
+      }
+  }
+}
+
+/** Release listener if no other interface needs it.
+ *
+ * @return 1 if released, 0 if still required
+ */
+static int release_listener(struct listener *l)
+{
+  if (l->used > 1)
+    {
+      struct irec *iface;
+      for (iface = daemon->interfaces; iface; iface = iface->next)
+	if (iface->done && sockaddr_isequal(&l->addr, &iface->addr))
+	  {
+	    if (iface->found)
+	      {
+		/* update listener to point to active interface instead */
+		if (!l->iface->found)
+		  l->iface = iface;
+	      }
+	    else
+	      {
+		l->used--;
+		iface->done = 0;
+	      }
+	  }
+
+      /* Someone is still using this listener, skip its deletion */
+      if (l->used > 0)
+	return 0;
+    }
+
+  if (l->iface->done)
+    {
+      int port;
+
+      port = prettyprint_addr(&l->iface->addr, daemon->addrbuff);
+      my_syslog(LOG_DEBUG, _("stopped listening on %s(#%d): %s port %d"),
+		l->iface->name, l->iface->index, daemon->addrbuff, port);
+      /* In case it ever returns */
+      l->iface->done = 0;
+    }
+
+  if (l->fd != -1)
+    close(l->fd);
+  if (l->tcpfd != -1)
+    close(l->tcpfd);
+  if (l->tftpfd != -1)
+    close(l->tftpfd);
+
+  free(l);
+  return 1;
+}
+
 int enumerate_interfaces(int reset)
 {
   static struct addrlist *spare = NULL;
@@ -630,6 +706,7 @@ int enumerate_interfaces(int reset)
 	 in OPT_CLEVERBIND mode, that at listener will just disappear after
 	 a call to enumerate_interfaces, this is checked OK on all calls. */
       struct listener *l, *tmp, **up;
+      int freed = 0;
       
       for (up = &daemon->listeners, l = daemon->listeners; l; l = tmp)
 	{
@@ -637,25 +714,17 @@ int enumerate_interfaces(int reset)
 	  
 	  if (!l->iface || l->iface->found)
 	    up = &l->next;
-	  else
+	  else if (release_listener(l))
 	    {
-	      *up = l->next;
-	      
-	      /* In case it ever returns */
-	      l->iface->done = 0;
-	      
-	      if (l->fd != -1)
-		close(l->fd);
-	      if (l->tcpfd != -1)
-		close(l->tcpfd);
-	      if (l->tftpfd != -1)
-		close(l->tftpfd);
-	      
-	      free(l);
+	      *up = tmp;
+	      freed = 1;
 	    }
 	}
+
+      if (freed)
+	clean_interfaces();
     }
-  
+
   errno = errsave;
   spare = param.spare;
     
@@ -895,10 +964,11 @@ static struct listener *create_listeners(union mysockaddr *addr, int do_tftp, in
     {
       l = safe_malloc(sizeof(struct listener));
       l->next = NULL;
-      l->family = addr->sa.sa_family;
       l->fd = fd;
       l->tcpfd = tcpfd;
-      l->tftpfd = tftpfd;	
+      l->tftpfd = tftpfd;
+      l->addr = *addr;
+      l->used = 1;
       l->iface = NULL;
     }
 
@@ -937,20 +1007,48 @@ void create_wildcard_listeners(void)
   daemon->listeners = l;
 }
 
+static struct listener *find_listener(union mysockaddr *addr)
+{
+  struct listener *l;
+  for (l = daemon->listeners; l; l = l->next)
+    if (sockaddr_isequal(&l->addr, addr))
+      return l;
+  return NULL;
+}
+
 void create_bound_listeners(int dienow)
 {
   struct listener *new;
   struct irec *iface;
   struct iname *if_tmp;
+  struct listener *existing;
 
   for (iface = daemon->interfaces; iface; iface = iface->next)
-    if (!iface->done && !iface->dad && iface->found &&
-	(new = create_listeners(&iface->addr, iface->tftp_ok, dienow)))
+    if (!iface->done && !iface->dad && iface->found)
       {
-	new->iface = iface;
-	new->next = daemon->listeners;
-	daemon->listeners = new;
-	iface->done = 1;
+	existing = find_listener(&iface->addr);
+	if (existing)
+	  {
+	    iface->done = 1;
+	    existing->used++; /* increase usage counter */
+	  }
+	else if ((new = create_listeners(&iface->addr, iface->tftp_ok, dienow)))
+	  {
+	    new->iface = iface;
+	    new->next = daemon->listeners;
+	    daemon->listeners = new;
+	    iface->done = 1;
+
+	    /* Don't log the initial set of listen addresses created
+               at startup, since this is happening before the logging
+               system is initialised and the sign-on printed. */
+            if (!dienow)
+              {
+		int port = prettyprint_addr(&iface->addr, daemon->addrbuff);
+		my_syslog(LOG_DEBUG, _("listening on %s(#%d): %s port %d"),
+			  iface->name, iface->index, daemon->addrbuff, port);
+	      }
+	  }
       }
 
   /* Check for --listen-address options that haven't been used because there's
@@ -970,6 +1068,12 @@ void create_bound_listeners(int dienow)
       {
 	new->next = daemon->listeners;
 	daemon->listeners = new;
+
+	if (!dienow)
+	  {
+	    int port = prettyprint_addr(&if_tmp->addr, daemon->addrbuff);
+	    my_syslog(LOG_DEBUG, _("listening on %s port %d"), daemon->addrbuff, port);
+	  }
       }
 }
 
@@ -1301,7 +1405,7 @@ void pre_allocate_sfds(void)
 	errno != 0 &&
 	option_bool(OPT_NOWILD))
       {
-	prettyprint_addr(&srv->source_addr, daemon->namebuff);
+	(void)prettyprint_addr(&srv->source_addr, daemon->namebuff);
 	if (srv->interface[0] != 0)
 	  {
 	    strcat(daemon->namebuff, " ");

src/option.c

@@ -1225,7 +1225,7 @@ static int parse_dhcp_opt(char *errstr, char *arg, int flags)
 	  } 
 	else if (c == '.')	
 	  {
-	    is_addr6 = is_dec = is_hex = 0;
+	    is_dec = is_hex = 0;
 	    dots++;
 	  }
 	else if (c == '-')
@@ -2991,7 +2991,6 @@ static int one_opt(int option, char *arg, char *errstr, char *gen_err, int comma
 	struct dhcp_context *new = opt_malloc(sizeof(struct dhcp_context));
 	
 	memset (new, 0, sizeof(*new));
-	new->lease_time = DEFLEASE;
 	
 	while(1)
 	  {
@@ -3041,6 +3040,7 @@ static int one_opt(int option, char *arg, char *errstr, char *gen_err, int comma
 	if (inet_pton(AF_INET, a[0], &new->start))
 	  {
 	    new->next = daemon->dhcp;
+	    new->lease_time = DEFLEASE;
 	    daemon->dhcp = new;
 	    new->end = new->start;
 	    if (strcmp(a[1], "static") == 0)
@@ -3088,6 +3088,7 @@ static int one_opt(int option, char *arg, char *errstr, char *gen_err, int comma
 	    new->flags |= CONTEXT_V6; 
 	    new->prefix = 64; /* default */
 	    new->end6 = new->start6;
+	    new->lease_time = DEFLEASE6;
 	    new->next = daemon->dhcp6;
 	    daemon->dhcp6 = new;
 
@@ -3187,7 +3188,10 @@ static int one_opt(int option, char *arg, char *errstr, char *gen_err, int comma
 	      }
 	    
 	    if (strcmp(a[leasepos], "infinite") == 0)
-	      new->lease_time = 0xffffffff;
+	      {
+		new->lease_time = 0xffffffff;
+		new->flags |= CONTEXT_SETLEASE;
+	      }
 	    else if (strcmp(a[leasepos], "deprecated") == 0)
 	      new->flags |= CONTEXT_DEPRECATE;
 	    else
@@ -3226,6 +3230,7 @@ static int one_opt(int option, char *arg, char *errstr, char *gen_err, int comma
 		      ret_err_free(_("bad dhcp-range"), new);
 		    
 		    new->lease_time = atoi(a[leasepos]) * fac;
+		    new->flags |= CONTEXT_SETLEASE;
 		    /* Leases of a minute or less confuse
 		       some clients, notably Apple's */
 		    if (new->lease_time < 120)
@@ -3233,6 +3238,7 @@ static int one_opt(int option, char *arg, char *errstr, char *gen_err, int comma
 		  }
 	      }
 	  }
+
 	break;
       }
 

src/radv.c

@@ -626,8 +626,11 @@ static int add_prefixes(struct in6_addr *local,  int prefix,
 		    real_prefix = context->prefix;
 		  }
 
-		/* find floor time, don't reduce below 3 * RA interval. */
-		if (time > context->lease_time)
+		/* find floor time, don't reduce below 3 * RA interval.
+		   If the lease time has been left as default, don't
+		   use that as a floor. */
+		if ((context->flags & CONTEXT_SETLEASE) &&
+		    time > context->lease_time)
 		  {
 		    time = context->lease_time;
 		    if (time < ((unsigned int)(3 * param->adv_interval)))

src/tftp.c

@@ -61,8 +61,9 @@ void tftp_request(struct listener *listen, time_t now)
   char *prefix = daemon->tftp_prefix;
   struct tftp_prefix *pref;
   union all_addr addra;
+  int family = listen->addr.sa.sa_family;
   /* Can always get recvd interface for IPv6 */
-  int check_dest = !option_bool(OPT_NOWILD) || listen->family == AF_INET6;
+  int check_dest = !option_bool(OPT_NOWILD) || family == AF_INET6;
   union {
     struct cmsghdr align; /* this ensures alignment */
     char control6[CMSG_SPACE(sizeof(struct in6_pktinfo))];
@@ -121,10 +122,10 @@ void tftp_request(struct listener *listen, time_t now)
       if (msg.msg_controllen < sizeof(struct cmsghdr))
         return;
       
-      addr.sa.sa_family = listen->family;
+      addr.sa.sa_family = family;
       
 #if defined(HAVE_LINUX_NETWORK)
-      if (listen->family == AF_INET)
+      if (family == AF_INET)
 	for (cmptr = CMSG_FIRSTHDR(&msg); cmptr; cmptr = CMSG_NXTHDR(&msg, cmptr))
 	  if (cmptr->cmsg_level == IPPROTO_IP && cmptr->cmsg_type == IP_PKTINFO)
 	    {
@@ -138,7 +139,7 @@ void tftp_request(struct listener *listen, time_t now)
 	    }
       
 #elif defined(HAVE_SOLARIS_NETWORK)
-      if (listen->family == AF_INET)
+      if (family == AF_INET)
 	for (cmptr = CMSG_FIRSTHDR(&msg); cmptr; cmptr = CMSG_NXTHDR(&msg, cmptr))
 	  {
 	    union {
@@ -154,7 +155,7 @@ void tftp_request(struct listener *listen, time_t now)
 	  }
       
 #elif defined(IP_RECVDSTADDR) && defined(IP_RECVIF)
-      if (listen->family == AF_INET)
+      if (family == AF_INET)
 	for (cmptr = CMSG_FIRSTHDR(&msg); cmptr; cmptr = CMSG_NXTHDR(&msg, cmptr))
 	  {
 	    union {
@@ -171,7 +172,7 @@ void tftp_request(struct listener *listen, time_t now)
 	  
 #endif
 
-      if (listen->family == AF_INET6)
+      if (family == AF_INET6)
         {
           for (cmptr = CMSG_FIRSTHDR(&msg); cmptr; cmptr = CMSG_NXTHDR(&msg, cmptr))
             if (cmptr->cmsg_level == IPPROTO_IPV6 && cmptr->cmsg_type == daemon->v6pktinfo)
@@ -194,7 +195,7 @@ void tftp_request(struct listener *listen, time_t now)
       
       addra.addr4 = addr.in.sin_addr;
 
-      if (listen->family == AF_INET6)
+      if (family == AF_INET6)
 	addra.addr6 = addr.in6.sin6_addr;
 
       if (daemon->tftp_interfaces)
@@ -210,12 +211,12 @@ void tftp_request(struct listener *listen, time_t now)
       else
 	{
 	  /* Do the same as DHCP */
-	  if (!iface_check(listen->family, &addra, name, NULL))
+	  if (!iface_check(family, &addra, name, NULL))
 	    {
 	      if (!option_bool(OPT_CLEVERBIND))
 		enumerate_interfaces(0); 
-	      if (!loopback_exception(listen->tftpfd, listen->family, &addra, name) &&
-		  !label_exception(if_index, listen->family, &addra))
+	      if (!loopback_exception(listen->tftpfd, family, &addra, name) &&
+		  !label_exception(if_index, family, &addra))
 		return;
 	    }
 	  
@@ -281,7 +282,7 @@ void tftp_request(struct listener *listen, time_t now)
 	  prefix = pref->prefix;  
     }
 
-  if (listen->family == AF_INET)
+  if (family == AF_INET)
     {
       addr.in.sin_port = htons(port);
 #ifdef HAVE_SOCKADDR_SA_LEN
@@ -304,7 +305,7 @@ void tftp_request(struct listener *listen, time_t now)
   
   if (option_bool(OPT_SINGLE_PORT))
     transfer->sockfd = listen->tftpfd;
-  else if ((transfer->sockfd = socket(listen->family, SOCK_DGRAM, 0)) == -1)
+  else if ((transfer->sockfd = socket(family, SOCK_DGRAM, 0)) == -1)
     {
       free(transfer);
       return;
@@ -322,7 +323,7 @@ void tftp_request(struct listener *listen, time_t now)
   transfer->opt_blocksize = transfer->opt_transize = 0;
   transfer->netascii = transfer->carrylf = 0;
  
-  prettyprint_addr(&peer, daemon->addrbuff);
+  (void)prettyprint_addr(&peer, daemon->addrbuff);
   
   /* if we have a nailed-down range, iterate until we find a free one. */
   while (!option_bool(OPT_SINGLE_PORT))
@@ -337,7 +338,7 @@ void tftp_request(struct listener *listen, time_t now)
 	    {
 	      if (++port <= daemon->end_tftp_port)
 		{ 
-		  if (listen->family == AF_INET)
+		  if (family == AF_INET)
 		    addr.in.sin_port = htons(port);
 		  else
 		    addr.in6.sin6_port = htons(port);
@@ -375,7 +376,7 @@ void tftp_request(struct listener *listen, time_t now)
 	      if ((opt = next(&p, end)) && !option_bool(OPT_TFTP_NOBLOCK))
 		{
 		  /* 32 bytes for IP, UDP and TFTP headers, 52 bytes for IPv6 */
-		  int overhead = (listen->family == AF_INET) ? 32 : 52;
+		  int overhead = (family == AF_INET) ? 32 : 52;
 		  transfer->blocksize = atoi(opt);
 		  if (transfer->blocksize < 1)
 		    transfer->blocksize = 1;
@@ -624,7 +625,7 @@ void check_tftp_listeners(time_t now)
 	    {
 	      strcpy(daemon->namebuff, transfer->file->filename);
 	      sanitise(daemon->namebuff);
-	      prettyprint_addr(&transfer->peer, daemon->addrbuff);
+	      (void)prettyprint_addr(&transfer->peer, daemon->addrbuff);
 	      my_syslog(MS_TFTP | LOG_INFO, endcon ? _("failed sending %s to %s") : _("sent %s to %s"), daemon->namebuff, daemon->addrbuff);
 	      /* unlink */
 	      *up = tmp;
@@ -667,7 +668,7 @@ static void handle_tftp(time_t now, struct tftp_transfer *transfer, ssize_t len)
 	  char *end = daemon->packet + len;
 	  char *err = next(&p, end);
 	  
-	  prettyprint_addr(&transfer->peer, daemon->addrbuff);
+	  (void)prettyprint_addr(&transfer->peer, daemon->addrbuff);
 	  
 	  /* Sanitise error message */
 	  if (!err)