-This version of tinc supports multiple virtual networks at once. To use this
-feature, you may supply a netname via the -n or --net options. The standard
-locations for the config files will then be /etc/tinc/<net>/. Because of this
-feature, tinc will send packets directly to their destinations, instead of to
-the uplink. If this behaviour is undesirable (for instance because of firewalls
-or other restrictions), please use an older version of tinc (I would recommend
-tinc-0.2.19).
-
-In order to force the kernel to accept received packets, the destination MAC
-address will be set to FE:FD:00:00:00:00 upon reception. The MAC address of the
-ethertap or tun/tap interface must also be set to this address. See the manual
-for more detailed information.
-
-tincd regenerates its encryption key pairs. It does this on the first activity
-after the keys have expired. This period is adjustable in the configuration
-file, and the default time is 3600 seconds (one hour).
-
-This version supports multiple subnets at once. They are also sorted on subnet
-mask size. This means that it is possible to have overlapping subnets on the
-VPN, as long as their subnet mask sizes differ.
+This version of tinc supports multiple virtual networks at once. To
+use this feature, you may supply a netname via the -n or --net
+options. The standard locations for the config files will then be
+/etc/tinc/<net>/.
+
+tincd regenerates its encryption key pairs. It does this on the first
+activity after the keys have expired. This period is adjustable in the
+configuration file, and the default time is 3600 seconds (one hour).
+
+This version supports multiple subnets at once. They are also sorted
+on subnet mask size. This means that it is possible to have
+overlapping subnets on the VPN, as long as their subnet mask sizes
+differ.
+
+Since pre5, tinc can operate in several routing modes. The default mode,
+"router", works exactly like the older version, and uses Subnet lines to
+determine the destination of packets. The other two modes, "switch" and "hub",
+allow the tinc daemons to work together like a single network switch or hub.
+This is useful for bridging networks. The latter modes only work properly on
+Linux, FreeBSD and Windows.
+
+The algorithms used for encryption and generating message authentication codes
+can now be changed in the configuration files. All cipher and digest algorithms
+supported by OpenSSL can be used. Useful ciphers are "blowfish" (default),
+"bf-ofb", "des", "des3", etcetera. Useful digests are "sha1" (default), "md5",
+etcetera.
+
+Support for routing IPv6 packets has been added. Just add Subnet lines with
+IPv6 addresses (without using :: abbreviations) and use ifconfig or ip (from
+the iproute package) to give the virtual network interface corresponding IPv6
+addresses. tinc does not provide autoconfiguration for IPv6 hosts, if you need
+it use radvd or zebra.
+
+It is also possible to make tunnels to other tinc daemons over IPv6 networks,
+if the operating system supports IPv6. tinc will automatically use both IPv6
+and IPv4 when available, but this can be changed by adding the option
+"AddressFamily = ipv4" or "AddressFamily = ipv6" to the tinc.conf file.
+
+Normally, when started tinc will detach and run in the background. In a native
+Windows environment this means tinc will intall itself as a service, which will
+restart after reboots. To prevent tinc from detaching or running as a service,
+use the -D option.
+