note the new method of changing link layer addresses using a PF_LINK

socket ioctl.

share/man/man4/tap.4

@@ -1,4 +1,4 @@
-.\" $NetBSD: tap.4,v 1.9 2009/03/09 18:09:39 plunky Exp $
+.\" $NetBSD: tap.4,v 1.10 2009/03/12 11:34:54 plunky Exp $
 .\"
 .\"  Copyright (c) 2004, 2005 The NetBSD Foundation.
 .\"  All rights reserved.
@@ -24,7 +24,7 @@
 .\"  ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
 .\"  POSSIBILITY OF SUCH DAMAGE.
 .\"
-.Dd March 9, 2009
+.Dd March 10, 2009
 .Dt TAP 4
 .Os
 .Sh NAME
@@ -151,23 +151,13 @@ When a
 .Nm
 device is created, it is assigned an Ethernet address
 of the form f2:0b:a4:xx:xx:xx.
-This address can later be changed in two ways:
-through a sysctl node, or an ioctl call.
-.Pp
-The sysctl node is net.link.tap.\*[Lt]iface\*[Gt].
-Any string of six colon-separated hexadecimal numbers will be accepted.
-Reading that node will provide a string representation of the current
-Ethernet address.
-.Pp
-The address can also be changed with the
-.Dv SIOCSIFPHYADDR
-ioctl, which is used the same way as with
-.Xr gif 4 .
-The difference is in the family of the address which is passed inside the
-.Ft struct ifaliasreq
-argument, which should be set to
-.Dv AF_LINK .
-This ioctl call should be made on a socket, as it is not available on
+This address can later be changed using
+.Xr ifconfig 8
+to add an active link layer address, or directly via the 
+.Dv SIOCALIFADDR
+ioctl on a
+.Dv PF_LINK
+socket, as it is not available on
 the ioctl handler of the character device interface.
 .Sh FILES
 .Bl -tag -compact -width /dev/tap[0-9]*
@@ -179,7 +169,6 @@ individual character device nodes
 .Sh SEE ALSO
 .Xr bridge 4 ,
 .Xr etherip 4 ,
-.Xr gif 4 ,
 .Xr tun 4 ,
 .Xr ifconfig 8
 .Sh HISTORY