Index of /vet

      Name                    Last modified       Size  Description


[DIR] Parent Directory        01-Jul-2010 18:29      -  
[TXT] iproute2_diff.txt       30-May-2008 16:52     5k  
[TXT] vet_diff_linux_2_6_2..> 30-May-2008 16:52    14k  
[TXT] vet_diff_net-2_6.txt    31-May-2008 21:28    13k  

The files vet_diff_net-2_6.txt and iproute2_diff.txt include source code
changes to implement the Virtual Ethernet (VET) interface abstraction,
which is documented in 'draft-templin-autoconf-dhcp'. Diffs are relative
to the Linux 2.6 network developer's (netdev) and shemminger iproute2
'git' trees captured on 5/6/08. 'git' commands used were:

  # git clone git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6.git
  # git clone git://git.kernel.org/pub/scm/linux/kernel/git/shemminger/iproute2.git

Build the vet_diff_net-2.6.txt diffs into the kernel source tree
using 'patch' as-normal, and fix-up any errors. Do the same with the
iproute2_diff.txt diffs wrt the iproute2 source tree. Build and install
both the kernel and 'ip', then boot to enable VET.

***********************************************************************
NEW: The kernel diffs have now been back-ported to the public-release
'linux-2.6.25.3' kernel. The diff file is vet_diff_linux_2_6_25_3.txt.
The '2.6.25.3' kernel is obtained from:

ftp://ftp.kernel.org/pub/linux/kernel/v2.6/linux-2.6.25.3.tar.bz2

and is patched exactly as for net-2.6 above.
***********************************************************************

The VET changes add a new tunnel virtual interface abstration that
is a superset of the existing ISATAP abstraction. It adds multicast
support vis-a-vis RFC2529 (6over4) multicast mapping and retains
the unicast features of ISATAP. To enabele a VET interface, first
configure an IPv4 interface (e.g., eth0) using either manual config
or DHCP to add an IPv4 address (e.g., 192.168.1.100) to the interface.
Next, use the modified 'ip' tool to add and enable a vet interface
as follows:

  # ip tunnel add vet0 mode vet ttl 255 local 192.168.1.100 dev eth0
  # ip link set dev vet0 up

The "dev eth0" in the 'ip tunnel add' binds the vet device to the
underlying eth0 device. This is the device over which IPv6 MLD will
be mapped to IPv4 IGMP. After the vet0 interface is enabled, IPv6
multicast will work over the vet interface as-normal. The only
difference is that the vet interface does not permit multicast
RS/RA exchanges and uses static calculations for address resolution
instead of multicast NS/NA. So, IPv6 ND is handled via unicast
(ISATAP-mode) while all other multicast is disseminated throughout
the site (6over4-mode).

Site border routers need to filter IPv4 multicast groups 239.192/16
and not let them go through to other networks.