To the Linux folks having problems, please check out www.linux-ipv6.org . This is home
of the USAGI project (part of WIDE project in Japan, headed by Dr. Jun Murai). Following
this is the description of USAGI from this site.
 
Note that WIDE also does the KAME IPv6 stack, which is included by default in *BSD.
It is the best and most complete IPv6 I have run across. USAGI is an attempt to provide
the same basic functionality for Linux, although they seem to run a bit behind KAME.
 
To those who don't speak Japanese, you may be interested to learn that Kame means
"turtle", and Usagi means "hare". I fyou know your Aesop, you will understand why
the "turtle" is ahead in this race. ;-) The expansion of USAGI as an acronym is
clearly an after-the-fact rationalization, and would make it USAGIP anyway.
 
If it is an option for you, we have found by experience that FreeBSD & OpenBSD are
better and more complete for IPv6 work than Linux as is, or even Linux with USAGI
stack. Please no flames from Linux fans! The KAME stack is recognized worldwide
as the reference implementation of IPv6, and WIDE has done an amazing amount of
really quality work on it, and I am stating this as my opinion, based on research and
testing.
 
If your Linux distro happens to already include the USAGI stack, you can ignore this
message. Any real Linux gurus out there - do you know of any distros that do include
USAGI by default? If not, any experience using the Linux stack as is, or the USAGI
stack? If it USAGI is not already included by default, perhaps you can encourage
your favorite distro to include it?
 
Description from USAGI site follows:

Currently we have an IPv6 implementation in Linux kernel source tree. Only enabling "Internet Protocol Version 6" option in the Networking section, we can enjoy IPv6 life.

However, once you begin to use IPv6 on Linux box, you will soon be aware that the implementation have some problems... Because an existing Linux implementation is too old and not so well-tested, it has many bugs and unimplemented functions.

Then we decided to start USAGI Project(UniverSAl playGround for Ipv6 Project) with WIDE Project, KAME Project and TAHI Project. The project aim to improve IPv6 environment on Linux and deploy the IPv6 Internet on the world. We've started to hack the kernel, libraries and applications aggressively and will provide our product freely for Linux and IPv6 community. In the near future we would contribute and merge our code into the main trunks of Linux kernel and glibc.

Because of the contribution for main trunks, we have a policy that we don't make and accept changes which depend on Linux distributions. Instead the policy, we will provide binary packages for Linux distributions on every stable release. Let's try out USAGI IPv6 environment with us !!

 



From: [EMAIL PROTECTED] on behalf of Bellino, Phil
Sent: Wed 4/6/2005 12:33 AM
To: Users-IPv6; Users-Usagi; Users-Deepspace
Subject: Trouble with 2.6.11 Linux

Hello,
       
I have a 2.6.5 Linux running router radvd.
I also have 2.6.5 clients(and a 2.4.20 client) that accept the router advertisements from the router and acquire a Link-Global address and also autoconfigures their Link-Local address.

Their configs:
ipv6.conf.eth0.accept_ra=1
ipv6.conf.all.accept_ra=1
ipv6.conf.default.accept_ra=1

I have a 2.6.11 client host that does not accept any router advertisements even though it's config is the same as above. (I have compared the "sysctl -a" output on both the 2.6.5 and 2.6.11 and they are identical).  In fact the following is what occurs at boot time:

1.  When I boot up this client, eth0 does not have the inet6 Link-local address.
If I then issue:
ifconfig eth0 down
ifconfig eth0 up

The inet6 Link-local address then appears.

2. My 2.6.11 host does not learn this prefix and as a result there is no Link-Global address.

Has anyone experienced this issue.

Thank you,
Phil Bellino


============================
Phil Bellino
MRV Communications, Inc.
Boston Product Division
295 Foster St.
Littleton,MA 01460
Tel: (978)952-4807
Email: [EMAIL PROTECTED]
============================

Reply via email to