(usagi-users 02896) Re: USAGI MIPL problems

[ognjen dobrijenic <ognjen_d@xxxxxxxxx>]

Hi, 

--- namviet <tools4the@xxxxxxxxx> wrote:
> There are some slight differences in
> network-mip6.conf
> and MIP configuration from version 0.9.x (bundled
> with
> USAGI) and the latest 1.0. So, follow the
> corresponding README carefully.

 I have configured my network-mip6.conf according to
USAGI's docs/HOWTO/MIP6 so I think this shouldn't
cause any problems.
 
> After starting MN, you will always have 2 IPv6
> addresses at the mobile interface: stateless
> autoconfig address and home address (as defined in
> network-mip6.conf). 
>
> When your MN is away at foreign link, it will
> acquire
> another stateless autoconfig address with foreign
> prefix as care-of-address  (so virtually 3 IPv6
> addresses at mobile interface).

 While I was using HUT's MIPL 0.9.5.1 this wasn't the
case, after starting MN, my mobile interface had only
had one address - home address, and of course, while
away from home, it would get stateless autoconfig
address with foreign prefix. And this is supposed to
be almost the same MIP6 stack, as the one bundled in
USAGI, or? Anyhow, I don't see the need, for MN, to
get stateless autoconfig address immediately after
starting. Or is this done by default?   

> Only the home address will be binding-updated to HA.
> I
> dont know what you mean by 2 "bindings" at HA. Is
> that
> the stateless autoconfig address with home link
> prefix
> also binding-updated???

No, this is the most strange thing. Both "bindings" at
HA say that a home address is being
binding-updated?!?! As I could figure it out from my
Ethereal capture file, MN sends BU, HA replys with BA,
but then MN sends another BU?! This second one is, as
I have mentioned, properly chached at HA, and HA
doesn't send second BA. I guess the source is somehow
corrupted or?  
 
> In my opinion, you may eliminate one IPv6 address by
> specifying homeaddress in network-mip6.conf exactly
> the same as stateless autoconf address with homelink
> prefix. Otherwise you may delete the autoconf
> address
> manually.

Thanks for the tip, that's what I also had in mind...
 
> Personally I have had a lot of troubles with MIPv6
> stack bundled in USAGI January stable kit. So it may
> be better to stick with the earlier snapshot dated
> 03
> Aug 2003. Note a minor mistake in MIPv6 how-to in
> 20030804 snapshot: IPv6 subtrees option should
> always
> be enabled (even the CN).
> 

While compiling kernel for stable release 5, did you
enable that option at both HA and MN? Because it
doesn't say so...

Thanks a lot for your tips, hope all these questions
won't bother you too much :-)  

              Regards, 
                      Ognjen


> --- ognjen dobrijenic <ognjen_d@xxxxxxxxx> wrote:
> > Hi Dave,
> >  first of all, thank you for your quick
> response...
> > 
> > This concerning recognition of BU/BA, I had
> > previously
> > installed HUT's MIPL 0.9.5.1 and 1.0 and the
> version
> > of Ethereal that I have been using from the start
> > was
> > always capable of recognizing BAs/BUs. Do you
> think
> > perhaps I should degrade my Ethereal to some of
> the
> > previous versions? I don't think so, because
> MIPL's
> > 0.9.5(.1) is integrated in USAGI so I assume that
> > Mobility Header's definition is the same.
> > 
> > About this second thing, your assumption is
> > completely
> > accurate. I had totally forgotten to mention this
> in
> > my previous message. The thing is, after boot my
> > MN's
> > interface really has 2 addresses, one that gets
> from
> > ifcfg-eth1 script and another that is defined
> after
> > MIP6 script starts. This latter one is a care-of
> > address (assigned to the interface during boot)
> made
> > with the home-link prefix. It has confused me a
> bit,
> > because I didn't encounter it before and I
> obviously
> > don't need it. What about this care-of address, do
> > you
> > perhaps know how to turn off this stateless
> > autoconfiguration during boot?
> >  
> >     Best regards,
> >                   Ognjen     
> >     
> > --- "Craig, Dave" <dwcraig@xxxxxxxxxxxx> wrote:
> > > I also encountered problems recognizing BA/BU
> > > packets when using an
> > > older version of Ethereal.  Upgrading to the
> > latest
> > > version of Ethereal
> > > has the latest definition for Mobility Header
> > (135).
> > > 
> > > I have not seen the dual binding problem before.
> 
> > A
> > > binding is supposed
> > > to replace the previous binding.  If there are
> two
> > > Binding Updates sent,
> > > the HA would only store one binding.  Do you
> have
> > > two addresses
> > > configured?
> > > 
> > > 	Dave
> > 
> > __________________________________
> > Do you Yahoo!?
> > Yahoo! Small Business $15K Web Design Giveaway 
> > http://promotions.yahoo.com/design_giveaway/
> > 
> > 
> 
> 
> __________________________________
> Do you Yahoo!?
> Yahoo! Tax Center - File online by April 15th
> http://taxes.yahoo.com/filing.html
> 
> 


__________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online by April 15th
http://taxes.yahoo.com/filing.html