[support] Destination unreachable from HA on BU???

[Mattias Blomqvist]

Hello again,

Just an update. After moving to kernel 2.6.29 it works very well over ethernet.
Running MN on debian 5.0 kernel 2.6.26-1 was the source of the problem.
And it resulted in very very strange behaviour...

Back to see if I can get it to work over a tunnel interface...

BR,
Mattias

On Wed, Mar 25, 2009 at 2:25 PM, Mattias Blomqvist
<mattias.blomqvist at gmail.com> wrote:
> Hi
>
> Yes I was trying to run this over a tunnel (tun driver).
> But now I have moved back to ethernet with the same result.
> Yes, UMIP on MN as well. Config after backing to ethernet is:
>
> NodeConfig MN;
> DebugLevel 10;
> DoRouteOptimizationCN disabled;
> DoRouteOptimizationMN disabled;
> UseCnBuAck enabled;
> MnDiscardHaParamProb enabled;
> Interface "eth1" {
>        Tunnel disabled;
> }
> MnRouterProbes 1;
> MnHomeLink "eth1" {
>        HomeAddress 2a03:a0a:0:1300::1/32;
>        HomeAgentAddress 2a03:a0a::1;
> }
> UseMnHaIPsec disabled;
> KeyMngMobCapability disabled;
>
>
> When I moved back to ethernet and plugged the mobile node into the
> home network it seemed to work fine: The following is a BU from that
> case:
> 14:52:40.670292 IP6 (hlim 64, next-header Mobility (135) payload
> length: 32) xxxx:xxxx:0:1300::1 > xxxx:xxxx::1: mobility: BU
> seq#=26946 AH lifetime=0(padn)(alt-CoA: xxxx:xxxx:0:1300::1)
>        0x0000:  6000 0000 0020 8740 xxxx xxxx 0000 1300
>        0x0010:  0000 0000 0000 0001 xxxx xxxx 0000 0000
>        0x0020:  0000 0000 0000 0001 3b03 0500 cfd7 6942
>        0x0030:  c000 0000 0100 0310 xxxx xxxx 0000 1300
>        0x0040:  0000 0000 0000 0001
> And its BA:
> 14:52:40.674322 IP6 (hlim 64, next-header Mobility (135) payload
> length: 16) xxxx:xxxx::1 > xxxx:xxxx:0:1300::1: mobility: BA
> status=133 seq#=26946 lifetime=0(padn)
>        0x0000:  6000 0000 0010 8740 xxxx xxxx 0000 0000
>        0x0010:  0000 0000 0000 0001 xxxx xxxx 0000 1300
>        0x0020:  0000 0000 0000 0001 3b01 0600 5406 8500
>        0x0030:  6942 0000 0102 0000
>
>
> But when I moved it behind a router it started sending BU's like it did before:
> 14:53:31.666555 IP6 (hlim 63, next-header IPv6 (41) payload length:
> 72) yyyy:yyyy::230:59ff:fe04:98a6 > xxxx:xxxx::1: IP6 (hlim 64,
> next-header Mobility (135) payload length: 32) xxxx:xxxx:0:1300::1 >
> xxxx:xxxx::1: mobility: BU seq#=4249 AH lifetime=262140(padn)(alt-CoA:
> 2a03:b0b::230:59ff:fe04:98a6)
>        0x0000:  6000 0000 0048 293f yyyy yyyy 0000 0000
>        0x0010:  0230 59ff fe04 98a6 xxxx xxxx 0000 0000
>        0x0020:  0000 0000 0000 0001 6000 0000 0020 8740
>        0x0030:  xxxx xxxx 0000 1300 0000 0000 0000 0001
>        0x0040:  xxxx xxxx 0000 0000 0000 0000 0000 0001
>        0x0050:  3b03 0500 47a6 1099 c000 ffff 0100 0310
>        0x0060:  yyyy yyyy 0000 0000 0230 59ff fe04 98a6
>
>> Having  2001:db9::13 in the altcoa option and in the *external* IPv6
>> header is just weird. 2001:db8:0:1300::1 is not an autoconfigured
>> address, so Additional questions about your MN:
>
> 2001:db9::13 was the CoA in the tunnel config. When running it over
> ethernet (see tcpdumps above), the same thing happens. alt-CoA and
> source address are the same.
>
>> - how does it get that 2001:db8:0:1300::1 address?
> This is a static address used as HoA.
>> - you are also running UMIP on the MN? can you send the config?
> Yes. umip 0.4-12 from natisbad.org.
>> - On a recent kernel?
> Currently on debian 5.0 standard kernel 2.6.26-1. Compiling 2.6.29 for MN now.
>> - Any exotic network config (i.e. not ethernet or wifi), like some
>>  tunneling mechanism ?
> See above.
>
> I really appreciate all the help.
>
> BR,
> Mattias Blomqvist
>
>
>
> On Wed, Mar 25, 2009 at 1:31 PM, Arnaud Ebalard <arno at natisbad.org> wrote:
>> Hi,
>>
>> Romain KUNTZ <kuntz at lsiit.u-strasbg.fr> writes:
>>
>>> On 2009/03/25, at 11:34, Mattias Blomqvist wrote:
>>>> Tcpdump of a received BU:
>>>> HA:~# tcpdump -vvv -x -i eth0 -s 200 host xxxx:xxxx::1
>>>> tcpdump: listening on eth0, link-type EN10MB (Ethernet), capture
>>>> size 200 bytes
>>>> 12:16:42.450564 IP6 (hlim 63, next-header IPv6 (41)
>>>
>>> Next header should be destination option header (0x3c)
>>
>> Yep, that's weird. The layout of the packet imported in Scapy is the
>> following (replaced xxxx xxxx and yyyy yyyy values by 2001:db8:: and
>> 2001:db8::):
>>
>> ###[ IPv6 ]###
>>  version= 6L
>>  tc= 0L
>>  fl= 0L
>>  plen= 72
>>  nh= IPv6
>>  hlim= 63
>>  src= 2001:db9::13
>>  dst= 2001:db8::1
>> ###[ IPv6 ]###
>>     version= 6L
>>     tc= 0L
>>     fl= 0L
>>     plen= 32
>>     nh= Mobility Header
>>     hlim= 64
>>     src= 2001:db8:0:1300::1
>>     dst= 2001:db8::1
>> ###[ IPv6 Mobility Header - Binding Update ]###
>>        nh= No Next Header
>>        len= 3
>>        mhtype= BU
>>        res= 0
>>        cksum= 0x20a
>>        seq= 0x48fd
>>        flags= HA
>>        reserved= 0x0L
>>        mhtime= 262140 sec
>>        autopad= On
>>        \options\
>>         |###[ PadN ]###
>>         |  otype= PadN [00: skip, 0: Don't change en-route]
>>         |  optlen= 0
>>         |  optdata= ''
>>         |###[ MIPv6 Option - Alternate Care-of Address ]###
>>         |  otype= Alternate Care-of Address
>>         |  olen= 16
>>         |  acoa= 2001:db9::13
>>
>> Having  2001:db9::13 in the altcoa option and in the *external* IPv6
>> header is just weird. 2001:db8:0:1300::1 is not an autoconfigured
>> address, so Additional questions about your MN:
>>
>> - how does it get that 2001:db8:0:1300::1 address?
>> - you are also running UMIP on the MN? can you send the config?
>> - On a recent kernel?
>> - Any exotic network config (i.e. not ethernet or wifi), like some
>>  tunneling mechanism ?
>>
>> Cheers,
>>
>> a+
>>
>