Re: [IPv6 Users] ipv6 routing, problem with neighbour solicitation

2007-06-24 Thread Volker Böhme
You made my day. Thank you for your lead using the fe80:: for the gateway 
setting. When i use these addresses for the routing everything works (on the 
real hardware as on the xen machine). Thank you very much.

Also thank you to everyone else helping me with my problem. I hope i did not 
bother you to much. 

Greets,
Volker


___
Users mailing list
[email protected]
https://lists.ipv6.org/mailman/listinfo/users


Re: [IPv6 Users] ipv6 routing, problem with neighbour solicitation

2007-06-24 Thread Volker Böhme
On Sunday 24 June 2007 09:53:34 Peter Bieringer wrote:
> You have IPv4 addresses as linklayer addresses?

Yes. I am working on different testbeds. The one i used was an xen maschine, 
where the ipv6 data is send over on ipv4 gre tunnel. The interface on which 
the gre tunnel is created has an ipv4 address. That is why the 
linklayeraddress for the wldev device is an ipv4 address.

> Which kernel version do you use?

The used kernel version is 2.6.16.29-xen but i also tried 2.6.18-xen. In the 
real testbed (not emulated with xen) kernel 2.6.16.29.

> Can you send the output on host B of
> # sysctl net.ipv6.conf.eth0

Shure, but why do you whant the output of eth0? eth0 does not have an ipv6 
address. Here the output of 

#sysctl net.ipv6.conf
net.ipv6.conf.wldev.max_addresses = 16
net.ipv6.conf.wldev.max_desync_factor = 600
net.ipv6.conf.wldev.regen_max_retry = 5
net.ipv6.conf.wldev.temp_prefered_lft = 86400
net.ipv6.conf.wldev.temp_valid_lft = 604800
net.ipv6.conf.wldev.use_tempaddr = 0
net.ipv6.conf.wldev.force_mld_version = 0
net.ipv6.conf.wldev.router_solicitation_delay = 1
net.ipv6.conf.wldev.router_solicitation_interval = 4
net.ipv6.conf.wldev.router_solicitations = 3
net.ipv6.conf.wldev.dad_transmits = 1
net.ipv6.conf.wldev.autoconf = 1
net.ipv6.conf.wldev.accept_redirects = 1
net.ipv6.conf.wldev.accept_ra = 1
net.ipv6.conf.wldev.mtu = 1472
net.ipv6.conf.wldev.hop_limit = 64
net.ipv6.conf.wldev.forwarding = 1
net.ipv6.conf.default.max_addresses = 16
net.ipv6.conf.default.max_desync_factor = 600
net.ipv6.conf.default.regen_max_retry = 5
net.ipv6.conf.default.temp_prefered_lft = 86400
net.ipv6.conf.default.temp_valid_lft = 604800
net.ipv6.conf.default.use_tempaddr = 0
net.ipv6.conf.default.force_mld_version = 0
net.ipv6.conf.default.router_solicitation_delay = 1
net.ipv6.conf.default.router_solicitation_interval = 4
net.ipv6.conf.default.router_solicitations = 3
net.ipv6.conf.default.dad_transmits = 1
net.ipv6.conf.default.autoconf = 1
net.ipv6.conf.default.accept_redirects = 1
net.ipv6.conf.default.accept_ra = 1
net.ipv6.conf.default.mtu = 1280
net.ipv6.conf.default.hop_limit = 64
net.ipv6.conf.default.forwarding = 1
net.ipv6.conf.all.max_addresses = 16
net.ipv6.conf.all.max_desync_factor = 600
net.ipv6.conf.all.regen_max_retry = 5
net.ipv6.conf.all.temp_prefered_lft = 86400
net.ipv6.conf.all.temp_valid_lft = 604800
net.ipv6.conf.all.use_tempaddr = 0
net.ipv6.conf.all.force_mld_version = 0
net.ipv6.conf.all.router_solicitation_delay = 1
net.ipv6.conf.all.router_solicitation_interval = 4
net.ipv6.conf.all.router_solicitations = 3
net.ipv6.conf.all.dad_transmits = 1
net.ipv6.conf.all.autoconf = 1
net.ipv6.conf.all.accept_redirects = 1
net.ipv6.conf.all.accept_ra = 1
net.ipv6.conf.all.mtu = 1280
net.ipv6.conf.all.hop_limit = 64
net.ipv6.conf.all.forwarding = 1
net.ipv6.conf.lo.max_addresses = 16
net.ipv6.conf.lo.max_desync_factor = 600
net.ipv6.conf.lo.regen_max_retry = 5
net.ipv6.conf.lo.temp_prefered_lft = 86400
net.ipv6.conf.lo.temp_valid_lft = 604800
net.ipv6.conf.lo.use_tempaddr = -1
net.ipv6.conf.lo.force_mld_version = 0
net.ipv6.conf.lo.router_solicitation_delay = 1
net.ipv6.conf.lo.router_solicitation_interval = 4
net.ipv6.conf.lo.router_solicitations = 3
net.ipv6.conf.lo.dad_transmits = 1
net.ipv6.conf.lo.autoconf = 1
net.ipv6.conf.lo.accept_redirects = 1
net.ipv6.conf.lo.accept_ra = 1
net.ipv6.conf.lo.mtu = 16436
net.ipv6.conf.lo.hop_limit = 64
net.ipv6.conf.lo.forwarding = 1

What are you looking for?

Greets
Volker

___
Users mailing list
[email protected]
https://lists.ipv6.org/mailman/listinfo/users


Re: [IPv6 Users] ipv6 routing, problem with neighbour solicitation

2007-06-24 Thread Volker Böhme
On Friday 22 June 2007 16:42:20 Jun-ichiro itojun Hagino wrote:
>   so in your configuration, A, B and C are all on the same link but
>   A and C has to use B as the gateway.  B would send A/B ICMPv6 redirects.

I have checked with tcpdump, B is not sending any ICMPv6 redirect messages.

>   for ICMPv6 to work correctly, you MUST use link-local addresses
>   (fe80::blah) for the gateway setting.  this is a hidden rule in
>   the RFCs.

Ok, i'll try that.

Greets,
Volker


___
Users mailing list
[email protected]
https://lists.ipv6.org/mailman/listinfo/users


Re: [IPv6 Users] ipv6 routing, problem with neighbour solicitation

2007-06-24 Thread Volker Böhme
On Saturday 23 June 2007 14:53:26 Jeroen Massar wrote:
> Volker Böhme wrote:
> > On Friday 22 June 2007 15:47:17 Jeroen Massar wrote:
>
> [..]
>
> > ip -6 route show
> > 2001:db80::11 via 2001:db80::12 dev wldev  metric 1024  expires -240sec
> > mtu 1472 advmss 1412 hoplimit 4294967295
> > 2001:db80::/32 dev wldev  metric 256  expires -440sec mtu 1472 advmss
> > 1412 hoplimit 4294967295
>
> 2001:db80::/32? I think you should be using 2001:db8::/32 there.
> Also, 2001:db8::/32 is the FULL prefix. On a link your should only be
> using a single /64 per link.

Oh, didn't realised that i was using 2001:db80:: instead of 2001:db8::/32. 
Don't know where the 0 came frome. Sorry for that. I have just tried it again 
with 2001:db80 but this does not change anything in my problem.

> Might I actually suggest that before continuing to play with this that
> you pick up a good book about IPv6? As you clearly miss some of the
> fundamentals of why this might break and how should work and moreover:
> the why behind it.
>
> Deploying IPv6 (http://www.deployingipv6.net/)
> Which should have been here:
> http://books.google.com/books?id=Q4yGCAAJ&dq=Deploying+IPv6
>
> or "Ipv6 Network Administration", see:
> http://books.google.com/books?id=UFxm48LOyugC
>
> or "IPv6 Essentials":
> http://books.google.com/books?id=925YA6QQdu0C&hl=en
>
> or most likely very appropriate for this: "Running IPv6"
> http://books.google.com/books?id=H5-yAQAACAAJ&dq=running+ipv6
>
> or "IPv6 in Practice" which might also be appropriate:
> http://books.google.com/books?id=KKKBCAAJ&dq=ipv6+in+practice
>
> Actually, I can recommend all of the books, they all have their niche.
> The IPv6 Essentials one is really good if you take a look on IPv6 from a
> more non-technical perspective but do want to get to know all the techy
> details as they are really well explained in there. The Running IPv6 one
> is very technical and explains really well on how to configure various
> operating systems and also handles routing etc. "IPv6 Practice" is the
> newest one and also covers a great deal of information.

Thanks for your advice, i allready ordered some of these books, but 
unfortunately it will take some time till they get here.

> [..]
>
> > ip -6 nei show
> > 2001:db80::12 dev wldev lladdr 137.226.18.145 router STALE
>
> Ehmm, that really is not good. a lldaddr should be in fe80::/10 or
> actually more specifically in fe80::/64. What kind of interface is
> 'wldev' and on what kind of hardware?
>

I am working on different testbeds. The one i used was an xen maschine, where 
the ipv6 data is send over on ipv4 gre tunnel. The interface on which the gre 
tunnel is created has an ipv4 address. That is why the linklayeraddress for 
the wldev device is an ipv4 address. I also tried everything on "real" 
hardware (without tunneling ipv6 over ipv4) but there it does not work too.
On the real hardware i get

ip -6 nei show
2001:db8::6 dev ath0 lladdr 00:0b:6b:4e:b9:49 router STALE

>
> > ping6 from A to C -> Destination unreachable: Address unreachable
> > ping6 from C to A -> Destination unreachable: Address unreachable
>
> But the big question is WHAT is sending the "Address Unreachable" ?

B is senden the Destination unreachable.
Here the output from tcpdump:
tcpdump -i wldev -vv ip6
tcpdump: WARNING: arptype 778 not supported by libpcap - falling back to 
cooked socket
tcpdump: listening on wldev, link-type LINUX_SLL (Linux cooked), capture size 
96 bytes
14:04:56.734426 IP6 (hlim 64, next-header: ICMPv6 (58), length: 64) 
2001:db8::11 > 2001:db8::13: ICMP6, echo request, length 64, seq 41
14:04:57.734305 IP6 (hlim 64, next-header: ICMPv6 (58), length: 64) 
2001:db8::11 > 2001:db8::13: ICMP6, echo request, length 64, seq 42
14:04:57.754238 IP6 (hlim 64, next-header: ICMPv6 (58), length: 112) 
2001:db8::12 > 2001:db8::11: ICMP6, destination unreachable, length 112[|
icmp6]
14:04:57.754297 IP6 (hlim 64, next-header: ICMPv6 (58), length: 112) 
2001:db8::12 > 2001:db8::11: ICMP6, destination unreachable, length 112[|
icmp6]
14:04:57.754374 IP6 (hlim 64, next-header: ICMPv6 (58), length: 112) 
2001:db8::12 > 2001:db8::11: ICMP6, destination unreachable, length 112[|
icmp6]
14:04:58.734283 IP6 (hlim 64, next-header: ICMPv6 (58), length: 64) 
2001:db8::11 > 2001:db8::13: ICMP6, echo request, length 64, seq 43
14:04:59.734296 IP6 (hlim 64, next-header: ICMPv6 (58), length: 64) 
2001:db8::11 > 2001:db8::13: ICMP6, echo request, length 64, seq 44
14:05:00.734268 IP6 (hlim 64, next-header: ICMPv6 (58), length: 64) 
2001:db8::11 > 2001:db8::13: ICMP6, echo request, length 64, seq 45
14:05:01.734294 IP6 (hlim 64, next-header: ICMPv6 (58), length: 64) 
2001:db8::11 > 2001:db8::13: ICMP6, echo request, length 64, seq 46
14:05:01.754231 IP6 (hlim 255, next-header: ICMPv6 (58), length: 32) 
2001:db8::12 > ff02::1:ff00:11: [icmp6 sum ok] ICMP6, neighbor solicitation, 
length 32, who has 2001:db8::11
  source link-address option (1), length 8 (1): 89:e2:12:91:00:

Re: [IPv6 Users] ipv6 routing, problem with neighbour solicitation

2007-06-24 Thread Peter Bieringer
Volker Böhme schrieb:

> ip -6 nei show
> 2001:2::1 dev wldev lladdr 137.226.18.145 router REACHABLE

You have IPv4 addresses as linklayer addresses?

> When i run tcpdump on host B (while running ping6 from A to C), i can see, 
> that the ping gets to host B but host B does not request the linklayeraddress 
> from host C. That is what i am thinking is causing the problem. But i do not 
> know why B is not requesting the linklayeraddress. 

> I can get the ping6 from A to C and from C to A work when i manually add the 
> linklayeraddress from A and C into the cache from B with "ip -6 neigh add 
>  lladdr  dev ". 

Which kernel version do you use?

Can you send the output on host B of
# sysctl net.ipv6.conf.eth0

Peter
___
Users mailing list
[email protected]
https://lists.ipv6.org/mailman/listinfo/users


Re: [IPv6 Users] ipv6 routing, problem with neighbour solicitation

2007-06-23 Thread Volker Böhme
On Friday 22 June 2007 15:47:17 Jeroen Massar wrote:
> If you want to use example addresses, use 2001:db8::/32 or generate a ULA
> prefix using RFC4193 (See also http://www.sixxs.net/tools/grh/ula/).

Ok, i will use 2001:db8::/32 then. When everythin is working i will gernerate 
a ULA prefix.

> The big interesting thing you are not providing is the prefix lengths.

I was using /128. In the beneath example i am using 2001:db8::/32. In my other 
mail as answer to Ed Kapitein i used 2001:1::/64, 2001:2::/64 and 2001:3::/64 
as he proposed. The results are the same.

> Providing the output of "ip -6 ro sho" and "ip -6 nei sho" on all the
> boxes is helpful. The other thing you might want to try to do is:
>  - ping6 from A to B
>  - ping6 from B to A
>  - ping6 from B to C
>  - ping6 from C to B
>  - ping6 from A to C
>  - ping6 from C to A
>
> then you can pinpoint where it goes wrong. Running a tcpdump on the links
> involved while doing these tests helps.

Ok here you go:

On host A:
ip -6 addr show
2: lo:  mtu 16436
inet6 ::1/128 scope host
   valid_lft forever preferred_lft forever
35: [EMAIL PROTECTED]:  mtu 1472
inet6 2001:db80::11/32 scope global
   valid_lft forever preferred_lft forever

ip -6 route show
2001:db80::13 via 2001:db80::12 dev wldev  metric 1024  expires -168sec mtu 
1472 advmss 1412 hoplimit 4294967295
2001:db80::/32 dev wldev  metric 256  expires -362sec mtu 1472 advmss 1412 
hoplimit 4294967295
fe80::/64 dev wldev  metric 256  expires -362sec mtu 1472 advmss 1412 hoplimit 
4294967295
ff00::/8 dev wldev  metric 256  expires -362sec mtu 1472 advmss 1412 hoplimit 
4294967295
unreachable default dev lo  proto none  metric -1  error -101 hoplimit 255

ip -6 nei show
2001:db80::12 dev wldev lladdr 137.226.18.145 router STALE

On host B:
ip -6 add show
2: lo:  mtu 16436
inet6 ::1/128 scope host
   valid_lft forever preferred_lft forever
25: [EMAIL PROTECTED]:  mtu 1472
inet6 2001:db80::12/32 scope global
   valid_lft forever preferred_lft forever

ip -6 route show
2001:db80::/32 dev wldev  metric 256  expires -404sec mtu 1472 advmss 1412 
hoplimit 4294967295
fe80::/64 dev wldev  metric 256  expires -404sec mtu 1472 advmss 1412 hoplimit 
4294967295
ff00::/8 dev wldev  metric 256  expires -404sec mtu 1472 advmss 1412 hoplimit 
4294967295
unreachable default dev lo  proto none  metric -1  error -101 hoplimit 255

ip -6 nei show -> shows nothing

On host C:
ip -6 addr show
2: lo:  mtu 16436
inet6 ::1/128 scope host
   valid_lft forever preferred_lft forever
24: [EMAIL PROTECTED]:  mtu 1472
inet6 2001:db80::13/32 scope global
   valid_lft forever preferred_lft forever

ip -6 route show
2001:db80::11 via 2001:db80::12 dev wldev  metric 1024  expires -240sec mtu 
1472 advmss 1412 hoplimit 4294967295
2001:db80::/32 dev wldev  metric 256  expires -440sec mtu 1472 advmss 1412 
hoplimit 4294967295
fe80::/64 dev wldev  metric 256  expires -440sec mtu 1472 advmss 1412 hoplimit 
4294967295
ff00::/8 dev wldev  metric 256  expires -440sec mtu 1472 advmss 1412 hoplimit 
4294967295
unreachable default dev lo  proto none  metric -1  error -101 hoplimit 255

ip -6 nei show
2001:db80::12 dev wldev lladdr 137.226.18.145 router STALE


Here the results of ping6:
ping6 from A to B -> works
ping6 from B to A -> works
ping6 from B to C -> works
ping6 from C to B -> works
ping6 from A to C -> Destination unreachable: Address unreachable
ping6 from C to A -> Destination unreachable: Address unreachable

Like i have also written in another mail, when i run tcpdump on host B (while 
running ping6 from A to C), i can see, that the ping gets to host B but host 
B does not request the linklayeraddress from host C. That is what i am 
thinking is causing the problem. But i do not know why B is not requesting 
the linklayeraddress. 
I can get the ping6 from A to C and from C to A work when i manually add the 
linklayeraddress from A and C into the cache from B with "ip -6 neigh add 
 lladdr  dev ". 

> Also remember that there is ICMP Redirects which can instruct the host
> to still use the link directly instead of going via another.

I have checked that with tcpdum, no ICMP Redirects are send.

> Also don't forget to check to check sysctl's and firewall settings.

Here the firewall setting (the same on host A,B and C:
ip6tables -L
Chain INPUT (policy ACCEPT)
target prot opt source   destination

Chain FORWARD (policy ACCEPT)
target prot opt source   destination

Chain OUTPUT (policy ACCEPT)
target prot opt source   destination

With the sysctl settings, perhaps theres something wrong. I did enable 
forwarding, but do i have to change some other parameters to? 


Greets,
Volker
___
Users mailing list
[email protected]
https://lists.ipv6.org/mailman/listinfo/users


Re: [IPv6 Users] ipv6 routing, problem with neighbour solicitation

2007-06-23 Thread Volker Böhme
Hi,

On Friday 22 June 2007 16:15:07 Ed Kapitein wrote:
> Hi Volker,
>
> I am missing the mask you use for the various networks.
> i think you would get it to work with a setup like this:
> A has Ipv6-Address 2001:1::1
> B has Ipv6-Address 2001:2::1
> C has Ipv6-Address 2001:3::1
> and a /64 mask
>
> I think in your setup al ip addresses are in the same subnet, so routing
> will not be done.

I dont think think the problem is caused by using one subnet. I am working on 
routing in wireless adhoc networks and there it can happen that routing has 
to be done in the same subnet  (not each client can hear each other because 
of the wireless connections). But nevertheless i have tried your 
configuration. Unfortunately my problem still exists. Here some outputs which 
can be usefull:

On host A:
ip -6 addr show
2: lo:  mtu 16436
inet6 ::1/128 scope host
   valid_lft forever preferred_lft forever
34: [EMAIL PROTECTED]:  mtu 1472
inet6 2001:1::1/64 scope global
   valid_lft forever preferred_lft forever

ip -6 route show
2001:1::/64 dev wldev  metric 256  expires -73726sec mtu 1472 advmss 1412 
hoplimit 4294967295
2001:2::/64 dev wldev  metric 1  expires -73515sec mtu 1472 advmss 1412 
hoplimit 4294967295
2001:3::/64 via 2001:2::1 dev wldev  metric 1024  expires -73397sec mtu 1472 
advmss 1412 hoplimit 4294967295
fe80::/64 dev wldev  metric 256  expires -73726sec mtu 1472 advmss 1412 
hoplimit 4294967295
ff00::/8 dev wldev  metric 256  expires -73726sec mtu 1472 advmss 1412 
hoplimit 4294967295
unreachable default dev lo  proto none  metric -1  error -101 hoplimit 255

ip -6 nei show
2001:2::1 dev wldev lladdr 137.226.18.145 router REACHABLE

On host B:
ip -6 add show
2: lo:  mtu 16436
inet6 ::1/128 scope host
   valid_lft forever preferred_lft forever
24: [EMAIL PROTECTED]:  mtu 1472
inet6 2001:2::1/64 scope global
   valid_lft forever preferred_lft forever

ip -6 route show
2001:1::/64 dev wldev  metric 1  expires -73429sec mtu 1472 advmss 1412 
hoplimit 4294967295
2001:2::/64 dev wldev  metric 256  expires -73800sec mtu 1472 advmss 1412 
hoplimit 4294967295
2001:3::/64 dev wldev  metric 1  expires -73424sec mtu 1472 advmss 1412 
hoplimit 4294967295
fe80::/64 dev wldev  metric 256  expires -73800sec mtu 1472 advmss 1412 
hoplimit 4294967295
ff00::/8 dev wldev  metric 256  expires -73800sec mtu 1472 advmss 1412 
hoplimit 4294967295
unreachable default dev lo  proto none  metric -1  error -101 hoplimit 255

ip -6 nei show -> shows nothing

On host C:
ip -6 addr show
2: lo:  mtu 16436
inet6 ::1/128 scope host
   valid_lft forever preferred_lft forever
23: [EMAIL PROTECTED]:  mtu 1472
inet6 2001:3::1/64 scope global
   valid_lft forever preferred_lft forever

ip -6 route show
2001:1::/64 via 2001:2::1 dev wldev  metric 1024  expires -73651sec mtu 1472 
advmss 1412 hoplimit 4294967295
2001:2::/64 dev wldev  metric 1  expires -73701sec mtu 1472 advmss 1412 
hoplimit 4294967295
2001:3::/64 dev wldev  metric 1  expires -73544sec mtu 1472 advmss 1412 
hoplimit 4294967295
fe80::/64 dev wldev  metric 256  expires -73782sec mtu 1472 advmss 1412 
hoplimit 4294967295
ff00::/8 dev wldev  metric 256  expires -73782sec mtu 1472 advmss 1412 
hoplimit 4294967295
unreachable default dev lo  proto none  metric -1  error -101 hoplimit 255

ip -6 nei show
2001:2::1 dev wldev lladdr 137.226.18.145 router REACHABLE

Here the results of some ping6:
ping6 from A to B -> works
ping6 from A to C -> Destination unreachable: Address unreachable
ping6 from B to A -> works
ping6 from B to C -> works
ping6 from C to B -> works
ping6 from C to A -> Destination unreachable: Address unreachable

When i run tcpdump on host B (while running ping6 from A to C), i can see, 
that the ping gets to host B but host B does not request the linklayeraddress 
from host C. That is what i am thinking is causing the problem. But i do not 
know why B is not requesting the linklayeraddress. 
I can get the ping6 from A to C and from C to A work when i manually add the 
linklayeraddress from A and C into the cache from B with "ip -6 neigh add 
 lladdr  dev ". 

Hope you now have all information you need and you have an idea to help me.

Kind regards
Volker

___
Users mailing list
[email protected]
https://lists.ipv6.org/mailman/listinfo/users


Re: [IPv6 Users] ipv6 routing, problem with neighbour solicitation

2007-06-23 Thread Volker Böhme
On Friday 22 June 2007 16:16:42 Peter Bieringer wrote:
> Which prefix lengths do you use? Usually, each link has a /64, don't
> play around with /128 on local networks.

I was using /128. Why shouldn't i use this prefix? 
But like you can see in my other mail in reply to Ed Kapitein i have the same 
problem when i use /64 and different subnets.

> Improve your IPv6 setup by a more realistic configuration with proper
> prefixes and try again (and think about whether you want to use routing
> or bridging).

I definitly want to use routing. In my examples i set my routes by myself (for 
better debugging) but actually the routes are set with use of  olsr-routing. 

> BTW: you should at least use fec0:: addresses for local testing or take
> the prefix from the documentation space (2001:db80::/32) or use a Unique
> Local prefix, if you have no proper global /48 prefix assigned to you.

The addresses i used, were only for testing. I thought perhaps the cause of my 
problem lies in the used type of addresses. I have tried all different 
address types then (ULAN, site-local, global)  but i have the same problem 
with all address types.

 Kind regards
Volker
___
Users mailing list
[email protected]
https://lists.ipv6.org/mailman/listinfo/users


Re: [IPv6 Users] ipv6 routing, problem with neighbour solicitation

2007-06-22 Thread Ed Kapitein
Hi Volker,

I am missing the mask you use for the various networks.
i think you would get it to work with a setup like this:
A has Ipv6-Address 2001:1::1
B has Ipv6-Address 2001:2::1
C has Ipv6-Address 2001:3::1
and a /64 mask

I think in your setup al ip addresses are in the same subnet, so routing
will not be done.

I hope this helps.

Kind regards,
@

On Fri, 2007-06-22 at 15:30 +0200, Volker Böhme wrote:
> Hello,
> 
> I have a problem with routing and ipv6. 
> Consider the following senario: i have three PCs setup in a chain (A - B - 
> C ). All PCs do only have one NIC. 
> 
> A has Ipv6-Address 2001::1
> B has Ipv6-Address 2001::2
> C has Ipv6-Address 2001::3
> 
> The following routes exist on PC A:
> 2001::2 via eth0
> 2001::3 via gateway 2001::2
> 
> On PC B i have the this routes:
> 2001::1 via eth0
> 2001::3 via eth0
> 
> On PC C:
> 2001::2 via eth0
> 2001::1 via gateway 2001::2
> 
> PC A can talk to PC B and PC B can talk to PC C. No direct communikation 
> between PC A and PC C is possible (i am simulating wireless connections). 
> I can send a ping6 from A->B, B->A, B->C, C->B. My problem now is, that i 
> cannot send a ping6 from A->C. Forwarding has been enabled by doing "echo 1 
> > /proc/sys/net/ipv6/conf/all/forwarding". I think i have found out why this 
> is not possible. When PC B gehts the data 
> from PC A it has to look in is neighbour solicitaion table to get the 
> linklayeraddress from PC C. Now we have the case, that there is no entry for 
> PC C and so PC B is sendeing a ICMP6 host unreachable message back to A. This 
> is where the error occurs. In my opinion PC B first have to try to get the 
> linklayeraddress from PC C by sending a neighbour solicitation message but it 
> is no doing that it directly sends back the host unreachable message.  When i 
> add PC A and PC B in the neighbour solicitation list on PC B by hand 
> everything works fine.
> 
> My question now is do i have forget to configure something? Do i have done 
> something wrong? Or can this be a bug in the kernel?
> 
> Kind regards
> Volker
> ___
> Users mailing list
> [email protected]
> https://lists.ipv6.org/mailman/listinfo/users

___
Users mailing list
[email protected]
https://lists.ipv6.org/mailman/listinfo/users


Re: [IPv6 Users] ipv6 routing, problem with neighbour solicitation

2007-06-22 Thread Jun-ichiro itojun Hagino
> I have a problem with routing and ipv6. 
> Consider the following senario: i have three PCs setup in a chain (A - B - 
> C ). All PCs do only have one NIC. 
(snip)

so in your configuration, A, B and C are all on the same link but
A and C has to use B as the gateway.  B would send A/B ICMPv6 redirects.

for ICMPv6 to work correctly, you MUST use link-local addresses
(fe80::blah) for the gateway setting.  this is a hidden rule in
the RFCs.

itojun
___
Users mailing list
[email protected]
https://lists.ipv6.org/mailman/listinfo/users


Re: [IPv6 Users] ipv6 routing, problem with neighbour solicitation

2007-06-22 Thread Peter Bieringer
Hi,

At 22.06.2007 15:30, Volker Böhme wrote:
> Hello,
> 
> I have a problem with routing and ipv6. 
> Consider the following senario: i have three PCs setup in a chain (A - B - 
> C ). All PCs do only have one NIC. 
> 
> A has Ipv6-Address 2001::1
> B has Ipv6-Address 2001::2
> C has Ipv6-Address 2001::3
> 
> The following routes exist on PC A:
> 2001::2 via eth0
> 2001::3 via gateway 2001::2
> 
> On PC B i have the this routes:
> 2001::1 via eth0
> 2001::3 via eth0
> 
> On PC C:
> 2001::2 via eth0
> 2001::1 via gateway 2001::2
> 
> PC A can talk to PC B and PC B can talk to PC C. No direct communikation 
> between PC A and PC C is possible (i am simulating wireless connections). 
> I can send a ping6 from A->B, B->A, B->C, C->B. My problem now is, that i 
> cannot send a ping6 from A->C. Forwarding has been enabled by doing "echo 1 
>> /proc/sys/net/ipv6/conf/all/forwarding". I think i have found out why this 
> is not possible. When PC B gehts the data 
> from PC A it has to look in is neighbour solicitaion table to get the 
> linklayeraddress from PC C. Now we have the case, that there is no entry for 
> PC C and so PC B is sendeing a ICMP6 host unreachable message back to A. This 
> is where the error occurs. In my opinion PC B first have to try to get the 
> linklayeraddress from PC C by sending a neighbour solicitation message but it 
> is no doing that it directly sends back the host unreachable message.  When i 
> add PC A and PC B in the neighbour solicitation list on PC B by hand 
> everything works fine.
> 
> My question now is do i have forget to configure something? Do i have done 
> something wrong? Or can this be a bug in the kernel?

Which prefix lengths do you use? Usually, each link has a /64, don't
play around with /128 on local networks.

Improve your IPv6 setup by a more realistic configuration with proper
prefixes and try again (and think about whether you want to use routing
or bridging).

BTW: you should at least use fec0:: addresses for local testing or take
the prefix from the documentation space (2001:db80::/32) or use a Unique
Local prefix, if you have no proper global /48 prefix assigned to you.

Peter
-- 
Dr. Peter Bieringer http://www.bieringer.de/pb/
GPG/PGP Key 0x958F422D   mailto:[EMAIL PROTECTED]
Deep Space 6 Co-Founder and Core Member  http://www.deepspace6.net/
Xing/OpenBC http://www.xing.com/hp/Peter_Bieringer/
Personal invitation to Xing  http://www.xing.com/go/invita/3889
___
Users mailing list
[email protected]
https://lists.ipv6.org/mailman/listinfo/users