Hi again!,
More tests!! Different machines as the previous tests (just in case),
again from scratch (althought with the "release" version of hipl).
It doesn't get frozen/stopped, but also fails to connect.
Client-gai says:
nx9420 ~ # conntest-client-gai theta tcp 1111
name='theta' service='1111'
flags: 800
not IPv4 or IPv6 address, resolve name (!AI_NUMERICHOST)
HIP_TRANSPARENT_API: fetch HIT addresses
HIP_TRANSPARENT_API: AI_HIP set: do not get IPv6 addresses
Dumping the structure
AF_INET6 in6_addr=0x20 01 00 7a 33 80 66 0f e9 ae 1f e1 71 11 ed d7
AF_INET6 in6_addr=0x3f fe 00 00 00 00 00 00 00 00 00 00 00 00 00 02
HIT: 2001:007a:3380:660f:e9ae:1fe1:7111:edd7
IP: 3ffe:0000:0000:0000:0000:0000:0000:0002
connect failed
HIP: AI_HIP set: remove IP addresses
Dumping the structure after removing IP addreses
AF_INET6 in6_addr=0x20 01 00 7a 33 80 66 0f e9 ae 1f e1 71 11 ed d7
Input some text, press enter and ctrl+d
hi
!!!! conntest.c Connecting...
ai_flags 2048
ai_family 10
ai_socktype 1
ai_protocol6
ai_addrlen 28
ai_canonname (null)
Trying to connect to 2001:7a:3380:660f:e9ae:1fe1:7111:edd7
After call conntest.c: connect to 2001:7a:3380:660f:e9ae:1fe1:7111:edd7
connect ret=-1 errno=101
trying next
failed to connect
Where it says a first "connect failed" and then "Remove IP address"... Why
does it fail to connect? where is it trying to connect to? I don't see a
single packet with tcpdump (in general, nor just about esp or proto 253).
Then it seems it's using the HIT as a "real IP address" to connect to :S
Obviusly, it fails.
The log from hipd is in the attached file. There are some strange errors
like:
error(nlink.c:160@netlink_talk): Cannot talk to rtnetlink Message too long
error(beet.c:186@hip_xfrm_policy_delete): No associated policies to be
deleted
error(nlink.c:160@netlink_talk): Cannot talk to rtnetlink Message too long
error(beet.c:186@hip_xfrm_policy_delete): No associated policies to be
deleted
What is netlink exactly for? And that "rtnetlink"? Why the mesage is too
long? Which message? Is this normal? :S
Mmmm, and this? More errors:
error(nlink.c:1225@rtnl_dump_filter): OVERRUN Bad address
Another strang thing is that, even if I began from scratch and I did the
same steps from the same sources in both machines, it seems that one of them
is "doing more things" than the other one, at least hipd offers more debug
output than the other... ??? How is this possible? :S
Thanks a lot again.
Hope you can reproduce it since I change the machines and I got the same
errors than before, just following the steps in your "manual".
Regards,
--
Jesús Rojo Martínez.
Human Resources responsible BEST Stockholm - Kungliga Tekniska Högskolan
BEST - Board of European Students of Technology (www.BEST.eu.org)
e-mail: jrojomartinez@xxxxxxxxx
phone: +46704369273
MSN: jrojomartinez@xxxxxxx
