[Babel-users] babels bug with uninitialized data somewhere?

[Dave Taht]

ok, well, it kind of appears that on the wire is "a221:b7ff:feac:e456"
as part of a larger prefix. I note that I must have at least 5
different babel versions installed on the overall network here, but
shouldn't the code
be seeing and printing the entire 2001 prefix?

root at ida:~/git/tcpdump-babels# ./tcpdump -n -v -r
/home/d/testit/babelwifi.cap.pcapng | grep a221
reading from file /home/d/testit/babelwifi.cap.pcapng, link-type
EN10MB (Ethernet)
    Update/prefix/id 2001:470:8236:261:a221:b7ff:feac:e456/128 metric
96 seqno 25994 interval 160.0s sub-diversity empty
    Update/prefix/id 2001:470:8236:261:a221:b7ff:feac:e456/128 metric
360 seqno 25994 interval 160.0s sub-diversity 11



On Thu, Jun 26, 2014 at 3:10 PM, Dave Taht <dave.taht at gmail.com> wrote:
> lacking tools and clue I put up captures from the ethernet and wifi
> interfaces that took place during a rejecting martian route episode.
>
> the two captures are at: http://snapon.lab.bufferbloat.net/~cero2/babcap/
>
>
> On Thu, Jun 26, 2014 at 3:03 PM, Dave Taht <dave.taht at gmail.com> wrote:
>> On Thu, Jun 26, 2014 at 2:43 PM, Matthieu Boutier
>> <boutier at pps.univ-paris-diderot.fr> wrote:
>>>> Rejecting martian route to ::/0 from ::/128 through
>>>> a221:b7ff:feac:e456:80:9601::.
>>>
>>> This log message comes from update_route, which is call when an update is received.  The martian route is the source address (::/128).
>>>
>>> Are you sure the problem is not in one of your config file, probably with a rule like
>>>
>>>     redistribute ... src-prefix ::/128
>>>
>>> or
>>>
>>>     in ... src-prefix ::/128
>>>
>>> If it's in the latter case, it is more interesting, since probably you would have to write:
>>>
>>>     in src-eq 128 allow
>>>
>>> which allow all (ipv6) routes with a prefix length of 128.
>>
>> There is no conf file.
>>
>> ./babeld -z3 eth0 wlan0 eth6
>>
>>>
>>>> picking up unitialized data somewhere?
>>>
>>> Uninitialized at exactly 128 seems strange.
>>
>> Yes, but the constant change in the last few bits strongly implies something
>> is getting corrupted somewhere.
>>
>> This network has 128 routes installed, and probably 5x that many feasible.
>>
>>> Since it's after when receiving packets, you should be able to see routes with ::/128 source prefix on tcpdump. (?)
>>>
>>>     http://git.wifi.pps.univ-paris-diderot.fr/?p=tcpdump-babels.git;a=summary
>>
>> OK, will look harder and get a dump. Various guesses at the right git
>> clone url for the above failed, though... git clone... what?
>>
>> d at ida:~/git$ git clone
>> http://git.wifi.pps.univ-paris-diderot.fr/projects/tcpdump-babels.git
>> Cloning into 'tcpdump-babels'...
>> fatal: repository
>> 'http://git.wifi.pps.univ-paris-diderot.fr/projects/tcpdump-babels.git/'
>> not found
>> d at ida:~/git$ git clone
>> http://git.wifi.pps.univ-paris-diderot.fr/tcpdump-babels.git
>> Cloning into 'tcpdump-babels'...
>> fatal: repository
>> 'http://git.wifi.pps.univ-paris-diderot.fr/tcpdump-babels.git/' not
>> found
>> d at ida:~/git$ git clone http://git.wifi.pps.univ-paris-diderot.fr/tcpdump-babels
>> Cloning into 'tcpdump-babels'...
>> fatal: repository
>> 'http://git.wifi.pps.univ-paris-diderot.fr/tcpdump-babels/' not found
>> d at ida:~/git$ git clone
>> http://git.wifi.pps.univ-paris-diderot.fr/projects/tcpdump-babels.git
>> Cloning into 'tcpdump-babels'...
>> fatal: repository
>> 'http://git.wifi.pps.univ-paris-diderot.fr/projects/tcpdump-babels.git/'
>> not found
>>
>>
>>> Matthieu
>>>
>>
>>
>>
>> --
>> Dave Täht
>>
>> NSFW: https://w2.eff.org/Censorship/Internet_censorship_bills/russell_0296_indecent.article
>
>
>
> --
> Dave Täht
>
> NSFW: https://w2.eff.org/Censorship/Internet_censorship_bills/russell_0296_indecent.article



-- 
Dave Täht

NSFW: https://w2.eff.org/Censorship/Internet_censorship_bills/russell_0296_indecent.article