[nos-bbs] Presence of IPV6 on tun interface & random JNOS crashes ?

[Jay]

Greetings Maiko,

On Wed, 1 Feb 2023, Maiko (Personal) wrote:

> There is a small chance that if you have IPV6 enabled, JNOS will
> receive IPV6 packets over the tun interface, even if you have no
> desire to run IPV6, I noticed in my trace when I enabled IPV6 on
> my linux host system. Suddenly started seeing odd warnings, ie :
>
> Tue Jan 31 19:15:52 2023 - tun1 recv:
> IP: bad header
> 0000  60 09 0f 93 00 28 06 40 fd 00 00 04 00 00 00 00  `....(.@}.......
> 0010  00 00 00 00 00 00 00 01 fd 00 00 04 00 00 00 00  ........}.......
> 0020  00 00 00 00 00 00 00 02 97 22 00 17 d5 a9 df 4a  ........."..U)_J
> 0030  00 00 00 00 a0 02 fd 20 b9 38 00 00 02 04 05 a0  .... .} 98.....
> 0040  04 02 08 0a 0c 59 3f 27 00 00 00 00 01 03 03 07  .....Y?'........
>
> The reason I suggest it 'might' be contributing to random crashes is
> that the IP version is only checked after IP header extraction, which
> means it had to parse data it was not meant to receive to find out the
> type of data. Chances are probably it's not a big deal, but perhaps it
> is something to rule out for those having random crash issues.
>
> IF you can disable ipv6 on the tun interface between linux host and JNOS
> itself, then see if that helps. I'll leave it up to you to figure out.


    There is a 'reserved' field in the IP Header that will contain a binary 
value of 4 (0100) if the frame is in IPv4 format, or a 6 (0110) 
if the frame is in IPv6 format.

    You can test this field and decide whether to drop any further 
processing of the frame's contents, or if it is an IPv4 frame and further 
processing will be necessary.

    We might want an "IPv6 Yes/No" flag option to be set in the 'header' 
file at compile time.  Some day in the future IPv6 *may* become 
universally recognized, but for now I think IPv6 won't see much use in the 
AMPRnet.

    Enjoy!
       --- Jay Nugent  WB8TKL
           Ypsilanti, MI