netdev - Re: [PATCH net] ip_tunnel: allow to inherit from VLAN encapsulated IP frames

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <331695e3-bfa3-9ea7-3ba9-aebd0689251c@westermo.com>
Date:   Tue, 12 Jul 2022 00:06:16 +0200
From:   Matthias May <matthias.may@...termo.com>
To:     Jakub Kicinski <kuba@...nel.org>
CC:     <netdev@...r.kernel.org>, <davem@...emloft.net>,
        <yoshfuji@...ux-ipv6.org>, <dsahern@...nel.org>,
        <edumazet@...gle.com>, <pabeni@...hat.com>
Subject: Re: [PATCH net] ip_tunnel: allow to inherit from VLAN encapsulated IP
 frames

On 11/07/2022 20:29, Jakub Kicinski wrote:
> On Sat, 9 Jul 2022 22:09:12 +0200 Matthias May wrote:
>>>> How should i go forward with this?
>>>
>>> I think your example above shows that "tos 0xa0" does not work but the
>>> conversation was about inheritance, does "tos inherit" not work either?
>>
>> Yes inherit does not work either. This is why i started setting it statically.
>> However I think I figured out what is going on.
>> Setting the TOS statically to 0xa0 does work... when the payload is IPv4 or IPv6,
>> which is also when inheriting works. For everything other type of payload, it is always 0x00.
>> This is different than with an IPv4 tunnel.
>> Should i consider this a bug that needs to be fixed, or is that the intended behaviour?
> 
> Yes, most likely a bug if you ask me :S

Yeah i treated it as such, and sent a fix with the series.
I wasn't successful yet getting a selftest running.
I hope to get some feedback on the posted series, before i get the selftest running.
Especially this "bugfix", I'm not really sure if i did the right thing.
Essentially i just copied the prepare function for IPv6, stripped everything
out that i thought didn't fit and call it in the "other" path.
 From the manual tests i did, i couldn't find anything that i broke, but I'm not sure of that...

One thing that puzzles me a bit: Is there any reason why the IPv6 version of ip tunnels is so... distributed?
The IPv4 version does everything in a single function in ip_tunnels, while the IPv6 delegates some? of the parsing to
the respective tunnel types, but then does some of the parsing again in ip6_tunnel (e.g the ttl parsing).

BR
Matthias

Download attachment "OpenPGP_signature" of type "application/pgp-signature" (237 bytes)