netdev - Re: [PATCH v11 nf-next 1/2] netfilter: bridge: Add conntrack double vlan and pppoe

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: <20250408163931.GA11581@breakpoint.cc>
Date: Tue, 8 Apr 2025 18:39:31 +0200
From: Florian Westphal <fw@...len.de>
To: Eric Woudstra <ericwouds@...il.com>
Cc: Pablo Neira Ayuso <pablo@...filter.org>,
	Jozsef Kadlecsik <kadlec@...filter.org>,
	Nikolay Aleksandrov <razor@...ckwall.org>,
	Ido Schimmel <idosch@...dia.com>,
	"David S. Miller" <davem@...emloft.net>,
	Eric Dumazet <edumazet@...gle.com>,
	Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
	Simon Horman <horms@...nel.org>, netfilter-devel@...r.kernel.org,
	bridge@...ts.linux.dev, netdev@...r.kernel.org
Subject: Re: [PATCH v11 nf-next 1/2] netfilter: bridge: Add conntrack double
 vlan and pppoe

Eric Woudstra <ericwouds@...il.com> wrote:
> This adds the capability to conntrack 802.1ad, QinQ, PPPoE and PPPoE-in-Q
> packets that are passing a bridge.

Conntrack is l2 agnostic, so this either requires distinct
ip addresses in the vlans/pppoe tunneled traffic or users
need to configure connection tracking zones manually to
ensure there are no collisions or traffic merges (i.e.,
packet x from PPPoE won't be merged with frag from a vlan).

Actually reading  nf_ct_br_defrag4/6 it seems existing
code already has this bug :/

I currently don't see a fix for this problem.
Can't add L2 addresses to conntrack since those aren't
unique accross vlans/tunnels and they can change anyway
even mid-stream, we can't add ifindexes into the mix
as we'd miss all reply traffic, can't use the vlan tag
since it can be vlan-in-vlan etc.

So likely, we have to live with this.

Maybe refuse to track (i.e. ACCEPT) vlan/8021ad qinq, etc.
traffic if the skb has no template with a zone attached to it?

This would at least push 'address collisions' into the
'incorrect ruleset configuration' domain.