From: Daniel Borkmann <daniel@iogearbox.net>
To: Pablo Neira Ayuso <pablo@netfilter.org>
Cc: daniel@zonque.org, fw@strlen.de, a.perevalov@samsung.com,
netfilter-devel@vger.kernel.org
Subject: Re: [PATCH nf-next 2/2] netfilter: x_tables: fix cgroup's NF_INET_LOCAL_IN sk lookups
Date: Wed, 25 Mar 2015 22:34:50 +0100 [thread overview]
Message-ID: <551329FA.4030002@iogearbox.net> (raw)
In-Reply-To: <20150325202659.GA27374@salvia>
On 03/25/2015 09:26 PM, Pablo Neira Ayuso wrote:
> On Tue, Mar 24, 2015 at 04:30:29PM +0100, Daniel Borkmann wrote:
>> While originally only being intended for outgoing traffic, commit
>> a00e76349f35 ("netfilter: x_tables: allow to use cgroup match for
>> LOCAL_IN nf hooks") enabled xt_cgroups for the NF_INET_LOCAL_IN hook
>> as well, in order to allow for nfacct accounting.
>>
>> This basically was under the assumption that socket early demux will
>> resolve it. It's correct that demux happens after PRE_ROUTING, but
>> before LOCAL_IN.
>>
>> However, that as-is only partially works, i.e. it works for the case
>> of established TCP and connected UDP sockets when early demux is
>> enabled, but not for various other ingress scenarios e.g. unconnected
>> UDP, request sockets, etc.
>>
>> Instead of reverting commit a00e76349f35, I think it's worth to fix
>> it up as there are applications requiring xt_cgroup to match on
>> ingress and egress side. In order to do so, we need to perform a
>> full lookup on skb->sk (ingress) miss, similarly as being done in
>> xt_socket.
>>
>> Therefore, we need to make use of shared helpers xt_sk_lookup() and
>> xt_sk_lookup6(). Thanks to Daniel for the report and also additional
>> testing.
>
> So this is basically needed when early demux is disabled?
>
> This is a rather large rework, I would like to know what scenarios
> we're not currently catching with the existing code.
Hm, perhaps Daniel can elaborate better, what I have seen in my
testing when xt_cgroup fails to match the cgroup on ingress traffic
is i) early demux sysctl disabled, ii) udp on unconnected sockets
(which I understand is the majority of udp traffic), iii) tcp and
udp (any kind) on localhost communications. Daniel's original report
can be found here [1].
Thanks,
Daniel
[1] http://thread.gmane.org/gmane.linux.network/355527
next prev parent reply other threads:[~2015-03-25 21:34 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-03-24 15:30 [PATCH nf-next 0/2] xt_cgroups fix Daniel Borkmann
2015-03-24 15:30 ` [PATCH nf-next 1/2] netfilter: x_tables: refactor lookup helpers from xt_socket Daniel Borkmann
2015-03-24 15:30 ` [PATCH nf-next 2/2] netfilter: x_tables: fix cgroup's NF_INET_LOCAL_IN sk lookups Daniel Borkmann
2015-03-25 16:03 ` Pablo Neira Ayuso
2015-03-25 16:39 ` Daniel Borkmann
2015-03-25 17:17 ` Pablo Neira Ayuso
2015-03-25 17:27 ` Daniel Borkmann
2015-03-25 20:26 ` Pablo Neira Ayuso
2015-03-25 21:34 ` Daniel Borkmann [this message]
2015-03-25 21:54 ` Daniel Mack
2015-03-24 15:42 ` [PATCH nf-next 0/2] xt_cgroups fix Florian Westphal
2015-03-24 15:58 ` Daniel Borkmann
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=551329FA.4030002@iogearbox.net \
--to=daniel@iogearbox.net \
--cc=a.perevalov@samsung.com \
--cc=daniel@zonque.org \
--cc=fw@strlen.de \
--cc=netfilter-devel@vger.kernel.org \
--cc=pablo@netfilter.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).