Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Thu, 14 Nov 2019 13:29:49 -0300
From:   Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
To:     Aaron Conole <aconole@redhat.com>
Cc:     netdev@vger.kernel.org, Pravin B Shelar <pshelar@ovn.org>,
        "David S . Miller" <davem@davemloft.net>,
        Jamal Hadi Salim <jhs@mojatatu.com>,
        Cong Wang <xiyou.wangcong@gmail.com>,
        Jiri Pirko <jiri@resnulli.us>, dev@openvswitch.org,
        linux-kernel@vger.kernel.org, paulb@mellanox.com
Subject: Re: [PATCH net 2/2] act_ct: support asymmetric conntrack
Message-ID: <20191114162949.GB3419@localhost.localdomain>
References: <20191108210714.12426-1-aconole@redhat.com>
 <20191108210714.12426-2-aconole@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20191108210714.12426-2-aconole@redhat.com>
User-Agent: Mutt/1.12.1 (2019-06-15)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Fri, Nov 08, 2019 at 04:07:14PM -0500, Aaron Conole wrote:
> The act_ct TC module shares a common conntrack and NAT infrastructure
> exposed via netfilter.  It's possible that a packet needs both SNAT and
> DNAT manipulation, due to e.g. tuple collision.  Netfilter can support
> this because it runs through the NAT table twice - once on ingress and
> again after egress.  The act_ct action doesn't have such capability.
> 
> Like netfilter hook infrastructure, we should run through NAT twice to
> keep the symmetry.
> 
> Fixes: b57dc7c13ea9 ("net/sched: Introduce action ct")
> 
> Signed-off-by: Aaron Conole <aconole@redhat.com>
> ---
>  net/sched/act_ct.c | 13 ++++++++++++-
>  1 file changed, 12 insertions(+), 1 deletion(-)
> 
> diff --git a/net/sched/act_ct.c b/net/sched/act_ct.c
> index fcc46025e790..f3232a00970f 100644
> --- a/net/sched/act_ct.c
> +++ b/net/sched/act_ct.c
> @@ -329,6 +329,7 @@ static int tcf_ct_act_nat(struct sk_buff *skb,
>  			  bool commit)
>  {
>  #if IS_ENABLED(CONFIG_NF_NAT)
> +	int err;
>  	enum nf_nat_manip_type maniptype;
>  
>  	if (!(ct_action & TCA_CT_ACT_NAT))
> @@ -359,7 +360,17 @@ static int tcf_ct_act_nat(struct sk_buff *skb,
>  		return NF_ACCEPT;
>  	}
>  
> -	return ct_nat_execute(skb, ct, ctinfo, range, maniptype);
> +	err = ct_nat_execute(skb, ct, ctinfo, range, maniptype);
> +	if (err == NF_ACCEPT &&
> +	    ct->status & IPS_SRC_NAT && ct->status & IPS_DST_NAT) {
> +		if (maniptype == NF_NAT_MANIP_SRC)
> +			maniptype = NF_NAT_MANIP_DST;
> +		else
> +			maniptype = NF_NAT_MANIP_SRC;
> +
> +		err = ct_nat_execute(skb, ct, ctinfo, range, maniptype);
> +	}

I keep thinking about this and I'm not entirely convinced that this
shouldn't be simpler. More like:

if (DNAT)
	DNAT
if (SNAT)
	SNAT

So it always does DNAT before SNAT, similarly to what iptables would
do on PRE/POSTROUTING chains.

> +	return err;
>  #else
>  	return NF_ACCEPT;
>  #endif
> -- 
> 2.21.0
>