From mboxrd@z Thu Jan  1 00:00:00 1970
From: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Subject: Re: [PATCH net-next v3] route: allow to route in a peer netns via
 lwt framework
Date: Wed, 29 Jul 2015 23:17:51 +0200
Message-ID: <55B942FF.2020303@6wind.com>
References: <55B68CE2.10008@6wind.com>	 <1438175774-4408-1-git-send-email-nicolas.dichtel@6wind.com> <1438183251.20182.93.camel@edumazet-glaptop2.roam.corp.google.com>
Reply-To: nicolas.dichtel@6wind.com
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8;
	format=flowed
Content-Transfer-Encoding: QUOTED-PRINTABLE
Cc: davem@davemloft.net, netdev@vger.kernel.org,
	roopa@cumulusnetworks.com, tgraf@suug.ch,
	alexei.starovoitov@gmail.com
To: Eric Dumazet <eric.dumazet@gmail.com>
Return-path: <netdev-owner@vger.kernel.org>
Received: from mail-wi0-f172.google.com ([209.85.212.172]:33139 "EHLO
	mail-wi0-f172.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1753340AbbG2VRz (ORCPT
	<rfc822;netdev@vger.kernel.org>); Wed, 29 Jul 2015 17:17:55 -0400
Received: by wicmv11 with SMTP id mv11so234668287wic.0
        for <netdev@vger.kernel.org>; Wed, 29 Jul 2015 14:17:54 -0700 (PDT)
In-Reply-To: <1438183251.20182.93.camel@edumazet-glaptop2.roam.corp.google.com>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

Le 29/07/2015 17:20, Eric Dumazet a =C3=A9crit :
> On Wed, 2015-07-29 at 15:16 +0200, Nicolas Dichtel wrote:
>> This patch takes advantage of the newly added lwtunnel framework to
>> allow the user to set routes that point to a peer netns.
>>
>> Packets are injected to the peer netns via the loopback device. It w=
orks
>> only when the output device is 'lo'.
>>
>> Example:
>> ip route add 40.1.1.1/32 encap netns nsid 5 via dev lo
>>
>> The goal is to be scalable when the number of netns is high (10k or =
more).
>> Which this patch, we can save two interfaces (veth) per netns, which=
 helps
>> to to reduce memory consumption and the time needed to create a netn=
s.
>
>
> Really this is a hack Nicolas.
>
> get_net_ns_by_id() was not meant to be used in data (fast ???) path.
>
> Same for get_net() and put_net()
>
> Plumbing like that should not happen in lo start_xmit()
Yes, I think you're right. I was a bit too enthusiasm with this new fra=
mework.
Do you think it would be acceptable if the netns was directly reference=
d
instead of a nsid?