From mboxrd@z Thu Jan 1 00:00:00 1970 From: Benjamin Thery Subject: Re: [PATCH 0/8] netns: make IPv6 multicast forwarding per-namespace Date: Tue, 09 Dec 2008 17:56:03 +0100 Message-ID: <493EA323.5050806@bull.net> References: <20081208160156.343758157@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: YOSHIFUJI Hideaki , netdev , Alexey Dobriyan , Daniel Lezcano To: Dave Miller Return-path: Received: from ecfrec.frec.bull.fr ([129.183.4.8]:43913 "EHLO ecfrec.frec.bull.fr" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753398AbYLIQ4G (ORCPT ); Tue, 9 Dec 2008 11:56:06 -0500 In-Reply-To: <20081208160156.343758157@localhost.localdomain> Sender: netdev-owner@vger.kernel.org List-ID: Dave, While doing some more tests on the patchset I've realized that the queue of unresolved entries (mfc_unres_queue) should also be allocated per-netns (or at least counter cache_resolve_queue_len should be). With the current patchset, the cache for unresolved entries is limited to 10 entries for all namespaces. D'oh. I'm looking at this issue and will resend the patchset with the missing patch ASAP. Benjamin Benjamin Thery wrote: > This patchset introduces the support for network namespaces in IPv6 > multicast forwarding code (net/ipv6/ip6mr.c). > > The first patches in the series moves global data from ip6mr.c into > struct netns_ipv6 to prepare netns support. Data are still referenced in > init_net only. One of these patches makes the related /proc entries > per-namespace. > > The last patch does the main job and enables the network namespace support > by replacing all the init_net references with the proper net retrieved > from sockets or net devices. > > The patchset applies on top of net-next-2.6 > > Regards, > Benjamin > > > -- B e n j a m i n T h e r y - BULL/DT/Open Software R&D http://www.bull.com