From mboxrd@z Thu Jan 1 00:00:00 1970 From: Florian Westphal Subject: Re: [PATCH v3] netfilter: conntrack: remove timer from ecache extension Date: Wed, 25 Jun 2014 19:13:27 +0200 Message-ID: <20140625171327.GE6328@breakpoint.cc> References: <1402434776-23461-1-git-send-email-fw@strlen.de> <20140611085547.GA4131@localhost> <20140611092015.GA12405@breakpoint.cc> <20140625170636.GA16571@localhost> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Florian Westphal , netfilter-devel@vger.kernel.org To: Pablo Neira Ayuso Return-path: Received: from Chamillionaire.breakpoint.cc ([80.244.247.6]:57443 "EHLO Chamillionaire.breakpoint.cc" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S964834AbaFYRN2 (ORCPT ); Wed, 25 Jun 2014 13:13:28 -0400 Content-Disposition: inline In-Reply-To: <20140625170636.GA16571@localhost> Sender: netfilter-devel-owner@vger.kernel.org List-ID: Pablo Neira Ayuso wrote: > This patch looks good to me. I made some experiments again in my > testbed. I see no packet drops and the result of 'conntrack -L dying | > wc -l' per second shows small peaks of entries waiting in the dying > list to be delivered which is followed by empty lists by conntrackd at > some point. Thanks a lot for all the testing Pablo. > However, perhaps the per-second watch is not enough to get > a real evolution on how the dying list population evolves along time. Yes, you might be right. IFF you think its useful, we could add a counter to keep track of the maximum length of entries that the ecache worker has seen and export that via ctnetlink (like we do for number of conntracks, search list restarts, etc etc). > Anyway, I think this is good enough to get rid of the extra timer. We > can enqueue this patch to 3.17, unless you have any objection. No objections from me, thanks Pablo!