From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: Extensible hashing and RCU Date: Wed, 21 Feb 2007 05:37:21 -0800 (PST) Message-ID: <20070221.053721.112287244.davem@davemloft.net> References: <200702201209.52388.dada1@cosmosbay.com> <200702211419.30856.dada1@cosmosbay.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: andi@firstfloor.org, johnpol@2ka.mipt.ru, akepner@sgi.com, linux@horizon.com, netdev@vger.kernel.org, bcrl@kvack.org To: dada1@cosmosbay.com Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:36627 "EHLO sunset.davemloft.net" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1161232AbXBUNhW (ORCPT ); Wed, 21 Feb 2007 08:37:22 -0500 In-Reply-To: <200702211419.30856.dada1@cosmosbay.com> Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org From: Eric Dumazet Date: Wed, 21 Feb 2007 14:19:30 +0100 > Now, when the rate of lookups/inserts/delete is high, with totally random > endpoints and cache *always* cold , 'tree structures' are not welcome (not > cache friendly) But what about if tree lookup were free :-) This is why I consider Robert Olsson's trash work the most promising, if we stick sockets into his full flow identified routing cache trie entries, we can eliminate lookup altogether. Just like how he already uses traffic inspection to kill cache entries when FIN shutdown sequence is snooped, we can explicitly flush such entries when socket is closed fully on local system. Really, possibilities are truly endless with that thing.