From: David Miller Subject: [PATCH 0/7]: Gradual conversion to sk_buff list-head.. Date: Mon, 22 Sep 2008 19:12:30 -0700 (PDT) Message-ID: <20080922.191230.127783294.davem@davemloft.net> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: netdev@vger.kernel.org Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:37169 "EHLO sunset.davemloft.net" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1753702AbYIWCMm (ORCPT ); Mon, 22 Sep 2008 22:12:42 -0400 Received: from localhost (localhost [127.0.0.1]) by sunset.davemloft.net (Postfix) with ESMTP id 64435C8C181 for ; Mon, 22 Sep 2008 19:12:30 -0700 (PDT) Sender: netdev-owner@vger.kernel.org List-ID: As I mentioned in my INFO patch the other day I'm trying to convert sk_buff to use list_head's. To make the transition smoother I've started to hit places one by one that directly reference the sk_buff_head list handling implementation details and make them use only the provided interfaces. I've added new interfaces where necessary to add more parity with what linux/list.h provides. Of note is that there are now list splice interfaces, as well as a "light weight" sk_buff_head initializer for on-stack local list cases. There is more to do, but once it's all handled it should be a matter of only editing linux/skbuff.h and net/core/skbuff.c to perform the final conversion. The trickiest part so far is the PPP MP fragmentation reassembly code. I converted ISDN's copy, and am working on ppp_generic.c's at the moment in my local tree. Really we shouldn't have two copies (their implementations are even slightly different wrt. handling unfinished frames and sequence gaps). But such consolidation is a task for later. All of this is pushed to net-next-2.6, so if you encounter some errors let me know and I'll get them fixed up as fast as possible.