From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Morton Subject: Re: [PATCH] Execute tasklets in the same order they were queued Date: Mon, 11 Feb 2008 14:49:27 -0800 Message-ID: <20080211144927.55920e56.akpm@linux-foundation.org> References: <20080211222813.GA14316@lixom.net> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: linux-kernel@vger.kernel.org, netdev@vger.kernel.org To: Olof Johansson Return-path: Received: from smtp2.linux-foundation.org ([207.189.120.14]:40033 "EHLO smtp2.linux-foundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753330AbYBKWu1 (ORCPT ); Mon, 11 Feb 2008 17:50:27 -0500 In-Reply-To: <20080211222813.GA14316@lixom.net> Sender: netdev-owner@vger.kernel.org List-ID: On Mon, 11 Feb 2008 16:28:13 -0600 Olof Johansson wrote: > I noticed this when looking at an openswan issue. Openswan (ab?)uses > the tasklet API to defer processing of packets in some situations, > with one packet per tasklet_action(). I started noticing sequences of > reverse-ordered sequence numbers coming over the wire, since new tasklets > are always queued at the head of the list but processed sequentially. > > Convert it to instead append new entries to the tail of the list. As an > extra bonus, the splicing code in takeover_tasklets() no longer has to > iterate over the list. > hm, I'd have thought that this would already have caused problems in networking. And perhaps this change might have effects on networking too? Probably it won't have _much_ effect on networking because networking probably isn't queueing one tasklet per packet(!) but perhaps with bonded channels or something like that?