From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [PATCH 1/2] net: reimplement completion_queue as a FIFO queue Date: Mon, 26 Apr 2010 09:21:49 +0200 Message-ID: <1272266509.2346.2.camel@edumazet-laptop> References: <1272258841.2069.968.camel@edumazet-laptop> <1272262154.2069.1032.camel@edumazet-laptop> <20100425.232758.101478192.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: xiaosuo@gmail.com, netdev@vger.kernel.org To: David Miller Return-path: Received: from mail-bw0-f219.google.com ([209.85.218.219]:47893 "EHLO mail-bw0-f219.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752704Ab0DZHVz (ORCPT ); Mon, 26 Apr 2010 03:21:55 -0400 Received: by bwz19 with SMTP id 19so179449bwz.21 for ; Mon, 26 Apr 2010 00:21:53 -0700 (PDT) In-Reply-To: <20100425.232758.101478192.davem@davemloft.net> Sender: netdev-owner@vger.kernel.org List-ID: Le dimanche 25 avril 2010 =C3=A0 23:27 -0700, David Miller a =C3=A9crit= : > No matter what is faster, we have to process packets in > FIFO order, otherwise we get reordering within a flow > which is to be absolutely avoided. Hmm, completion queue is about freeing skb, and its name is misleading. This queue is feed by dev_kfree_skb_irq() only