From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [net-next-2.6 PATCH v2] net: implement mechanism for HW based QOS Date: Mon, 20 Dec 2010 23:58:51 +0100 Message-ID: <1292885931.2627.128.camel@edumazet-laptop> References: <20101217165622.26608.24819.stgit@jf-dev1-dcblab> <1292885287.3055.22.camel@bwh-desktop> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: John Fastabend , davem@davemloft.net, netdev@vger.kernel.org, hadi@cyberus.ca, shemminger@vyatta.com, tgraf@infradead.org, nhorman@tuxdriver.com To: Ben Hutchings Return-path: Received: from mail-wy0-f174.google.com ([74.125.82.174]:55859 "EHLO mail-wy0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757821Ab0LTW64 (ORCPT ); Mon, 20 Dec 2010 17:58:56 -0500 Received: by wyb28 with SMTP id 28so3385220wyb.19 for ; Mon, 20 Dec 2010 14:58:55 -0800 (PST) In-Reply-To: <1292885287.3055.22.camel@bwh-desktop> Sender: netdev-owner@vger.kernel.org List-ID: Le lundi 20 d=C3=A9cembre 2010 =C3=A0 22:48 +0000, Ben Hutchings a =C3=A9= crit : > That seems like a fair amount of data to add to every net device, > considering that users may create e.g. a lot of VLAN devices and they > won't use this state at all. Have you considered putting these in a > structure that is accessed indirectly? Yes, this was like this in previous patch, but I asked John to put it i= n netdevice instead, in order to avoid a false sharing and this indirection. This adds 64 bytes on a netdevice. Most machines dont have more than 16 netdevices, it seems a reasonable cost. (We removed the ingress bloat some time ago, so no significant increase anyway)