From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andi Kleen Subject: Re: [PATCH] Disable TSO for non standard qdiscs Date: Fri, 1 Feb 2008 05:31:13 +0100 Message-ID: <20080201043113.GA10720@one.firstfloor.org> References: <20080131124632.GA25299@basil.nowhere.org> <47A212CB.1060403@hp.com> <20080131190326.GF4671@one.firstfloor.org> <47A214FE.3050200@hp.com> <20080131192521.GG4671@one.firstfloor.org> <47A21E1A.6000006@hp.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Andi Kleen , netdev@vger.kernel.org, davem@davemloft.net To: Rick Jones Return-path: Received: from one.firstfloor.org ([213.235.205.2]:53318 "EHLO one.firstfloor.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756099AbYBAD4s (ORCPT ); Thu, 31 Jan 2008 22:56:48 -0500 Content-Disposition: inline In-Reply-To: <47A21E1A.6000006@hp.com> Sender: netdev-owner@vger.kernel.org List-ID: On Thu, Jan 31, 2008 at 11:14:34AM -0800, Rick Jones wrote: > Sounds like the functionality needs to be in the DSL bridge :) (or the > "router" in the same case) Particularly since it might be getting used > by more than one host on the GbE switch. Possible, but it is not usually in the real world. Setups like WonderShaper which do this on the host side are pretty common. > then the qdisc could/should place a cap on the size of a 'TSO' based on > the bitrate (and perhaps input as to how much time any one "burst" of > data should be allowed to consume on the network) and pass that up the > stack? right now you seem to be proposing what is effectively a cap of > 1 MSS. Hmm, that would probably be possible for TBF, but I'm not sure this can be really done in a useful way for the more complicated qdiscs. Especially since they would likely need to turn on/off GSO regularly when dynamic circumstances change and there is not really a good way to affect a socket after it was created. -Andi