From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andi Kleen Subject: Re: [PATCH] Disable TSO for non standard qdiscs Date: Sun, 3 Feb 2008 10:35:30 +0100 Message-ID: <20080203093530.GA20908@one.firstfloor.org> References: <20080131185328.GD4671@one.firstfloor.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Andi Kleen , shemminger@linux-foundation.org, kaber@trash.net, netdev@vger.kernel.org To: Herbert Xu Return-path: Received: from one.firstfloor.org ([213.235.205.2]:46983 "EHLO one.firstfloor.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S936331AbYBCJAv (ORCPT ); Sun, 3 Feb 2008 04:00:51 -0500 Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Sun, Feb 03, 2008 at 09:57:10AM +1100, Herbert Xu wrote: > Andi Kleen wrote: > >> Then change TBF to use skb_gso_segment? Be careful, the fact that > > > > That doesn't help because it wants to interleave packets > > from different streams to get everything fair and smooth. The only > > good way to handle that is to split it up and the simplest way to do > > this is to just tell TCP to not do GSO in the first place. > > Actually if we're going to do this I'd prefer you to call skb_gso_segment > instead because that lets us at least bypass netfilter which is one of > the key benefits of software GSO. Ok. The only problem I see right now is that skb_segment() seems to overuse GFP_ATOMIC which would make it unreliable, but that is something that can be fixed. -Andi