From mboxrd@z Thu Jan 1 00:00:00 1970 From: Hagen Paul Pfeifer Subject: Re: [RFC] netem: de-GSO packets before enqueing Date: Sat, 14 Jan 2012 17:06:46 +0100 Message-ID: <20120114160646.GB2966@hell> References: <1326400930.2740.10.camel@bwh-desktop> <20120112220328.GA10813@gondor.apana.org.au> <20120112165720.4545c670@nehalam.linuxnetplumber.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Herbert Xu , Ben Hutchings , netdev , netfilter-devel To: Stephen Hemminger Return-path: Content-Disposition: inline In-Reply-To: <20120112165720.4545c670@nehalam.linuxnetplumber.net> Sender: netdev-owner@vger.kernel.org List-Id: netfilter-devel.vger.kernel.org * Stephen Hemminger | 2012-01-12 16:57:20 [-0800]: >Probably something like this is needed (untested). > >This issue was discovered when looking at the skb_checksum path for the >netem corruption operation, but it is a general problem. >Network emulation operations like corruption and drop want to operate >on a per-packet (not per-segment) basis. This patch does GSO in software >if necessary to break up packets. Code is similar to logic in xfrm_output. > >Although it appears that the operation is not work conserving, it is okay >because the higher level qdisc operations account for packets by incrementing >by gso_size. Thanks Stephen! Corruption is currently unusable if GSO is enabled. We disable GSO/TSO on all our test machines therefore. I have a larger set of patches for the corruption option for this merge window. We will now test your patch (and the hopefully following qdisc len updated patch). Hagen