From mboxrd@z Thu Jan 1 00:00:00 1970 From: Neil Horman Subject: Re: [PATCH] pktgen: Clone skb to avoid corruption of skbs in ndo_start_xmit methods Date: Thu, 21 Jul 2011 19:50:49 -0400 Message-ID: <20110721235049.GA29489@hmsreliant.think-freely.org> References: <20110720151827.GD12349@hmsreliant.think-freely.org> <20110721.150107.25773992475689131.davem@davemloft.net> <4E28A4C8.8040707@candelatech.com> <20110721.151903.297506479006061401.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: greearb@candelatech.com, eric.dumazet@gmail.com, jpirko@redhat.com, netdev@vger.kernel.org, adobriyan@gmail.com, robert.olsson@its.uu.se To: David Miller Return-path: Received: from charlotte.tuxdriver.com ([70.61.120.58]:38198 "EHLO smtp.tuxdriver.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752716Ab1GUXvD (ORCPT ); Thu, 21 Jul 2011 19:51:03 -0400 Content-Disposition: inline In-Reply-To: <20110721.151903.297506479006061401.davem@davemloft.net> Sender: netdev-owner@vger.kernel.org List-ID: On Thu, Jul 21, 2011 at 03:19:03PM -0700, David Miller wrote: > From: Ben Greear > Date: Thu, 21 Jul 2011 15:14:32 -0700 >=20 > > On 07/21/2011 03:01 PM, David Miller wrote: > >> From: Neil Horman > >> Date: Wed, 20 Jul 2011 11:18:27 -0400 > >> > >>> On Wed, Jul 20, 2011 at 06:24:15AM +0200, Eric Dumazet wrote: > >>>> Le mardi 19 juillet 2011 =E0 22:07 -0400, Neil Horman a =E9crit = : > >>>>>> > >>>>> I think this is a good idea. It lets pktgen dynamically make t= he > >>>>> clone/share > >>>>> decision dynamically and only impacts performance for those sys= tems. > >>>>> > >>>> > >>>> Just let pktgen refuse to use clone_skb command for these device= s. > >>>> > >>> copy that, This is by no means final, but what do you think of th= is? > >>> If its > >>> agreeable to you, Ben, et al. I can add this to my local tree and > >>> start auditing > >>> all the drivers that may need to have the flag set. > >> > >> I think there is a much simpler solution. > >> > >> Set a flag in the SKB when pktgen does SKB sharing. > >> > >> In dev_queue_xmit() (or perhaps, dev_hard_start_xmit()), check the > >> flag > >> and if it's set then we copy the SKB. > >> > >> If this works, then we fix the crash and no driver changes are > >> necessary both now and in the future. > >=20 > > Doesn't that make clone-skb in pktgen much less efficient > > in all cases? >=20 > No, the copy only happens if we enter dev_queue_xmit() which pktgen > doesn't do, it calls the driver's ->ndo_start_xmit() method directly. >=20 > That's the whole idea. Only these encapsulating software devices > will trigger the condition. >=20 I'm happy to go down this route Dave, and agree, its a more solid solut= ion, but I think the problem with it (which Ben may have been alluding to previo= usly) is that pktgen doesn't use dev_queue_xmit or dev_hard_start_xmit to send f= rames. It mimics the locking of dev_hard_start_xmit (but ignores the other che= cks that function does), and just calls the ndo_start_xmit routine of the driver directly. So theres no common code that an skb traverses from pktgen t= o a given driver where we can check such a flag Again, I'm happy to change that so that pktgen uses dev_hard_start_xmit= , but I wonder if thats going to get the same sort of pushback about performanc= e that my origional patch did. Eric, et al., thoughts? Regards Neil