From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: Kernel memory leak in bnx2x driver with vxlan tunnel Date: Wed, 20 Jan 2016 16:34:55 -0800 Message-ID: <1453336495.1223.341.camel@edumazet-glaptop2.roam.corp.google.com> References: <5697D833.5010506@hpe.com> <1453243654.1223.290.camel@edumazet-glaptop2.roam.corp.google.com> <1453247464.1223.297.camel@edumazet-glaptop2.roam.corp.google.com> <1453249067.1223.300.camel@edumazet-glaptop2.roam.corp.google.com> <20160120013128.GA6326@pox.localdomain> <56A01BBF.1010301@hpe.com> <1453334825.1223.332.camel@edumazet-glaptop2.roam.corp.google.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: John , Thomas Graf , Linux Kernel Network Developers , Tom Herbert , david.roth@hpe.com, Pravin B Shelar To: Jesse Gross Return-path: Received: from mail-pf0-f177.google.com ([209.85.192.177]:35009 "EHLO mail-pf0-f177.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752101AbcAUAe5 (ORCPT ); Wed, 20 Jan 2016 19:34:57 -0500 Received: by mail-pf0-f177.google.com with SMTP id 65so13217752pff.2 for ; Wed, 20 Jan 2016 16:34:57 -0800 (PST) In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Wed, 2016-01-20 at 16:19 -0800, Jesse Gross wrote: > I have a patch that implements the comparison between dsts. For > packets without a dst, there isn't really a cost and if we do have a > dst then GRO is still a benefit. So it seems like it is worth doing, > even if it is more expensive than is ideal. You guys really want to kill GRO performance. Really the aggregation should happen at the first layer (ethernet device), instead of doing it after tunnel decap. This is already done for GRE, IPIP, SIT, ... GRO having to access metadata looks wrong, if you think about trying to do the same function in hardware (offload)