From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [Patch net v2] mlx5: fixup checksum for short ethernet frame padding Date: Wed, 28 Nov 2018 16:14:43 -0800 Message-ID: References: <20181128061013.3885-1-xiyou.wangcong@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Cc: Linux Kernel Network Developers , Saeed Mahameed To: Cong Wang , Eric Dumazet Return-path: Received: from mail-pf1-f176.google.com ([209.85.210.176]:37723 "EHLO mail-pf1-f176.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726328AbeK2LSN (ORCPT ); Thu, 29 Nov 2018 06:18:13 -0500 Received: by mail-pf1-f176.google.com with SMTP id y126so75863pfb.4 for ; Wed, 28 Nov 2018 16:14:49 -0800 (PST) In-Reply-To: Content-Language: en-US Sender: netdev-owner@vger.kernel.org List-ID: On 11/28/2018 04:05 PM, Cong Wang wrote: > While we are on this page, mlx5e_lro_update_hdr() incorrectly assumes > TCP header is located right after struct iphdr, which is wrong if we could > have IP options on this path. > > It could the hardware which already verified this corner case though. > GRO makes sure IPv4 header has no options, so I would not be surprised that LRO on mlx5 has the same logic.