From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alexander Duyck Subject: Re: [net-next 1/8] e1000e: Minimum packet size must be 17 bytes Date: Tue, 23 Oct 2012 09:37:08 -0700 Message-ID: <5086C7B4.6060202@intel.com> References: <1350987887-16161-1-git-send-email-jeffrey.t.kirsher@intel.com> <1350987887-16161-2-git-send-email-jeffrey.t.kirsher@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Cc: Jeff Kirsher , davem@davemloft.net, Tushar Dave , netdev@vger.kernel.org, gospo@redhat.com, sassmann@redhat.com To: David Laight Return-path: Received: from mga01.intel.com ([192.55.52.88]:21482 "EHLO mga01.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757305Ab2JWRHS (ORCPT ); Tue, 23 Oct 2012 13:07:18 -0400 In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On 10/23/2012 08:25 AM, David Laight wrote: >> This is a HW requirement. Although a buffer as short as 1 byte is allowed, >> the total length of packet before, padding and CRC insertion, must be at >> least 17 bytes. So pad all small packets manually up to 17 bytes before >> delivering them to HW. > Where do such very short packets come from? > The shortest one I know of have: > 6 bytes dest-mac > 6 bytes src-mac > 2 bytes length > 3 bytes llc header (eg reflect request). > 17 bytes total. > > David > This mostly has to do with security fuzz testers sending raw packets at the interface trying to see what can cause it to hang. As far as I know nothing from the stack will be that small, but we still have to support raw packets that size. Thanks, Alex