From: David Miller Subject: Re: [PATCH 1/2] myri10ge: allow LRO to be enabled via ethtool Date: Tue, 19 May 2009 15:26:08 -0700 (PDT) Message-ID: <20090519.152608.149436309.davem@davemloft.net> References: <4A131364.5050402@myri.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org To: brice@myri.com Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:49865 "EHLO sunset.davemloft.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752919AbZESW0K (ORCPT ); Tue, 19 May 2009 18:26:10 -0400 In-Reply-To: <4A131364.5050402@myri.com> Sender: netdev-owner@vger.kernel.org List-ID: From: Brice Goglin Date: Tue, 19 May 2009 22:15:32 +0200 > Allow myri10ge LRO to be enabled/disabled via ethtool > (and by the stack for packet forwarding). > > Signed-off-by: Brice Goglin So I applied these two patches for now to net-next-2.6 but in the long term I think it's probably better to convert this driver to GRO as we're trying to get rid of LRO and GRO handles forwarding and all of that stuff transparently. And I also still want the myri10ge_lro module option removed regardless of whether you do a GRO conversion or not. If LRO/GRO causes a performance regression in some situation, we should fix it instead of papering over this junk with module options. Thanks.