From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eliezer Tamir Subject: Re: [PATCH v3 net-next 3/4] ixgbe: Add support for ndo_ll_poll Date: Tue, 21 May 2013 13:27:25 +0300 Message-ID: <519B4C0D.3040602@linux.intel.com> References: <1369120003.25971.2.camel@lb-tlvb-eilong.il.broadcom.com> <20130521.001444.1361294042663568537.davem@davemloft.net> <20130521.013912.96249995742935192.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: David Miller , eilong@broadcom.com, linux-kernel@vger.kernel.org, netdev@vger.kernel.org, jesse.brandeburg@intel.com, donald.c.skidmore@intel.com, e1000-devel@lists.sourceforge.net, willemb@google.com, andi@firstfloor.org, hpa@zytor.com, eliezer@tamir.org.il To: Or Gerlitz Return-path: In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On 21/05/2013 11:43, Or Gerlitz wrote: > On Tue, May 21, 2013 at 11:39 AM, David Miller wrote: > >> It's one of those "implementation details", I hate it too. > > Maybe if we bake it on this list little further we can see how to get > away from that, or what's the most non ugly way for that? I'm all for proper review and fixing any issues before forcing "ugliness" and "black magic" on unsuspecting users. Having said that, you failed to mention that your company sells userspace stack replacements. Informal testing I did convinces me that for a given HW, the latencies you get with this patchset and with userspace busy-polling are about the same. (This is my personal opinion, I'm not authorized to talk on behalf of my company or anyone else.) Why don't you try it out and tell us what you find.