From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Michael S. Tsirkin" Subject: Re: [PATCH v2] net-tun: restructure tun_do_read for better sleep/wakeup efficiency Date: Tue, 20 May 2014 09:35:51 +0300 Message-ID: <20140520063551.GC6653@redhat.com> References: <1400278308-25372-1-git-send-email-xii@google.com> <5379CE8D.2030405@redhat.com> <1400508587.5367.37.camel@edumazet-glaptop2.roam.corp.google.com> <537ADD91.4000609@redhat.com> <1400561559.5367.87.camel@edumazet-glaptop2.roam.corp.google.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Jason Wang , Xi Wang , "David S. Miller" , netdev@vger.kernel.org, Maxim Krasnyansky , Neal Cardwell , Eric Dumazet To: Eric Dumazet Return-path: Received: from mx1.redhat.com ([209.132.183.28]:25183 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750769AbaETGhC (ORCPT ); Tue, 20 May 2014 02:37:02 -0400 Content-Disposition: inline In-Reply-To: <1400561559.5367.87.camel@edumazet-glaptop2.roam.corp.google.com> Sender: netdev-owner@vger.kernel.org List-ID: On Mon, May 19, 2014 at 09:52:39PM -0700, Eric Dumazet wrote: > On Tue, 2014-05-20 at 12:44 +0800, Jason Wang wrote: > > > But this changes (sk_data_ready()) has nothing related to switching to > > use __skb_recv_datagram() > > I think I will give up at this point, this is becoming too much time to > gain only 15% of cpu. > Ouch I did like the patch, less code and more performance. I hope we'll be able to save it still, please be patient Jason's just being cautious trying to make sure we don't break existing userspace. -- MST