From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759299Ab3KMRlK (ORCPT ); Wed, 13 Nov 2013 12:41:10 -0500 Received: from mail-pa0-f48.google.com ([209.85.220.48]:38007 "EHLO mail-pa0-f48.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758428Ab3KMRkx (ORCPT ); Wed, 13 Nov 2013 12:40:53 -0500 Message-ID: <5283B9A0.2080502@linaro.org> Date: Wed, 13 Nov 2013 09:40:48 -0800 From: John Stultz User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.1.0 MIME-Version: 1.0 To: Yuanhan Liu CC: Fengguang Wu , Huang Ying , linux-kernel@vger.kernel.org, Ingo Molnar , Peter Zijlstra Subject: Re: performance regressions by "seqcount: Add lockdep functionality to seqcount/seqlock structures" References: <20131113091408.GH22256@yliu-dev.sh.intel.com> In-Reply-To: <20131113091408.GH22256@yliu-dev.sh.intel.com> X-Enigmail-Version: 1.6 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 11/13/2013 01:14 AM, Yuanhan Liu wrote: > Hi, > > FYI, we found some performance regressions caused by commit 1ca7d67c > ("seqcount: Add lockdep functionality to seqcount/seqlock structures") So this is expected. seqlock readers are usually very very cheap operations, and we're now doing lockdep tracking on every iteration around the loop. As the lockdep help states: | If you say Y here, the lock dependency engine will do | | additional runtime checks to debug itself, at the price | | of more runtime overhead. So now since we're also tracking seqlocks in addition to spinlocks, it creates more overhead. Disabling CONFIG_LOCKDEP should restore performance. thanks -john