From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752932AbdLSWLs (ORCPT ); Tue, 19 Dec 2017 17:11:48 -0500 Received: from mga06.intel.com ([134.134.136.31]:14310 "EHLO mga06.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750918AbdLSWLr (ORCPT ); Tue, 19 Dec 2017 17:11:47 -0500 X-Amp-Result: UNSCANNABLE X-Amp-File-Uploaded: False X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.45,428,1508828400"; d="scan'208";a="3195284" Date: Tue, 19 Dec 2017 14:11:39 -0800 From: Andi Kleen To: Peter Zijlstra Cc: "Liang, Kan" , mingo@redhat.com, acme@kernel.org, linux-kernel@vger.kernel.org, tglx@linutronix.de, jolsa@redhat.com, eranian@google.com Subject: Re: [PATCH 2/4] perf/x86/intel: fix event update for auto-reload Message-ID: <20171219221139.GF25156@tassilo.jf.intel.com> References: <1513596891-12362-1-git-send-email-kan.liang@linux.intel.com> <1513596891-12362-3-git-send-email-kan.liang@linux.intel.com> <20171219185810.d4yqz6qwqmx3f5o5@hirez.programming.kicks-ass.net> <20171219220709.sq7kwvg7l2ojltvr@hirez.programming.kicks-ass.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20171219220709.sq7kwvg7l2ojltvr@hirez.programming.kicks-ass.net> User-Agent: Mutt/1.9.1 (2017-09-22) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Dec 19, 2017 at 11:07:09PM +0100, Peter Zijlstra wrote: > On Tue, Dec 19, 2017 at 03:08:58PM -0500, Liang, Kan wrote: > > > This all looks very wrong... In auto reload we should never call > > > intel_pmu_save_and_restore() in the first place I think. > > > > > > Things like x86_perf_event_update() and x86_perf_event_set_period() > > > simply _cannot_ do the right thing when we auto reload the counter. > > > > > > > I think it should be OK to call it in first place. > > For x86_perf_event_update(), the reload_times will tell if it's auto reload. > > Both period_left and event->count are carefully recalculated for auto > > reload. > > How does prev_count make sense when we've already reloaded a bunch of > times? We can figure it out how often there was a reload based on the PEBS index. -Andi