From patchwork Mon Nov 23 13:06:45 2015 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: plongepe X-Patchwork-Id: 7681061 Return-Path: X-Original-To: patchwork-linux-pm@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork2.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork2.web.kernel.org (Postfix) with ESMTP id 7A1F0BF90C for ; Mon, 23 Nov 2015 13:06:33 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id A648E20268 for ; Mon, 23 Nov 2015 13:06:32 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id B3B58202FE for ; Mon, 23 Nov 2015 13:06:31 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753909AbbKWNG3 (ORCPT ); Mon, 23 Nov 2015 08:06:29 -0500 Received: from mga09.intel.com ([134.134.136.24]:14848 "EHLO mga09.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753586AbbKWNG3 (ORCPT ); Mon, 23 Nov 2015 08:06:29 -0500 Received: from orsmga002.jf.intel.com ([10.7.209.21]) by orsmga102.jf.intel.com with ESMTP; 23 Nov 2015 05:06:03 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.20,337,1444719600"; d="scan'208";a="857572218" Received: from tllab185.tl.intel.com ([10.102.161.63]) by orsmga002.jf.intel.com with ESMTP; 23 Nov 2015 05:06:02 -0800 From: Philippe Longepe To: linux-pm@vger.kernel.org Cc: srinivas.pandruvada@linux.intel.com, Philippe Longepe , Stephane Gasparini Subject: [PATCH v1 5/5] cpufreq: intel_pstate: try load instead of busy_scaled Date: Mon, 23 Nov 2015 14:06:45 +0100 Message-Id: <1448284006-13596-11-git-send-email-philippe.longepe@linux.intel.com> X-Mailer: git-send-email 1.9.1 In-Reply-To: <1448284006-13596-1-git-send-email-philippe.longepe@linux.intel.com> References: <1448284006-13596-1-git-send-email-philippe.longepe@linux.intel.com> Sender: linux-pm-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-pm@vger.kernel.org X-Spam-Status: No, score=-7.5 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_HI, RP_MATCHES_RCVD, UNPARSEABLE_RELAY autolearn=unavailable version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP From: Philippe Longepe - use the load instead of busy scales - reactivate ioboost The time spent in iowait is converted into cycles at max freq, so it increases the load during IOs. Signed-off-by: Philippe Longepe Signed-off-by: Stephane Gasparini --- drivers/cpufreq/intel_pstate.c | 38 +++++++++++++++----------------------- 1 file changed, 15 insertions(+), 23 deletions(-) diff --git a/drivers/cpufreq/intel_pstate.c b/drivers/cpufreq/intel_pstate.c index 4f08fa7..61b31d4 100644 --- a/drivers/cpufreq/intel_pstate.c +++ b/drivers/cpufreq/intel_pstate.c @@ -114,6 +114,7 @@ struct cpudata { u64 prev_mperf; u64 prev_tsc; struct sample sample; + u64 prev_cummulative_iowait; }; static struct cpudata **all_cpu_data; @@ -931,12 +932,22 @@ static inline void intel_pstate_set_sample_time(struct cpudata *cpu) mod_timer_pinned(&cpu->timer, jiffies + delay); } - static inline int32_t intel_pstate_calc_scaled_busy(struct cpudata *cpu) { + u64 cummulative_iowait, delta_iowait_us; + u64 delta_iowait_mperf; + u64 mperf, now; struct sample *sample = &cpu->sample; struct pstate_data *pstate = &cpu->pstate; - int64_t core_busy_ratio; + + cummulative_iowait = get_cpu_iowait_time_us(cpu->cpu, &now); + /* Convert iowait time into number of IO cycles spent at max_freq */ + delta_iowait_us = cummulative_iowait - cpu->prev_cummulative_iowait; + delta_iowait_mperf = div64_u64(delta_iowait_us * cpu->pstate.scaling * + cpu->pstate.max_pstate, 1000); + + mperf = cpu->sample.mperf + delta_iowait_mperf; + cpu->prev_cummulative_iowait = cummulative_iowait; /* * The load can be estimated as the ratio of the mperf counter @@ -944,30 +955,11 @@ static inline int32_t intel_pstate_calc_scaled_busy(struct cpudata *cpu) * (C0) and the time stamp counter running at the same frequency * also during C-states. */ - sample->cpu_load = div64_u64(100 * sample->mperf, sample->tsc); - - /* - * The target P-state can be estimated with the following formula: - * PercentPerformance = PercentBusy * (delta_aperf/delta_mperf); - * (see Section 14.2 from Intel Software Developer Manual) - * with PercentBusy = 100 * (delta_mperf / delta_tsc) and - * PercentPerformance can be simplified with: - * (delta_mperf * delta_aperf) / (delta_tsc * delta_mperf) = - * delta_aperf / delta_tsc. Finally, we normalize core_busy_ratio, - * which was our actual percent performance to what we requested - * during the last sample period. The result will be a percentage of - * busy at a specified pstate. - */ - core_busy_ratio = div64_u64(int_tofp(100) * sample->aperf * - pstate->max_pstate, sample->tsc * pstate->current_pstate); + sample->cpu_load = div64_u64(100 * mperf, sample->tsc); - sample->freq = div64_u64(sample->aperf * pstate->max_pstate * - pstate->scaling, sample->mperf); - - return core_busy_ratio; + return sample->cpu_load; } - static inline int32_t intel_pstate_get_scaled_busy_estimate(struct cpudata *cpu) { int32_t core_busy, max_pstate, current_pstate, sample_ratio;