From mboxrd@z Thu Jan 1 00:00:00 1970 From: Viresh Kumar Subject: Re: [PATCH] cpufreq, intel_pstate.c, Fix rounding errors Date: Thu, 19 Nov 2015 10:16:39 +0530 Message-ID: <20151119044639.GD3737@ubuntu> References: <1447862144-7188-1-git-send-email-prarit@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mail-pa0-f51.google.com ([209.85.220.51]:35253 "EHLO mail-pa0-f51.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932541AbbKSEqo (ORCPT ); Wed, 18 Nov 2015 23:46:44 -0500 Received: by pacej9 with SMTP id ej9so67853972pac.2 for ; Wed, 18 Nov 2015 20:46:44 -0800 (PST) Content-Disposition: inline In-Reply-To: <1447862144-7188-1-git-send-email-prarit@redhat.com> Sender: linux-pm-owner@vger.kernel.org List-Id: linux-pm@vger.kernel.org To: Prarit Bhargava Cc: linux-kernel@vger.kernel.org, Srinivas Pandruvada , Len Brown , Alexandra Yates , Kristen Carlson Accardi , "Rafael J. Wysocki" , linux-pm@vger.kernel.org On 18-11-15, 10:55, Prarit Bhargava wrote: > diff --git a/drivers/cpufreq/intel_pstate.c b/drivers/cpufreq/intel_pstate.c > index 2e31d09..686f024 100644 > --- a/drivers/cpufreq/intel_pstate.c > +++ b/drivers/cpufreq/intel_pstate.c > @@ -1233,6 +1233,8 @@ static int intel_pstate_set_policy(struct cpufreq_policy *policy) > struct cpudata *cpu; > int i; > #endif > + int max_policy_calc; > + > pr_debug("intel_pstate: %s max %u policy->max %u\n", __func__, > policy->cpuinfo.max_freq, policy->max); > if (!policy->cpuinfo.max_freq) > @@ -1249,7 +1251,10 @@ static int intel_pstate_set_policy(struct cpufreq_policy *policy) > limits = &powersave_limits; > limits->min_policy_pct = (policy->min * 100) / policy->cpuinfo.max_freq; > limits->min_policy_pct = clamp_t(int, limits->min_policy_pct, 0 , 100); > - limits->max_policy_pct = (policy->max * 100) / policy->cpuinfo.max_freq; > + > + max_policy_calc = (policy->max * 1000) / policy->cpuinfo.max_freq; > + limits->max_policy_pct = DIV_ROUND_UP(max_policy_calc, 10); > + Nice catch, but why can't we do this instead: limits->max_policy_pct = DIV_ROUND_UP(policy->max * 100, policy->cpuinfo.max_freq); > limits->max_policy_pct = clamp_t(int, limits->max_policy_pct, 0 , 100); > > /* Normalize user input to [min_policy_pct, max_policy_pct] */ > @@ -1269,6 +1274,7 @@ static int intel_pstate_set_policy(struct cpufreq_policy *policy) > int_tofp(100)); > limits->max_perf = div_fp(int_tofp(limits->max_perf_pct), > int_tofp(100)); > + limits->max_perf = round_up(limits->max_perf, 8); Perhaps you should fix this in a separate patch. -- viresh