From mboxrd@z Thu Jan 1 00:00:00 1970 From: Srinivas Pandruvada Subject: Re: [PATCH] cpufreq: intel_pstate: Set P-state upfront in performance mode Date: Fri, 21 Oct 2016 08:08:39 -0700 Message-ID: <1477062519.31557.2.camel@linux.intel.com> References: <1724762.8Kf4o5e5cm@vostro.rjw.lan> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 8bit Return-path: Received: from mga09.intel.com ([134.134.136.24]:19554 "EHLO mga09.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S934759AbcJUPLj (ORCPT ); Fri, 21 Oct 2016 11:11:39 -0400 In-Reply-To: <1724762.8Kf4o5e5cm@vostro.rjw.lan> Sender: linux-pm-owner@vger.kernel.org List-Id: linux-pm@vger.kernel.org To: "Rafael J. Wysocki" , Linux PM list Cc: Linux Kernel Mailing List On Wed, 2016-10-19 at 02:57 +0200, Rafael J. Wysocki wrote: > From: Rafael J. Wysocki > > After commit a4675fbc4a7a (cpufreq: intel_pstate: Replace timers with > utilization update callbacks) the cpufreq governor callbacks may not > be invoked on NOHZ_FULL CPUs and, in particular, switching to the > "performance" policy via sysfs may not have any effect on them.  That > is a problem, because it usually is desirable to squeeze the last > bit of performance out of those CPUs, so work around it by setting > the maximum P-state (within the limits) in intel_pstate_set_policy() > upfront when the policy is CPUFREQ_POLICY_PERFORMANCE. > > Fixes: a4675fbc4a7a (cpufreq: intel_pstate: Replace timers with > utilization update callbacks) > Signed-off-by: Rafael J. Wysocki Acked-by: Srinivas Pandruvada > --- >  drivers/cpufreq/intel_pstate.c |   29 +++++++++++++++++++++++++---- >  1 file changed, 25 insertions(+), 4 deletions(-) > > Index: linux-pm/drivers/cpufreq/intel_pstate.c > =================================================================== > --- linux-pm.orig/drivers/cpufreq/intel_pstate.c > +++ linux-pm/drivers/cpufreq/intel_pstate.c > @@ -1137,10 +1137,8 @@ static void intel_pstate_get_min_max(str >   *min = clamp_t(int, min_perf, cpu->pstate.min_pstate, > max_perf); >  } >   > -static void intel_pstate_set_min_pstate(struct cpudata *cpu) > +static void intel_pstate_set_pstate(struct cpudata *cpu, int pstate) >  { > - int pstate = cpu->pstate.min_pstate; > - >   trace_cpu_frequency(pstate * cpu->pstate.scaling, cpu->cpu); >   cpu->pstate.current_pstate = pstate; >   /* > @@ -1152,6 +1150,20 @@ static void intel_pstate_set_min_pstate( >         pstate_funcs.get_val(cpu, pstate)); >  } >   > +static void intel_pstate_set_min_pstate(struct cpudata *cpu) > +{ > + intel_pstate_set_pstate(cpu, cpu->pstate.min_pstate); > +} > + > +static void intel_pstate_max_within_limits(struct cpudata *cpu) > +{ > + int min_pstate, max_pstate; > + > + update_turbo_state(); > + intel_pstate_get_min_max(cpu, &min_pstate, &max_pstate); > + intel_pstate_set_pstate(cpu, max_pstate); > +} > + >  static void intel_pstate_get_cpu_pstates(struct cpudata *cpu) >  { >   cpu->pstate.min_pstate = pstate_funcs.get_min(); > @@ -1485,7 +1497,7 @@ static int intel_pstate_set_policy(struc >   pr_debug("set_policy cpuinfo.max %u policy->max %u\n", >    policy->cpuinfo.max_freq, policy->max); >   > - cpu = all_cpu_data[0]; > + cpu = all_cpu_data[policy->cpu]; >   if (cpu->pstate.max_pstate_physical > cpu->pstate.max_pstate > && >       policy->max < policy->cpuinfo.max_freq && >       policy->max > cpu->pstate.max_pstate * cpu- > >pstate.scaling) { > @@ -1529,6 +1541,15 @@ static int intel_pstate_set_policy(struc >   limits->max_perf = round_up(limits->max_perf, FRAC_BITS); >   >   out: > + if (policy->policy == CPUFREQ_POLICY_PERFORMANCE) { > + /* > +  * NOHZ_FULL CPUs need this as the governor callback > may not > +  * be invoked on them. > +  */ > + intel_pstate_clear_update_util_hook(policy->cpu); > + intel_pstate_max_within_limits(cpu); > + } > + >   intel_pstate_set_update_util_hook(policy->cpu); >   >   intel_pstate_hwp_set_policy(policy); >