From patchwork Wed Sep 8 14:05:22 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Vincent Donnefort X-Patchwork-Id: 508165 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI, SPF_HELO_NONE, SPF_PASS, USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id E8ADCC433EF for ; Wed, 8 Sep 2021 14:05:44 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id CED6D6113D for ; Wed, 8 Sep 2021 14:05:44 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1349505AbhIHOGv (ORCPT ); Wed, 8 Sep 2021 10:06:51 -0400 Received: from foss.arm.com ([217.140.110.172]:47084 "EHLO foss.arm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229600AbhIHOGv (ORCPT ); Wed, 8 Sep 2021 10:06:51 -0400 Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.121.207.14]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id 8C089D6E; Wed, 8 Sep 2021 07:05:43 -0700 (PDT) Received: from e120877-lin.cambridge.arm.com (e120877-lin.cambridge.arm.com [10.1.194.43]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPA id 3BF3F3F766; Wed, 8 Sep 2021 07:05:42 -0700 (PDT) From: Vincent Donnefort To: rjw@rjwysocki.net, viresh.kumar@linaro.org, vincent.guittot@linaro.org, qperret@google.com Cc: linux-pm@vger.kernel.org, ionela.voinescu@arm.com, lukasz.luba@arm.com, dietmar.eggemann@arm.com, mka@chromium.org, Vincent Donnefort Subject: [PATCH v7 1/9] PM / EM: Fix inefficient states detection Date: Wed, 8 Sep 2021 15:05:22 +0100 Message-Id: <1631109930-290049-2-git-send-email-vincent.donnefort@arm.com> X-Mailer: git-send-email 2.7.4 In-Reply-To: <1631109930-290049-1-git-send-email-vincent.donnefort@arm.com> References: <1631109930-290049-1-git-send-email-vincent.donnefort@arm.com> Precedence: bulk List-ID: X-Mailing-List: linux-pm@vger.kernel.org Currently, a debug message is printed if an inefficient state is detected in the Energy Model. Unfortunately, it won't detect if the first state is inefficient or if two successive states are. Fix this behavior. Fixes: 27871f7a8a34 (PM: Introduce an Energy Model management framework) Signed-off-by: Vincent Donnefort Reviewed-by: Quentin Perret Reviewed-by: Lukasz Luba Reviewed-by: Matthias Kaehlcke diff --git a/kernel/power/energy_model.c b/kernel/power/energy_model.c index a332ccd829e2..97e62469a6b3 100644 --- a/kernel/power/energy_model.c +++ b/kernel/power/energy_model.c @@ -107,8 +107,7 @@ static void em_debug_remove_pd(struct device *dev) {} static int em_create_perf_table(struct device *dev, struct em_perf_domain *pd, int nr_states, struct em_data_callback *cb) { - unsigned long opp_eff, prev_opp_eff = ULONG_MAX; - unsigned long power, freq, prev_freq = 0; + unsigned long power, freq, prev_freq = 0, prev_cost = ULONG_MAX; struct em_perf_state *table; int i, ret; u64 fmax; @@ -153,27 +152,21 @@ static int em_create_perf_table(struct device *dev, struct em_perf_domain *pd, table[i].power = power; table[i].frequency = prev_freq = freq; - - /* - * The hertz/watts efficiency ratio should decrease as the - * frequency grows on sane platforms. But this isn't always - * true in practice so warn the user if a higher OPP is more - * power efficient than a lower one. - */ - opp_eff = freq / power; - if (opp_eff >= prev_opp_eff) - dev_dbg(dev, "EM: hertz/watts ratio non-monotonically decreasing: em_perf_state %d >= em_perf_state%d\n", - i, i - 1); - prev_opp_eff = opp_eff; } /* Compute the cost of each performance state. */ fmax = (u64) table[nr_states - 1].frequency; - for (i = 0; i < nr_states; i++) { + for (i = nr_states - 1; i >= 0; i--) { unsigned long power_res = em_scale_power(table[i].power); table[i].cost = div64_u64(fmax * power_res, table[i].frequency); + if (table[i].cost >= prev_cost) { + dev_dbg(dev, "EM: OPP:%lu is inefficient\n", + table[i].frequency); + } else { + prev_cost = table[i].cost; + } } pd->table = table; From patchwork Wed Sep 8 14:05:24 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Vincent Donnefort X-Patchwork-Id: 508164 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI, SPF_HELO_NONE, SPF_PASS, USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 2F2E4C433EF for ; Wed, 8 Sep 2021 14:05:48 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 1836460C3E for ; Wed, 8 Sep 2021 14:05:48 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1351755AbhIHOGz (ORCPT ); Wed, 8 Sep 2021 10:06:55 -0400 Received: from foss.arm.com ([217.140.110.172]:47112 "EHLO foss.arm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229600AbhIHOGy (ORCPT ); Wed, 8 Sep 2021 10:06:54 -0400 Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.121.207.14]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id A086C1063; Wed, 8 Sep 2021 07:05:46 -0700 (PDT) Received: from e120877-lin.cambridge.arm.com (e120877-lin.cambridge.arm.com [10.1.194.43]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPA id 4FA043F766; Wed, 8 Sep 2021 07:05:45 -0700 (PDT) From: Vincent Donnefort To: rjw@rjwysocki.net, viresh.kumar@linaro.org, vincent.guittot@linaro.org, qperret@google.com Cc: linux-pm@vger.kernel.org, ionela.voinescu@arm.com, lukasz.luba@arm.com, dietmar.eggemann@arm.com, mka@chromium.org, Vincent Donnefort Subject: [PATCH v7 3/9] PM / EM: Extend em_perf_domain with a flag field Date: Wed, 8 Sep 2021 15:05:24 +0100 Message-Id: <1631109930-290049-4-git-send-email-vincent.donnefort@arm.com> X-Mailer: git-send-email 2.7.4 In-Reply-To: <1631109930-290049-1-git-send-email-vincent.donnefort@arm.com> References: <1631109930-290049-1-git-send-email-vincent.donnefort@arm.com> Precedence: bulk List-ID: X-Mailing-List: linux-pm@vger.kernel.org Merge the current "milliwatts" option into a "flag" field. This intends to prepare the extension of this structure for inefficient states support in the Energy Model. Signed-off-by: Vincent Donnefort Reviewed-by: Lukasz Luba diff --git a/include/linux/energy_model.h b/include/linux/energy_model.h index 629f5f63a7d7..ac2f7d0ab946 100644 --- a/include/linux/energy_model.h +++ b/include/linux/energy_model.h @@ -40,8 +40,7 @@ struct em_perf_state { * em_perf_domain - Performance domain * @table: List of performance states, in ascending order * @nr_perf_states: Number of performance states - * @milliwatts: Flag indicating the power values are in milli-Watts - * or some other scale. + * @flags: See "em_perf_domain flags" * @cpus: Cpumask covering the CPUs of the domain. It's here * for performance reasons to avoid potential cache * misses during energy calculations in the scheduler @@ -56,10 +55,18 @@ struct em_perf_state { struct em_perf_domain { struct em_perf_state *table; int nr_perf_states; - int milliwatts; + unsigned long flags; unsigned long cpus[]; }; +/* + * em_perf_domain flags: + * + * EM_PERF_DOMAIN_MILLIWATTS: The power values are in milli-Watts or some + * other scale. + */ +#define EM_PERF_DOMAIN_MILLIWATTS BIT(0) + #define em_span_cpus(em) (to_cpumask((em)->cpus)) #ifdef CONFIG_ENERGY_MODEL diff --git a/kernel/power/energy_model.c b/kernel/power/energy_model.c index 6d8438347535..3a7d1573b214 100644 --- a/kernel/power/energy_model.c +++ b/kernel/power/energy_model.c @@ -56,7 +56,8 @@ DEFINE_SHOW_ATTRIBUTE(em_debug_cpus); static int em_debug_units_show(struct seq_file *s, void *unused) { struct em_perf_domain *pd = s->private; - char *units = pd->milliwatts ? "milliWatts" : "bogoWatts"; + char *units = (pd->flags & EM_PERF_DOMAIN_MILLIWATTS) ? + "milliWatts" : "bogoWatts"; seq_printf(s, "%s\n", units); @@ -330,7 +331,8 @@ int em_dev_register_perf_domain(struct device *dev, unsigned int nr_states, if (ret) goto unlock; - dev->em_pd->milliwatts = milliwatts; + if (milliwatts) + dev->em_pd->flags |= EM_PERF_DOMAIN_MILLIWATTS; em_debug_create_pd(dev); dev_info(dev, "EM: created perf domain\n"); From patchwork Wed Sep 8 14:05:26 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Vincent Donnefort X-Patchwork-Id: 508163 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI, SPF_HELO_NONE, SPF_PASS, USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 86F44C433FE for ; Wed, 8 Sep 2021 14:05:51 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 6E63A60C3E for ; Wed, 8 Sep 2021 14:05:51 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S235140AbhIHOG6 (ORCPT ); Wed, 8 Sep 2021 10:06:58 -0400 Received: from foss.arm.com ([217.140.110.172]:47140 "EHLO foss.arm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1351761AbhIHOG5 (ORCPT ); Wed, 8 Sep 2021 10:06:57 -0400 Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.121.207.14]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id CAD94D6E; Wed, 8 Sep 2021 07:05:49 -0700 (PDT) Received: from e120877-lin.cambridge.arm.com (e120877-lin.cambridge.arm.com [10.1.194.43]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPA id 7A69E3F766; Wed, 8 Sep 2021 07:05:48 -0700 (PDT) From: Vincent Donnefort To: rjw@rjwysocki.net, viresh.kumar@linaro.org, vincent.guittot@linaro.org, qperret@google.com Cc: linux-pm@vger.kernel.org, ionela.voinescu@arm.com, lukasz.luba@arm.com, dietmar.eggemann@arm.com, mka@chromium.org, Vincent Donnefort Subject: [PATCH v7 5/9] cpufreq: Make policy min/max hard requirements Date: Wed, 8 Sep 2021 15:05:26 +0100 Message-Id: <1631109930-290049-6-git-send-email-vincent.donnefort@arm.com> X-Mailer: git-send-email 2.7.4 In-Reply-To: <1631109930-290049-1-git-send-email-vincent.donnefort@arm.com> References: <1631109930-290049-1-git-send-email-vincent.donnefort@arm.com> Precedence: bulk List-ID: X-Mailing-List: linux-pm@vger.kernel.org When applying the policy min/max limits, the requested frequency is simply clamped to not be out of range. It means, however, if one of the boundaries isn't an available frequency, the frequency resolution can return a value out of those limits, depending on the relation used. e.g. freq{0,1,2} being available frequencies. freq0 policy->min freq1 policy->max freq2 | | | | | 17kHz 18kHz 19kHz 20kHz 21kHz __resolve_freq(21kHz, CPUFREQ_RELATION_L) -> 21kHz (out of bounds) __resolve_freq(17kHz, CPUFREQ_RELATION_H) -> 17kHz (out of bounds) If, during the policy init, we resolve the requested min/max to existing frequencies, we ensure that any CPUFREQ_RELATION_* would resolve to a frequency which is inside the policy min/max range. Making the policy limits rigid helps to introduce the inefficient frequencies support. Resolving an inefficient frequency to an efficient one should not transgress policy->max (which can be set for thermal reason) and having a value we can trust simplify this comparison. Signed-off-by: Vincent Donnefort diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c index 7d5f170ecad1..95b0464249d6 100644 --- a/drivers/cpufreq/cpufreq.c +++ b/drivers/cpufreq/cpufreq.c @@ -2527,8 +2527,15 @@ static int cpufreq_set_policy(struct cpufreq_policy *policy, if (ret) return ret; + /* + * Resolve policy min/max to available frequencies. It ensures + * no frequency resolution will neither overshoot the requested maximum + * nor undershoot the requested minimum. + */ policy->min = new_data.min; policy->max = new_data.max; + policy->min = __resolve_freq(policy, policy->min, CPUFREQ_RELATION_L); + policy->max = __resolve_freq(policy, policy->max, CPUFREQ_RELATION_H); trace_cpu_frequency_limits(policy); policy->cached_target_freq = UINT_MAX; From patchwork Wed Sep 8 14:05:28 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Vincent Donnefort X-Patchwork-Id: 508162 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI, SPF_HELO_NONE, SPF_PASS, USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 90263C433F5 for ; Wed, 8 Sep 2021 14:05:54 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 76F5660C3E for ; Wed, 8 Sep 2021 14:05:54 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1351761AbhIHOHB (ORCPT ); Wed, 8 Sep 2021 10:07:01 -0400 Received: from foss.arm.com ([217.140.110.172]:47166 "EHLO foss.arm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1351804AbhIHOHA (ORCPT ); Wed, 8 Sep 2021 10:07:00 -0400 Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.121.207.14]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id DB3A11042; Wed, 8 Sep 2021 07:05:52 -0700 (PDT) Received: from e120877-lin.cambridge.arm.com (e120877-lin.cambridge.arm.com [10.1.194.43]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPA id 8BC053F766; Wed, 8 Sep 2021 07:05:51 -0700 (PDT) From: Vincent Donnefort To: rjw@rjwysocki.net, viresh.kumar@linaro.org, vincent.guittot@linaro.org, qperret@google.com Cc: linux-pm@vger.kernel.org, ionela.voinescu@arm.com, lukasz.luba@arm.com, dietmar.eggemann@arm.com, mka@chromium.org, Vincent Donnefort Subject: [PATCH v7 7/9] cpufreq: Introducing CPUFREQ_RELATION_E Date: Wed, 8 Sep 2021 15:05:28 +0100 Message-Id: <1631109930-290049-8-git-send-email-vincent.donnefort@arm.com> X-Mailer: git-send-email 2.7.4 In-Reply-To: <1631109930-290049-1-git-send-email-vincent.donnefort@arm.com> References: <1631109930-290049-1-git-send-email-vincent.donnefort@arm.com> Precedence: bulk List-ID: X-Mailing-List: linux-pm@vger.kernel.org This newly introduced flag can be applied by a governor to a CPUFreq relation, when looking for a frequency within the policy table. The resolution would then only walk through efficient frequencies. Even with the flag set, the policy max limit will still be honoured. If no efficient frequencies can be found within the limits of the policy, an inefficient one would be returned. Signed-off-by: Vincent Donnefort diff --git a/drivers/cpufreq/acpi-cpufreq.c b/drivers/cpufreq/acpi-cpufreq.c index b49612895c78..e88b6b5dd913 100644 --- a/drivers/cpufreq/acpi-cpufreq.c +++ b/drivers/cpufreq/acpi-cpufreq.c @@ -470,7 +470,8 @@ static unsigned int acpi_cpufreq_fast_switch(struct cpufreq_policy *policy, if (policy->cached_target_freq == target_freq) index = policy->cached_resolved_idx; else - index = cpufreq_table_find_index_dl(policy, target_freq); + index = cpufreq_table_find_index_dl(policy, target_freq, + false); entry = &policy->freq_table[index]; next_freq = entry->frequency; diff --git a/drivers/cpufreq/amd_freq_sensitivity.c b/drivers/cpufreq/amd_freq_sensitivity.c index d0b10baf039a..6448e03bcf48 100644 --- a/drivers/cpufreq/amd_freq_sensitivity.c +++ b/drivers/cpufreq/amd_freq_sensitivity.c @@ -91,7 +91,8 @@ static unsigned int amd_powersave_bias_target(struct cpufreq_policy *policy, unsigned int index; index = cpufreq_table_find_index_h(policy, - policy->cur - 1); + policy->cur - 1, + relation & CPUFREQ_RELATION_E); freq_next = policy->freq_table[index].frequency; } diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c index 95b0464249d6..9edf4abc9fa0 100644 --- a/drivers/cpufreq/cpufreq.c +++ b/drivers/cpufreq/cpufreq.c @@ -2264,8 +2264,16 @@ int __cpufreq_driver_target(struct cpufreq_policy *policy, !(cpufreq_driver->flags & CPUFREQ_NEED_UPDATE_LIMITS)) return 0; - if (cpufreq_driver->target) + if (cpufreq_driver->target) { + /* + * If the driver hasn't setup a single inefficient frequency, + * it's unlikely it knows how to decode CPUFREQ_RELATION_E. + */ + if (!policy->efficiencies_available) + relation &= ~CPUFREQ_RELATION_E; + return cpufreq_driver->target(policy, target_freq, relation); + } if (!cpufreq_driver->target_index) return -EINVAL; diff --git a/drivers/cpufreq/cpufreq_ondemand.c b/drivers/cpufreq/cpufreq_ondemand.c index eb4320b619c9..f68cad9abd11 100644 --- a/drivers/cpufreq/cpufreq_ondemand.c +++ b/drivers/cpufreq/cpufreq_ondemand.c @@ -83,9 +83,11 @@ static unsigned int generic_powersave_bias_target(struct cpufreq_policy *policy, freq_avg = freq_req - freq_reduc; /* Find freq bounds for freq_avg in freq_table */ - index = cpufreq_table_find_index_h(policy, freq_avg); + index = cpufreq_table_find_index_h(policy, freq_avg, + relation & CPUFREQ_RELATION_E); freq_lo = freq_table[index].frequency; - index = cpufreq_table_find_index_l(policy, freq_avg); + index = cpufreq_table_find_index_l(policy, freq_avg, + relation & CPUFREQ_RELATION_E); freq_hi = freq_table[index].frequency; /* Find out how long we have to be in hi and lo freqs */ diff --git a/drivers/cpufreq/powernv-cpufreq.c b/drivers/cpufreq/powernv-cpufreq.c index 23a06cba392c..ed5b317574ec 100644 --- a/drivers/cpufreq/powernv-cpufreq.c +++ b/drivers/cpufreq/powernv-cpufreq.c @@ -933,7 +933,7 @@ static void powernv_cpufreq_work_fn(struct work_struct *work) policy = cpufreq_cpu_get(cpu); if (!policy) continue; - index = cpufreq_table_find_index_c(policy, policy->cur); + index = cpufreq_table_find_index_c(policy, policy->cur, false); powernv_cpufreq_target_index(policy, index); cpumask_andnot(&mask, &mask, policy->cpus); cpufreq_cpu_put(policy); @@ -1021,7 +1021,7 @@ static unsigned int powernv_fast_switch(struct cpufreq_policy *policy, int index; struct powernv_smp_call_data freq_data; - index = cpufreq_table_find_index_dl(policy, target_freq); + index = cpufreq_table_find_index_dl(policy, target_freq, false); freq_data.pstate_id = powernv_freqs[index].driver_data; freq_data.gpstate_id = powernv_freqs[index].driver_data; set_pstate(&freq_data); diff --git a/drivers/cpufreq/s5pv210-cpufreq.c b/drivers/cpufreq/s5pv210-cpufreq.c index ad7d4f272ddc..76c888ed8d16 100644 --- a/drivers/cpufreq/s5pv210-cpufreq.c +++ b/drivers/cpufreq/s5pv210-cpufreq.c @@ -243,7 +243,7 @@ static int s5pv210_target(struct cpufreq_policy *policy, unsigned int index) new_freq = s5pv210_freq_table[index].frequency; /* Finding current running level index */ - priv_index = cpufreq_table_find_index_h(policy, old_freq); + priv_index = cpufreq_table_find_index_h(policy, old_freq, false); arm_volt = dvs_conf[index].arm_volt; int_volt = dvs_conf[index].int_volt; diff --git a/include/linux/cpufreq.h b/include/linux/cpufreq.h index 18b486a1c816..045565fe5b29 100644 --- a/include/linux/cpufreq.h +++ b/include/linux/cpufreq.h @@ -117,6 +117,13 @@ struct cpufreq_policy { bool strict_target; /* + * Set if inefficient frequencies were found in the frequency table. + * This indicates if the relation flag CPUFREQ_RELATION_E can be + * honored. + */ + bool efficiencies_available; + + /* * Preferred average time interval between consecutive invocations of * the driver to set the frequency for this policy. To be set by the * scaling driver (0, which is the default, means no preference). @@ -271,6 +278,8 @@ static inline void cpufreq_stats_record_transition(struct cpufreq_policy *policy #define CPUFREQ_RELATION_L 0 /* lowest frequency at or above target */ #define CPUFREQ_RELATION_H 1 /* highest frequency below or at target */ #define CPUFREQ_RELATION_C 2 /* closest frequency to target */ +/* relation flags */ +#define CPUFREQ_RELATION_E BIT(2) /* Get if possible an efficient frequency */ struct freq_attr { struct attribute attr; @@ -742,6 +751,22 @@ static inline void dev_pm_opp_free_cpufreq_table(struct device *dev, continue; \ else +/** + * cpufreq_for_each_efficient_entry_idx - iterate with index over a cpufreq + * frequency_table excluding CPUFREQ_ENTRY_INVALID and + * CPUFREQ_INEFFICIENT_FREQ frequencies. + * @pos: the &struct cpufreq_frequency_table to use as a loop cursor. + * @table: the &struct cpufreq_frequency_table to iterate over. + * @idx: the table entry currently being processed. + * @efficiencies: set to true to only iterate over efficient frequencies. + */ + +#define cpufreq_for_each_efficient_entry_idx(pos, table, idx, efficiencies) \ + cpufreq_for_each_valid_entry_idx(pos, table, idx) \ + if (efficiencies && (pos->flags & CPUFREQ_INEFFICIENT_FREQ)) \ + continue; \ + else + int cpufreq_frequency_table_cpuinfo(struct cpufreq_policy *policy, struct cpufreq_frequency_table *table); @@ -766,14 +791,15 @@ bool policy_has_boost_freq(struct cpufreq_policy *policy); /* Find lowest freq at or above target in a table in ascending order */ static inline int cpufreq_table_find_index_al(struct cpufreq_policy *policy, - unsigned int target_freq) + unsigned int target_freq, + bool efficiencies) { struct cpufreq_frequency_table *table = policy->freq_table; struct cpufreq_frequency_table *pos; unsigned int freq; int idx, best = -1; - cpufreq_for_each_valid_entry_idx(pos, table, idx) { + cpufreq_for_each_efficient_entry_idx(pos, table, idx, efficiencies) { freq = pos->frequency; if (freq >= target_freq) @@ -787,14 +813,15 @@ static inline int cpufreq_table_find_index_al(struct cpufreq_policy *policy, /* Find lowest freq at or above target in a table in descending order */ static inline int cpufreq_table_find_index_dl(struct cpufreq_policy *policy, - unsigned int target_freq) + unsigned int target_freq, + bool efficiencies) { struct cpufreq_frequency_table *table = policy->freq_table; struct cpufreq_frequency_table *pos; unsigned int freq; int idx, best = -1; - cpufreq_for_each_valid_entry_idx(pos, table, idx) { + cpufreq_for_each_efficient_entry_idx(pos, table, idx, efficiencies) { freq = pos->frequency; if (freq == target_freq) @@ -817,26 +844,30 @@ static inline int cpufreq_table_find_index_dl(struct cpufreq_policy *policy, /* Works only on sorted freq-tables */ static inline int cpufreq_table_find_index_l(struct cpufreq_policy *policy, - unsigned int target_freq) + unsigned int target_freq, + bool efficiencies) { target_freq = clamp_val(target_freq, policy->min, policy->max); if (policy->freq_table_sorted == CPUFREQ_TABLE_SORTED_ASCENDING) - return cpufreq_table_find_index_al(policy, target_freq); + return cpufreq_table_find_index_al(policy, target_freq, + efficiencies); else - return cpufreq_table_find_index_dl(policy, target_freq); + return cpufreq_table_find_index_dl(policy, target_freq, + efficiencies); } /* Find highest freq at or below target in a table in ascending order */ static inline int cpufreq_table_find_index_ah(struct cpufreq_policy *policy, - unsigned int target_freq) + unsigned int target_freq, + bool efficiencies) { struct cpufreq_frequency_table *table = policy->freq_table; struct cpufreq_frequency_table *pos; unsigned int freq; int idx, best = -1; - cpufreq_for_each_valid_entry_idx(pos, table, idx) { + cpufreq_for_each_efficient_entry_idx(pos, table, idx, efficiencies) { freq = pos->frequency; if (freq == target_freq) @@ -859,14 +890,15 @@ static inline int cpufreq_table_find_index_ah(struct cpufreq_policy *policy, /* Find highest freq at or below target in a table in descending order */ static inline int cpufreq_table_find_index_dh(struct cpufreq_policy *policy, - unsigned int target_freq) + unsigned int target_freq, + bool efficiencies) { struct cpufreq_frequency_table *table = policy->freq_table; struct cpufreq_frequency_table *pos; unsigned int freq; int idx, best = -1; - cpufreq_for_each_valid_entry_idx(pos, table, idx) { + cpufreq_for_each_efficient_entry_idx(pos, table, idx, efficiencies) { freq = pos->frequency; if (freq <= target_freq) @@ -880,26 +912,30 @@ static inline int cpufreq_table_find_index_dh(struct cpufreq_policy *policy, /* Works only on sorted freq-tables */ static inline int cpufreq_table_find_index_h(struct cpufreq_policy *policy, - unsigned int target_freq) + unsigned int target_freq, + bool efficiencies) { target_freq = clamp_val(target_freq, policy->min, policy->max); if (policy->freq_table_sorted == CPUFREQ_TABLE_SORTED_ASCENDING) - return cpufreq_table_find_index_ah(policy, target_freq); + return cpufreq_table_find_index_ah(policy, target_freq, + efficiencies); else - return cpufreq_table_find_index_dh(policy, target_freq); + return cpufreq_table_find_index_dh(policy, target_freq, + efficiencies); } /* Find closest freq to target in a table in ascending order */ static inline int cpufreq_table_find_index_ac(struct cpufreq_policy *policy, - unsigned int target_freq) + unsigned int target_freq, + bool efficiencies) { struct cpufreq_frequency_table *table = policy->freq_table; struct cpufreq_frequency_table *pos; unsigned int freq; int idx, best = -1; - cpufreq_for_each_valid_entry_idx(pos, table, idx) { + cpufreq_for_each_efficient_entry_idx(pos, table, idx, efficiencies) { freq = pos->frequency; if (freq == target_freq) @@ -926,14 +962,15 @@ static inline int cpufreq_table_find_index_ac(struct cpufreq_policy *policy, /* Find closest freq to target in a table in descending order */ static inline int cpufreq_table_find_index_dc(struct cpufreq_policy *policy, - unsigned int target_freq) + unsigned int target_freq, + bool efficiencies) { struct cpufreq_frequency_table *table = policy->freq_table; struct cpufreq_frequency_table *pos; unsigned int freq; int idx, best = -1; - cpufreq_for_each_valid_entry_idx(pos, table, idx) { + cpufreq_for_each_efficient_entry_idx(pos, table, idx, efficiencies) { freq = pos->frequency; if (freq == target_freq) @@ -960,35 +997,58 @@ static inline int cpufreq_table_find_index_dc(struct cpufreq_policy *policy, /* Works only on sorted freq-tables */ static inline int cpufreq_table_find_index_c(struct cpufreq_policy *policy, - unsigned int target_freq) + unsigned int target_freq, + bool efficiencies) { target_freq = clamp_val(target_freq, policy->min, policy->max); if (policy->freq_table_sorted == CPUFREQ_TABLE_SORTED_ASCENDING) - return cpufreq_table_find_index_ac(policy, target_freq); + return cpufreq_table_find_index_ac(policy, target_freq, + efficiencies); else - return cpufreq_table_find_index_dc(policy, target_freq); + return cpufreq_table_find_index_dc(policy, target_freq, + efficiencies); } static inline int cpufreq_frequency_table_target(struct cpufreq_policy *policy, unsigned int target_freq, unsigned int relation) { + bool efficiencies = policy->efficiencies_available && + (relation & CPUFREQ_RELATION_E); + int idx; + + /* cpufreq_table_index_unsorted() has no use for this flag anyway */ + relation &= ~CPUFREQ_RELATION_E; + if (unlikely(policy->freq_table_sorted == CPUFREQ_TABLE_UNSORTED)) return cpufreq_table_index_unsorted(policy, target_freq, relation); - +retry: switch (relation) { case CPUFREQ_RELATION_L: - return cpufreq_table_find_index_l(policy, target_freq); + idx = cpufreq_table_find_index_l(policy, target_freq, + efficiencies); + break; case CPUFREQ_RELATION_H: - return cpufreq_table_find_index_h(policy, target_freq); + idx = cpufreq_table_find_index_h(policy, target_freq, + efficiencies); + break; case CPUFREQ_RELATION_C: - return cpufreq_table_find_index_c(policy, target_freq); + idx = cpufreq_table_find_index_c(policy, target_freq, + efficiencies); + break; default: WARN_ON_ONCE(1); return 0; } + + if (idx < 0 && efficiencies) { + efficiencies = false; + goto retry; + } + + return idx; } static inline int cpufreq_table_count_valid_entries(const struct cpufreq_policy *policy) @@ -1028,6 +1088,7 @@ cpufreq_table_set_inefficient(struct cpufreq_policy *policy, cpufreq_for_each_valid_entry(pos, policy->freq_table) { if (pos->frequency == frequency) { pos->flags |= CPUFREQ_INEFFICIENT_FREQ; + policy->efficiencies_available = true; return 0; } } From patchwork Wed Sep 8 14:05:30 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Vincent Donnefort X-Patchwork-Id: 508161 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI, SPF_HELO_NONE, SPF_PASS, USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 1CCFFC433EF for ; Wed, 8 Sep 2021 14:05:58 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id F3CE9600AA for ; Wed, 8 Sep 2021 14:05:57 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1351804AbhIHOHE (ORCPT ); Wed, 8 Sep 2021 10:07:04 -0400 Received: from foss.arm.com ([217.140.110.172]:47194 "EHLO foss.arm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1351775AbhIHOHD (ORCPT ); Wed, 8 Sep 2021 10:07:03 -0400 Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.121.207.14]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id EC04D6D; Wed, 8 Sep 2021 07:05:55 -0700 (PDT) Received: from e120877-lin.cambridge.arm.com (e120877-lin.cambridge.arm.com [10.1.194.43]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPA id 9C35C3F766; Wed, 8 Sep 2021 07:05:54 -0700 (PDT) From: Vincent Donnefort To: rjw@rjwysocki.net, viresh.kumar@linaro.org, vincent.guittot@linaro.org, qperret@google.com Cc: linux-pm@vger.kernel.org, ionela.voinescu@arm.com, lukasz.luba@arm.com, dietmar.eggemann@arm.com, mka@chromium.org, Vincent Donnefort Subject: [PATCH v7 9/9] PM / EM: Mark inefficiencies in CPUFreq Date: Wed, 8 Sep 2021 15:05:30 +0100 Message-Id: <1631109930-290049-10-git-send-email-vincent.donnefort@arm.com> X-Mailer: git-send-email 2.7.4 In-Reply-To: <1631109930-290049-1-git-send-email-vincent.donnefort@arm.com> References: <1631109930-290049-1-git-send-email-vincent.donnefort@arm.com> Precedence: bulk List-ID: X-Mailing-List: linux-pm@vger.kernel.org The Energy Model has a 1:1 mapping between OPPs and performance states (em_perf_state). If a CPUFreq driver registers an Energy Model, inefficiencies found by the latter can be applied to CPUFreq. Signed-off-by: Vincent Donnefort diff --git a/kernel/power/energy_model.c b/kernel/power/energy_model.c index d353ef29e37f..0153b0ca7b23 100644 --- a/kernel/power/energy_model.c +++ b/kernel/power/energy_model.c @@ -10,6 +10,7 @@ #define pr_fmt(fmt) "energy_model: " fmt #include +#include #include #include #include @@ -231,6 +232,43 @@ static int em_create_pd(struct device *dev, int nr_states, return 0; } +static void em_cpufreq_update_efficiencies(struct device *dev) +{ + struct em_perf_domain *pd = dev->em_pd; + struct em_perf_state *table; + struct cpufreq_policy *policy; + int found = 0; + int i; + + if (!_is_cpu_device(dev) || !pd) + return; + + policy = cpufreq_cpu_get(cpumask_first(em_span_cpus(pd))); + if (!policy) { + dev_warn(dev, "EM: Access to CPUFreq policy failed"); + return; + } + + table = pd->table; + + for (i = 0; i < pd->nr_perf_states; i++) { + if (!(table[i].flags & EM_PERF_STATE_INEFFICIENT)) + continue; + + if (!cpufreq_table_set_inefficient(policy, table[i].frequency)) + found++; + } + + if (!found) + return; + + /* + * Efficiencies have been installed in CPUFreq, inefficient frequencies + * will be skipped. The EM can do the same. + */ + pd->flags |= EM_PERF_DOMAIN_SKIP_INEFFICIENCIES; +} + /** * em_pd_get() - Return the performance domain for a device * @dev : Device to find the performance domain for @@ -347,6 +385,8 @@ int em_dev_register_perf_domain(struct device *dev, unsigned int nr_states, if (milliwatts) dev->em_pd->flags |= EM_PERF_DOMAIN_MILLIWATTS; + em_cpufreq_update_efficiencies(dev); + em_debug_create_pd(dev); dev_info(dev, "EM: created perf domain\n");