Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756964AbcCRO7c (ORCPT ); Fri, 18 Mar 2016 10:59:32 -0400 Received: from e23smtp09.au.ibm.com ([202.81.31.142]:47804 "EHLO e23smtp09.au.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752982AbcCRO73 (ORCPT ); Fri, 18 Mar 2016 10:59:29 -0400 X-IBM-Helo: d23dlp03.au.ibm.com X-IBM-MailFrom: shilpa.bhat@linux.vnet.ibm.com X-IBM-RcptTo: linux-kernel@vger.kernel.org;linux-pm@vger.kernel.org From: Shilpasri G Bhat To: linuxppc-dev@ozlabs.org, linux-kernel@vger.kernel.org Cc: ego@linux.vnet.ibm.com, linux-pm@vger.kernel.org, viresh.kumar@linaro.org, rjw@rjwysocki.net, pc@us.ibm.com, shreyas@linux.vnet.ibm.com, anton@samba.org, mikey@neuling.org, shilpa.bhat@linux.vnet.ibm.com Subject: [PATCH] cpufreq: powernv: Define per_cpu chip pointer to optimize hot-path Date: Fri, 18 Mar 2016 20:28:20 +0530 Message-Id: <1458313100-27676-1-git-send-email-shilpa.bhat@linux.vnet.ibm.com> X-Mailer: git-send-email 1.9.3 In-Reply-To: <1458274308.6622.79.camel@neuling.org> References: <1458274308.6622.79.camel@neuling.org> X-TM-AS-MML: disable X-Content-Scanned: Fidelis XPS MAILER x-cbid: 16031814-0033-0000-0000-0000039106BE Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 4242 Lines: 142 From: Michael Neuling "cpufreq: powernv: Remove cpu_to_chip_id() from hot-path" introduced 'core_to_chip_map' array to cache the chip-id of all cores. Replace this with per_cpu variable that stores the pointer to the chip-array. This removes the linear lookup and provides a neater and simpler solution. Signed-off-by: Michael Neuling Tested-by: Shilpasri G Bhat --- - Rebased the patch on top of linux-pm/linux-next - nr_chips is defined static, so it will be initialized to zero - Moved the initialization of the per_cpu variable after 'chips' is allocated - Removed 'core_to_chip_map' drivers/cpufreq/powernv-cpufreq.c | 50 +++++++++++++-------------------------- 1 file changed, 17 insertions(+), 33 deletions(-) diff --git a/drivers/cpufreq/powernv-cpufreq.c b/drivers/cpufreq/powernv-cpufreq.c index 50bf120..a00bcc2 100644 --- a/drivers/cpufreq/powernv-cpufreq.c +++ b/drivers/cpufreq/powernv-cpufreq.c @@ -44,7 +44,6 @@ static struct cpufreq_frequency_table powernv_freqs[POWERNV_MAX_PSTATES+1]; static bool rebooting, throttled, occ_reset; -static unsigned int *core_to_chip_map; static const char * const throttle_reason[] = { "No throttling", @@ -65,6 +64,7 @@ static struct chip { } *chips; static int nr_chips; +static DEFINE_PER_CPU(struct chip *, chip_info); /* * Note: The set of pstates consists of contiguous integers, the @@ -324,34 +324,31 @@ static inline unsigned int get_nominal_index(void) static void powernv_cpufreq_throttle_check(void *data) { + struct chip *chip; unsigned int cpu = smp_processor_id(); - unsigned int chip_id = core_to_chip_map[cpu_core_index_of_thread(cpu)]; unsigned long pmsr; - int pmsr_pmax, i; + int pmsr_pmax; pmsr = get_pmspr(SPRN_PMSR); - - for (i = 0; i < nr_chips; i++) - if (chips[i].id == chip_id) - break; + chip = this_cpu_read(chip_info); /* Check for Pmax Capping */ pmsr_pmax = (s8)PMSR_MAX(pmsr); if (pmsr_pmax != powernv_pstate_info.max) { - if (chips[i].throttled) + if (chip->throttled) goto next; - chips[i].throttled = true; + chip->throttled = true; if (pmsr_pmax < powernv_pstate_info.nominal) pr_warn_once("CPU %d on Chip %u has Pmax reduced below nominal frequency (%d < %d)\n", - cpu, chips[i].id, pmsr_pmax, + cpu, chip->id, pmsr_pmax, powernv_pstate_info.nominal); - trace_powernv_throttle(chips[i].id, - throttle_reason[chips[i].throttle_reason], + trace_powernv_throttle(chip->id, + throttle_reason[chip->throttle_reason], pmsr_pmax); - } else if (chips[i].throttled) { - chips[i].throttled = false; - trace_powernv_throttle(chips[i].id, - throttle_reason[chips[i].throttle_reason], + } else if (chip->throttled) { + chip->throttled = false; + trace_powernv_throttle(chip->id, + throttle_reason[chip->throttle_reason], pmsr_pmax); } @@ -558,47 +555,34 @@ static int init_chip_info(void) unsigned int chip[256]; unsigned int cpu, i; unsigned int prev_chip_id = UINT_MAX; - cpumask_t cpu_mask; - int ret = -ENOMEM; - - core_to_chip_map = kcalloc(cpu_nr_cores(), sizeof(unsigned int), - GFP_KERNEL); - if (!core_to_chip_map) - goto out; - cpumask_copy(&cpu_mask, cpu_possible_mask); - for_each_cpu(cpu, &cpu_mask) { + for_each_possible_cpu(cpu) { unsigned int id = cpu_to_chip_id(cpu); if (prev_chip_id != id) { prev_chip_id = id; chip[nr_chips++] = id; } - core_to_chip_map[cpu_core_index_of_thread(cpu)] = id; - cpumask_andnot(&cpu_mask, &cpu_mask, cpu_sibling_mask(cpu)); } chips = kcalloc(nr_chips, sizeof(struct chip), GFP_KERNEL); if (!chips) - goto free_chip_map; + return -ENOMEM; for (i = 0; i < nr_chips; i++) { chips[i].id = chip[i]; cpumask_copy(&chips[i].mask, cpumask_of_node(chip[i])); INIT_WORK(&chips[i].throttle, powernv_cpufreq_work_fn); + for_each_cpu(cpu, &chips[i].mask) + per_cpu(chip_info, cpu) = &chips[i]; } return 0; -free_chip_map: - kfree(core_to_chip_map); -out: - return ret; } static inline void clean_chip_info(void) { kfree(chips); - kfree(core_to_chip_map); } static inline void unregister_all_notifiers(void) -- 1.9.3