From: Sumit Gupta <sumitg@nvidia.com> To: <rafael@kernel.org>, <rui.zhang@intel.com>, <lenb@kernel.org>, <lpieralisi@kernel.org>, <guohanjun@huawei.com>, <sudeep.holla@arm.com>, <linux-acpi@vger.kernel.org>, <linux-tegra@vger.kernel.org>, <linux-kernel@vger.kernel.org>, <linux-arm-kernel@lists.infradead.org> Cc: <treding@nvidia.com>, <jonathanh@nvidia.com>, <bbasu@nvidia.com>, <sumitg@nvidia.com>, <sanjayc@nvidia.com>, <ksitaraman@nvidia.com>, <srikars@nvidia.com>, <jbrasen@nvidia.com> Subject: [Patch v6 2/2] ACPI: processor: reduce CPUFREQ thermal reduction pctg for Tegra241 Date: Fri, 10 Nov 2023 00:03:22 +0530 [thread overview] Message-ID: <20231109183322.28039-3-sumitg@nvidia.com> (raw) In-Reply-To: <20231109183322.28039-1-sumitg@nvidia.com> From: Srikar Srimath Tirumala <srikars@nvidia.com> Current implementation of processor_thermal performs software throttling in fixed steps of "20%" which can be too coarse for some platforms. We observed some performance gain after reducing the throttle percentage. Change the CPUFREQ thermal reduction percentage and maximum thermal steps to be configurable. Also, update the default values of both for Nvidia Tegra241 (Grace) SoC. The thermal reduction percentage is reduced to "5%" and accordingly the maximum number of thermal steps are increased as they are derived from the reduction percentage. Signed-off-by: Srikar Srimath Tirumala <srikars@nvidia.com> Co-developed-by: Sumit Gupta <sumitg@nvidia.com> Signed-off-by: Sumit Gupta <sumitg@nvidia.com> --- drivers/acpi/arm64/Makefile | 1 + drivers/acpi/arm64/thermal_cpufreq.c | 22 +++++++++++++ drivers/acpi/internal.h | 9 +++++ drivers/acpi/processor_thermal.c | 49 +++++++++++++++++++++++----- 4 files changed, 72 insertions(+), 9 deletions(-) create mode 100644 drivers/acpi/arm64/thermal_cpufreq.c diff --git a/drivers/acpi/arm64/Makefile b/drivers/acpi/arm64/Makefile index 143debc1ba4a..726944648c9b 100644 --- a/drivers/acpi/arm64/Makefile +++ b/drivers/acpi/arm64/Makefile @@ -5,3 +5,4 @@ obj-$(CONFIG_ACPI_GTDT) += gtdt.o obj-$(CONFIG_ACPI_APMT) += apmt.o obj-$(CONFIG_ARM_AMBA) += amba.o obj-y += dma.o init.o +obj-y += thermal_cpufreq.o diff --git a/drivers/acpi/arm64/thermal_cpufreq.c b/drivers/acpi/arm64/thermal_cpufreq.c new file mode 100644 index 000000000000..40d5806ed528 --- /dev/null +++ b/drivers/acpi/arm64/thermal_cpufreq.c @@ -0,0 +1,22 @@ +// SPDX-License-Identifier: GPL-2.0-only +#include <linux/acpi.h> + +#include "../internal.h" + +#ifdef CONFIG_HAVE_ARM_SMCCC_DISCOVERY +#define SMCCC_SOC_ID_T241 0x036b0241 + +int acpi_arch_thermal_cpufreq_pctg(void) +{ + s32 soc_id = arm_smccc_get_soc_id_version(); + + /* + * Check JEP106 code for NVIDIA Tegra241 chip (036b:0241) and + * reduce the CPUFREQ Thermal reduction percentage to 5%. + */ + if (soc_id == SMCCC_SOC_ID_T241) + return 5; + + return 0; +} +#endif diff --git a/drivers/acpi/internal.h b/drivers/acpi/internal.h index 866c7c4ed233..ee213a8cddc5 100644 --- a/drivers/acpi/internal.h +++ b/drivers/acpi/internal.h @@ -85,6 +85,15 @@ bool acpi_scan_is_offline(struct acpi_device *adev, bool uevent); acpi_status acpi_sysfs_table_handler(u32 event, void *table, void *context); void acpi_scan_table_notify(void); +#ifdef CONFIG_HAVE_ARM_SMCCC_DISCOVERY +int acpi_arch_thermal_cpufreq_pctg(void); +#else +static inline int acpi_arch_thermal_cpufreq_pctg(void) +{ + return 0; +} +#endif + /* -------------------------------------------------------------------------- Device Node Initialization / Removal -------------------------------------------------------------------------- */ diff --git a/drivers/acpi/processor_thermal.c b/drivers/acpi/processor_thermal.c index b7c6287eccca..1219adb11ab9 100644 --- a/drivers/acpi/processor_thermal.c +++ b/drivers/acpi/processor_thermal.c @@ -17,6 +17,8 @@ #include <acpi/processor.h> #include <linux/uaccess.h> +#include "internal.h" + #ifdef CONFIG_CPU_FREQ /* If a passive cooling situation is detected, primarily CPUfreq is used, as it @@ -26,12 +28,21 @@ */ #define CPUFREQ_THERMAL_MIN_STEP 0 -#define CPUFREQ_THERMAL_MAX_STEP 3 -static DEFINE_PER_CPU(unsigned int, cpufreq_thermal_reduction_pctg); +static int cpufreq_thermal_max_step __read_mostly = 3; + +/* + * Minimum throttle percentage for processor_thermal cooling device. + * The processor_thermal driver uses it to calculate the percentage amount by + * which cpu frequency must be reduced for each cooling state. This is also used + * to calculate the maximum number of throttling steps or cooling states. + */ +static int cpufreq_thermal_reduction_pctg __read_mostly = 20; -#define reduction_pctg(cpu) \ - per_cpu(cpufreq_thermal_reduction_pctg, phys_package_first_cpu(cpu)) +static DEFINE_PER_CPU(unsigned int, cpufreq_thermal_reduction_step); + +#define reduction_step(cpu) \ + per_cpu(cpufreq_thermal_reduction_step, phys_package_first_cpu(cpu)) /* * Emulate "per package data" using per cpu data (which should really be @@ -71,7 +82,7 @@ static int cpufreq_get_max_state(unsigned int cpu) if (!cpu_has_cpufreq(cpu)) return 0; - return CPUFREQ_THERMAL_MAX_STEP; + return cpufreq_thermal_max_step; } static int cpufreq_get_cur_state(unsigned int cpu) @@ -79,7 +90,7 @@ static int cpufreq_get_cur_state(unsigned int cpu) if (!cpu_has_cpufreq(cpu)) return 0; - return reduction_pctg(cpu); + return reduction_step(cpu); } static int cpufreq_set_cur_state(unsigned int cpu, int state) @@ -92,7 +103,7 @@ static int cpufreq_set_cur_state(unsigned int cpu, int state) if (!cpu_has_cpufreq(cpu)) return 0; - reduction_pctg(cpu) = state; + reduction_step(cpu) = state; /* * Update all the CPUs in the same package because they all @@ -113,7 +124,8 @@ static int cpufreq_set_cur_state(unsigned int cpu, int state) if (!policy) return -EINVAL; - max_freq = (policy->cpuinfo.max_freq * (100 - reduction_pctg(i) * 20)) / 100; + max_freq = (policy->cpuinfo.max_freq * + (100 - reduction_step(i) * cpufreq_thermal_reduction_pctg)) / 100; cpufreq_cpu_put(policy); @@ -126,10 +138,29 @@ static int cpufreq_set_cur_state(unsigned int cpu, int state) return 0; } +static void acpi_thermal_cpufreq_config(void) +{ + int cpufreq_pctg = acpi_arch_thermal_cpufreq_pctg(); + + if (!cpufreq_pctg) + return; + + cpufreq_thermal_reduction_pctg = cpufreq_pctg; + + /* + * Derive the MAX_STEP from minimum throttle percentage so that the reduction + * percentage doesn't end up becoming negative. Also, cap the MAX_STEP so that + * the CPU performance doesn't become 0. + */ + cpufreq_thermal_max_step = (100 / cpufreq_pctg) - 2; +} + void acpi_thermal_cpufreq_init(struct cpufreq_policy *policy) { unsigned int cpu; + acpi_thermal_cpufreq_config(); + for_each_cpu(cpu, policy->related_cpus) { struct acpi_processor *pr = per_cpu(processors, cpu); int ret; @@ -190,7 +221,7 @@ static int acpi_processor_max_state(struct acpi_processor *pr) /* * There exists four states according to - * cpufreq_thermal_reduction_pctg. 0, 1, 2, 3 + * cpufreq_thermal_reduction_step. 0, 1, 2, 3 */ max_state += cpufreq_get_max_state(pr->id); if (pr->flags.throttling) -- 2.17.1
WARNING: multiple messages have this Message-ID (diff)
From: Sumit Gupta <sumitg@nvidia.com> To: <rafael@kernel.org>, <rui.zhang@intel.com>, <lenb@kernel.org>, <lpieralisi@kernel.org>, <guohanjun@huawei.com>, <sudeep.holla@arm.com>, <linux-acpi@vger.kernel.org>, <linux-tegra@vger.kernel.org>, <linux-kernel@vger.kernel.org>, <linux-arm-kernel@lists.infradead.org> Cc: <treding@nvidia.com>, <jonathanh@nvidia.com>, <bbasu@nvidia.com>, <sumitg@nvidia.com>, <sanjayc@nvidia.com>, <ksitaraman@nvidia.com>, <srikars@nvidia.com>, <jbrasen@nvidia.com> Subject: [Patch v6 2/2] ACPI: processor: reduce CPUFREQ thermal reduction pctg for Tegra241 Date: Fri, 10 Nov 2023 00:03:22 +0530 [thread overview] Message-ID: <20231109183322.28039-3-sumitg@nvidia.com> (raw) In-Reply-To: <20231109183322.28039-1-sumitg@nvidia.com> From: Srikar Srimath Tirumala <srikars@nvidia.com> Current implementation of processor_thermal performs software throttling in fixed steps of "20%" which can be too coarse for some platforms. We observed some performance gain after reducing the throttle percentage. Change the CPUFREQ thermal reduction percentage and maximum thermal steps to be configurable. Also, update the default values of both for Nvidia Tegra241 (Grace) SoC. The thermal reduction percentage is reduced to "5%" and accordingly the maximum number of thermal steps are increased as they are derived from the reduction percentage. Signed-off-by: Srikar Srimath Tirumala <srikars@nvidia.com> Co-developed-by: Sumit Gupta <sumitg@nvidia.com> Signed-off-by: Sumit Gupta <sumitg@nvidia.com> --- drivers/acpi/arm64/Makefile | 1 + drivers/acpi/arm64/thermal_cpufreq.c | 22 +++++++++++++ drivers/acpi/internal.h | 9 +++++ drivers/acpi/processor_thermal.c | 49 +++++++++++++++++++++++----- 4 files changed, 72 insertions(+), 9 deletions(-) create mode 100644 drivers/acpi/arm64/thermal_cpufreq.c diff --git a/drivers/acpi/arm64/Makefile b/drivers/acpi/arm64/Makefile index 143debc1ba4a..726944648c9b 100644 --- a/drivers/acpi/arm64/Makefile +++ b/drivers/acpi/arm64/Makefile @@ -5,3 +5,4 @@ obj-$(CONFIG_ACPI_GTDT) += gtdt.o obj-$(CONFIG_ACPI_APMT) += apmt.o obj-$(CONFIG_ARM_AMBA) += amba.o obj-y += dma.o init.o +obj-y += thermal_cpufreq.o diff --git a/drivers/acpi/arm64/thermal_cpufreq.c b/drivers/acpi/arm64/thermal_cpufreq.c new file mode 100644 index 000000000000..40d5806ed528 --- /dev/null +++ b/drivers/acpi/arm64/thermal_cpufreq.c @@ -0,0 +1,22 @@ +// SPDX-License-Identifier: GPL-2.0-only +#include <linux/acpi.h> + +#include "../internal.h" + +#ifdef CONFIG_HAVE_ARM_SMCCC_DISCOVERY +#define SMCCC_SOC_ID_T241 0x036b0241 + +int acpi_arch_thermal_cpufreq_pctg(void) +{ + s32 soc_id = arm_smccc_get_soc_id_version(); + + /* + * Check JEP106 code for NVIDIA Tegra241 chip (036b:0241) and + * reduce the CPUFREQ Thermal reduction percentage to 5%. + */ + if (soc_id == SMCCC_SOC_ID_T241) + return 5; + + return 0; +} +#endif diff --git a/drivers/acpi/internal.h b/drivers/acpi/internal.h index 866c7c4ed233..ee213a8cddc5 100644 --- a/drivers/acpi/internal.h +++ b/drivers/acpi/internal.h @@ -85,6 +85,15 @@ bool acpi_scan_is_offline(struct acpi_device *adev, bool uevent); acpi_status acpi_sysfs_table_handler(u32 event, void *table, void *context); void acpi_scan_table_notify(void); +#ifdef CONFIG_HAVE_ARM_SMCCC_DISCOVERY +int acpi_arch_thermal_cpufreq_pctg(void); +#else +static inline int acpi_arch_thermal_cpufreq_pctg(void) +{ + return 0; +} +#endif + /* -------------------------------------------------------------------------- Device Node Initialization / Removal -------------------------------------------------------------------------- */ diff --git a/drivers/acpi/processor_thermal.c b/drivers/acpi/processor_thermal.c index b7c6287eccca..1219adb11ab9 100644 --- a/drivers/acpi/processor_thermal.c +++ b/drivers/acpi/processor_thermal.c @@ -17,6 +17,8 @@ #include <acpi/processor.h> #include <linux/uaccess.h> +#include "internal.h" + #ifdef CONFIG_CPU_FREQ /* If a passive cooling situation is detected, primarily CPUfreq is used, as it @@ -26,12 +28,21 @@ */ #define CPUFREQ_THERMAL_MIN_STEP 0 -#define CPUFREQ_THERMAL_MAX_STEP 3 -static DEFINE_PER_CPU(unsigned int, cpufreq_thermal_reduction_pctg); +static int cpufreq_thermal_max_step __read_mostly = 3; + +/* + * Minimum throttle percentage for processor_thermal cooling device. + * The processor_thermal driver uses it to calculate the percentage amount by + * which cpu frequency must be reduced for each cooling state. This is also used + * to calculate the maximum number of throttling steps or cooling states. + */ +static int cpufreq_thermal_reduction_pctg __read_mostly = 20; -#define reduction_pctg(cpu) \ - per_cpu(cpufreq_thermal_reduction_pctg, phys_package_first_cpu(cpu)) +static DEFINE_PER_CPU(unsigned int, cpufreq_thermal_reduction_step); + +#define reduction_step(cpu) \ + per_cpu(cpufreq_thermal_reduction_step, phys_package_first_cpu(cpu)) /* * Emulate "per package data" using per cpu data (which should really be @@ -71,7 +82,7 @@ static int cpufreq_get_max_state(unsigned int cpu) if (!cpu_has_cpufreq(cpu)) return 0; - return CPUFREQ_THERMAL_MAX_STEP; + return cpufreq_thermal_max_step; } static int cpufreq_get_cur_state(unsigned int cpu) @@ -79,7 +90,7 @@ static int cpufreq_get_cur_state(unsigned int cpu) if (!cpu_has_cpufreq(cpu)) return 0; - return reduction_pctg(cpu); + return reduction_step(cpu); } static int cpufreq_set_cur_state(unsigned int cpu, int state) @@ -92,7 +103,7 @@ static int cpufreq_set_cur_state(unsigned int cpu, int state) if (!cpu_has_cpufreq(cpu)) return 0; - reduction_pctg(cpu) = state; + reduction_step(cpu) = state; /* * Update all the CPUs in the same package because they all @@ -113,7 +124,8 @@ static int cpufreq_set_cur_state(unsigned int cpu, int state) if (!policy) return -EINVAL; - max_freq = (policy->cpuinfo.max_freq * (100 - reduction_pctg(i) * 20)) / 100; + max_freq = (policy->cpuinfo.max_freq * + (100 - reduction_step(i) * cpufreq_thermal_reduction_pctg)) / 100; cpufreq_cpu_put(policy); @@ -126,10 +138,29 @@ static int cpufreq_set_cur_state(unsigned int cpu, int state) return 0; } +static void acpi_thermal_cpufreq_config(void) +{ + int cpufreq_pctg = acpi_arch_thermal_cpufreq_pctg(); + + if (!cpufreq_pctg) + return; + + cpufreq_thermal_reduction_pctg = cpufreq_pctg; + + /* + * Derive the MAX_STEP from minimum throttle percentage so that the reduction + * percentage doesn't end up becoming negative. Also, cap the MAX_STEP so that + * the CPU performance doesn't become 0. + */ + cpufreq_thermal_max_step = (100 / cpufreq_pctg) - 2; +} + void acpi_thermal_cpufreq_init(struct cpufreq_policy *policy) { unsigned int cpu; + acpi_thermal_cpufreq_config(); + for_each_cpu(cpu, policy->related_cpus) { struct acpi_processor *pr = per_cpu(processors, cpu); int ret; @@ -190,7 +221,7 @@ static int acpi_processor_max_state(struct acpi_processor *pr) /* * There exists four states according to - * cpufreq_thermal_reduction_pctg. 0, 1, 2, 3 + * cpufreq_thermal_reduction_step. 0, 1, 2, 3 */ max_state += cpufreq_get_max_state(pr->id); if (pr->flags.throttling) -- 2.17.1 _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
next prev parent reply other threads:[~2023-11-09 18:34 UTC|newest] Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top 2023-11-09 18:33 [Patch v6 0/2] Add support for _TFP and change throttle pctg Sumit Gupta 2023-11-09 18:33 ` Sumit Gupta 2023-11-09 18:33 ` [Patch v6 1/2] ACPI: thermal: Add Thermal fast Sampling Period (_TFP) support Sumit Gupta 2023-11-09 18:33 ` Sumit Gupta 2023-11-21 19:59 ` Rafael J. Wysocki 2023-11-21 19:59 ` Rafael J. Wysocki 2023-11-22 12:55 ` Sumit Gupta 2023-11-22 12:55 ` Sumit Gupta 2023-11-22 13:34 ` Rafael J. Wysocki 2023-11-22 13:34 ` Rafael J. Wysocki 2023-11-09 18:33 ` Sumit Gupta [this message] 2023-11-09 18:33 ` [Patch v6 2/2] ACPI: processor: reduce CPUFREQ thermal reduction pctg for Tegra241 Sumit Gupta 2023-11-10 10:15 ` Sudeep Holla 2023-11-10 10:15 ` Sudeep Holla 2023-11-17 10:38 ` Sumit Gupta 2023-11-17 10:38 ` Sumit Gupta
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20231109183322.28039-3-sumitg@nvidia.com \ --to=sumitg@nvidia.com \ --cc=bbasu@nvidia.com \ --cc=guohanjun@huawei.com \ --cc=jbrasen@nvidia.com \ --cc=jonathanh@nvidia.com \ --cc=ksitaraman@nvidia.com \ --cc=lenb@kernel.org \ --cc=linux-acpi@vger.kernel.org \ --cc=linux-arm-kernel@lists.infradead.org \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-tegra@vger.kernel.org \ --cc=lpieralisi@kernel.org \ --cc=rafael@kernel.org \ --cc=rui.zhang@intel.com \ --cc=sanjayc@nvidia.com \ --cc=srikars@nvidia.com \ --cc=sudeep.holla@arm.com \ --cc=treding@nvidia.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.