From mboxrd@z Thu Jan 1 00:00:00 1970 From: Viresh Kumar Date: Thu, 22 Oct 2020 04:53:46 +0000 Subject: Re: [PATCH] sched/fair: check for idle core Message-Id: <20201022044146.4n2jl6jzyycfhfzg@vireshk-i7> List-Id: References: <1603211879-1064-1-git-send-email-Julia.Lawall@inria.fr> <20201021121950.GF2628@hirez.programming.kicks-ass.net> <20201021131026.GY2651@hirez.programming.kicks-ass.net> <6145907.dAfbsivgMF@kreacher> In-Reply-To: <6145907.dAfbsivgMF@kreacher> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: "Rafael J. Wysocki" Cc: Peter Zijlstra , Julia Lawall , Mel Gorman , Ingo Molnar , kernel-janitors@vger.kernel.org, Juri Lelli , Vincent Guittot , Dietmar Eggemann , Steven Rostedt , Ben Segall , Daniel Bristot de Oliveira , linux-kernel@vger.kernel.org, Valentin Schneider , Gilles Muller , srinivas.pandruvada@linux.intel.com On 21-10-20, 20:11, Rafael J. Wysocki wrote: > On Wednesday, October 21, 2020 3:10:26 PM CEST Peter Zijlstra wrote: > > On Wed, Oct 21, 2020 at 02:19:50PM +0200, Peter Zijlstra wrote: > > > On Wed, Oct 21, 2020 at 01:56:55PM +0200, Julia Lawall wrote: > > > > Prior to 5.8, my machine was using intel_pstate and had few background > > > > tasks. Thus the problem wasn't visible in practice. Starting with 5.8 > > > > the kernel decided that intel_cpufreq would be more appropriate, which > > > > introduced kworkers every 0.004 seconds on all cores. > > > > > > That still doesn't make any sense. Are you running the legacy on-demand > > > thing or something? > > > > > > Rafael, Srinivas, Viresh, how come it defaults to that? > > > > Does we want something like this? > > > > --- > > arch/x86/configs/i386_defconfig | 3 +-- > > arch/x86/configs/x86_64_defconfig | 3 +-- > > drivers/cpufreq/Kconfig | 7 +++++-- > > 3 files changed, 7 insertions(+), 6 deletions(-) > > > > diff --git a/arch/x86/configs/i386_defconfig b/arch/x86/configs/i386_defconfig > > index 78210793d357..c343ad459737 100644 > > --- a/arch/x86/configs/i386_defconfig > > +++ b/arch/x86/configs/i386_defconfig > > @@ -41,8 +41,7 @@ CONFIG_PM_DEBUG=y > > CONFIG_PM_TRACE_RTC=y > > CONFIG_ACPI_DOCK=y > > CONFIG_ACPI_BGRT=y > > -CONFIG_CPU_FREQ_DEFAULT_GOV_USERSPACE=y > > -CONFIG_CPU_FREQ_GOV_ONDEMAND=y > > +CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL=y > > CONFIG_X86_ACPI_CPUFREQ=y > > CONFIG_EFI_VARS=y > > CONFIG_KPROBES=y > > diff --git a/arch/x86/configs/x86_64_defconfig b/arch/x86/configs/x86_64_defconfig > > index 9936528e1939..23e1ea85c1ec 100644 > > --- a/arch/x86/configs/x86_64_defconfig > > +++ b/arch/x86/configs/x86_64_defconfig > > @@ -38,8 +38,7 @@ CONFIG_PM_DEBUG=y > > CONFIG_PM_TRACE_RTC=y > > CONFIG_ACPI_DOCK=y > > CONFIG_ACPI_BGRT=y > > -CONFIG_CPU_FREQ_DEFAULT_GOV_USERSPACE=y > > -CONFIG_CPU_FREQ_GOV_ONDEMAND=y > > +CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL=y > > CONFIG_X86_ACPI_CPUFREQ=y > > CONFIG_IA32_EMULATION=y > > CONFIG_EFI_VARS=y > > diff --git a/drivers/cpufreq/Kconfig b/drivers/cpufreq/Kconfig > > index 2c7171e0b001..8dfca6e9b836 100644 > > --- a/drivers/cpufreq/Kconfig > > +++ b/drivers/cpufreq/Kconfig > > @@ -37,8 +37,7 @@ config CPU_FREQ_STAT > > choice > > prompt "Default CPUFreq governor" > > default CPU_FREQ_DEFAULT_GOV_USERSPACE if ARM_SA1100_CPUFREQ || ARM_SA1110_CPUFREQ > > - default CPU_FREQ_DEFAULT_GOV_SCHEDUTIL if ARM64 || ARM > > - default CPU_FREQ_DEFAULT_GOV_SCHEDUTIL if X86_INTEL_PSTATE && SMP > > + default CPU_FREQ_DEFAULT_GOV_SCHEDUTIL if SMP > > default CPU_FREQ_DEFAULT_GOV_PERFORMANCE > > help > > This option sets which CPUFreq governor shall be loaded at > > @@ -71,6 +70,7 @@ config CPU_FREQ_DEFAULT_GOV_USERSPACE > > > > config CPU_FREQ_DEFAULT_GOV_ONDEMAND > > bool "ondemand" > > + depends on !SMP > > select CPU_FREQ_GOV_ONDEMAND > > select CPU_FREQ_GOV_PERFORMANCE > > help > > @@ -83,6 +83,7 @@ config CPU_FREQ_DEFAULT_GOV_ONDEMAND > > > > config CPU_FREQ_DEFAULT_GOV_CONSERVATIVE > > bool "conservative" > > + depends on !SMP > > select CPU_FREQ_GOV_CONSERVATIVE > > select CPU_FREQ_GOV_PERFORMANCE > > help > > The changes above should work. > > > @@ -144,6 +145,7 @@ config CPU_FREQ_GOV_USERSPACE > > > > config CPU_FREQ_GOV_ONDEMAND > > tristate "'ondemand' cpufreq policy governor" > > + depends on !SMP > > But I don't think that we can do this and the one below. I have exactly the same comments. > > select CPU_FREQ_GOV_COMMON > > help > > 'ondemand' - This driver adds a dynamic cpufreq policy governor. > > @@ -163,6 +165,7 @@ config CPU_FREQ_GOV_ONDEMAND > > config CPU_FREQ_GOV_CONSERVATIVE > > tristate "'conservative' cpufreq governor" > > depends on CPU_FREQ > > + depends on !SMP > > select CPU_FREQ_GOV_COMMON > > help > > 'conservative' - this driver is rather similar to the 'ondemand' > > > > > -- viresh