From mboxrd@z Thu Jan 1 00:00:00 1970 From: Srinivas Pandruvada Subject: [PATCH 02/11] sched, x86: Add SD_ASYM_PACKING flags to x86 cpu topology for cpus supporting Intel Turbo Boost Max Technology Date: Thu, 18 Aug 2016 15:36:43 -0700 Message-ID: <1471559812-19967-3-git-send-email-srinivas.pandruvada@linux.intel.com> References: <1471559812-19967-1-git-send-email-srinivas.pandruvada@linux.intel.com> Return-path: In-Reply-To: <1471559812-19967-1-git-send-email-srinivas.pandruvada@linux.intel.com> Sender: linux-kernel-owner@vger.kernel.org To: mingo@redhat.com, tglx@linutronix.de, hpa@zytor.com, rjw@rjwysocki.net, peterz@infradead.org Cc: x86@kernel.org, bp@suse.de, sudeep.holla@arm.com, ak@linux.intel.com, linux-acpi@vger.kernel.org, linux-pm@vger.kernel.org, alexey.klimov@arm.com, viresh.kumar@linaro.org, akpm@linux-foundation.org, linux-kernel@vger.kernel.org, lenb@kernel.org, tim.c.chen@linux.intel.com, srinivas.pandruvada@linux.intel.com, paul.gortmaker@windriver.com, jpoimboe@redhat.com, mcgrof@kernel.org, jgross@suse.com, robert.moore@intel.com, dvyukov@google.com, jeyu@redhat.com List-Id: linux-pm@vger.kernel.org From: Tim Chen We uses ASYM_PACKING feature in the scheduler to move tasks to more capable cpus that can be boosted to higher frequency. We mark the sched domain topology level with SD_ASYM_PACKING flag for such systems. Signed-off-by: Tim Chen Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Srinivas Pandruvada --- arch/x86/kernel/smpboot.c | 77 ++++++++++++++++++++++++++++++++++++----------- kernel/sched/core.c | 3 ++ 2 files changed, 62 insertions(+), 18 deletions(-) diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 2a6e84a..255f64e 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -464,31 +464,59 @@ static bool match_die(struct cpuinfo_x86 *c, struct cpuinfo_x86 *o) return false; } -static struct sched_domain_topology_level numa_inside_package_topology[] = { +#ifdef CONFIG_SCHED_ITMT +extern unsigned int sysctl_sched_itmt_enabled; + +static int x86_core_flags(void) +{ + int flags = cpu_core_flags(); + + if (sysctl_sched_itmt_enabled) + flags |= SD_ASYM_PACKING; + + return flags; +} + +static int x86_smt_flags(void) +{ + int flags = cpu_smt_flags(); + + if (sysctl_sched_itmt_enabled) + flags |= SD_ASYM_PACKING; + + return flags; +} +#else +#define x86_core_flags cpu_core_flags +#define x86_smt_flags cpu_smt_flags +#endif + +static struct sched_domain_topology_level x86_topology[] = { +#ifdef CONFIG_SCHED_SMT + { cpu_smt_mask, x86_smt_flags, SD_INIT_NAME(SMT) }, +#endif +#ifdef CONFIG_SCHED_MC + { cpu_coregroup_mask, x86_core_flags, SD_INIT_NAME(MC) }, +#endif + { cpu_cpu_mask, SD_INIT_NAME(DIE) }, + { NULL, }, +}; + +static struct sched_domain_topology_level x86_numa_in_package_topology[] = { #ifdef CONFIG_SCHED_SMT - { cpu_smt_mask, cpu_smt_flags, SD_INIT_NAME(SMT) }, + { cpu_smt_mask, x86_smt_flags, SD_INIT_NAME(SMT) }, #endif #ifdef CONFIG_SCHED_MC - { cpu_coregroup_mask, cpu_core_flags, SD_INIT_NAME(MC) }, + { cpu_coregroup_mask, x86_core_flags, SD_INIT_NAME(MC) }, #endif { NULL, }, }; + /* - * set_sched_topology() sets the topology internal to a CPU. The - * NUMA topologies are layered on top of it to build the full - * system topology. - * - * If NUMA nodes are observed to occur within a CPU package, this - * function should be called. It forces the sched domain code to - * only use the SMT level for the CPU portion of the topology. - * This essentially falls back to relying on NUMA information - * from the SRAT table to describe the entire system topology - * (except for hyperthreads). + * Set if a package/die has multiple NUMA nodes inside. + * AMD Magny-Cours and Intel Cluster-on-Die have this. */ -static void primarily_use_numa_for_topology(void) -{ - set_sched_topology(numa_inside_package_topology); -} +static bool x86_has_numa_in_package = false; void set_cpu_sibling_map(int cpu) { @@ -551,7 +579,7 @@ void set_cpu_sibling_map(int cpu) c->booted_cores = cpu_data(i).booted_cores; } if (match_die(c, o) && !topology_same_node(c, o)) - primarily_use_numa_for_topology(); + x86_has_numa_in_package = true; } threads = cpumask_weight(topology_sibling_cpumask(cpu)); @@ -1297,6 +1325,16 @@ void __init native_smp_prepare_cpus(unsigned int max_cpus) zalloc_cpumask_var(&per_cpu(cpu_core_map, i), GFP_KERNEL); zalloc_cpumask_var(&per_cpu(cpu_llc_shared_map, i), GFP_KERNEL); } + + /* + * Set 'default' x86 topology, this matches default_topology() in that + * it has NUMA nodes as a topology level. See also + * native_smp_cpus_done(). + * + * Must be done before set_cpus_sibling_map() is ran. + */ + set_sched_topology(x86_topology); + set_cpu_sibling_map(0); switch (smp_sanity_check(max_cpus)) { @@ -1363,6 +1401,9 @@ void __init native_smp_cpus_done(unsigned int max_cpus) { pr_debug("Boot done\n"); + if (x86_has_numa_in_package) + set_sched_topology(x86_numa_in_package_topology); + nmi_selftest(); impress_friends(); setup_ioapic_dest(); diff --git a/kernel/sched/core.c b/kernel/sched/core.c index ec752da..342eca9 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -6487,6 +6487,9 @@ static struct sched_domain_topology_level *sched_domain_topology = void set_sched_topology(struct sched_domain_topology_level *tl) { + if (WARN_ON_ONCE(sched_smp_initialized)) + return; + sched_domain_topology = tl; } -- 2.7.4