diff options
Diffstat (limited to 'arch/powerpc/kernel/smp.c')
-rw-r--r-- | arch/powerpc/kernel/smp.c | 876 |
1 files changed, 626 insertions, 250 deletions
diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c index ea6adbf6a221..12e53b3d7923 100644 --- a/arch/powerpc/kernel/smp.c +++ b/arch/powerpc/kernel/smp.c @@ -33,6 +33,9 @@ #include <linux/processor.h> #include <linux/random.h> #include <linux/stackprotector.h> +#include <linux/pgtable.h> +#include <linux/clockchips.h> +#include <linux/kexec.h> #include <asm/ptrace.h> #include <linux/atomic.h> @@ -41,11 +44,10 @@ #include <asm/kvm_ppc.h> #include <asm/dbell.h> #include <asm/page.h> -#include <asm/pgtable.h> -#include <asm/prom.h> #include <asm/smp.h> #include <asm/time.h> #include <asm/machdep.h> +#include <asm/mmu_context.h> #include <asm/cputhreads.h> #include <asm/cputable.h> #include <asm/mpic.h> @@ -55,10 +57,12 @@ #endif #include <asm/vdso.h> #include <asm/debug.h> -#include <asm/kexec.h> -#include <asm/asm-prototypes.h> #include <asm/cpu_has_feature.h> #include <asm/ftrace.h> +#include <asm/kup.h> +#include <asm/fadump.h> + +#include <trace/events/ipi.h> #ifdef DEBUG #include <asm/udbg.h> @@ -73,12 +77,16 @@ static DEFINE_PER_CPU(int, cpu_state) = { 0 }; #endif struct task_struct *secondary_current; -bool has_big_cores; +bool has_big_cores __ro_after_init; +bool coregroup_enabled __ro_after_init; +bool thread_group_shares_l2 __ro_after_init; +bool thread_group_shares_l3 __ro_after_init; DEFINE_PER_CPU(cpumask_var_t, cpu_sibling_map); DEFINE_PER_CPU(cpumask_var_t, cpu_smallcore_map); DEFINE_PER_CPU(cpumask_var_t, cpu_l2_cache_map); DEFINE_PER_CPU(cpumask_var_t, cpu_core_map); +static DEFINE_PER_CPU(cpumask_var_t, cpu_coregroup_map); EXPORT_PER_CPU_SYMBOL(cpu_sibling_map); EXPORT_PER_CPU_SYMBOL(cpu_l2_cache_map); @@ -87,6 +95,7 @@ EXPORT_SYMBOL_GPL(has_big_cores); #define MAX_THREAD_LIST_SIZE 8 #define THREAD_GROUP_SHARE_L1 1 +#define THREAD_GROUP_SHARE_L2_L3 2 struct thread_groups { unsigned int property; unsigned int nr_groups; @@ -94,11 +103,33 @@ struct thread_groups { unsigned int thread_list[MAX_THREAD_LIST_SIZE]; }; +/* Maximum number of properties that groups of threads within a core can share */ +#define MAX_THREAD_GROUP_PROPERTIES 2 + +struct thread_groups_list { + unsigned int nr_properties; + struct thread_groups property_tgs[MAX_THREAD_GROUP_PROPERTIES]; +}; + +static struct thread_groups_list tgl[NR_CPUS] __initdata; /* - * On big-cores system, cpu_l1_cache_map for each CPU corresponds to + * On big-cores system, thread_group_l1_cache_map for each CPU corresponds to * the set its siblings that share the L1-cache. */ -DEFINE_PER_CPU(cpumask_var_t, cpu_l1_cache_map); +DEFINE_PER_CPU(cpumask_var_t, thread_group_l1_cache_map); + +/* + * On some big-cores system, thread_group_l2_cache_map for each CPU + * corresponds to the set its siblings within the core that share the + * L2-cache. + */ +DEFINE_PER_CPU(cpumask_var_t, thread_group_l2_cache_map); + +/* + * On P10, thread_group_l3_cache_map for each CPU is equal to the + * thread_group_l2_cache_map + */ +DEFINE_PER_CPU(cpumask_var_t, thread_group_l3_cache_map); /* SMP operations for this machine */ struct smp_ops_t *smp_ops; @@ -252,7 +283,7 @@ void smp_muxed_ipi_set_message(int cpu, int msg) * Order previous accesses before accesses in the IPI handler. */ smp_mb(); - message[msg] = 1; + WRITE_ONCE(message[msg], 1); } void smp_muxed_ipi_message_pass(int cpu, int msg) @@ -311,7 +342,7 @@ irqreturn_t smp_ipi_demux_relaxed(void) if (all & IPI_MESSAGE(PPC_MSG_NMI_IPI)) nmi_ipi_action(0, NULL); #endif - } while (info->messages); + } while (READ_ONCE(info->messages)); return IRQ_HANDLED; } @@ -327,12 +358,12 @@ static inline void do_message_pass(int cpu, int msg) #endif } -void smp_send_reschedule(int cpu) +void arch_smp_send_reschedule(int cpu) { if (likely(smp_ops)) do_message_pass(cpu, PPC_MSG_RESCHEDULE); } -EXPORT_SYMBOL_GPL(smp_send_reschedule); +EXPORT_SYMBOL_GPL(arch_smp_send_reschedule); void arch_send_call_function_single_ipi(int cpu) { @@ -374,32 +405,32 @@ static struct cpumask nmi_ipi_pending_mask; static bool nmi_ipi_busy = false; static void (*nmi_ipi_function)(struct pt_regs *) = NULL; -static void nmi_ipi_lock_start(unsigned long *flags) +noinstr static void nmi_ipi_lock_start(unsigned long *flags) { raw_local_irq_save(*flags); hard_irq_disable(); - while (atomic_cmpxchg(&__nmi_ipi_lock, 0, 1) == 1) { + while (raw_atomic_cmpxchg(&__nmi_ipi_lock, 0, 1) == 1) { raw_local_irq_restore(*flags); - spin_until_cond(atomic_read(&__nmi_ipi_lock) == 0); + spin_until_cond(raw_atomic_read(&__nmi_ipi_lock) == 0); raw_local_irq_save(*flags); hard_irq_disable(); } } -static void nmi_ipi_lock(void) +noinstr static void nmi_ipi_lock(void) { - while (atomic_cmpxchg(&__nmi_ipi_lock, 0, 1) == 1) - spin_until_cond(atomic_read(&__nmi_ipi_lock) == 0); + while (raw_atomic_cmpxchg(&__nmi_ipi_lock, 0, 1) == 1) + spin_until_cond(raw_atomic_read(&__nmi_ipi_lock) == 0); } -static void nmi_ipi_unlock(void) +noinstr static void nmi_ipi_unlock(void) { smp_mb(); - WARN_ON(atomic_read(&__nmi_ipi_lock) != 1); - atomic_set(&__nmi_ipi_lock, 0); + WARN_ON(raw_atomic_read(&__nmi_ipi_lock) != 1); + raw_atomic_set(&__nmi_ipi_lock, 0); } -static void nmi_ipi_unlock_end(unsigned long *flags) +noinstr static void nmi_ipi_unlock_end(unsigned long *flags) { nmi_ipi_unlock(); raw_local_irq_restore(*flags); @@ -408,7 +439,7 @@ static void nmi_ipi_unlock_end(unsigned long *flags) /* * Platform NMI handler calls this to ack */ -int smp_handle_nmi_ipi(struct pt_regs *regs) +noinstr int smp_handle_nmi_ipi(struct pt_regs *regs) { void (*fn)(struct pt_regs *) = NULL; unsigned long flags; @@ -546,7 +577,7 @@ void tick_broadcast(const struct cpumask *mask) #endif #ifdef CONFIG_DEBUGGER -void debugger_ipi_callback(struct pt_regs *regs) +static void debugger_ipi_callback(struct pt_regs *regs) { debugger_ipi(regs); } @@ -557,7 +588,7 @@ void smp_send_debugger_break(void) } #endif -#ifdef CONFIG_KEXEC_CORE +#ifdef CONFIG_CRASH_DUMP void crash_send_ipi(void (*crash_ipi_callback)(struct pt_regs *)) { int cpu; @@ -582,12 +613,42 @@ void crash_send_ipi(void (*crash_ipi_callback)(struct pt_regs *)) } #endif +void crash_smp_send_stop(void) +{ + static bool stopped = false; + + /* + * In case of fadump, register data for all CPUs is captured by f/w + * on ibm,os-term rtas call. Skip IPI callbacks to other CPUs before + * this rtas call to avoid tricky post processing of those CPUs' + * backtraces. + */ + if (should_fadump_crash()) + return; + + if (stopped) + return; + + stopped = true; + +#ifdef CONFIG_CRASH_DUMP + if (kexec_crash_image) { + crash_kexec_prepare(); + return; + } +#endif + + smp_send_stop(); +} + #ifdef CONFIG_NMI_IPI static void nmi_stop_this_cpu(struct pt_regs *regs) { /* * IRQs are already hard disabled by the smp_handle_nmi_ipi. */ + set_cpu_online(smp_processor_id(), false); + spin_begin(); while (1) spin_cpu_relax(); @@ -603,6 +664,15 @@ void smp_send_stop(void) static void stop_this_cpu(void *dummy) { hard_irq_disable(); + + /* + * Offlining CPUs in stop_this_cpu can result in scheduler warnings, + * (see commit de6e5d38417e), but printk_safe_flush_on_panic() wants + * to know other CPUs are offline before it breaks locks to flush + * printk buffers, in case we panic()ed while holding the lock. + */ + set_cpu_online(smp_processor_id(), false); + spin_begin(); while (1) spin_cpu_relax(); @@ -627,12 +697,12 @@ void smp_send_stop(void) } #endif /* CONFIG_NMI_IPI */ -struct task_struct *current_set[NR_CPUS]; +static struct task_struct *current_set[NR_CPUS]; static void smp_store_cpu_info(int id) { per_cpu(cpu_pvr, id) = mfspr(SPRN_PVR); -#ifdef CONFIG_PPC_FSL_BOOK3E +#ifdef CONFIG_PPC_E500 per_cpu(next_tlbcam_idx, id) = (mfspr(SPRN_TLB1CFG) & TLBnCFG_N_ENTRY) - 1; #endif @@ -659,83 +729,124 @@ static void set_cpus_unrelated(int i, int j, #endif /* + * Extends set_cpus_related. Instead of setting one CPU at a time in + * dstmask, set srcmask at oneshot. dstmask should be super set of srcmask. + */ +static void or_cpumasks_related(int i, int j, struct cpumask *(*srcmask)(int), + struct cpumask *(*dstmask)(int)) +{ + struct cpumask *mask; + int k; + + mask = srcmask(j); + for_each_cpu(k, srcmask(i)) + cpumask_or(dstmask(k), dstmask(k), mask); + + if (i == j) + return; + + mask = srcmask(i); + for_each_cpu(k, srcmask(j)) + cpumask_or(dstmask(k), dstmask(k), mask); +} + +/* * parse_thread_groups: Parses the "ibm,thread-groups" device tree * property for the CPU device node @dn and stores - * the parsed output in the thread_groups - * structure @tg if the ibm,thread-groups[0] - * matches @property. + * the parsed output in the thread_groups_list + * structure @tglp. * * @dn: The device node of the CPU device. - * @tg: Pointer to a thread group structure into which the parsed + * @tglp: Pointer to a thread group list structure into which the parsed * output of "ibm,thread-groups" is stored. - * @property: The property of the thread-group that the caller is - * interested in. * * ibm,thread-groups[0..N-1] array defines which group of threads in * the CPU-device node can be grouped together based on the property. * - * ibm,thread-groups[0] tells us the property based on which the + * This array can represent thread groupings for multiple properties. + * + * ibm,thread-groups[i + 0] tells us the property based on which the * threads are being grouped together. If this value is 1, it implies - * that the threads in the same group share L1, translation cache. + * that the threads in the same group share L1, translation cache. If + * the value is 2, it implies that the threads in the same group share + * the same L2 cache. * - * ibm,thread-groups[1] tells us how many such thread groups exist. + * ibm,thread-groups[i+1] tells us how many such thread groups exist for the + * property ibm,thread-groups[i] * - * ibm,thread-groups[2] tells us the number of threads in each such + * ibm,thread-groups[i+2] tells us the number of threads in each such * group. + * Suppose k = (ibm,thread-groups[i+1] * ibm,thread-groups[i+2]), then, * - * ibm,thread-groups[3..N-1] is the list of threads identified by + * ibm,thread-groups[i+3..i+k+2] (is the list of threads identified by * "ibm,ppc-interrupt-server#s" arranged as per their membership in * the grouping. * - * Example: If ibm,thread-groups = [1,2,4,5,6,7,8,9,10,11,12] it - * implies that there are 2 groups of 4 threads each, where each group - * of threads share L1, translation cache. + * Example: + * If "ibm,thread-groups" = [1,2,4,8,10,12,14,9,11,13,15,2,2,4,8,10,12,14,9,11,13,15] + * This can be decomposed up into two consecutive arrays: + * a) [1,2,4,8,10,12,14,9,11,13,15] + * b) [2,2,4,8,10,12,14,9,11,13,15] + * + * where in, * - * The "ibm,ppc-interrupt-server#s" of the first group is {5,6,7,8} - * and the "ibm,ppc-interrupt-server#s" of the second group is {9, 10, - * 11, 12} structure + * a) provides information of Property "1" being shared by "2" groups, + * each with "4" threads each. The "ibm,ppc-interrupt-server#s" of + * the first group is {8,10,12,14} and the + * "ibm,ppc-interrupt-server#s" of the second group is + * {9,11,13,15}. Property "1" is indicative of the thread in the + * group sharing L1 cache, translation cache and Instruction Data + * flow. + * + * b) provides information of Property "2" being shared by "2" groups, + * each group with "4" threads. The "ibm,ppc-interrupt-server#s" of + * the first group is {8,10,12,14} and the + * "ibm,ppc-interrupt-server#s" of the second group is + * {9,11,13,15}. Property "2" indicates that the threads in each + * group share the L2-cache. * * Returns 0 on success, -EINVAL if the property does not exist, * -ENODATA if property does not have a value, and -EOVERFLOW if the * property data isn't large enough. */ static int parse_thread_groups(struct device_node *dn, - struct thread_groups *tg, - unsigned int property) + struct thread_groups_list *tglp) { - int i; - u32 thread_group_array[3 + MAX_THREAD_LIST_SIZE]; - u32 *thread_list; + unsigned int property_idx = 0; + u32 *thread_group_array; size_t total_threads; - int ret; + int ret = 0, count; + u32 *thread_list; + int i = 0; + count = of_property_count_u32_elems(dn, "ibm,thread-groups"); + thread_group_array = kcalloc(count, sizeof(u32), GFP_KERNEL); ret = of_property_read_u32_array(dn, "ibm,thread-groups", - thread_group_array, 3); + thread_group_array, count); if (ret) - return ret; - - tg->property = thread_group_array[0]; - tg->nr_groups = thread_group_array[1]; - tg->threads_per_group = thread_group_array[2]; - if (tg->property != property || - tg->nr_groups < 1 || - tg->threads_per_group < 1) - return -ENODATA; + goto out_free; - total_threads = tg->nr_groups * tg->threads_per_group; + while (i < count && property_idx < MAX_THREAD_GROUP_PROPERTIES) { + int j; + struct thread_groups *tg = &tglp->property_tgs[property_idx++]; - ret = of_property_read_u32_array(dn, "ibm,thread-groups", - thread_group_array, - 3 + total_threads); - if (ret) - return ret; + tg->property = thread_group_array[i]; + tg->nr_groups = thread_group_array[i + 1]; + tg->threads_per_group = thread_group_array[i + 2]; + total_threads = tg->nr_groups * tg->threads_per_group; - thread_list = &thread_group_array[3]; + thread_list = &thread_group_array[i + 3]; - for (i = 0 ; i < total_threads; i++) - tg->thread_list[i] = thread_list[i]; + for (j = 0; j < total_threads; j++) + tg->thread_list[j] = thread_list[j]; + i = i + 3 + total_threads; + } - return 0; + tglp->nr_properties = property_idx; + +out_free: + kfree(thread_group_array); + return ret; } /* @@ -746,7 +857,7 @@ static int parse_thread_groups(struct device_node *dn, * @tg : The thread-group structure of the CPU node which @cpu belongs * to. * - * Returns the index to tg->thread_list that points to the the start + * Returns the index to tg->thread_list that points to the start * of the thread_group that @cpu belongs to. * * Returns -1 if cpu doesn't belong to any of the groups pointed to by @@ -771,59 +882,188 @@ static int get_cpu_thread_group_start(int cpu, struct thread_groups *tg) return -1; } -static int init_cpu_l1_cache_map(int cpu) - +static struct thread_groups *__init get_thread_groups(int cpu, + int group_property, + int *err) { struct device_node *dn = of_get_cpu_node(cpu, NULL); - struct thread_groups tg = {.property = 0, - .nr_groups = 0, - .threads_per_group = 0}; - int first_thread = cpu_first_thread_sibling(cpu); - int i, cpu_group_start = -1, err = 0; + struct thread_groups_list *cpu_tgl = &tgl[cpu]; + struct thread_groups *tg = NULL; + int i; + *err = 0; - if (!dn) - return -ENODATA; + if (!dn) { + *err = -ENODATA; + return NULL; + } - err = parse_thread_groups(dn, &tg, THREAD_GROUP_SHARE_L1); - if (err) - goto out; + if (!cpu_tgl->nr_properties) { + *err = parse_thread_groups(dn, cpu_tgl); + if (*err) + goto out; + } - zalloc_cpumask_var_node(&per_cpu(cpu_l1_cache_map, cpu), - GFP_KERNEL, - cpu_to_node(cpu)); + for (i = 0; i < cpu_tgl->nr_properties; i++) { + if (cpu_tgl->property_tgs[i].property == group_property) { + tg = &cpu_tgl->property_tgs[i]; + break; + } + } - cpu_group_start = get_cpu_thread_group_start(cpu, &tg); + if (!tg) + *err = -EINVAL; +out: + of_node_put(dn); + return tg; +} - if (unlikely(cpu_group_start == -1)) { - WARN_ON_ONCE(1); - err = -ENODATA; - goto out; - } +static int __init update_mask_from_threadgroup(cpumask_var_t *mask, struct thread_groups *tg, + int cpu, int cpu_group_start) +{ + int first_thread = cpu_first_thread_sibling(cpu); + int i; + + zalloc_cpumask_var_node(mask, GFP_KERNEL, cpu_to_node(cpu)); for (i = first_thread; i < first_thread + threads_per_core; i++) { - int i_group_start = get_cpu_thread_group_start(i, &tg); + int i_group_start = get_cpu_thread_group_start(i, tg); if (unlikely(i_group_start == -1)) { WARN_ON_ONCE(1); - err = -ENODATA; - goto out; + return -ENODATA; } if (i_group_start == cpu_group_start) - cpumask_set_cpu(i, per_cpu(cpu_l1_cache_map, cpu)); + cpumask_set_cpu(i, *mask); } -out: - of_node_put(dn); - return err; + return 0; +} + +static int __init init_thread_group_cache_map(int cpu, int cache_property) + +{ + int cpu_group_start = -1, err = 0; + struct thread_groups *tg = NULL; + cpumask_var_t *mask = NULL; + + if (cache_property != THREAD_GROUP_SHARE_L1 && + cache_property != THREAD_GROUP_SHARE_L2_L3) + return -EINVAL; + + tg = get_thread_groups(cpu, cache_property, &err); + + if (!tg) + return err; + + cpu_group_start = get_cpu_thread_group_start(cpu, tg); + + if (unlikely(cpu_group_start == -1)) { + WARN_ON_ONCE(1); + return -ENODATA; + } + + if (cache_property == THREAD_GROUP_SHARE_L1) { + mask = &per_cpu(thread_group_l1_cache_map, cpu); + update_mask_from_threadgroup(mask, tg, cpu, cpu_group_start); + } + else if (cache_property == THREAD_GROUP_SHARE_L2_L3) { + mask = &per_cpu(thread_group_l2_cache_map, cpu); + update_mask_from_threadgroup(mask, tg, cpu, cpu_group_start); + mask = &per_cpu(thread_group_l3_cache_map, cpu); + update_mask_from_threadgroup(mask, tg, cpu, cpu_group_start); + } + + + return 0; +} + +static bool shared_caches __ro_after_init; + +#ifdef CONFIG_SCHED_SMT +/* cpumask of CPUs with asymmetric SMT dependency */ +static int powerpc_smt_flags(void) +{ + int flags = SD_SHARE_CPUCAPACITY | SD_SHARE_LLC; + + if (cpu_has_feature(CPU_FTR_ASYM_SMT)) { + printk_once(KERN_INFO "Enabling Asymmetric SMT scheduling\n"); + flags |= SD_ASYM_PACKING; + } + return flags; +} +#endif + +/* + * On shared processor LPARs scheduled on a big core (which has two or more + * independent thread groups per core), prefer lower numbered CPUs, so + * that workload consolidates to lesser number of cores. + */ +static __ro_after_init DEFINE_STATIC_KEY_FALSE(splpar_asym_pack); + +/* + * P9 has a slightly odd architecture where pairs of cores share an L2 cache. + * This topology makes it *much* cheaper to migrate tasks between adjacent cores + * since the migrated task remains cache hot. We want to take advantage of this + * at the scheduler level so an extra topology level is required. + */ +static int powerpc_shared_cache_flags(void) +{ + if (static_branch_unlikely(&splpar_asym_pack)) + return SD_SHARE_LLC | SD_ASYM_PACKING; + + return SD_SHARE_LLC; +} + +static int powerpc_shared_proc_flags(void) +{ + if (static_branch_unlikely(&splpar_asym_pack)) + return SD_ASYM_PACKING; + + return 0; +} + +/* + * We can't just pass cpu_l2_cache_mask() directly because + * returns a non-const pointer and the compiler barfs on that. + */ +static const struct cpumask *shared_cache_mask(int cpu) +{ + return per_cpu(cpu_l2_cache_map, cpu); +} + +#ifdef CONFIG_SCHED_SMT +static const struct cpumask *smallcore_smt_mask(int cpu) +{ + return cpu_smallcore_mask(cpu); +} +#endif + +static struct cpumask *cpu_coregroup_mask(int cpu) +{ + return per_cpu(cpu_coregroup_map, cpu); +} + +static bool has_coregroup_support(void) +{ + /* Coregroup identification not available on shared systems */ + if (is_shared_processor()) + return 0; + + return coregroup_enabled; } -static int init_big_cores(void) +static const struct cpumask *cpu_mc_mask(int cpu) +{ + return cpu_coregroup_mask(cpu); +} + +static int __init init_big_cores(void) { int cpu; for_each_possible_cpu(cpu) { - int err = init_cpu_l1_cache_map(cpu); + int err = init_thread_group_cache_map(cpu, THREAD_GROUP_SHARE_L1); if (err) return err; @@ -834,17 +1074,29 @@ static int init_big_cores(void) } has_big_cores = true; + + for_each_possible_cpu(cpu) { + int err = init_thread_group_cache_map(cpu, THREAD_GROUP_SHARE_L2_L3); + + if (err) + return err; + } + + thread_group_shares_l2 = true; + thread_group_shares_l3 = true; + pr_debug("L2/L3 cache only shared by the threads in the small core\n"); + return 0; } void __init smp_prepare_cpus(unsigned int max_cpus) { - unsigned int cpu; + unsigned int cpu, num_threads; DBG("smp_prepare_cpus\n"); /* - * setup_cpu may need to be called on the boot cpu. We havent + * setup_cpu may need to be called on the boot cpu. We haven't * spun any cpus up but lets be paranoid. */ BUG_ON(boot_cpuid != smp_processor_id()); @@ -860,6 +1112,11 @@ void __init smp_prepare_cpus(unsigned int max_cpus) GFP_KERNEL, cpu_to_node(cpu)); zalloc_cpumask_var_node(&per_cpu(cpu_core_map, cpu), GFP_KERNEL, cpu_to_node(cpu)); + if (has_coregroup_support()) + zalloc_cpumask_var_node(&per_cpu(cpu_coregroup_map, cpu), + GFP_KERNEL, cpu_to_node(cpu)); + +#ifdef CONFIG_NUMA /* * numa_node_id() works after this. */ @@ -868,6 +1125,7 @@ void __init smp_prepare_cpus(unsigned int max_cpus) set_cpu_numa_mem(cpu, local_memory_node(numa_cpu_lookup_table[cpu])); } +#endif } /* Init the cpumasks so the boot CPU is related to itself */ @@ -875,14 +1133,37 @@ void __init smp_prepare_cpus(unsigned int max_cpus) cpumask_set_cpu(boot_cpuid, cpu_l2_cache_mask(boot_cpuid)); cpumask_set_cpu(boot_cpuid, cpu_core_mask(boot_cpuid)); + if (has_coregroup_support()) + cpumask_set_cpu(boot_cpuid, cpu_coregroup_mask(boot_cpuid)); + init_big_cores(); if (has_big_cores) { cpumask_set_cpu(boot_cpuid, cpu_smallcore_mask(boot_cpuid)); } + if (cpu_to_chip_id(boot_cpuid) != -1) { + int idx = DIV_ROUND_UP(num_possible_cpus(), threads_per_core); + + /* + * All threads of a core will all belong to the same core, + * chip_id_lookup_table will have one entry per core. + * Assumption: if boot_cpuid doesn't have a chip-id, then no + * other CPUs, will also not have chip-id. + */ + chip_id_lookup_table = kcalloc(idx, sizeof(int), GFP_KERNEL); + if (chip_id_lookup_table) + memset(chip_id_lookup_table, -1, sizeof(int) * idx); + } + if (smp_ops && smp_ops->probe) smp_ops->probe(); + + // Initalise the generic SMT topology support + num_threads = 1; + if (smt_enabled_at_boot) + num_threads = smt_enabled_at_boot; + cpu_smt_set_num_threads(num_threads, threads_per_core); } void smp_prepare_boot_cpu(void) @@ -980,15 +1261,20 @@ static void cpu_idle_thread_init(unsigned int cpu, struct task_struct *idle) #ifdef CONFIG_PPC64 paca_ptrs[cpu]->__current = idle; paca_ptrs[cpu]->kstack = (unsigned long)task_stack_page(idle) + - THREAD_SIZE - STACK_FRAME_OVERHEAD; + THREAD_SIZE - STACK_FRAME_MIN_SIZE; #endif - idle->cpu = cpu; + task_thread_info(idle)->cpu = cpu; secondary_current = current_set[cpu] = idle; } int __cpu_up(unsigned int cpu, struct task_struct *tidle) { - int rc, c; + const unsigned long boot_spin_ms = 5 * MSEC_PER_SEC; + const bool booting = system_state < SYSTEM_RUNNING; + const unsigned long hp_spin_ms = 1; + unsigned long deadline; + int rc; + const unsigned long spin_wait_ms = booting ? boot_spin_ms : hp_spin_ms; /* * Don't allow secondary threads to come online if inhibited @@ -1033,22 +1319,23 @@ int __cpu_up(unsigned int cpu, struct task_struct *tidle) } /* - * wait to see if the cpu made a callin (is actually up). - * use this value that I found through experimentation. - * -- Cort + * At boot time, simply spin on the callin word until the + * deadline passes. + * + * At run time, spin for an optimistic amount of time to avoid + * sleeping in the common case. */ - if (system_state < SYSTEM_RUNNING) - for (c = 50000; c && !cpu_callin_map[cpu]; c--) - udelay(100); -#ifdef CONFIG_HOTPLUG_CPU - else - /* - * CPUs can take much longer to come up in the - * hotplug case. Wait five seconds. - */ - for (c = 5000; c && !cpu_callin_map[cpu]; c--) - msleep(1); -#endif + deadline = jiffies + msecs_to_jiffies(spin_wait_ms); + spin_until_cond(cpu_callin_map[cpu] || time_is_before_jiffies(deadline)); + + if (!cpu_callin_map[cpu] && system_state >= SYSTEM_RUNNING) { + const unsigned long sleep_interval_us = 10 * USEC_PER_MSEC; + const unsigned long sleep_wait_ms = 100 * MSEC_PER_SEC; + + deadline = jiffies + msecs_to_jiffies(sleep_wait_ms); + while (!cpu_callin_map[cpu] && time_is_after_jiffies(deadline)) + fsleep(sleep_interval_us); + } if (!cpu_callin_map[cpu]) { printk(KERN_ERR "Processor %u is stuck.\n", cpu); @@ -1072,18 +1359,13 @@ int __cpu_up(unsigned int cpu, struct task_struct *tidle) int cpu_to_core_id(int cpu) { struct device_node *np; - const __be32 *reg; int id = -1; np = of_get_cpu_node(cpu, NULL); if (!np) goto out; - reg = of_get_property(np, "reg", NULL); - if (!reg) - goto out; - - id = be32_to_cpup(reg); + id = of_get_cpu_hwid(np, 0); out: of_node_put(np); return id; @@ -1125,26 +1407,68 @@ static struct device_node *cpu_to_l2cache(int cpu) return cache; } -static bool update_mask_by_l2(int cpu, struct cpumask *(*mask_fn)(int)) +static bool update_mask_by_l2(int cpu, cpumask_var_t *mask) { + struct cpumask *(*submask_fn)(int) = cpu_sibling_mask; struct device_node *l2_cache, *np; int i; + if (has_big_cores) + submask_fn = cpu_smallcore_mask; + + /* + * If the threads in a thread-group share L2 cache, then the + * L2-mask can be obtained from thread_group_l2_cache_map. + */ + if (thread_group_shares_l2) { + cpumask_set_cpu(cpu, cpu_l2_cache_mask(cpu)); + + for_each_cpu(i, per_cpu(thread_group_l2_cache_map, cpu)) { + if (cpu_online(i)) + set_cpus_related(i, cpu, cpu_l2_cache_mask); + } + + /* Verify that L1-cache siblings are a subset of L2 cache-siblings */ + if (!cpumask_equal(submask_fn(cpu), cpu_l2_cache_mask(cpu)) && + !cpumask_subset(submask_fn(cpu), cpu_l2_cache_mask(cpu))) { + pr_warn_once("CPU %d : Inconsistent L1 and L2 cache siblings\n", + cpu); + } + + return true; + } + l2_cache = cpu_to_l2cache(cpu); - if (!l2_cache) + if (!l2_cache || !*mask) { + /* Assume only core siblings share cache with this CPU */ + for_each_cpu(i, cpu_sibling_mask(cpu)) + set_cpus_related(cpu, i, cpu_l2_cache_mask); + return false; + } + + cpumask_and(*mask, cpu_online_mask, cpu_cpu_mask(cpu)); + + /* Update l2-cache mask with all the CPUs that are part of submask */ + or_cpumasks_related(cpu, cpu, submask_fn, cpu_l2_cache_mask); + + /* Skip all CPUs already part of current CPU l2-cache mask */ + cpumask_andnot(*mask, *mask, cpu_l2_cache_mask(cpu)); - for_each_cpu(i, cpu_online_mask) { + for_each_cpu(i, *mask) { /* * when updating the marks the current CPU has not been marked * online, but we need to update the cache masks */ np = cpu_to_l2cache(i); - if (!np) - continue; - if (np == l2_cache) - set_cpus_related(cpu, i, mask_fn); + /* Skip all CPUs already part of current CPU l2-cache */ + if (np == l2_cache) { + or_cpumasks_related(cpu, i, submask_fn, cpu_l2_cache_mask); + cpumask_andnot(*mask, *mask, submask_fn(i)); + } else { + cpumask_andnot(*mask, *mask, cpu_l2_cache_mask(i)); + } of_node_put(np); } @@ -1156,89 +1480,159 @@ static bool update_mask_by_l2(int cpu, struct cpumask *(*mask_fn)(int)) #ifdef CONFIG_HOTPLUG_CPU static void remove_cpu_from_masks(int cpu) { + struct cpumask *(*mask_fn)(int) = cpu_sibling_mask; int i; - /* NB: cpu_core_mask is a superset of the others */ - for_each_cpu(i, cpu_core_mask(cpu)) { - set_cpus_unrelated(cpu, i, cpu_core_mask); + unmap_cpu_from_node(cpu); + + if (shared_caches) + mask_fn = cpu_l2_cache_mask; + + for_each_cpu(i, mask_fn(cpu)) { set_cpus_unrelated(cpu, i, cpu_l2_cache_mask); set_cpus_unrelated(cpu, i, cpu_sibling_mask); if (has_big_cores) set_cpus_unrelated(cpu, i, cpu_smallcore_mask); } + + for_each_cpu(i, cpu_core_mask(cpu)) + set_cpus_unrelated(cpu, i, cpu_core_mask); + + if (has_coregroup_support()) { + for_each_cpu(i, cpu_coregroup_mask(cpu)) + set_cpus_unrelated(cpu, i, cpu_coregroup_mask); + } } #endif static inline void add_cpu_to_smallcore_masks(int cpu) { - struct cpumask *this_l1_cache_map = per_cpu(cpu_l1_cache_map, cpu); - int i, first_thread = cpu_first_thread_sibling(cpu); + int i; if (!has_big_cores) return; cpumask_set_cpu(cpu, cpu_smallcore_mask(cpu)); - for (i = first_thread; i < first_thread + threads_per_core; i++) { - if (cpu_online(i) && cpumask_test_cpu(i, this_l1_cache_map)) + for_each_cpu(i, per_cpu(thread_group_l1_cache_map, cpu)) { + if (cpu_online(i)) set_cpus_related(i, cpu, cpu_smallcore_mask); } } +static void update_coregroup_mask(int cpu, cpumask_var_t *mask) +{ + struct cpumask *(*submask_fn)(int) = cpu_sibling_mask; + int coregroup_id = cpu_to_coregroup_id(cpu); + int i; + + if (shared_caches) + submask_fn = cpu_l2_cache_mask; + + if (!*mask) { + /* Assume only siblings are part of this CPU's coregroup */ + for_each_cpu(i, submask_fn(cpu)) + set_cpus_related(cpu, i, cpu_coregroup_mask); + + return; + } + + cpumask_and(*mask, cpu_online_mask, cpu_cpu_mask(cpu)); + + /* Update coregroup mask with all the CPUs that are part of submask */ + or_cpumasks_related(cpu, cpu, submask_fn, cpu_coregroup_mask); + + /* Skip all CPUs already part of coregroup mask */ + cpumask_andnot(*mask, *mask, cpu_coregroup_mask(cpu)); + + for_each_cpu(i, *mask) { + /* Skip all CPUs not part of this coregroup */ + if (coregroup_id == cpu_to_coregroup_id(i)) { + or_cpumasks_related(cpu, i, submask_fn, cpu_coregroup_mask); + cpumask_andnot(*mask, *mask, submask_fn(i)); + } else { + cpumask_andnot(*mask, *mask, cpu_coregroup_mask(i)); + } + } +} + static void add_cpu_to_masks(int cpu) { + struct cpumask *(*submask_fn)(int) = cpu_sibling_mask; int first_thread = cpu_first_thread_sibling(cpu); - int chipid = cpu_to_chip_id(cpu); + cpumask_var_t mask; + int chip_id = -1; + bool ret; int i; /* * This CPU will not be in the online mask yet so we need to manually * add it to it's own thread sibling mask. */ + map_cpu_to_node(cpu, cpu_to_node(cpu)); cpumask_set_cpu(cpu, cpu_sibling_mask(cpu)); + cpumask_set_cpu(cpu, cpu_core_mask(cpu)); for (i = first_thread; i < first_thread + threads_per_core; i++) if (cpu_online(i)) set_cpus_related(i, cpu, cpu_sibling_mask); add_cpu_to_smallcore_masks(cpu); - /* - * Copy the thread sibling mask into the cache sibling mask - * and mark any CPUs that share an L2 with this CPU. - */ - for_each_cpu(i, cpu_sibling_mask(cpu)) - set_cpus_related(cpu, i, cpu_l2_cache_mask); - update_mask_by_l2(cpu, cpu_l2_cache_mask); - /* - * Copy the cache sibling mask into core sibling mask and mark - * any CPUs on the same chip as this CPU. - */ - for_each_cpu(i, cpu_l2_cache_mask(cpu)) - set_cpus_related(cpu, i, cpu_core_mask); + /* In CPU-hotplug path, hence use GFP_ATOMIC */ + ret = alloc_cpumask_var_node(&mask, GFP_ATOMIC, cpu_to_node(cpu)); + update_mask_by_l2(cpu, &mask); - if (chipid == -1) - return; + if (has_coregroup_support()) + update_coregroup_mask(cpu, &mask); - for_each_cpu(i, cpu_online_mask) - if (cpu_to_chip_id(i) == chipid) - set_cpus_related(cpu, i, cpu_core_mask); -} + if (chip_id_lookup_table && ret) + chip_id = cpu_to_chip_id(cpu); + + if (shared_caches) + submask_fn = cpu_l2_cache_mask; + + /* Update core_mask with all the CPUs that are part of submask */ + or_cpumasks_related(cpu, cpu, submask_fn, cpu_core_mask); + + /* Skip all CPUs already part of current CPU core mask */ + cpumask_andnot(mask, cpu_online_mask, cpu_core_mask(cpu)); -static bool shared_caches; + /* If chip_id is -1; limit the cpu_core_mask to within PKG */ + if (chip_id == -1) + cpumask_and(mask, mask, cpu_cpu_mask(cpu)); + + for_each_cpu(i, mask) { + if (chip_id == cpu_to_chip_id(i)) { + or_cpumasks_related(cpu, i, submask_fn, cpu_core_mask); + cpumask_andnot(mask, mask, submask_fn(i)); + } else { + cpumask_andnot(mask, mask, cpu_core_mask(i)); + } + } + + free_cpumask_var(mask); +} /* Activate a secondary processor. */ +__no_stack_protector void start_secondary(void *unused) { - unsigned int cpu = smp_processor_id(); - struct cpumask *(*sibling_mask)(int) = cpu_sibling_mask; + unsigned int cpu = raw_smp_processor_id(); + + /* PPC64 calls setup_kup() in early_setup_secondary() */ + if (IS_ENABLED(CONFIG_PPC32)) + setup_kup(); - mmgrab(&init_mm); + mmgrab_lazy_tlb(&init_mm); current->active_mm = &init_mm; + VM_WARN_ON(cpumask_test_cpu(smp_processor_id(), mm_cpumask(&init_mm))); + cpumask_set_cpu(cpu, mm_cpumask(&init_mm)); + inc_mm_active_cpus(&init_mm); smp_store_cpu_info(cpu); set_dec(tb_ticks_per_jiffy); - preempt_disable(); + rcutree_report_cpu_starting(cpu); cpu_callin_map[cpu] = 1; if (smp_ops->setup_cpu) @@ -1254,20 +1648,26 @@ void start_secondary(void *unused) vdso_getcpu_init(); #endif + set_numa_node(numa_cpu_lookup_table[cpu]); + set_numa_mem(local_memory_node(numa_cpu_lookup_table[cpu])); + /* Update topology CPU masks */ add_cpu_to_masks(cpu); - if (has_big_cores) - sibling_mask = cpu_smallcore_mask; /* * Check for any shared caches. Note that this must be done on a * per-core basis because one core in the pair might be disabled. */ - if (!cpumask_equal(cpu_l2_cache_mask(cpu), sibling_mask(cpu))) - shared_caches = true; + if (!shared_caches) { + struct cpumask *(*sibling_mask)(int) = cpu_sibling_mask; + struct cpumask *mask = cpu_l2_cache_mask(cpu); - set_numa_node(numa_cpu_lookup_table[cpu]); - set_numa_mem(local_memory_node(numa_cpu_lookup_table[cpu])); + if (has_big_cores) + sibling_mask = cpu_smallcore_mask; + + if (cpumask_weight(mask) > cpumask_weight(sibling_mask(cpu))) + shared_caches = true; + } smp_wmb(); notify_cpu_starting(cpu); @@ -1285,68 +1685,46 @@ void start_secondary(void *unused) BUG(); } -int setup_profiling_timer(unsigned int multiplier) -{ - return 0; -} +static struct sched_domain_topology_level powerpc_topology[6]; -#ifdef CONFIG_SCHED_SMT -/* cpumask of CPUs with asymetric SMT dependancy */ -static int powerpc_smt_flags(void) +static void __init build_sched_topology(void) { - int flags = SD_SHARE_CPUCAPACITY | SD_SHARE_PKG_RESOURCES; + int i = 0; - if (cpu_has_feature(CPU_FTR_ASYM_SMT)) { - printk_once(KERN_INFO "Enabling Asymmetric SMT scheduling\n"); - flags |= SD_ASYM_PACKING; - } - return flags; -} -#endif + if (is_shared_processor() && has_big_cores) + static_branch_enable(&splpar_asym_pack); -static struct sched_domain_topology_level powerpc_topology[] = { #ifdef CONFIG_SCHED_SMT - { cpu_smt_mask, powerpc_smt_flags, SD_INIT_NAME(SMT) }, + if (has_big_cores) { + pr_info("Big cores detected but using small core scheduling\n"); + powerpc_topology[i++] = (struct sched_domain_topology_level){ + smallcore_smt_mask, powerpc_smt_flags, SD_INIT_NAME(SMT) + }; + } else { + powerpc_topology[i++] = (struct sched_domain_topology_level){ + cpu_smt_mask, powerpc_smt_flags, SD_INIT_NAME(SMT) + }; + } #endif - { cpu_cpu_mask, SD_INIT_NAME(DIE) }, - { NULL, }, -}; - -/* - * P9 has a slightly odd architecture where pairs of cores share an L2 cache. - * This topology makes it *much* cheaper to migrate tasks between adjacent cores - * since the migrated task remains cache hot. We want to take advantage of this - * at the scheduler level so an extra topology level is required. - */ -static int powerpc_shared_cache_flags(void) -{ - return SD_SHARE_PKG_RESOURCES; -} + if (shared_caches) { + powerpc_topology[i++] = (struct sched_domain_topology_level){ + shared_cache_mask, powerpc_shared_cache_flags, SD_INIT_NAME(CACHE) + }; + } + if (has_coregroup_support()) { + powerpc_topology[i++] = (struct sched_domain_topology_level){ + cpu_mc_mask, powerpc_shared_proc_flags, SD_INIT_NAME(MC) + }; + } + powerpc_topology[i++] = (struct sched_domain_topology_level){ + cpu_cpu_mask, powerpc_shared_proc_flags, SD_INIT_NAME(PKG) + }; -/* - * We can't just pass cpu_l2_cache_mask() directly because - * returns a non-const pointer and the compiler barfs on that. - */ -static const struct cpumask *shared_cache_mask(int cpu) -{ - return cpu_l2_cache_mask(cpu); -} + /* There must be one trailing NULL entry left. */ + BUG_ON(i >= ARRAY_SIZE(powerpc_topology) - 1); -#ifdef CONFIG_SCHED_SMT -static const struct cpumask *smallcore_smt_mask(int cpu) -{ - return cpu_smallcore_mask(cpu); + set_sched_topology(powerpc_topology); } -#endif - -static struct sched_domain_topology_level power9_topology[] = { -#ifdef CONFIG_SCHED_SMT - { cpu_smt_mask, powerpc_smt_flags, SD_INIT_NAME(SMT) }, -#endif - { shared_cache_mask, powerpc_shared_cache_flags, SD_INIT_NAME(CACHE) }, - { cpu_cpu_mask, SD_INIT_NAME(DIE) }, - { NULL, }, -}; void __init smp_cpus_done(unsigned int max_cpus) { @@ -1359,31 +1737,21 @@ void __init smp_cpus_done(unsigned int max_cpus) if (smp_ops && smp_ops->bringup_done) smp_ops->bringup_done(); - /* - * On a shared LPAR, associativity needs to be requested. - * Hence, get numa topology before dumping cpu topology - */ - shared_proc_topology_init(); dump_numa_cpu_topology(); + build_sched_topology(); +} -#ifdef CONFIG_SCHED_SMT - if (has_big_cores) { - pr_info("Using small cores at SMT level\n"); - power9_topology[0].mask = smallcore_smt_mask; - powerpc_topology[0].mask = smallcore_smt_mask; - } -#endif - /* - * If any CPU detects that it's sharing a cache with another CPU then - * use the deeper topology that is aware of this sharing. - */ - if (shared_caches) { - pr_info("Using shared cache scheduler topology\n"); - set_sched_topology(power9_topology); - } else { - pr_info("Using standard scheduler topology\n"); - set_sched_topology(powerpc_topology); - } +/* + * For asym packing, by default lower numbered CPU has higher priority. + * On shared processors, pack to lower numbered core. However avoid moving + * between thread_groups within the same core. + */ +int arch_asym_cpu_priority(int cpu) +{ + if (static_branch_unlikely(&splpar_asym_pack)) + return -cpu / threads_per_core; + + return -cpu; } #ifdef CONFIG_HOTPLUG_CPU @@ -1409,11 +1777,19 @@ int __cpu_disable(void) void __cpu_die(unsigned int cpu) { + /* + * This could perhaps be a generic call in idlea_task_dead(), but + * that requires testing from all archs, so first put it here to + */ + VM_WARN_ON_ONCE(!cpumask_test_cpu(cpu, mm_cpumask(&init_mm))); + dec_mm_active_cpus(&init_mm); + cpumask_clear_cpu(cpu, mm_cpumask(&init_mm)); + if (smp_ops->cpu_die) smp_ops->cpu_die(cpu); } -void cpu_die(void) +void __noreturn arch_cpu_idle_dead(void) { /* * Disable on the down path. This will be re-enabled by @@ -1421,8 +1797,8 @@ void cpu_die(void) */ this_cpu_disable_ftrace(); - if (ppc_md.cpu_die) - ppc_md.cpu_die(); + if (smp_ops->cpu_offline_self) + smp_ops->cpu_offline_self(); /* If we return, we re-enter start_secondary */ start_secondary_resume(); |