diff options
Diffstat (limited to 'arch/arm64/kernel')
| -rw-r--r-- | arch/arm64/kernel/Makefile | 2 | ||||
| -rw-r--r-- | arch/arm64/kernel/armv8_deprecated.c | 2 | ||||
| -rw-r--r-- | arch/arm64/kernel/cpu_ops.c | 11 | ||||
| -rw-r--r-- | arch/arm64/kernel/cpufeature.c | 72 | ||||
| -rw-r--r-- | arch/arm64/kernel/cpuidle.c | 9 | ||||
| -rw-r--r-- | arch/arm64/kernel/entry-common.c | 2 | ||||
| -rw-r--r-- | arch/arm64/kernel/head.S | 1 | ||||
| -rw-r--r-- | arch/arm64/kernel/hibernate-asm.S | 2 | ||||
| -rw-r--r-- | arch/arm64/kernel/machine_kexec_file.c | 2 | ||||
| -rw-r--r-- | arch/arm64/kernel/perf_event.c | 338 | ||||
| -rw-r--r-- | arch/arm64/kernel/setup.c | 8 | ||||
| -rw-r--r-- | arch/arm64/kernel/smp.c | 151 | ||||
| -rw-r--r-- | arch/arm64/kernel/topology.c | 180 |
13 files changed, 539 insertions, 241 deletions
diff --git a/arch/arm64/kernel/Makefile b/arch/arm64/kernel/Makefile index fc6488660f64..4e5b8ee31442 100644 --- a/arch/arm64/kernel/Makefile +++ b/arch/arm64/kernel/Makefile @@ -21,7 +21,7 @@ obj-y := debug-monitors.o entry.o irq.o fpsimd.o \ smp.o smp_spin_table.o topology.o smccc-call.o \ syscall.o -extra-$(CONFIG_EFI) := efi-entry.o +targets += efi-entry.o OBJCOPYFLAGS := --prefix-symbols=__efistub_ $(obj)/%.stub.o: $(obj)/%.o FORCE diff --git a/arch/arm64/kernel/armv8_deprecated.c b/arch/arm64/kernel/armv8_deprecated.c index 7832b3216370..4cc581af2d96 100644 --- a/arch/arm64/kernel/armv8_deprecated.c +++ b/arch/arm64/kernel/armv8_deprecated.c @@ -630,7 +630,7 @@ static int __init armv8_deprecated_init(void) register_insn_emulation(&cp15_barrier_ops); if (IS_ENABLED(CONFIG_SETEND_EMULATION)) { - if(system_supports_mixed_endian_el0()) + if (system_supports_mixed_endian_el0()) register_insn_emulation(&setend_ops); else pr_info("setend instruction emulation is not supported on this system\n"); diff --git a/arch/arm64/kernel/cpu_ops.c b/arch/arm64/kernel/cpu_ops.c index 7e07072757af..e133011f64b5 100644 --- a/arch/arm64/kernel/cpu_ops.c +++ b/arch/arm64/kernel/cpu_ops.c @@ -15,10 +15,12 @@ #include <asm/smp_plat.h> extern const struct cpu_operations smp_spin_table_ops; +#ifdef CONFIG_ARM64_ACPI_PARKING_PROTOCOL extern const struct cpu_operations acpi_parking_protocol_ops; +#endif extern const struct cpu_operations cpu_psci_ops; -const struct cpu_operations *cpu_ops[NR_CPUS] __ro_after_init; +static const struct cpu_operations *cpu_ops[NR_CPUS] __ro_after_init; static const struct cpu_operations *const dt_supported_cpu_ops[] __initconst = { &smp_spin_table_ops, @@ -94,7 +96,7 @@ static const char *__init cpu_read_enable_method(int cpu) /* * Read a cpu's enable method and record it in cpu_ops. */ -int __init cpu_read_ops(int cpu) +int __init init_cpu_ops(int cpu) { const char *enable_method = cpu_read_enable_method(cpu); @@ -109,3 +111,8 @@ int __init cpu_read_ops(int cpu) return 0; } + +const struct cpu_operations *get_cpu_ops(int cpu) +{ + return cpu_ops[cpu]; +} diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c index 0b6715625cf6..38ebad880f5c 100644 --- a/arch/arm64/kernel/cpufeature.c +++ b/arch/arm64/kernel/cpufeature.c @@ -163,6 +163,7 @@ static const struct arm64_ftr_bits ftr_id_aa64pfr0[] = { ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_CSV3_SHIFT, 4, 0), ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_CSV2_SHIFT, 4, 0), ARM64_FTR_BITS(FTR_VISIBLE, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR0_DIT_SHIFT, 4, 0), + ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_AMU_SHIFT, 4, 0), ARM64_FTR_BITS(FTR_VISIBLE_IF_IS_ENABLED(CONFIG_ARM64_SVE), FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR0_SVE_SHIFT, 4, 0), ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR0_RAS_SHIFT, 4, 0), @@ -551,7 +552,7 @@ static void __init init_cpu_ftr_reg(u32 sys_reg, u64 new) BUG_ON(!reg); - for (ftrp = reg->ftr_bits; ftrp->width; ftrp++) { + for (ftrp = reg->ftr_bits; ftrp->width; ftrp++) { u64 ftr_mask = arm64_ftr_mask(ftrp); s64 ftr_new = arm64_ftr_value(ftrp, new); @@ -1222,6 +1223,57 @@ static bool has_hw_dbm(const struct arm64_cpu_capabilities *cap, #endif +#ifdef CONFIG_ARM64_AMU_EXTN + +/* + * The "amu_cpus" cpumask only signals that the CPU implementation for the + * flagged CPUs supports the Activity Monitors Unit (AMU) but does not provide + * information regarding all the events that it supports. When a CPU bit is + * set in the cpumask, the user of this feature can only rely on the presence + * of the 4 fixed counters for that CPU. But this does not guarantee that the + * counters are enabled or access to these counters is enabled by code + * executed at higher exception levels (firmware). + */ +static struct cpumask amu_cpus __read_mostly; + +bool cpu_has_amu_feat(int cpu) +{ + return cpumask_test_cpu(cpu, &amu_cpus); +} + +/* Initialize the use of AMU counters for frequency invariance */ +extern void init_cpu_freq_invariance_counters(void); + +static void cpu_amu_enable(struct arm64_cpu_capabilities const *cap) +{ + if (has_cpuid_feature(cap, SCOPE_LOCAL_CPU)) { + pr_info("detected CPU%d: Activity Monitors Unit (AMU)\n", + smp_processor_id()); + cpumask_set_cpu(smp_processor_id(), &amu_cpus); + init_cpu_freq_invariance_counters(); + } +} + +static bool has_amu(const struct arm64_cpu_capabilities *cap, + int __unused) +{ + /* + * The AMU extension is a non-conflicting feature: the kernel can + * safely run a mix of CPUs with and without support for the + * activity monitors extension. Therefore, unconditionally enable + * the capability to allow any late CPU to use the feature. + * + * With this feature unconditionally enabled, the cpu_enable + * function will be called for all CPUs that match the criteria, + * including secondary and hotplugged, marking this feature as + * present on that respective CPU. The enable function will also + * print a detection message. + */ + + return true; +} +#endif + #ifdef CONFIG_ARM64_VHE static bool runs_at_el2(const struct arm64_cpu_capabilities *entry, int __unused) { @@ -1499,6 +1551,24 @@ static const struct arm64_cpu_capabilities arm64_features[] = { .cpu_enable = cpu_clear_disr, }, #endif /* CONFIG_ARM64_RAS_EXTN */ +#ifdef CONFIG_ARM64_AMU_EXTN + { + /* + * The feature is enabled by default if CONFIG_ARM64_AMU_EXTN=y. + * Therefore, don't provide .desc as we don't want the detection + * message to be shown until at least one CPU is detected to + * support the feature. + */ + .capability = ARM64_HAS_AMU_EXTN, + .type = ARM64_CPUCAP_WEAK_LOCAL_CPU_FEATURE, + .matches = has_amu, + .sys_reg = SYS_ID_AA64PFR0_EL1, + .sign = FTR_UNSIGNED, + .field_pos = ID_AA64PFR0_AMU_SHIFT, + .min_field_value = ID_AA64PFR0_AMU, + .cpu_enable = cpu_amu_enable, + }, +#endif /* CONFIG_ARM64_AMU_EXTN */ { .desc = "Data cache clean to the PoU not required for I/D coherence", .capability = ARM64_HAS_CACHE_IDC, diff --git a/arch/arm64/kernel/cpuidle.c b/arch/arm64/kernel/cpuidle.c index e4d6af2fdec7..b512b5503f6e 100644 --- a/arch/arm64/kernel/cpuidle.c +++ b/arch/arm64/kernel/cpuidle.c @@ -18,11 +18,11 @@ int arm_cpuidle_init(unsigned int cpu) { + const struct cpu_operations *ops = get_cpu_ops(cpu); int ret = -EOPNOTSUPP; - if (cpu_ops[cpu] && cpu_ops[cpu]->cpu_suspend && - cpu_ops[cpu]->cpu_init_idle) - ret = cpu_ops[cpu]->cpu_init_idle(cpu); + if (ops && ops->cpu_suspend && ops->cpu_init_idle) + ret = ops->cpu_init_idle(cpu); return ret; } @@ -37,8 +37,9 @@ int arm_cpuidle_init(unsigned int cpu) int arm_cpuidle_suspend(int index) { int cpu = smp_processor_id(); + const struct cpu_operations *ops = get_cpu_ops(cpu); - return cpu_ops[cpu]->cpu_suspend(index); + return ops->cpu_suspend(index); } #ifdef CONFIG_ACPI diff --git a/arch/arm64/kernel/entry-common.c b/arch/arm64/kernel/entry-common.c index fde59981445c..c839b5bf1904 100644 --- a/arch/arm64/kernel/entry-common.c +++ b/arch/arm64/kernel/entry-common.c @@ -175,7 +175,7 @@ NOKPROBE_SYMBOL(el0_pc); static void notrace el0_sp(struct pt_regs *regs, unsigned long esr) { user_exit_irqoff(); - local_daif_restore(DAIF_PROCCTX_NOIRQ); + local_daif_restore(DAIF_PROCCTX); do_sp_pc_abort(regs->sp, esr, regs); } NOKPROBE_SYMBOL(el0_sp); diff --git a/arch/arm64/kernel/head.S b/arch/arm64/kernel/head.S index 989b1944cb71..f79023c9b374 100644 --- a/arch/arm64/kernel/head.S +++ b/arch/arm64/kernel/head.S @@ -404,7 +404,6 @@ __create_page_tables: ret x28 ENDPROC(__create_page_tables) - .ltorg /* * The following fragment of code is executed with the MMU enabled. diff --git a/arch/arm64/kernel/hibernate-asm.S b/arch/arm64/kernel/hibernate-asm.S index 38bcd4d4e43b..6532105b3e32 100644 --- a/arch/arm64/kernel/hibernate-asm.S +++ b/arch/arm64/kernel/hibernate-asm.S @@ -110,8 +110,6 @@ ENTRY(swsusp_arch_suspend_exit) cbz x24, 3f /* Do we need to re-initialise EL2? */ hvc #0 3: ret - - .ltorg ENDPROC(swsusp_arch_suspend_exit) /* diff --git a/arch/arm64/kernel/machine_kexec_file.c b/arch/arm64/kernel/machine_kexec_file.c index dd3ae8081b38..b40c3b0def92 100644 --- a/arch/arm64/kernel/machine_kexec_file.c +++ b/arch/arm64/kernel/machine_kexec_file.c @@ -121,7 +121,7 @@ static int setup_dtb(struct kimage *image, /* add kaslr-seed */ ret = fdt_delprop(dtb, off, FDT_PROP_KASLR_SEED); - if (ret == -FDT_ERR_NOTFOUND) + if (ret == -FDT_ERR_NOTFOUND) ret = 0; else if (ret) goto out; diff --git a/arch/arm64/kernel/perf_event.c b/arch/arm64/kernel/perf_event.c index e40b65645c86..4d7879484cec 100644 --- a/arch/arm64/kernel/perf_event.c +++ b/arch/arm64/kernel/perf_event.c @@ -285,6 +285,17 @@ static struct attribute_group armv8_pmuv3_format_attr_group = { #define ARMV8_IDX_COUNTER_LAST(cpu_pmu) \ (ARMV8_IDX_CYCLE_COUNTER + cpu_pmu->num_events - 1) + +/* + * We unconditionally enable ARMv8.5-PMU long event counter support + * (64-bit events) where supported. Indicate if this arm_pmu has long + * event counter support. + */ +static bool armv8pmu_has_long_event(struct arm_pmu *cpu_pmu) +{ + return (cpu_pmu->pmuver >= ID_AA64DFR0_PMUVER_8_5); +} + /* * We must chain two programmable counters for 64 bit events, * except when we have allocated the 64bit cycle counter (for CPU @@ -294,9 +305,11 @@ static struct attribute_group armv8_pmuv3_format_attr_group = { static inline bool armv8pmu_event_is_chained(struct perf_event *event) { int idx = event->hw.idx; + struct arm_pmu *cpu_pmu = to_arm_pmu(event->pmu); return !WARN_ON(idx < 0) && armv8pmu_event_is_64bit(event) && + !armv8pmu_has_long_event(cpu_pmu) && (idx != ARMV8_IDX_CYCLE_COUNTER); } @@ -345,7 +358,7 @@ static inline void armv8pmu_select_counter(int idx) isb(); } -static inline u32 armv8pmu_read_evcntr(int idx) +static inline u64 armv8pmu_read_evcntr(int idx) { armv8pmu_select_counter(idx); return read_sysreg(pmxevcntr_el0); @@ -362,6 +375,44 @@ static inline u64 armv8pmu_read_hw_counter(struct perf_event *event) return val; } +/* + * The cycle counter is always a 64-bit counter. When ARMV8_PMU_PMCR_LP + * is set the event counters also become 64-bit counters. Unless the + * user has requested a long counter (attr.config1) then we want to + * interrupt upon 32-bit overflow - we achieve this by applying a bias. + */ +static bool armv8pmu_event_needs_bias(struct perf_event *event) +{ + struct arm_pmu *cpu_pmu = to_arm_pmu(event->pmu); + struct hw_perf_event *hwc = &event->hw; + int idx = hwc->idx; + + if (armv8pmu_event_is_64bit(event)) + return false; + + if (armv8pmu_has_long_event(cpu_pmu) || + idx == ARMV8_IDX_CYCLE_COUNTER) + return true; + + return false; +} + +static u64 armv8pmu_bias_long_counter(struct perf_event *event, u64 value) +{ + if (armv8pmu_event_needs_bias(event)) + value |= GENMASK(63, 32); + + return value; +} + +static u64 armv8pmu_unbias_long_counter(struct perf_event *event, u64 value) +{ + if (armv8pmu_event_needs_bias(event)) + value &= ~GENMASK(63, 32); + + return value; +} + static u64 armv8pmu_read_counter(struct perf_event *event) { struct arm_pmu *cpu_pmu = to_arm_pmu(event->pmu); @@ -377,10 +428,10 @@ static u64 armv8pmu_read_counter(struct perf_event *event) else value = armv8pmu_read_hw_counter(event); - return value; + return armv8pmu_unbias_long_counter(event, value); } -static inline void armv8pmu_write_evcntr(int idx, u32 value) +static inline void armv8pmu_write_evcntr(int idx, u64 value) { armv8pmu_select_counter(idx); write_sysreg(value, pmxevcntr_el0); @@ -405,20 +456,14 @@ static void armv8pmu_write_counter(struct perf_event *event, u64 value) struct hw_perf_event *hwc = &event->hw; int idx = hwc->idx; + value = armv8pmu_bias_long_counter(event, value); + if (!armv8pmu_counter_valid(cpu_pmu, idx)) pr_err("CPU%u writing wrong counter %d\n", smp_processor_id(), idx); - else if (idx == ARMV8_IDX_CYCLE_COUNTER) { - /* - * The cycles counter is really a 64-bit counter. - * When treating it as a 32-bit counter, we only count - * the lower 32 bits, and set the upper 32-bits so that - * we get an interrupt upon 32-bit overflow. - */ - if (!armv8pmu_event_is_64bit(event)) - value |= 0xffffffff00000000ULL; + else if (idx == ARMV8_IDX_CYCLE_COUNTER) write_sysreg(value, pmccntr_el0); - } else + else armv8pmu_write_hw_counter(event, value); } @@ -450,86 +495,74 @@ static inline void armv8pmu_write_event_type(struct perf_event *event) } } -static inline int armv8pmu_enable_counter(int idx) +static u32 armv8pmu_event_cnten_mask(struct perf_event *event) { - u32 counter = ARMV8_IDX_TO_COUNTER(idx); - write_sysreg(BIT(counter), pmcntenset_el0); - return idx; + int counter = ARMV8_IDX_TO_COUNTER(event->hw.idx); + u32 mask = BIT(counter); + + if (armv8pmu_event_is_chained(event)) + mask |= BIT(counter - 1); + return mask; +} + +static inline void armv8pmu_enable_counter(u32 mask) +{ + write_sysreg(mask, pmcntenset_el0); } static inline void armv8pmu_enable_event_counter(struct perf_event *event) { struct perf_event_attr *attr = &event->attr; - int idx = event->hw.idx; - u32 counter_bits = BIT(ARMV8_IDX_TO_COUNTER(idx)); + u32 mask = armv8pmu_event_cnten_mask(event); - if (armv8pmu_event_is_chained(event)) - counter_bits |= BIT(ARMV8_IDX_TO_COUNTER(idx - 1)); - - kvm_set_pmu_events(counter_bits, attr); + kvm_set_pmu_events(mask, attr); /* We rely on the hypervisor switch code to enable guest counters */ - if (!kvm_pmu_counter_deferred(attr)) { - armv8pmu_enable_counter(idx); - if (armv8pmu_event_is_chained(event)) - armv8pmu_enable_counter(idx - 1); - } + if (!kvm_pmu_counter_deferred(attr)) + armv8pmu_enable_counter(mask); } -static inline int armv8pmu_disable_counter(int idx) +static inline void armv8pmu_disable_counter(u32 mask) { - u32 counter = ARMV8_IDX_TO_COUNTER(idx); - write_sysreg(BIT(counter), pmcntenclr_el0); - return idx; + write_sysreg(mask, pmcntenclr_el0); } static inline void armv8pmu_disable_event_counter(struct perf_event *event) { - struct hw_perf_event *hwc = &event->hw; struct perf_event_attr *attr = &event->attr; - int idx = hwc->idx; - u32 counter_bits = BIT(ARMV8_IDX_TO_COUNTER(idx)); - - if (armv8pmu_event_is_chained(event)) - counter_bits |= BIT(ARMV8_IDX_TO_COUNTER(idx - 1)); + u32 mask = armv8pmu_event_cnten_mask(event); - kvm_clr_pmu_events(counter_bits); + kvm_clr_pmu_events(mask); /* We rely on the hypervisor switch code to disable guest counters */ - if (!kvm_pmu_counter_deferred(attr)) { - if (armv8pmu_event_is_chained(event)) - armv8pmu_disable_counter(idx - 1); - armv8pmu_disable_counter(idx); - } + if (!kvm_pmu_counter_deferred(attr)) + armv8pmu_disable_counter(mask); } -static inline int armv8pmu_enable_intens(int idx) +static inline void armv8pmu_enable_intens(u32 mask) { - u32 counter = ARMV8_IDX_TO_COUNTER(idx); - write_sysreg(BIT(counter), pmintenset_el1); - return idx; + write_sysreg(mask, pmintenset_el1); } -static inline int armv8pmu_enable_event_irq(struct perf_event *event) +static inline void armv8pmu_enable_event_irq(struct perf_event *event) { - return armv8pmu_enable_intens(event->hw.idx); + u32 counter = ARMV8_IDX_TO_COUNTER(event->hw.idx); + armv8pmu_enable_intens(BIT(counter)); } -static inline int armv8pmu_disable_intens(int idx) +static inline void armv8pmu_disable_intens(u32 mask) { - u32 counter = ARMV8_IDX_TO_COUNTER(idx); - write_sysreg(BIT(counter), pmintenclr_el1); + write_sysreg(mask, pmintenclr_el1); isb(); /* Clear the overflow flag in case an interrupt is pending. */ - write_sysreg(BIT(counter), pmovsclr_el0); + write_sysreg(mask, pmovsclr_el0); isb(); - - return idx; } -static inline int armv8pmu_disable_event_irq(struct perf_event *event) +static inline void armv8pmu_disable_event_irq(struct perf_event *event) { - return armv8pmu_disable_intens(event->hw.idx); + u32 counter = ARMV8_IDX_TO_COUNTER(event->hw.idx); + armv8pmu_disable_intens(BIT(counter)); } static inline u32 armv8pmu_getreset_flags(void) @@ -743,7 +776,8 @@ static int armv8pmu_get_event_idx(struct pmu_hw_events *cpuc, /* * Otherwise use events counters */ - if (armv8pmu_event_is_64bit(event)) + if (armv8pmu_event_is_64bit(event) && + !armv8pmu_has_long_event(cpu_pmu)) return armv8pmu_get_chain_idx(cpuc, cpu_pmu); else return armv8pmu_get_single_idx(cpuc, cpu_pmu); @@ -815,13 +849,11 @@ static int armv8pmu_filter_match(struct perf_event *event) static void armv8pmu_reset(void *info) { struct arm_pmu *cpu_pmu = (struct arm_pmu *)info; - u32 idx, nb_cnt = cpu_pmu->num_events; + u32 pmcr; /* The counter and interrupt enable registers are unknown at reset. */ - for (idx = ARMV8_IDX_CYCLE_COUNTER; idx < nb_cnt; ++idx) { - armv8pmu_disable_counter(idx); - armv8pmu_disable_intens(idx); - } + armv8pmu_disable_counter(U32_MAX); + armv8pmu_disable_intens(U32_MAX); /* Clear the counters we flip at guest entry/exit */ kvm_clr_pmu_events(U32_MAX); @@ -830,8 +862,13 @@ static void armv8pmu_reset(void *info) * Initialize & Reset PMNC. Request overflow interrupt for * 64 bit cycle counter but cheat in armv8pmu_write_counter(). */ - armv8pmu_pmcr_write(ARMV8_PMU_PMCR_P | ARMV8_PMU_PMCR_C | - ARMV8_PMU_PMCR_LC); + pmcr = ARMV8_PMU_PMCR_P | ARMV8_PMU_PMCR_C | ARMV8_PMU_PMCR_LC; + + /* Enable long event counter support where available */ + if (armv8pmu_has_long_event(cpu_pmu)) + pmcr |= ARMV8_PMU_PMCR_LP; + + armv8pmu_pmcr_write(pmcr); } static int __armv8_pmuv3_map_event(struct perf_event *event, @@ -914,6 +951,7 @@ static void __armv8pmu_probe_pmu(void *info) if (pmuver == 0xf || pmuver == 0) return; + cpu_pmu->pmuver = pmuver; probe->present = true; /* Read the nb of CNTx counters supported from PMNC */ @@ -953,7 +991,10 @@ static int armv8pmu_probe_pmu(struct arm_pmu *cpu_pmu) return probe.present ? 0 : -ENODEV; } -static int armv8_pmu_init(struct arm_pmu *cpu_pmu) +static int armv8_pmu_init(struct arm_pmu *cpu_pmu, char *name, + int (*map_event)(struct perf_event *event), + const struct attribute_group *events, + const struct attribute_group *format) { int ret = armv8pmu_probe_pmu(cpu_pmu); if (ret) @@ -972,144 +1013,127 @@ static int armv8_pmu_init(struct arm_pmu *cpu_pmu) cpu_pmu->set_event_filter = armv8pmu_set_event_filter; cpu_pmu->filter_match = armv8pmu_filter_match; + cpu_pmu->name = name; + cpu_pmu->map_event = map_event; + cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_EVENTS] = events ? + events : &armv8_pmuv3_events_attr_group; + cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_FORMATS] = format ? + format : &armv8_pmuv3_format_attr_group; + return 0; } static int armv8_pmuv3_init(struct arm_pmu *cpu_pmu) { - int ret = armv8_pmu_init(cpu_pmu); - if (ret) - return ret; - - cpu_pmu->name = "armv8_pmuv3"; - cpu_pmu->map_event = armv8_pmuv3_map_event; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_EVENTS] = - &armv8_pmuv3_events_attr_group; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_FORMATS] = - &armv8_pmuv3_format_attr_group; + return armv8_pmu_init(cpu_pmu, "armv8_pmuv3", + armv8_pmuv3_map_event, NULL, NULL); +} - return 0; +static int armv8_a34_pmu_init(struct arm_pmu *cpu_pmu) +{ + return armv8_pmu_init(cpu_pmu, "armv8_cortex_a34", + armv8_pmuv3_map_event, NULL, NULL); } static int armv8_a35_pmu_init(struct arm_pmu *cpu_pmu) { - int ret = armv8_pmu_init(cpu_pmu); - if (ret) - return ret; - - cpu_pmu->name = "armv8_cortex_a35"; - cpu_pmu->map_event = armv8_a53_map_event; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_EVENTS] = - &armv8_pmuv3_events_attr_group; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_FORMATS] = - &armv8_pmuv3_format_attr_group; - - return 0; + return armv8_pmu_init(cpu_pmu, "armv8_cortex_a35", + armv8_a53_map_event, NULL, NULL); } static int armv8_a53_pmu_init(struct arm_pmu *cpu_pmu) { - int ret = armv8_pmu_init(cpu_pmu); - if (ret) - return ret; - - cpu_pmu->name = "armv8_cortex_a53"; - cpu_pmu->map_event = armv8_a53_map_event; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_EVENTS] = - &armv8_pmuv3_events_attr_group; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_FORMATS] = - &armv8_pmuv3_format_attr_group; + return armv8_pmu_init(cpu_pmu, "armv8_cortex_a53", + armv8_a53_map_event, NULL, NULL); +} - return 0; +static int armv8_a55_pmu_init(struct arm_pmu *cpu_pmu) +{ + return armv8_pmu_init(cpu_pmu, "armv8_cortex_a55", + armv8_pmuv3_map_event, NULL, NULL); } static int armv8_a57_pmu_init(struct arm_pmu *cpu_pmu) { - int ret = armv8_pmu_init(cpu_pmu); - if (ret) - return ret; - - cpu_pmu->name = "armv8_cortex_a57"; - cpu_pmu->map_event = armv8_a57_map_event; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_EVENTS] = - &armv8_pmuv3_events_attr_group; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_FORMATS] = - &armv8_pmuv3_format_attr_group; + return armv8_pmu_init(cpu_pmu, "armv8_cortex_a57", + armv8_a57_map_event, NULL, NULL); +} - return 0; +static int armv8_a65_pmu_init(struct arm_pmu *cpu_pmu) +{ + return armv8_pmu_init(cpu_pmu, "armv8_cortex_a65", + armv8_pmuv3_map_event, NULL, NULL); } static int armv8_a72_pmu_init(struct arm_pmu *cpu_pmu) { - int ret = armv8_pmu_init(cpu_pmu); - if (ret) - return ret; - - cpu_pmu->name = "armv8_cortex_a72"; - cpu_pmu->map_event = armv8_a57_map_event; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_EVENTS] = - &armv8_pmuv3_events_attr_group; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_FORMATS] = - &armv8_pmuv3_format_attr_group; - - return 0; + return armv8_pmu_init(cpu_pmu, "armv8_cortex_a72", + armv8_a57_map_event, NULL, NULL); } static int armv8_a73_pmu_init(struct arm_pmu *cpu_pmu) { - int ret = armv8_pmu_init(cpu_pmu); - if (ret) - return ret; - - cpu_pmu->name = "armv8_cortex_a73"; - cpu_pmu->map_event = armv8_a73_map_event; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_EVENTS] = - &armv8_pmuv3_events_attr_group; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_FORMATS] = - &armv8_pmuv3_format_attr_group; + return armv8_pmu_init(cpu_pmu, "armv8_cortex_a73", + armv8_a73_map_event, NULL, NULL); +} - return 0; +static int armv8_a75_pmu_init(struct arm_pmu *cpu_pmu) +{ + return armv8_pmu_init(cpu_pmu, "armv8_cortex_a75", + armv8_pmuv3_map_event, NULL, NULL); } -static int armv8_thunder_pmu_init(struct arm_pmu *cpu_pmu) +static int armv8_a76_pmu_init(struct arm_pmu *cpu_pmu) { - int ret = armv8_pmu_init(cpu_pmu); - if (ret) - return ret; + return armv8_pmu_init(cpu_pmu, "armv8_cortex_a76", + armv8_pmuv3_map_event, NULL, NULL); +} - cpu_pmu->name = "armv8_cavium_thunder"; - cpu_pmu->map_event = armv8_thunder_map_event; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_EVENTS] = - &armv8_pmuv3_events_attr_group; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_FORMATS] = - &armv8_pmuv3_format_attr_group; +static int armv8_a77_pmu_init(struct arm_pmu *cpu_pmu) +{ + return armv8_pmu_init(cpu_pmu, "armv8_cortex_a77", + armv8_pmuv3_map_event, NULL, NULL); +} - return 0; +static int armv8_e1_pmu_init(struct arm_pmu *cpu_pmu) +{ + return armv8_pmu_init(cpu_pmu, "armv8_neoverse_e1", + armv8_pmuv3_map_event, NULL, NULL); } -static int armv8_vulcan_pmu_init(struct arm_pmu *cpu_pmu) +static int armv8_n1_pmu_init(struct arm_pmu *cpu_pmu) { - int ret = armv8_pmu_init(cpu_pmu); - if (ret) - return ret; + return armv8_pmu_init(cpu_pmu, "armv8_neoverse_n1", + armv8_pmuv3_map_event, NULL, NULL); +} - cpu_pmu->name = "armv8_brcm_vulcan"; - cpu_pmu->map_event = armv8_vulcan_map_event; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_EVENTS] = - &armv8_pmuv3_events_attr_group; - cpu_pmu->attr_groups[ARMPMU_ATTR_GROUP_FORMATS] = - &armv8_pmuv3_format_attr_group; +static int armv8_thunder_pmu_init(struct arm_pmu *cpu_pmu) +{ + return armv8_pmu_init(cpu_pmu, "armv8_cavium_thunder", + armv8_thunder_map_event, NULL, NULL); +} - return 0; +static int armv8_vulcan_pmu_init(struct arm_pmu *cpu_pmu) +{ + return armv8_pmu_init(cpu_pmu, "armv8_brcm_vulcan", + armv8_vulcan_map_event, NULL, NULL); } static const struct of_device_id armv8_pmu_of_device_ids[] = { {.compatible = "arm,armv8-pmuv3", .data = armv8_pmuv3_init}, + {.compatible = "arm,cortex-a34-pmu", .data = armv8_a34_pmu_init}, {.compatible = "arm,cortex-a35-pmu", .data = armv8_a35_pmu_init}, {.compatible = "arm,cortex-a53-pmu", .data = armv8_a53_pmu_init}, + {.compatible = "arm,cortex-a55-pmu", .data = armv8_a55_pmu_init}, {.compatible = "arm,cortex-a57-pmu", .data = armv8_a57_pmu_init}, + {.compatible = "arm,cortex-a65-pmu", .data = armv8_a65_pmu_init}, {.compatible = "arm,cortex-a72-pmu", .data = armv8_a72_pmu_init}, {.compatible = "arm,cortex-a73-pmu", .data = armv8_a73_pmu_init}, + {.compatible = "arm,cortex-a75-pmu", .data = armv8_a75_pmu_init}, + {.compatible = "arm,cortex-a76-pmu", .data = armv8_a76_pmu_init}, + {.compatible = "arm,cortex-a77-pmu", .data = armv8_a77_pmu_init}, + {.compatible = "arm,neoverse-e1-pmu", .data = armv8_e1_pmu_init}, + {.compatible = "arm,neoverse-n1-pmu", .data = armv8_n1_pmu_init}, {.compatible = "cavium,thunder-pmu", .data = armv8_thunder_pmu_init}, {.compatible = "brcm,vulcan-pmu", .data = armv8_vulcan_pmu_init}, {}, diff --git a/arch/arm64/kernel/setup.c b/arch/arm64/kernel/setup.c index a34890bf309f..3fd2c11c09fc 100644 --- a/arch/arm64/kernel/setup.c +++ b/arch/arm64/kernel/setup.c @@ -344,7 +344,7 @@ void __init setup_arch(char **cmdline_p) else psci_acpi_init(); - cpu_read_bootcpu_ops(); + init_bootcpu_ops(); smp_init_cpus(); smp_build_mpidr_hash(); @@ -371,8 +371,10 @@ void __init setup_arch(char **cmdline_p) static inline bool cpu_can_disable(unsigned int cpu) { #ifdef CONFIG_HOTPLUG_CPU - if (cpu_ops[cpu] && cpu_ops[cpu]->cpu_can_disable) - return cpu_ops[cpu]->cpu_can_disable(cpu); + const struct cpu_operations *ops = get_cpu_ops(cpu); + + if (ops && ops->cpu_can_disable) + return ops->cpu_can_disable(cpu); #endif return false; } diff --git a/arch/arm64/kernel/smp.c b/arch/arm64/kernel/smp.c index d4ed9a19d8fe..034806725598 100644 --- a/arch/arm64/kernel/smp.c +++ b/arch/arm64/kernel/smp.c @@ -93,8 +93,10 @@ static inline int op_cpu_kill(unsigned int cpu) */ static int boot_secondary(unsigned int cpu, struct task_struct *idle) { - if (cpu_ops[cpu]->cpu_boot) - return cpu_ops[cpu]->cpu_boot(cpu); + const struct cpu_operations *ops = get_cpu_ops(cpu); + + if (ops->cpu_boot) + return ops->cpu_boot(cpu); return -EOPNOTSUPP; } @@ -115,60 +117,55 @@ int __cpu_up(unsigned int cpu, struct task_struct *idle) update_cpu_boot_status(CPU_MMU_OFF); __flush_dcache_area(&secondary_data, sizeof(secondary_data)); - /* - * Now bring the CPU into our world. - */ + /* Now bring the CPU into our world */ ret = boot_secondary(cpu, idle); - if (ret == 0) { - /* - * CPU was successfully started, wait for it to come online or - * time out. - */ - wait_for_completion_timeout(&cpu_running, - msecs_to_jiffies(5000)); - - if (!cpu_online(cpu)) { - pr_crit("CPU%u: failed to come online\n", cpu); - ret = -EIO; - } - } else { + if (ret) { pr_err("CPU%u: failed to boot: %d\n", cpu, ret); return ret; } + /* + * CPU was successfully started, wait for it to come online or + * time out. + */ + wait_for_completion_timeout(&cpu_running, + msecs_to_jiffies(5000)); + if (cpu_online(cpu)) + return 0; + + pr_crit("CPU%u: failed to come online\n", cpu); secondary_data.task = NULL; secondary_data.stack = NULL; __flush_dcache_area(&secondary_data, sizeof(secondary_data)); status = READ_ONCE(secondary_data.status); - if (ret && status) { + if (status == CPU_MMU_OFF) + status = READ_ONCE(__early_cpu_boot_status); - if (status == CPU_MMU_OFF) - status = READ_ONCE(__early_cpu_boot_status); - - switch (status & CPU_BOOT_STATUS_MASK) { - default: - pr_err("CPU%u: failed in unknown state : 0x%lx\n", - cpu, status); - cpus_stuck_in_kernel++; - break; - case CPU_KILL_ME: - if (!op_cpu_kill(cpu)) { - pr_crit("CPU%u: died during early boot\n", cpu); - break; - } - pr_crit("CPU%u: may not have shut down cleanly\n", cpu); - /* Fall through */ - case CPU_STUCK_IN_KERNEL: - pr_crit("CPU%u: is stuck in kernel\n", cpu); - if (status & CPU_STUCK_REASON_52_BIT_VA) - pr_crit("CPU%u: does not support 52-bit VAs\n", cpu); - if (status & CPU_STUCK_REASON_NO_GRAN) - pr_crit("CPU%u: does not support %luK granule \n", cpu, PAGE_SIZE / SZ_1K); - cpus_stuck_in_kernel++; + switch (status & CPU_BOOT_STATUS_MASK) { + default: + pr_err("CPU%u: failed in unknown state : 0x%lx\n", + cpu, status); + cpus_stuck_in_kernel++; + break; + case CPU_KILL_ME: + if (!op_cpu_kill(cpu)) { + pr_crit("CPU%u: died during early boot\n", cpu); break; - case CPU_PANIC_KERNEL: - panic("CPU%u detected unsupported configuration\n", cpu); } + pr_crit("CPU%u: may not have shut down cleanly\n", cpu); + /* Fall through */ + case CPU_STUCK_IN_KERNEL: + pr_crit("CPU%u: is stuck in kernel\n", cpu); + if (status & CPU_STUCK_REASON_52_BIT_VA) + pr_crit("CPU%u: does not support 52-bit VAs\n", cpu); + if (status & CPU_STUCK_REASON_NO_GRAN) { + pr_crit("CPU%u: does not support %luK granule\n", + cpu, PAGE_SIZE / SZ_1K); + } + cpus_stuck_in_kernel++; + break; + case CPU_PANIC_KERNEL: + panic("CPU%u detected unsupported configuration\n", cpu); } return ret; @@ -196,6 +193,7 @@ asmlinkage notrace void secondary_start_kernel(void) { u64 mpidr = read_cpuid_mpidr() & MPIDR_HWID_BITMASK; struct mm_struct *mm = &init_mm; + const struct cpu_operations *ops; unsigned int cpu; cpu = task_cpu(current); @@ -227,8 +225,9 @@ asmlinkage notrace void secondary_start_kernel(void) */ check_local_cpu_capabilities(); - if (cpu_ops[cpu]->cpu_postboot) - cpu_ops[cpu]->cpu_postboot(); + ops = get_cpu_ops(cpu); + if (ops->cpu_postboot) + ops->cpu_postboot(); /* * Log the CPU info before it is marked online and might get read. @@ -266,19 +265,21 @@ asmlinkage notrace void secondary_start_kernel(void) #ifdef CONFIG_HOTPLUG_CPU static int op_cpu_disable(unsigned int cpu) { + const struct cpu_operations *ops = get_cpu_ops(cpu); + /* * If we don't have a cpu_die method, abort before we reach the point * of no return. CPU0 may not have an cpu_ops, so test for it. */ - if (!cpu_ops[cpu] || !cpu_ops[cpu]->cpu_die) + if (!ops || !ops->cpu_die) return -EOPNOTSUPP; /* * We may need to abort a hot unplug for some other mechanism-specific * reason. */ - if (cpu_ops[cpu]->cpu_disable) - return cpu_ops[cpu]->cpu_disable(cpu); + if (ops->cpu_disable) + return ops->cpu_disable(cpu); return 0; } @@ -314,15 +315,17 @@ int __cpu_disable(void) static int op_cpu_kill(unsigned int cpu) { + const struct cpu_operations *ops = get_cpu_ops(cpu); + /* * If we have no means of synchronising with the dying CPU, then assume * that it is really dead. We can only wait for an arbitrary length of * time and hope that it's dead, so let's skip the wait and just hope. */ - if (!cpu_ops[cpu]->cpu_kill) + if (!ops->cpu_kill) return 0; - return cpu_ops[cpu]->cpu_kill(cpu); + return ops->cpu_kill(cpu); } /* @@ -357,6 +360,7 @@ void __cpu_die(unsigned int cpu) void cpu_die(void) { unsigned int cpu = smp_processor_id(); + const struct cpu_operations *ops = get_cpu_ops(cpu); idle_task_exit(); @@ -370,12 +374,22 @@ void cpu_die(void) * mechanism must perform all required cache maintenance to ensure that * no dirty lines are lost in the process of shutting down the CPU. */ - cpu_ops[cpu]->cpu_die(cpu); + ops->cpu_die(cpu); BUG(); } #endif +static void __cpu_try_die(int cpu) +{ +#ifdef CONFIG_HOTPLUG_CPU + const struct cpu_operations *ops = get_cpu_ops(cpu); + + if (ops && ops->cpu_die) + ops->cpu_die(cpu); +#endif +} + /* * Kill the calling secondary CPU, early in bringup before it is turned * online. @@ -389,12 +403,11 @@ void cpu_die_early(void) /* Mark this CPU absent */ set_cpu_present(cpu, 0); -#ifdef CONFIG_HOTPLUG_CPU - update_cpu_boot_status(CPU_KILL_ME); - /* Check if we can park ourselves */ - if (cpu_ops[cpu] && cpu_ops[cpu]->cpu_die) - cpu_ops[cpu]->cpu_die(cpu); -#endif + if (IS_ENABLED(CONFIG_HOTPLUG_CPU)) { + update_cpu_boot_status(CPU_KILL_ME); + __cpu_try_die(cpu); + } + update_cpu_boot_status(CPU_STUCK_IN_KERNEL); cpu_park_loop(); @@ -488,10 +501,13 @@ static bool __init is_mpidr_duplicate(unsigned int cpu, u64 hwid) */ static int __init smp_cpu_setup(int cpu) { - if (cpu_read_ops(cpu)) + const struct cpu_operations *ops; + + if (init_cpu_ops(cpu)) return -ENODEV; - if (cpu_ops[cpu]->cpu_init(cpu)) + ops = get_cpu_ops(cpu); + if (ops->cpu_init(cpu)) return -ENODEV; set_cpu_possible(cpu, true); @@ -714,6 +730,7 @@ void __init smp_init_cpus(void) void __init smp_prepare_cpus(unsigned int max_cpus) { + const struct cpu_operations *ops; int err; unsigned int cpu; unsigned int this_cpu; @@ -744,10 +761,11 @@ void __init smp_prepare_cpus(unsigned int max_cpus) if (cpu == smp_processor_id()) continue; - if (!cpu_ops[cpu]) + ops = get_cpu_ops(cpu); + if (!ops) continue; - err = cpu_ops[cpu]->cpu_prepare(cpu); + err = ops->cpu_prepare(cpu); if (err) continue; @@ -863,10 +881,8 @@ static void ipi_cpu_crash_stop(unsigned int cpu, struct pt_regs *regs) local_irq_disable(); sdei_mask_local_cpu(); -#ifdef CONFIG_HOTPLUG_CPU - if (cpu_ops[cpu]->cpu_die) - cpu_ops[cpu]->cpu_die(cpu); -#endif + if (IS_ENABLED(CONFIG_HOTPLUG_CPU)) + __cpu_try_die(cpu); /* just in case */ cpu_park_loop(); @@ -1044,8 +1060,9 @@ static bool have_cpu_die(void) { #ifdef CONFIG_HOTPLUG_CPU int any_cpu = raw_smp_processor_id(); + const struct cpu_operations *ops = get_cpu_ops(any_cpu); - if (cpu_ops[any_cpu] && cpu_ops[any_cpu]->cpu_die) + if (ops && ops->cpu_die) return true; #endif return false; diff --git a/arch/arm64/kernel/topology.c b/arch/arm64/kernel/topology.c index fa9528dfd0ce..0801a0f3c156 100644 --- a/arch/arm64/kernel/topology.c +++ b/arch/arm64/kernel/topology.c @@ -14,6 +14,7 @@ #include <linux/acpi.h> #include <linux/arch_topology.h> #include <linux/cacheinfo.h> +#include <linux/cpufreq.h> #include <linux/init.h> #include <linux/percpu.h> @@ -120,4 +121,183 @@ int __init parse_acpi_topology(void) } #endif +#ifdef CONFIG_ARM64_AMU_EXTN +#undef pr_fmt +#define pr_fmt(fmt) "AMU: " fmt + +static DEFINE_PER_CPU_READ_MOSTLY(unsigned long, arch_max_freq_scale); +static DEFINE_PER_CPU(u64, arch_const_cycles_prev); +static DEFINE_PER_CPU(u64, arch_core_cycles_prev); +static cpumask_var_t amu_fie_cpus; + +/* Initialize counter reference per-cpu variables for the current CPU */ +void init_cpu_freq_invariance_counters(void) +{ + this_cpu_write(arch_core_cycles_prev, + read_sysreg_s(SYS_AMEVCNTR0_CORE_EL0)); + this_cpu_write(arch_const_cycles_prev, + read_sysreg_s(SYS_AMEVCNTR0_CONST_EL0)); +} + +static int validate_cpu_freq_invariance_counters(int cpu) +{ + u64 max_freq_hz, ratio; + + if (!cpu_has_amu_feat(cpu)) { + pr_debug("CPU%d: counters are not supported.\n", cpu); + return -EINVAL; + } + + if (unlikely(!per_cpu(arch_const_cycles_prev, cpu) || + !per_cpu(arch_core_cycles_prev, cpu))) { + pr_debug("CPU%d: cycle counters are not enabled.\n", cpu); + return -EINVAL; + } + + /* Convert maximum frequency from KHz to Hz and validate */ + max_freq_hz = cpufreq_get_hw_max_freq(cpu) * 1000; + if (unlikely(!max_freq_hz)) { + pr_debug("CPU%d: invalid maximum frequency.\n", cpu); + return -EINVAL; + } + + /* + * Pre-compute the fixed ratio between the frequency of the constant + * counter and the maximum frequency of the CPU. + * + * const_freq + * arch_max_freq_scale = ---------------- * SCHED_CAPACITY_SCALE² + * cpuinfo_max_freq + * + * We use a factor of 2 * SCHED_CAPACITY_SHIFT -> SCHED_CAPACITY_SCALE² + * in order to ensure a good resolution for arch_max_freq_scale for + * very low arch timer frequencies (down to the KHz range which should + * be unlikely). + */ + ratio = (u64)arch_timer_get_rate() << (2 * SCHED_CAPACITY_SHIFT); + ratio = div64_u64(ratio, max_freq_hz); + if (!ratio) { + WARN_ONCE(1, "System timer frequency too low.\n"); + return -EINVAL; + } + + per_cpu(arch_max_freq_scale, cpu) = (unsigned long)ratio; + + return 0; +} + +static inline bool +enable_policy_freq_counters(int cpu, cpumask_var_t valid_cpus) +{ + struct cpufreq_policy *policy = cpufreq_cpu_get(cpu); + + if (!policy) { + pr_debug("CPU%d: No cpufreq policy found.\n", cpu); + return false; + } + + if (cpumask_subset(policy->related_cpus, valid_cpus)) + cpumask_or(amu_fie_cpus, policy->related_cpus, + amu_fie_cpus); + + cpufreq_cpu_put(policy); + + return true; +} + +static DEFINE_STATIC_KEY_FALSE(amu_fie_key); +#define amu_freq_invariant() static_branch_unlikely(&amu_fie_key) + +static int __init init_amu_fie(void) +{ + cpumask_var_t valid_cpus; + bool have_policy = false; + int ret = 0; + int cpu; + + if (!zalloc_cpumask_var(&valid_cpus, GFP_KERNEL)) + return -ENOMEM; + + if (!zalloc_cpumask_var(&amu_fie_cpus, GFP_KERNEL)) { + ret = -ENOMEM; + goto free_valid_mask; + } + + for_each_present_cpu(cpu) { + if (validate_cpu_freq_invariance_counters(cpu)) + continue; + cpumask_set_cpu(cpu, valid_cpus); + have_policy |= enable_policy_freq_counters(cpu, valid_cpus); + } + + /* + * If we are not restricted by cpufreq policies, we only enable + * the use of the AMU feature for FIE if all CPUs support AMU. + * Otherwise, enable_policy_freq_counters has already enabled + * policy cpus. + */ + if (!have_policy && cpumask_equal(valid_cpus, cpu_present_mask)) + cpumask_or(amu_fie_cpus, amu_fie_cpus, valid_cpus); + + if (!cpumask_empty(amu_fie_cpus)) { + pr_info("CPUs[%*pbl]: counters will be used for FIE.", + cpumask_pr_args(amu_fie_cpus)); + static_branch_enable(&amu_fie_key); + } + +free_valid_mask: + free_cpumask_var(valid_cpus); + + return ret; +} +late_initcall_sync(init_amu_fie); + +bool arch_freq_counters_available(struct cpumask *cpus) +{ + return amu_freq_invariant() && + cpumask_subset(cpus, amu_fie_cpus); +} + +void topology_scale_freq_tick(void) +{ + u64 prev_core_cnt, prev_const_cnt; + u64 core_cnt, const_cnt, scale; + int cpu = smp_processor_id(); + + if (!amu_freq_invariant()) + return; + + if (!cpumask_test_cpu(cpu, amu_fie_cpus)) + return; + + const_cnt = read_sysreg_s(SYS_AMEVCNTR0_CONST_EL0); + core_cnt = read_sysreg_s(SYS_AMEVCNTR0_CORE_EL0); + prev_const_cnt = this_cpu_read(arch_const_cycles_prev); + prev_core_cnt = this_cpu_read(arch_core_cycles_prev); + + if (unlikely(core_cnt <= prev_core_cnt || + const_cnt <= prev_const_cnt)) + goto store_and_exit; + + /* + * /\core arch_max_freq_scale + * scale = ------- * -------------------- + * /\const SCHED_CAPACITY_SCALE + * + * See validate_cpu_freq_invariance_counters() for details on + * arch_max_freq_scale and the use of SCHED_CAPACITY_SHIFT. + */ + scale = core_cnt - prev_core_cnt; + scale *= this_cpu_read(arch_max_freq_scale); + scale = div64_u64(scale >> SCHED_CAPACITY_SHIFT, + const_cnt - prev_const_cnt); + + scale = min_t(unsigned long, scale, SCHED_CAPACITY_SCALE); + this_cpu_write(freq_scale, (unsigned long)scale); + +store_and_exit: + this_cpu_write(arch_core_cycles_prev, core_cnt); + this_cpu_write(arch_const_cycles_prev, const_cnt); +} +#endif /* CONFIG_ARM64_AMU_EXTN */ |