diff options
Diffstat (limited to 'arch/x86/events/rapl.c')
-rw-r--r-- | arch/x86/events/rapl.c | 149 |
1 files changed, 94 insertions, 55 deletions
diff --git a/arch/x86/events/rapl.c b/arch/x86/events/rapl.c index fb2b1961e5a3..a481a939862e 100644 --- a/arch/x86/events/rapl.c +++ b/arch/x86/events/rapl.c @@ -64,6 +64,7 @@ #include "perf_event.h" #include "probe.h" +MODULE_DESCRIPTION("Support Intel/AMD RAPL energy consumption counters"); MODULE_LICENSE("GPL"); /* @@ -102,6 +103,19 @@ static struct perf_pmu_events_attr event_attr_##v = { \ .event_str = str, \ }; +/* + * RAPL Package energy counter scope: + * 1. AMD/HYGON platforms have a per-PKG package energy counter + * 2. For Intel platforms + * 2.1. CLX-AP is multi-die and its RAPL MSRs are die-scope + * 2.2. Other Intel platforms are single die systems so the scope can be + * considered as either pkg-scope or die-scope, and we are considering + * them as die-scope. + */ +#define rapl_pmu_is_pkg_scope() \ + (boot_cpu_data.x86_vendor == X86_VENDOR_AMD || \ + boot_cpu_data.x86_vendor == X86_VENDOR_HYGON) + struct rapl_pmu { raw_spinlock_t lock; int n_active; @@ -114,8 +128,8 @@ struct rapl_pmu { struct rapl_pmus { struct pmu pmu; - unsigned int maxdie; - struct rapl_pmu *pmus[] __counted_by(maxdie); + unsigned int nr_rapl_pmu; + struct rapl_pmu *pmus[] __counted_by(nr_rapl_pmu); }; enum rapl_unit_quirk { @@ -139,15 +153,31 @@ static unsigned int rapl_cntr_mask; static u64 rapl_timer_ms; static struct perf_msr *rapl_msrs; +/* + * Helper functions to get the correct topology macros according to the + * RAPL PMU scope. + */ +static inline unsigned int get_rapl_pmu_idx(int cpu) +{ + return rapl_pmu_is_pkg_scope() ? topology_logical_package_id(cpu) : + topology_logical_die_id(cpu); +} + +static inline const struct cpumask *get_rapl_pmu_cpumask(int cpu) +{ + return rapl_pmu_is_pkg_scope() ? topology_core_cpumask(cpu) : + topology_die_cpumask(cpu); +} + static inline struct rapl_pmu *cpu_to_rapl_pmu(unsigned int cpu) { - unsigned int dieid = topology_logical_die_id(cpu); + unsigned int rapl_pmu_idx = get_rapl_pmu_idx(cpu); /* * The unsigned check also catches the '-1' return value for non * existent mappings in the topology map. */ - return dieid < rapl_pmus->maxdie ? rapl_pmus->pmus[dieid] : NULL; + return rapl_pmu_idx < rapl_pmus->nr_rapl_pmu ? rapl_pmus->pmus[rapl_pmu_idx] : NULL; } static inline u64 rapl_read_counter(struct perf_event *event) @@ -551,7 +581,7 @@ static int rapl_cpu_offline(unsigned int cpu) pmu->cpu = -1; /* Find a new cpu to collect rapl events */ - target = cpumask_any_but(topology_die_cpumask(cpu), cpu); + target = cpumask_any_but(get_rapl_pmu_cpumask(cpu), cpu); /* Migrate rapl events to the new target */ if (target < nr_cpu_ids) { @@ -564,6 +594,11 @@ static int rapl_cpu_offline(unsigned int cpu) static int rapl_cpu_online(unsigned int cpu) { + s32 rapl_pmu_idx = get_rapl_pmu_idx(cpu); + if (rapl_pmu_idx < 0) { + pr_err("topology_logical_(package/die)_id() returned a negative value"); + return -EINVAL; + } struct rapl_pmu *pmu = cpu_to_rapl_pmu(cpu); int target; @@ -578,14 +613,14 @@ static int rapl_cpu_online(unsigned int cpu) pmu->timer_interval = ms_to_ktime(rapl_timer_ms); rapl_hrtimer_init(pmu); - rapl_pmus->pmus[topology_logical_die_id(cpu)] = pmu; + rapl_pmus->pmus[rapl_pmu_idx] = pmu; } /* * Check if there is an online cpu in the package which collects rapl * events already. */ - target = cpumask_any_and(&rapl_cpu_mask, topology_die_cpumask(cpu)); + target = cpumask_any_and(&rapl_cpu_mask, get_rapl_pmu_cpumask(cpu)); if (target < nr_cpu_ids) return 0; @@ -658,7 +693,7 @@ static void cleanup_rapl_pmus(void) { int i; - for (i = 0; i < rapl_pmus->maxdie; i++) + for (i = 0; i < rapl_pmus->nr_rapl_pmu; i++) kfree(rapl_pmus->pmus[i]); kfree(rapl_pmus); } @@ -674,15 +709,16 @@ static const struct attribute_group *rapl_attr_update[] = { static int __init init_rapl_pmus(void) { - int maxdie = topology_max_packages() * topology_max_dies_per_package(); - size_t size; + int nr_rapl_pmu = topology_max_packages(); + + if (!rapl_pmu_is_pkg_scope()) + nr_rapl_pmu *= topology_max_dies_per_package(); - size = sizeof(*rapl_pmus) + maxdie * sizeof(struct rapl_pmu *); - rapl_pmus = kzalloc(size, GFP_KERNEL); + rapl_pmus = kzalloc(struct_size(rapl_pmus, pmus, nr_rapl_pmu), GFP_KERNEL); if (!rapl_pmus) return -ENOMEM; - rapl_pmus->maxdie = maxdie; + rapl_pmus->nr_rapl_pmu = nr_rapl_pmu; rapl_pmus->pmu.attr_groups = rapl_attr_groups; rapl_pmus->pmu.attr_update = rapl_attr_update; rapl_pmus->pmu.task_ctx_nr = perf_invalid_context; @@ -766,48 +802,51 @@ static struct rapl_model model_amd_hygon = { }; static const struct x86_cpu_id rapl_model_match[] __initconst = { - X86_MATCH_FEATURE(X86_FEATURE_RAPL, &model_amd_hygon), - X86_MATCH_INTEL_FAM6_MODEL(SANDYBRIDGE, &model_snb), - X86_MATCH_INTEL_FAM6_MODEL(SANDYBRIDGE_X, &model_snbep), - X86_MATCH_INTEL_FAM6_MODEL(IVYBRIDGE, &model_snb), - X86_MATCH_INTEL_FAM6_MODEL(IVYBRIDGE_X, &model_snbep), - X86_MATCH_INTEL_FAM6_MODEL(HASWELL, &model_hsw), - X86_MATCH_INTEL_FAM6_MODEL(HASWELL_X, &model_hsx), - X86_MATCH_INTEL_FAM6_MODEL(HASWELL_L, &model_hsw), - X86_MATCH_INTEL_FAM6_MODEL(HASWELL_G, &model_hsw), - X86_MATCH_INTEL_FAM6_MODEL(BROADWELL, &model_hsw), - X86_MATCH_INTEL_FAM6_MODEL(BROADWELL_G, &model_hsw), - X86_MATCH_INTEL_FAM6_MODEL(BROADWELL_X, &model_hsx), - X86_MATCH_INTEL_FAM6_MODEL(BROADWELL_D, &model_hsx), - X86_MATCH_INTEL_FAM6_MODEL(XEON_PHI_KNL, &model_knl), - X86_MATCH_INTEL_FAM6_MODEL(XEON_PHI_KNM, &model_knl), - X86_MATCH_INTEL_FAM6_MODEL(SKYLAKE_L, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(SKYLAKE, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(SKYLAKE_X, &model_hsx), - X86_MATCH_INTEL_FAM6_MODEL(KABYLAKE_L, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(KABYLAKE, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(CANNONLAKE_L, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(ATOM_GOLDMONT, &model_hsw), - X86_MATCH_INTEL_FAM6_MODEL(ATOM_GOLDMONT_D, &model_hsw), - X86_MATCH_INTEL_FAM6_MODEL(ATOM_GOLDMONT_PLUS, &model_hsw), - X86_MATCH_INTEL_FAM6_MODEL(ICELAKE_L, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(ICELAKE, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(ICELAKE_D, &model_hsx), - X86_MATCH_INTEL_FAM6_MODEL(ICELAKE_X, &model_hsx), - X86_MATCH_INTEL_FAM6_MODEL(COMETLAKE_L, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(COMETLAKE, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(TIGERLAKE_L, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(TIGERLAKE, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(ALDERLAKE, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(ALDERLAKE_L, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(ATOM_GRACEMONT, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(SAPPHIRERAPIDS_X, &model_spr), - X86_MATCH_INTEL_FAM6_MODEL(EMERALDRAPIDS_X, &model_spr), - X86_MATCH_INTEL_FAM6_MODEL(RAPTORLAKE, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(RAPTORLAKE_P, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(RAPTORLAKE_S, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(METEORLAKE, &model_skl), - X86_MATCH_INTEL_FAM6_MODEL(METEORLAKE_L, &model_skl), + X86_MATCH_FEATURE(X86_FEATURE_RAPL, &model_amd_hygon), + X86_MATCH_VFM(INTEL_SANDYBRIDGE, &model_snb), + X86_MATCH_VFM(INTEL_SANDYBRIDGE_X, &model_snbep), + X86_MATCH_VFM(INTEL_IVYBRIDGE, &model_snb), + X86_MATCH_VFM(INTEL_IVYBRIDGE_X, &model_snbep), + X86_MATCH_VFM(INTEL_HASWELL, &model_hsw), + X86_MATCH_VFM(INTEL_HASWELL_X, &model_hsx), + X86_MATCH_VFM(INTEL_HASWELL_L, &model_hsw), + X86_MATCH_VFM(INTEL_HASWELL_G, &model_hsw), + X86_MATCH_VFM(INTEL_BROADWELL, &model_hsw), + X86_MATCH_VFM(INTEL_BROADWELL_G, &model_hsw), + X86_MATCH_VFM(INTEL_BROADWELL_X, &model_hsx), + X86_MATCH_VFM(INTEL_BROADWELL_D, &model_hsx), + X86_MATCH_VFM(INTEL_XEON_PHI_KNL, &model_knl), + X86_MATCH_VFM(INTEL_XEON_PHI_KNM, &model_knl), + X86_MATCH_VFM(INTEL_SKYLAKE_L, &model_skl), + X86_MATCH_VFM(INTEL_SKYLAKE, &model_skl), + X86_MATCH_VFM(INTEL_SKYLAKE_X, &model_hsx), + X86_MATCH_VFM(INTEL_KABYLAKE_L, &model_skl), + X86_MATCH_VFM(INTEL_KABYLAKE, &model_skl), + X86_MATCH_VFM(INTEL_CANNONLAKE_L, &model_skl), + X86_MATCH_VFM(INTEL_ATOM_GOLDMONT, &model_hsw), + X86_MATCH_VFM(INTEL_ATOM_GOLDMONT_D, &model_hsw), + X86_MATCH_VFM(INTEL_ATOM_GOLDMONT_PLUS, &model_hsw), + X86_MATCH_VFM(INTEL_ICELAKE_L, &model_skl), + X86_MATCH_VFM(INTEL_ICELAKE, &model_skl), + X86_MATCH_VFM(INTEL_ICELAKE_D, &model_hsx), + X86_MATCH_VFM(INTEL_ICELAKE_X, &model_hsx), + X86_MATCH_VFM(INTEL_COMETLAKE_L, &model_skl), + X86_MATCH_VFM(INTEL_COMETLAKE, &model_skl), + X86_MATCH_VFM(INTEL_TIGERLAKE_L, &model_skl), + X86_MATCH_VFM(INTEL_TIGERLAKE, &model_skl), + X86_MATCH_VFM(INTEL_ALDERLAKE, &model_skl), + X86_MATCH_VFM(INTEL_ALDERLAKE_L, &model_skl), + X86_MATCH_VFM(INTEL_ATOM_GRACEMONT, &model_skl), + X86_MATCH_VFM(INTEL_SAPPHIRERAPIDS_X, &model_spr), + X86_MATCH_VFM(INTEL_EMERALDRAPIDS_X, &model_spr), + X86_MATCH_VFM(INTEL_RAPTORLAKE, &model_skl), + X86_MATCH_VFM(INTEL_RAPTORLAKE_P, &model_skl), + X86_MATCH_VFM(INTEL_RAPTORLAKE_S, &model_skl), + X86_MATCH_VFM(INTEL_METEORLAKE, &model_skl), + X86_MATCH_VFM(INTEL_METEORLAKE_L, &model_skl), + X86_MATCH_VFM(INTEL_ARROWLAKE_H, &model_skl), + X86_MATCH_VFM(INTEL_ARROWLAKE, &model_skl), + X86_MATCH_VFM(INTEL_LUNARLAKE_M, &model_skl), {}, }; MODULE_DEVICE_TABLE(x86cpu, rapl_model_match); |