Fix: timer_expire_entry changed in 4.19.312
[lttng-modules.git] / lttng-context-perf-counters.c
diff --git a/lttng-context-perf-counters.c b/lttng-context-perf-counters.c
deleted file mode 100644 (file)
index 260e5d0..0000000
+++ /dev/null
@@ -1,377 +0,0 @@
-/*
- * lttng-context-perf-counters.c
- *
- * LTTng performance monitoring counters (perf-counters) integration module.
- *
- * Copyright (C) 2009-2012 Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; only
- * version 2.1 of the License.
- *
- * This library is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
-
-#include <linux/module.h>
-#include <linux/slab.h>
-#include <linux/perf_event.h>
-#include <linux/list.h>
-#include <linux/string.h>
-#include <linux/cpu.h>
-#include <lttng-events.h>
-#include <wrapper/ringbuffer/frontend_types.h>
-#include <wrapper/vmalloc.h>
-#include <wrapper/perf.h>
-#include <lttng-tracer.h>
-
-static
-size_t perf_counter_get_size(size_t offset)
-{
-       size_t size = 0;
-
-       size += lib_ring_buffer_align(offset, lttng_alignof(uint64_t));
-       size += sizeof(uint64_t);
-       return size;
-}
-
-static
-void perf_counter_record(struct lttng_ctx_field *field,
-                        struct lib_ring_buffer_ctx *ctx,
-                        struct lttng_channel *chan)
-{
-       struct perf_event *event;
-       uint64_t value;
-
-       event = field->u.perf_counter->e[ctx->cpu];
-       if (likely(event)) {
-               if (unlikely(event->state == PERF_EVENT_STATE_ERROR)) {
-                       value = 0;
-               } else {
-                       event->pmu->read(event);
-                       value = local64_read(&event->count);
-               }
-       } else {
-               /*
-                * Perf chooses not to be clever and not to support enabling a
-                * perf counter before the cpu is brought up. Therefore, we need
-                * to support having events coming (e.g. scheduler events)
-                * before the counter is setup. Write an arbitrary 0 in this
-                * case.
-                */
-               value = 0;
-       }
-       lib_ring_buffer_align_ctx(ctx, lttng_alignof(value));
-       chan->ops->event_write(ctx, &value, sizeof(value));
-}
-
-#if defined(CONFIG_PERF_EVENTS) && (LINUX_VERSION_CODE >= KERNEL_VERSION(3,0,99))
-static
-void overflow_callback(struct perf_event *event,
-                      struct perf_sample_data *data,
-                      struct pt_regs *regs)
-{
-}
-#else
-static
-void overflow_callback(struct perf_event *event, int nmi,
-                      struct perf_sample_data *data,
-                      struct pt_regs *regs)
-{
-}
-#endif
-
-static
-void lttng_destroy_perf_counter_field(struct lttng_ctx_field *field)
-{
-       struct perf_event **events = field->u.perf_counter->e;
-
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,10,0))
-       {
-               int ret;
-
-               ret = cpuhp_state_remove_instance(lttng_hp_online,
-                       &field->u.perf_counter->cpuhp_online.node);
-               WARN_ON(ret);
-               ret = cpuhp_state_remove_instance(lttng_hp_prepare,
-                       &field->u.perf_counter->cpuhp_prepare.node);
-               WARN_ON(ret);
-       }
-#else /* #if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,10,0)) */
-       {
-               int cpu;
-
-               get_online_cpus();
-               for_each_online_cpu(cpu)
-                       perf_event_release_kernel(events[cpu]);
-               put_online_cpus();
-#ifdef CONFIG_HOTPLUG_CPU
-               unregister_cpu_notifier(&field->u.perf_counter->nb);
-#endif
-       }
-#endif /* #else #if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,10,0)) */
-       kfree(field->event_field.name);
-       kfree(field->u.perf_counter->attr);
-       lttng_kvfree(events);
-       kfree(field->u.perf_counter);
-}
-
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,10,0))
-
-int lttng_cpuhp_perf_counter_online(unsigned int cpu,
-               struct lttng_cpuhp_node *node)
-{
-       struct lttng_perf_counter_field *perf_field =
-               container_of(node, struct lttng_perf_counter_field,
-                               cpuhp_online);
-       struct perf_event **events = perf_field->e;
-       struct perf_event_attr *attr = perf_field->attr;
-       struct perf_event *pevent;
-
-       pevent = wrapper_perf_event_create_kernel_counter(attr,
-                       cpu, NULL, overflow_callback);
-       if (!pevent || IS_ERR(pevent))
-               return -EINVAL;
-       if (pevent->state == PERF_EVENT_STATE_ERROR) {
-               perf_event_release_kernel(pevent);
-               return -EINVAL;
-       }
-       barrier();      /* Create perf counter before setting event */
-       events[cpu] = pevent;
-       return 0;
-}
-
-int lttng_cpuhp_perf_counter_dead(unsigned int cpu,
-               struct lttng_cpuhp_node *node)
-{
-       struct lttng_perf_counter_field *perf_field =
-               container_of(node, struct lttng_perf_counter_field,
-                               cpuhp_prepare);
-       struct perf_event **events = perf_field->e;
-       struct perf_event *pevent;
-
-       pevent = events[cpu];
-       events[cpu] = NULL;
-       barrier();      /* NULLify event before perf counter teardown */
-       perf_event_release_kernel(pevent);
-       return 0;
-}
-
-#else /* #if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,10,0)) */
-
-#ifdef CONFIG_HOTPLUG_CPU
-
-/**
- *     lttng_perf_counter_hp_callback - CPU hotplug callback
- *     @nb: notifier block
- *     @action: hotplug action to take
- *     @hcpu: CPU number
- *
- *     Returns the success/failure of the operation. (%NOTIFY_OK, %NOTIFY_BAD)
- *
- * We can setup perf counters when the cpu is online (up prepare seems to be too
- * soon).
- */
-static
-int lttng_perf_counter_cpu_hp_callback(struct notifier_block *nb,
-                                                unsigned long action,
-                                                void *hcpu)
-{
-       unsigned int cpu = (unsigned long) hcpu;
-       struct lttng_perf_counter_field *perf_field =
-               container_of(nb, struct lttng_perf_counter_field, nb);
-       struct perf_event **events = perf_field->e;
-       struct perf_event_attr *attr = perf_field->attr;
-       struct perf_event *pevent;
-
-       if (!perf_field->hp_enable)
-               return NOTIFY_OK;
-
-       switch (action) {
-       case CPU_ONLINE:
-       case CPU_ONLINE_FROZEN:
-               pevent = wrapper_perf_event_create_kernel_counter(attr,
-                               cpu, NULL, overflow_callback);
-               if (!pevent || IS_ERR(pevent))
-                       return NOTIFY_BAD;
-               if (pevent->state == PERF_EVENT_STATE_ERROR) {
-                       perf_event_release_kernel(pevent);
-                       return NOTIFY_BAD;
-               }
-               barrier();      /* Create perf counter before setting event */
-               events[cpu] = pevent;
-               break;
-       case CPU_UP_CANCELED:
-       case CPU_UP_CANCELED_FROZEN:
-       case CPU_DEAD:
-       case CPU_DEAD_FROZEN:
-               pevent = events[cpu];
-               events[cpu] = NULL;
-               barrier();      /* NULLify event before perf counter teardown */
-               perf_event_release_kernel(pevent);
-               break;
-       }
-       return NOTIFY_OK;
-}
-
-#endif
-
-#endif /* #else #if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,10,0)) */
-
-int lttng_add_perf_counter_to_ctx(uint32_t type,
-                                 uint64_t config,
-                                 const char *name,
-                                 struct lttng_ctx **ctx)
-{
-       struct lttng_ctx_field *field;
-       struct lttng_perf_counter_field *perf_field;
-       struct perf_event **events;
-       struct perf_event_attr *attr;
-       int ret;
-       char *name_alloc;
-
-       events = lttng_kvzalloc(num_possible_cpus() * sizeof(*events), GFP_KERNEL);
-       if (!events)
-               return -ENOMEM;
-
-       attr = kzalloc(sizeof(struct perf_event_attr), GFP_KERNEL);
-       if (!attr) {
-               ret = -ENOMEM;
-               goto error_attr;
-       }
-
-       attr->type = type;
-       attr->config = config;
-       attr->size = sizeof(struct perf_event_attr);
-       attr->pinned = 1;
-       attr->disabled = 0;
-
-       perf_field = kzalloc(sizeof(struct lttng_perf_counter_field), GFP_KERNEL);
-       if (!perf_field) {
-               ret = -ENOMEM;
-               goto error_alloc_perf_field;
-       }
-       perf_field->e = events;
-       perf_field->attr = attr;
-
-       name_alloc = kstrdup(name, GFP_KERNEL);
-       if (!name_alloc) {
-               ret = -ENOMEM;
-               goto name_alloc_error;
-       }
-
-       field = lttng_append_context(ctx);
-       if (!field) {
-               ret = -ENOMEM;
-               goto append_context_error;
-       }
-       if (lttng_find_context(*ctx, name_alloc)) {
-               ret = -EEXIST;
-               goto find_error;
-       }
-
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,10,0))
-
-       perf_field->cpuhp_prepare.component = LTTNG_CONTEXT_PERF_COUNTERS;
-       ret = cpuhp_state_add_instance(lttng_hp_prepare,
-               &perf_field->cpuhp_prepare.node);
-       if (ret)
-               goto cpuhp_prepare_error;
-
-       perf_field->cpuhp_online.component = LTTNG_CONTEXT_PERF_COUNTERS;
-       ret = cpuhp_state_add_instance(lttng_hp_online,
-               &perf_field->cpuhp_online.node);
-       if (ret)
-               goto cpuhp_online_error;
-
-#else  /* #if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,10,0)) */
-       {
-               int cpu;
-
-#ifdef CONFIG_HOTPLUG_CPU
-               perf_field->nb.notifier_call =
-                       lttng_perf_counter_cpu_hp_callback;
-               perf_field->nb.priority = 0;
-               register_cpu_notifier(&perf_field->nb);
-#endif
-               get_online_cpus();
-               for_each_online_cpu(cpu) {
-                       events[cpu] = wrapper_perf_event_create_kernel_counter(attr,
-                                               cpu, NULL, overflow_callback);
-                       if (!events[cpu] || IS_ERR(events[cpu])) {
-                               ret = -EINVAL;
-                               goto counter_error;
-                       }
-                       if (events[cpu]->state == PERF_EVENT_STATE_ERROR) {
-                               ret = -EBUSY;
-                               goto counter_busy;
-                       }
-               }
-               put_online_cpus();
-               perf_field->hp_enable = 1;
-       }
-#endif /* #else #if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,10,0)) */
-
-       field->destroy = lttng_destroy_perf_counter_field;
-
-       field->event_field.name = name_alloc;
-       field->event_field.type.atype = atype_integer;
-       field->event_field.type.u.basic.integer.size = sizeof(uint64_t) * CHAR_BIT;
-       field->event_field.type.u.basic.integer.alignment = lttng_alignof(uint64_t) * CHAR_BIT;
-       field->event_field.type.u.basic.integer.signedness = lttng_is_signed_type(uint64_t);
-       field->event_field.type.u.basic.integer.reverse_byte_order = 0;
-       field->event_field.type.u.basic.integer.base = 10;
-       field->event_field.type.u.basic.integer.encoding = lttng_encode_none;
-       field->get_size = perf_counter_get_size;
-       field->record = perf_counter_record;
-       field->u.perf_counter = perf_field;
-       lttng_context_update(*ctx);
-
-       wrapper_vmalloc_sync_all();
-       return 0;
-
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,10,0))
-cpuhp_online_error:
-       {
-               int remove_ret;
-
-               remove_ret = cpuhp_state_remove_instance(lttng_hp_prepare,
-                               &perf_field->cpuhp_prepare.node);
-               WARN_ON(remove_ret);
-       }
-cpuhp_prepare_error:
-#else  /* #if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,10,0)) */
-counter_busy:
-counter_error:
-       {
-               int cpu;
-
-               for_each_online_cpu(cpu) {
-                       if (events[cpu] && !IS_ERR(events[cpu]))
-                               perf_event_release_kernel(events[cpu]);
-               }
-               put_online_cpus();
-#ifdef CONFIG_HOTPLUG_CPU
-               unregister_cpu_notifier(&perf_field->nb);
-#endif
-       }
-#endif /* #else #if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,10,0)) */
-find_error:
-       lttng_remove_context_field(ctx, field);
-append_context_error:
-       kfree(name_alloc);
-name_alloc_error:
-       kfree(perf_field);
-error_alloc_perf_field:
-       kfree(attr);
-error_attr:
-       lttng_kvfree(events);
-       return ret;
-}
This page took 0.027474 seconds and 4 git commands to generate.