提交 92205c23 编写于 作者: S Steven Rostedt 提交者: Thomas Gleixner

ftrace: user raw_spin_lock in tracing

Lock debugging enabled cause huge performance problems for tracing. Having
the lock verification happening for every function that is called
because mcount calls spin_lock can cripple the system.

This patch converts the spin_locks used by ftrace into raw_spin_locks.
Signed-off-by: NSteven Rostedt <srostedt@redhat.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
上级 c5f888ca
...@@ -133,7 +133,8 @@ static const char *trace_options[] = { ...@@ -133,7 +133,8 @@ static const char *trace_options[] = {
NULL NULL
}; };
static DEFINE_SPINLOCK(ftrace_max_lock); static raw_spinlock_t ftrace_max_lock =
(raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
/* /*
* Copy the new maximum trace into the separate maximum-trace * Copy the new maximum trace into the separate maximum-trace
...@@ -335,7 +336,7 @@ update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu) ...@@ -335,7 +336,7 @@ update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
int i; int i;
WARN_ON_ONCE(!irqs_disabled()); WARN_ON_ONCE(!irqs_disabled());
spin_lock(&ftrace_max_lock); __raw_spin_lock(&ftrace_max_lock);
/* clear out all the previous traces */ /* clear out all the previous traces */
for_each_possible_cpu(i) { for_each_possible_cpu(i) {
data = tr->data[i]; data = tr->data[i];
...@@ -344,7 +345,7 @@ update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu) ...@@ -344,7 +345,7 @@ update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
} }
__update_max_tr(tr, tsk, cpu); __update_max_tr(tr, tsk, cpu);
spin_unlock(&ftrace_max_lock); __raw_spin_unlock(&ftrace_max_lock);
} }
/** /**
...@@ -360,7 +361,7 @@ update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu) ...@@ -360,7 +361,7 @@ update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
int i; int i;
WARN_ON_ONCE(!irqs_disabled()); WARN_ON_ONCE(!irqs_disabled());
spin_lock(&ftrace_max_lock); __raw_spin_lock(&ftrace_max_lock);
for_each_possible_cpu(i) for_each_possible_cpu(i)
tracing_reset(max_tr.data[i]); tracing_reset(max_tr.data[i]);
...@@ -368,7 +369,7 @@ update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu) ...@@ -368,7 +369,7 @@ update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
tracing_reset(data); tracing_reset(data);
__update_max_tr(tr, tsk, cpu); __update_max_tr(tr, tsk, cpu);
spin_unlock(&ftrace_max_lock); __raw_spin_unlock(&ftrace_max_lock);
} }
int register_tracer(struct tracer *type) int register_tracer(struct tracer *type)
...@@ -652,13 +653,15 @@ trace_function(struct trace_array *tr, struct trace_array_cpu *data, ...@@ -652,13 +653,15 @@ trace_function(struct trace_array *tr, struct trace_array_cpu *data,
struct trace_entry *entry; struct trace_entry *entry;
unsigned long irq_flags; unsigned long irq_flags;
spin_lock_irqsave(&data->lock, irq_flags); raw_local_irq_save(irq_flags);
__raw_spin_lock(&data->lock);
entry = tracing_get_trace_entry(tr, data); entry = tracing_get_trace_entry(tr, data);
tracing_generic_entry_update(entry, flags); tracing_generic_entry_update(entry, flags);
entry->type = TRACE_FN; entry->type = TRACE_FN;
entry->fn.ip = ip; entry->fn.ip = ip;
entry->fn.parent_ip = parent_ip; entry->fn.parent_ip = parent_ip;
spin_unlock_irqrestore(&data->lock, irq_flags); __raw_spin_unlock(&data->lock);
raw_local_irq_restore(irq_flags);
} }
void void
...@@ -678,14 +681,16 @@ __trace_special(void *__tr, void *__data, ...@@ -678,14 +681,16 @@ __trace_special(void *__tr, void *__data,
struct trace_entry *entry; struct trace_entry *entry;
unsigned long irq_flags; unsigned long irq_flags;
spin_lock_irqsave(&data->lock, irq_flags); raw_local_irq_save(irq_flags);
__raw_spin_lock(&data->lock);
entry = tracing_get_trace_entry(tr, data); entry = tracing_get_trace_entry(tr, data);
tracing_generic_entry_update(entry, 0); tracing_generic_entry_update(entry, 0);
entry->type = TRACE_SPECIAL; entry->type = TRACE_SPECIAL;
entry->special.arg1 = arg1; entry->special.arg1 = arg1;
entry->special.arg2 = arg2; entry->special.arg2 = arg2;
entry->special.arg3 = arg3; entry->special.arg3 = arg3;
spin_unlock_irqrestore(&data->lock, irq_flags); __raw_spin_unlock(&data->lock);
raw_local_irq_restore(irq_flags);
trace_wake_up(); trace_wake_up();
} }
...@@ -725,7 +730,8 @@ tracing_sched_switch_trace(struct trace_array *tr, ...@@ -725,7 +730,8 @@ tracing_sched_switch_trace(struct trace_array *tr,
struct trace_entry *entry; struct trace_entry *entry;
unsigned long irq_flags; unsigned long irq_flags;
spin_lock_irqsave(&data->lock, irq_flags); raw_local_irq_save(irq_flags);
__raw_spin_lock(&data->lock);
entry = tracing_get_trace_entry(tr, data); entry = tracing_get_trace_entry(tr, data);
tracing_generic_entry_update(entry, flags); tracing_generic_entry_update(entry, flags);
entry->type = TRACE_CTX; entry->type = TRACE_CTX;
...@@ -736,7 +742,8 @@ tracing_sched_switch_trace(struct trace_array *tr, ...@@ -736,7 +742,8 @@ tracing_sched_switch_trace(struct trace_array *tr,
entry->ctx.next_prio = next->prio; entry->ctx.next_prio = next->prio;
entry->ctx.next_state = next->state; entry->ctx.next_state = next->state;
__trace_stack(tr, data, flags, 4); __trace_stack(tr, data, flags, 4);
spin_unlock_irqrestore(&data->lock, irq_flags); __raw_spin_unlock(&data->lock);
raw_local_irq_restore(irq_flags);
} }
void void
...@@ -749,7 +756,8 @@ tracing_sched_wakeup_trace(struct trace_array *tr, ...@@ -749,7 +756,8 @@ tracing_sched_wakeup_trace(struct trace_array *tr,
struct trace_entry *entry; struct trace_entry *entry;
unsigned long irq_flags; unsigned long irq_flags;
spin_lock_irqsave(&data->lock, irq_flags); raw_local_irq_save(irq_flags);
__raw_spin_lock(&data->lock);
entry = tracing_get_trace_entry(tr, data); entry = tracing_get_trace_entry(tr, data);
tracing_generic_entry_update(entry, flags); tracing_generic_entry_update(entry, flags);
entry->type = TRACE_WAKE; entry->type = TRACE_WAKE;
...@@ -760,7 +768,8 @@ tracing_sched_wakeup_trace(struct trace_array *tr, ...@@ -760,7 +768,8 @@ tracing_sched_wakeup_trace(struct trace_array *tr,
entry->ctx.next_prio = wakee->prio; entry->ctx.next_prio = wakee->prio;
entry->ctx.next_state = wakee->state; entry->ctx.next_state = wakee->state;
__trace_stack(tr, data, flags, 5); __trace_stack(tr, data, flags, 5);
spin_unlock_irqrestore(&data->lock, irq_flags); __raw_spin_unlock(&data->lock);
raw_local_irq_restore(irq_flags);
trace_wake_up(); trace_wake_up();
} }
...@@ -1824,7 +1833,8 @@ tracing_cpumask_write(struct file *filp, const char __user *ubuf, ...@@ -1824,7 +1833,8 @@ tracing_cpumask_write(struct file *filp, const char __user *ubuf,
if (err) if (err)
goto err_unlock; goto err_unlock;
spin_lock_irq(&ftrace_max_lock); raw_local_irq_disable();
__raw_spin_lock(&ftrace_max_lock);
for_each_possible_cpu(cpu) { for_each_possible_cpu(cpu) {
/* /*
* Increase/decrease the disabled counter if we are * Increase/decrease the disabled counter if we are
...@@ -1839,7 +1849,8 @@ tracing_cpumask_write(struct file *filp, const char __user *ubuf, ...@@ -1839,7 +1849,8 @@ tracing_cpumask_write(struct file *filp, const char __user *ubuf,
atomic_dec(&global_trace.data[cpu]->disabled); atomic_dec(&global_trace.data[cpu]->disabled);
} }
} }
spin_unlock_irq(&ftrace_max_lock); __raw_spin_unlock(&ftrace_max_lock);
raw_local_irq_enable();
tracing_cpumask = tracing_cpumask_new; tracing_cpumask = tracing_cpumask_new;
...@@ -2299,7 +2310,7 @@ tracing_read_pipe(struct file *filp, char __user *ubuf, ...@@ -2299,7 +2310,7 @@ tracing_read_pipe(struct file *filp, char __user *ubuf,
for_each_cpu_mask(cpu, mask) { for_each_cpu_mask(cpu, mask) {
data = iter->tr->data[cpu]; data = iter->tr->data[cpu];
spin_lock(&data->lock); __raw_spin_lock(&data->lock);
} }
while (find_next_entry_inc(iter) != NULL) { while (find_next_entry_inc(iter) != NULL) {
...@@ -2320,7 +2331,7 @@ tracing_read_pipe(struct file *filp, char __user *ubuf, ...@@ -2320,7 +2331,7 @@ tracing_read_pipe(struct file *filp, char __user *ubuf,
for_each_cpu_mask(cpu, mask) { for_each_cpu_mask(cpu, mask) {
data = iter->tr->data[cpu]; data = iter->tr->data[cpu];
spin_unlock(&data->lock); __raw_spin_unlock(&data->lock);
} }
for_each_cpu_mask(cpu, mask) { for_each_cpu_mask(cpu, mask) {
...@@ -2538,8 +2549,7 @@ static int trace_alloc_page(void) ...@@ -2538,8 +2549,7 @@ static int trace_alloc_page(void)
/* Now that we successfully allocate a page per CPU, add them */ /* Now that we successfully allocate a page per CPU, add them */
for_each_possible_cpu(i) { for_each_possible_cpu(i) {
data = global_trace.data[i]; data = global_trace.data[i];
spin_lock_init(&data->lock); data->lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
lockdep_set_class(&data->lock, &data->lock_key);
page = list_entry(pages.next, struct page, lru); page = list_entry(pages.next, struct page, lru);
list_del_init(&page->lru); list_del_init(&page->lru);
list_add_tail(&page->lru, &data->trace_pages); list_add_tail(&page->lru, &data->trace_pages);
...@@ -2547,8 +2557,7 @@ static int trace_alloc_page(void) ...@@ -2547,8 +2557,7 @@ static int trace_alloc_page(void)
#ifdef CONFIG_TRACER_MAX_TRACE #ifdef CONFIG_TRACER_MAX_TRACE
data = max_tr.data[i]; data = max_tr.data[i];
spin_lock_init(&data->lock); data->lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
lockdep_set_class(&data->lock, &data->lock_key);
page = list_entry(pages.next, struct page, lru); page = list_entry(pages.next, struct page, lru);
list_del_init(&page->lru); list_del_init(&page->lru);
list_add_tail(&page->lru, &data->trace_pages); list_add_tail(&page->lru, &data->trace_pages);
......
...@@ -76,7 +76,7 @@ struct trace_entry { ...@@ -76,7 +76,7 @@ struct trace_entry {
struct trace_array_cpu { struct trace_array_cpu {
struct list_head trace_pages; struct list_head trace_pages;
atomic_t disabled; atomic_t disabled;
spinlock_t lock; raw_spinlock_t lock;
struct lock_class_key lock_key; struct lock_class_key lock_key;
/* these fields get copied into max-trace: */ /* these fields get copied into max-trace: */
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册