[RFC 1/4] perf, ftrace: Add perf support to use function tracepoint

From: Jiri Olsa
Date: Mon Jul 11 2011 - 09:23:48 EST

Adding perf support to use function tracepoint.

The ftrace:function tracepoint could be now open within the perf and
returns number of processed kernel functions like:

# ./perf stat -e ftrace:function sleep 1

Performance counter stats for 'sleep 1':

7,488 ftrace:function

1.002918678 seconds time elapsed

Details:
- added "struct ftrace_ops" into "struct perf_event", so each event
could define its own filter

- added TRACE_REG_PERF_OPEN/TRACE_REG_PERF_CLOSE in addition to
TRACE_REG_PERF_REGISTER/TRACE_REG_PERF_UNREGISTER cases.

The OPEN/CLOSE cases are called within the event init/destroy
right after the REG/UNREG cases.

The reason is that REG/UNREG are processed only for for first and last
events of the same type. While for function trace we need each event
to get initialized and register the ftrace_ops struct.

- added "void *data" to the class::reg function to pass the "struct perf_event"
for the ftrace:function register function

---
include/linux/ftrace_event.h | 6 +-
include/linux/perf_event.h | 1 +
kernel/trace/trace.h | 5 +
kernel/trace/trace_event_perf.c | 183 ++++++++++++++++++++++++++++++---------
kernel/trace/trace_events.c | 10 ++-
kernel/trace/trace_export.c | 21 +++++
kernel/trace/trace_kprobe.c | 6 +-
kernel/trace/trace_syscalls.c | 14 ++-
8 files changed, 195 insertions(+), 51 deletions(-)

diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h
index b1e69ee..061c267 100644
--- a/include/linux/ftrace_event.h
+++ b/include/linux/ftrace_event.h
@@ -145,6 +145,8 @@ enum trace_reg {
TRACE_REG_UNREGISTER,
TRACE_REG_PERF_REGISTER,
TRACE_REG_PERF_UNREGISTER,
+ TRACE_REG_PERF_OPEN,
+ TRACE_REG_PERF_CLOSE,
};

struct ftrace_event_call;
@@ -156,7 +158,7 @@ struct ftrace_event_class {
void *perf_probe;
#endif
int (*reg)(struct ftrace_event_call *event,
- enum trace_reg type);
+ enum trace_reg type, void *data);
int (*define_fields)(struct ftrace_event_call *);
struct list_head *(*get_fields)(struct ftrace_event_call *);
struct list_head fields;
@@ -164,7 +166,7 @@ struct ftrace_event_class {
};

extern int ftrace_event_reg(struct ftrace_event_call *event,
- enum trace_reg type);
+ enum trace_reg type, void *data);

enum {
TRACE_EVENT_FL_ENABLED_BIT,
diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index 3f2711c..bb708d1 100644
--- a/include/linux/perf_event.h
+++ b/include/linux/perf_event.h
@@ -844,6 +844,7 @@ struct perf_event {
#ifdef CONFIG_EVENT_TRACING
struct ftrace_event_call *tp_event;
struct event_filter *filter;
+ struct ftrace_ops ftrace_ops;
#endif

#ifdef CONFIG_CGROUP_PERF
diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index 30a94c2..9278d74 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -810,4 +810,9 @@ extern const char *__stop___trace_bprintk_fmt[];
#define trace_recursion_clear(bit) do { (current)->trace_recursion &= ~(bit); } while (0)
#define trace_recursion_test(bit) ((current)->trace_recursion & (bit))

+#ifdef CONFIG_PERF_EVENTS
+int perf_ftrace_event_register(struct ftrace_event_call *call,
+ enum trace_reg type, void *data);
+#endif
+
#endif /* _LINUX_KERNEL_TRACE_H */
diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c
index 19a359d..2b56b81 100644
--- a/kernel/trace/trace_event_perf.c
+++ b/kernel/trace/trace_event_perf.c
@@ -44,23 +44,17 @@ static int perf_trace_event_perm(struct ftrace_event_call *tp_event,
return 0;
}

-static int perf_trace_event_init(struct ftrace_event_call *tp_event,
- struct perf_event *p_event)
+static int perf_trace_event_reg(struct ftrace_event_call *tp_event,
+ struct perf_event *p_event)
{
struct hlist_head __percpu *list;
- int ret;
+ int ret = -ENOMEM;
int cpu;

- ret = perf_trace_event_perm(tp_event, p_event);
- if (ret)
- return ret;
-
p_event->tp_event = tp_event;
if (tp_event->perf_refcount++ > 0)
return 0;

- ret = -ENOMEM;
-
list = alloc_percpu(struct hlist_head);
if (!list)
goto fail;
@@ -83,7 +77,7 @@ static int perf_trace_event_init(struct ftrace_event_call *tp_event,
}
}

- ret = tp_event->class->reg(tp_event, TRACE_REG_PERF_REGISTER);
+ ret = tp_event->class->reg(tp_event, TRACE_REG_PERF_REGISTER, NULL);
if (ret)
goto fail;

@@ -108,6 +102,69 @@ fail:
return ret;
}

+static void perf_trace_event_unreg(struct perf_event *p_event)
+{
+ struct ftrace_event_call *tp_event = p_event->tp_event;
+ int i;
+
+ if (--tp_event->perf_refcount > 0)
+ goto out;
+
+ tp_event->class->reg(tp_event, TRACE_REG_PERF_UNREGISTER, NULL);
+
+ /*
+ * Ensure our callback won't be called anymore. The buffers
+ * will be freed after that.
+ */
+ tracepoint_synchronize_unregister();
+
+ free_percpu(tp_event->perf_events);
+ tp_event->perf_events = NULL;
+
+ if (!--total_ref_count) {
+ for (i = 0; i < PERF_NR_CONTEXTS; i++) {
+ free_percpu(perf_trace_buf[i]);
+ perf_trace_buf[i] = NULL;
+ }
+ }
+out:
+ module_put(tp_event->mod);
+}
+
+static int perf_trace_event_open(struct perf_event *p_event)
+{
+ struct ftrace_event_call *tp_event = p_event->tp_event;
+ return tp_event->class->reg(tp_event, TRACE_REG_PERF_OPEN, p_event);
+}
+
+static void perf_trace_event_close(struct perf_event *p_event)
+{
+ struct ftrace_event_call *tp_event = p_event->tp_event;
+ tp_event->class->reg(tp_event, TRACE_REG_PERF_CLOSE, p_event);
+}
+
+static int perf_trace_event_init(struct ftrace_event_call *tp_event,
+ struct perf_event *p_event)
+{
+ int ret;
+
+ ret = perf_trace_event_perm(tp_event, p_event);
+ if (ret)
+ return ret;
+
+ ret = perf_trace_event_reg(tp_event, p_event);
+ if (ret)
+ return ret;
+
+ ret = perf_trace_event_open(p_event);
+ if (ret) {
+ perf_trace_event_unreg(p_event);
+ return ret;
+ }
+
+ return 0;
+}
+
int perf_trace_init(struct perf_event *p_event)
{
struct ftrace_event_call *tp_event;
@@ -130,6 +187,14 @@ int perf_trace_init(struct perf_event *p_event)
return ret;
}

+void perf_trace_destroy(struct perf_event *p_event)
+{
+ mutex_lock(&event_mutex);
+ perf_trace_event_close(p_event);
+ perf_trace_event_unreg(p_event);
+ mutex_unlock(&event_mutex);
+}
+
int perf_trace_add(struct perf_event *p_event, int flags)
{
struct ftrace_event_call *tp_event = p_event->tp_event;
@@ -154,37 +219,6 @@ void perf_trace_del(struct perf_event *p_event, int flags)
hlist_del_rcu(&p_event->hlist_entry);
}

-void perf_trace_destroy(struct perf_event *p_event)
-{
- struct ftrace_event_call *tp_event = p_event->tp_event;
- int i;
-
- mutex_lock(&event_mutex);
- if (--tp_event->perf_refcount > 0)
- goto out;
-
- tp_event->class->reg(tp_event, TRACE_REG_PERF_UNREGISTER);
-
- /*
- * Ensure our callback won't be called anymore. The buffers
- * will be freed after that.
- */
- tracepoint_synchronize_unregister();
-
- free_percpu(tp_event->perf_events);
- tp_event->perf_events = NULL;
-
- if (!--total_ref_count) {
- for (i = 0; i < PERF_NR_CONTEXTS; i++) {
- free_percpu(perf_trace_buf[i]);
- perf_trace_buf[i] = NULL;
- }
- }
-out:
- module_put(tp_event->mod);
- mutex_unlock(&event_mutex);
-}
-
__kprobes void *perf_trace_buf_prepare(int size, unsigned short type,
struct pt_regs *regs, int *rctxp)
{
@@ -214,3 +248,70 @@ __kprobes void *perf_trace_buf_prepare(int size, unsigned short type,
return raw_data;
}
EXPORT_SYMBOL_GPL(perf_trace_buf_prepare);
+
+
+static void
+perf_ftrace_function_call(unsigned long ip, unsigned long parent_ip)
+{
+ struct ftrace_entry *entry;
+ struct hlist_head *head;
+ unsigned long flags;
+ int rctx;
+
+ local_irq_save(flags);
+
+ BUILD_BUG_ON(sizeof(*entry) > PERF_MAX_TRACE_SIZE);
+
+ entry = (struct ftrace_entry *) perf_trace_buf_prepare(sizeof(*entry),
+ TRACE_FN, NULL, &rctx);
+ if (!entry)
+ return;
+
+ entry->ip = ip;
+ entry->parent_ip = parent_ip;
+
+ head = this_cpu_ptr(event_function.perf_events);
+ perf_trace_buf_submit(entry, sizeof(*entry), rctx, 0,
+ 1, NULL, head);
+
+ local_irq_restore(flags);
+}
+
+static int perf_ftrace_function_register(struct ftrace_event_call *call,
+ struct perf_event *event)
+{
+ struct ftrace_ops *ops = &event->ftrace_ops;
+ ops->func = perf_ftrace_function_call;
+ return register_ftrace_function(ops);
+}
+
+static int perf_ftrace_function_unregister(struct ftrace_event_call *call,
+ struct perf_event *event)
+{
+ struct ftrace_ops *ops = &event->ftrace_ops;
+ return unregister_ftrace_function(ops);
+}
+
+int perf_ftrace_event_register(struct ftrace_event_call *call,
+ enum trace_reg type, void *data)
+{
+ int etype = call->event.type;
+
+ if (etype != TRACE_FN)
+ return -EINVAL;
+
+ switch (type) {
+ case TRACE_REG_REGISTER:
+ case TRACE_REG_UNREGISTER:
+ break;
+ case TRACE_REG_PERF_REGISTER:
+ case TRACE_REG_PERF_UNREGISTER:
+ return 0;
+ case TRACE_REG_PERF_OPEN:
+ return perf_ftrace_function_register(call, data);
+ case TRACE_REG_PERF_CLOSE:
+ return perf_ftrace_function_unregister(call, data);
+ }
+
+ return -EINVAL;
+}
diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index 581876f..ad4bf55 100644
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -147,7 +147,8 @@ int trace_event_raw_init(struct ftrace_event_call *call)
}
EXPORT_SYMBOL_GPL(trace_event_raw_init);

-int ftrace_event_reg(struct ftrace_event_call *call, enum trace_reg type)
+int ftrace_event_reg(struct ftrace_event_call *call,
+ enum trace_reg type, void *data)
{
switch (type) {
case TRACE_REG_REGISTER:
@@ -170,6 +171,9 @@ int ftrace_event_reg(struct ftrace_event_call *call, enum trace_reg type)
call->class->perf_probe,
call);
return 0;
+ case TRACE_REG_PERF_OPEN:
+ case TRACE_REG_PERF_CLOSE:
+ return 0;
#endif
}
return 0;
@@ -209,7 +213,7 @@ static int ftrace_event_enable_disable(struct ftrace_event_call *call,
tracing_stop_cmdline_record();
call->flags &= ~TRACE_EVENT_FL_RECORDED_CMD;
}
- call->class->reg(call, TRACE_REG_UNREGISTER);
+ call->class->reg(call, TRACE_REG_UNREGISTER, NULL);
}
break;
case 1:
@@ -218,7 +222,7 @@ static int ftrace_event_enable_disable(struct ftrace_event_call *call,
tracing_start_cmdline_record();
call->flags |= TRACE_EVENT_FL_RECORDED_CMD;
}
- ret = call->class->reg(call, TRACE_REG_REGISTER);
+ ret = call->class->reg(call, TRACE_REG_REGISTER, NULL);
if (ret) {
tracing_stop_cmdline_record();
pr_info("event trace: Could not enable event "
diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c
index bbeec31..7b7193c 100644
--- a/kernel/trace/trace_export.c
+++ b/kernel/trace/trace_export.c
@@ -131,6 +131,26 @@ ftrace_define_fields_##name(struct ftrace_event_call *event_call) \

#include "trace_entries.h"

+static int ftrace_event_class_register(struct ftrace_event_call *call,
+ enum trace_reg type, void *data)
+{
+ switch (type) {
+ case TRACE_REG_PERF_REGISTER:
+ case TRACE_REG_PERF_UNREGISTER:
+ return 0;
+ case TRACE_REG_PERF_OPEN:
+ case TRACE_REG_PERF_CLOSE:
+#ifdef CONFIG_PERF_EVENTS
+ return perf_ftrace_event_register(call, type, data);
+#endif
+ case TRACE_REG_REGISTER:
+ case TRACE_REG_UNREGISTER:
+ break;
+ }
+
+ return -EINVAL;
+}
+
#undef __entry
#define __entry REC

@@ -159,6 +179,7 @@ struct ftrace_event_class event_class_ftrace_##call = { \
.system = __stringify(TRACE_SYSTEM), \
.define_fields = ftrace_define_fields_##call, \
.fields = LIST_HEAD_INIT(event_class_ftrace_##call.fields),\
+ .reg = ftrace_event_class_register, \
}; \
\
struct ftrace_event_call __used event_##call = { \
diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c
index 7db7b68..bfcf609 100644
--- a/kernel/trace/trace_kprobe.c
+++ b/kernel/trace/trace_kprobe.c
@@ -1744,7 +1744,8 @@ static void probe_perf_disable(struct ftrace_event_call *call)
#endif /* CONFIG_PERF_EVENTS */

static __kprobes
-int kprobe_register(struct ftrace_event_call *event, enum trace_reg type)
+int kprobe_register(struct ftrace_event_call *event,
+ enum trace_reg type, void *data)
{
switch (type) {
case TRACE_REG_REGISTER:
@@ -1759,6 +1760,9 @@ int kprobe_register(struct ftrace_event_call *event, enum trace_reg type)
case TRACE_REG_PERF_UNREGISTER:
probe_perf_disable(event);
return 0;
+ case TRACE_REG_PERF_OPEN:
+ case TRACE_REG_PERF_CLOSE:
+ return 0;
#endif
}
return 0;
diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c
index ee7b5a0..61ebe6a 100644
--- a/kernel/trace/trace_syscalls.c
+++ b/kernel/trace/trace_syscalls.c
@@ -16,9 +16,9 @@ static DECLARE_BITMAP(enabled_enter_syscalls, NR_syscalls);
static DECLARE_BITMAP(enabled_exit_syscalls, NR_syscalls);

static int syscall_enter_register(struct ftrace_event_call *event,
- enum trace_reg type);
+ enum trace_reg type, void *data);
static int syscall_exit_register(struct ftrace_event_call *event,
- enum trace_reg type);
+ enum trace_reg type, void *data);

static int syscall_enter_define_fields(struct ftrace_event_call *call);
static int syscall_exit_define_fields(struct ftrace_event_call *call);
@@ -648,7 +648,7 @@ void perf_sysexit_disable(struct ftrace_event_call *call)
#endif /* CONFIG_PERF_EVENTS */

static int syscall_enter_register(struct ftrace_event_call *event,
- enum trace_reg type)
+ enum trace_reg type, void *data)
{
switch (type) {
case TRACE_REG_REGISTER:
@@ -663,13 +663,16 @@ static int syscall_enter_register(struct ftrace_event_call *event,
case TRACE_REG_PERF_UNREGISTER:
perf_sysenter_disable(event);
return 0;
+ case TRACE_REG_PERF_OPEN:
+ case TRACE_REG_PERF_CLOSE:
+ return 0;
#endif
}
return 0;
}

static int syscall_exit_register(struct ftrace_event_call *event,
- enum trace_reg type)
+ enum trace_reg type, void *data)
{
switch (type) {
case TRACE_REG_REGISTER:
@@ -684,6 +687,9 @@ static int syscall_exit_register(struct ftrace_event_call *event,
case TRACE_REG_PERF_UNREGISTER:
perf_sysexit_disable(event);
return 0;
+ case TRACE_REG_PERF_OPEN:
+ case TRACE_REG_PERF_CLOSE:
+ return 0;
#endif
}
return 0;
--
1.7.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/