ftrace: kill the dead code in trace_sched_switch.c

PATCH New Oleg Nesterov 1 participant 3 messages Jul 23, 2014 Tracing

lore

Share & tools

lore

Oleg Nesterov

author

2014-07-23 19:36 UTC #1 lore raw

PATCH 1/2 ftrace: Kill tracing_{start,stop}_sched_switch_record() and tracing_sched_switch_assign_trace()

In reply to Oleg Nesterov

tracing_{start,stop}_sched_switch_record() have no callers since 87d80de2800d "tracing: Remove obsolete sched_switch tracer".

The last caller of tracing_sched_switch_assign_trace() was removed by 30dbb20e68e6 "tracing: Remove boot tracer".

Signed-off-by Oleg Nesterov <[email protected]>

---
 kernel/trace/trace.h              |    3 --
 kernel/trace/trace_sched_switch.c |   48 -------------------------------------
 2 files changed, 0 insertions(+), 51 deletions(-)

kernel/trace/trace.h +0 -3

index 9258f5a..df2b33b 100644
@@ -595,9 +595,6 @@ void set_graph_array(struct trace_array *tr);
 
 void tracing_start_cmdline_record(void);
 void tracing_stop_cmdline_record(void);
-void tracing_sched_switch_assign_trace(struct trace_array *tr);
-void tracing_stop_sched_switch_record(void);
-void tracing_start_sched_switch_record(void);
 int register_tracer(struct tracer *type);
 int is_tracing_stopped(void);

kernel/trace/trace_sched_switch.c +0 ~~-48~~

51 lines not on this page · raw message

Signature · 1 line

1.5.5.1

Oleg Nesterov

author

2014-07-23 19:36 UTC #2 lore raw

PATCH 2/2 ftrace: Kill the dead code in probe_sched_switch() and probe_sched_wakeup()

In reply to Oleg Nesterov

After the previous patch it is clear that "tracer_enabled" can never be true, we can remove the "if (tracer_enabled)" code in probe_sched_switch() and probe_sched_wakeup(). Plus we can obviously remove tracer_enabled, ctx_trace, and sched_stopped as well.

Signed-off-by Oleg Nesterov <[email protected]>

---
 kernel/trace/trace_sched_switch.c |   40 -------------------------------------
 1 files changed, 0 insertions(+), 40 deletions(-)

kernel/trace/trace_sched_switch.c +0 ~~-40~~

index 3b60301..f7c7f4f 100644
@@ -14,12 +14,8 @@
 
 #include "trace.h"
 
-static struct trace_array	*ctx_trace;
-static int __read_mostly	tracer_enabled;
 static int			sched_ref;
 static DEFINE_MUTEX(sched_register_mutex);
-static int			sched_stopped;
-
 
 void
 tracing_sched_switch_trace(struct trace_array *tr,

@@ -52,29 +48,11 @@ tracing_sched_switch_trace(struct trace_array *tr,
 static void
 probe_sched_switch(void *ignore, struct task_struct *prev, struct task_struct *next)
 {
-	struct trace_array_cpu *data;
-	unsigned long flags;
-	int cpu;
-	int pc;
-
 	if (unlikely(!sched_ref))
 		return;
 
 	tracing_record_cmdline(prev);
 	tracing_record_cmdline(next);
-
-	if (!tracer_enabled || sched_stopped)
-		return;
-
-	pc = preempt_count();
-	local_irq_save(flags);
-	cpu = raw_smp_processor_id();
-	data = per_cpu_ptr(ctx_trace->trace_buffer.data, cpu);
-
-	if (likely(!atomic_read(&data->disabled)))
-		tracing_sched_switch_trace(ctx_trace, prev, next, flags, pc);
-
-	local_irq_restore(flags);
 }
 
 void

@@ -108,28 +86,10 @@ tracing_sched_wakeup_trace(struct trace_array *tr,
 static void
 probe_sched_wakeup(void *ignore, struct task_struct *wakee, int success)
 {
-	struct trace_array_cpu *data;
-	unsigned long flags;
-	int cpu, pc;
-
 	if (unlikely(!sched_ref))
 		return;
 
 	tracing_record_cmdline(current);
-
-	if (!tracer_enabled || sched_stopped)
-		return;
-
-	pc = preempt_count();
-	local_irq_save(flags);
-	cpu = raw_smp_processor_id();
-	data = per_cpu_ptr(ctx_trace->trace_buffer.data, cpu);
-
-	if (likely(!atomic_read(&data->disabled)))
-		tracing_sched_wakeup_trace(ctx_trace, wakee, current,
-					   flags, pc);
-
-	local_irq_restore(flags);
 }
 
 static int tracing_sched_register(void)

Signature · 1 line

1.5.5.1

Oleg Nesterov

author

2014-07-23 19:55 UTC #0 lore raw

cover ftrace: kill the dead code in trace_sched_switch.c

Quoted 4 lines · Oleg Nesterov

This motivated me to try to find something else unused. And it seems that almost all code in kernel/trace/trace_sched_switch.c is dead? Hmm, and probably even event_context_switch... I'll recheck and send the patch.

No, event_context_switch can't simply go away. But afaics the only user is tracing_sched_switch_trace(), and the only caller of this function is probe_wakeup_sched_switch(). Looks a bit strange.

After this patch we should probably move tracing_sched_switch_trace() and tracing_sched_wakeup_trace() from trace_sched_switch.c to trace_sched_wakeup.c and make them static. And perhaps we can also move the rest of this code somewhere else and rm this file...

Oleg.

 kernel/trace/trace.h              |    3 -
 kernel/trace/trace_sched_switch.c |   88 -------------------------------------
 2 files changed, 0 insertions(+), 91 deletions(-)