Re: [PATCH 1/2] tracing: Move enabling tracepoints to just after rcu_init()

[Wang Nan <wangnan0@huawei.com>]

Hi Steven Rostedt,

During studying your code we find a problem, please see below.

> 
> From: "Steven Rostedt (Red Hat)" <rostedt@goodmis.org>
> 
> Enabling tracepoints at boot up can be very useful. The tracepoint
> can be initialized right after RCU has been. There's no need to
> wait for the early_initcall() to be called. That's too late for some
> things that can use tracepoints for debugging. Move the logic to
> enable tracepoints out of the initcalls and into init/main.c to
> right after rcu_init().
> 
> This also allows trace_printk() to be used early too.
> 
> Link: http://lkml.kernel.org/r/alpine.DEB.2.11.1412121539300.16494@nanos
> Link: http://lkml.kernel.org/r/20141214164104.307127356@goodmis.org
> 
> Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> Suggested-by: Thomas Gleixner <tglx@linutronix.de>
> Tested-by: Thomas Gleixner <tglx@linutronix.de>
> Acked-by: Thomas Gleixner <tglx@linutronix.de>
> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>

[...]

> +void __init trace_init(void)
> +{
> +	tracer_alloc_buffers();
> +	init_ftrace_syscalls();
> +	trace_event_init();	
> +}
> +

[...]

> +
> +void __init trace_event_init(void)
> +{
> +	event_trace_memsetup();
> +	init_ftrace_syscalls();
> +	event_trace_enable();
> +}
> +

init_ftrace_syscalls() get called twice by trace_init() and trace_event_init(), some resources are wasted.
At lease one of them can be removed.

In addition, could you please have a look at my early kprobe patch series?

http://lists.infradead.org/pipermail/linux-arm-kernel/2015-January/313835.html

Which enables kprobe very early, even before memory initialized. I think it is possible to combine these
early tracing facilities together.

Thank you!