Re: [PATCH v2] ftrace: Add a C-state tracer to help power optimization

[Ingo Molnar <mingo@elte.hu>]

* Steven Rostedt <rostedt@goodmis.org> wrote:

> 
> My comments are more towards Ingo than to Arjan.
> 
> On Sat, 4 Oct 2008, Arjan van de Ven wrote:
> > +++ b/kernel/trace/Kconfig
> > @@ -134,6 +134,17 @@ config BOOT_TRACER
> >  	  be enabled if this tracer is selected since only one tracer
> >  	  should touch the tracing buffer at a time.
> >  
> > +config CSTATE_TRACER
> > +	bool "Trace C-state behavior"
> > +	depends on HAVE_FTRACE
> 
> Ingo, here's the confusion again between the FTRACE infrastructure and 
> the ftrace "function tracer". the HAVE_FTRACE should really be 
> HAVE_FUNCTION_TRACER. The infrastructure of a lot of the ftrace 
> tracers do not depend on the function tracer.
> 
> The CSTATE_TRACER also does not depend on HAVE_FTRACE. Think we should 
> rename this? Otherwise more kernel developers will get confused by it. 
> :-(

agreed, please send a rename patch to make it more consistent.

> > +static enum print_line_t cstate_print_line(struct trace_iterator *iter)
> > +{
> > +	int ret;
> > +	struct trace_entry *entry = iter->ent;
> > +	struct trace_cstate *field = (struct trace_cstate *)entry;
> 
> Ingo, isn't the trace_assign_type in your latest tree?

yeah, it is, and it should be used here instead of a direct cast.

	Ingo