Re: [PATCH 2/4] ftrace - add function_duration tracer

[Ingo Molnar <mingo@elte.hu>]

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

> On Thu, 2009-12-10 at 13:03 +0100, Frederic Weisbecker wrote:
> 
> > This makes me feel I'm going to try converting the function graph 
> > tracer into an event during the next cycle. It does not mean I could 
> > make it usable as a perf event right away in the same shot that 
> > said, as you can guess this is not a trivial plug. The current perf 
> > fast path is not yet adapted for that.
> 
> I curious how you plan on doing this. The current event system shows 
> one event per trace point. A straight forward approach would make 
> every entry and exit of a function a trace point and that would lead 
> to a very large kernel to handle that.
> 
> Perhaps we could abstract out all entries and exits. We need to be 
> able to link to a single point (entry or exit) not all.  This also has 
> the added issue of using the ftrace infrastructure of nop the mcount 
> call. We also need a way to enable a set of functions.

I think a good solution would be what i suggested (and what the patch i 
sent does in essence), i.e. to add a special software event for 
functions:

  PERF_COUNT_SW_FUNCTION_CALLS

Next step would be to extend the ioctl interface to allow per function 
tracepoints to be turned on/off (and set other per function attributes). 
The attributes would be system-wide (at least initially) - we dont want 
to live patch functions at context switch time.

We'd basically keep the function tracer, just re-wire it somewhat and 
expose it via the unified interface and make it available to a much 
larger set of usecases.

	Ingo