[ltt-dev] LTTng finds abnormally long APIC interrupt handler : 58.2 ms

[Steven Rostedt]

On Thu, 7 Aug 2008, Frank Ch. Eigler wrote:
> Steven Rostedt <rostedt at goodmis.org> writes:
> >
> > I've been thinking of exporting it, and make it so that once you 
> > register something to the trace_ops, I'll up the module count, and never 
> > let that module unload.
> 
> This issue will need some sort of resolution if systemtap is to be
> able to hook up to ftrace.  Perhaps we could have a second/nested API
> that suffers the costs of explicit reference counting, but does permit
> dynamic attach and robust detach.

I've also thought about having a second registration method that would 
hook the mcount function to an internal one, one that is always there. 
Here we could add logic to know that a driver has registered a function. 
It will just have the added overhead of that logic.

It's not hard to do. I just didn't want to clutter ftrace too much before 
it made it into a major release. This means any patches to do this will 
happen after 27.

-- Steve