Re: [PATCH 2/3] tracing/profile: add ref count for registeringprofile events

[Steven Rostedt]

On Fri, 2009-09-11 at 16:37 +0200, Peter Zijlstra wrote:

> > I originally tried to do it in the tracepoint logic, but that broke a
> > lot of assumptions about tracepoints that Mathieu pointed out. This is
> > not a normal use of tracepoints. It is expected that if you register a
> > probe in a module, you will unregister it before exiting.
> > 
> > I can't remember all the details, but at the end, it seemed that the fix
> > belonged at the ftrace level.
> 
> Right, Mathieu thinks its sane to be able to attach to
> tracepoints/markers before they exist, so you can put them in module
> init code. I disagree.
> 
> ftrace doesn't mirror this behaviour, that is the source of the problem.
> If it did the ftrace structures wouldn't go away on unload and there
> wouldn't be no crash.
> 
> But if you want to maintain this disparity between the two frameworks
> then yes Li's patch, or yours (they're identical) seems the way to solve
> it.

I'll take Li's patch then. He wrote it first.

> 
> Still think its daft though.
> 

Well, it needs to be fixed, and I don't want to put up with a flamewar
about changing the way tracepoints currently work. I'll take Li's patch
and send out another pull request.

-- Steve


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/