Re: [RFC PATCH v2] Tracepoint: register/unregister struct tracepoint

[Mathieu Desnoyers <mathieu.desnoyers@efficios.com>]

----- Original Message -----
> From: "Frank Ch. Eigler" <fche@redhat.com>
> To: "Mathieu Desnoyers" <mathieu.desnoyers@efficios.com>
> Cc: "Steven Rostedt" <rostedt@goodmis.org>, linux-kernel@vger.kernel.org, "Ingo Molnar" <mingo@kernel.org>, "Frederic
> Weisbecker" <fweisbec@gmail.com>, "Andrew Morton" <akpm@linux-foundation.org>, "Johannes Berg"
> <johannes.berg@intel.com>
> Sent: Thursday, March 13, 2014 12:35:31 PM
> Subject: Re: [RFC PATCH v2] Tracepoint: register/unregister struct tracepoint
> 
> Hi -
> 
> On Thu, Mar 13, 2014 at 12:10:48PM -0400, Mathieu Desnoyers wrote:
> 
> > [...]  Moreover, tracers are responsible for unregistering the probe
> > before the module containing its associated tracepoint is unloaded.
> 
> Could you spell out please how a tracer is supposed to know early
> enough that the module is going to be unloaded?

There are two ways this can be done.

One use-case is when the probe and the callsite are within the same module,
or if the module containing the probe has a symbol dependency on the callsite.
In this case the probe normally unregisters itself from a module exit function
before the module unloads.

The other use-case is if the tracer has a module coming/going notifier
tracking the module's tracepoint callsites. The going notifier should
be run before the tracepoint.c going notifier. A notifier with negative
priority should have this effect, since the tracepoint.c notifier has
priority 0.

Thoughts ?

Thanks,

Mathieu

-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com