Messages in this thread |  | | | Date | Thu, 11 Feb 2010 19:57:48 +0100 | | From | Frederic Weisbecker <> | | Subject | Re: [PATCH 02/11] tracing: Introduce TRACE_EVENT_INJECT |
| |
On Wed, Feb 10, 2010 at 09:05:06AM -0500, Steven Rostedt wrote: > On Wed, 2010-02-10 at 11:04 +0100, Frederic Weisbecker wrote: > > > > We can add a register_event_injector later. For now, why not just add > > > the TRACE_EVENT() and then hook to it in perf using the normal > > > tracepoint mechanism. > > > > > > You could add some macro around the trace_init_lock_class() call that > > > would facilitate finding all the locks you need. This would probably be > > > a bit more straight forward than to overload TRACE_EVENT() again. > > > > > > > > I'm not sure what you mean by that. > > Looking back at where trace_lock_class_init() was placed, you don't need > a wrapper. But instead, why not just have perf hook directly into the > trace_lock_class_init trace point, and do its injection directly. > Instead of needing to add some hack to TRACE_EVENT()?
We could. But we expect further events may need such catch-up by injection system. And ftrace will also need to perform this injection.
That's why I think something more generic would be more reasonable.
|  |