Re: [PATCH -next] tracepoint: ipi: don't trace IPI on offline CPUs

[Sudeep Holla]

On 09/03/16 16:05, Steven Rostedt wrote:
> On Wed,  9 Mar 2016 12:22:22 +0000
> Sudeep Holla <sudeep.holla@xxxxxxx> wrote:
>
>
> > Hi Steven,
> >
> > I observed that in "include/linux/tracepoint.h", we have
> > #define __DO_TRACE(tp, proto, args, cond, prercu, postrcu)
> > ...
> >                   if (!cpu_online(raw_smp_processor_id()))
> >                           return;
> >
> >                   if (!(cond))
> >                           return;
> > ...
> >
> > where !cond check seems reduntant if it's cpu_online check.
> > So, does this patch handle the warning correctly or is there any better
> > way ? I did see few traces with same condition, just thought of checking
> > with you.
>
> Bah, I forgot that we have lockdep checks for when the event isn't
> enabled.

Yes I was about to ask you the same. I did further digging to check if I
was missing something after seeing your series[1] especially patch 2/12
(tracing: Remove duplicate checks for online CPUs)

> Can you try this patch:

It works. Thanks for the quick fix.

Tested-by: Sudeep Holla <sudeep.holla@xxxxxxx>

--
Regards,
Sudeep

[1] http://www.spinics.net/lists/kernel/msg2208604.html