Re: [PATCH tip/core/rcu 15/16] scripts: Make checkpatch.pl warn on expedited RCU grace periods

[Joe Perches <joe@perches.com>]

On Mon, 2015-07-20 at 18:29 -0700, Paul E. McKenney wrote:
> On Mon, Jul 20, 2015 at 06:06:19PM -0700, Joe Perches wrote:
> > On Mon, 2015-07-20 at 20:55 -0400, Steven Rostedt wrote:
> > > On Fri, 17 Jul 2015 15:37:54 -0700
> > > Joe Perches <joe@perches.com> wrote:
> > > 
> > > > > +		if ($line =~ /\b(synchronize_rcu_expedited|synchronize_sched_expedited)\(/) {
> > > > 
> > > > It'd be faster perl without capture groups:
> > > 
> > > Is checkpatch such a critical code path that we could possibly notice
> > > the difference in speed? Although the above may be few microseconds
> > > slower, to me, the above looks a hell of a lot more readable.
> > 
> > Regardless, it's not as capable as it wouldn't find
> > any miswritten form like
> > 
> > 	synchronize_<foo>_expedited ( args... )
> 
> Hmmm...  My current patch doesn't complain about this:
> 
> ------------------------------------------------------------------------
> 
> diff --git a/kernel/rcu/rcutorture.c b/kernel/rcu/rcutorture.c
> index 77192953dee5..89577b8d3ba6 100644
> --- a/kernel/rcu/rcutorture.c
> +++ b/kernel/rcu/rcutorture.c
> @@ -492,6 +492,8 @@ static void srcu_read_delay(struct torture_random_state *rrsp)
>  	const long uspertick = 1000000 / HZ;
>  	const long longdelay = 10;
>  
> +	synchronize_srcu_expedited(srcu_ctlp);
> +
>  	/* We want there to be long-running readers, but not all the time. */
>  
>  	delay = torture_random(rrsp) %
> 
> ------------------------------------------------------------------------
> 
> But it does correctly complain when I add synchronize_rcu_expedited()
> or synchronize_sched_expedited().
> 
> So what am I missing here?

A space between the function name and the open parenthesis