Re: [PATCH -tip] v3 Consolidate sparse and lockdep declarations in include/linux/rcupdate.h

["Paul E. McKenney" <paulmck@linux.vnet.ibm.com>]

On Mon, Aug 24, 2009 at 03:34:46PM +0800, Lai Jiangshan wrote:
> Paul E. McKenney wrote:
> > From: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> > 
> > Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> > ---

[ . . . ]

> >   * rcu_read_lock_sched - mark the beginning of a RCU-classic critical section
> > @@ -160,10 +190,14 @@ extern int rcu_scheduler_active;
> >  static inline void rcu_read_lock_sched(void)
> >  {
> >  	preempt_disable();
> > +	__acquire(RCU_SCHED);
> > +	rcu_read_acquire();
> >  }
> >  static inline void rcu_read_lock_sched_notrace(void)
> 
> It need be marked "notrace" too.

I guess I just need to grep for "_notrace(" in my code.  ;-)

> >  {
> >  	preempt_disable_notrace();
> > +	__acquire(RCU_SCHED);
> > +	rcu_read_acquire();
> >  }
> 
> It may cause infinity recursion.
> rcu_read_acquire() calls rcu_read_lock_sched_notrace()
> before current->lockdep_recursion is set to 1 when tracing in on,
> thus infinity recursion occurs.

That might explain Ingo's test results.

Thank you again, good catch!!!

							Thanx, Paul