Re: [PATCH tip/core/rcu 0/2] rcu: fixes for RCU lockdep

[Ingo Molnar <mingo@elte.hu>]

* Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:

> On Fri, Feb 26, 2010 at 12:50:18PM +0100, Ingo Molnar wrote:
> > 
> > * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> > 
> > > Hello!
> > > 
> > > This set fixes a couple of problems with the RCU-based lockdep patches:
> > > 
> > > 1.	Make the lockdep_rcu_dereference() message less alarmist
> > > 
> > > 2.	Make rcu_read_lock_sched_held() take boot-time lack of
> > > 	preemption into account.
> > > 
> > > 							Thanx, Paul
> > > 
> > > ------------------------------------------------------------------------
> > > 
> > >  include/linux/rcupdate.h |    4 +++-
> > >  include/linux/rcutiny.h  |    4 ----
> > >  include/linux/rcutree.h  |    1 -
> > >  kernel/lockdep.c         |    7 +++----
> > >  kernel/rcupdate.c        |   18 ++++++++++++++++++
> > >  kernel/rcutree.c         |   20 --------------------
> > >  6 files changed, 24 insertions(+), 30 deletions(-)
> > 
> > Applied, thanks Paul!
> > 
> > Here's a new warning:
> > 
> > [    0.000000] 
> > [    0.000000] ===================================================
> > [    0.000000] [ INFO: suspicious rcu_dereference_check() usage. ]
> > [    0.000000] ---------------------------------------------------
> > [    0.000000] kernel/sched_fair.c:3469 invoked rcu_dereference_check() without protection!
> 
> Hmmm...  Could you please tell me what commit this test ran against?

One version of tip:master :-/

ob'kconfig - i think this header bit of the .config:

> > # Automatically generated make config: don't edit
> > # Linux kernel version: 2.6.33
> > # Fri Feb 26 14:51:28 2010

should contain the last-seen kernel Git sha1 as well. It would ease things 
considerably.

> My guesses thus far either don't have that many lines in kernel/sched_fair.c 
> or have some unrelated code at line 3469.

I'll post the sha1 next time i encounter this warning.

Thanks,

	Ingo