Re: linux-next-20110923: warning kernel/rcutree.c:1833

[Frederic Weisbecker <fweisbec@gmail.com>]

On Mon, Sep 26, 2011 at 03:34:26PM -0700, Paul E. McKenney wrote:
> On Mon, Sep 26, 2011 at 11:39:41AM +0200, Frederic Weisbecker wrote:
> > On Sun, Sep 25, 2011 at 06:41:18PM -0700, Paul E. McKenney wrote:
> > > On Sun, Sep 25, 2011 at 06:26:11PM -0700, Paul E. McKenney wrote:
> > > > On Mon, Sep 26, 2011 at 03:10:33AM +0200, Frederic Weisbecker wrote:
> > > > > 2011/9/26 Frederic Weisbecker <fweisbec@gmail.com>:
> > > > > > On Sun, Sep 25, 2011 at 09:48:04AM -0700, Paul E. McKenney wrote:
> > > > > >> This is required for RCU_FAST_NO_HZ, which checks to see whether the
> > > > > >> current CPU can accelerate the current grace period so as to enter
> > > > > >> dyntick-idle mode sooner than it would otherwise.  This takes effect
> > > > > >> in the situation where rcu_needs_cpu() sees that there are callbacks.
> > > > > >> It then notes a quiescent state (which is illegal in an RCU read-side
> > > > > >> critical section), calls force_quiescent_state(), and so on.  For this
> > > > > >> to work, the current CPU must be in an RCU read-side critical section.
> > > > > >
> > > > > > You mean it must *not* be in an RCU read-side critical section (ie: in a
> > > > > > quiescent state)?
> > > > > >
> > > > > > That assumption at least fails anytime in idle for the RCU
> > > > > > sched flavour given that preemption is disabled in the idle loop.
> > > > > >
> > > > > >> If this cannot be made to work, another option is to call a new RCU
> > > > > >> function in the case where rcu_needs_cpu() returned false, but after
> > > > > >> the RCU read-side critical section has exited.
> > > > > >
> > > > > > You mean when rcu_needs_cpu() returns true (when we have callbacks
> > > > > > enqueued)?
> > > > > >
> > > > > >> This new RCU function
> > > > > >> could then attempt to rearrange RCU so as to allow the CPU to enter
> > > > > >> dyntick-idle mode more quickly.  It is more important for this to
> > > > > >> happen when the CPU is going idle than when it is executing a user
> > > > > >> process.
> > > > > >>
> > > > > >> So, is this doable?
> > > > > >
> > > > > > At least not when we have RCU sched callbacks enqueued, given preemption
> > > > > > is disabled. But that sounds plausible in order to accelerate the switch
> > > > > > to dyntick-idle mode when we only have rcu and/or rcu bh callbacks.
> > > > > 
> > > > > But the RCU sched case could be dealt with if we embrace every use of
> > > > > it with rcu_read_lock_sched() and rcu_read_unlock_sched(), or some light
> > > > > version that just increases a local counter that rcu_needs_cpu() could check.
> > > > > 
> > > > > It's an easy thing to add: we can ensure preempt is disabled when we call it
> > > > > and we can force rcu_dereference_sched() to depend on it.
> > > > 
> > > > Or just check to see if this is the first level of interrupt from the
> > > > idle task after the scheduler is up.
> > > 
> > > Hmmm...  Is it the case that rcu_needs_cpu() gets called from within an
> > > RCU read-side critical section only when called from an interrupt that
> > > interrupted an RCU read-side critical section (keeping in mind that the
> > > idle loop is a quiescent state regardless of preemption)?
> > 
> > Yeah. rcu_needs_cpu() can be called from an irq that either interrupted
> > an rcu read side critical section or a bh one. But not a sched one if
> > we forbid rcu sched uses in the preempt offset race windows I described
> > in a previous mail.
> 
> But can't I just assume that if rcu_needs_cpu is invoked within
> a second-level interrupt handler that it might be in any type of
> RCU read-side critical section?  I could determine this by checking
> RCU's dyntick-idle nesting state.

No, rcu_needs_cpu() can only be called from the first level of interrupt.

> 
> Such checks are not necessary if CONFIG_NO_HZ=n because in that
> case rcu_needs_cpu() is just checking the callback queues, with
> no assumptions about quiescent states.

I believe it's not even called when CONFIG_NO_HZ=n

> 
> > > If so, I should be able to do the appropriate checks within
> > > rcu_needs_cpu().
> > 
> > Right. But to know if you interrupted an rcu read side, don't you
> > need a specific counter when !CONFIG_PREEMPT?
> 
> Not if it is OK to assume that rcu_needs_cpu() can only be called from
> within an RCU read-side interrupt handler if it is invoked from within a
> second-level interrupt handler or if it interrupted some non-dyntick-idle
> process-level code.
> 
> So, is this assumption valid?

Not sure I understand what you mean. But currently it can only be called from:

- idle
- first interrupt level, interrupting idle, but at a time where in_interrupt() returns 0

With idle beeing or not in extended quiescent state.