On Thu, 09 Nov 2017 15:17:28 +0000 Ben Hutchings <ben.hutchings@xxxxxxxxxxxxxxx> wrote: > On Mon, 2017-10-16 at 18:11 +0200, gregkh@xxxxxxxxxxxxxxxxxxx wrote: > > 4.4-stable review patch. If anyone has any objections, please let me know. > > > > ------------------ > > > > From: Paul E. McKenney <paulmck@xxxxxxxxxxxxxxxxxx> > > > > commit 28585a832602747cbfa88ad8934013177a3aae38 upstream. > > > > A number of architecture invoke rcu_irq_enter() on exception entry in > > order to allow RCU read-side critical sections in the exception handler > > when the exception is from an idle or nohz_full CPU. This works, at > > least unless the exception happens in an NMI handler. In that case, > > rcu_nmi_enter() would already have exited the extended quiescent state, > > which would mean that rcu_irq_enter() would (incorrectly) cause RCU > > to think that it is again in an extended quiescent state. This will > > in turn result in lockdep splats in response to later RCU read-side > > critical sections. > > > > This commit therefore causes rcu_irq_enter() and rcu_irq_exit() to > > take no action if there is an rcu_nmi_enter() in effect, thus avoiding > > the unscheduled return to RCU quiescent state. This in turn should > > make the kernel safe for on-demand RCU voyeurism. > > > > Link: http://lkml.kernel.org/r/20170922211022.GA18084@xxxxxxxxxxxxxxxxxx > > > > Cc: stable@xxxxxxxxxxxxxxx > > Fixes: 0be964be0 ("module: Sanitize RCU usage and locking") > > > Reported-by: Steven Rostedt <rostedt@xxxxxxxxxxx> > > > Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxxxxxxxxxx> > > > Signed-off-by: Steven Rostedt (VMware) <rostedt@xxxxxxxxxxx> > > > Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> > > --- > > kernel/rcu/tree.c | 12 ++++++++++++ > > 1 file changed, 12 insertions(+) > > > > --- a/kernel/rcu/tree.c > > +++ b/kernel/rcu/tree.c > > @@ -759,6 +759,12 @@ void rcu_irq_exit(void) > > > > local_irq_save(flags); > > rdtp = this_cpu_ptr(&rcu_dynticks); > > + > > + /* Page faults can happen in NMI handlers, so check... */ > > + if (READ_ONCE(rdtp->dynticks_nmi_nesting)) > > + return; > > Shouldn't there be a local_irq_restore() on this return path? Or does > this condition imply that IRQs were already disabled? It does, but there still should be a local_irq_restore(), because it will might confuse lockdep (lockdep keeps track of these), and it just looks bad. > > > + RCU_LOCKDEP_WARN(!irqs_disabled(), "rcu_irq_exit() invoked with irqs enabled!!!"); > > I don't see why you added RCU_LOCKDEP_WARN() here. Prior to 4.5 it's > not an error to call this function with IRQs disabled. And after > calling local_irq_save(), it's redundant to assert that IRQs are > disabled. good point. > > > oldval = rdtp->dynticks_nesting; > > rdtp->dynticks_nesting--; > > WARN_ON_ONCE(IS_ENABLED(CONFIG_RCU_EQS_DEBUG) && > > @@ -887,6 +893,12 @@ void rcu_irq_enter(void) > > > > local_irq_save(flags); > > rdtp = this_cpu_ptr(&rcu_dynticks); > > + > > + /* Page faults can happen in NMI handlers, so check... */ > > + if (READ_ONCE(rdtp->dynticks_nmi_nesting)) > > + return; > > + > > + RCU_LOCKDEP_WARN(!irqs_disabled(), "rcu_irq_enter() invoked with irqs enabled!!!"); > > Same problems here. -- Steve