Re: [paulmck-rcu:dev.2023.11.08a] [EXP locktorture] 1254a620b4: WARNING:at_kernel/rcu/tree_stall.h:#rcu_stall_chain_notifier_register

[Paul E. McKenney]

On Tue, Dec 12, 2023 at 03:43:44PM +0800, Oliver Sang wrote:
> hi, Paul,
> 
> On Mon, Dec 11, 2023 at 08:26:12PM -0800, Paul E. McKenney wrote:
> > On Tue, Dec 12, 2023 at 10:06:36AM +0800, Oliver Sang wrote:
> > > hi, Paul,
> > > 
> > > On Mon, Dec 11, 2023 at 08:59:16AM -0800, Paul E. McKenney wrote:
> > > > On Thu, Dec 07, 2023 at 04:19:56PM +0800, kernel test robot wrote:
> > > > > 
> > > > > 
> > > > > Hello,
> > > > > 
> > > > > kernel test robot noticed "WARNING:at_kernel/rcu/tree_stall.h:#rcu_stall_chain_notifier_register" on:
> > > > > 
> > > > > commit: 1254a620b4a3832e65ac01bcef769b99e34515b2 ("EXP locktorture: Add RCU CPU stall-warning notifier stub")
> > > > > https://git.kernel.org/cgit/linux/kernel/git/paulmck/linux-rcu.git dev.2023.11.08a
> > > > 
> > > > Thank you for your testing efforts!
> > > > 
> > > > This one is expected behavior by explicit request from Linus Torvalds.
> > > > The concern is that people might use this hook without understanding
> > > > the risks of losing RCU CPU stall warnings.
> > > > 
> > > > One fix would be to never specify the rcupdate.rcu_cpu_stall_notifiers
> > > > kernel boot parameter.  Another would be to forgive this warning when
> > > > that boot parameter was specified.  Your choice!  ;-)
> > > 
> > > Thanks a lot for information!
> > > 
> > > this commit (1254a620b4) is a test for this warning, am I right?
> > > when this warning mechanism goes into upstream, do you want us still report
> > > for similar cases? or we could just ignore them? Thanks!
> > 
> > This 1254a620b4 ("EXP locktorture: Add RCU CPU stall-warning notifier
> > stub") commit is a debug-only use of this facility that will never go
> > upstream, as signified by the "EXP" at the beginning of the subject line.
> > 
> > Or is there some better way than "EXP" to mark commits that are not
> > intended for mainline?
> > 
> > 							Thanx, Paul
> >
> 
> sorry, seems I didn't state it very well. let me clarify.
> 
> the 'EXP' and "This not-for-mainline commit" is very good for us to know it
> will not go into mainline.
> 
> what I asked is if this warning is triggered by other usages (not this
> debug-only test), could we still ignore them by following below guidance you
> gave us?

Ah, got it!  I could imagine someone pushing out a not-for-mainline use
of this facility in order to allow others to run their test.  Actually,
I could also imagine me doing that.  ;-)

							Thanx, Paul

> > > > One fix would be to never specify the rcupdate.rcu_cpu_stall_notifiers
> > > > kernel boot parameter.  Another would be to forgive this warning when
> > > > that boot parameter was specified.  Your choice!  ;-)