Re: [PATCH] tick/nohz: Narrow down noise while setting current task's tick dependency

From: Frederic Weisbecker
Date: Sun May 17 2020 - 12:20:19 EST


On Sun, May 17, 2020 at 08:53:22AM -0700, Paul E. McKenney wrote:
> On Sun, May 17, 2020 at 03:31:16PM +0200, Frederic Weisbecker wrote:
> > On Fri, May 15, 2020 at 08:07:18PM -0700, Paul E. McKenney wrote:
> > > On Fri, May 15, 2020 at 02:34:29AM +0200, Frederic Weisbecker wrote:
> > > > So far setting a tick dependency on any task, including current, used to
> > > > trigger an IPI to all CPUs. That's of course suboptimal but it wasn't
> > > > an issue as long as it was only used by posix-cpu-timers on nohz_full,
> > > > a combo that nobody seemed to use in real life.
> > > >
> > > > But RCU started to use task tick dependency on current task to fix
> > > > stall issues on callbacks processing. These trigger regular and
> > > > undesired system wide IPIs on nohz_full.
> > > >
> > > > The fix is very easy while setting a tick dependency on the current
> > > > task, only its CPU needs an IPI.
> > >
> > > This passes moderate rcutorture testing. If you want me to take it, please
> > > let me know, and otherwise:
> > >
> > > Tested-by: Paul E. McKenney <paulmck@xxxxxxxxxx>
> >
> > If you already have a pending urgent queue, I'd love you to take it.
> > If not I can take it.
>
> Nothing urgent yet in -rcu, so if you would like it in the next merge
> window, please take it through your normal upstream path.

Got it, thanks!