Re: [PATCH 2/4] irq_work: Fix irq_work_claim() ordering

From: Frederic Weisbecker
Date: Mon Nov 11 2019 - 18:17:11 EST


On Mon, Nov 11, 2019 at 08:20:05AM +0100, Ingo Molnar wrote:
>
> * Frederic Weisbecker <frederic@xxxxxxxxxx> wrote:
>
> > When irq_work_claim() finds IRQ_WORK_PENDING flag already set, we just
> > return and don't raise a new IPI. We expect the destination to see
> > and handle our latest updades thanks to the pairing atomic_xchg()
> > in irq_work_run_list().
> >
> > But cmpxchg() doesn't guarantee a full memory barrier upon failure. So
> > it's possible that the destination misses our latest updates.
> >
> > So use atomic_fetch_or() instead that is unconditionally fully ordered
> > and also performs exactly what we want here and simplify the code.
>
> Just curious, how was this bug found - in the wild, or via code review?

Well, I wanted to make sure the nohz kcpustat patches are safe and I had
a last minute doubt about that irq work scenario. So I would say code
review :)