On Sat, Oct 05, 2024 at 02:56:26PM -0400, Mathieu Desnoyers wrote:
> On 2024-10-05 18:07, Peter Zijlstra wrote:
> > On Sat, Oct 05, 2024 at 06:04:44PM +0200, Peter Zijlstra wrote:
> > > On Fri, Oct 04, 2024 at 02:27:33PM -0400, Mathieu Desnoyers wrote:
> >
> > > > +void hp_scan(struct hp_slot __percpu *percpu_slots, void *addr,
> > > > + void (*retire_cb)(int cpu, struct hp_slot *slot, void
> > > > *addr))
> > > > +{
> > > > + int cpu;
> > > > +
> > > > + /*
> > > > + * Store A precedes hp_scan(): it unpublishes addr (sets it to
> > > > + * NULL or to a different value), and thus hides it from hazard
> > > > + * pointer readers.
> > > > + */
> >
> > This should probably assert we're in a preemptible context. Otherwise
> > people will start using this in non-preemptible context and then we get
> > to unfuck things later.
>
> Something like this ?
>
> + /* Should only be called from preemptible context. */
> + WARN_ON_ONCE(in_atomic());
lockdep_assert_preemption_enabled();
that also checks local IRQ state IIRC.
