[Intel-gfx] [PATCH 07/20] drm/i915/gt: Shrink the critical section for irq signaling

[Chris Wilson]

Quoting Matthew Brost (2020-12-10 17:37:13)
> On Mon, Dec 07, 2020 at 07:38:11PM +0000, Chris Wilson wrote:
> > Let's only wait for the list iterator when decoupling the virtual
> > breadcrumb, as the signaling of all the requests may take a long time,
> > during which we do not want to keep the tasklet spinning.
> > 
> > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > ---
> >  drivers/gpu/drm/i915/gt/intel_breadcrumbs.c       | 2 ++
> >  drivers/gpu/drm/i915/gt/intel_breadcrumbs_types.h | 1 +
> >  drivers/gpu/drm/i915/gt/intel_lrc.c               | 3 ++-
> >  3 files changed, 5 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c b/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c
> > index 63900edbde88..ac1e5f6c3c2c 100644
> > --- a/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c
> > +++ b/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c
> > @@ -239,6 +239,7 @@ static void signal_irq_work(struct irq_work *work)
> >               intel_breadcrumbs_disarm_irq(b);
> >  
> >       rcu_read_lock();
> > +     atomic_inc(&b->signaler_active);
> >       list_for_each_entry_rcu(ce, &b->signalers, signal_link) {
> >               struct i915_request *rq;
> >  
> > @@ -274,6 +275,7 @@ static void signal_irq_work(struct irq_work *work)
> >                       }
> >               }
> >       }
> > +     atomic_dec(&b->signaler_active);
> >       rcu_read_unlock();
> >  
> >       llist_for_each_safe(signal, sn, signal) {
> > diff --git a/drivers/gpu/drm/i915/gt/intel_breadcrumbs_types.h b/drivers/gpu/drm/i915/gt/intel_breadcrumbs_types.h
> > index a74bb3062bd8..f672053d694d 100644
> > --- a/drivers/gpu/drm/i915/gt/intel_breadcrumbs_types.h
> > +++ b/drivers/gpu/drm/i915/gt/intel_breadcrumbs_types.h
> > @@ -35,6 +35,7 @@ struct intel_breadcrumbs {
> >       spinlock_t signalers_lock; /* protects the list of signalers */
> >       struct list_head signalers;
> >       struct llist_head signaled_requests;
> > +     atomic_t signaler_active;
> >  
> >       spinlock_t irq_lock; /* protects the interrupt from hardirq context */
> >       struct irq_work irq_work; /* for use from inside irq_lock */
> > diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c
> > index b3db16b2a5a4..35cded25c6c1 100644
> > --- a/drivers/gpu/drm/i915/gt/intel_lrc.c
> > +++ b/drivers/gpu/drm/i915/gt/intel_lrc.c
> > @@ -1401,7 +1401,8 @@ static void kick_siblings(struct i915_request *rq, struct intel_context *ce)
> >                * ce->signal_link.
> >                */
> >               i915_request_cancel_breadcrumb(rq);
> > -             irq_work_sync(&engine->breadcrumbs->irq_work);
> > +             while (atomic_read(&engine->breadcrumbs->signaler_active))
> > +                     cpu_relax();
> 
> Would a 'cond_resched' be better here? I trust your opinion on which to
> use but thought I'd mention it.

While by this point we've removed the spinlock from around us, we are
still inside a softirq bh. That will disallow sleeping. Fortunately,
we only have to busy wait on the list iteration.
-Chris