We use the completion of the last active breadcrumb to retire the
requests along a timeline. This is purely opportunistic as nothing
guarantees that any particular timeline is terminated by a breadcrumb;
except for the parking the engine. We explicitly add a breadcrumb to
parking the engine so that we park quickly and do an explicit retire
upon signaling to reduce the latency dramatically.
With scheduling, we anticipate retiring completed timelines as a matter
of course. Performing the same action from inside the breadcrumbs is
intended to provide similar functionality for legacy ringbuffer
submission.
Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
---
drivers/gpu/drm/i915/gt/intel_breadcrumbs.c | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c b/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c
index 7137b6f24f55..6996e22ba65b 100644
--- a/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c
+++ b/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c
@@ -257,17 +257,19 @@ static void signal_irq_work(struct irq_work *work)
list_del_rcu(&rq->signal_link);
release = remove_signaling_context(b, ce);
spin_unlock(&ce->signal_lock);
+ if (release) {
+ if (list_is_last_rcu(&rq->link,
+ &ce->timeline->requests))
+ add_retire(b, ce->timeline);
+
+ intel_context_put(ce);
+ }
if (__dma_fence_signal(&rq->fence))
/* We own signal_node now, xfer to local list */
signal = slist_add(&rq->signal_node, signal);
else
i915_request_put(rq);
-
- if (release) {
- add_retire(b, ce->timeline);
- intel_context_put(ce);
- }
}
}
atomic_dec(&b->signaler_active);
--
2.20.1
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
