Chris Wilson <[email protected]> writes: > Only the first request added to the execlist queue, can be submitted. If > this request is not the first request on the queue, it means that there > are already higher priority requests waiting upon the tasklet and > kicking it will make no difference. > > This is more relevant for a later patch, where we more eagerly try and > kick the tasklet to handle the submission of new requests. > > Signed-off-by: Chris Wilson <[email protected]> > Cc: Mika Kuoppala <[email protected]> > Cc: Tvrtko Ursulin <[email protected]
Reviewed-by: Mika Kuoppala <[email protected]> > --- > drivers/gpu/drm/i915/intel_lrc.c | 7 ++++--- > 1 file changed, 4 insertions(+), 3 deletions(-) > > diff --git a/drivers/gpu/drm/i915/intel_lrc.c > b/drivers/gpu/drm/i915/intel_lrc.c > index a46a05f0e79c..9bf8858ecdcf 100644 > --- a/drivers/gpu/drm/i915/intel_lrc.c > +++ b/drivers/gpu/drm/i915/intel_lrc.c > @@ -651,10 +651,11 @@ static void execlists_submit_request(struct > drm_i915_gem_request *request) > /* Will be called from irq-context when using foreign fences. */ > spin_lock_irqsave(&engine->timeline->lock, flags); > > - if (insert_request(&request->priotree, &engine->execlist_queue)) > + if (insert_request(&request->priotree, &engine->execlist_queue)) { > engine->execlist_first = &request->priotree.node; > - if (execlists_elsp_idle(engine)) > - tasklet_hi_schedule(&engine->irq_tasklet); > + if (execlists_elsp_idle(engine)) > + tasklet_hi_schedule(&engine->irq_tasklet); > + } > > spin_unlock_irqrestore(&engine->timeline->lock, flags); > } > -- > 2.11.0 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
