xiaomi-sm8450/android_kernel_xiaomi_sm8450
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

drm/i915: Drop unused engine->irq_seqno_barrier w/a

Browse Source 
Now that we have eliminated the CPU-side irq_seqno_barrier by moving the
delays on the GPU before emitting the MI_USER_INTERRUPT, we can remove
the engine->irq_seqno_barrier infrastructure. Though intentionally
slowing down the GPU is nasty, so is the code we can now remove!

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20181228171641.16531-6-chris@chris-wilson.co.uk
This commit is contained in:
Chris Wilson
2018-12-28 17:16:41 +00:00
parent 835051d397
commit 1216e3c3af
9 changed files with 1 additions and 161 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
drivers/gpu/drm/i915/intel_breadcrumbs.c
View File

@@ -166,12 +166,6 @@ static void irq_enable(struct intel_engine_cs *engine)
*/
GEM_BUG_ON(!intel_irqs_enabled(engine->i915));
/* Enabling the IRQ may miss the generation of the interrupt, but
* we still need to force the barrier before reading the seqno,
* just in case.
*/
set_bit(ENGINE_IRQ_BREADCRUMB, &engine->irq_posted);
/* Caller disables interrupts */
if (engine->irq_enable) {
spin_lock(&engine->i915->irq_lock);
@@ -683,16 +677,6 @@ static int intel_breadcrumbs_signaler(void *arg)
}
if (unlikely(do_schedule)) {
/* Before we sleep, check for a missed seqno */
if (current->state & TASK_NORMAL &&
!list_empty(&b->signals) &&
engine->irq_seqno_barrier &&
test_and_clear_bit(ENGINE_IRQ_BREADCRUMB,
&engine->irq_posted)) {
engine->irq_seqno_barrier(engine);
intel_engine_wakeup(engine);
}
sleep:
if (kthread_should_park())
kthread_parkme();
@@ -859,16 +843,6 @@ void intel_engine_reset_breadcrumbs(struct intel_engine_cs *engine)
else
irq_disable(engine);
/*
* We set the IRQ_BREADCRUMB bit when we enable the irq presuming the
* GPU is active and may have already executed the MI_USER_INTERRUPT
* before the CPU is ready to receive. However, the engine is currently
* idle (we haven't started it yet), there is no possibility for a
* missed interrupt as we enabled the irq and so we can clear the
* immediate wakeup (until a real interrupt arrives for the waiter).
*/
clear_bit(ENGINE_IRQ_BREADCRUMB, &engine->irq_posted);
spin_unlock_irqrestore(&b->irq_lock, flags);
}