We can forgo queuing the hangcheck from the start of every request to
until we wait upon a request. This reduces the overhead of every
request, but may increase the latency of detecting a hang. Howeever, if
nothing every waits upon a hang, did it ever hang? It also improves the
robustness of the wait-request by ensuring that the hangchecker is
indeed running before we sleep indefinitely (and thereby ensuring that
we never actually sleep forever waiting for a dead GPU).
Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
---
drivers/gpu/drm/i915/i915_gem.c | 9 +++++----
drivers/gpu/drm/i915/i915_irq.c | 10 ++++------
2 files changed, 9 insertions(+), 10 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index c99d1b2c65d4..4de6eafaec28 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1310,6 +1310,9 @@ int __i915_wait_request(struct drm_i915_gem_request *req,
break;
}
+ /* Ensure that even if the GPU hangs, we get woken up. */
+ i915_queue_hangcheck(dev_priv);
+
timer.function = NULL;
if (timeout || missed_irq(dev_priv, engine)) {
unsigned long expire;
@@ -2650,8 +2653,6 @@ void __i915_add_request(struct drm_i915_gem_request *request,
/* Not allowed to fail! */
WARN(ret, "emit|add_request failed: %d!\n", ret);
- i915_queue_hangcheck(engine->i915);
-
queue_delayed_work(dev_priv->wq,
&dev_priv->mm.retire_work,
round_jiffies_up_relative(HZ));
@@ -2995,8 +2996,8 @@ i915_gem_retire_requests(struct drm_i915_private *dev_priv)
if (idle)
mod_delayed_work(dev_priv->wq,
- &dev_priv->mm.idle_work,
- msecs_to_jiffies(100));
+ &dev_priv->mm.idle_work,
+ msecs_to_jiffies(100));
return idle;
}
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 8864ee19154f..1544a5677ac8 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -3167,10 +3167,10 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
intel_uncore_arm_unclaimed_mmio_detection(dev_priv);
for_each_engine_id(engine, dev_priv, id) {
+ bool busy = waitqueue_active(&engine->irq_queue);
u64 acthd;
u32 seqno;
unsigned user_interrupts;
- bool busy = true;
semaphore_clear_deadlocks(dev_priv);
@@ -3193,12 +3193,11 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
if (engine->hangcheck.seqno == seqno) {
if (ring_idle(engine, seqno)) {
engine->hangcheck.action = HANGCHECK_IDLE;
- if (waitqueue_active(&engine->irq_queue)) {
+ if (busy) {
/* Safeguard against driver failure */
user_interrupts = kick_waiters(engine);
engine->hangcheck.score += BUSY;
- } else
- busy = false;
+ }
} else {
/* We always increment the hangcheck score
* if the ring is busy and still processing
@@ -3272,9 +3271,8 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
goto out;
}
+ /* Reset timer in case GPU hangs without another request being added */
if (busy_count)
- /* Reset timer case chip hangs without another request
- * being added */
i915_queue_hangcheck(dev_priv);
out:
--
2.8.1
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
