drm/i915: Always run hangcheck while the GPU is busy

Web:        https://git.kernel.org/torvalds/c/b26a32a82a901f894c79d1ec8b1ac94dde83e83c
Commit:     b26a32a82a901f894c79d1ec8b1ac94dde83e83c
Parent:     b5a756a722286af9702d565501e1f690d075d16b
Refname:    refs/heads/master
Author:     Chris Wilson <[email protected]>
AuthorDate: Mon Jan 29 14:41:04 2018 +0000
Committer:  Rodrigo Vivi <[email protected]>
CommitDate: Thu Feb 1 07:33:02 2018 -0800

    drm/i915: Always run hangcheck while the GPU is busy
    
    Previously, we relied on only running the hangcheck while somebody was
    waiting on the GPU, in order to minimise the amount of time hangcheck
    had to run. (If nobody was watching the GPU, nobody would notice if the
    GPU wasn't responding -- eventually somebody would care and so kick
    hangcheck into action.) However, this falls apart from around commit
    4680816be336 ("drm/i915: Wait first for submission, before waiting for
    request completion"), as not all waiters declare themselves to hangcheck
    and so we could switch off hangcheck and miss GPU hangs even when
    waiting under the struct_mutex.
    
    If we enable hangcheck from the first request submission, and let it run
    until the GPU is idle again, we forgo all the complexity involved with
    only enabling around waiters. We just have to remember to be careful that
    we do not declare a GPU hang when idly waiting for the next request to
    be come ready, as we will run hangcheck continuously even when the
    engines are stalled waiting for external events. This should be true
    already as we should only be tracking requests submitted to hardware for
    execution as an indicator that the engine is busy.
    
    Fixes: 4680816be336 ("drm/i915: Wait first for submission, before waiting for request completion"
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104840
    Signed-off-by: Chris Wilson <[email protected]>
    Cc: Chris Wilson <[email protected]>
    Cc: Mika Kuoppala <[email protected]>
    Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
    Reviewed-by: Mika Kuoppala <[email protected]>
    (cherry picked from commit 889230489b6b138ba97ba2f13fc9644a3d16d0d2)
    Signed-off-by: Rodrigo Vivi <[email protected]>
---
 drivers/gpu/drm/i915/i915_gem.c          |  7 +++----
 drivers/gpu/drm/i915/i915_gem_request.c  |  2 ++
 drivers/gpu/drm/i915/intel_breadcrumbs.c | 11 -----------
 drivers/gpu/drm/i915/intel_hangcheck.c   |  7 +------
 4 files changed, 6 insertions(+), 21 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 1135a77b383a..dd89abd2263d 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -3323,16 +3323,15 @@ i915_gem_retire_work_handler(struct work_struct *work)
 		mutex_unlock(&dev->struct_mutex);
 	}
 
-	/* Keep the retire handler running until we are finally idle.
+	/*
+	 * Keep the retire handler running until we are finally idle.
 	 * We do not need to do this test under locking as in the worst-case
 	 * we queue the retire worker once too often.
 	 */
-	if (READ_ONCE(dev_priv->gt.awake)) {
-		i915_queue_hangcheck(dev_priv);
+	if (READ_ONCE(dev_priv->gt.awake))
 		queue_delayed_work(dev_priv->wq,
 				   &dev_priv->gt.retire_work,
 				   round_jiffies_up_relative(HZ));
-	}
 }
 
 static inline bool
diff --git a/drivers/gpu/drm/i915/i915_gem_request.c b/drivers/gpu/drm/i915/i915_gem_request.c
index d575109f7a7f..e09d18df8b7f 100644
--- a/drivers/gpu/drm/i915/i915_gem_request.c
+++ b/drivers/gpu/drm/i915/i915_gem_request.c
@@ -276,6 +276,8 @@ static void mark_busy(struct drm_i915_private *i915)
 
 	intel_engines_unpark(i915);
 
+	i915_queue_hangcheck(i915);
+
 	queue_delayed_work(i915->wq,
 			   &i915->gt.retire_work,
 			   round_jiffies_up_relative(HZ));
diff --git a/drivers/gpu/drm/i915/intel_breadcrumbs.c b/drivers/gpu/drm/i915/intel_breadcrumbs.c
index 58c624f982d9..bd40fea16b4f 100644
--- a/drivers/gpu/drm/i915/intel_breadcrumbs.c
+++ b/drivers/gpu/drm/i915/intel_breadcrumbs.c
@@ -149,17 +149,6 @@ static void intel_breadcrumbs_fake_irq(struct timer_list *t)
 		return;
 
 	mod_timer(&b->fake_irq, jiffies + 1);
-
-	/* Ensure that even if the GPU hangs, we get woken up.
-	 *
-	 * However, note that if no one is waiting, we never notice
-	 * a gpu hang. Eventually, we will have to wait for a resource
-	 * held by the GPU and so trigger a hangcheck. In the most
-	 * pathological case, this will be upon memory starvation! To
-	 * prevent this, we also queue the hangcheck from the retire
-	 * worker.
-	 */
-	i915_queue_hangcheck(engine->i915);
 }
 
 static void irq_enable(struct intel_engine_cs *engine)
diff --git a/drivers/gpu/drm/i915/intel_hangcheck.c b/drivers/gpu/drm/i915/intel_hangcheck.c
index 31f01d64c021..348a4f7ffb67 100644
--- a/drivers/gpu/drm/i915/intel_hangcheck.c
+++ b/drivers/gpu/drm/i915/intel_hangcheck.c
@@ -411,7 +411,6 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
 	struct intel_engine_cs *engine;
 	enum intel_engine_id id;
 	unsigned int hung = 0, stuck = 0;
-	int busy_count = 0;
 
 	if (!i915_modparams.enable_hangcheck)
 		return;
@@ -429,7 +428,6 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
 	intel_uncore_arm_unclaimed_mmio_detection(dev_priv);
 
 	for_each_engine(engine, dev_priv, id) {
-		const bool busy = intel_engine_has_waiter(engine);
 		struct intel_engine_hangcheck hc;
 
 		semaphore_clear_deadlocks(dev_priv);
@@ -443,16 +441,13 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
 			if (hc.action != ENGINE_DEAD)
 				stuck |= intel_engine_flag(engine);
 		}