drm/i915: Separate out reset flags from the reset counter

In preparation for introducing a per-engine reset, we can first separate
the mixing of the reset state from the global reset counter.

The loss of atomicity in updating the reset state poses a small problem
for handling the waiters. For requests, this is solved by advancing the
seqno so that a waiter waking up after the reset knows the request is
complete. For pending flips, we still rely on the increment of the
global reset epoch (as well as the reset-in-progress flag) to signify
when the hardware was reset.

The advantage, now that we do not inspect the reset state during reset
itself i.e. we no longer emit requests during reset, is that we can use
the atomic updates of the state flags to ensure that only one reset
worker is active.

v2: Mika spotted that I transformed the i915_gem_wait_for_error() wakeup
into a waiter wakeup.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Arun Siluvery <arun.siluvery@linux.intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1470414607-32453-6-git-send-email-arun.siluvery@linux.intel.com
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160909131201.16673-7-chris@chris-wilson.co.uk

1 files changed, 46 insertions, 57 deletions

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 82358d4e0cc2..ed172d7beecb 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -2501,53 +2501,41 @@ static void i915_reset_and_wakeup(struct drm_i915_private *dev_priv)
 
 	kobject_uevent_env(kobj, KOBJ_CHANGE, error_event);
 
+	DRM_DEBUG_DRIVER("resetting chip\n");
+	kobject_uevent_env(kobj, KOBJ_CHANGE, reset_event);
+
 	/*
-	 * Note that there's only one work item which does gpu resets, so we
-	 * need not worry about concurrent gpu resets potentially incrementing
-	 * error->reset_counter twice. We only need to take care of another
-	 * racing irq/hangcheck declaring the gpu dead for a second time. A
-	 * quick check for that is good enough: schedule_work ensures the
-	 * correct ordering between hang detection and this work item, and since
-	 * the reset in-progress bit is only ever set by code outside of this
-	 * work we don't need to worry about any other races.
+	 * In most cases it's guaranteed that we get here with an RPM
+	 * reference held, for example because there is a pending GPU
+	 * request that won't finish until the reset is done. This
+	 * isn't the case at least when we get here by doing a
+	 * simulated reset via debugs, so get an RPM reference.
 	 */
-	if (i915_reset_in_progress(&dev_priv->gpu_error)) {
-		DRM_DEBUG_DRIVER("resetting chip\n");
-		kobject_uevent_env(kobj, KOBJ_CHANGE, reset_event);
-
-		/*
-		 * In most cases it's guaranteed that we get here with an RPM
-		 * reference held, for example because there is a pending GPU
-		 * request that won't finish until the reset is done. This
-		 * isn't the case at least when we get here by doing a
-		 * simulated reset via debugs, so get an RPM reference.
-		 */
-		intel_runtime_pm_get(dev_priv);
+	intel_runtime_pm_get(dev_priv);
 
-		intel_prepare_reset(dev_priv);
+	intel_prepare_reset(dev_priv);
 
-		/*
-		 * All state reset _must_ be completed before we update the
-		 * reset counter, for otherwise waiters might miss the reset
-		 * pending state and not properly drop locks, resulting in
-		 * deadlocks with the reset work.
-		 */
-		ret = i915_reset(dev_priv);
+	/*
+	 * All state reset _must_ be completed before we update the
+	 * reset counter, for otherwise waiters might miss the reset
+	 * pending state and not properly drop locks, resulting in
+	 * deadlocks with the reset work.
+	 */
+	ret = i915_reset(dev_priv);
 
-		intel_finish_reset(dev_priv);
+	intel_finish_reset(dev_priv);
 
-		intel_runtime_pm_put(dev_priv);
+	intel_runtime_pm_put(dev_priv);
 
-		if (ret == 0)
-			kobject_uevent_env(kobj,
-					   KOBJ_CHANGE, reset_done_event);
+	if (ret == 0)
+		kobject_uevent_env(kobj,
+				   KOBJ_CHANGE, reset_done_event);
 
-		/*
-		 * Note: The wake_up also serves as a memory barrier so that
-		 * waiters see the update value of the reset counter atomic_t.
-		 */
-		wake_up_all(&dev_priv->gpu_error.reset_queue);
-	}
+	/*
+	 * Note: The wake_up also serves as a memory barrier so that
+	 * waiters see the updated value of the dev_priv->gpu_error.
+	 */
+	wake_up_all(&dev_priv->gpu_error.reset_queue);
 }
 
 static void i915_report_and_clear_eir(struct drm_i915_private *dev_priv)
@@ -2666,25 +2654,26 @@ void i915_handle_error(struct drm_i915_private *dev_priv,
 	i915_capture_error_state(dev_priv, engine_mask, error_msg);
 	i915_report_and_clear_eir(dev_priv);
 
-	if (engine_mask) {
-		atomic_or(I915_RESET_IN_PROGRESS_FLAG,
-				&dev_priv->gpu_error.reset_counter);
+	if (!engine_mask)
+		return;
 
-		/*
-		 * Wakeup waiting processes so that the reset function
-		 * i915_reset_and_wakeup doesn't deadlock trying to grab
-		 * various locks. By bumping the reset counter first, the woken
-		 * processes will see a reset in progress and back off,
-		 * releasing their locks and then wait for the reset completion.
-		 * We must do this for _all_ gpu waiters that might hold locks
-		 * that the reset work needs to acquire.
-		 *
-		 * Note: The wake_up serves as the required memory barrier to
-		 * ensure that the waiters see the updated value of the reset
-		 * counter atomic_t.
-		 */
-		i915_error_wake_up(dev_priv);
-	}
+	if (test_and_set_bit(I915_RESET_IN_PROGRESS,
+			     &dev_priv->gpu_error.flags))
+		return;
+
+	/*
+	 * Wakeup waiting processes so that the reset function
+	 * i915_reset_and_wakeup doesn't deadlock trying to grab
+	 * various locks. By bumping the reset counter first, the woken
+	 * processes will see a reset in progress and back off,
+	 * releasing their locks and then wait for the reset completion.
+	 * We must do this for _all_ gpu waiters that might hold locks
+	 * that the reset work needs to acquire.
+	 *
+	 * Note: The wake_up also provides a memory barrier to ensure that the
+	 * waiters see the updated value of the reset flags.
+	 */
+	i915_error_wake_up(dev_priv);
 
 	i915_reset_and_wakeup(dev_priv);
 }