[Intel-gfx] [PATCH] Revert "drm/i915: Drop forcewake w/a for missed interrupts/seqno on Sandybridge"

[Chris Wilson]

This reverts commit 67e5871be82fec1451801d448b51d9a403d1ffac.

The missed interrupts are back, in a big fashion. I am observing stalls
followed by the missing interrupt warning on every Sandybridge+ machine
I have (a mixture of Sandybridge, Ivybridge and Haswell).

Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/intel_ringbuffer.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index db9c64b..e5d1845 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -1282,6 +1282,11 @@ gen6_ring_get_irq(struct intel_engine_cs *ring)
 	if (!dev->irq_enabled)
 	       return false;
 
+	/* It looks like we need to prevent the gt from suspending while waiting
+	 * for an notifiy irq, otherwise irqs seem to get lost on at least the
+	 * blt/bsd rings on ivb. */
+	gen6_gt_force_wake_get(dev_priv, FORCEWAKE_ALL);
+
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
 	if (ring->irq_refcount++ == 0) {
 		if (HAS_L3_DPF(dev) && ring->id == RCS)
@@ -1313,6 +1318,8 @@ gen6_ring_put_irq(struct intel_engine_cs *ring)
 		gen5_disable_gt_irq(dev_priv, ring->irq_enable_mask);
 	}
 	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+
+	gen6_gt_force_wake_put(dev_priv, FORCEWAKE_ALL);
 }
 
 static bool
-- 
1.9.1