drm/i915: use semaphores for the display plane

In theory this will have performance and power improvements. Performance
because we don't need to stall when the scanout BO is busy, and power
because we don't have to stall when the BO is busy (and the ring can
even go to sleep if the HW supports it).

v2:
squash 2 patches into 1 (me)
un-inline the enable_semaphores function (Daniel)
remove comment about SNB hangs from i915_gem_object_sync (Chris)
rename intel_enable_semaphores to i915_semaphore_is_enabled (me)
removed page flip comment; "no why" (Chris)

To address other comments from Daniel (irc):
update the comment to say 'vt-d is crap, don't enable semaphores'
  - I think you misinterpreted Chris' comment, it already exists.
checking out whether we can pageflip on the render ring on ivb (didn't
work on early silicon)
  - We don't want to enable workarounds for early silicon unless we have
    to.
  - I can't find any references in the docs about this.
optionally use it if the fb is already busy on the render ring
  - This should be how the code already worked, unless I am
    misunderstanding your meaning.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

4 files changed, 64 insertions, 66 deletions

diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index c33b0a4..96f8efc 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -394,6 +394,21 @@ void intel_detect_pch(struct drm_device *dev)
 	}
 }
 
+bool i915_semaphore_is_enabled(struct drm_device *dev)
+{
+	if (INTEL_INFO(dev)->gen < 6)
+		return 0;
+
+	if (i915_semaphores >= 0)
+		return i915_semaphores;
+
+	/* Enable semaphores on SNB when IO remapping is off */
+	if (INTEL_INFO(dev)->gen == 6)
+		return !intel_iommu_enabled;
+
+	return 1;
+}
+
 void __gen6_gt_force_wake_get(struct drm_i915_private *dev_priv)
 {
 	int count;
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 9bc64d2..7dcdccb 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -38,6 +38,7 @@
 #include <linux/i2c-algo-bit.h>
 #include <drm/intel-gtt.h>
 #include <linux/backlight.h>
+#include <linux/intel-iommu.h>
 
 /* General customization:
  */
@@ -1230,6 +1231,8 @@ void i915_gem_lastclose(struct drm_device *dev);
 
 int __must_check i915_mutex_lock_interruptible(struct drm_device *dev);
 int __must_check i915_gem_object_wait_rendering(struct drm_i915_gem_object *obj);
+int i915_gem_object_sync(struct drm_i915_gem_object *obj,
+			 struct intel_ring_buffer *to);
 void i915_gem_object_move_to_active(struct drm_i915_gem_object *obj,
 				    struct intel_ring_buffer *ring,
 				    u32 seqno);
@@ -1439,6 +1442,7 @@ extern void gen6_set_rps(struct drm_device *dev, u8 val);
 extern void intel_detect_pch(struct drm_device *dev);
 extern int intel_trans_dp_port_sel(struct drm_crtc *crtc);
 
+extern bool i915_semaphore_is_enabled(struct drm_device *dev);
 extern void __gen6_gt_force_wake_get(struct drm_i915_private *dev_priv);
 extern void __gen6_gt_force_wake_mt_get(struct drm_i915_private *dev_priv);
 extern void __gen6_gt_force_wake_put(struct drm_i915_private *dev_priv);
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index ad71751..d75a657 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1953,6 +1953,48 @@ i915_gem_object_wait_rendering(struct drm_i915_gem_object *obj)
 	return 0;
 }
 
+int
+i915_gem_object_sync(struct drm_i915_gem_object *obj,
+		     struct intel_ring_buffer *to)
+{
+	struct intel_ring_buffer *from = obj->ring;
+	u32 seqno;
+	int ret, idx;
+
+	if (from == NULL || to == from)
+		return 0;
+
+	if (!i915_semaphore_is_enabled(obj->base.dev))
+		return i915_gem_object_wait_rendering(obj);
+
+	idx = intel_ring_sync_index(from, to);
+
+	seqno = obj->last_rendering_seqno;
+	if (seqno <= from->sync_seqno[idx])
+		return 0;
+
+	if (seqno == from->outstanding_lazy_request) {
+		struct drm_i915_gem_request *request;
+
+		request = kzalloc(sizeof(*request), GFP_KERNEL);
+		if (request == NULL)
+			return -ENOMEM;
+
+		ret = i915_add_request(from, NULL, request);
+		if (ret) {
+			kfree(request);
+			return ret;
+		}
+
+		seqno = request->seqno;
+	}
+
+	from->sync_seqno[idx] = seqno;
+
+	return to->sync_to(to, from, seqno - 1);
+
+}
+
 static void i915_gem_object_finish_gtt(struct drm_i915_gem_object *obj)
 {
 	u32 old_write_domain, old_read_domains;
@@ -2926,11 +2968,6 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
  * Prepare buffer for display plane (scanout, cursors, etc).
  * Can be called from an uninterruptible phase (modesetting) and allows
  * any flushes to be pipelined (for pageflips).
- *
- * For the display plane, we want to be in the GTT but out of any write
- * domains. So in many ways this looks like set_to_gtt_domain() apart from the
- * ability to pipeline the waits, pinning and any additional subtleties
- * that may differentiate the display plane from ordinary buffers.
  */
 int
 i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj,
@@ -2945,8 +2982,8 @@ i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj,
 		return ret;
 
 	if (pipelined != obj->ring) {
-		ret = i915_gem_object_wait_rendering(obj);
-		if (ret == -ERESTARTSYS)
+		ret = i915_gem_object_sync(obj, pipelined);
+		if (ret)
 			return ret;
 	}
 
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 56d1001..2a24d0c 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -835,64 +835,6 @@ i915_gem_execbuffer_flush(struct drm_device *dev,
 	return 0;
 }
 
-static bool
-intel_enable_semaphores(struct drm_device *dev)
-{
-	if (INTEL_INFO(dev)->gen < 6)
-		return 0;
-
-	if (i915_semaphores >= 0)
-		return i915_semaphores;
-
-	/* Disable semaphores on SNB */
-	if (INTEL_INFO(dev)->gen == 6)
-		return 0;
-
-	return 1;
-}
-
-static int
-i915_gem_execbuffer_sync_rings(struct drm_i915_gem_object *obj,
-			       struct intel_ring_buffer *to)
-{
-	struct intel_ring_buffer *from = obj->ring;
-	u32 seqno;
-	int ret, idx;
-
-	if (from == NULL || to == from)
-		return 0;
-
-	/* XXX gpu semaphores are implicated in various hard hangs on SNB */
-	if (!intel_enable_semaphores(obj->base.dev))
-		return i915_gem_object_wait_rendering(obj);
-
-	idx = intel_ring_sync_index(from, to);
-
-	seqno = obj->last_rendering_seqno;
-	if (seqno <= from->sync_seqno[idx])
-		return 0;
-
-	if (seqno == from->outstanding_lazy_request) {
-		struct drm_i915_gem_request *request;
-
-		request = kzalloc(sizeof(*request), GFP_KERNEL);
-		if (request == NULL)
-			return -ENOMEM;
-
-		ret = i915_add_request(from, NULL, request);
-		if (ret) {
-			kfree(request);
-			return ret;
-		}
-
-		seqno = request->seqno;
-	}
-
-	from->sync_seqno[idx] = seqno;
-
-	return to->sync_to(to, from, seqno - 1);
-}
-
 static int
 i915_gem_execbuffer_wait_for_flips(struct intel_ring_buffer *ring, u32 flips)
 {
@@ -954,7 +896,7 @@ i915_gem_execbuffer_move_to_gpu(struct intel_ring_buffer *ring,
 	}
 
 	list_for_each_entry(obj, objects, exec_list) {
-		ret = i915_gem_execbuffer_sync_rings(obj, ring);
+		ret = i915_gem_object_sync(obj, ring);
 		if (ret)
 			return ret;
 	}