[Intel-gfx] [PATCH 11/11] drm/i915: Maintain LRU order of inactive objects upon access by CPU

[Chris Wilson]

In order to reduce the penalty of fallbacks under memory pressure and to
avoid a potential immediate ping-pong of evicting a mmaped buffer, we
move the object to the tail of the inactive list when a page is freshly
faulted or the object is moved into the CPU domain.

We choose not to protect the CPU objects from casual eviction,
preferring to keep the GPU active for as long as possible.

Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/i915_gem.c |    7 +++++++
 1 files changed, 7 insertions(+), 0 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index bb7d6a9..26d0345 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1036,6 +1036,10 @@ i915_gem_set_domain_ioctl(struct drm_device *dev, void *data,
 		ret = i915_gem_object_set_to_cpu_domain(obj, write_domain != 0);
 	}
 
+	/* Maintain LRU order of "inactive" objects */
+	if (ret == 0 && obj_priv->gtt_space && !obj_priv->active)
+		list_move_tail(&obj_priv->list, &dev_priv->mm.inactive_list);
+
 	drm_gem_object_unreference(obj);
 	mutex_unlock(&dev->struct_mutex);
 	return ret;
@@ -1169,6 +1173,9 @@ int i915_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 			goto unlock;
 	}
 
+	if (!obj_priv->active)
+		list_move_tail(&obj_priv->list, &dev_priv->mm.inactive_list);
+
 	pfn = ((dev->agp->base + obj_priv->gtt_offset) >> PAGE_SHIFT) +
 		page_offset;
 
-- 
1.7.1