[Intel-gfx] [PATCH 3/3] drm/i915: Don't waste fences for scanout when not needed
ville.syrjala at linux.intel.com
ville.syrjala at linux.intel.com
Tue Jan 28 13:51:56 CET 2014
From: Ville Syrjälä <ville.syrjala at linux.intel.com>
Gen2 and gen3 require fences for tiled scanout, but gen4+ only require a
fence for FBC GTT modification tracking. Install a fence for scanout
only when it might actually be required.
This could potentially be optimized even further, but this seems like
a fair comporomise for now.
Signed-off-by: Ville Syrjälä <ville.syrjala at linux.intel.com>
---
drivers/gpu/drm/i915/intel_display.c | 19 ++++++++++++++-----
1 file changed, 14 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index fd48c2e..71dd041 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -1972,6 +1972,13 @@ int intel_pin_fenced_fb_obj(struct drm_device *dev,
u32 alignment;
int ret;
+ /*
+ * Pre-i965 always needs a fence for tiled scanout, whereas
+ * 965+ only requires a fence if using framebuffer compression.
+ */
+ if (INTEL_INFO(dev)->gen >= 4 && !HAS_FBC(dev))
+ return intel_pin_fb_obj(dev, obj, pipelined);
+
ret = intel_fb_obj_alignment(dev, obj, &alignment);
if (ret)
return ret;
@@ -1981,11 +1988,6 @@ int intel_pin_fenced_fb_obj(struct drm_device *dev,
if (ret)
goto err_interruptible;
- /* Install a fence for tiled scan-out. Pre-i965 always needs a
- * fence, whereas 965+ only requires a fence if using
- * framebuffer compression. For simplicity, we always install
- * a fence as the cost is not that onerous.
- */
ret = i915_gem_object_get_fence(obj);
if (ret)
goto err_unpin;
@@ -2004,6 +2006,13 @@ err_interruptible:
void intel_unpin_fenced_fb_obj(struct drm_i915_gem_object *obj)
{
+ struct drm_device *dev = obj->base.dev;
+
+ if (INTEL_INFO(dev)->gen >= 4 && !HAS_FBC(dev)) {
+ intel_unpin_fb_obj(obj);
+ return;
+ }
+
i915_gem_object_unpin_fence(obj);
i915_gem_object_unpin_from_display_plane(obj);
}
--
1.8.3.2
More information about the Intel-gfx
mailing list