After moving the write flushing into the move_to_ring, we can simplify
the execbuffer flush by not having to compute inter-ring flushes.
Signed-off-by: Chris Wilson <ch...@chris-wilson.co.uk>
---
drivers/gpu/drm/i915/i915_gem_execbuffer.c | 73 +++++++---------------------
1 files changed, 17 insertions(+), 56 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 3c54911..d04bd3d 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -36,7 +36,6 @@
struct change_domains {
uint32_t invalidate_domains;
uint32_t flush_domains;
- uint32_t flush_rings;
uint32_t flips;
};
@@ -156,7 +155,7 @@ i915_gem_object_set_to_gpu_domain(struct
drm_i915_gem_object *obj,
struct intel_ring_buffer *ring,
struct change_domains *cd)
{
- uint32_t invalidate_domains = 0, flush_domains = 0;
+ uint32_t flush = 0;
/*
* If the object isn't moving to a new write domain,
@@ -172,22 +171,17 @@ i915_gem_object_set_to_gpu_domain(struct
drm_i915_gem_object *obj,
* write domain
*/
if (obj->base.write_domain &&
- (obj->base.write_domain != obj->base.pending_read_domains ||
- obj->ring != ring)) {
- flush_domains |= obj->base.write_domain;
- invalidate_domains |=
- obj->base.pending_read_domains &
~obj->base.write_domain;
- }
+ obj->base.write_domain != obj->base.pending_read_domains)
+ flush = obj->base.write_domain;
/*
* Invalidate any read caches which may have
* stale data. That is, any new read domains.
*/
- invalidate_domains |= obj->base.pending_read_domains &
~obj->base.read_domains;
- if ((flush_domains | invalidate_domains) & I915_GEM_DOMAIN_CPU)
+ if (flush & I915_GEM_DOMAIN_CPU)
i915_gem_clflush_object(obj);
/* blow away mappings if mapped through GTT */
- if ((flush_domains | invalidate_domains) & I915_GEM_DOMAIN_GTT)
+ if (flush & I915_GEM_DOMAIN_GTT)
i915_gem_release_mmap(obj);
if (obj->base.pending_write_domain)
@@ -199,15 +193,12 @@ i915_gem_object_set_to_gpu_domain(struct
drm_i915_gem_object *obj,
* write_domains). So if we have a current write domain that we
* aren't changing, set pending_write_domain to that.
*/
- if (flush_domains == 0 && obj->base.pending_write_domain == 0)
+ if (flush == 0 && obj->base.pending_write_domain == 0)
obj->base.pending_write_domain = obj->base.write_domain;
- cd->invalidate_domains |= invalidate_domains;
- cd->flush_domains |= flush_domains;
- if (flush_domains & I915_GEM_GPU_DOMAINS)
- cd->flush_rings |= obj->ring->id;
- if (invalidate_domains & I915_GEM_GPU_DOMAINS)
- cd->flush_rings |= ring->id;
+ cd->flush_domains |= flush;
+ cd->invalidate_domains |=
+ obj->base.pending_read_domains & ~obj->base.read_domains;
}
struct eb_objects {
@@ -710,35 +701,6 @@ err:
}
static int
-i915_gem_execbuffer_flush(struct drm_device *dev,
- uint32_t invalidate_domains,
- uint32_t flush_domains,
- uint32_t flush_rings)
-{
- drm_i915_private_t *dev_priv = dev->dev_private;
- int i, ret;
-
- if (flush_domains & I915_GEM_DOMAIN_CPU)
- intel_gtt_chipset_flush();
-
- if (flush_domains & I915_GEM_DOMAIN_GTT)
- wmb();
-
- if ((flush_domains | invalidate_domains) & I915_GEM_GPU_DOMAINS) {
- for (i = 0; i < I915_NUM_RINGS; i++)
- if (flush_rings & (1 << i)) {
- ret = i915_gem_flush_ring(&dev_priv->ring[i],
- invalidate_domains,
- flush_domains);
- if (ret)
- return ret;
- }
- }
-
- return 0;
-}
-
-static int
i915_gem_execbuffer_wait_for_flips(struct intel_ring_buffer *ring, u32 flips)
{
u32 plane, flip_mask;
@@ -783,14 +745,11 @@ i915_gem_execbuffer_move_to_gpu(struct intel_ring_buffer
*ring,
list_for_each_entry(obj, objects, exec_list)
i915_gem_object_set_to_gpu_domain(obj, ring, &cd);
- if (cd.invalidate_domains | cd.flush_domains) {
- ret = i915_gem_execbuffer_flush(ring->dev,
- cd.invalidate_domains,
- cd.flush_domains,
- cd.flush_rings);
- if (ret)
- return ret;
- }
+ if (cd.flush_domains & I915_GEM_DOMAIN_CPU)
+ intel_gtt_chipset_flush();
+
+ if (cd.flush_domains & I915_GEM_DOMAIN_GTT)
+ wmb();
if (cd.flips) {
ret = i915_gem_execbuffer_wait_for_flips(ring, cd.flips);
@@ -804,7 +763,9 @@ i915_gem_execbuffer_move_to_gpu(struct intel_ring_buffer
*ring,
return ret;
}
- return 0;
+ return i915_gem_flush_ring(ring,
+ cd.invalidate_domains,
+ cd.flush_domains);
}
static bool
--
1.7.4.1
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
