difference is that after calling finish_gpu() the object no longer
resides in any GPU domain, and so will cause the GPU caches to be
invalidated if it is ever used again.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
drivers/gpu/drm/i915/i915_drv.h | 2 +-
drivers/gpu/drm/i915/i915_gem.c | 29 +++++++++++++++++++----------
drivers/gpu/drm/i915/intel_display.c | 2 +-
3 files changed, 21 insertions(+), 12 deletions(-)
@@ -1163,7 +1163,7 @@ void i915_gem_clflush_object(struct drm_i915_gem_object *obj);
int __must_check i915_gem_object_set_domain(struct drm_i915_gem_object *obj,
uint32_t read_domains,
uint32_t write_domain);
-int __must_check i915_gem_object_flush_gpu(struct drm_i915_gem_object *obj);
+int __must_check i915_gem_object_finish_gpu(struct drm_i915_gem_object *obj);
int __must_check i915_gem_init_ringbuffer(struct drm_device *dev);
void i915_gem_cleanup_ringbuffer(struct drm_device *dev);
void i915_gem_do_init(struct drm_device *dev,
@@ -2167,23 +2167,29 @@ i915_gem_object_unbind(struct drm_i915_gem_object *obj)
return -EINVAL;
}
+ ret = i915_gem_object_finish_gpu(obj);
+ if (ret == -ERESTARTSYS)
+ return ret;
+ /* Continue on if we fail due to EIO, the GPU is hung so we
+ * should be safe and we need to cleanup or else we might
+ * cause memory corruption through use-after-free.
+ */
+
/* blow away mappings if mapped through GTT */
i915_gem_release_mmap(obj);
/* Move the object to the CPU domain to ensure that
* any possible CPU writes while it's not in the GTT
- * are flushed when we go to remap it. This will
- * also ensure that all pending GPU writes are finished
- * before we unbind.
+ * are flushed when we go to remap it.
*/
- ret = i915_gem_object_set_to_cpu_domain(obj, 1);
+ if (ret == 0)
+ ret = i915_gem_object_set_to_cpu_domain(obj, 1);
if (ret == -ERESTARTSYS)
return ret;
- /* Continue on if we fail due to EIO, the GPU is hung so we
- * should be safe and we need to cleanup or else we might
- * cause memory corruption through use-after-free.
- */
if (ret) {
+ /* In the event of a disaster, abandon all caches and
+ * hope for the best.
+ */
i915_gem_clflush_object(obj);
obj->base.read_domains = obj->base.write_domain = I915_GEM_DOMAIN_CPU;
}
@@ -3045,11 +3051,11 @@ i915_gem_object_set_to_display_plane(struct drm_i915_gem_object *obj,
}
int
-i915_gem_object_flush_gpu(struct drm_i915_gem_object *obj)
+i915_gem_object_finish_gpu(struct drm_i915_gem_object *obj)
{
int ret;
- if (!obj->active)
+ if ((obj->base.read_domains & I915_GEM_GPU_DOMAINS) == 0)
return 0;
if (obj->base.write_domain & I915_GEM_GPU_DOMAINS) {
@@ -3058,6 +3064,9 @@ i915_gem_object_flush_gpu(struct drm_i915_gem_object *obj)
return ret;
}
+ /* Ensure that we invalidate the GPU's caches and TLBs. */
+ obj->base.read_domains &= I915_GEM_GPU_DOMAINS;
+
return i915_gem_object_wait_rendering(obj);
}
@@ -1966,7 +1966,7 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
* This should only fail upon a hung GPU, in which case we
* can safely continue.
*/
- ret = i915_gem_object_flush_gpu(obj);
+ ret = i915_gem_object_finish_gpu(obj);
(void) ret;
}