[Intel-gfx] [PATCH 08/12] drm/i915: Support for pread/pwrite from/to non shmem backed objects

[Chris Wilson]

On Wed, Apr 20, 2016 at 04:47:35PM +0530, ankitprasad.r.sharma at intel.com wrote:
>  static int
>  i915_gem_shmem_pread(struct drm_device *dev,
>  		     struct drm_i915_gem_object *obj,
> @@ -656,6 +795,9 @@ i915_gem_shmem_pread(struct drm_device *dev,
>  	int needs_clflush = 0;
>  	struct sg_page_iter sg_iter;
>  
> +	if (!obj->base.filp)
> +		return -ENODEV;

This is now:

if ((obj->ops->flags & I915_GEM_OBJECT_HAS_STRUCT_PAGE) == 0)
	return -ENODEV;

> @@ -866,22 +1009,34 @@ i915_gem_gtt_pwrite_fast(struct drm_i915_private *i915,
>  		unsigned page_length = PAGE_SIZE - page_offset;
>  		page_length = remain < page_length ? remain : page_length;
>  		if (node.allocated) {
> -			wmb();
> +			wmb(); /* flush the write before we modify the GGTT */
>  			ggtt->base.insert_page(&ggtt->base,
>  					       i915_gem_object_get_dma_address(obj, offset >> PAGE_SHIFT),
>  					       node.start, I915_CACHE_NONE, 0);
> -			wmb();
> +			wmb(); /* flush modifications to the GGTT (insert_page) */

We might as delete this wmb(); then we don't even need to document it ;)
Hmm. On reflection, we can punt wmb() (from insert_page()) to the caller
as well as it is much clearer, so go back to patch 1 and remove both the
wmb() and chipset flushes.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre