Kenneth Graunke <kenn...@whitecape.org> writes:
> On 02/14/2014 03:00 PM, Eric Anholt wrote:
> [snip]
>> @@ -666,78 +667,6 @@ intel_miptree_create_for_bo(struct brw_context *brw,
>> return mt;
>> }
>>
>> -
>> -/**
>> - * For a singlesample DRI2 buffer, this simply wraps the given region with
>> a miptree.
>> - *
>> - * For a multisample DRI2 buffer, this wraps the given region with
>> - * a singlesample miptree, then creates a multisample miptree into which the
>> - * singlesample miptree is embedded as a child.
>> - */
>> -struct intel_mipmap_tree*
>> -intel_miptree_create_for_dri2_buffer(struct brw_context *brw,
>> - unsigned dri_attachment,
>> - mesa_format format,
>> - uint32_t num_samples,
>> - struct intel_region *region)
>> -{
>> - struct intel_mipmap_tree *singlesample_mt = NULL;
>> - struct intel_mipmap_tree *multisample_mt = NULL;
>> -
>> - /* Only the front and back buffers, which are color buffers, are shared
>> - * through DRI2.
>> - */
>> - assert(dri_attachment == __DRI_BUFFER_BACK_LEFT ||
>> - dri_attachment == __DRI_BUFFER_FRONT_LEFT ||
>> - dri_attachment == __DRI_BUFFER_FAKE_FRONT_LEFT);
>> - assert(_mesa_get_format_base_format(format) == GL_RGB ||
>> - _mesa_get_format_base_format(format) == GL_RGBA);
>> -
>> - singlesample_mt = intel_miptree_create_for_bo(brw,
>> - region->bo,
>> - format,
>> - 0,
>> - region->width,
>> - region->height,
>> - region->pitch,
>> - region->tiling);
>> - if (!singlesample_mt)
>> - return NULL;
>> - singlesample_mt->region->name = region->name;
>
> The singlesample_mt->region->name = region->name line is missing from
> the new code. I doubt it actually matters, but figured I'd point it out
> in case it wasn't intentional.It wasn't intentional, and it would have broken the optimization in intel_process_dri2_buffer(). Thanks!
pgpo4GpQjlGhn.pgp
Description: PGP signature
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
