[Bug 95472] [i915] Feature request: Add support for fencing for PRIME setups

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=95472

--- Comment #44 from Axel Davy <axel.davy at ens.fr> ---
(In reply to Chris Wilson from comment #43)
> There's no synchronisation point in dri3 with X:
> 
> diff --git a/src/loader/loader_dri3_helper.c
> b/src/loader/loader_dri3_helper.c
> index a55719a..6b5736b 100644
> --- a/src/loader/loader_dri3_helper.c
> +++ b/src/loader/loader_dri3_helper.c
> @@ -1399,6 +1399,7 @@ loader_dri3_get_buffers(__DRIdrawable *driDrawable,
>        buffers->back = back->image;
>     }
>  
> +   loader_dri3_wait_x(draw);
>     draw->stamp = stamp;
>  
>     return true;
> 
> brings it into line with dri2.

I don't see why waitX is required.

The call has only an impact when a fake front buffer is used.
And when fake front buffer is used, the same buffer is always used and it seems
correctly updated when needed.

(In reply to Mike Lothian from comment #42)
Are you sure the dedicated card was not used for this test ?

It looks like the problem is the compositor copies the damaged windows contents
before the blitImage calls are executed on the gpu. Could it be a problem with
the flushing behaviour in intel_blit_image ?

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the bug.
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20160816/9d388573/attachment.html>