[Intel-gfx] [PATCH 2/3] drm/i915: add syncobj timeline support

[Daniel Vetter]

On Fri, Jul 31, 2020 at 9:07 PM Lionel Landwerlin
<lionel.g.landwerlin at intel.com> wrote:
>
> On 31/07/2020 17:30, Chris Wilson wrote:
> > Quoting Lionel Landwerlin (2020-07-31 14:45:52)
> >> -               drm_syncobj_replace_fence(syncobj, fence);
> >> +               if (eb->fences[n].chain_fence) {
> >> +                       drm_syncobj_add_point(syncobj, eb->fences[n].chain_fence,
> >> +                                             fence, eb->fences[n].value);
> > This function remains not acceptable. It is trivial to write an igt test
> > that causes the DRM_ERROR. A user controllable error message is still
> > not allowed. If you do not have such a test in your igt series, then that
> > too is lacking.
> > -Chris
>
> As far as I remember there are IGT tests for this (*unordered* subtests).
>
> So CI should report a fairlure. The IGT test itself won't fail though.

CI did catch it.

> I thought we removed that DRM_ERROR a while ago.
>
> I'll send a patch to remove it. Thanks.

Typed it already (since I didn't see yours), as soon as it's pushed
you can just hit the retest button with your current series and then
we see whether it's all gone.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch