[igt-dev] [PATCH i-g-t 1/2] lib/media_fill: Use RENDER engine flag to work on Xe

[Zbigniew Kempczyński]

On Thu, Aug 03, 2023 at 11:11:19PM +0200, Manszewski, Christoph wrote:
> Hi Zbigniew,
> 
> On 3.08.2023 09:49, Zbigniew Kempczyński wrote:
> > Due to explicit requirement of running on valid hw engine on xe
> > change I915_EXEC_DEFAULT to I915_EXEC_RENDER. This change doesn't
> > break i915 media pipeline allowing to create and assign valid xe
> > engine to execute media fill test.
> > 
> > Signed-off-by: Zbigniew Kempczyński <zbigniew.kempczynski at intel.com>
> > Cc: Christoph Manszewski <christoph.manszewski at intel.com>
> > ---
> >   lib/media_fill.c | 2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/lib/media_fill.c b/lib/media_fill.c
> > index e80dae1505..2520995d97 100644
> > --- a/lib/media_fill.c
> > +++ b/lib/media_fill.c
> > @@ -281,7 +281,7 @@ __gen9_media_fillfunc(int i915,
> >   	intel_bb_ptr_align(ibb, 32);
> >   	intel_bb_exec(ibb, intel_bb_offset(ibb),
> > -		      I915_EXEC_DEFAULT | I915_EXEC_NO_RELOC, true);
> > +		      I915_EXEC_RENDER | I915_EXEC_NO_RELOC, true);
> 
> As discussed offline - since we are the ones doing the i915 to xe engine
> type conversion, we could as well translate 'I915_EXEC_DEFAULT' to
> 'DRM_XE_ENGINE_CLASS_RENDER'. But I think it's a minor thing, besides I
> already reviewed a similar change in gpgpu_fill so:
> 
> Reviewed-by: Christoph Manszewski <christoph.manszewski at intel.com>
> 
> Christoph

Thank you for the review. If I'm not wrong blitter didn't have dedicated
ring for gen < 6 (see HAS_BLT_RING() macro in igt code for reference).

--
Zbigniew

> 
> 
> >   	intel_bb_destroy(ibb);
> >   }