[Intel-gfx] [PATCH 04/23] drm/i915: Push the ring creation flags to the backend

[Chris Wilson]

Quoting Tvrtko Ursulin (2019-08-01 09:42:15)
> 
> On 30/07/2019 10:38, Chris Wilson wrote:
> > Quoting Tvrtko Ursulin (2019-07-29 13:59:22)
> >>
> >> On 26/07/2019 09:43, Chris Wilson wrote:
> >>> Quoting Tvrtko Ursulin (2019-07-24 12:11:54)
> >>>>
> >>>> On 23/07/2019 19:38, Chris Wilson wrote:
> >>>>> Push the ring creation flags from the outer GEM context to the inner
> >>>>> intel_cotnext to avoid an unsightly back-reference from inside the
> >>>>> backend.
> >>>>
> >>>> Sorry I find this quite ugly. Passing in integers in pointers filed and
> >>>> casting back and forth.
> >>>
> >>> But better than a union, since once the intel_context is active, the
> >>> ring is always a ring.
> >>
> >> Unless it is u64 size. I am not buying it. :)
> > 
> > We don't need u64 size? I don't understand what you mean.
> 
> I complained about very unobvious and questionable hack of passing the 
> size in the pointer field and you said it is better than an union. For 
> me union certainly rates way higher than the casing hack with a macro.

We store values inside pointer slots all the time, sometimes even memory
addresses.
-Chris