[Mesa-dev] [PATCH 10/10] intel: Request DRI2 buffers for separate stencil and hiz

[Chad Versace]

On Tue, 07 Jun 2011 08:15:51 -0700, Kenneth Graunke <kenneth at whitecape.org> wrote:
> On 06/07/2011 01:33 AM, Julien Cristau wrote:
> > On Mon, Jun  6, 2011 at 18:56:20 -0700, Kenneth Graunke wrote:
> >
> >> I'd make this an assertion.  must_use_separate_stencil is only set for
> >> gen>= 7, and I assert that try_separate_stencil ought to be true in
> >> that case (old DDX doesn't support IVB, new DDX requires new dri2proto
> >> to compile).
> >>
> >> Only unreleased git-snapshots of the DDX should hit this (and only on
> >> unreleased hardware), so hitting an assertion seems like a friendly way
> >> to say "hey, update your DDX".
> >>
> > There's nothing friendly about hitting an assertion, so if there's
> > another option then please avoid assert...
> >
> > Cheers,
> > Julien
> 
> Well...my point was that I think only Jesse, Eric, and I could possibly 
> hit it, and hitting an assert is a lot easier to debug than random 
> behavior.  In that sense, it's more helpful.  I'd agree with you if it 
> were possible to trip it on released hardware.
> 
> --Kenneth

Agreeing with Ken here. Hitting an assertion is easier than debugging
random behaviour that should never occur.

Assertions are used in cases in which the supposedly impossible has
happened despite our understanding of the situation. Once Gen7 hardware
is released, the if's condition should be impossible to trigger.