[Mesa-dev] Last UBO patchbomb!

[Kenneth Graunke]

On 07/31/2012 03:01 PM, Eric Anholt wrote:
> I've tested it with my piglit series and with Intel's oglconform.  My piglit
> series complains about negative-bindbuffer-buffer's expectations being
> violated by patch 6, but I think that testcase wants to be replced by one that
> is aware of gl >=3.1 versus <=3.0.  Intel's oglconform complains only
> about one testcase that's transform feedbacking into a UBO and failing, so
> probably a little driver bug I hope.
> 
> Other than that, I've got a few testing TODOs, but I think this is ready to
> land and unblock GL 3.1.

Except for patch 5, this series gets a:
Reviewed-by: Kenneth Graunke <kenneth at whitecape.org>

I would like to review v2 of patch 5.

Nice work!  I do like the UBO load lowering pass.  It encapsulates all
the ugly offset computation in one place and makes the backend code so
much simpler.