On Mon, Nov 28, 2011 at 05:55:58PM -0800, Eric Anholt wrote: > On Mon, 28 Nov 2011 22:24:54 +0100, Daniel Vetter <daniel.vetter at ffwll.ch> wrote: > > And track the existence of such a binding similar to the aliasing > > ppgtt case. Speeds up binding/unbinding in the common case where we > > only need a ppgtt binding (which is accessed in a cpu coherent fashion > > by the gpu) and no gloabl gtt binding (which needs uc writes for the > > ptes). > > For one month from mid-june to mid-july, Mesa master was setting > I915_GEM_DOMAIN_GTT in the workaround pipe control. That was changed to > be _INSTRUCTION, because _GTT caused kernel oopses, and query objects > had already been doing _INSTRUCTION since forever. There are no uses of > MI_STOREs in the driver. > > I'd say the minor ABI change potential in here is safe. Can I interpret this as an Acked-by for the (fixed up and actually tested) patch? -Daniel -- Daniel Vetter Mail: daniel at ffwll.ch Mobile: +41 (0)79 365 57 48
