I've been working on a diagram to try and lay out how Gallium3D will pass data to the Graphics Card.
https://docs.google.com/drawings/d/1KG4wuspMHZWl0398J_m2BXCWxkNGYxcLZwxE9Xk3LV0/editRight now i'm thinking one of the following are going to be our options (from most likely to least likely)
BGLRenderer -> DRI2 translation "thing" -> Accelerant (ring buffer) BGLRenderer -> Accelerant (ring buffer) BGLRenderer -> Driver -> Accelerant (ring buffer) (blah, i hate this)The ring buffer has to be in the Accelerant due to it needing AtomBIOS (which the accelerant provides)
Any thoughts welcome. Please keep in mind Hardware rendering is post-R1 as mainline Mesa/Gallium
will *not* compile under gcc2 without us forking it :) Comments / Thoughts / Flames *very* welcome. Thanks! -- Alex