On 06/01/2014 11:29 PM, Chris Forbes wrote: > This series adds support for ARB_compressed_texture_pixel_storage from GL4.2. > This > provides a convenient way to update a subrectangle of a compressed texture > from client memory or a pixel buffer object, without having to rearrange the > client copy. > > (Well, as convenient as GL pixel store state ever is to work with...) > > All the work is in the core mesa fallbacks (which are called eventually, even > in > the cases where a custom driver func is present for these cases), so this is > enabled > on all drivers. > > Patch 1 does some preliminary cleanup of PixelStorei. Once upon a time, this > function > had flushes etc in it, and so tried to avoid spurious updates. The flushes > were > removed, but the update-avoidance machinery (two copies of it in some cases) > was all left in place. Just rip it out, merge some error paths, etc. > > Patch 2 tidies up some inconsistent error reporting in CompressedTex*Image. > > Patches 3-6 add extension plumbing and new pack/unpack state. > > Patches 7-10 make the new state actually work. > > Error generation for imageSize isn't quite right, when the user-provided > block dimensions disagree with the actual format's block dimensions, but this > is complete enough to be useful, so I figured I'd send it out with an RFC tag.
Which patch is that? I can go back and look closer. > No regressions in quick.py on Ivybridge. > > > _______________________________________________ > mesa-dev mailing list > mesa-dev@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/mesa-dev _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
