On 08/23/2012 05:00 PM, Paul Berry wrote: > This test provokes a bug in Mesa as of 8/23/12 (commit 1cb07bd): when > fast depth clears are performed on multiple layers of a multilayer > depth texture (such as a 2D array or a cubemap texture), the > implementation doesn't properly track the need for depth resolves, > resulting in incorrect data when the texture is later used for > texturing. See https://bugs.freedesktop.org/show_bug.cgi?id=50270. > > Although the bug was initially discovered using cubemap textures, this > test provokes it using a 2D array texture, so that we don't have to > reason about the coordinate transformations involved in cube map > texture lookup. > --- > tests/all.tests | 1 + > tests/texturing/CMakeLists.gl.txt | 1 + > tests/texturing/array-depth-roundtrip.c | 203 > +++++++++++++++++++++++++++++++ > 3 files changed, 205 insertions(+), 0 deletions(-) > create mode 100644 tests/texturing/array-depth-roundtrip.c
Reviewed-by: Chad Versace <chad.vers...@linux.intel.com> _______________________________________________ Piglit mailing list Piglit@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/piglit