The test fails for me with the setup below: test: dEQP-EGL.functional.image.modify.tex_rgb5_a1_tex_subimage_rgba8 mesa: master@7a4ea95 deqp: nougat-cts-dev@4acf725 and master@7363696 deqp-target: x11_egl hw: Broadwell 0x1616
The test failed because in glTexImage2D(format=GL_RGBA, type=GL_UNSIGNED_SHORT_5_5_5_1) Mesa chose BRW_SURFACE_FORMAT_R8G8B8A8_UNORM as the storage format. The two formats have incompatible precision in the alpha channel. One provides binary alpha; the other provides gradient alpha. Intel hardware supports BRW_SURFACE_FORMAT_B5G5R5A1_UNORM, so we need to teach glTexImage2D to choose that instead.
Fix submitted. https://lists.freedesktop.org/archives/mesa-dev/2016-December/138921.html https://patchwork.freedesktop.org/series/17153/
Patches pushed.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.