System Environment: -------------------------- Platform: Broadwell Libdrm: (master)libdrm-2.4.53-6-gd4083dc762a2974c35ecd16be995272bbb6d57b4 Mesa: (master)34a68345e228b7cbc1bc6909704a89c08bf5368e Xserver: (master)xorg-server-1.15.99.902-47-ge924034269532979f1e88947ee9c33d30461edae Xf86_video_intel:(master)2.99.911-94-gbb5b21e34f6cbe44a249e566471754b45c500e60 Libva: (staging)8577efbd3b1d4e8f113f78cf13d2182aa681b738 Libva_intel_driver:(staging)42258e128f19b93aa102672d5f61eb73d9f9808f Kernel: (drm-intel-nightly)1e771b84e47085ef9b6efea1321e7cb5a8b2c065 Bug detailed description: ----------------------------- It fails on Broadwell with Mesa master branch, but works well on Haswell. spec_ARB_framebuffer_object_fbo-drawbuffers-none_glClearBuffer also fails on Broadwell. output: libGL: screen 0 does not appear to be DRI3 capable libGL: pci id for fd 4: 8086:1616, driver i965 libGL: OpenDriver: trying /opt/X11R7/lib/dri/i965_dri.so Testing glBlitFramebuffer. Probe color at (0,0) Expected: 1.000000 0.000000 0.000000 1.000000 Observed: 0.200000 0.200000 0.200000 0.200000 from color attachment 0, config (NONE, COLOR_ATTACHMENT1, COLOR_ATTACHMENT0, COLOR_ATTACHMENT2) Probe color at (0,0) Expected: 1.000000 0.000000 0.000000 1.000000 Observed: 0.200000 0.200000 0.200000 0.200000 from color attachment 1 [details] [review], config (NONE, COLOR_ATTACHMENT1, COLOR_ATTACHMENT0, COLOR_ATTACHMENT2) Probe color at (0,0) Expected: 1.000000 0.000000 0.000000 1.000000 Observed: 0.200000 0.200000 0.200000 0.200000 from color attachment 2 [details] [review], config (NONE, COLOR_ATTACHMENT1, COLOR_ATTACHMENT0, COLOR_ATTACHMENT2) Probe color at (0,0) Expected: 1.000000 0.000000 0.000000 1.000000 Observed: 0.200000 0.200000 0.200000 0.200000 from color attachment 2 [details] [review], config (NONE, NONE, COLOR_ATTACHMENT3, COLOR_ATTACHMENT2) Probe color at (0,0) Expected: 1.000000 0.000000 0.000000 1.000000 Observed: 0.200000 0.200000 0.200000 0.200000 from color attachment 3 [details] [review], config (NONE, NONE, COLOR_ATTACHMENT3, COLOR_ATTACHMENT2) Probe color at (0,0) Expected: 1.000000 0.000000 0.000000 1.000000 Observed: 0.200000 0.200000 0.200000 0.200000 from color attachment 0, config (NONE, COLOR_ATTACHMENT0, NONE, COLOR_ATTACHMENT1) Probe color at (0,0) Expected: 1.000000 0.000000 0.000000 1.000000 Observed: 0.200000 0.200000 0.200000 0.200000 from color attachment 1 [details] [review], config (NONE, COLOR_ATTACHMENT0, NONE, COLOR_ATTACHMENT1) Probe color at (0,0) Expected: 1.000000 0.000000 0.000000 1.000000 Observed: 0.200000 0.200000 0.200000 0.200000 from color attachment 0, config (NONE, COLOR_ATTACHMENT2, COLOR_ATTACHMENT0, NONE) Probe color at (0,0) Expected: 1.000000 0.000000 0.000000 1.000000 Observed: 0.200000 0.200000 0.200000 0.200000 from color attachment 2 [details] [review], config (NONE, COLOR_ATTACHMENT2, COLOR_ATTACHMENT0, NONE) Probe color at (0,0) Expected: 1.000000 0.000000 0.000000 1.000000 Observed: 0.200000 0.200000 0.200000 0.200000 from color attachment 1 [details] [review], config (NONE, NONE, NONE, COLOR_ATTACHMENT1) Probe color at (0,0) Expected: 1.000000 0.000000 0.000000 1.000000 Observed: 0.200000 0.200000 0.200000 0.200000 from color attachment 2 [details] [review], config (NONE, NONE, COLOR_ATTACHMENT2, NONE) Probe color at (0,0) Expected: 1.000000 0.000000 0.000000 1.000000 Observed: 0.200000 0.200000 0.200000 0.200000 from color attachment 0, config (NONE, COLOR_ATTACHMENT0, NONE, NONE) PIGLIT: {'result': 'fail' } Reproduce steps: ---------------------------- 1. xinit 2. bin/fbo-drawbuffers-none glBlitFramebuffer -fbo -auto
The "fbo-drawbuffers-none glClearBuffer" failure is because the _mesa_meta_glsl_Clear path doesn't pay attention to the set of buffers it was asked to clear. (This is basically #76832 - Iago fixed the problem for the blorp path, but not the meta path.) The "clearfix" branch of my tree has some proposed patches to fix this, but they regress basically every glean test currently. Eric thinks that's due to an unrelated bug my code happened to uncover. More work required...
Seems to be passing with latest Mesa master.
Verified.Fixed.
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.