Created attachment 119052 [details] Dmesg logs The following test cases are failing: deqp-gles2@functional@texture@vertex@cube@wrap@clamp_clamp deqp-gles2@functional@texture@vertex@cube@wrap@clamp_mirror deqp-gles2@functional@texture@vertex@cube@wrap@clamp_repeat deqp-gles2@functional@texture@vertex@cube@wrap@mirror_clamp deqp-gles2@functional@texture@vertex@cube@wrap@mirror_mirror deqp-gles2@functional@texture@vertex@cube@wrap@mirror_repeat Configuration: Broadwell-U Hardware Platform: NUC5i7RYB Processor: Intel(R) Core(TM) i7-5557U CPU @ 3.10GHz GPU: Broadwell-U Integrated Graphics Software Linux distribution: Ubuntu 14.04 LTS 64 bits kernel 4.3.0-rc5-drm-intel-nightly+ from git://anongit.freedesktop.org/drm-intel Mesa - 11.0-branchpoint-1379-g6f39556 http://cgit.freedesktop.org/mesa/mesa/ xf86-video-intel - 2.99.917 from http://cgit.freedesktop.org/xorg/driver/xf86-video-intel/ Libdrm - 2.4.65 from http://cgit.freedesktop.org/mesa/drm/ Libva - 1.6.1 from http://cgit.freedesktop.org/libva/ vaapi intel-driver - 1.6.1 from http://cgit.freedesktop.org/vaapi/intel-driver Cairo - 1.14.2 from http://cgit.freedesktop.org/cairo Xorg Xserver - 1.17.2 from http://cgit.freedesktop.org/xorg/xserver Spected result: Test cases should pass Actual result: Test cases are failing, Attahing terminal output and dmesg "drm.debug=0x06"
Created attachment 119053 [details] Terminal Output
This same tests also fails on HSW-U with the following configuration : ==Test Environment== -------------------------------------------------- xorg-server-1.17.2 libdrm-2.4.65 xf86-video-intel-2.99.917 mesa-11.1.0-devel (git 6f39546) libva-1.6.1 intel-driver-1.6.1 cairo-1.14.2 kernel: 4.3.0-rc5-drm-intel-nightly+ commit ce4630de6e6ba63c0477e753510f6294fac9d30c Author: Matt Roper <matthew.d.roper@intel.com> Date: Wed Oct 21 15:05:45 2015 -0700
This issue on same tests was reproduced on BYT with the following configuration: kernel: 4.3.0-rc5-drm-intel-nightly xorg-server-1.17.2 libdrm-2.4.65 xf86-video-intel-2.99.917 mesa-11.1.0-devel (git 6f39546) libva-1.6.1 intel-driver-1.6.1 cairo-1.14.2
The following tests keeps fail on BDW-U with the next configuration : Case list : deqp-gles2@functional@texture@vertex@cube@wrap@clamp_clamp deqp-gles2@functional@texture@vertex@cube@wrap@clamp_mirror deqp-gles2@functional@texture@vertex@cube@wrap@clamp_repeat deqp-gles2@functional@texture@vertex@cube@wrap@mirror_clamp deqp-gles2@functional@texture@vertex@cube@wrap@mirror_mirror deqp-gles2@functional@texture@vertex@cube@wrap@mirror_repeat kernel drm-intel-testing: commit 91587c722c28c4116dedbfbf08aa874377bc76f8 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Fri Dec 4 17:35:54 2015 +0100 drm-intel-nightly: 2015y-12m-04d-16h-35m-07s UTC integration manifest kernel version : 4.4.0-rc3 git url : git://anongit.freedesktop.org/drm-intel git branch : drm-intel-testing git describe : drm-intel-next-2015-11-20-rebased-13721-g91587c7 kernel drm-intel-nightly commit 79686f613b3955a4ed09cee936e7f70ec4e61b67 Author: Jani Nikula <jani.nikula@intel.com> Date: Wed Dec 30 14:00:24 2015 +0200 drm-intel-nightly: 2015y-12m-30d-11h-59m-54s UTC integration manifest kernel version : 4.4.0-rc6 git url : git://anongit.freedesktop.org/drm-intel git branch : drm-intel-nightly git describe : drm-intel-next-2015-12-18-1500-g79686f6 cairo version: 1.15.2 / commit : db8a7f1 drm version : libdrm-2.4.66 / commit : b38a4b2 intel-driver : 1.6.2 / commit: 683edee libva version : libva-1.6.2 / commit : 304bc13 mesa version : mesa-11.0.8 / commit : 261daab xf86-video-intel version : 2.99.917 / commit : baec802 xserver version :xorg-server-1.18.0 / commit :7921764
If I run this on git master of mesa (1106e79ed97) then all 9 of the tests fail. I think this has something to do with this patch here: http://cgit.freedesktop.org/mesa/mesa/commit/?id=6148e3aae7e0d36b5975 Since then we are enabling seamless cube textures on GLES2 which presumably the test suite isn't expecting. However, if I force it to always disable seamless cube textures I get the results below. These are different from those reported in this bug, so something doesn't quite add up. dEQP-GLES2.functional.texture.vertex.cube.wrap.clamp_repeat Fail dEQP-GLES2.functional.texture.vertex.cube.wrap.mirror_repeat Fail dEQP-GLES2.functional.texture.vertex.cube.wrap.repeat_clamp Fail dEQP-GLES2.functional.texture.vertex.cube.wrap.repeat_mirror Fail dEQP-GLES2.functional.texture.vertex.cube.wrap.repeat_repeat Fail dEQP-GLES2.functional.texture.vertex.cube.wrap.clamp_clamp Pass dEQP-GLES2.functional.texture.vertex.cube.wrap.clamp_mirror Pass dEQP-GLES2.functional.texture.vertex.cube.wrap.mirror_clamp Pass dEQP-GLES2.functional.texture.vertex.cube.wrap.mirror_mirror Pass
I don't see this test in the result from Mesa's CI runs. Can you include the commit of the dEQP source tree that you are using? It may be a new test.
Ok, I think the thing about all of the tests failing is a red herring, sorry about that. For some reason I was passing --deqp-surface-height=100 to dEQP (I guess I must've copied it from somewhere) and it broke everything. If I don't use any options then I get the same results as Elio. However, if I tweak _mesa_init_texture to always set ctx->Texture.CubeMapSeamless to false then it fixes all of the tests except this one: dEQP-GLES2.functional.texture.vertex.cube.wrap.repeat_repeat I don't know why the tests don't appear in the CI. As far as I can tell they were added since the first import commit in the dEQP repo. I will investigate a bit more.
Created attachment 121597 [details] [review] Patch for dEQP to stop the tests from skipping Mark, is the CI system using Chad's dEQP branch with --deqp-surface-type=fbo? If I set the surface type to FBO like that then all of these tests end up as NotSupported. Maybe that is why they're not appearing. This seems like a bug in dEQP because it seems to be getting confused about multisampling. I'm attaching a patch for dEQP which could fix it.
bugzilla-daemon@freedesktop.org writes: > https://bugs.freedesktop.org/show_bug.cgi?id=92579 > > --- Comment #8 from Neil Roberts <neil@linux.intel.com> --- > Created attachment 121597 [details] [review] > --> https://bugs.freedesktop.org/attachment.cgi?id=121597&action=edit > Patch for dEQP to stop the tests from skipping > > Mark, is the CI system using Chad's dEQP branch with > --deqp-surface-type=fbo? yes. > If I set the surface type to FBO like that then all of these tests end > up as NotSupported. Maybe that is why they're not appearing. This > seems like a bug in dEQP because it seems to be getting confused about > multisampling. I'm attaching a patch for dEQP which could fix it. thanks, I'll apply the patch to Chad's branch. > -- > You are receiving this mail because: > You are on the CC list for the bug.
Created attachment 121598 [details] [review] Hacky patch to make the tests pass Here's a hacky patch which makes all the tests pass. However I don't think we can really do this properly without also breaking GLES3. I'm inclined to say we shouldn't bother fixing this and we should consider the tests broken because they are testing behaviour that has changed in GLES3. See the patch commit message for more details.
Neil, thank you for investigating this. However, I'm concerned that investigating bugs in dEQP is a poor use of your time and expertise. Until the Mesa team has prioritized 100% compliance with dEQP as a target for the product, the validation has been instructed to write up bugs against Mesa only when: 1) a specific mesa commit has been bisected which triggers the regression 2) failing that, a clear regression has been shown between mesa versions. I do not see either of those criteria met in the history of this bug. I hope that Neil's investigation will be useful if and when Mesa begins a coordinated effort to fix known dEQP failures.
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.