| Summary: | [HSW] Mesa 11.0, lots of failing UBO conformance tests | ||
|---|---|---|---|
| Product: | Mesa | Reporter: | Tapani Pälli <lemody> |
| Component: | Drivers/DRI/i965 | Assignee: | Ian Romanick <idr> |
| Status: | RESOLVED WORKSFORME | QA Contact: | Intel 3D Bugs Mailing List <intel-3d-bugs> |
| Severity: | major | ||
| Priority: | high | ||
| Version: | git | ||
| Hardware: | Other | ||
| OS: | All | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
|
Description
Tapani Pälli
2015-09-25 11:37:16 UTC
With Mesa master there are fewer errors, as example following results this morning Mesa (64831832791139328a67b80387f062d39e304d24) --- 8< --- ES3-CTS.shaders.uniform_block.single_basic_type.shared.highp_uvec4 ES3-CTS.shaders.uniform_block.single_basic_type.shared.highp_mat3 ES3-CTS.shaders.uniform_block.single_basic_type.shared.row_major_lowp_mat3 ES3-CTS.shaders.uniform_block.single_basic_type.shared.row_major_mediump_mat3 ES3-CTS.shaders.uniform_block.single_basic_type.shared.row_major_highp_mat3 ES3-CTS.shaders.uniform_block.single_basic_type.shared.row_major_mediump_mat4 ES3-CTS.shaders.uniform_block.single_basic_type.shared.row_major_lowp_mat2x3 ES3-CTS.shaders.uniform_block.single_basic_type.shared.highp_mat2x4 ES3-CTS.shaders.uniform_block.single_basic_type.shared.row_major_lowp_mat2x4 ES3-CTS.shaders.uniform_block.single_basic_type.shared.row_major_mediump_mat2x4 ES3-CTS.shaders.uniform_block.single_basic_type.shared.row_major_highp_mat2x4 ES3-CTS.shaders.uniform_block.single_basic_type.shared.column_major_lowp_mat2x4 ES3-CTS.shaders.uniform_block.single_basic_type.shared.column_major_mediump_mat2x4 ES3-CTS.shaders.uniform_block.single_basic_type.shared.column_major_highp_mat2x4 ES3-CTS.shaders.uniform_block.single_basic_type.shared.mediump_mat3x2 ES3-CTS.shaders.uniform_block.single_basic_type.shared.column_major_mediump_mat4x2 ES3-CTS.shaders.uniform_block.single_basic_type.packed.highp_float ES3-CTS.shaders.uniform_block.single_basic_type.packed.highp_mat4 ES3-CTS.shaders.uniform_block.single_basic_type.packed.lowp_mat2x3 ES3-CTS.shaders.uniform_block.single_basic_type.std140.row_major_mediump_mat2x3 ES3-CTS.shaders.uniform_block.single_basic_array.shared.ivec2 ES3-CTS.shaders.uniform_block.single_basic_array.packed.vec3 ES3-CTS.shaders.uniform_block.single_basic_array.std140.ivec2 ES3-CTS.shaders.uniform_block.instance_array_basic_type.shared.vec2 ES3-CTS.shaders.uniform_block.instance_array_basic_type.shared.ivec2 ES3-CTS.shaders.uniform_block.instance_array_basic_type.shared.ivec4 ES3-CTS.shaders.uniform_block.instance_array_basic_type.packed.vec3 ES3-CTS.shaders.uniform_block.instance_array_basic_type.std140.vec2 ES3-CTS.shaders.uniform_block.instance_array_basic_type.std140.ivec2 ES3-CTS.shaders.uniform_block.random.scalar_types.6 ES3-CTS.shaders.uniform_block.random.scalar_types.9 On my HSW laptop I cannot reproduce any of these fails. I get 100% pass ratio for ES3-CTS.shaders.uniform_block.* I'm using CTS 3.1.1.0 though. Should this be tested specifically with CTS-3.0 (e.g ESCTS-3.0.4.0) to reproduce the fails? Otherwise, I think this can be close as already-fixed. (In reply to Eduardo Lima Mitev from comment #2) > On my HSW laptop I cannot reproduce any of these fails. I get 100% pass > ratio for ES3-CTS.shaders.uniform_block.* > > I'm using CTS 3.1.1.0 though. Should this be tested specifically with > CTS-3.0 (e.g ESCTS-3.0.4.0) to reproduce the fails? Otherwise, I think this > can be close as already-fixed. individually they pass but if you run the whole ES3 tests like this: ./glcts --deqp-case=ES3-CTS* then they fail. (In reply to Tapani Pälli from comment #3) > (In reply to Eduardo Lima Mitev from comment #2) > > On my HSW laptop I cannot reproduce any of these fails. I get 100% pass > > ratio for ES3-CTS.shaders.uniform_block.* > > > > I'm using CTS 3.1.1.0 though. Should this be tested specifically with > > CTS-3.0 (e.g ESCTS-3.0.4.0) to reproduce the fails? Otherwise, I think this > > can be close as already-fixed. > > individually they pass but if you run the whole ES3 tests like this: > > ./glcts --deqp-case=ES3-CTS* > > then they fail. Just tried that, and I get this: Test run totals: Passed: 2996/2997 (99.97%) Failed: 0/2997 (0.00%) Not supported: 1/2997 (0.03%) Warnings: 0/2997 (0.00%) This was with mesa HEAD git~edf774b (master from this morning), ESCTS-3.1.1.0, on my HSW laptop. (In reply to Eduardo Lima Mitev from comment #4) > (In reply to Tapani Pälli from comment #3) > > (In reply to Eduardo Lima Mitev from comment #2) > > > On my HSW laptop I cannot reproduce any of these fails. I get 100% pass > > > ratio for ES3-CTS.shaders.uniform_block.* > > > > > > I'm using CTS 3.1.1.0 though. Should this be tested specifically with > > > CTS-3.0 (e.g ESCTS-3.0.4.0) to reproduce the fails? Otherwise, I think this > > > can be close as already-fixed. > > > > individually they pass but if you run the whole ES3 tests like this: > > > > ./glcts --deqp-case=ES3-CTS* > > > > then they fail. > > Just tried that, and I get this: > > Test run totals: > Passed: 2996/2997 (99.97%) > Failed: 0/2997 (0.00%) > Not supported: 1/2997 (0.03%) > Warnings: 0/2997 (0.00%) > > This was with mesa HEAD git~edf774b (master from this morning), > ESCTS-3.1.1.0, on my HSW laptop. I'm running on same Mesa, CTS is a snapshot (not master but should not be very old) from gitlab and says 'GL-CTS-2.0' as releaseName. I have 2887 passing, 109 failing (all of these ubo tests) and 1 not supported. This is Haswell desktop GT2 (0x0412). (In reply to Tapani Pälli from comment #5) > (In reply to Eduardo Lima Mitev from comment #4) > > (In reply to Tapani Pälli from comment #3) > > > (In reply to Eduardo Lima Mitev from comment #2) > > > > On my HSW laptop I cannot reproduce any of these fails. I get 100% pass > > > > ratio for ES3-CTS.shaders.uniform_block.* > > > > > > > > I'm using CTS 3.1.1.0 though. Should this be tested specifically with > > > > CTS-3.0 (e.g ESCTS-3.0.4.0) to reproduce the fails? Otherwise, I think this > > > > can be close as already-fixed. > > > > > > individually they pass but if you run the whole ES3 tests like this: > > > > > > ./glcts --deqp-case=ES3-CTS* > > > > > > then they fail. > > > > Just tried that, and I get this: > > > > Test run totals: > > Passed: 2996/2997 (99.97%) > > Failed: 0/2997 (0.00%) > > Not supported: 1/2997 (0.03%) > > Warnings: 0/2997 (0.00%) > > > > This was with mesa HEAD git~edf774b (master from this morning), > > ESCTS-3.1.1.0, on my HSW laptop. > > I'm running on same Mesa, CTS is a snapshot (not master but should not be > very old) from gitlab and says 'GL-CTS-2.0' as releaseName. I have 2887 > passing, 109 failing (all of these ubo tests) and 1 not supported. This is > Haswell desktop GT2 (0x0412). This CTS snapshot is from last September and has a set of internal patches to enable using gbm + Intel support for image external extension. I'm resolving this as WORKSFORME as I cannot reproduce this anymore. There are numerous ES3-CTS tests failing for HSW but these ones are passing. |
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.