| Summary: | program interface queries not returning right data for UBO / GL_BLOCK_INDEX | ||
|---|---|---|---|
| Product: | Mesa | Reporter: | Ilia Mirkin <imirkin> |
| Component: | Mesa core | Assignee: | Kenneth Graunke <kenneth> |
| Status: | RESOLVED FIXED | QA Contact: | mesa-dev |
| Severity: | normal | ||
| Priority: | medium | CC: | t_arceri |
| Version: | git | ||
| Hardware: | Other | ||
| OS: | All | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
| Bug Depends on: | |||
| Bug Blocks: | 94448 | ||
| Attachments: |
Test results for dEQP-GLES31.functional.program_interface_query.uniform.block_index.*
Partial fix |
||
Created attachment 121732 [details] [review] Partial fix This is a partial fix although I wonder if it needs to apply to SSBO too. The problem is we are trying to match the program resource index not the block index. I'm working on a fix for this. Fixed by: commit 0fbd073dc284f952ea7df691941a65ddc89b7554 glsl: store ubo or ssbo index in block index Previously we store the buffer block index i.e the index of a combined ubo/ssbo list. Fixes several dEQP-GLES31.functional tests: - program_interface_query.uniform.block_index.block_array - program_interface_query.uniform.block_index.named_block - program_interface_query.uniform.block_index.unnamed_block - program_interface_query.uniform.random.10 - program_interface_query.uniform.random.15 - program_interface_query.uniform.random.22 - program_interface_query.uniform.random.24 - program_interface_query.uniform.random.26 - program_interface_query.uniform.random.28 - program_interface_query.uniform.random.3 - program_interface_query.uniform.random.31 - program_interface_query.uniform.random.38 - program_interface_query.uniform.random.5 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94116 Reviewed-by: Kenneth Graunke <kenneth@whitecape.org> |
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.
Created attachment 121715 [details] Test results for dEQP-GLES31.functional.program_interface_query.uniform.block_index.* Not 100% sure that my diagnosis is correct, but it seems that there's some disagreement between the GL_BLOCK_INDEX being returned by glGetProgramResourceiv and the resource index accepted by glGetProgramResourceName. Note the output of these dEQP tests: MESA_GLES_VERSION_OVERRIDE=3.1 ./deqp-gles31 --deqp-visibility=hidden --deqp-case='dEQP-GLES31.functional.program_interface_query.uniform.block_index.*' dEQP Core git-0c2b7cdc3a0455e0486f88574021d3ccd22f28a8 (0x0c2b7cdc) starting.. target implementation = 'X11 EGL/GLX' Test case 'dEQP-GLES31.functional.program_interface_query.uniform.block_index.default_block'.. Compute shader compile time = 2.354000 ms Link time = 0.581000 ms Test case duration in microseconds = 6153 us Pass (Pass) Test case 'dEQP-GLES31.functional.program_interface_query.uniform.block_index.named_block'.. Compute shader compile time = 0.319000 ms Link time = 0.522000 ms Mesa: User error: GL_INVALID_VALUE in glGetProgramResourceName(index 1) Test case duration in microseconds = 2495 us Fail (query block name: glGetError() returned GL_INVALID_VALUE at es31fProgramInterfaceQueryTestCase.cpp:609) Test case 'dEQP-GLES31.functional.program_interface_query.uniform.block_index.unnamed_block'.. Compute shader compile time = 0.291000 ms Link time = 0.526000 ms Mesa: User error: GL_INVALID_VALUE in glGetProgramResourceName(index 1) Test case duration in microseconds = 2408 us Fail (query block name: glGetError() returned GL_INVALID_VALUE at es31fProgramInterfaceQueryTestCase.cpp:609) Test case 'dEQP-GLES31.functional.program_interface_query.uniform.block_index.block_array'.. Compute shader compile time = 0.398000 ms Link time = 1.887000 ms Test case duration in microseconds = 3985 us Fail (resource block index invalid) DONE! Now the difference between the default_block one and the others is that the others have a layout(binding=0). For default block, it returns GL_BLOCK_INDEX: -1, and it's all happy about that. I would sort of have assumed that for the others, the block index should be 0, but I'm not 100% sure. When it goes to use this block index as an argument to glGetProgramResouceName it errors out for the first 2, and for the last one it gets: <Text>Block name with index 1 is "TargetInterface[1]"</Text> <Text> Error, expected TargetInterface[0]</Text> So I suspect there's an off-by-one somewhere. Attaching the test result output from dEQP which includes the shaders.