Re: [Mesa-dev] [PATCH 00/10] Various fixes for dEQP failing tests
On 12/01/2014 05:29 PM, Brian Paul wrote: On 12/01/2014 06:04 AM, Eduardo Lima Mitev wrote: This set of (unrelated) patches fixes over 230 tests from the dEQP test suite [1]. [...] The non-GLSL changes look good to me and are: Reviewed-by: Brian Paul bri...@vmware.com The GLSL changes look OK too but should probably be reviewed by someone more active in that area. Thanks Brian, We fixed the issues you commented, and after we get a go for the GLSL ones, Iago will push them. cheers, Eduardo ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH 00/10] Various fixes for dEQP failing tests
This set of (unrelated) patches fixes over 230 tests from the dEQP test suite [1]. While the tests failures were gathered on i965 (gen8) against 10.3.3, there are several driver and version agnostic fixes. There is one piglit regression brought by patch 10 glsl: fail when a shader's input var has not an equivalent out var in previous, specifically in test: bin/glslparsertest \ tests/spec/glsl-1.50/compiler/gs-also-uses-smooth-flat-noperspective.geom pass \ 1.50 --check-link The piglit test seems to be wrong in this case because it contradicts the spec, please check that commit's log for details. It would be good to hear other people's comments on this in case we are missing something. The plan is to continue sending these random fixes in batches like this one, with the purpose of improving the pass rate of dEQP tests against Mesa. Eduardo Lima [1] https://android.googlesource.com/platform/external/deqp/ Patch summary: Eduardo Lima Mitev (4): mesa: Considers GL_DEPTH_STENCIL_ATTACHMENT a valid argument for FBO invalidation under GLES3 mesa: Enables GL_RGB and GL_RGBA unsized internal formats for OpenGL ES 3.0 mesa: Returns zero samples when querying GL_NUM_SAMPLE_COUNTS when internal format is integer glsl: Don't allow gl_FragData[i], with i0 in GLES shaders Iago Toral Quiroga (1): mesa: Recompute LegalTypesMask if the GL API has changed Samuel Iglesias Gonsalvez (5): glsl: invariant qualifier is not valid for shader inputs in GLSL ES 3.00 glsl: don't allow invariant qualifiers for interface blocks in GLSL ES glsl: A shader cannot redefine or overload built-in functions in GLSL ES 3.00 main/glsles: return two minor digits for SHADING_LANGUAGE_VERSION glsl: fail when a shader's input var has not an equivalent out var in previous src/glsl/ast_array_index.cpp | 21 +++- src/glsl/ast_to_hir.cpp| 22 + src/glsl/builtin_functions.cpp | 11 +++ src/glsl/glsl_parser.yy| 27 ++ src/glsl/ir.h | 4 src/glsl/link_varyings.cpp | 18 - src/mesa/main/fbobject.c | 8 src/mesa/main/formatquery.c| 44 ++ src/mesa/main/getstring.c | 2 +- src/mesa/main/mtypes.h | 3 ++- src/mesa/main/varray.c | 3 ++- 11 files changed, 141 insertions(+), 22 deletions(-) -- 2.1.3 ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH 00/10] Various fixes for dEQP failing tests
On 12/01/2014 06:04 AM, Eduardo Lima Mitev wrote: This set of (unrelated) patches fixes over 230 tests from the dEQP test suite [1]. While the tests failures were gathered on i965 (gen8) against 10.3.3, there are several driver and version agnostic fixes. There is one piglit regression brought by patch 10 glsl: fail when a shader's input var has not an equivalent out var in previous, specifically in test: bin/glslparsertest \ tests/spec/glsl-1.50/compiler/gs-also-uses-smooth-flat-noperspective.geom pass \ 1.50 --check-link The piglit test seems to be wrong in this case because it contradicts the spec, please check that commit's log for details. It would be good to hear other people's comments on this in case we are missing something. The plan is to continue sending these random fixes in batches like this one, with the purpose of improving the pass rate of dEQP tests against Mesa. Eduardo Lima [1] https://urldefense.proofpoint.com/v2/url?u=https-3A__android.googlesource.com_platform_external_deqp_d=AAIGaQc=Sqcl0Ez6M0X8aeM67LKIiDJAXVeAw-YihVMNtXt-uEsr=T0t4QG7chq2ZwJo6wilkFznRSFy-8uDKartPGbomVj8m=fAE0uqehtpO4ulnoKJnU_Eg9a2TCw24OVXuq_gLWGbMs=45bTrgaG9RYA41Cln3IrEM3mgeyiZmCYRGndRq3Ovy0e= Patch summary: Eduardo Lima Mitev (4): mesa: Considers GL_DEPTH_STENCIL_ATTACHMENT a valid argument for FBO invalidation under GLES3 mesa: Enables GL_RGB and GL_RGBA unsized internal formats for OpenGL ES 3.0 mesa: Returns zero samples when querying GL_NUM_SAMPLE_COUNTS when internal format is integer glsl: Don't allow gl_FragData[i], with i0 in GLES shaders Iago Toral Quiroga (1): mesa: Recompute LegalTypesMask if the GL API has changed Samuel Iglesias Gonsalvez (5): glsl: invariant qualifier is not valid for shader inputs in GLSL ES 3.00 glsl: don't allow invariant qualifiers for interface blocks in GLSL ES glsl: A shader cannot redefine or overload built-in functions in GLSL ES 3.00 main/glsles: return two minor digits for SHADING_LANGUAGE_VERSION glsl: fail when a shader's input var has not an equivalent out var in previous src/glsl/ast_array_index.cpp | 21 +++- src/glsl/ast_to_hir.cpp| 22 + src/glsl/builtin_functions.cpp | 11 +++ src/glsl/glsl_parser.yy| 27 ++ src/glsl/ir.h | 4 src/glsl/link_varyings.cpp | 18 - src/mesa/main/fbobject.c | 8 src/mesa/main/formatquery.c| 44 ++ src/mesa/main/getstring.c | 2 +- src/mesa/main/mtypes.h | 3 ++- src/mesa/main/varray.c | 3 ++- 11 files changed, 141 insertions(+), 22 deletions(-) The non-GLSL changes look good to me and are: Reviewed-by: Brian Paul bri...@vmware.com The GLSL changes look OK too but should probably be reviewed by someone more active in that area. -Brian ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
