path: root/src/compiler
AgeCommit message (Expand)AuthorFilesLines
2020-10-14scons: fix SPIR-V -> NIR buildRhys Perry1-0/+7
2020-10-14android: fix SPIR-V -> NIR buildRhys Perry3-2/+8
2020-10-14spirv: replace discard with demote for incorrect HLSL->SPIR-V translationsRhys Perry3-4/+31
2020-10-14spirv: add and use a generator id enumRhys Perry6-2/+278
2020-10-14glsl/ fix importsDylan Baker1-2/+1
2020-10-14nir/opt_load_store_vectorize: Use bit sizes when checking mask compatibilityJason Ekstrand1-2/+2
2020-10-14glsl: don't duplicate state vars as uniforms in the NIR linkerTimothy Arceri1-0/+6
2020-10-14nir/cf: Better handle intra-block splitsJason Ekstrand1-7/+24
2020-09-30nir/lower_io_arrays: Fix xfb_offset bugConnor Abbott1-1/+1
2020-09-28spirv: fix emitting switch cases that directly jump to the merge blockSamuel Pitoiset1-0/+7
2020-09-28spirv: extract switch parsing into its own functionKarol Herbst2-83/+99
2020-09-27nir/liveness: Consider if uses in nir_ssa_defs_interfereJason Ekstrand1-0/+9
2020-09-23nir/lower_samplers: Clamp out-of-bounds access to array of samplersDanylo Piliaiev1-1/+21
2020-09-23nir/large_constants: Eliminate out-of-bounds writes to large constantsDanylo Piliaiev1-2/+5
2020-09-23nir/lower_io: Eliminate oob writes and return zero for oob readsDanylo Piliaiev1-0/+31
2020-09-23glsl_type: Add packed to structure type comparison for hash mapJesse Natalie1-0/+3
2020-09-16nir: More NIR_MAX_VEC_COMPONENTS fixesJason Ekstrand9-10/+19
2020-09-16spirv: Run repair_ssa if there are discard instructionsJason Ekstrand2-1/+3
2020-09-16nir/opt_if: fix opt_if_merge when destination branch has a jumpRhys Perry1-0/+7
2020-09-02nir: Add a nir_metadata_all enum valueJason Ekstrand1-0/+9
2020-09-02nir/algebraic: mark some optimizations with fsat(NaN) as inexactSamuel Pitoiset1-3/+9
2020-09-02glsl: Eliminate assigments to out-of-bounds elements of vectorDanylo Piliaiev1-8/+24
2020-09-02ir_constant: Return zero on out-of-bounds vector accessesDanylo Piliaiev1-0/+14
2020-09-02glsl: Eliminate out-of-bounds triop_vector_insertDanylo Piliaiev1-1/+32
2020-08-19spirv: Don't emit RMW for vector indexing in shared or globalJason Ekstrand1-5/+17
2020-08-19spirv: Only require bare types to match when copying variablesDanylo Piliaiev1-1/+2
2020-08-19nir/lower_idiv: Port recent LLVM fixes to emit_udivTony Wasserka1-22/+23
2020-08-05nir/deref: Don't try to compare derefs containing castsJason Ekstrand1-4/+6
2020-07-22nir/opt_if: Fix opt_if_simplification when else branch has jumpDanylo Piliaiev1-0/+11
2020-07-22nir/lower_int64: lower 64-bit amulRhys Perry1-4/+22
2020-07-22spirv: Allow block-decorated struct types for constantsJason Ekstrand2-1/+3
2020-07-22spirv: Skip phis in unreachable blocks in the second phi passJason Ekstrand1-1/+8
2020-07-22glsl: only allow 32 bits atomic operations on imagesPierre-Eric Pelloux-Prayer1-0/+44
2020-07-22glsl: don't expose imageAtomicIncWrap for signed imagePierre-Eric Pelloux-Prayer1-15/+29
2020-07-22glsl: reject size1x8 for image variable with floating-point data typesPierre-Eric Pelloux-Prayer1-5/+7
2020-07-08nir/algebraic: Don't distrubte absolute-value into dot-productsIan Romanick1-7/+7
2020-07-08nir: add missing break to nir_opt_access()Timothy Arceri1-0/+1
2020-07-08glsl: fix constant packing for 64-bit big endian.Dave Airlie2-10/+25
2020-06-24nir: only uniforms with dynamically_uniform offset are dynamically_uniformJose Maria Casanova Crespo1-2/+2
2020-06-24spirv: fix using OpSampledImage with OpUndef instead of OpType{Image,Sampler}Samuel Pitoiset1-4/+22
2020-06-24glsl: fix incorrect optimisation in opt_constant_variable()Timothy Arceri1-0/+9
2020-06-10glsl: fix crash on glsl macro redefinitionAndrii Simiklit1-0/+3
2020-06-10glsl: inline functions with unsupported return type before converting to nirDanylo Piliaiev1-0/+6
2020-06-10nir/lower_explicit_io: fix NON_UNIFORM access for UBO loadsSamuel Pitoiset1-1/+1
2020-06-10glsl: fix potential slow compile times for GLSLOptimizeConservativelyTimothy Arceri1-1/+14
2020-06-10nir: reuse existing psiz-variableErik Faye-Lund1-6/+20
2020-06-10glsl: Don't replace lrp pattern with lrp if arguments are not floatsDanylo Piliaiev1-1/+2
2020-06-10nir: lower_tex: Don't normalize coordinates for TXF with RECTGert Wollny1-1/+2
2020-06-10spirv,radv,anv: implement no-op VK_GOOGLE_user_typeSamuel Pitoiset2-0/+6
2020-05-27glsl: fix slow linking of uniforms in the nir linkerTimothy Arceri1-9/+81