X-Git-Url: https://git.libre-soc.org/?a=blobdiff_plain;f=src%2Fglsl%2Fbuiltin_functions.cpp;h=9973a7630870bdaea24d24c421df1e23c02fc93a;hb=f821a3ec4f8dda55722f326f26855c2b24ca186c;hp=84bbdc2a4c6c8d9108765c190770d3b9d3646c9b;hpb=c16c719647664abf7531896770c31eda1dbe7a58;p=mesa.git diff --git a/src/glsl/builtin_functions.cpp b/src/glsl/builtin_functions.cpp index 84bbdc2a4c6..9973a763087 100644 --- a/src/glsl/builtin_functions.cpp +++ b/src/glsl/builtin_functions.cpp @@ -60,7 +60,7 @@ #include "ir_builder.h" #include "glsl_parser_extras.h" #include "program/prog_instruction.h" -#include +#include #define M_PIf ((float) M_PI) #define M_PI_2f ((float) M_PI_2) @@ -135,6 +135,19 @@ v140(const _mesa_glsl_parse_state *state) return state->is_version(140, 0); } +static bool +v400_fs_only(const _mesa_glsl_parse_state *state) +{ + return state->is_version(400, 0) && + state->stage == MESA_SHADER_FRAGMENT; +} + +static bool +es31(const _mesa_glsl_parse_state *state) +{ + return state->is_version(0, 310); +} + static bool texture_rectangle(const _mesa_glsl_parse_state *state) { @@ -194,14 +207,15 @@ shader_bit_encoding(const _mesa_glsl_parse_state *state) static bool shader_integer_mix(const _mesa_glsl_parse_state *state) { - return v130(state) && state->EXT_shader_integer_mix_enable; + return state->is_version(450, 310) || + (v130(state) && state->EXT_shader_integer_mix_enable); } static bool shader_packing_or_es3(const _mesa_glsl_parse_state *state) { return state->ARB_shading_language_packing_enable || - state->is_version(400, 300); + state->is_version(420, 300); } static bool @@ -219,10 +233,17 @@ gpu_shader5(const _mesa_glsl_parse_state *state) } static bool -shader_packing_or_gpu_shader5(const _mesa_glsl_parse_state *state) +gpu_shader5_or_es31(const _mesa_glsl_parse_state *state) +{ + return state->is_version(400, 310) || state->ARB_gpu_shader5_enable; +} + +static bool +shader_packing_or_es31_or_gpu_shader5(const _mesa_glsl_parse_state *state) { return state->ARB_shading_language_packing_enable || - gpu_shader5(state); + state->ARB_gpu_shader5_enable || + state->is_version(400, 310); } static bool @@ -256,10 +277,32 @@ texture_array(const _mesa_glsl_parse_state *state) static bool texture_multisample(const _mesa_glsl_parse_state *state) { - return state->is_version(150, 0) || + return state->is_version(150, 310) || state->ARB_texture_multisample_enable; } +static bool +texture_multisample_array(const _mesa_glsl_parse_state *state) +{ + return state->is_version(150, 320) || + state->ARB_texture_multisample_enable || + state->OES_texture_storage_multisample_2d_array_enable; +} + +static bool +texture_samples_identical(const _mesa_glsl_parse_state *state) +{ + return texture_multisample(state) && + state->EXT_shader_samples_identical_enable; +} + +static bool +texture_samples_identical_array(const _mesa_glsl_parse_state *state) +{ + return texture_multisample_array(state) && + state->EXT_shader_samples_identical_enable; +} + static bool fs_texture_cube_map_array(const _mesa_glsl_parse_state *state) { @@ -297,15 +340,24 @@ texture_gather(const _mesa_glsl_parse_state *state) state->ARB_gpu_shader5_enable; } +static bool +texture_gather_or_es31(const _mesa_glsl_parse_state *state) +{ + return state->is_version(400, 310) || + state->ARB_texture_gather_enable || + state->ARB_gpu_shader5_enable; +} + /* Only ARB_texture_gather but not GLSL 4.0 or ARB_gpu_shader5. * used for relaxation of const offset requirements. */ static bool -texture_gather_only(const _mesa_glsl_parse_state *state) +texture_gather_only_or_es31(const _mesa_glsl_parse_state *state) { return !state->is_version(400, 0) && !state->ARB_gpu_shader5_enable && - state->ARB_texture_gather_enable; + (state->ARB_texture_gather_enable || + state->is_version(0, 310)); } /* Desktop GL or OES_standard_derivatives + fragment shader only */ @@ -359,7 +411,19 @@ tex3d_lod(const _mesa_glsl_parse_state *state) static bool shader_atomic_counters(const _mesa_glsl_parse_state *state) { - return state->ARB_shader_atomic_counters_enable; + return state->has_atomic_counters(); +} + +static bool +shader_clock(const _mesa_glsl_parse_state *state) +{ + return state->ARB_shader_clock_enable; +} + +static bool +shader_storage_buffer_object(const _mesa_glsl_parse_state *state) +{ + return state->has_shader_storage_buffer_objects(); } static bool @@ -370,11 +434,32 @@ shader_trinary_minmax(const _mesa_glsl_parse_state *state) static bool shader_image_load_store(const _mesa_glsl_parse_state *state) +{ + return (state->is_version(420, 310) || + state->ARB_shader_image_load_store_enable); +} + +static bool +shader_image_atomic(const _mesa_glsl_parse_state *state) { return (state->is_version(420, 0) || state->ARB_shader_image_load_store_enable); } +static bool +shader_image_size(const _mesa_glsl_parse_state *state) +{ + return state->is_version(430, 310) || + state->ARB_shader_image_size_enable; +} + +static bool +shader_samples(const _mesa_glsl_parse_state *state) +{ + return state->is_version(450, 0) || + state->ARB_shader_texture_image_samples_enable; +} + static bool gs_streams(const _mesa_glsl_parse_state *state) { @@ -387,6 +472,25 @@ fp64(const _mesa_glsl_parse_state *state) return state->has_double(); } +static bool +compute_shader(const _mesa_glsl_parse_state *state) +{ + return state->stage == MESA_SHADER_COMPUTE; +} + +static bool +buffer_atomics_supported(const _mesa_glsl_parse_state *state) +{ + return compute_shader(state) || shader_storage_buffer_object(state); +} + +static bool +barrier_supported(const _mesa_glsl_parse_state *state) +{ + return compute_shader(state) || + state->stage == MESA_SHADER_TESS_CTRL; +} + /** @} */ /******************************************************************************/ @@ -462,13 +566,19 @@ private: /** Create a new function and add the given signatures. */ void add_function(const char *name, ...); + typedef ir_function_signature *(builtin_builder::*image_prototype_ctr)(const glsl_type *image_type, + unsigned num_arguments, + unsigned flags); + enum image_function_flags { IMAGE_FUNCTION_EMIT_STUB = (1 << 0), IMAGE_FUNCTION_RETURNS_VOID = (1 << 1), IMAGE_FUNCTION_HAS_VECTOR_DATA_TYPE = (1 << 2), IMAGE_FUNCTION_SUPPORTS_FLOAT_DATA_TYPE = (1 << 3), IMAGE_FUNCTION_READ_ONLY = (1 << 4), - IMAGE_FUNCTION_WRITE_ONLY = (1 << 5) + IMAGE_FUNCTION_WRITE_ONLY = (1 << 5), + IMAGE_FUNCTION_AVAIL_ATOMIC = (1 << 6), + IMAGE_FUNCTION_MS_ONLY = (1 << 7), }; /** @@ -477,6 +587,7 @@ private: */ void add_image_function(const char *name, const char *intrinsic_name, + image_prototype_ctr prototype, unsigned num_arguments, unsigned flags); @@ -501,8 +612,8 @@ private: ir_expression_operation opcode, const glsl_type *return_type, const glsl_type *param_type); - ir_function_signature *binop(ir_expression_operation opcode, - builtin_available_predicate avail, + ir_function_signature *binop(builtin_available_predicate avail, + ir_expression_operation opcode, const glsl_type *return_type, const glsl_type *param0_type, const glsl_type *param1_type); @@ -600,10 +711,8 @@ private: B1(any); B1(all); B1(not); - B2(textureSize); - ir_function_signature *_textureSize(builtin_available_predicate avail, - const glsl_type *return_type, - const glsl_type *sampler_type); + BA2(textureSize); + B1(textureSamples); /** Flags to _texture() */ #define TEX_PROJECT 1 @@ -631,9 +740,11 @@ private: const glsl_type *stream_type); ir_function_signature *_EndStreamPrimitive(builtin_available_predicate avail, const glsl_type *stream_type); + B0(barrier) - B2(textureQueryLod); + BA2(textureQueryLod); B1(textureQueryLevels); + BA2(textureSamplesIdentical); B1(dFdx); B1(dFdy); B1(fwidth); @@ -665,27 +776,49 @@ private: B1(interpolateAtOffset) B1(interpolateAtSample) - ir_function_signature *_atomic_intrinsic(builtin_available_predicate avail); - ir_function_signature *_atomic_op(const char *intrinsic, - builtin_available_predicate avail); + ir_function_signature *_atomic_counter_intrinsic(builtin_available_predicate avail); + ir_function_signature *_atomic_counter_op(const char *intrinsic, + builtin_available_predicate avail); + + ir_function_signature *_atomic_intrinsic2(builtin_available_predicate avail, + const glsl_type *type); + ir_function_signature *_atomic_op2(const char *intrinsic, + builtin_available_predicate avail, + const glsl_type *type); + ir_function_signature *_atomic_intrinsic3(builtin_available_predicate avail, + const glsl_type *type); + ir_function_signature *_atomic_op3(const char *intrinsic, + builtin_available_predicate avail, + const glsl_type *type); B1(min3) B1(max3) B1(mid3) ir_function_signature *_image_prototype(const glsl_type *image_type, - const char *intrinsic_name, unsigned num_arguments, unsigned flags); - ir_function_signature *_image(const glsl_type *image_type, + ir_function_signature *_image_size_prototype(const glsl_type *image_type, + unsigned num_arguments, + unsigned flags); + ir_function_signature *_image_samples_prototype(const glsl_type *image_type, + unsigned num_arguments, + unsigned flags); + ir_function_signature *_image(image_prototype_ctr prototype, + const glsl_type *image_type, const char *intrinsic_name, unsigned num_arguments, unsigned flags); ir_function_signature *_memory_barrier_intrinsic( builtin_available_predicate avail); - ir_function_signature *_memory_barrier( - builtin_available_predicate avail); + ir_function_signature *_memory_barrier(const char *intrinsic_name, + builtin_available_predicate avail); + + ir_function_signature *_shader_clock_intrinsic(builtin_available_predicate avail, + const glsl_type *type); + ir_function_signature *_shader_clock(builtin_available_predicate avail, + const glsl_type *type); #undef B0 #undef B1 @@ -794,13 +927,62 @@ void builtin_builder::create_intrinsics() { add_function("__intrinsic_atomic_read", - _atomic_intrinsic(shader_atomic_counters), + _atomic_counter_intrinsic(shader_atomic_counters), NULL); add_function("__intrinsic_atomic_increment", - _atomic_intrinsic(shader_atomic_counters), + _atomic_counter_intrinsic(shader_atomic_counters), NULL); add_function("__intrinsic_atomic_predecrement", - _atomic_intrinsic(shader_atomic_counters), + _atomic_counter_intrinsic(shader_atomic_counters), + NULL); + + add_function("__intrinsic_atomic_add", + _atomic_intrinsic2(buffer_atomics_supported, + glsl_type::uint_type), + _atomic_intrinsic2(buffer_atomics_supported, + glsl_type::int_type), + NULL); + add_function("__intrinsic_atomic_min", + _atomic_intrinsic2(buffer_atomics_supported, + glsl_type::uint_type), + _atomic_intrinsic2(buffer_atomics_supported, + glsl_type::int_type), + NULL); + add_function("__intrinsic_atomic_max", + _atomic_intrinsic2(buffer_atomics_supported, + glsl_type::uint_type), + _atomic_intrinsic2(buffer_atomics_supported, + glsl_type::int_type), + NULL); + add_function("__intrinsic_atomic_and", + _atomic_intrinsic2(buffer_atomics_supported, + glsl_type::uint_type), + _atomic_intrinsic2(buffer_atomics_supported, + glsl_type::int_type), + NULL); + add_function("__intrinsic_atomic_or", + _atomic_intrinsic2(buffer_atomics_supported, + glsl_type::uint_type), + _atomic_intrinsic2(buffer_atomics_supported, + glsl_type::int_type), + NULL); + add_function("__intrinsic_atomic_xor", + _atomic_intrinsic2(buffer_atomics_supported, + glsl_type::uint_type), + _atomic_intrinsic2(buffer_atomics_supported, + glsl_type::int_type), + NULL); + add_function("__intrinsic_atomic_exchange", + _atomic_intrinsic2(buffer_atomics_supported, + glsl_type::uint_type), + _atomic_intrinsic2(buffer_atomics_supported, + glsl_type::int_type), + NULL); + add_function("__intrinsic_atomic_comp_swap", + _atomic_intrinsic3(buffer_atomics_supported, + glsl_type::uint_type), + _atomic_intrinsic3(buffer_atomics_supported, + glsl_type::int_type), NULL); add_image_functions(false); @@ -808,6 +990,26 @@ builtin_builder::create_intrinsics() add_function("__intrinsic_memory_barrier", _memory_barrier_intrinsic(shader_image_load_store), NULL); + add_function("__intrinsic_group_memory_barrier", + _memory_barrier_intrinsic(compute_shader), + NULL); + add_function("__intrinsic_memory_barrier_atomic_counter", + _memory_barrier_intrinsic(compute_shader), + NULL); + add_function("__intrinsic_memory_barrier_buffer", + _memory_barrier_intrinsic(compute_shader), + NULL); + add_function("__intrinsic_memory_barrier_image", + _memory_barrier_intrinsic(compute_shader), + NULL); + add_function("__intrinsic_memory_barrier_shared", + _memory_barrier_intrinsic(compute_shader), + NULL); + + add_function("__intrinsic_shader_clock", + _shader_clock_intrinsic(shader_clock, + glsl_type::uvec2_type), + NULL); } /** @@ -1140,7 +1342,7 @@ builtin_builder::create_builtins() _smoothstep(fp64, glsl_type::dvec3_type, glsl_type::dvec3_type), _smoothstep(fp64, glsl_type::dvec4_type, glsl_type::dvec4_type), NULL); - + FD130(isnan) FD130(isinf) @@ -1161,12 +1363,12 @@ builtin_builder::create_builtins() add_function("packUnorm2x16", _packUnorm2x16(shader_packing_or_es3_or_gpu_shader5), NULL); add_function("packSnorm2x16", _packSnorm2x16(shader_packing_or_es3), NULL); - add_function("packUnorm4x8", _packUnorm4x8(shader_packing_or_gpu_shader5), NULL); - add_function("packSnorm4x8", _packSnorm4x8(shader_packing_or_gpu_shader5), NULL); + add_function("packUnorm4x8", _packUnorm4x8(shader_packing_or_es31_or_gpu_shader5), NULL); + add_function("packSnorm4x8", _packSnorm4x8(shader_packing_or_es31_or_gpu_shader5), NULL); add_function("unpackUnorm2x16", _unpackUnorm2x16(shader_packing_or_es3_or_gpu_shader5), NULL); add_function("unpackSnorm2x16", _unpackSnorm2x16(shader_packing_or_es3), NULL); - add_function("unpackUnorm4x8", _unpackUnorm4x8(shader_packing_or_gpu_shader5), NULL); - add_function("unpackSnorm4x8", _unpackSnorm4x8(shader_packing_or_gpu_shader5), NULL); + add_function("unpackUnorm4x8", _unpackUnorm4x8(shader_packing_or_es31_or_gpu_shader5), NULL); + add_function("unpackSnorm4x8", _unpackSnorm4x8(shader_packing_or_es31_or_gpu_shader5), NULL); add_function("packHalf2x16", _packHalf2x16(shader_packing_or_es3), NULL); add_function("unpackHalf2x16", _unpackHalf2x16(shader_packing_or_es3), NULL); add_function("packDouble2x32", _packDouble2x32(fp64), NULL); @@ -1177,7 +1379,7 @@ builtin_builder::create_builtins() FD(distance) FD(dot) - add_function("cross", _cross(always_available, glsl_type::vec3_type), + add_function("cross", _cross(always_available, glsl_type::vec3_type), _cross(fp64, glsl_type::dvec3_type), NULL); FD(normalize) @@ -1336,9 +1538,19 @@ builtin_builder::create_builtins() _textureSize(texture_multisample, glsl_type::ivec2_type, glsl_type::isampler2DMS_type), _textureSize(texture_multisample, glsl_type::ivec2_type, glsl_type::usampler2DMS_type), - _textureSize(texture_multisample, glsl_type::ivec3_type, glsl_type::sampler2DMSArray_type), - _textureSize(texture_multisample, glsl_type::ivec3_type, glsl_type::isampler2DMSArray_type), - _textureSize(texture_multisample, glsl_type::ivec3_type, glsl_type::usampler2DMSArray_type), + _textureSize(texture_multisample_array, glsl_type::ivec3_type, glsl_type::sampler2DMSArray_type), + _textureSize(texture_multisample_array, glsl_type::ivec3_type, glsl_type::isampler2DMSArray_type), + _textureSize(texture_multisample_array, glsl_type::ivec3_type, glsl_type::usampler2DMSArray_type), + NULL); + + add_function("textureSamples", + _textureSamples(glsl_type::sampler2DMS_type), + _textureSamples(glsl_type::isampler2DMS_type), + _textureSamples(glsl_type::usampler2DMS_type), + + _textureSamples(glsl_type::sampler2DMSArray_type), + _textureSamples(glsl_type::isampler2DMSArray_type), + _textureSamples(glsl_type::usampler2DMSArray_type), NULL); add_function("texture", @@ -1601,9 +1813,9 @@ builtin_builder::create_builtins() _texelFetch(texture_multisample, glsl_type::ivec4_type, glsl_type::isampler2DMS_type, glsl_type::ivec2_type), _texelFetch(texture_multisample, glsl_type::uvec4_type, glsl_type::usampler2DMS_type, glsl_type::ivec2_type), - _texelFetch(texture_multisample, glsl_type::vec4_type, glsl_type::sampler2DMSArray_type, glsl_type::ivec3_type), - _texelFetch(texture_multisample, glsl_type::ivec4_type, glsl_type::isampler2DMSArray_type, glsl_type::ivec3_type), - _texelFetch(texture_multisample, glsl_type::uvec4_type, glsl_type::usampler2DMSArray_type, glsl_type::ivec3_type), + _texelFetch(texture_multisample_array, glsl_type::vec4_type, glsl_type::sampler2DMSArray_type, glsl_type::ivec3_type), + _texelFetch(texture_multisample_array, glsl_type::ivec4_type, glsl_type::isampler2DMSArray_type, glsl_type::ivec3_type), + _texelFetch(texture_multisample_array, glsl_type::uvec4_type, glsl_type::usampler2DMSArray_type, glsl_type::ivec3_type), NULL); add_function("texelFetchOffset", @@ -1910,42 +2122,80 @@ builtin_builder::create_builtins() _EndStreamPrimitive(gs_streams, glsl_type::uint_type), _EndStreamPrimitive(gs_streams, glsl_type::int_type), NULL); + add_function("barrier", _barrier(), NULL); add_function("textureQueryLOD", - _textureQueryLod(glsl_type::sampler1D_type, glsl_type::float_type), - _textureQueryLod(glsl_type::isampler1D_type, glsl_type::float_type), - _textureQueryLod(glsl_type::usampler1D_type, glsl_type::float_type), - - _textureQueryLod(glsl_type::sampler2D_type, glsl_type::vec2_type), - _textureQueryLod(glsl_type::isampler2D_type, glsl_type::vec2_type), - _textureQueryLod(glsl_type::usampler2D_type, glsl_type::vec2_type), - - _textureQueryLod(glsl_type::sampler3D_type, glsl_type::vec3_type), - _textureQueryLod(glsl_type::isampler3D_type, glsl_type::vec3_type), - _textureQueryLod(glsl_type::usampler3D_type, glsl_type::vec3_type), - - _textureQueryLod(glsl_type::samplerCube_type, glsl_type::vec3_type), - _textureQueryLod(glsl_type::isamplerCube_type, glsl_type::vec3_type), - _textureQueryLod(glsl_type::usamplerCube_type, glsl_type::vec3_type), - - _textureQueryLod(glsl_type::sampler1DArray_type, glsl_type::float_type), - _textureQueryLod(glsl_type::isampler1DArray_type, glsl_type::float_type), - _textureQueryLod(glsl_type::usampler1DArray_type, glsl_type::float_type), - - _textureQueryLod(glsl_type::sampler2DArray_type, glsl_type::vec2_type), - _textureQueryLod(glsl_type::isampler2DArray_type, glsl_type::vec2_type), - _textureQueryLod(glsl_type::usampler2DArray_type, glsl_type::vec2_type), - - _textureQueryLod(glsl_type::samplerCubeArray_type, glsl_type::vec3_type), - _textureQueryLod(glsl_type::isamplerCubeArray_type, glsl_type::vec3_type), - _textureQueryLod(glsl_type::usamplerCubeArray_type, glsl_type::vec3_type), + _textureQueryLod(texture_query_lod, glsl_type::sampler1D_type, glsl_type::float_type), + _textureQueryLod(texture_query_lod, glsl_type::isampler1D_type, glsl_type::float_type), + _textureQueryLod(texture_query_lod, glsl_type::usampler1D_type, glsl_type::float_type), + + _textureQueryLod(texture_query_lod, glsl_type::sampler2D_type, glsl_type::vec2_type), + _textureQueryLod(texture_query_lod, glsl_type::isampler2D_type, glsl_type::vec2_type), + _textureQueryLod(texture_query_lod, glsl_type::usampler2D_type, glsl_type::vec2_type), + + _textureQueryLod(texture_query_lod, glsl_type::sampler3D_type, glsl_type::vec3_type), + _textureQueryLod(texture_query_lod, glsl_type::isampler3D_type, glsl_type::vec3_type), + _textureQueryLod(texture_query_lod, glsl_type::usampler3D_type, glsl_type::vec3_type), + + _textureQueryLod(texture_query_lod, glsl_type::samplerCube_type, glsl_type::vec3_type), + _textureQueryLod(texture_query_lod, glsl_type::isamplerCube_type, glsl_type::vec3_type), + _textureQueryLod(texture_query_lod, glsl_type::usamplerCube_type, glsl_type::vec3_type), + + _textureQueryLod(texture_query_lod, glsl_type::sampler1DArray_type, glsl_type::float_type), + _textureQueryLod(texture_query_lod, glsl_type::isampler1DArray_type, glsl_type::float_type), + _textureQueryLod(texture_query_lod, glsl_type::usampler1DArray_type, glsl_type::float_type), + + _textureQueryLod(texture_query_lod, glsl_type::sampler2DArray_type, glsl_type::vec2_type), + _textureQueryLod(texture_query_lod, glsl_type::isampler2DArray_type, glsl_type::vec2_type), + _textureQueryLod(texture_query_lod, glsl_type::usampler2DArray_type, glsl_type::vec2_type), + + _textureQueryLod(texture_query_lod, glsl_type::samplerCubeArray_type, glsl_type::vec3_type), + _textureQueryLod(texture_query_lod, glsl_type::isamplerCubeArray_type, glsl_type::vec3_type), + _textureQueryLod(texture_query_lod, glsl_type::usamplerCubeArray_type, glsl_type::vec3_type), + + _textureQueryLod(texture_query_lod, glsl_type::sampler1DShadow_type, glsl_type::float_type), + _textureQueryLod(texture_query_lod, glsl_type::sampler2DShadow_type, glsl_type::vec2_type), + _textureQueryLod(texture_query_lod, glsl_type::samplerCubeShadow_type, glsl_type::vec3_type), + _textureQueryLod(texture_query_lod, glsl_type::sampler1DArrayShadow_type, glsl_type::float_type), + _textureQueryLod(texture_query_lod, glsl_type::sampler2DArrayShadow_type, glsl_type::vec2_type), + _textureQueryLod(texture_query_lod, glsl_type::samplerCubeArrayShadow_type, glsl_type::vec3_type), + NULL); - _textureQueryLod(glsl_type::sampler1DShadow_type, glsl_type::float_type), - _textureQueryLod(glsl_type::sampler2DShadow_type, glsl_type::vec2_type), - _textureQueryLod(glsl_type::samplerCubeShadow_type, glsl_type::vec3_type), - _textureQueryLod(glsl_type::sampler1DArrayShadow_type, glsl_type::float_type), - _textureQueryLod(glsl_type::sampler2DArrayShadow_type, glsl_type::vec2_type), - _textureQueryLod(glsl_type::samplerCubeArrayShadow_type, glsl_type::vec3_type), + add_function("textureQueryLod", + _textureQueryLod(v400_fs_only, glsl_type::sampler1D_type, glsl_type::float_type), + _textureQueryLod(v400_fs_only, glsl_type::isampler1D_type, glsl_type::float_type), + _textureQueryLod(v400_fs_only, glsl_type::usampler1D_type, glsl_type::float_type), + + _textureQueryLod(v400_fs_only, glsl_type::sampler2D_type, glsl_type::vec2_type), + _textureQueryLod(v400_fs_only, glsl_type::isampler2D_type, glsl_type::vec2_type), + _textureQueryLod(v400_fs_only, glsl_type::usampler2D_type, glsl_type::vec2_type), + + _textureQueryLod(v400_fs_only, glsl_type::sampler3D_type, glsl_type::vec3_type), + _textureQueryLod(v400_fs_only, glsl_type::isampler3D_type, glsl_type::vec3_type), + _textureQueryLod(v400_fs_only, glsl_type::usampler3D_type, glsl_type::vec3_type), + + _textureQueryLod(v400_fs_only, glsl_type::samplerCube_type, glsl_type::vec3_type), + _textureQueryLod(v400_fs_only, glsl_type::isamplerCube_type, glsl_type::vec3_type), + _textureQueryLod(v400_fs_only, glsl_type::usamplerCube_type, glsl_type::vec3_type), + + _textureQueryLod(v400_fs_only, glsl_type::sampler1DArray_type, glsl_type::float_type), + _textureQueryLod(v400_fs_only, glsl_type::isampler1DArray_type, glsl_type::float_type), + _textureQueryLod(v400_fs_only, glsl_type::usampler1DArray_type, glsl_type::float_type), + + _textureQueryLod(v400_fs_only, glsl_type::sampler2DArray_type, glsl_type::vec2_type), + _textureQueryLod(v400_fs_only, glsl_type::isampler2DArray_type, glsl_type::vec2_type), + _textureQueryLod(v400_fs_only, glsl_type::usampler2DArray_type, glsl_type::vec2_type), + + _textureQueryLod(v400_fs_only, glsl_type::samplerCubeArray_type, glsl_type::vec3_type), + _textureQueryLod(v400_fs_only, glsl_type::isamplerCubeArray_type, glsl_type::vec3_type), + _textureQueryLod(v400_fs_only, glsl_type::usamplerCubeArray_type, glsl_type::vec3_type), + + _textureQueryLod(v400_fs_only, glsl_type::sampler1DShadow_type, glsl_type::float_type), + _textureQueryLod(v400_fs_only, glsl_type::sampler2DShadow_type, glsl_type::vec2_type), + _textureQueryLod(v400_fs_only, glsl_type::samplerCubeShadow_type, glsl_type::vec3_type), + _textureQueryLod(v400_fs_only, glsl_type::sampler1DArrayShadow_type, glsl_type::float_type), + _textureQueryLod(v400_fs_only, glsl_type::sampler2DArrayShadow_type, glsl_type::vec2_type), + _textureQueryLod(v400_fs_only, glsl_type::samplerCubeArrayShadow_type, glsl_type::vec3_type), NULL); add_function("textureQueryLevels", @@ -1981,6 +2231,16 @@ builtin_builder::create_builtins() NULL); + add_function("textureSamplesIdenticalEXT", + _textureSamplesIdentical(texture_samples_identical, glsl_type::sampler2DMS_type, glsl_type::ivec2_type), + _textureSamplesIdentical(texture_samples_identical, glsl_type::isampler2DMS_type, glsl_type::ivec2_type), + _textureSamplesIdentical(texture_samples_identical, glsl_type::usampler2DMS_type, glsl_type::ivec2_type), + + _textureSamplesIdentical(texture_samples_identical_array, glsl_type::sampler2DMSArray_type, glsl_type::ivec3_type), + _textureSamplesIdentical(texture_samples_identical_array, glsl_type::isampler2DMSArray_type, glsl_type::ivec3_type), + _textureSamplesIdentical(texture_samples_identical_array, glsl_type::usampler2DMSArray_type, glsl_type::ivec3_type), + NULL); + add_function("texture1D", _texture(ir_tex, v110, glsl_type::vec4_type, glsl_type::sampler1D_type, glsl_type::float_type), _texture(ir_txb, v110_fs_only, glsl_type::vec4_type, glsl_type::sampler1D_type, glsl_type::float_type), @@ -2202,61 +2462,69 @@ builtin_builder::create_builtins() NULL); add_function("textureGather", - _texture(ir_tg4, texture_gather, glsl_type::vec4_type, glsl_type::sampler2D_type, glsl_type::vec2_type), - _texture(ir_tg4, texture_gather, glsl_type::ivec4_type, glsl_type::isampler2D_type, glsl_type::vec2_type), - _texture(ir_tg4, texture_gather, glsl_type::uvec4_type, glsl_type::usampler2D_type, glsl_type::vec2_type), + _texture(ir_tg4, texture_gather_or_es31, glsl_type::vec4_type, glsl_type::sampler2D_type, glsl_type::vec2_type), + _texture(ir_tg4, texture_gather_or_es31, glsl_type::ivec4_type, glsl_type::isampler2D_type, glsl_type::vec2_type), + _texture(ir_tg4, texture_gather_or_es31, glsl_type::uvec4_type, glsl_type::usampler2D_type, glsl_type::vec2_type), _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::sampler2DRect_type, glsl_type::vec2_type), _texture(ir_tg4, gpu_shader5, glsl_type::ivec4_type, glsl_type::isampler2DRect_type, glsl_type::vec2_type), _texture(ir_tg4, gpu_shader5, glsl_type::uvec4_type, glsl_type::usampler2DRect_type, glsl_type::vec2_type), - _texture(ir_tg4, texture_gather, glsl_type::vec4_type, glsl_type::sampler2DArray_type, glsl_type::vec3_type), - _texture(ir_tg4, texture_gather, glsl_type::ivec4_type, glsl_type::isampler2DArray_type, glsl_type::vec3_type), - _texture(ir_tg4, texture_gather, glsl_type::uvec4_type, glsl_type::usampler2DArray_type, glsl_type::vec3_type), + _texture(ir_tg4, texture_gather_or_es31, glsl_type::vec4_type, glsl_type::sampler2DArray_type, glsl_type::vec3_type), + _texture(ir_tg4, texture_gather_or_es31, glsl_type::ivec4_type, glsl_type::isampler2DArray_type, glsl_type::vec3_type), + _texture(ir_tg4, texture_gather_or_es31, glsl_type::uvec4_type, glsl_type::usampler2DArray_type, glsl_type::vec3_type), - _texture(ir_tg4, texture_gather, glsl_type::vec4_type, glsl_type::samplerCube_type, glsl_type::vec3_type), - _texture(ir_tg4, texture_gather, glsl_type::ivec4_type, glsl_type::isamplerCube_type, glsl_type::vec3_type), - _texture(ir_tg4, texture_gather, glsl_type::uvec4_type, glsl_type::usamplerCube_type, glsl_type::vec3_type), + _texture(ir_tg4, texture_gather_or_es31, glsl_type::vec4_type, glsl_type::samplerCube_type, glsl_type::vec3_type), + _texture(ir_tg4, texture_gather_or_es31, glsl_type::ivec4_type, glsl_type::isamplerCube_type, glsl_type::vec3_type), + _texture(ir_tg4, texture_gather_or_es31, glsl_type::uvec4_type, glsl_type::usamplerCube_type, glsl_type::vec3_type), _texture(ir_tg4, texture_gather, glsl_type::vec4_type, glsl_type::samplerCubeArray_type, glsl_type::vec4_type), _texture(ir_tg4, texture_gather, glsl_type::ivec4_type, glsl_type::isamplerCubeArray_type, glsl_type::vec4_type), _texture(ir_tg4, texture_gather, glsl_type::uvec4_type, glsl_type::usamplerCubeArray_type, glsl_type::vec4_type), - _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::sampler2D_type, glsl_type::vec2_type, TEX_COMPONENT), - _texture(ir_tg4, gpu_shader5, glsl_type::ivec4_type, glsl_type::isampler2D_type, glsl_type::vec2_type, TEX_COMPONENT), - _texture(ir_tg4, gpu_shader5, glsl_type::uvec4_type, glsl_type::usampler2D_type, glsl_type::vec2_type, TEX_COMPONENT), + _texture(ir_tg4, gpu_shader5_or_es31, glsl_type::vec4_type, glsl_type::sampler2D_type, glsl_type::vec2_type, TEX_COMPONENT), + _texture(ir_tg4, gpu_shader5_or_es31, glsl_type::ivec4_type, glsl_type::isampler2D_type, glsl_type::vec2_type, TEX_COMPONENT), + _texture(ir_tg4, gpu_shader5_or_es31, glsl_type::uvec4_type, glsl_type::usampler2D_type, glsl_type::vec2_type, TEX_COMPONENT), _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::sampler2DRect_type, glsl_type::vec2_type, TEX_COMPONENT), _texture(ir_tg4, gpu_shader5, glsl_type::ivec4_type, glsl_type::isampler2DRect_type, glsl_type::vec2_type, TEX_COMPONENT), _texture(ir_tg4, gpu_shader5, glsl_type::uvec4_type, glsl_type::usampler2DRect_type, glsl_type::vec2_type, TEX_COMPONENT), - _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::sampler2DArray_type, glsl_type::vec3_type, TEX_COMPONENT), - _texture(ir_tg4, gpu_shader5, glsl_type::ivec4_type, glsl_type::isampler2DArray_type, glsl_type::vec3_type, TEX_COMPONENT), - _texture(ir_tg4, gpu_shader5, glsl_type::uvec4_type, glsl_type::usampler2DArray_type, glsl_type::vec3_type, TEX_COMPONENT), + _texture(ir_tg4, gpu_shader5_or_es31, glsl_type::vec4_type, glsl_type::sampler2DArray_type, glsl_type::vec3_type, TEX_COMPONENT), + _texture(ir_tg4, gpu_shader5_or_es31, glsl_type::ivec4_type, glsl_type::isampler2DArray_type, glsl_type::vec3_type, TEX_COMPONENT), + _texture(ir_tg4, gpu_shader5_or_es31, glsl_type::uvec4_type, glsl_type::usampler2DArray_type, glsl_type::vec3_type, TEX_COMPONENT), - _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::samplerCube_type, glsl_type::vec3_type, TEX_COMPONENT), - _texture(ir_tg4, gpu_shader5, glsl_type::ivec4_type, glsl_type::isamplerCube_type, glsl_type::vec3_type, TEX_COMPONENT), - _texture(ir_tg4, gpu_shader5, glsl_type::uvec4_type, glsl_type::usamplerCube_type, glsl_type::vec3_type, TEX_COMPONENT), + _texture(ir_tg4, gpu_shader5_or_es31, glsl_type::vec4_type, glsl_type::samplerCube_type, glsl_type::vec3_type, TEX_COMPONENT), + _texture(ir_tg4, gpu_shader5_or_es31, glsl_type::ivec4_type, glsl_type::isamplerCube_type, glsl_type::vec3_type, TEX_COMPONENT), + _texture(ir_tg4, gpu_shader5_or_es31, glsl_type::uvec4_type, glsl_type::usamplerCube_type, glsl_type::vec3_type, TEX_COMPONENT), _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::samplerCubeArray_type, glsl_type::vec4_type, TEX_COMPONENT), _texture(ir_tg4, gpu_shader5, glsl_type::ivec4_type, glsl_type::isamplerCubeArray_type, glsl_type::vec4_type, TEX_COMPONENT), _texture(ir_tg4, gpu_shader5, glsl_type::uvec4_type, glsl_type::usamplerCubeArray_type, glsl_type::vec4_type, TEX_COMPONENT), - _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::sampler2DShadow_type, glsl_type::vec2_type), - _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::sampler2DArrayShadow_type, glsl_type::vec3_type), - _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::samplerCubeShadow_type, glsl_type::vec3_type), + _texture(ir_tg4, gpu_shader5_or_es31, glsl_type::vec4_type, glsl_type::sampler2DShadow_type, glsl_type::vec2_type), + _texture(ir_tg4, gpu_shader5_or_es31, glsl_type::vec4_type, glsl_type::sampler2DArrayShadow_type, glsl_type::vec3_type), + _texture(ir_tg4, gpu_shader5_or_es31, glsl_type::vec4_type, glsl_type::samplerCubeShadow_type, glsl_type::vec3_type), _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::samplerCubeArrayShadow_type, glsl_type::vec4_type), _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::sampler2DRectShadow_type, glsl_type::vec2_type), NULL); add_function("textureGatherOffset", - _texture(ir_tg4, texture_gather_only, glsl_type::vec4_type, glsl_type::sampler2D_type, glsl_type::vec2_type, TEX_OFFSET), - _texture(ir_tg4, texture_gather_only, glsl_type::ivec4_type, glsl_type::isampler2D_type, glsl_type::vec2_type, TEX_OFFSET), - _texture(ir_tg4, texture_gather_only, glsl_type::uvec4_type, glsl_type::usampler2D_type, glsl_type::vec2_type, TEX_OFFSET), + _texture(ir_tg4, texture_gather_only_or_es31, glsl_type::vec4_type, glsl_type::sampler2D_type, glsl_type::vec2_type, TEX_OFFSET), + _texture(ir_tg4, texture_gather_only_or_es31, glsl_type::ivec4_type, glsl_type::isampler2D_type, glsl_type::vec2_type, TEX_OFFSET), + _texture(ir_tg4, texture_gather_only_or_es31, glsl_type::uvec4_type, glsl_type::usampler2D_type, glsl_type::vec2_type, TEX_OFFSET), + + _texture(ir_tg4, texture_gather_only_or_es31, glsl_type::vec4_type, glsl_type::sampler2DArray_type, glsl_type::vec3_type, TEX_OFFSET), + _texture(ir_tg4, texture_gather_only_or_es31, glsl_type::ivec4_type, glsl_type::isampler2DArray_type, glsl_type::vec3_type, TEX_OFFSET), + _texture(ir_tg4, texture_gather_only_or_es31, glsl_type::uvec4_type, glsl_type::usampler2DArray_type, glsl_type::vec3_type, TEX_OFFSET), + + _texture(ir_tg4, es31, glsl_type::vec4_type, glsl_type::sampler2D_type, glsl_type::vec2_type, TEX_OFFSET | TEX_COMPONENT), + _texture(ir_tg4, es31, glsl_type::ivec4_type, glsl_type::isampler2D_type, glsl_type::vec2_type, TEX_OFFSET | TEX_COMPONENT), + _texture(ir_tg4, es31, glsl_type::uvec4_type, glsl_type::usampler2D_type, glsl_type::vec2_type, TEX_OFFSET | TEX_COMPONENT), - _texture(ir_tg4, texture_gather_only, glsl_type::vec4_type, glsl_type::sampler2DArray_type, glsl_type::vec3_type, TEX_OFFSET), - _texture(ir_tg4, texture_gather_only, glsl_type::ivec4_type, glsl_type::isampler2DArray_type, glsl_type::vec3_type, TEX_OFFSET), - _texture(ir_tg4, texture_gather_only, glsl_type::uvec4_type, glsl_type::usampler2DArray_type, glsl_type::vec3_type, TEX_OFFSET), + _texture(ir_tg4, es31, glsl_type::vec4_type, glsl_type::sampler2DArray_type, glsl_type::vec3_type, TEX_OFFSET | TEX_COMPONENT), + _texture(ir_tg4, es31, glsl_type::ivec4_type, glsl_type::isampler2DArray_type, glsl_type::vec3_type, TEX_OFFSET | TEX_COMPONENT), + _texture(ir_tg4, es31, glsl_type::uvec4_type, glsl_type::usampler2DArray_type, glsl_type::vec3_type, TEX_OFFSET | TEX_COMPONENT), _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::sampler2D_type, glsl_type::vec2_type, TEX_OFFSET_NONCONST), _texture(ir_tg4, gpu_shader5, glsl_type::ivec4_type, glsl_type::isampler2D_type, glsl_type::vec2_type, TEX_OFFSET_NONCONST), @@ -2285,6 +2553,9 @@ builtin_builder::create_builtins() _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::sampler2DShadow_type, glsl_type::vec2_type, TEX_OFFSET_NONCONST), _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::sampler2DArrayShadow_type, glsl_type::vec3_type, TEX_OFFSET_NONCONST), _texture(ir_tg4, gpu_shader5, glsl_type::vec4_type, glsl_type::sampler2DRectShadow_type, glsl_type::vec2_type, TEX_OFFSET_NONCONST), + + _texture(ir_tg4, es31, glsl_type::vec4_type, glsl_type::sampler2DShadow_type, glsl_type::vec2_type, TEX_OFFSET), + _texture(ir_tg4, es31, glsl_type::vec4_type, glsl_type::sampler2DArrayShadow_type, glsl_type::vec3_type, TEX_OFFSET), NULL); add_function("textureGatherOffsets", @@ -2404,16 +2675,81 @@ builtin_builder::create_builtins() NULL); add_function("atomicCounter", - _atomic_op("__intrinsic_atomic_read", - shader_atomic_counters), + _atomic_counter_op("__intrinsic_atomic_read", + shader_atomic_counters), NULL); add_function("atomicCounterIncrement", - _atomic_op("__intrinsic_atomic_increment", - shader_atomic_counters), + _atomic_counter_op("__intrinsic_atomic_increment", + shader_atomic_counters), NULL); add_function("atomicCounterDecrement", - _atomic_op("__intrinsic_atomic_predecrement", - shader_atomic_counters), + _atomic_counter_op("__intrinsic_atomic_predecrement", + shader_atomic_counters), + NULL); + + add_function("atomicAdd", + _atomic_op2("__intrinsic_atomic_add", + buffer_atomics_supported, + glsl_type::uint_type), + _atomic_op2("__intrinsic_atomic_add", + buffer_atomics_supported, + glsl_type::int_type), + NULL); + add_function("atomicMin", + _atomic_op2("__intrinsic_atomic_min", + buffer_atomics_supported, + glsl_type::uint_type), + _atomic_op2("__intrinsic_atomic_min", + buffer_atomics_supported, + glsl_type::int_type), + NULL); + add_function("atomicMax", + _atomic_op2("__intrinsic_atomic_max", + buffer_atomics_supported, + glsl_type::uint_type), + _atomic_op2("__intrinsic_atomic_max", + buffer_atomics_supported, + glsl_type::int_type), + NULL); + add_function("atomicAnd", + _atomic_op2("__intrinsic_atomic_and", + buffer_atomics_supported, + glsl_type::uint_type), + _atomic_op2("__intrinsic_atomic_and", + buffer_atomics_supported, + glsl_type::int_type), + NULL); + add_function("atomicOr", + _atomic_op2("__intrinsic_atomic_or", + buffer_atomics_supported, + glsl_type::uint_type), + _atomic_op2("__intrinsic_atomic_or", + buffer_atomics_supported, + glsl_type::int_type), + NULL); + add_function("atomicXor", + _atomic_op2("__intrinsic_atomic_xor", + buffer_atomics_supported, + glsl_type::uint_type), + _atomic_op2("__intrinsic_atomic_xor", + buffer_atomics_supported, + glsl_type::int_type), + NULL); + add_function("atomicExchange", + _atomic_op2("__intrinsic_atomic_exchange", + buffer_atomics_supported, + glsl_type::uint_type), + _atomic_op2("__intrinsic_atomic_exchange", + buffer_atomics_supported, + glsl_type::int_type), + NULL); + add_function("atomicCompSwap", + _atomic_op3("__intrinsic_atomic_comp_swap", + buffer_atomics_supported, + glsl_type::uint_type), + _atomic_op3("__intrinsic_atomic_comp_swap", + buffer_atomics_supported, + glsl_type::int_type), NULL); add_function("min3", @@ -2470,7 +2806,33 @@ builtin_builder::create_builtins() add_image_functions(true); add_function("memoryBarrier", - _memory_barrier(shader_image_load_store), + _memory_barrier("__intrinsic_memory_barrier", + shader_image_load_store), + NULL); + add_function("groupMemoryBarrier", + _memory_barrier("__intrinsic_group_memory_barrier", + compute_shader), + NULL); + add_function("memoryBarrierAtomicCounter", + _memory_barrier("__intrinsic_memory_barrier_atomic_counter", + compute_shader), + NULL); + add_function("memoryBarrierBuffer", + _memory_barrier("__intrinsic_memory_barrier_buffer", + compute_shader), + NULL); + add_function("memoryBarrierImage", + _memory_barrier("__intrinsic_memory_barrier_image", + compute_shader), + NULL); + add_function("memoryBarrierShared", + _memory_barrier("__intrinsic_memory_barrier_shared", + compute_shader), + NULL); + + add_function("clock2x32ARB", + _shader_clock(shader_clock, + glsl_type::uvec2_type), NULL); #undef F @@ -2509,6 +2871,7 @@ builtin_builder::add_function(const char *name, ...) void builtin_builder::add_image_function(const char *name, const char *intrinsic_name, + image_prototype_ctr prototype, unsigned num_arguments, unsigned flags) { @@ -2547,12 +2910,15 @@ builtin_builder::add_image_function(const char *name, glsl_type::uimage2DMS_type, glsl_type::uimage2DMSArray_type }; + ir_function *f = new(mem_ctx) ir_function(name); for (unsigned i = 0; i < ARRAY_SIZE(types); ++i) { - if (types[i]->sampler_type != GLSL_TYPE_FLOAT || - (flags & IMAGE_FUNCTION_SUPPORTS_FLOAT_DATA_TYPE)) - f->add_signature(_image(types[i], intrinsic_name, + if ((types[i]->sampler_type != GLSL_TYPE_FLOAT || + (flags & IMAGE_FUNCTION_SUPPORTS_FLOAT_DATA_TYPE)) && + (types[i]->sampler_dimensionality == GLSL_SAMPLER_DIM_MS || + !(flags & IMAGE_FUNCTION_MS_ONLY))) + f->add_signature(_image(prototype, types[i], intrinsic_name, num_arguments, flags)); } @@ -2565,43 +2931,66 @@ builtin_builder::add_image_functions(bool glsl) const unsigned flags = (glsl ? IMAGE_FUNCTION_EMIT_STUB : 0); add_image_function(glsl ? "imageLoad" : "__intrinsic_image_load", - "__intrinsic_image_load", 0, - (flags | IMAGE_FUNCTION_HAS_VECTOR_DATA_TYPE | + "__intrinsic_image_load", + &builtin_builder::_image_prototype, 0, + (flags | IMAGE_FUNCTION_HAS_VECTOR_DATA_TYPE | IMAGE_FUNCTION_SUPPORTS_FLOAT_DATA_TYPE | IMAGE_FUNCTION_READ_ONLY)); add_image_function(glsl ? "imageStore" : "__intrinsic_image_store", - "__intrinsic_image_store", 1, + "__intrinsic_image_store", + &builtin_builder::_image_prototype, 1, (flags | IMAGE_FUNCTION_RETURNS_VOID | IMAGE_FUNCTION_HAS_VECTOR_DATA_TYPE | IMAGE_FUNCTION_SUPPORTS_FLOAT_DATA_TYPE | IMAGE_FUNCTION_WRITE_ONLY)); + const unsigned atom_flags = flags | IMAGE_FUNCTION_AVAIL_ATOMIC; + add_image_function(glsl ? "imageAtomicAdd" : "__intrinsic_image_atomic_add", - "__intrinsic_image_atomic_add", 1, flags); + "__intrinsic_image_atomic_add", + &builtin_builder::_image_prototype, 1, atom_flags); add_image_function(glsl ? "imageAtomicMin" : "__intrinsic_image_atomic_min", - "__intrinsic_image_atomic_min", 1, flags); + "__intrinsic_image_atomic_min", + &builtin_builder::_image_prototype, 1, atom_flags); add_image_function(glsl ? "imageAtomicMax" : "__intrinsic_image_atomic_max", - "__intrinsic_image_atomic_max", 1, flags); + "__intrinsic_image_atomic_max", + &builtin_builder::_image_prototype, 1, atom_flags); add_image_function(glsl ? "imageAtomicAnd" : "__intrinsic_image_atomic_and", - "__intrinsic_image_atomic_and", 1, flags); + "__intrinsic_image_atomic_and", + &builtin_builder::_image_prototype, 1, atom_flags); add_image_function(glsl ? "imageAtomicOr" : "__intrinsic_image_atomic_or", - "__intrinsic_image_atomic_or", 1, flags); + "__intrinsic_image_atomic_or", + &builtin_builder::_image_prototype, 1, atom_flags); add_image_function(glsl ? "imageAtomicXor" : "__intrinsic_image_atomic_xor", - "__intrinsic_image_atomic_xor", 1, flags); + "__intrinsic_image_atomic_xor", + &builtin_builder::_image_prototype, 1, atom_flags); add_image_function((glsl ? "imageAtomicExchange" : "__intrinsic_image_atomic_exchange"), - "__intrinsic_image_atomic_exchange", 1, flags); + "__intrinsic_image_atomic_exchange", + &builtin_builder::_image_prototype, 1, atom_flags); add_image_function((glsl ? "imageAtomicCompSwap" : "__intrinsic_image_atomic_comp_swap"), - "__intrinsic_image_atomic_comp_swap", 2, flags); + "__intrinsic_image_atomic_comp_swap", + &builtin_builder::_image_prototype, 2, atom_flags); + + add_image_function(glsl ? "imageSize" : "__intrinsic_image_size", + "__intrinsic_image_size", + &builtin_builder::_image_size_prototype, 1, + flags | IMAGE_FUNCTION_SUPPORTS_FLOAT_DATA_TYPE); + + add_image_function(glsl ? "imageSamples" : "__intrinsic_image_samples", + "__intrinsic_image_samples", + &builtin_builder::_image_samples_prototype, 1, + flags | IMAGE_FUNCTION_SUPPORTS_FLOAT_DATA_TYPE | + IMAGE_FUNCTION_MS_ONLY); } ir_variable * @@ -2731,8 +3120,8 @@ builtin_builder::_##NAME(builtin_available_predicate avail, const glsl_type *typ } ir_function_signature * -builtin_builder::binop(ir_expression_operation opcode, - builtin_available_predicate avail, +builtin_builder::binop(builtin_available_predicate avail, + ir_expression_operation opcode, const glsl_type *return_type, const glsl_type *param0_type, const glsl_type *param1_type) @@ -3028,7 +3417,7 @@ builtin_builder::_atanh(const glsl_type *type) ir_function_signature * builtin_builder::_pow(const glsl_type *type) { - return binop(ir_binop_pow, always_available, type, type, type); + return binop(always_available, ir_binop_pow, type, type, type); } UNOP(exp, ir_unop_exp, always_available) @@ -3052,7 +3441,7 @@ UNOPA(fract, ir_unop_fract) ir_function_signature * builtin_builder::_mod(const glsl_type *x_type, const glsl_type *y_type) { - return binop(ir_binop_mod, always_available, x_type, x_type, y_type); + return binop(always_available, ir_binop_mod, x_type, x_type, y_type); } ir_function_signature * @@ -3074,14 +3463,14 @@ ir_function_signature * builtin_builder::_min(builtin_available_predicate avail, const glsl_type *x_type, const glsl_type *y_type) { - return binop(ir_binop_min, avail, x_type, x_type, y_type); + return binop(avail, ir_binop_min, x_type, x_type, y_type); } ir_function_signature * builtin_builder::_max(builtin_available_predicate avail, const glsl_type *x_type, const glsl_type *y_type) { - return binop(ir_binop_max, avail, x_type, x_type, y_type); + return binop(avail, ir_binop_max, x_type, x_type, y_type); } ir_function_signature * @@ -3215,7 +3604,16 @@ builtin_builder::_isinf(builtin_available_predicate avail, const glsl_type *type ir_constant_data infinities; for (int i = 0; i < type->vector_elements; i++) { - infinities.f[i] = std::numeric_limits::infinity(); + switch (type->base_type) { + case GLSL_TYPE_FLOAT: + infinities.f[i] = INFINITY; + break; + case GLSL_TYPE_DOUBLE: + infinities.d[i] = INFINITY; + break; + default: + unreachable("unknown type"); + } } body.emit(ret(equal(abs(x), imm(type, infinities)))); @@ -3401,9 +3799,9 @@ ir_function_signature * builtin_builder::_dot(builtin_available_predicate avail, const glsl_type *type) { if (type->vector_elements == 1) - return binop(ir_binop_mul, avail, type, type, type); + return binop(avail, ir_binop_mul, type, type, type); - return binop(ir_binop_dot, avail, + return binop(avail, ir_binop_dot, type->get_base_type(), type, type); } @@ -3919,7 +4317,7 @@ ir_function_signature * builtin_builder::_lessThan(builtin_available_predicate avail, const glsl_type *type) { - return binop(ir_binop_less, avail, + return binop(avail, ir_binop_less, glsl_type::bvec(type->vector_elements), type, type); } @@ -3927,7 +4325,7 @@ ir_function_signature * builtin_builder::_lessThanEqual(builtin_available_predicate avail, const glsl_type *type) { - return binop(ir_binop_lequal, avail, + return binop(avail, ir_binop_lequal, glsl_type::bvec(type->vector_elements), type, type); } @@ -3935,7 +4333,7 @@ ir_function_signature * builtin_builder::_greaterThan(builtin_available_predicate avail, const glsl_type *type) { - return binop(ir_binop_greater, avail, + return binop(avail, ir_binop_greater, glsl_type::bvec(type->vector_elements), type, type); } @@ -3943,7 +4341,7 @@ ir_function_signature * builtin_builder::_greaterThanEqual(builtin_available_predicate avail, const glsl_type *type) { - return binop(ir_binop_gequal, avail, + return binop(avail, ir_binop_gequal, glsl_type::bvec(type->vector_elements), type, type); } @@ -3951,7 +4349,7 @@ ir_function_signature * builtin_builder::_equal(builtin_available_predicate avail, const glsl_type *type) { - return binop(ir_binop_equal, avail, + return binop(avail, ir_binop_equal, glsl_type::bvec(type->vector_elements), type, type); } @@ -3959,7 +4357,7 @@ ir_function_signature * builtin_builder::_notEqual(builtin_available_predicate avail, const glsl_type *type) { - return binop(ir_binop_nequal, avail, + return binop(avail, ir_binop_nequal, glsl_type::bvec(type->vector_elements), type, type); } @@ -4035,6 +4433,19 @@ builtin_builder::_textureSize(builtin_available_predicate avail, return sig; } +ir_function_signature * +builtin_builder::_textureSamples(const glsl_type *sampler_type) +{ + ir_variable *s = in_var(sampler_type, "sampler"); + MAKE_SIG(glsl_type::int_type, shader_samples, 1, s); + + ir_texture *tex = new(mem_ctx) ir_texture(ir_texture_samples); + tex->set_sampler(new(mem_ctx) ir_dereference_variable(s), glsl_type::int_type); + body.emit(ret(tex)); + + return sig; +} + ir_function_signature * builtin_builder::_texture(ir_texture_opcode opcode, builtin_available_predicate avail, @@ -4262,13 +4673,23 @@ builtin_builder::_EndStreamPrimitive(builtin_available_predicate avail, } ir_function_signature * -builtin_builder::_textureQueryLod(const glsl_type *sampler_type, +builtin_builder::_barrier() +{ + MAKE_SIG(glsl_type::void_type, barrier_supported, 0); + + body.emit(new(mem_ctx) ir_barrier()); + return sig; +} + +ir_function_signature * +builtin_builder::_textureQueryLod(builtin_available_predicate avail, + const glsl_type *sampler_type, const glsl_type *coord_type) { ir_variable *s = in_var(sampler_type, "sampler"); ir_variable *coord = in_var(coord_type, "coord"); /* The sampler and coordinate always exist; add optional parameters later. */ - MAKE_SIG(glsl_type::vec2_type, texture_query_lod, 2, s, coord); + MAKE_SIG(glsl_type::vec2_type, avail, 2, s, coord); ir_texture *tex = new(mem_ctx) ir_texture(ir_lod); tex->coordinate = var_ref(coord); @@ -4294,6 +4715,25 @@ builtin_builder::_textureQueryLevels(const glsl_type *sampler_type) return sig; } +ir_function_signature * +builtin_builder::_textureSamplesIdentical(builtin_available_predicate avail, + const glsl_type *sampler_type, + const glsl_type *coord_type) +{ + ir_variable *s = in_var(sampler_type, "sampler"); + ir_variable *P = in_var(coord_type, "P"); + const glsl_type *return_type = glsl_type::bool_type; + MAKE_SIG(return_type, avail, 2, s, P); + + ir_texture *tex = new(mem_ctx) ir_texture(ir_samples_identical); + tex->coordinate = var_ref(P); + tex->set_sampler(var_ref(s), return_type); + + body.emit(ret(tex)); + + return sig; +} + UNOP(dFdx, ir_unop_dFdx, fs_oes_derivatives) UNOP(dFdxCoarse, ir_unop_dFdx_coarse, fs_derivative_control) UNOP(dFdxFine, ir_unop_dFdx_fine, fs_derivative_control) @@ -4445,7 +4885,7 @@ builtin_builder::_bitfieldExtract(const glsl_type *type) ir_variable *value = in_var(type, "value"); ir_variable *offset = in_var(glsl_type::int_type, "offset"); ir_variable *bits = in_var(glsl_type::int_type, "bits"); - MAKE_SIG(type, gpu_shader5, 3, value, offset, bits); + MAKE_SIG(type, gpu_shader5_or_es31, 3, value, offset, bits); body.emit(ret(expr(ir_triop_bitfield_extract, value, offset, bits))); @@ -4459,33 +4899,33 @@ builtin_builder::_bitfieldInsert(const glsl_type *type) ir_variable *insert = in_var(type, "insert"); ir_variable *offset = in_var(glsl_type::int_type, "offset"); ir_variable *bits = in_var(glsl_type::int_type, "bits"); - MAKE_SIG(type, gpu_shader5, 4, base, insert, offset, bits); + MAKE_SIG(type, gpu_shader5_or_es31, 4, base, insert, offset, bits); body.emit(ret(bitfield_insert(base, insert, offset, bits))); return sig; } -UNOP(bitfieldReverse, ir_unop_bitfield_reverse, gpu_shader5) +UNOP(bitfieldReverse, ir_unop_bitfield_reverse, gpu_shader5_or_es31) ir_function_signature * builtin_builder::_bitCount(const glsl_type *type) { - return unop(gpu_shader5, ir_unop_bit_count, + return unop(gpu_shader5_or_es31, ir_unop_bit_count, glsl_type::ivec(type->vector_elements), type); } ir_function_signature * builtin_builder::_findLSB(const glsl_type *type) { - return unop(gpu_shader5, ir_unop_find_lsb, + return unop(gpu_shader5_or_es31, ir_unop_find_lsb, glsl_type::ivec(type->vector_elements), type); } ir_function_signature * builtin_builder::_findMSB(const glsl_type *type) { - return unop(gpu_shader5, ir_unop_find_msb, + return unop(gpu_shader5_or_es31, ir_unop_find_msb, glsl_type::ivec(type->vector_elements), type); } @@ -4505,7 +4945,8 @@ builtin_builder::_fma(builtin_available_predicate avail, const glsl_type *type) ir_function_signature * builtin_builder::_ldexp(const glsl_type *x_type, const glsl_type *exp_type) { - return binop(ir_binop_ldexp, x_type->base_type == GLSL_TYPE_DOUBLE ? fp64 : gpu_shader5, x_type, x_type, exp_type); + return binop(x_type->base_type == GLSL_TYPE_DOUBLE ? fp64 : gpu_shader5_or_es31, + ir_binop_ldexp, x_type, x_type, exp_type); } ir_function_signature * @@ -4526,7 +4967,7 @@ builtin_builder::_frexp(const glsl_type *x_type, const glsl_type *exp_type) { ir_variable *x = in_var(x_type, "x"); ir_variable *exponent = out_var(exp_type, "exp"); - MAKE_SIG(x_type, gpu_shader5, 2, x, exponent); + MAKE_SIG(x_type, gpu_shader5_or_es31, 2, x, exponent); const unsigned vec_elem = x_type->vector_elements; const glsl_type *bvec = glsl_type::get_instance(GLSL_TYPE_BOOL, vec_elem, 1); @@ -4575,7 +5016,7 @@ builtin_builder::_uaddCarry(const glsl_type *type) ir_variable *x = in_var(type, "x"); ir_variable *y = in_var(type, "y"); ir_variable *carry = out_var(type, "carry"); - MAKE_SIG(type, gpu_shader5, 3, x, y, carry); + MAKE_SIG(type, gpu_shader5_or_es31, 3, x, y, carry); body.emit(assign(carry, ir_builder::carry(x, y))); body.emit(ret(add(x, y))); @@ -4589,7 +5030,7 @@ builtin_builder::_usubBorrow(const glsl_type *type) ir_variable *x = in_var(type, "x"); ir_variable *y = in_var(type, "y"); ir_variable *borrow = out_var(type, "borrow"); - MAKE_SIG(type, gpu_shader5, 3, x, y, borrow); + MAKE_SIG(type, gpu_shader5_or_es31, 3, x, y, borrow); body.emit(assign(borrow, ir_builder::borrow(x, y))); body.emit(ret(sub(x, y))); @@ -4607,7 +5048,7 @@ builtin_builder::_mulExtended(const glsl_type *type) ir_variable *y = in_var(type, "y"); ir_variable *msb = out_var(type, "msb"); ir_variable *lsb = out_var(type, "lsb"); - MAKE_SIG(glsl_type::void_type, gpu_shader5, 4, x, y, msb, lsb); + MAKE_SIG(glsl_type::void_type, gpu_shader5_or_es31, 4, x, y, msb, lsb); body.emit(assign(msb, imul_high(x, y))); body.emit(assign(lsb, mul(x, y))); @@ -4654,7 +5095,7 @@ builtin_builder::_interpolateAtSample(const glsl_type *type) } ir_function_signature * -builtin_builder::_atomic_intrinsic(builtin_available_predicate avail) +builtin_builder::_atomic_counter_intrinsic(builtin_available_predicate avail) { ir_variable *counter = in_var(glsl_type::atomic_uint_type, "counter"); MAKE_INTRINSIC(glsl_type::uint_type, avail, 1, counter); @@ -4662,8 +5103,29 @@ builtin_builder::_atomic_intrinsic(builtin_available_predicate avail) } ir_function_signature * -builtin_builder::_atomic_op(const char *intrinsic, - builtin_available_predicate avail) +builtin_builder::_atomic_intrinsic2(builtin_available_predicate avail, + const glsl_type *type) +{ + ir_variable *atomic = in_var(type, "atomic"); + ir_variable *data = in_var(type, "data"); + MAKE_INTRINSIC(type, avail, 2, atomic, data); + return sig; +} + +ir_function_signature * +builtin_builder::_atomic_intrinsic3(builtin_available_predicate avail, + const glsl_type *type) +{ + ir_variable *atomic = in_var(type, "atomic"); + ir_variable *data1 = in_var(type, "data1"); + ir_variable *data2 = in_var(type, "data2"); + MAKE_INTRINSIC(type, avail, 3, atomic, data1, data2); + return sig; +} + +ir_function_signature * +builtin_builder::_atomic_counter_op(const char *intrinsic, + builtin_available_predicate avail) { ir_variable *counter = in_var(glsl_type::atomic_uint_type, "atomic_counter"); MAKE_SIG(glsl_type::uint_type, avail, 1, counter); @@ -4675,6 +5137,39 @@ builtin_builder::_atomic_op(const char *intrinsic, return sig; } +ir_function_signature * +builtin_builder::_atomic_op2(const char *intrinsic, + builtin_available_predicate avail, + const glsl_type *type) +{ + ir_variable *atomic = in_var(type, "atomic_var"); + ir_variable *data = in_var(type, "atomic_data"); + MAKE_SIG(type, avail, 2, atomic, data); + + ir_variable *retval = body.make_temp(type, "atomic_retval"); + body.emit(call(shader->symbols->get_function(intrinsic), retval, + sig->parameters)); + body.emit(ret(retval)); + return sig; +} + +ir_function_signature * +builtin_builder::_atomic_op3(const char *intrinsic, + builtin_available_predicate avail, + const glsl_type *type) +{ + ir_variable *atomic = in_var(type, "atomic_var"); + ir_variable *data1 = in_var(type, "atomic_data1"); + ir_variable *data2 = in_var(type, "atomic_data2"); + MAKE_SIG(type, avail, 3, atomic, data1, data2); + + ir_variable *retval = body.make_temp(type, "atomic_retval"); + body.emit(call(shader->symbols->get_function(intrinsic), retval, + sig->parameters)); + body.emit(ret(retval)); + return sig; +} + ir_function_signature * builtin_builder::_min3(const glsl_type *type) { @@ -4719,7 +5214,6 @@ builtin_builder::_mid3(const glsl_type *type) ir_function_signature * builtin_builder::_image_prototype(const glsl_type *image_type, - const char *intrinsic_name, unsigned num_arguments, unsigned flags) { @@ -4735,8 +5229,10 @@ builtin_builder::_image_prototype(const glsl_type *image_type, ir_variable *coord = in_var( glsl_type::ivec(image_type->coordinate_components()), "coord"); - ir_function_signature *sig = new_sig( - ret_type, shader_image_load_store, 2, image, coord); + const builtin_available_predicate avail = + (flags & IMAGE_FUNCTION_AVAIL_ATOMIC ? shader_image_atomic : + shader_image_load_store); + ir_function_signature *sig = new_sig(ret_type, avail, 2, image, coord); /* Sample index for multisample images. */ if (image_type->sampler_dimensionality == GLSL_SAMPLER_DIM_MS) @@ -4766,13 +5262,79 @@ builtin_builder::_image_prototype(const glsl_type *image_type, } ir_function_signature * -builtin_builder::_image(const glsl_type *image_type, +builtin_builder::_image_size_prototype(const glsl_type *image_type, + unsigned /* num_arguments */, + unsigned /* flags */) +{ + const glsl_type *ret_type; + unsigned num_components = image_type->coordinate_components(); + + /* From the ARB_shader_image_size extension: + * "Cube images return the dimensions of one face." + */ + if (image_type->sampler_dimensionality == GLSL_SAMPLER_DIM_CUBE && + !image_type->sampler_array) { + num_components = 2; + } + + /* FIXME: Add the highp precision qualifier for GLES 3.10 when it is + * supported by mesa. + */ + ret_type = glsl_type::get_instance(GLSL_TYPE_INT, num_components, 1); + + ir_variable *image = in_var(image_type, "image"); + ir_function_signature *sig = new_sig(ret_type, shader_image_size, 1, image); + + /* Set the maximal set of qualifiers allowed for this image + * built-in. Function calls with arguments having fewer + * qualifiers than present in the prototype are allowed by the + * spec, but not with more, i.e. this will make the compiler + * accept everything that needs to be accepted, and reject cases + * like loads from write-only or stores to read-only images. + */ + image->data.image_read_only = true; + image->data.image_write_only = true; + image->data.image_coherent = true; + image->data.image_volatile = true; + image->data.image_restrict = true; + + return sig; +} + +ir_function_signature * +builtin_builder::_image_samples_prototype(const glsl_type *image_type, + unsigned /* num_arguments */, + unsigned /* flags */) +{ + ir_variable *image = in_var(image_type, "image"); + ir_function_signature *sig = + new_sig(glsl_type::int_type, shader_samples, 1, image); + + /* Set the maximal set of qualifiers allowed for this image + * built-in. Function calls with arguments having fewer + * qualifiers than present in the prototype are allowed by the + * spec, but not with more, i.e. this will make the compiler + * accept everything that needs to be accepted, and reject cases + * like loads from write-only or stores to read-only images. + */ + image->data.image_read_only = true; + image->data.image_write_only = true; + image->data.image_coherent = true; + image->data.image_volatile = true; + image->data.image_restrict = true; + + return sig; +} + +ir_function_signature * +builtin_builder::_image(image_prototype_ctr prototype, + const glsl_type *image_type, const char *intrinsic_name, unsigned num_arguments, unsigned flags) { - ir_function_signature *sig = _image_prototype(image_type, intrinsic_name, - num_arguments, flags); + ir_function_signature *sig = (this->*prototype)(image_type, + num_arguments, flags); if (flags & IMAGE_FUNCTION_EMIT_STUB) { ir_factory body(&sig->body, mem_ctx); @@ -4804,14 +5366,37 @@ builtin_builder::_memory_barrier_intrinsic(builtin_available_predicate avail) } ir_function_signature * -builtin_builder::_memory_barrier(builtin_available_predicate avail) +builtin_builder::_memory_barrier(const char *intrinsic_name, + builtin_available_predicate avail) { MAKE_SIG(glsl_type::void_type, avail, 0); - body.emit(call(shader->symbols->get_function("__intrinsic_memory_barrier"), + body.emit(call(shader->symbols->get_function(intrinsic_name), NULL, sig->parameters)); return sig; } +ir_function_signature * +builtin_builder::_shader_clock_intrinsic(builtin_available_predicate avail, + const glsl_type *type) +{ + MAKE_INTRINSIC(type, avail, 0); + return sig; +} + +ir_function_signature * +builtin_builder::_shader_clock(builtin_available_predicate avail, + const glsl_type *type) +{ + MAKE_SIG(type, avail, 0); + + ir_variable *retval = body.make_temp(type, "clock_retval"); + + body.emit(call(shader->symbols->get_function("__intrinsic_shader_clock"), + retval, sig->parameters)); + body.emit(ret(retval)); + return sig; +} + /** @} */ /******************************************************************************/ @@ -4852,8 +5437,7 @@ _mesa_glsl_find_builtin_function(_mesa_glsl_parse_state *state, } ir_function * -_mesa_glsl_find_builtin_function_by_name(_mesa_glsl_parse_state *state, - const char *name) +_mesa_glsl_find_builtin_function_by_name(const char *name) { ir_function *f; mtx_lock(&builtins_lock); @@ -4868,4 +5452,32 @@ _mesa_glsl_get_builtin_function_shader() return builtins.shader; } + +/** + * Get the function signature for main from a shader + */ +ir_function_signature * +_mesa_get_main_function_signature(gl_shader *sh) +{ + ir_function *const f = sh->symbols->get_function("main"); + if (f != NULL) { + exec_list void_parameters; + + /* Look for the 'void main()' signature and ensure that it's defined. + * This keeps the linker from accidentally pick a shader that just + * contains a prototype for main. + * + * We don't have to check for multiple definitions of main (in multiple + * shaders) because that would have already been caught above. + */ + ir_function_signature *sig = + f->matching_signature(NULL, &void_parameters, false); + if ((sig != NULL) && sig->is_defined) { + return sig; + } + } + + return NULL; +} + /** @} */