- Wine-Commits - list.winehq.org

Francisco Casas : tests: Test for common type conversion for element-wise intrinsics.
by Alexandre Julliard Jan. 11, 2023

Jan. 11, 2023

Module: vkd3d Branch: master Commit: 86c35fc79da7c5206ed095c47ab7e3479a3a95e9 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/86c35fc79da7c5206ed095c47ab7e… Author: Francisco Casas <fcasas(a)codeweavers.com> Date: Mon Nov 7 11:54:11 2022 -0300 tests: Test for common type conversion for element-wise intrinsics. Some intrinsics have different rules for the allowed data types than expressions: - Vectors and matrices at the same time are not allowed, regardless of their dimensions. Even if they have the same number of components. - Any combination of matrices is always allowed, even those when no matrix fits inside another, e.g.: float2x3 is compatible with float3x2, resulting in float 2x2. The common data type is the min on each dimension. This is the case for max, pow, ldexp, clamp and smoothstep; which suggest that it is the case for all intrinsics where the operation is applied element-wise. Tests for mul() are also added as a counter-example where the operation is not element-wise. --- tests/hlsl-clamp.shader_test | 28 ++++++++++++++++++++++++++++ tests/hlsl-ldexp.shader_test | 26 ++++++++++++++++++++++++++ tests/hlsl-lerp.shader_test | 28 ++++++++++++++++++++++++++++ tests/hlsl-mul.shader_test | 30 ++++++++++++++++++++++++++++++ tests/max.shader_test | 27 +++++++++++++++++++++++++++ tests/pow.shader_test | 26 ++++++++++++++++++++++++++ 6 files changed, 165 insertions(+) diff --git a/tests/hlsl-clamp.shader_test b/tests/hlsl-clamp.shader_test index 8e26270c..cc198735 100644 --- a/tests/hlsl-clamp.shader_test +++ b/tests/hlsl-clamp.shader_test @@ -8,3 +8,31 @@ float4 main(uniform float3 u) : sv_target uniform 0 float4 -0.3 -0.1 0.7 0.0 draw quad probe all rgba (-0.1, 0.7, -0.3, 0.3) + + +[pixel shader todo] +float4 main() : sv_target +{ + float3x2 a = {6, 5, 4, 3, 2, 1}; + float2x3 b = {1, 2, 3, 4.2, 5.2, 6.2}; + float3x4 c = 5.5; + + float2x2 r = clamp(a, b, c); + return float4(r); +} + +[test] +todo draw quad +todo probe all rgba (5.5, 5.0, 4.2, 5.2) + + +[pixel shader fail todo] +float4 main() : sv_target +{ + float2x2 a = {3.1, 3.1, 3.1, 3.1}; + float2x2 b = {1, 2, 3, 4}; + float4 c = {5.5, 4.5, 3.5, 2.5}; + + clamp(a, b, c); + return 0; +} diff --git a/tests/hlsl-ldexp.shader_test b/tests/hlsl-ldexp.shader_test index 0873fc9e..bea97953 100644 --- a/tests/hlsl-ldexp.shader_test +++ b/tests/hlsl-ldexp.shader_test @@ -30,3 +30,29 @@ uniform 0 int4 2 3 4 5 uniform 4 int4 0 -10 10 100 draw quad probe all rgba (2.0, 0.00292968750, 4096.0, 6.33825300e+030) + + +[pixel shader todo] +float4 main() : sv_target +{ + float2x3 a = {1, 2, 3, 4, 5, 6}; + float3x2 b = {6, 5, 4, 3, 2, 1}; + + float2x2 r = ldexp(a, b); + return float4(r); +} + +[test] +todo draw quad +todo probe all rgba (64.0, 64.0, 64.0, 40.0) + + +[pixel shader fail todo] +float4 main() : sv_target +{ + float2x2 a = {1, 2, 3, 4}; + float1 b = {2}; + + ldexp(a, b); + return 0; +} diff --git a/tests/hlsl-lerp.shader_test b/tests/hlsl-lerp.shader_test index 3f93b02d..3cd10ec1 100644 --- a/tests/hlsl-lerp.shader_test +++ b/tests/hlsl-lerp.shader_test @@ -34,3 +34,31 @@ uniform 4 int4 0 -10 10 1000000 uniform 8 int4 0 1 -1 1000000 draw quad probe all rgba (2.0, -10.0, -2.0, 1e12) + + +[pixel shader todo] +float4 main() : sv_target +{ + float3x2 a = {6, 5, 4, 3, 2, 1}; + float2x3 b = {1, 2, 3, 4.2, 5.2, 6.2}; + float3x4 c = 2.4; + + float2x2 r = lerp(a, b, c); + return float4(r); +} + +[test] +todo draw quad +todo probe all rgba (-6.0, -2.2, 4.48, 8.28) + + +[pixel shader fail todo] +float4 main() : sv_target +{ + float2x2 a = {0, 1, 2, 3}; + float2x2 b = {1, 2, 3, 4}; + float4 c = {0.5, 0.5, 0.5, 0.5}; + + lerp(a, b, c); + return 0; +} diff --git a/tests/hlsl-mul.shader_test b/tests/hlsl-mul.shader_test index 7b453187..cb104a9e 100644 --- a/tests/hlsl-mul.shader_test +++ b/tests/hlsl-mul.shader_test @@ -288,3 +288,33 @@ float4 main(float4 pos : sv_position) : sv_target [test] draw quad probe all rgba (78.0, 96.0, 114.0, 0.0) + + +[pixel shader] +float4 main() : sv_target +{ + float2x3 a = float2x3(1, 2, 3, 4, 5, 6); + float3x2 b = float3x2(6, 5, 4, 3, 2, 1); + + float2x2 r = mul(a, b); + return float4(r); +} + +[test] +draw quad +probe all rgba (20.0, 14.0, 56.0, 41.0) + + +[pixel shader] +float4 main() : sv_target +{ + float2x2 a = float2x2(1, 2, 3, 4); + float2 b = float2(1, 2); + + float2 r = mul(a, b); + return float4(r, 0, 0); +} + +[test] +draw quad +probe all rgba (5.0, 11.0, 0.0, 0.0) diff --git a/tests/max.shader_test b/tests/max.shader_test index 50083f33..7a917ec5 100644 --- a/tests/max.shader_test +++ b/tests/max.shader_test @@ -9,6 +9,7 @@ uniform 0 float4 0.7 -0.1 0.0 0.0 draw quad probe all rgba (0.7, 2.1, 2.0, -1.0) + [pixel shader] float4 main(uniform float4 u) : sv_target { @@ -21,3 +22,29 @@ float4 main(uniform float4 u) : sv_target uniform 0 float4 0.7 -0.1 0.4 0.8 draw quad probe all rgba (0.7, 0.8, 0.7, 0.2) + + +[pixel shader todo] +float4 main() : sv_target +{ + float2x3 a = {1, 2, 3, 4, 5, 6}; + float3x2 b = {6, 5, 4, 3, 2, 1}; + + float2x2 r = max(a, b); + return float4(r); +} + +[test] +todo draw quad +todo probe all rgba (6.0, 5.0, 4.0, 5.0) + + +[pixel shader fail todo] +float4 main() : sv_target +{ + float2x2 a = {1, 2, 3, 4}; + float4 b = {4, 3, 2, 1}; + + max(a, b); + return 0; +} diff --git a/tests/pow.shader_test b/tests/pow.shader_test index 6470494e..6f2b2741 100644 --- a/tests/pow.shader_test +++ b/tests/pow.shader_test @@ -8,3 +8,29 @@ float4 main(uniform float4 u) : sv_target uniform 0 float4 0.4 0.8 2.5 2.0 draw quad probe all rgba (0.512, 0.101192884, 0.64, 0.25) 4 + + +[pixel shader todo] +float4 main() : sv_target +{ + float2x3 a = {1, 2, 3, 4, 5, 6}; + float3x2 b = {6, 5, 4, 3, 2, 1}; + + float2x2 r = pow(a, b); + return float4(r); +} + +[test] +todo draw quad +todo probe all rgba (1.0, 32.0, 256.0, 125.0) + + +[pixel shader fail todo] +float4 main() : sv_target +{ + float2x2 a = {1, 2, 3, 4}; + float4 b = {1, 2, 3, 4}; + + pow(a, b); + return 0; +}

1 0

Francisco Casas : vkd3d-shader/hlsl: Support smoothstep() intrinsic.
by Alexandre Julliard Jan. 11, 2023

Jan. 11, 2023

Module: vkd3d Branch: master Commit: a7bb5a0835d4008553aa3f405fe989270fb0a5c1 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/a7bb5a0835d4008553aa3f405fe98… Author: Francisco Casas <fcasas(a)codeweavers.com> Date: Thu Nov 3 20:54:48 2022 -0300 vkd3d-shader/hlsl: Support smoothstep() intrinsic. --- Makefile.am | 1 + libs/vkd3d-shader/hlsl.y | 77 ++++++++++++++++++ tests/hlsl-smoothstep.shader_test | 166 ++++++++++++++++++++++++++++++++++++++ 3 files changed, 244 insertions(+) diff --git a/Makefile.am b/Makefile.am index ff817f0a..9459bad9 100644 --- a/Makefile.am +++ b/Makefile.am @@ -102,6 +102,7 @@ vkd3d_shader_tests = \ tests/hlsl-return-void.shader_test \ tests/hlsl-shape.shader_test \ tests/hlsl-single-numeric-initializer.shader_test \ + tests/hlsl-smoothstep.shader_test \ tests/hlsl-state-block-syntax.shader_test \ tests/hlsl-static-initializer.shader_test \ tests/hlsl-storage-qualifiers.shader_test \ diff --git a/libs/vkd3d-shader/hlsl.y b/libs/vkd3d-shader/hlsl.y index a18394a6..82a9711b 100644 --- a/libs/vkd3d-shader/hlsl.y +++ b/libs/vkd3d-shader/hlsl.y @@ -2596,6 +2596,82 @@ static bool intrinsic_saturate(struct hlsl_ctx *ctx, return !!add_unary_arithmetic_expr(ctx, params->instrs, HLSL_OP1_SAT, arg, loc); } +/* smoothstep(a, b, x) = p^2 (3 - 2p), where p = saturate((x - a)/(b - a)) */ +static bool intrinsic_smoothstep(struct hlsl_ctx *ctx, + const struct parse_initializer *params, const struct vkd3d_shader_location *loc) +{ + struct hlsl_ir_node *min_arg, *max_arg, *x_arg, *p, *p_num, *p_denom, *res; + struct hlsl_ir_constant *one, *minus_two, *three; + enum hlsl_type_class common_class; + struct hlsl_type *common_type; + unsigned int dimx, dimy; + + min_arg = params->args[0]; + max_arg = params->args[1]; + x_arg = params->args[2]; + + if (!expr_common_shape(ctx, min_arg->data_type, max_arg->data_type, loc, &common_class, &dimx, &dimy)) + return false; + common_type = hlsl_get_numeric_type(ctx, common_class, HLSL_TYPE_FLOAT, dimx, dimy); + + if (!expr_common_shape(ctx, common_type, x_arg->data_type, loc, &common_class, &dimx, &dimy)) + return false; + common_type = hlsl_get_numeric_type(ctx, common_class, HLSL_TYPE_FLOAT, dimx, dimy); + + if (!(min_arg = add_implicit_conversion(ctx, params->instrs, min_arg, common_type, loc))) + return false; + + if (!(max_arg = add_implicit_conversion(ctx, params->instrs, max_arg, common_type, loc))) + return false; + + if (!(x_arg = add_implicit_conversion(ctx, params->instrs, x_arg, common_type, loc))) + return false; + + if (!(min_arg = add_unary_arithmetic_expr(ctx, params->instrs, HLSL_OP1_NEG, min_arg, loc))) + return false; + + if (!(p_num = add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_ADD, x_arg, min_arg, loc))) + return false; + + if (!(p_denom = add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_ADD, max_arg, min_arg, loc))) + return false; + + if (!(one = hlsl_new_float_constant(ctx, 1.0, loc))) + return false; + list_add_tail(params->instrs, &one->node.entry); + + if (!(p_denom = add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_DIV, &one->node, p_denom, loc))) + return false; + + if (!(p = add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_MUL, p_num, p_denom, loc))) + return false; + + if (!(p = add_unary_arithmetic_expr(ctx, params->instrs, HLSL_OP1_SAT, p, loc))) + return false; + + if (!(minus_two = hlsl_new_float_constant(ctx, -2.0, loc))) + return false; + list_add_tail(params->instrs, &minus_two->node.entry); + + if (!(three = hlsl_new_float_constant(ctx, 3.0, loc))) + return false; + list_add_tail(params->instrs, &three->node.entry); + + if (!(res = add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_MUL, &minus_two->node, p, loc))) + return false; + + if (!(res = add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_ADD, &three->node, res, loc))) + return false; + + if (!(p = add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_MUL, p, p, loc))) + return false; + + if (!(res = add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_MUL, p, res, loc))) + return false; + + return true; +} + static bool intrinsic_transpose(struct hlsl_ctx *ctx, const struct parse_initializer *params, const struct vkd3d_shader_location *loc) { @@ -2681,6 +2757,7 @@ intrinsic_functions[] = {"pow", 2, true, intrinsic_pow}, {"round", 1, true, intrinsic_round}, {"saturate", 1, true, intrinsic_saturate}, + {"smoothstep", 3, true, intrinsic_smoothstep}, {"transpose", 1, true, intrinsic_transpose}, }; diff --git a/tests/hlsl-smoothstep.shader_test b/tests/hlsl-smoothstep.shader_test new file mode 100644 index 00000000..fc1c856a --- /dev/null +++ b/tests/hlsl-smoothstep.shader_test @@ -0,0 +1,166 @@ + + +[pixel shader] +float4 main() : sv_target +{ + float4 a = {1, -1, -1, 10}; + float4 b = {2, 1, 1, 20}; + float4 x = {0.3, 0.4, 2, 15.4}; + + return smoothstep(a, b, x); +} + +[test] +draw quad +probe all rgba (0, 0.784, 1.0, 0.559872) 1 + + +[pixel shader] +float4 main() : sv_target +{ + float a = 1; + float b = 2; + float4 x = {0.9, 1.2, 1.8, 2.1}; + + return smoothstep(a, b, x); +} + +[test] +draw quad +probe all rgba (0, 0.104, 0.896, 1.000000) 5 + + +[pixel shader] +float4 main() : sv_target +{ + float4 a = {1, 10, 100, 1000}; + float4 b = {2, 20, 200, 2000}; + float x = 14; + + return smoothstep(a, b, x); +} + +[test] +draw quad +probe all rgba (1.0, 0.352, 0, 0) 1 + + +[pixel shader] +float4 main() : sv_target +{ + float2 a = {1, 10}; + float3 b = {2, 20, 200}; + float4 x = {1.4, 14, 140, 1400}; + + float2 res = smoothstep(a, b, x); + return float4(res, 0, 0); +} + +[test] +draw quad +probe all rgba (0.352, 0.352, 0, 0) 1 + + +[pixel shader] +float4 main() : sv_target +{ + float3 a = {1, 10, 100}; + float2 b = {2, 20}; + float4 x = {1.4, 14, 140, 1400}; + + float2 res = smoothstep(a, b, x); + return float4(res, 0, 0); +} + +[test] +draw quad +probe all rgba (0.352, 0.352, 0, 0) 1 + + +[pixel shader] +float4 main() : sv_target +{ + float4 a = {1, 10, 100, 1000}; + float4 b = {2, 20, 200, 2000}; + float2 x = {14, 140}; + + float2 res = smoothstep(a, b, x); + return float4(res, 0, 0); +} + +[test] +draw quad +probe all rgba (1.0, 1.0, 0, 0) 1 + + +[pixel shader todo] +float4 main() : sv_target +{ + float2x3 a = {1, 1, 1, 1, 1, 1}; + float3x2 b = {2, 2, 2, 2, 2, 2}; + float4x2 x = {1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8}; + + float2x2 r = smoothstep(a, b, x); + return r; +} + +[test] +todo draw quad +todo probe all rgba (0.028, 0.104, 0.216, 0.352) 1 + + +[pixel shader] +// 4 division by zero warnings. +// Only test compilation because result is implementation-dependent. +float4 main() : sv_target +{ + float4 a = {0, 0, 0, 0}; + float4 b = {-1, -1, 0, 0}; + float4 x = {0, -0.25, 0, 1}; + + return smoothstep(a, b, x); +} + + +[pixel shader] +float4 main() : sv_target +{ + float4x1 a = {0.0, 0.0, 0.0, 0.0}; + float b = 1.0; + float3x1 x = {0.5, 0.5, 0.5}; + + float3x1 r = smoothstep(a, b, x); + return float4(r, 0); +} + +[test] +draw quad +probe all rgba (0.5, 0.5, 0.5, 0.0) + + +[pixel shader todo] +float4 main() : sv_target +{ + float4x1 a = {0.0, 0.0, 0.0, 0.0}; + float2x2 b = {1.0, 1.0, 1.0, 1.0}; + float3x1 x = {0.5, 0.5, 0.5}; + + float2x1 r = smoothstep(a, b, x); + return float4(r, r); +} + +[test] +todo draw quad +todo probe all rgba (0.5, 0.5, 0.5, 0.5) + + +[pixel shader fail todo] +float4 main() : sv_target +{ + float2x2 a = {0.0, 0.0, 0.0, 0.0}; + float4 b = 1.0; + float2x2 x = {0.5, 0.5, 0.5, 0.5}; + + smoothstep(a, b, x); + return 0; +}

1 0

Francisco Casas : vkd3d-shader/hlsl: Support transpose() intrinsic.
by Alexandre Julliard Jan. 11, 2023

Jan. 11, 2023

Module: vkd3d Branch: master Commit: 09e7218539a3c636bcb5282c573b7295f5432f8b URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/09e7218539a3c636bcb5282c573b7… Author: Francisco Casas <fcasas(a)codeweavers.com> Date: Mon Mar 28 12:55:14 2022 -0300 vkd3d-shader/hlsl: Support transpose() intrinsic. --- Makefile.am | 1 + libs/vkd3d-shader/hlsl.y | 59 +++++++++++++++++++++++++++++++ tests/hlsl-transpose.shader_test | 75 ++++++++++++++++++++++++++++++++++++++++ 3 files changed, 135 insertions(+) diff --git a/Makefile.am b/Makefile.am index 8c05338b..ff817f0a 100644 --- a/Makefile.am +++ b/Makefile.am @@ -108,6 +108,7 @@ vkd3d_shader_tests = \ tests/hlsl-struct-array.shader_test \ tests/hlsl-struct-assignment.shader_test \ tests/hlsl-struct-semantics.shader_test \ + tests/hlsl-transpose.shader_test \ tests/hlsl-vector-indexing.shader_test \ tests/hlsl-vector-indexing-uniform.shader_test \ tests/logic-operations.shader_test \ diff --git a/libs/vkd3d-shader/hlsl.y b/libs/vkd3d-shader/hlsl.y index eedc85bd..a18394a6 100644 --- a/libs/vkd3d-shader/hlsl.y +++ b/libs/vkd3d-shader/hlsl.y @@ -2596,6 +2596,64 @@ static bool intrinsic_saturate(struct hlsl_ctx *ctx, return !!add_unary_arithmetic_expr(ctx, params->instrs, HLSL_OP1_SAT, arg, loc); } +static bool intrinsic_transpose(struct hlsl_ctx *ctx, + const struct parse_initializer *params, const struct vkd3d_shader_location *loc) +{ + struct hlsl_ir_node *arg = params->args[0]; + struct hlsl_type *arg_type = arg->data_type; + struct hlsl_deref var_deref; + struct hlsl_type *mat_type; + struct hlsl_ir_load *load; + struct hlsl_ir_var *var; + unsigned int i, j; + + if (arg_type->type != HLSL_CLASS_SCALAR && arg_type->type != HLSL_CLASS_MATRIX) + { + struct vkd3d_string_buffer *string; + + if ((string = hlsl_type_to_string(ctx, arg_type))) + hlsl_error(ctx, &arg->loc, VKD3D_SHADER_ERROR_HLSL_INVALID_TYPE, + "Wrong type for argument 1 of transpose(): expected a matrix or scalar type, but got '%s'.\n", + string->buffer); + hlsl_release_string_buffer(ctx, string); + return false; + } + + if (arg_type->type == HLSL_CLASS_SCALAR) + { + list_add_tail(params->instrs, &arg->entry); + return true; + } + + mat_type = hlsl_get_matrix_type(ctx, arg_type->base_type, arg_type->dimy, arg_type->dimx); + + if (!(var = hlsl_new_synthetic_var(ctx, "transpose", mat_type, loc))) + return false; + hlsl_init_simple_deref_from_var(&var_deref, var); + + for (i = 0; i < arg_type->dimx; ++i) + { + for (j = 0; j < arg_type->dimy; ++j) + { + struct hlsl_ir_store *store; + struct hlsl_block block; + + if (!(load = add_load_component(ctx, params->instrs, arg, j * arg->data_type->dimx + i, loc))) + return false; + + if (!(store = hlsl_new_store_component(ctx, &block, &var_deref, i * var->data_type->dimx + j, &load->node))) + return false; + list_move_tail(params->instrs, &block.instrs); + } + } + + if (!(load = hlsl_new_var_load(ctx, var, *loc))) + return false; + list_add_tail(params->instrs, &load->node.entry); + + return true; +} + static const struct intrinsic_function { const char *name; @@ -2623,6 +2681,7 @@ intrinsic_functions[] = {"pow", 2, true, intrinsic_pow}, {"round", 1, true, intrinsic_round}, {"saturate", 1, true, intrinsic_saturate}, + {"transpose", 1, true, intrinsic_transpose}, }; static int intrinsic_function_name_compare(const void *a, const void *b) diff --git a/tests/hlsl-transpose.shader_test b/tests/hlsl-transpose.shader_test new file mode 100644 index 00000000..83852fa1 --- /dev/null +++ b/tests/hlsl-transpose.shader_test @@ -0,0 +1,75 @@ +[pixel shader] +float4 main() : sv_target +{ + return transpose(5); +} + +[test] +draw quad +probe all rgba (5.0, 5.0, 5.0, 5.0) + + +[pixel shader] +float4 main() : sv_target +{ + float1x1 x = 5; + + return transpose(x); +} + +[test] +draw quad +probe all rgba (5.0, 5.0, 5.0, 5.0) + + +[pixel shader fail] +float4 main() : sv_target +{ + float4 x = float4(1, 2, 3, 4); + + return transpose(x); +} + +[pixel shader] +float4 main() : sv_target +{ + float1x4 x = float1x4(1.0, 2.0, 3.0, 4.0); + + return transpose(x); +} + +[test] +draw quad +probe all rgba (1.0, 2.0, 3.0, 4.0) + + +[pixel shader] +float4 main() : sv_target +{ + float4x3 m = float4x3(1.0, 2.0, 3.0, + 4.0, 5.0, 6.0, + 7.0, 8.0, 9.0, + 10.0, 11.0, 12.0); + + return transpose(m)[1]; +} + +[test] +draw quad +probe all rgba (2.0, 5.0, 8.0, 11.0) + + +[pixel shader] +float4 main() : sv_target +{ + row_major float4x3 m = float4x3(1.0, 2.0, 3.0, + 4.0, 5.0, 6.0, + 7.0, 8.0, 9.0, + 10.0, 11.0, 12.0); + + return transpose(m)[1]; +} + +[test] +draw quad +probe all rgba (2.0, 5.0, 8.0, 11.0)

1 0

Francisco Casas : vkd3d-shader/hlsl: Convert elementwise intrinsics args to the proper common type.
by Alexandre Julliard Jan. 11, 2023

Jan. 11, 2023

Module: vkd3d Branch: master Commit: 9ceed76a9c7f4b7bba8822b33974276922a50104 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/9ceed76a9c7f4b7bba8822b339742… Author: Francisco Casas <fcasas(a)codeweavers.com> Date: Mon Nov 7 19:05:52 2022 -0300 vkd3d-shader/hlsl: Convert elementwise intrinsics args to the proper common type. --- libs/vkd3d-shader/hlsl.y | 111 ++++++++++++++++++++++++++++++++++---- tests/hlsl-clamp.shader_test | 8 +-- tests/hlsl-ldexp.shader_test | 8 +-- tests/hlsl-lerp.shader_test | 8 +-- tests/hlsl-smoothstep.shader_test | 14 ++--- tests/max.shader_test | 8 +-- tests/pow.shader_test | 2 +- 7 files changed, 126 insertions(+), 33 deletions(-)

1 0

Francisco Casas : vkd3d-shader/hlsl: Introduce elementwise_intrinsic_float_convert_args().
by Alexandre Julliard Jan. 11, 2023

Jan. 11, 2023

Module: vkd3d Branch: master Commit: 6770ecbdf42b602c1f593f66c880b093691564c8 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/6770ecbdf42b602c1f593f66c880b… Author: Francisco Casas <fcasas(a)codeweavers.com> Date: Tue Dec 13 20:13:56 2022 -0300 vkd3d-shader/hlsl: Introduce elementwise_intrinsic_float_convert_args(). --- libs/vkd3d-shader/hlsl.y | 56 ++++++++++++++++++++---------------------------- 1 file changed, 23 insertions(+), 33 deletions(-) diff --git a/libs/vkd3d-shader/hlsl.y b/libs/vkd3d-shader/hlsl.y index c0ec779f..cbc59bc5 100644 --- a/libs/vkd3d-shader/hlsl.y +++ b/libs/vkd3d-shader/hlsl.y @@ -2303,6 +2303,19 @@ static bool elementwise_intrinsic_convert_args(struct hlsl_ctx *ctx, return convert_args(ctx, params, common_type, loc); } +static bool elementwise_intrinsic_float_convert_args(struct hlsl_ctx *ctx, + const struct parse_initializer *params, const struct vkd3d_shader_location *loc) +{ + struct hlsl_type *type; + + if (!(type = elementwise_intrinsic_get_common_type(ctx, params, loc))) + return false; + + type = hlsl_get_numeric_type(ctx, type->type, HLSL_TYPE_FLOAT, type->dimx, type->dimy); + + return convert_args(ctx, params, type, loc); +} + static bool intrinsic_abs(struct hlsl_ctx *ctx, const struct parse_initializer *params, const struct vkd3d_shader_location *loc) { @@ -2442,13 +2455,10 @@ static bool intrinsic_ldexp(struct hlsl_ctx *ctx, { struct hlsl_ir_node *arg; - if (!elementwise_intrinsic_convert_args(ctx, params, loc)) + if (!elementwise_intrinsic_float_convert_args(ctx, params, loc)) return false; - if (!(arg = intrinsic_float_convert_arg(ctx, params, params->args[1], loc))) - return false; - - if (!(arg = add_unary_arithmetic_expr(ctx, params->instrs, HLSL_OP1_EXP2, arg, loc))) + if (!(arg = add_unary_arithmetic_expr(ctx, params->instrs, HLSL_OP1_EXP2, params->args[1], loc))) return false; return !!add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_MUL, params->args[0], arg, loc); @@ -2482,15 +2492,12 @@ static bool intrinsic_length(struct hlsl_ctx *ctx, static bool intrinsic_lerp(struct hlsl_ctx *ctx, const struct parse_initializer *params, const struct vkd3d_shader_location *loc) { - struct hlsl_ir_node *arg, *neg, *add, *mul; - - if (!elementwise_intrinsic_convert_args(ctx, params, loc)) - return false; + struct hlsl_ir_node *neg, *add, *mul; - if (!(arg = intrinsic_float_convert_arg(ctx, params, params->args[0], loc))) + if (!elementwise_intrinsic_float_convert_args(ctx, params, loc)) return false; - if (!(neg = add_unary_arithmetic_expr(ctx, params->instrs, HLSL_OP1_NEG, arg, loc))) + if (!(neg = add_unary_arithmetic_expr(ctx, params->instrs, HLSL_OP1_NEG, params->args[0], loc))) return false; if (!(add = add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_ADD, params->args[1], neg, loc))) @@ -2499,7 +2506,7 @@ static bool intrinsic_lerp(struct hlsl_ctx *ctx, if (!(mul = add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_MUL, params->args[2], add, loc))) return false; - return !!add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_ADD, arg, mul, loc); + return !!add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_ADD, params->args[0], mul, loc); } static bool intrinsic_max(struct hlsl_ctx *ctx, @@ -2649,15 +2656,12 @@ static bool intrinsic_normalize(struct hlsl_ctx *ctx, static bool intrinsic_pow(struct hlsl_ctx *ctx, const struct parse_initializer *params, const struct vkd3d_shader_location *loc) { - struct hlsl_ir_node *log, *exp, *arg, *mul; - - if (!elementwise_intrinsic_convert_args(ctx, params, loc)) - return false; + struct hlsl_ir_node *log, *exp, *mul; - if (!(arg = intrinsic_float_convert_arg(ctx, params, params->args[0], loc))) + if (!elementwise_intrinsic_float_convert_args(ctx, params, loc)) return false; - if (!(log = hlsl_new_unary_expr(ctx, HLSL_OP1_LOG2, arg, *loc))) + if (!(log = hlsl_new_unary_expr(ctx, HLSL_OP1_LOG2, params->args[0], *loc))) return false; list_add_tail(params->instrs, &log->entry); @@ -2698,28 +2702,14 @@ static bool intrinsic_smoothstep(struct hlsl_ctx *ctx, { struct hlsl_ir_node *min_arg, *max_arg, *x_arg, *p, *p_num, *p_denom, *res; struct hlsl_ir_constant *one, *minus_two, *three; - struct hlsl_type *common_type; - if (!elementwise_intrinsic_convert_args(ctx, params, loc)) + if (!elementwise_intrinsic_float_convert_args(ctx, params, loc)) return false; min_arg = params->args[0]; max_arg = params->args[1]; x_arg = params->args[2]; - common_type = x_arg->data_type; - common_type = hlsl_get_numeric_type(ctx, common_type->type, HLSL_TYPE_FLOAT, common_type->dimx, - common_type->dimy); - - if (!(min_arg = add_implicit_conversion(ctx, params->instrs, min_arg, common_type, loc))) - return false; - - if (!(max_arg = add_implicit_conversion(ctx, params->instrs, max_arg, common_type, loc))) - return false; - - if (!(x_arg = add_implicit_conversion(ctx, params->instrs, x_arg, common_type, loc))) - return false; - if (!(min_arg = add_unary_arithmetic_expr(ctx, params->instrs, HLSL_OP1_NEG, min_arg, loc))) return false;

1 0

Jacek Caban : include: Add stdarg.h header.
by Alexandre Julliard Jan. 10, 2023

Jan. 10, 2023

Module: wine Branch: master Commit: 4d8091cccc46053c605e68eccd27b94ec297f28e URL: https://gitlab.winehq.org/wine/wine/-/commit/4d8091cccc46053c605e68eccd27b9… Author: Jacek Caban <jacek(a)codeweavers.com> Date: Thu Dec 8 17:52:23 2022 +0100 include: Add stdarg.h header. Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=54263 --- include/Makefile.in | 1 + include/msvcrt/stdarg.h | 31 +++++++++++++++++++++++++++++++ 2 files changed, 32 insertions(+) diff --git a/include/Makefile.in b/include/Makefile.in index c24ba7eff07..54cbf4d955c 100644 --- a/include/Makefile.in +++ b/include/Makefile.in @@ -512,6 +512,7 @@ SOURCES = \ msvcrt/setjmp.h \ msvcrt/share.h \ msvcrt/signal.h \ + msvcrt/stdarg.h \ msvcrt/stdbool.h \ msvcrt/stddef.h \ msvcrt/stdint.h \ diff --git a/include/msvcrt/stdarg.h b/include/msvcrt/stdarg.h new file mode 100644 index 00000000000..6c6b6048351 --- /dev/null +++ b/include/msvcrt/stdarg.h @@ -0,0 +1,31 @@ +/* + * Variable argument definitions + * + * Copyright 2022 Jacek Caban + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA + */ + +#ifndef _INC_STDARG +#define _INC_STDARG + +#include <vadefs.h> + +#define va_start(v,l) _crt_va_start(v,l) +#define va_arg(v,l) _crt_va_arg(v,l) +#define va_end(v) _crt_va_end(v) +#define va_copy(d,s) _crt_va_copy(d,s) + +#endif /* _INC_STDARG */

1 0

Alexandre Julliard : wow64: Initialize win32u syscall table when it is mapped.
by Alexandre Julliard Jan. 10, 2023

Jan. 10, 2023

Module: wine Branch: master Commit: f2a07805b9b027da9a1caaf6023abc90bc950459 URL: https://gitlab.winehq.org/wine/wine/-/commit/f2a07805b9b027da9a1caaf6023abc… Author: Alexandre Julliard <julliard(a)winehq.org> Date: Tue Jan 10 15:35:53 2023 +0100 wow64: Initialize win32u syscall table when it is mapped. --- dlls/wow64/syscall.c | 75 +++++++++++++--------------------------------------- 1 file changed, 18 insertions(+), 57 deletions(-) diff --git a/dlls/wow64/syscall.c b/dlls/wow64/syscall.c index 8be5a1c300c..5ba9cd14954 100644 --- a/dlls/wow64/syscall.c +++ b/dlls/wow64/syscall.c @@ -88,6 +88,7 @@ SYSTEM_DLL_INIT_BLOCK *pLdrSystemDllInitBlock = NULL; /* wow64win syscall table */ static const SYSTEM_SERVICE_TABLE *psdwhwin32; +static HMODULE win32u_module; /* cpu backend dll functions */ static void * (WINAPI *pBTCpuGetBopCode)(void); @@ -392,17 +393,6 @@ static DWORD get_syscall_num( const BYTE *syscall ) } -/********************************************************************** - * init_image_mapping - */ -void init_image_mapping( HMODULE module ) -{ - void **ptr = RtlFindExportedRoutineByName( module, "Wow64Transition" ); - - if (ptr) *ptr = pBTCpuGetBopCode(); -} - - /********************************************************************** * init_syscall_table */ @@ -467,6 +457,23 @@ static void init_syscall_table( HMODULE module, ULONG idx, const SYSTEM_SERVICE_ } +/********************************************************************** + * init_image_mapping + */ +void init_image_mapping( HMODULE module ) +{ + void **ptr = RtlFindExportedRoutineByName( module, "Wow64Transition" ); + + if (!ptr) return; + *ptr = pBTCpuGetBopCode(); + if (!win32u_module && RtlFindExportedRoutineByName( module, "NtUserInitializeClientPfnArrays" )) + { + win32u_module = module; + init_syscall_table( win32u_module, 1, psdwhwin32 ); + } +} + + /********************************************************************** * load_64bit_module */ @@ -489,48 +496,6 @@ static HMODULE load_64bit_module( const WCHAR *name ) } -/********************************************************************** - * load_32bit_module - */ -static HMODULE load_32bit_module( const WCHAR *name ) -{ - NTSTATUS status; - UNICODE_STRING str; - WCHAR path[MAX_PATH]; - HANDLE file, mapping; - LARGE_INTEGER size; - IO_STATUS_BLOCK io; - OBJECT_ATTRIBUTES attr; - SIZE_T len = 0; - void *module = NULL; - const WCHAR *dir = get_machine_wow64_dir( current_machine ); - - swprintf( path, MAX_PATH, L"%s\\%s", dir, name ); - RtlInitUnicodeString( &str, path ); - InitializeObjectAttributes( &attr, &str, OBJ_CASE_INSENSITIVE, 0, NULL ); - if ((status = NtOpenFile( &file, GENERIC_READ | SYNCHRONIZE, &attr, &io, - FILE_SHARE_READ | FILE_SHARE_DELETE, - FILE_SYNCHRONOUS_IO_NONALERT | FILE_NON_DIRECTORY_FILE ))) goto failed; - - size.QuadPart = 0; - status = NtCreateSection( &mapping, STANDARD_RIGHTS_REQUIRED | SECTION_QUERY | - SECTION_MAP_READ | SECTION_MAP_EXECUTE, - NULL, &size, PAGE_EXECUTE_READ, SEC_IMAGE, file ); - NtClose( file ); - if (status) goto failed; - - status = NtMapViewOfSection( mapping, GetCurrentProcess(), &module, 0, 0, NULL, &len, - ViewShare, 0, PAGE_EXECUTE_READ ); - NtClose( mapping ); - if (!status) return module; - -failed: - ERR( "failed to load dll %lx\n", status ); - NtTerminateProcess( GetCurrentProcess(), status ); - return NULL; -} - - /********************************************************************** * get_cpu_dll_name */ @@ -589,10 +554,6 @@ static DWORD WINAPI process_init( RTL_RUN_ONCE *once, void *param, void **contex *(void **)RtlFindExportedRoutineByName( module, "__wine_syscall_dispatcher" ) = pBTCpuGetBopCode(); *(void **)RtlFindExportedRoutineByName( module, "__wine_unix_call_dispatcher" ) = p__wine_get_unix_opcode(); - module = load_32bit_module( L"win32u.dll" ); - init_syscall_table( module, 1, psdwhwin32 ); - NtUnmapViewOfSection( GetCurrentProcess(), module ); - init_file_redirects(); return TRUE;

1 0

Jacek Caban : include: Add vadefs.h header.
by Alexandre Julliard Jan. 10, 2023

Jan. 10, 2023

Module: wine Branch: master Commit: 4c7e7ccac9f2b33d8685fd830e9468a1599f33fd URL: https://gitlab.winehq.org/wine/wine/-/commit/4c7e7ccac9f2b33d8685fd830e9468… Author: Jacek Caban <jacek(a)codeweavers.com> Date: Thu Dec 8 17:49:24 2022 +0100 include: Add vadefs.h header. Based on mingw-w64. --- include/Makefile.in | 1 + include/msvcrt/vadefs.h | 52 +++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 53 insertions(+) diff --git a/include/Makefile.in b/include/Makefile.in index 23a4265a437..c24ba7eff07 100644 --- a/include/Makefile.in +++ b/include/Makefile.in @@ -527,6 +527,7 @@ SOURCES = \ msvcrt/time.h \ msvcrt/uchar.h \ msvcrt/unistd.h \ + msvcrt/vadefs.h \ msvcrt/wchar.h \ msvcrt/wctype.h \ mswsock.h \ diff --git a/include/msvcrt/vadefs.h b/include/msvcrt/vadefs.h new file mode 100644 index 00000000000..2a612fc055c --- /dev/null +++ b/include/msvcrt/vadefs.h @@ -0,0 +1,52 @@ +/* + * Variable argument definitions + * + * Copyright 2022 Jacek Caban + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA + */ + +#ifndef _INC_VADEFS +#define _INC_VADEFS + +#ifdef __cplusplus +#define _ADDRESSOF(v) (&reinterpret_cast<const char &>(v)) +#else +#define _ADDRESSOF(v) (&(v)) +#endif + +#if defined(__GNUC__) || defined(__clang__) + +typedef __builtin_va_list va_list; +#define _crt_va_start(v,l) __builtin_va_start(v,l) +#define _crt_va_arg(v,l) __builtin_va_arg(v,l) +#define _crt_va_end(v) __builtin_va_end(v) +#define _crt_va_copy(d,s) __builtin_va_copy(d,s) + +#else + +typedef char *va_list; + +#if defined(__i386__) +#define _INTSIZEOF(n) ((sizeof(n) + sizeof(int) - 1) & ~(sizeof(int) - 1)) +#define _crt_va_start(v,l) ((v) = (va_list)_ADDRESSOF(l) + _INTSIZEOF(l)) +#define _crt_va_arg(v,l) (*(l *)(((v) += _INTSIZEOF(l)) - _INTSIZEOF(l))) +#define _crt_va_end(v) ((v) = (va_list)0) +#define _crt_va_copy(d,s) ((d) = (s)) +#endif + +#endif + +#endif /* _INC_VADEFS */

1 0

Paul Gofman : dwmapi: Fill qpcVBlank in DwmGetCompositionTimingInfo().
by Alexandre Julliard Jan. 10, 2023

Jan. 10, 2023

Module: wine Branch: master Commit: 62fa96df16a0285cce1694ce47eef946f75bfd42 URL: https://gitlab.winehq.org/wine/wine/-/commit/62fa96df16a0285cce1694ce47eef9… Author: Paul Gofman <pgofman(a)codeweavers.com> Date: Mon Jan 9 15:53:32 2023 -0600 dwmapi: Fill qpcVBlank in DwmGetCompositionTimingInfo(). --- dlls/dwmapi/dwmapi_main.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/dlls/dwmapi/dwmapi_main.c b/dlls/dwmapi/dwmapi_main.c index 56730f79cac..9ec03185f36 100644 --- a/dlls/dwmapi/dwmapi_main.c +++ b/dlls/dwmapi/dwmapi_main.c @@ -235,7 +235,7 @@ static int get_display_frequency(void) */ HRESULT WINAPI DwmGetCompositionTimingInfo(HWND hwnd, DWM_TIMING_INFO *info) { - LARGE_INTEGER performance_frequency; + LARGE_INTEGER performance_frequency, qpc; static int i, display_frequency; if (!info) @@ -258,6 +258,9 @@ HRESULT WINAPI DwmGetCompositionTimingInfo(HWND hwnd, DWM_TIMING_INFO *info) QueryPerformanceFrequency(&performance_frequency); info->qpcRefreshPeriod = performance_frequency.QuadPart / display_frequency; + QueryPerformanceCounter(&qpc); + info->qpcVBlank = (qpc.QuadPart / info->qpcRefreshPeriod) * info->qpcRefreshPeriod; + return S_OK; }

1 0

Stefan Dösinger : wined3d: Clamp the bound buffer range to the buffer size.
by Alexandre Julliard Jan. 10, 2023

Jan. 10, 2023

Module: wine Branch: master Commit: 09cafcc3b6eff3b046d3c08f64d9c1fb9ce54c47 URL: https://gitlab.winehq.org/wine/wine/-/commit/09cafcc3b6eff3b046d3c08f64d9c1… Author: Stefan Dösinger <stefan(a)codeweavers.com> Date: Wed Jan 4 11:16:14 2023 +0300 wined3d: Clamp the bound buffer range to the buffer size. Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=53671 --- dlls/wined3d/state.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/dlls/wined3d/state.c b/dlls/wined3d/state.c index eb25f01130e..a6d0b174e31 100644 --- a/dlls/wined3d/state.c +++ b/dlls/wined3d/state.c @@ -4618,7 +4618,9 @@ static void state_cb(struct wined3d_context *context, const struct wined3d_state buffer = buffer_state->buffer; bo_gl = wined3d_bo_gl(buffer->buffer_object); GL_EXTCALL(glBindBufferRange(GL_UNIFORM_BUFFER, base + i, - bo_gl->id, bo_gl->b.buffer_offset + buffer_state->offset, buffer_state->size)); + bo_gl->id, bo_gl->b.buffer_offset + buffer_state->offset, + min(buffer_state->size, buffer->resource.size - buffer_state->offset))); + buffer->bo_user.valid = true; } checkGLcall("bind constant buffers");

1 0