From: Francisco Casas fcasas@codeweavers.com
--- libs/vkd3d-shader/hlsl.y | 80 +++++++++++++++++++++++++++++++ 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, 104 insertions(+), 24 deletions(-)
diff --git a/libs/vkd3d-shader/hlsl.y b/libs/vkd3d-shader/hlsl.y index 4a551724..910261ba 100644 --- a/libs/vkd3d-shader/hlsl.y +++ b/libs/vkd3d-shader/hlsl.y @@ -2225,6 +2225,65 @@ static struct hlsl_ir_node *intrinsic_float_convert_arg(struct hlsl_ctx *ctx, return add_implicit_conversion(ctx, params->instrs, arg, type, loc); }
+static bool elementwise_intrinsic_convert_args_to_common_type(struct hlsl_ctx *ctx, + const struct parse_initializer *params, const struct vkd3d_shader_location *loc) +{ + enum hlsl_base_type base = params->args[0]->data_type->base_type; + bool vectors = false, matrices = false; + unsigned int dimx = 4, dimy = 4; + struct hlsl_type *common_type; + unsigned int i; + + for (i = 0; i < params->args_count; ++i) + { + struct hlsl_type *arg_type = params->args[i]->data_type; + + base = expr_common_base_type(base, arg_type->base_type); + + if (arg_type->type == HLSL_CLASS_VECTOR) + { + vectors = true; + dimx = min(dimx, arg_type->dimx); + } + else if (arg_type->type == HLSL_CLASS_MATRIX) + { + matrices = true; + dimx = min(dimx, arg_type->dimx); + dimy = min(dimy, arg_type->dimy); + } + } + + if (matrices && vectors) + { + hlsl_error(ctx, loc, VKD3D_SHADER_ERROR_HLSL_INVALID_TYPE, + "Cannot use both matrices and vectors in an elementwise intrinsic."); + return false; + } + else if (matrices) + { + common_type = hlsl_get_matrix_type(ctx, base, dimx, dimy); + } + else if (vectors) + { + common_type = hlsl_get_vector_type(ctx, base, dimx); + } + else + { + common_type = hlsl_get_scalar_type(ctx, base); + } + + for (i = 0; i < params->args_count; ++i) + { + struct hlsl_ir_node *new_arg; + + if (!(new_arg = add_implicit_conversion(ctx, params->instrs, params->args[i], common_type, loc))) + return NULL; + params->args[i] = new_arg; + } + + return true; +} + static bool intrinsic_abs(struct hlsl_ctx *ctx, const struct parse_initializer *params, const struct vkd3d_shader_location *loc) { @@ -2280,6 +2339,9 @@ static bool intrinsic_clamp(struct hlsl_ctx *ctx, { struct hlsl_ir_node *max;
+ if (!elementwise_intrinsic_convert_args_to_common_type(ctx, params, loc)) + return false; + if (!(max = add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_MAX, params->args[0], params->args[1], loc))) return false;
@@ -2361,6 +2423,9 @@ static bool intrinsic_ldexp(struct hlsl_ctx *ctx, { struct hlsl_ir_node *arg;
+ if (!elementwise_intrinsic_convert_args_to_common_type(ctx, params, loc)) + return false; + if (!(arg = intrinsic_float_convert_arg(ctx, params, params->args[1], loc))) return false;
@@ -2400,6 +2465,9 @@ static bool intrinsic_lerp(struct hlsl_ctx *ctx, { struct hlsl_ir_node *arg, *neg, *add, *mul;
+ if (!elementwise_intrinsic_convert_args_to_common_type(ctx, params, loc)) + return false; + if (!(arg = intrinsic_float_convert_arg(ctx, params, params->args[0], loc))) return false;
@@ -2418,12 +2486,18 @@ static bool intrinsic_lerp(struct hlsl_ctx *ctx, static bool intrinsic_max(struct hlsl_ctx *ctx, const struct parse_initializer *params, const struct vkd3d_shader_location *loc) { + if (!elementwise_intrinsic_convert_args_to_common_type(ctx, params, loc)) + return false; + return !!add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_MAX, params->args[0], params->args[1], loc); }
static bool intrinsic_min(struct hlsl_ctx *ctx, const struct parse_initializer *params, const struct vkd3d_shader_location *loc) { + if (!elementwise_intrinsic_convert_args_to_common_type(ctx, params, loc)) + return false; + return !!add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_MIN, params->args[0], params->args[1], loc); }
@@ -2558,6 +2632,9 @@ static bool intrinsic_pow(struct hlsl_ctx *ctx, { struct hlsl_ir_node *log, *exp, *arg, *mul;
+ if (!elementwise_intrinsic_convert_args_to_common_type(ctx, params, loc)) + return false; + if (!(arg = intrinsic_float_convert_arg(ctx, params, params->args[0], loc))) return false;
@@ -2606,6 +2683,9 @@ static bool intrinsic_smoothstep(struct hlsl_ctx *ctx, struct hlsl_type *common_type; unsigned int dimx, dimy;
+ if (!elementwise_intrinsic_convert_args_to_common_type(ctx, params, loc)) + return false; + min_arg = params->args[0]; max_arg = params->args[1]; x_arg = params->args[2]; diff --git a/tests/hlsl-clamp.shader_test b/tests/hlsl-clamp.shader_test index cc198735..1320c3dd 100644 --- a/tests/hlsl-clamp.shader_test +++ b/tests/hlsl-clamp.shader_test @@ -10,7 +10,7 @@ draw quad probe all rgba (-0.1, 0.7, -0.3, 0.3)
-[pixel shader todo] +[pixel shader] float4 main() : sv_target { float3x2 a = {6, 5, 4, 3, 2, 1}; @@ -22,11 +22,11 @@ float4 main() : sv_target }
[test] -todo draw quad -todo probe all rgba (5.5, 5.0, 4.2, 5.2) +draw quad +probe all rgba (5.5, 5.0, 4.2, 5.2)
-[pixel shader fail todo] +[pixel shader fail] float4 main() : sv_target { float2x2 a = {3.1, 3.1, 3.1, 3.1}; diff --git a/tests/hlsl-ldexp.shader_test b/tests/hlsl-ldexp.shader_test index bea97953..92988d37 100644 --- a/tests/hlsl-ldexp.shader_test +++ b/tests/hlsl-ldexp.shader_test @@ -32,7 +32,7 @@ draw quad probe all rgba (2.0, 0.00292968750, 4096.0, 6.33825300e+030)
-[pixel shader todo] +[pixel shader] float4 main() : sv_target { float2x3 a = {1, 2, 3, 4, 5, 6}; @@ -43,11 +43,11 @@ float4 main() : sv_target }
[test] -todo draw quad -todo probe all rgba (64.0, 64.0, 64.0, 40.0) +draw quad +probe all rgba (64.0, 64.0, 64.0, 40.0)
-[pixel shader fail todo] +[pixel shader fail] float4 main() : sv_target { float2x2 a = {1, 2, 3, 4}; diff --git a/tests/hlsl-lerp.shader_test b/tests/hlsl-lerp.shader_test index 3cd10ec1..15e90cef 100644 --- a/tests/hlsl-lerp.shader_test +++ b/tests/hlsl-lerp.shader_test @@ -36,7 +36,7 @@ draw quad probe all rgba (2.0, -10.0, -2.0, 1e12)
-[pixel shader todo] +[pixel shader] float4 main() : sv_target { float3x2 a = {6, 5, 4, 3, 2, 1}; @@ -48,11 +48,11 @@ float4 main() : sv_target }
[test] -todo draw quad -todo probe all rgba (-6.0, -2.2, 4.48, 8.28) +draw quad +probe all rgba (-6.0, -2.2, 4.48, 8.28) 1
-[pixel shader fail todo] +[pixel shader fail] float4 main() : sv_target { float2x2 a = {0, 1, 2, 3}; diff --git a/tests/hlsl-smoothstep.shader_test b/tests/hlsl-smoothstep.shader_test index ad53b50a..6dadfa8e 100644 --- a/tests/hlsl-smoothstep.shader_test +++ b/tests/hlsl-smoothstep.shader_test @@ -93,7 +93,7 @@ draw quad probe all rgba (1.0, 1.0, 0, 0) 1
-[pixel shader todo] +[pixel shader] float4 main() : sv_target { float2x3 a = {1, 1, 1, 1, 1, 1}; @@ -105,8 +105,8 @@ float4 main() : sv_target }
[test] -todo draw quad -todo probe all rgba (0.028, 0.104, 0.216, 0.352) 1 +draw quad +probe all rgba (0.028, 0.104, 0.216, 0.352) 6
[pixel shader] @@ -141,7 +141,7 @@ draw quad probe all rgba (0.5, 0.5, 0.5, 0.0)
-[pixel shader todo] +[pixel shader] float4 main() : sv_target { float4x1 a = {0.0, 0.0, 0.0, 0.0}; @@ -153,11 +153,11 @@ float4 main() : sv_target }
[test] -todo draw quad -todo probe all rgba (0.5, 0.5, 0.5, 0.5) +draw quad +probe all rgba (0.5, 0.5, 0.5, 0.5)
-[pixel shader fail todo] +[pixel shader fail] float4 main() : sv_target { float2x2 a = {0.0, 0.0, 0.0, 0.0}; diff --git a/tests/max.shader_test b/tests/max.shader_test index 7a917ec5..3a5c3125 100644 --- a/tests/max.shader_test +++ b/tests/max.shader_test @@ -24,7 +24,7 @@ draw quad probe all rgba (0.7, 0.8, 0.7, 0.2)
-[pixel shader todo] +[pixel shader] float4 main() : sv_target { float2x3 a = {1, 2, 3, 4, 5, 6}; @@ -35,11 +35,11 @@ float4 main() : sv_target }
[test] -todo draw quad -todo probe all rgba (6.0, 5.0, 4.0, 5.0) +draw quad +probe all rgba (6.0, 5.0, 4.0, 5.0)
-[pixel shader fail todo] +[pixel shader fail] float4 main() : sv_target { float2x2 a = {1, 2, 3, 4}; diff --git a/tests/pow.shader_test b/tests/pow.shader_test index 6f2b2741..0c1d7de3 100644 --- a/tests/pow.shader_test +++ b/tests/pow.shader_test @@ -25,7 +25,7 @@ todo draw quad todo probe all rgba (1.0, 32.0, 256.0, 125.0)
-[pixel shader fail todo] +[pixel shader fail] float4 main() : sv_target { float2x2 a = {1, 2, 3, 4};