February 2024 - Wine-commits

Petrichor Park : tests/shader-runner: Add tests for acos and asin trig intrinsics.
by Alexandre Julliard 06 Feb '24

06 Feb '24

Module: vkd3d Branch: master Commit: 92044d5e4114ae65b77b232fba442f347f21069d URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/92044d5e4114ae65b77b232fba442… Author: Petrichor Park <ppark(a)codeweavers.com> Date: Wed Sep 6 15:14:12 2023 -0500 tests/shader-runner: Add tests for acos and asin trig intrinsics. Extracted by Conor McCarthy from an HLSL patch, and modified to include SM 6 variations. --- Makefile.am | 1 + tests/hlsl/inverse-trig.shader_test | 92 +++++++++++++++++++++++++++++++++++++ 2 files changed, 93 insertions(+) diff --git a/Makefile.am b/Makefile.am index 90e7dcfc..1687d9a0 100644 --- a/Makefile.am +++ b/Makefile.am @@ -121,6 +121,7 @@ vkd3d_shader_tests = \ tests/hlsl/initializer-struct.shader_test \ tests/hlsl/intrinsic-override.shader_test \ tests/hlsl/invalid.shader_test \ + tests/hlsl/inverse-trig.shader_test \ tests/hlsl/is-front-face.shader_test \ tests/hlsl/ldexp.shader_test \ tests/hlsl/length.shader_test \ diff --git a/tests/hlsl/inverse-trig.shader_test b/tests/hlsl/inverse-trig.shader_test new file mode 100644 index 00000000..6077b4b5 --- /dev/null +++ b/tests/hlsl/inverse-trig.shader_test @@ -0,0 +1,92 @@ +% TPF does not define instructions for inverse trig; these intrinsics are +% decomposed into other instructions. FXC emits code which may vary wrt other +% implementations. DXIL defines intrinsics for inverse trig, to be implemented +% by the backend. + +[pixel shader todo] +uniform float4 a; + +float4 main() : sv_target +{ + return float4(acos(a.x), 0.0, 0.0, 0.0); +} + +[test] +uniform 0 float4 -1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (3.14159274, 0.0, 0.0, 0.0) 128 + +uniform 0 float4 -0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (2.094441441, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 0.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (1.57072878, 0.0, 0.0, 0.0) 1024 + +uniform 0 float4 0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (1.04715133, 0.0, 0.0, 0.0) 512 + +uniform 0 float4 1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (0.0, 0.0, 0.0, 0.0) 128 + +[pixel shader todo] +uniform float4 a; + +float4 main() : sv_target +{ + float4 result = float4(asin(a.x), 0.0, 0.0, 0.0); + // Quantize to cover implementation variations, which are quite large for asin(). + return round(result * 20000.0); +} + +[test] +uniform 0 float4 -1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (-31416.0, 0.0, 0.0, 0.0) + +[require] +shader model < 6.0 + +[test] +uniform 0 float4 -0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (-10473.0, 0.0, 0.0, 0.0) + +uniform 0 float4 0.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (1.0, 0.0, 0.0, 0.0) + +uniform 0 float4 0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (10473.0, 0.0, 0.0, 0.0) + +[require] +shader model >= 6.0 + +% We implement SM 6.0 inverse trig instructions using the native equivalents +% available in the backend. The values below are from the AMD Windows drivers, +% which are very close to those from Ubuntu's calculator app. Results from +% RADV are a bit lower than these, hence the large max ulp difference. +[test] +uniform 0 float4 -0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (-10472.0, 0.0, 0.0, 0.0) 4096 + +uniform 0 float4 0.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (0.0, 0.0, 0.0, 0.0) + +uniform 0 float4 0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (10472.0, 0.0, 0.0, 0.0) 4096 + +[require] +% reset requirements + +[test] +uniform 0 float4 1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (31416.0, 0.0, 0.0, 0.0)

1 0

Petrichor Park : tests/shader-runner: Add tests for atan and atan2 trig intrinsics.
by Alexandre Julliard 06 Feb '24

06 Feb '24

Module: vkd3d Branch: master Commit: e6746cbbdc649d8322c230ae6055cc5a44966551 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/e6746cbbdc649d8322c230ae6055c… Author: Petrichor Park <ppark(a)codeweavers.com> Date: Thu Dec 7 11:07:42 2023 -0600 tests/shader-runner: Add tests for atan and atan2 trig intrinsics. Extracted by Conor McCarthy from an HLSL patch, with ulp values doubled in some cases to cover SM 6 results. --- tests/hlsl/inverse-trig.shader_test | 104 ++++++++++++++++++++++++++++++++++++ 1 file changed, 104 insertions(+) diff --git a/tests/hlsl/inverse-trig.shader_test b/tests/hlsl/inverse-trig.shader_test index 6077b4b5..dd069fef 100644 --- a/tests/hlsl/inverse-trig.shader_test +++ b/tests/hlsl/inverse-trig.shader_test @@ -90,3 +90,107 @@ probe all rgba (10472.0, 0.0, 0.0, 0.0) 4096 uniform 0 float4 1.0 0.0 0.0 0.0 todo draw quad probe all rgba (31416.0, 0.0, 0.0, 0.0) + + +[pixel shader todo] +uniform float4 a; + +float4 main() : sv_target +{ + return float4(atan(a.x), 0.0, 0.0, 0.0); +} + +[test] +uniform 0 float4 -1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (-0.785409629, 0.0, 0.0, 0.0) 512 + +uniform 0 float4 -0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (-0.4636476, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 0.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (0.0, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (0.4636476, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (0.785409629, 0.0, 0.0, 0.0) 512 + +[pixel shader todo] +uniform float4 a; + +float4 main() : sv_target +{ + // The argument order is (y,x), and test case inputs are (y,x) also. + return float4(atan2(a.x, a.y), 0.0, 0.0, 0.0); +} + +[test] +% Non-degenerate cases +uniform 0 float4 1.0 1.0 0.0 0.0 +todo draw quad +probe all rgba (0.785385, 0.0, 0.0, 0.0) 512 + +uniform 0 float4 5.0 -5.0 0.0 0.0 +todo draw quad +probe all rgba (2.356194, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 -3.0 -3.0 0.0 0.0 +todo draw quad +probe all rgba (-2.356194, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (1.570796, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 -1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (-1.570796, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 0.0 1.0 0.0 0.0 +todo draw quad +probe all rgba (0.0, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 0.0 -1.0 0.0 0.0 +todo draw quad +probe all rgba (3.1415927, 0.0, 0.0, 0.0) 256 + +% Degenerate cases +uniform 0 float4 0.00001 0.00002 0.0 0.0 +todo draw quad +probe all rgba (0.463647, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 0.00001 -0.00002 0.0 0.0 +todo draw quad +probe all rgba (2.677945, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 -0.00001 100000.0 0.0 0.0 +todo draw quad +probe all rgba (-0.000000000099986595, 0.0, 0.0, 0.0) 2048 + +uniform 0 float4 10000000.0 0.00000001 0.0 0.0 +todo draw quad +probe all rgba (1.570796, 0.0, 0.0, 0.0) 256 + +% Negative zero behavior should be to treat it the +% same as normal zero. +uniform 0 float4 1000000000.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (1.570796, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 1000000000.0 -0.0 0.0 0.0 +todo draw quad +probe all rgba (1.570796, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 0.0 -1.0 0.0 0.0 +todo draw quad +probe all rgba (3.1415927, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 -0.0 -1.0 0.0 0.0 +todo draw quad +probe all rgba (3.1415927, 0.0, 0.0, 0.0) 256

1 0

Giovanni Mascellani : vkd3d-shader/dxil: Set the register before calling src_param_init_scalar().
by Alexandre Julliard 06 Feb '24

06 Feb '24

Module: vkd3d Branch: master Commit: 674f86042d67335603d3d1c4958b050ab697e9f4 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/674f86042d67335603d3d1c4958b0… Author: Giovanni Mascellani <gmascellani(a)codeweavers.com> Date: Tue Jan 30 11:22:45 2024 +0100 vkd3d-shader/dxil: Set the register before calling src_param_init_scalar(). Which uses the register to compute the appropriate swizzle. Fixes: 1f536238a89dc6bccb411c2db32d7b2010cc8fd7 --- libs/vkd3d-shader/dxil.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libs/vkd3d-shader/dxil.c b/libs/vkd3d-shader/dxil.c index ba951a6e..291012d1 100644 --- a/libs/vkd3d-shader/dxil.c +++ b/libs/vkd3d-shader/dxil.c @@ -4725,8 +4725,8 @@ static void sm6_parser_emit_extractval(struct sm6_parser *sm6, const struct dxil vsir_instruction_init(ins, &sm6->p.location, VKD3DSIH_MOV); src_param = instruction_src_params_alloc(ins, 1, sm6); - src_param_init_scalar(src_param, elem_idx); src_param->reg = src->u.reg; + src_param_init_scalar(src_param, elem_idx); instruction_dst_param_init_ssa_scalar(ins, sm6); }

1 0

Giovanni Mascellani : vkd3d-shader/spirv: Convert the swizzle according to the source bit width.
by Alexandre Julliard 06 Feb '24

06 Feb '24

Module: vkd3d Branch: master Commit: ee994e95dd06d00cbf16fe4ca550c1f6bf1dd015 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/ee994e95dd06d00cbf16fe4ca550c… Author: Giovanni Mascellani <gmascellani(a)codeweavers.com> Date: Mon Feb 5 12:12:02 2024 +0100 vkd3d-shader/spirv: Convert the swizzle according to the source bit width. Fixes: 1f536238a89dc6bccb411c2db32d7b2010cc8fd7 --- libs/vkd3d-shader/spirv.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libs/vkd3d-shader/spirv.c b/libs/vkd3d-shader/spirv.c index d4333e6d..3cc20b5c 100644 --- a/libs/vkd3d-shader/spirv.c +++ b/libs/vkd3d-shader/spirv.c @@ -7082,7 +7082,7 @@ static void spirv_compiler_emit_mov(struct spirv_compiler *compiler, } write_mask32 = data_type_is_64_bit(dst->reg.data_type) ? vsir_write_mask_32_from_64(dst->write_mask) : dst->write_mask; - swizzle32 = data_type_is_64_bit(dst->reg.data_type) ? vsir_swizzle_32_from_64(src->swizzle) : src->swizzle; + swizzle32 = data_type_is_64_bit(src->reg.data_type) ? vsir_swizzle_32_from_64(src->swizzle) : src->swizzle; component_count = vsir_write_mask_component_count(write_mask32); if (component_count != 1 && component_count != VKD3D_VEC4_SIZE && dst_reg_info.write_mask == VKD3DSP_WRITEMASK_ALL)

1 0

Giovanni Mascellani : vkd3d-shader/spirv: Move bool casting helpers above register loading helpers.
by Alexandre Julliard 06 Feb '24

06 Feb '24

Module: vkd3d Branch: master Commit: 49f0fd42b82b071a72dc98e129ced3b6061e0bc3 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/49f0fd42b82b071a72dc98e129ced… Author: Giovanni Mascellani <gmascellani(a)codeweavers.com> Date: Tue Jan 30 11:26:38 2024 +0100 vkd3d-shader/spirv: Move bool casting helpers above register loading helpers. --- libs/vkd3d-shader/spirv.c | 134 +++++++++++++++++++++++----------------------- 1 file changed, 67 insertions(+), 67 deletions(-) diff --git a/libs/vkd3d-shader/spirv.c b/libs/vkd3d-shader/spirv.c index 3cc20b5c..d836511c 100644 --- a/libs/vkd3d-shader/spirv.c +++ b/libs/vkd3d-shader/spirv.c @@ -3669,6 +3669,73 @@ static uint32_t spirv_compiler_emit_vector_shuffle(struct spirv_compiler *compil type_id, vector1_id, vector2_id, components, component_count); } +static uint32_t spirv_compiler_emit_int_to_bool(struct spirv_compiler *compiler, + enum vkd3d_shader_conditional_op condition, enum vkd3d_data_type data_type, + unsigned int component_count, uint32_t val_id) +{ + struct vkd3d_spirv_builder *builder = &compiler->spirv_builder; + uint32_t type_id; + SpvOp op; + + assert(!(condition & ~(VKD3D_SHADER_CONDITIONAL_OP_NZ | VKD3D_SHADER_CONDITIONAL_OP_Z))); + + type_id = vkd3d_spirv_get_type_id(builder, VKD3D_SHADER_COMPONENT_BOOL, component_count); + op = condition & VKD3D_SHADER_CONDITIONAL_OP_Z ? SpvOpIEqual : SpvOpINotEqual; + return vkd3d_spirv_build_op_tr2(builder, &builder->function_stream, op, type_id, val_id, + data_type == VKD3D_DATA_UINT64 + ? spirv_compiler_get_constant_uint64_vector(compiler, 0, component_count) + : spirv_compiler_get_constant_uint_vector(compiler, 0, component_count)); +} + +static uint32_t spirv_compiler_emit_bool_to_int(struct spirv_compiler *compiler, + unsigned int component_count, uint32_t val_id, bool signedness) +{ + struct vkd3d_spirv_builder *builder = &compiler->spirv_builder; + uint32_t type_id, true_id, false_id; + + true_id = spirv_compiler_get_constant_uint_vector(compiler, signedness ? 0xffffffff : 1, component_count); + false_id = spirv_compiler_get_constant_uint_vector(compiler, 0, component_count); + type_id = vkd3d_spirv_get_type_id(builder, VKD3D_SHADER_COMPONENT_UINT, component_count); + return vkd3d_spirv_build_op_select(builder, type_id, val_id, true_id, false_id); +} + +static uint32_t spirv_compiler_emit_bool_to_int64(struct spirv_compiler *compiler, + unsigned int component_count, uint32_t val_id, bool signedness) +{ + struct vkd3d_spirv_builder *builder = &compiler->spirv_builder; + uint32_t type_id, true_id, false_id; + + true_id = spirv_compiler_get_constant_uint64_vector(compiler, signedness ? UINT64_MAX : 1, + component_count); + false_id = spirv_compiler_get_constant_uint64_vector(compiler, 0, component_count); + type_id = vkd3d_spirv_get_type_id(builder, VKD3D_SHADER_COMPONENT_UINT64, component_count); + return vkd3d_spirv_build_op_select(builder, type_id, val_id, true_id, false_id); +} + +static uint32_t spirv_compiler_emit_bool_to_float(struct spirv_compiler *compiler, + unsigned int component_count, uint32_t val_id, bool signedness) +{ + struct vkd3d_spirv_builder *builder = &compiler->spirv_builder; + uint32_t type_id, true_id, false_id; + + true_id = spirv_compiler_get_constant_float_vector(compiler, signedness ? -1.0f : 1.0f, component_count); + false_id = spirv_compiler_get_constant_float_vector(compiler, 0.0f, component_count); + type_id = vkd3d_spirv_get_type_id(builder, VKD3D_SHADER_COMPONENT_FLOAT, component_count); + return vkd3d_spirv_build_op_select(builder, type_id, val_id, true_id, false_id); +} + +static uint32_t spirv_compiler_emit_bool_to_double(struct spirv_compiler *compiler, + unsigned int component_count, uint32_t val_id, bool signedness) +{ + struct vkd3d_spirv_builder *builder = &compiler->spirv_builder; + uint32_t type_id, true_id, false_id; + + true_id = spirv_compiler_get_constant_double_vector(compiler, signedness ? -1.0 : 1.0, component_count); + false_id = spirv_compiler_get_constant_double_vector(compiler, 0.0, component_count); + type_id = vkd3d_spirv_get_type_id(builder, VKD3D_SHADER_COMPONENT_DOUBLE, component_count); + return vkd3d_spirv_build_op_select(builder, type_id, val_id, true_id, false_id); +} + static uint32_t spirv_compiler_emit_load_constant(struct spirv_compiler *compiler, const struct vkd3d_shader_register *reg, uint32_t swizzle, uint32_t write_mask) { @@ -4383,73 +4450,6 @@ static void spirv_compiler_emit_interpolation_decorations(struct spirv_compiler } } -static uint32_t spirv_compiler_emit_int_to_bool(struct spirv_compiler *compiler, - enum vkd3d_shader_conditional_op condition, enum vkd3d_data_type data_type, - unsigned int component_count, uint32_t val_id) -{ - struct vkd3d_spirv_builder *builder = &compiler->spirv_builder; - uint32_t type_id; - SpvOp op; - - assert(!(condition & ~(VKD3D_SHADER_CONDITIONAL_OP_NZ | VKD3D_SHADER_CONDITIONAL_OP_Z))); - - type_id = vkd3d_spirv_get_type_id(builder, VKD3D_SHADER_COMPONENT_BOOL, component_count); - op = condition & VKD3D_SHADER_CONDITIONAL_OP_Z ? SpvOpIEqual : SpvOpINotEqual; - return vkd3d_spirv_build_op_tr2(builder, &builder->function_stream, op, type_id, val_id, - data_type == VKD3D_DATA_UINT64 - ? spirv_compiler_get_constant_uint64_vector(compiler, 0, component_count) - : spirv_compiler_get_constant_uint_vector(compiler, 0, component_count)); -} - -static uint32_t spirv_compiler_emit_bool_to_int(struct spirv_compiler *compiler, - unsigned int component_count, uint32_t val_id, bool signedness) -{ - struct vkd3d_spirv_builder *builder = &compiler->spirv_builder; - uint32_t type_id, true_id, false_id; - - true_id = spirv_compiler_get_constant_uint_vector(compiler, signedness ? 0xffffffff : 1, component_count); - false_id = spirv_compiler_get_constant_uint_vector(compiler, 0, component_count); - type_id = vkd3d_spirv_get_type_id(builder, VKD3D_SHADER_COMPONENT_UINT, component_count); - return vkd3d_spirv_build_op_select(builder, type_id, val_id, true_id, false_id); -} - -static uint32_t spirv_compiler_emit_bool_to_int64(struct spirv_compiler *compiler, - unsigned int component_count, uint32_t val_id, bool signedness) -{ - struct vkd3d_spirv_builder *builder = &compiler->spirv_builder; - uint32_t type_id, true_id, false_id; - - true_id = spirv_compiler_get_constant_uint64_vector(compiler, signedness ? UINT64_MAX : 1, - component_count); - false_id = spirv_compiler_get_constant_uint64_vector(compiler, 0, component_count); - type_id = vkd3d_spirv_get_type_id(builder, VKD3D_SHADER_COMPONENT_UINT64, component_count); - return vkd3d_spirv_build_op_select(builder, type_id, val_id, true_id, false_id); -} - -static uint32_t spirv_compiler_emit_bool_to_float(struct spirv_compiler *compiler, - unsigned int component_count, uint32_t val_id, bool signedness) -{ - struct vkd3d_spirv_builder *builder = &compiler->spirv_builder; - uint32_t type_id, true_id, false_id; - - true_id = spirv_compiler_get_constant_float_vector(compiler, signedness ? -1.0f : 1.0f, component_count); - false_id = spirv_compiler_get_constant_float_vector(compiler, 0.0f, component_count); - type_id = vkd3d_spirv_get_type_id(builder, VKD3D_SHADER_COMPONENT_FLOAT, component_count); - return vkd3d_spirv_build_op_select(builder, type_id, val_id, true_id, false_id); -} - -static uint32_t spirv_compiler_emit_bool_to_double(struct spirv_compiler *compiler, - unsigned int component_count, uint32_t val_id, bool signedness) -{ - struct vkd3d_spirv_builder *builder = &compiler->spirv_builder; - uint32_t type_id, true_id, false_id; - - true_id = spirv_compiler_get_constant_double_vector(compiler, signedness ? -1.0 : 1.0, component_count); - false_id = spirv_compiler_get_constant_double_vector(compiler, 0.0, component_count); - type_id = vkd3d_spirv_get_type_id(builder, VKD3D_SHADER_COMPONENT_DOUBLE, component_count); - return vkd3d_spirv_build_op_select(builder, type_id, val_id, true_id, false_id); -} - typedef uint32_t (*vkd3d_spirv_builtin_fixup_pfn)(struct spirv_compiler *compiler, uint32_t val_id);

1 0

Giovanni Mascellani : vkd3d-shader/spirv: Support bool TEMP registers.
by Alexandre Julliard 06 Feb '24

06 Feb '24

Module: vkd3d Branch: master Commit: eb723a8d2bc5240c6f2c33a3e1b60863f5df8af1 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/eb723a8d2bc5240c6f2c33a3e1b60… Author: Giovanni Mascellani <gmascellani(a)codeweavers.com> Date: Wed Jan 17 13:31:45 2024 +0100 vkd3d-shader/spirv: Support bool TEMP registers. --- libs/vkd3d-shader/spirv.c | 37 +++++++++++++++++++++++++++++++++---- 1 file changed, 33 insertions(+), 4 deletions(-) diff --git a/libs/vkd3d-shader/spirv.c b/libs/vkd3d-shader/spirv.c index d836511c..e9437c6f 100644 --- a/libs/vkd3d-shader/spirv.c +++ b/libs/vkd3d-shader/spirv.c @@ -3844,8 +3844,21 @@ static uint32_t spirv_compiler_emit_load_scalar(struct spirv_compiler *compiler, if (component_type != reg_info->component_type) { - type_id = vkd3d_spirv_get_type_id(builder, component_type, 1); - val_id = vkd3d_spirv_build_op_bitcast(builder, type_id, val_id); + if (component_type == VKD3D_SHADER_COMPONENT_BOOL) + { + if (reg_info->component_type != VKD3D_SHADER_COMPONENT_UINT) + { + type_id = vkd3d_spirv_get_type_id(builder, VKD3D_SHADER_COMPONENT_UINT, 1); + val_id = vkd3d_spirv_build_op_bitcast(builder, type_id, val_id); + } + val_id = spirv_compiler_emit_int_to_bool(compiler, VKD3D_SHADER_CONDITIONAL_OP_NZ, + VKD3D_DATA_UINT, 1, val_id); + } + else + { + type_id = vkd3d_spirv_get_type_id(builder, component_type, 1); + val_id = vkd3d_spirv_build_op_bitcast(builder, type_id, val_id); + } } return val_id; @@ -4032,8 +4045,21 @@ static uint32_t spirv_compiler_emit_load_reg(struct spirv_compiler *compiler, if (component_type != reg_info.component_type) { - type_id = vkd3d_spirv_get_type_id(builder, component_type, component_count); - val_id = vkd3d_spirv_build_op_bitcast(builder, type_id, val_id); + if (component_type == VKD3D_SHADER_COMPONENT_BOOL) + { + if (reg_info.component_type != VKD3D_SHADER_COMPONENT_UINT) + { + type_id = vkd3d_spirv_get_type_id(builder, VKD3D_SHADER_COMPONENT_UINT, component_count); + val_id = vkd3d_spirv_build_op_bitcast(builder, type_id, val_id); + } + val_id = spirv_compiler_emit_int_to_bool(compiler, VKD3D_SHADER_CONDITIONAL_OP_NZ, + VKD3D_DATA_UINT, component_count, val_id); + } + else + { + type_id = vkd3d_spirv_get_type_id(builder, component_type, component_count); + val_id = vkd3d_spirv_build_op_bitcast(builder, type_id, val_id); + } } return val_id; @@ -4227,6 +4253,9 @@ static void spirv_compiler_emit_store_reg(struct spirv_compiler *compiler, { if (data_type_is_64_bit(reg->data_type)) src_write_mask = vsir_write_mask_32_from_64(write_mask); + if (component_type == VKD3D_SHADER_COMPONENT_BOOL) + val_id = spirv_compiler_emit_bool_to_int(compiler, + vsir_write_mask_component_count(src_write_mask), val_id, false); type_id = vkd3d_spirv_get_type_id(builder, reg_info.component_type, vsir_write_mask_component_count(src_write_mask)); val_id = vkd3d_spirv_build_op_bitcast(builder, type_id, val_id);

1 0

Giovanni Mascellani : vkd3d-shader/ir: Materialize SSA registers to temporaries.
by Alexandre Julliard 06 Feb '24

06 Feb '24

Module: vkd3d Branch: master Commit: e0d3e9c376a9fc4f80bf9dc79201766030b191a7 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/e0d3e9c376a9fc4f80bf9dc792017… Author: Giovanni Mascellani <gmascellani(a)codeweavers.com> Date: Tue Jan 9 15:56:45 2024 +0100 vkd3d-shader/ir: Materialize SSA registers to temporaries. For simplicity PHI nodes are not currently handled. The goal for this pass is to make the CFG structurizer simpler, because it doesn't have to care about the more rigid rules SSA registers have to satisfy than TEMP registers. It is likely that the generated code will be harder for downstream compilers to optimize and execute efficiently, so once a complete structurizer is in place this pass should be removed, or at least greatly reduced in scope. --- libs/vkd3d-shader/ir.c | 81 ++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 81 insertions(+) diff --git a/libs/vkd3d-shader/ir.c b/libs/vkd3d-shader/ir.c index a0d6212c..b92a04b2 100644 --- a/libs/vkd3d-shader/ir.c +++ b/libs/vkd3d-shader/ir.c @@ -2637,6 +2637,84 @@ fail: return VKD3D_ERROR_OUT_OF_MEMORY; } +static void materialize_ssas_to_temps_process_src_param(struct vkd3d_shader_parser *parser, struct vkd3d_shader_src_param *src); + +/* This is idempotent: it can be safely applied more than once on the + * same register. */ +static void materialize_ssas_to_temps_process_reg(struct vkd3d_shader_parser *parser, struct vkd3d_shader_register *reg) +{ + unsigned int i; + + if (reg->type == VKD3DSPR_SSA) + { + reg->type = VKD3DSPR_TEMP; + reg->idx[0].offset += parser->program.temp_count; + } + + for (i = 0; i < reg->idx_count; ++i) + if (reg->idx[i].rel_addr) + materialize_ssas_to_temps_process_src_param(parser, reg->idx[i].rel_addr); +} + +static void materialize_ssas_to_temps_process_dst_param(struct vkd3d_shader_parser *parser, struct vkd3d_shader_dst_param *dst) +{ + materialize_ssas_to_temps_process_reg(parser, &dst->reg); +} + +static void materialize_ssas_to_temps_process_src_param(struct vkd3d_shader_parser *parser, struct vkd3d_shader_src_param *src) +{ + materialize_ssas_to_temps_process_reg(parser, &src->reg); +} + +static enum vkd3d_result materialize_ssas_to_temps(struct vkd3d_shader_parser *parser) +{ + struct vkd3d_shader_instruction *instructions = NULL; + size_t ins_capacity = 0, ins_count = 0, i; + + if (!reserve_instructions(&instructions, &ins_capacity, parser->program.instructions.count)) + goto fail; + + for (i = 0; i < parser->program.instructions.count; ++i) + { + struct vkd3d_shader_instruction *ins = &parser->program.instructions.elements[i]; + size_t j; + + if (ins->handler_idx == VKD3DSIH_PHI) + { + WARN("Unhandled PHI when materializing SSA registers.\n"); + vkd3d_shader_parser_error(parser, VKD3D_SHADER_ERROR_VSIR_NOT_IMPLEMENTED, + "Unhandled PHI when materializing SSA registers."); + vkd3d_free(instructions); + return VKD3D_ERROR_NOT_IMPLEMENTED; + } + + for (j = 0; j < ins->dst_count; ++j) + materialize_ssas_to_temps_process_dst_param(parser, &ins->dst[j]); + + for (j = 0; j < ins->src_count; ++j) + materialize_ssas_to_temps_process_src_param(parser, &ins->src[j]); + + if (!reserve_instructions(&instructions, &ins_capacity, ins_count + 1)) + goto fail; + + instructions[ins_count++] = *ins; + } + + vkd3d_free(parser->program.instructions.elements); + parser->program.instructions.elements = instructions; + parser->program.instructions.capacity = ins_capacity; + parser->program.instructions.count = ins_count; + parser->program.temp_count += parser->program.ssa_count; + parser->program.ssa_count = 0; + + return VKD3D_OK; + +fail: + vkd3d_free(instructions); + + return VKD3D_ERROR_OUT_OF_MEMORY; +} + enum vkd3d_result vkd3d_shader_normalise(struct vkd3d_shader_parser *parser, const struct vkd3d_shader_compile_info *compile_info) { @@ -2652,6 +2730,9 @@ enum vkd3d_result vkd3d_shader_normalise(struct vkd3d_shader_parser *parser, { if ((result = lower_switch_to_if_ladder(&parser->program)) < 0) return result; + + if ((result = materialize_ssas_to_temps(parser)) < 0) + return result; } else {

1 0

Giovanni Mascellani : vkd3d-shader/ir: Handle PHI nodes when materializing SSA registers.
by Alexandre Julliard 06 Feb '24

06 Feb '24

Module: vkd3d Branch: master Commit: 19aef21369c2e20d4239dd77c9c0c4e27b5dd18d URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/19aef21369c2e20d4239dd77c9c0c… Author: Giovanni Mascellani <gmascellani(a)codeweavers.com> Date: Tue Jan 9 15:56:45 2024 +0100 vkd3d-shader/ir: Handle PHI nodes when materializing SSA registers. PHI nodes cannot be used with TEMP registers, so they have to be converted to MOV/MOVC nodes and moved before the BRANCH node. --- libs/vkd3d-shader/ir.c | 170 +++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 163 insertions(+), 7 deletions(-) diff --git a/libs/vkd3d-shader/ir.c b/libs/vkd3d-shader/ir.c index b92a04b2..1f14f702 100644 --- a/libs/vkd3d-shader/ir.c +++ b/libs/vkd3d-shader/ir.c @@ -2666,27 +2666,110 @@ static void materialize_ssas_to_temps_process_src_param(struct vkd3d_shader_pars materialize_ssas_to_temps_process_reg(parser, &src->reg); } +static const struct vkd3d_shader_src_param *materialize_ssas_to_temps_compute_source(struct vkd3d_shader_instruction *ins, + unsigned int label) +{ + unsigned int i; + + assert(ins->handler_idx == VKD3DSIH_PHI); + + for (i = 0; i < ins->src_count; i += 2) + { + if (label_from_src_param(&ins->src[i + 1]) == label) + return &ins->src[i]; + } + + vkd3d_unreachable(); +} + +static bool materialize_ssas_to_temps_synthesize_mov(struct vkd3d_shader_parser *parser, + struct vkd3d_shader_instruction *instruction, const struct vkd3d_shader_location *loc, + const struct vkd3d_shader_dst_param *dest, const struct vkd3d_shader_src_param *cond, + const struct vkd3d_shader_src_param *source, bool invert) +{ + struct vkd3d_shader_src_param *src; + struct vkd3d_shader_dst_param *dst; + + if (!vsir_instruction_init_with_params(&parser->program, instruction, loc, + cond ? VKD3DSIH_MOVC : VKD3DSIH_MOV, 1, cond ? 3 : 1)) + return false; + + dst = instruction->dst; + src = instruction->src; + + dst[0] = *dest; + materialize_ssas_to_temps_process_dst_param(parser, &dst[0]); + + assert(dst[0].write_mask == VKD3DSP_WRITEMASK_0); + assert(dst[0].modifiers == 0); + assert(dst[0].shift == 0); + + if (cond) + { + src[0] = *cond; + src[1 + invert] = *source; + memset(&src[2 - invert], 0, sizeof(src[2 - invert])); + src[2 - invert].reg = dst[0].reg; + materialize_ssas_to_temps_process_src_param(parser, &src[1]); + materialize_ssas_to_temps_process_src_param(parser, &src[2]); + } + else + { + src[0] = *source; + materialize_ssas_to_temps_process_src_param(parser, &src[0]); + } + + return true; +} + static enum vkd3d_result materialize_ssas_to_temps(struct vkd3d_shader_parser *parser) { struct vkd3d_shader_instruction *instructions = NULL; + struct materialize_ssas_to_temps_block_data + { + size_t phi_begin; + size_t phi_count; + } *block_index = NULL; size_t ins_capacity = 0, ins_count = 0, i; + unsigned int current_label = 0; if (!reserve_instructions(&instructions, &ins_capacity, parser->program.instructions.count)) goto fail; + if (!(block_index = vkd3d_calloc(parser->program.block_count, sizeof(*block_index)))) + { + ERR("Failed to allocate block index.\n"); + goto fail; + } + for (i = 0; i < parser->program.instructions.count; ++i) { struct vkd3d_shader_instruction *ins = &parser->program.instructions.elements[i]; - size_t j; - if (ins->handler_idx == VKD3DSIH_PHI) + switch (ins->handler_idx) { - WARN("Unhandled PHI when materializing SSA registers.\n"); - vkd3d_shader_parser_error(parser, VKD3D_SHADER_ERROR_VSIR_NOT_IMPLEMENTED, - "Unhandled PHI when materializing SSA registers."); - vkd3d_free(instructions); - return VKD3D_ERROR_NOT_IMPLEMENTED; + case VKD3DSIH_LABEL: + current_label = label_from_src_param(&ins->src[0]); + break; + + case VKD3DSIH_PHI: + assert(current_label != 0); + assert(i != 0); + if (block_index[current_label - 1].phi_begin == 0) + block_index[current_label - 1].phi_begin = i; + block_index[current_label - 1].phi_count += 1; + break; + + default: + current_label = 0; + break; } + } + + for (i = 0; i < parser->program.instructions.count; ++i) + { + struct vkd3d_shader_instruction *ins = &parser->program.instructions.elements[i]; + size_t j; for (j = 0; j < ins->dst_count; ++j) materialize_ssas_to_temps_process_dst_param(parser, &ins->dst[j]); @@ -2694,6 +2777,77 @@ static enum vkd3d_result materialize_ssas_to_temps(struct vkd3d_shader_parser *p for (j = 0; j < ins->src_count; ++j) materialize_ssas_to_temps_process_src_param(parser, &ins->src[j]); + switch (ins->handler_idx) + { + case VKD3DSIH_LABEL: + current_label = label_from_src_param(&ins->src[0]); + break; + + case VKD3DSIH_BRANCH: + { + if (vsir_register_is_label(&ins->src[0].reg)) + { + const struct materialize_ssas_to_temps_block_data *data = &block_index[label_from_src_param(&ins->src[0]) - 1]; + + if (!reserve_instructions(&instructions, &ins_capacity, ins_count + data->phi_count)) + goto fail; + + for (j = data->phi_begin; j < data->phi_begin + data->phi_count; ++j) + { + const struct vkd3d_shader_src_param *source; + + source = materialize_ssas_to_temps_compute_source(&parser->program.instructions.elements[j], current_label); + if (!materialize_ssas_to_temps_synthesize_mov(parser, &instructions[ins_count], &ins->location, + &parser->program.instructions.elements[j].dst[0], NULL, source, false)) + goto fail; + + ++ins_count; + } + } + else + { + struct materialize_ssas_to_temps_block_data *data_true = &block_index[label_from_src_param(&ins->src[1]) - 1], + *data_false = &block_index[label_from_src_param(&ins->src[2]) - 1]; + const struct vkd3d_shader_src_param *cond = &ins->src[0]; + + if (!reserve_instructions(&instructions, &ins_capacity, + ins_count + data_true->phi_count + data_false->phi_count)) + goto fail; + + for (j = data_true->phi_begin; j < data_true->phi_begin + data_true->phi_count; ++j) + { + const struct vkd3d_shader_src_param *source; + + source = materialize_ssas_to_temps_compute_source(&parser->program.instructions.elements[j], current_label); + if (!materialize_ssas_to_temps_synthesize_mov(parser, &instructions[ins_count], &ins->location, + &parser->program.instructions.elements[j].dst[0], cond, source, false)) + goto fail; + + ++ins_count; + } + + for (j = data_false->phi_begin; j < data_false->phi_begin + data_false->phi_count; ++j) + { + const struct vkd3d_shader_src_param *source; + + source = materialize_ssas_to_temps_compute_source(&parser->program.instructions.elements[j], current_label); + if (!materialize_ssas_to_temps_synthesize_mov(parser, &instructions[ins_count], &ins->location, + &parser->program.instructions.elements[j].dst[0], cond, source, true)) + goto fail; + + ++ins_count; + } + } + break; + } + + case VKD3DSIH_PHI: + continue; + + default: + break; + } + if (!reserve_instructions(&instructions, &ins_capacity, ins_count + 1)) goto fail; @@ -2701,6 +2855,7 @@ static enum vkd3d_result materialize_ssas_to_temps(struct vkd3d_shader_parser *p } vkd3d_free(parser->program.instructions.elements); + vkd3d_free(block_index); parser->program.instructions.elements = instructions; parser->program.instructions.capacity = ins_capacity; parser->program.instructions.count = ins_count; @@ -2711,6 +2866,7 @@ static enum vkd3d_result materialize_ssas_to_temps(struct vkd3d_shader_parser *p fail: vkd3d_free(instructions); + vkd3d_free(block_index); return VKD3D_ERROR_OUT_OF_MEMORY; }

1 0

Giovanni Mascellani : vkd3d-shader/ir: Introduce a simple control flow graph structurizer.
by Alexandre Julliard 06 Feb '24

06 Feb '24

Module: vkd3d Branch: master Commit: 51f13391e6117592a4c77bbb457302c8c3bfa606 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/51f13391e6117592a4c77bbb45730… Author: Giovanni Mascellani <gmascellani(a)codeweavers.com> Date: Tue Jan 16 23:56:43 2024 +0100 vkd3d-shader/ir: Introduce a simple control flow graph structurizer. The structurizer is implemented along the lines of what is usually called the "structured program theorem": the control flow is completely virtualized by mean of an additional TEMP register which stores the block index which is currently running. The whole program is then converted to a huge switch construction enclosed in a loop, executing at each iteration the appropriate block and updating the register depending on block jump instruction. The algorithm's generality is also its major weakness: it accepts any input program, even if its CFG is not reducible, but the output program lacks any useful convergence information. It satisfies the letter of the SPIR-V requirements, but it is expected that it will be very inefficient to run on a GPU (unless a downstream compiler is able to devirtualize the control flow and do a proper convergence analysis pass). The algorithm is however very simple, and good enough to at least pass tests, enabling further development. A better alternative is expected to be upstreamed incrementally. Side note: the structured program theorem is often called the Böhm-Jacopini theorem; Böhm and Jacopini did indeed prove a variation of it, but their algorithm is different from what is commontly attributed to them and implemented here, so I opted for not using their name. --- libs/vkd3d-shader/ir.c | 147 ++++++++++++++++++++++++++++++++- tests/hlsl/conditional.shader_test | 2 +- tests/hlsl/for.shader_test | 2 +- tests/hlsl/function-return.shader_test | 44 +++++----- tests/hlsl/loop.shader_test | 6 +- tests/hlsl/return.shader_test | 32 +++---- tests/hlsl/sm6-ternary.shader_test | 4 +- tests/hlsl/switch.shader_test | 40 ++++----- 8 files changed, 209 insertions(+), 68 deletions(-)

1 0

Georg Lehmann : winevulkan: Prepare for VK_KHR_calibrated_timestamps.
by Alexandre Julliard 06 Feb '24

06 Feb '24

Module: wine Branch: master Commit: db03d2be88d80195635188cb51507f4c2d9c2b13 URL: https://gitlab.winehq.org/wine/wine/-/commit/db03d2be88d80195635188cb51507f… Author: Georg Lehmann <dadschoorse(a)gmail.com> Date: Sat Feb 3 17:23:52 2024 +0100 winevulkan: Prepare for VK_KHR_calibrated_timestamps. --- dlls/winevulkan/vulkan.c | 65 ++++++++++++++++++++++++++++++++---------------- 1 file changed, 44 insertions(+), 21 deletions(-) diff --git a/dlls/winevulkan/vulkan.c b/dlls/winevulkan/vulkan.c index 240ba3e8618..282c17f430e 100644 --- a/dlls/winevulkan/vulkan.c +++ b/dlls/winevulkan/vulkan.c @@ -1325,17 +1325,19 @@ static inline uint64_t convert_timestamp(VkTimeDomainEXT host_domain, VkTimeDoma return value; } -VkResult wine_vkGetCalibratedTimestampsEXT(VkDevice handle, uint32_t timestamp_count, - const VkCalibratedTimestampInfoEXT *timestamp_infos, - uint64_t *timestamps, uint64_t *max_deviation) +static VkResult wine_vk_get_timestamps(struct wine_device *device, uint32_t timestamp_count, + const VkCalibratedTimestampInfoEXT *timestamp_infos, + uint64_t *timestamps, uint64_t *max_deviation, + VkResult (*get_timestamps)(VkDevice, uint32_t, const VkCalibratedTimestampInfoEXT *, uint64_t *, uint64_t *)) { - struct wine_device *device = wine_device_from_handle(handle); VkCalibratedTimestampInfoEXT* host_timestamp_infos; unsigned int i; VkResult res; - TRACE("%p, %u, %p, %p, %p\n", device, timestamp_count, timestamp_infos, timestamps, max_deviation); - if (!(host_timestamp_infos = malloc(sizeof(VkCalibratedTimestampInfoEXT) * timestamp_count))) + if (timestamp_count == 0) + return VK_SUCCESS; + + if (!(host_timestamp_infos = calloc(sizeof(VkCalibratedTimestampInfoEXT), timestamp_count))) return VK_ERROR_OUT_OF_HOST_MEMORY; for (i = 0; i < timestamp_count; i++) @@ -1345,24 +1347,23 @@ VkResult wine_vkGetCalibratedTimestampsEXT(VkDevice handle, uint32_t timestamp_c host_timestamp_infos[i].timeDomain = map_to_host_time_domain(timestamp_infos[i].timeDomain); } - res = device->funcs.p_vkGetCalibratedTimestampsEXT(device->host_device, timestamp_count, host_timestamp_infos, - timestamps, max_deviation); - if (res != VK_SUCCESS) - return res; - - for (i = 0; i < timestamp_count; i++) - timestamps[i] = convert_timestamp(host_timestamp_infos[i].timeDomain, timestamp_infos[i].timeDomain, timestamps[i]); + res = get_timestamps(device->host_device, timestamp_count, host_timestamp_infos, timestamps, max_deviation); + if (res == VK_SUCCESS) + { + for (i = 0; i < timestamp_count; i++) + timestamps[i] = convert_timestamp(host_timestamp_infos[i].timeDomain, timestamp_infos[i].timeDomain, timestamps[i]); + } free(host_timestamp_infos); return res; } -VkResult wine_vkGetPhysicalDeviceCalibrateableTimeDomainsEXT(VkPhysicalDevice handle, - uint32_t *time_domain_count, - VkTimeDomainEXT *time_domains) +static VkResult wine_vk_get_time_domains(struct wine_phys_dev *phys_dev, + uint32_t *time_domain_count, + VkTimeDomainEXT *time_domains, + VkResult (*get_domains)(VkPhysicalDevice, uint32_t *, VkTimeDomainEXT *)) { - struct wine_phys_dev *phys_dev = wine_phys_dev_from_handle(handle); BOOL supports_device = FALSE, supports_monotonic = FALSE, supports_monotonic_raw = FALSE; const VkTimeDomainEXT performance_counter_domain = get_performance_counter_time_domain(); VkTimeDomainEXT *host_time_domains; @@ -1373,16 +1374,14 @@ VkResult wine_vkGetPhysicalDeviceCalibrateableTimeDomainsEXT(VkPhysicalDevice ha VkResult res; /* Find out the time domains supported on the host */ - res = phys_dev->instance->funcs.p_vkGetPhysicalDeviceCalibrateableTimeDomainsEXT(phys_dev->host_physical_device, - &host_time_domain_count, NULL); + res = get_domains(phys_dev->host_physical_device, &host_time_domain_count, NULL); if (res != VK_SUCCESS) return res; if (!(host_time_domains = malloc(sizeof(VkTimeDomainEXT) * host_time_domain_count))) return VK_ERROR_OUT_OF_HOST_MEMORY; - res = phys_dev->instance->funcs.p_vkGetPhysicalDeviceCalibrateableTimeDomainsEXT(phys_dev->host_physical_device, - &host_time_domain_count, host_time_domains); + res = get_domains(phys_dev->host_physical_device, &host_time_domain_count, host_time_domains); if (res != VK_SUCCESS) { free(host_time_domains); @@ -1432,6 +1431,30 @@ VkResult wine_vkGetPhysicalDeviceCalibrateableTimeDomainsEXT(VkPhysicalDevice ha return res; } +VkResult wine_vkGetCalibratedTimestampsEXT(VkDevice handle, uint32_t timestamp_count, + const VkCalibratedTimestampInfoEXT *timestamp_infos, + uint64_t *timestamps, uint64_t *max_deviation) +{ + struct wine_device *device = wine_device_from_handle(handle); + + TRACE("%p, %u, %p, %p, %p\n", device, timestamp_count, timestamp_infos, timestamps, max_deviation); + + return wine_vk_get_timestamps(device, timestamp_count, timestamp_infos, timestamps, max_deviation, + device->funcs.p_vkGetCalibratedTimestampsEXT); +} + +VkResult wine_vkGetPhysicalDeviceCalibrateableTimeDomainsEXT(VkPhysicalDevice handle, + uint32_t *time_domain_count, + VkTimeDomainEXT *time_domains) +{ + struct wine_phys_dev *phys_dev = wine_phys_dev_from_handle(handle); + + TRACE("%p, %p, %p\n", phys_dev, time_domain_count, time_domains); + + return wine_vk_get_time_domains(phys_dev, time_domain_count, time_domains, + phys_dev->instance->funcs.p_vkGetPhysicalDeviceCalibrateableTimeDomainsEXT); +} + void wine_vkGetPhysicalDeviceExternalSemaphoreProperties(VkPhysicalDevice phys_dev, const VkPhysicalDeviceExternalSemaphoreInfo *info, VkExternalSemaphoreProperties *properties)

1 0