- Wine-Devel - list.winehq.org

[PATCH vkd3d v3 3/8] tests: Test invalid number of arguments on initializers.
by Francisco Casas Jan. 19, 2022

Jan. 19, 2022

Signed-off-by: Francisco Casas <fcasas(a)codeweavers.com> --- Makefile.am | 2 + ...lsl-initializer-invalid-n-args.shader_test | 95 +++++++++++++++++++ 2 files changed, 97 insertions(+) create mode 100644 tests/hlsl-initializer-invalid-n-args.shader_test diff --git a/Makefile.am b/Makefile.am index 4a079c11..cdb59a85 100644 --- a/Makefile.am +++ b/Makefile.am @@ -68,6 +68,7 @@ vkd3d_shader_tests = \ tests/hlsl-duplicate-modifiers.shader_test \ tests/hlsl-for.shader_test \ tests/hlsl-function-overload.shader_test \ + tests/hlsl-initializer-invalid-n-args.shader_test \ tests/hlsl-initializer-local-array.shader_test \ tests/hlsl-initializer-static-array.shader_test \ tests/hlsl-initializer-struct.shader_test \ @@ -292,6 +293,7 @@ XFAIL_TESTS = \ tests/cast-to-uint.shader_test \ tests/conditional.shader_test \ tests/hlsl-array-dimension.shader_test \ + tests/hlsl-initializer-invalid-n-args.shader_test \ tests/hlsl-initializer-local-array.shader_test \ tests/hlsl-initializer-static-array.shader_test \ tests/hlsl-initializer-struct.shader_test \ diff --git a/tests/hlsl-initializer-invalid-n-args.shader_test b/tests/hlsl-initializer-invalid-n-args.shader_test new file mode 100644 index 00000000..bde6024f --- /dev/null +++ b/tests/hlsl-initializer-invalid-n-args.shader_test @@ -0,0 +1,95 @@ +[pixel shader] +float4 main() : sv_target +{ + float4 color[2][3] = { + 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24 + }; + return color[1][1]; +} + +[test] +draw quad +probe all rgba (17, 18, 19, 20) + + +[pixel shader fail] +float4 main() : sv_target +{ + float4 color[2][3] = { + 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23 + }; + return color[1][1]; +} + +[pixel shader fail] +float4 main() : sv_target +{ + float4 color[2][3] = { + 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25 + }; + return color[1][1]; +} + + +[pixel shader] +struct stu { + struct { + int bar; + float4 foo; + } bbb[2]; + int2 aaa; +}; + +float4 main() : sv_target +{ + struct stu val = { + 11, 12, 13, 14, 25, + 21, 22, 23, 24, 25, + 31, 32, + }; + return val.bbb[1].foo; +} + +[test] +draw quad +probe all rgba (22, 23, 24, 25) + + +[pixel shader fail] +struct stu { + struct { + int bar; + float4 foo; + } bbb[2]; + int2 aaa; +}; + +float4 main() : sv_target +{ + struct stu val = { + 11, 12, 13, 14, 25, + 21, 22, 23, 24, 25, + 31, 32, 33, + }; + return val.bbb[1].foo; +} + + +[pixel shader fail] +struct stu { + struct { + int bar; + float4 foo; + } bbb[2]; + int2 aaa; +}; + +float4 main() : sv_target +{ + struct stu val = { + 11, 12, 13, 14, 25, + 21, 22, 23, 24, 25, + 31, + }; + return val.bbb[1].foo; +} -- 2.25.1

1 0

[PATCH vkd3d v3 2/8] tests: Test struct initializers.
by Francisco Casas Jan. 19, 2022

Jan. 19, 2022

Signed-off-by: Francisco Casas <fcasas(a)codeweavers.com> --- Makefile.am | 2 + tests/hlsl-initializer-struct.shader_test | 78 +++++++++++++++++++++++ 2 files changed, 80 insertions(+) create mode 100644 tests/hlsl-initializer-struct.shader_test diff --git a/Makefile.am b/Makefile.am index efe99124..4a079c11 100644 --- a/Makefile.am +++ b/Makefile.am @@ -70,6 +70,7 @@ vkd3d_shader_tests = \ tests/hlsl-function-overload.shader_test \ tests/hlsl-initializer-local-array.shader_test \ tests/hlsl-initializer-static-array.shader_test \ + tests/hlsl-initializer-struct.shader_test \ tests/hlsl-invalid.shader_test \ tests/hlsl-majority-pragma.shader_test \ tests/hlsl-majority-typedef.shader_test \ @@ -293,6 +294,7 @@ XFAIL_TESTS = \ tests/hlsl-array-dimension.shader_test \ tests/hlsl-initializer-local-array.shader_test \ tests/hlsl-initializer-static-array.shader_test \ + tests/hlsl-initializer-struct.shader_test \ tests/hlsl-bool-cast.shader_test \ tests/hlsl-duplicate-modifiers.shader_test \ tests/hlsl-for.shader_test \ diff --git a/tests/hlsl-initializer-struct.shader_test b/tests/hlsl-initializer-struct.shader_test new file mode 100644 index 00000000..d5181f0e --- /dev/null +++ b/tests/hlsl-initializer-struct.shader_test @@ -0,0 +1,78 @@ +[pixel shader] +struct stu { + int3 aaa; + float3 bbb; + double2 ccc; + float4 ddd; + int eee; +}; + +float4 main() : sv_target +{ + struct stu val = { + 11, 12, 13, + 21, 22, 23, + 31, 32, + 41, 42, 43, 44, + 51 + }; + return val.ddd; +} + +[test] +draw quad +probe all rgba (41, 42, 43, 44) + + +[pixel shader] +struct stu { + int3 aaa; + float3 bbb[2]; + double2 ccc; + float4 ddd[3][2]; + int eee; +}; + +float4 main() : sv_target +{ + struct stu val = { + 11, 12, 13, + 211, 212, 213, 221, 222, 223, + 31, 32, + 4111, 4112, 4113, 4114, 4121, 4122, 4123, 4124, + 4211, 4212, 4213, 4214, 4221, 4222, 4223, 4224, + 4311, 4312, 4313, 4314, 4321, 4322, 4323, 4324, + 51 + }; + return val.ddd[2][0]; +} + +[test] +draw quad +probe all rgba (4311, 4312, 4313, 4314) + + +[pixel shader] +struct stu { + int3 aaa; + struct { + float4 foo; + int2 bar; + } bbb[3]; +}; + +float4 main() : sv_target +{ + struct stu val = { + 1, 2, 3, + 11, 12, 13, 14, 15, 16, + 21, 22, 23, 24, 25, 26, + 31, 32, 33, 34, 35, 36 + }; + return val.bbb[1].foo; +} + +[test] +draw quad +probe all rgba (21, 22, 23, 24) + -- 2.25.1

1 0

[PATCH vkd3d v3 01/12] vkd3d-shader/hlsl: Add support for sm4 instruction modifiers.
by Francisco Casas Jan. 19, 2022

Jan. 19, 2022

Signed-off-by: Francisco Casas <fcasas(a)codeweavers.com> --- libs/vkd3d-shader/hlsl_sm4.c | 54 ++++++++++++++++++++++++++++++++++++ 1 file changed, 54 insertions(+) diff --git a/libs/vkd3d-shader/hlsl_sm4.c b/libs/vkd3d-shader/hlsl_sm4.c index 2458018f..07bbd376 100644 --- a/libs/vkd3d-shader/hlsl_sm4.c +++ b/libs/vkd3d-shader/hlsl_sm4.c @@ -727,6 +727,46 @@ static enum vkd3d_sm4_resource_type sm4_resource_dimension(const struct hlsl_typ } } +struct sm4_instruction_modifier +{ + enum vkd3d_sm4_instruction_modifier type; + + union + { + struct + { + int u,v,w; + } aoffimmi; + }; +}; + +static uint32_t sm4_encode_instruction_modifier(struct sm4_instruction_modifier imod) +{ + uint32_t word = 0; + + word |= VKD3D_SM4_MODIFIER_MASK & imod.type; + + switch (imod.type) + { + case VKD3D_SM4_MODIFIER_AOFFIMMI: + assert(-8 <= imod.aoffimmi.u && imod.aoffimmi.u <= 7); + assert(-8 <= imod.aoffimmi.v && imod.aoffimmi.v <= 7); + assert(-8 <= imod.aoffimmi.w && imod.aoffimmi.w <= 7); + word |= ((uint32_t)imod.aoffimmi.u & 0xf) << VKD3D_SM4_AOFFIMMI_U_SHIFT; + word |= ((uint32_t)imod.aoffimmi.v & 0xf) << VKD3D_SM4_AOFFIMMI_V_SHIFT; + word |= ((uint32_t)imod.aoffimmi.w & 0xf) << VKD3D_SM4_AOFFIMMI_W_SHIFT; + break; + + default: + FIXME("Unhandled instruction modifier %#x.\n", imod.type); + return 0; + break; + } + + return word; +} + + struct sm4_register { enum vkd3d_sm4_register_type type; @@ -741,6 +781,9 @@ struct sm4_instruction { enum vkd3d_sm4_opcode opcode; + struct sm4_instruction_modifier modifiers[1]; + unsigned int modifier_count; + struct { struct sm4_register reg; @@ -913,6 +956,7 @@ static void write_sm4_instruction(struct vkd3d_bytecode_buffer *buffer, const st uint32_t token = instr->opcode; unsigned int size = 1, i, j; + size += instr->modifier_count; for (i = 0; i < instr->dst_count; ++i) size += sm4_register_order(&instr->dsts[i].reg); for (i = 0; i < instr->src_count; ++i) @@ -920,8 +964,18 @@ static void write_sm4_instruction(struct vkd3d_bytecode_buffer *buffer, const st size += instr->idx_count; token |= (size << VKD3D_SM4_INSTRUCTION_LENGTH_SHIFT); + + token |= ((0 < instr->modifier_count) << 31); + put_u32(buffer, token); + for (i = 0; i < instr->modifier_count; i++) + { + token = sm4_encode_instruction_modifier(instr->modifiers[i]); + token |= ((i + 1 < instr->modifier_count) << 31); + put_u32(buffer, token); + } + for (i = 0; i < instr->dst_count; ++i) { token = sm4_encode_register(&instr->dsts[i].reg); -- 2.25.1

5 7

[PATCH vkd3d] vkd3d-shader/hlsl: Handle constant value operations on a different file.
by Francisco Casas Jan. 19, 2022

Jan. 19, 2022

On constant folding, first switch on the op type, handling each operation using a function defined on hlsl_constant_ops.c. Each of these operations switches on the data type, if needed. Since constant value casting had to be rewritten, support was extended to all numeric types. Signed-off-by: Francisco Casas <fcasas(a)codeweavers.com> --- I suggest moving the definitions of the operations to this new file since they may end up being considerably large and similar. I am not sure if I should be on the copyright notice of the new files. Signed-off-by: Francisco Casas <fcasas(a)codeweavers.com> --- Makefile.am | 2 +- libs/vkd3d-shader/hlsl_codegen.c | 104 ++++--------- libs/vkd3d-shader/hlsl_constant_ops.c | 209 ++++++++++++++++++++++++++ libs/vkd3d-shader/hlsl_constant_ops.h | 30 ++++ 4 files changed, 268 insertions(+), 77 deletions(-) create mode 100644 libs/vkd3d-shader/hlsl_constant_ops.c create mode 100644 libs/vkd3d-shader/hlsl_constant_ops.h diff --git a/Makefile.am b/Makefile.am index 20fee06d..6bf8e1e6 100644 --- a/Makefile.am +++ b/Makefile.am @@ -194,6 +194,7 @@ libvkd3d_shader_la_SOURCES = \ libs/vkd3d-shader/hlsl.c \ libs/vkd3d-shader/hlsl.h \ libs/vkd3d-shader/hlsl_codegen.c \ + libs/vkd3d-shader/hlsl_constant_ops.c \ libs/vkd3d-shader/hlsl_sm1.c \ libs/vkd3d-shader/hlsl_sm4.c \ libs/vkd3d-shader/preproc.h \ @@ -304,7 +305,6 @@ XFAIL_TESTS = \ tests/hlsl-storage-qualifiers.shader_test \ tests/hlsl-vector-indexing.shader_test \ tests/hlsl-vector-indexing-uniform.shader_test \ - tests/max.shader_test \ tests/sampler.shader_test \ tests/texture-load.shader_test \ tests/texture-load-typed.shader_test \ diff --git a/libs/vkd3d-shader/hlsl_codegen.c b/libs/vkd3d-shader/hlsl_codegen.c index 75716bdf..9cc09738 100644 --- a/libs/vkd3d-shader/hlsl_codegen.c +++ b/libs/vkd3d-shader/hlsl_codegen.c @@ -19,6 +19,7 @@ */ #include "hlsl.h" +#include "hlsl_constant_ops.h" #include <stdio.h> /* Split uniforms into two variables representing the constant and temp @@ -652,11 +653,13 @@ static bool lower_narrowing_casts(struct hlsl_ctx *ctx, struct hlsl_ir_node *ins return false; } + static bool fold_constants(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context) { struct hlsl_ir_constant *arg1, *arg2 = NULL, *res; struct hlsl_ir_expr *expr; - unsigned int i, dimx; + unsigned int i; + bool success; if (instr->type != HLSL_IR_EXPR) return false; @@ -670,92 +673,41 @@ static bool fold_constants(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, voi arg1 = hlsl_ir_constant(expr->operands[0].node); if (expr->operands[1].node) arg2 = hlsl_ir_constant(expr->operands[1].node); - dimx = instr->data_type->dimx; if (!(res = hlsl_alloc(ctx, sizeof(*res)))) return false; init_node(&res->node, HLSL_IR_CONSTANT, instr->data_type, instr->loc); - switch (instr->data_type->base_type) + switch (expr->op) { - case HLSL_TYPE_FLOAT: - { - switch (expr->op) - { - case HLSL_OP1_CAST: - if (instr->data_type->dimx != arg1->node.data_type->dimx - || instr->data_type->dimy != arg1->node.data_type->dimy) - { - FIXME("Cast from %s to %s.\n", debug_hlsl_type(ctx, arg1->node.data_type), - debug_hlsl_type(ctx, instr->data_type)); - vkd3d_free(res); - return false; - } - - switch (arg1->node.data_type->base_type) - { - case HLSL_TYPE_INT: - for (i = 0; i < dimx; ++i) - res->value[i].f = arg1->value[i].i; - break; - - case HLSL_TYPE_UINT: - for (i = 0; i < dimx; ++i) - res->value[i].f = arg1->value[i].u; - break; - - default: - FIXME("Cast from %s to %s.\n", debug_hlsl_type(ctx, arg1->node.data_type), - debug_hlsl_type(ctx, instr->data_type)); - vkd3d_free(res); - return false; - } - break; - - default: - FIXME("Fold float op %#x.\n", expr->op); - vkd3d_free(res); - return false; - } + case HLSL_OP1_CAST: + success = constant_op1_cast(ctx, res, arg1); break; - } - - case HLSL_TYPE_UINT: - { - switch (expr->op) - { - case HLSL_OP1_NEG: - for (i = 0; i < instr->data_type->dimx; ++i) - res->value[i].u = -arg1->value[i].u; - break; - - case HLSL_OP2_ADD: - for (i = 0; i < instr->data_type->dimx; ++i) - res->value[i].u = arg1->value[i].u + arg2->value[i].u; - break; - - case HLSL_OP2_MUL: - for (i = 0; i < instr->data_type->dimx; ++i) - res->value[i].u = arg1->value[i].u * arg2->value[i].u; - break; - - default: - FIXME("Fold uint op %#x.\n", expr->op); - vkd3d_free(res); - return false; - } + case HLSL_OP1_NEG: + success = constant_op1_neg(ctx, res, arg1); + break; + case HLSL_OP2_ADD: + success = constant_op2_add(ctx, res, arg1, arg2); + break; + case HLSL_OP2_MUL: + success = constant_op2_mul(ctx, res, arg1, arg2); break; - } - default: - FIXME("Fold type %#x op %#x.\n", instr->data_type->base_type, expr->op); - vkd3d_free(res); - return false; + FIXME("Fold op %#x.\n", expr->op); + success = 0; } - list_add_before(&expr->node.entry, &res->node.entry); - replace_node(&expr->node, &res->node); - return true; + if (success) + { + list_add_before(&expr->node.entry, &res->node.entry); + replace_node(&expr->node, &res->node); + return true; + } + else + { + vkd3d_free(res); + return false; + } } static bool remove_trivial_swizzles(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context) diff --git a/libs/vkd3d-shader/hlsl_constant_ops.c b/libs/vkd3d-shader/hlsl_constant_ops.c new file mode 100644 index 00000000..406d1058 --- /dev/null +++ b/libs/vkd3d-shader/hlsl_constant_ops.c @@ -0,0 +1,209 @@ +/* + * HLSL constant value operations for constant folding + * + * Copyright 2022 Francisco Casas for CodeWeavers + * + * 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 + */ + +#include "hlsl_constant_ops.h" + +int constant_op1_cast(struct hlsl_ctx *ctx, struct hlsl_ir_constant *tgt, + struct hlsl_ir_constant *src) +{ + uint32_t u; + int32_t i; + double d; + float f; + + if (tgt->node.data_type->dimx != src->node.data_type->dimx || + tgt->node.data_type->dimy != src->node.data_type->dimy) + { + FIXME("Cast from %s to %s.\n", debug_hlsl_type(ctx, src->node.data_type), + debug_hlsl_type(ctx, tgt->node.data_type)); + return 0; + } + + for (int k = 0; k < 4; k++) + { + switch (src->node.data_type->base_type) + { + case HLSL_TYPE_FLOAT: + case HLSL_TYPE_HALF: + u = src->value[k].f; + i = src->value[k].f; + f = src->value[k].f; + d = src->value[k].f; + break; + case HLSL_TYPE_DOUBLE: + u = src->value[k].d; + i = src->value[k].d; + f = src->value[k].d; + d = src->value[k].d; + break; + case HLSL_TYPE_INT: + u = src->value[k].i; + i = src->value[k].i; + f = src->value[k].i; + d = src->value[k].i; + break; + case HLSL_TYPE_UINT: + u = src->value[k].u; + i = src->value[k].u; + f = src->value[k].u; + d = src->value[k].u; + break; + case HLSL_TYPE_BOOL: + u = !!src->value[k].u; + i = !!src->value[k].u; + f = !!src->value[k].u; + d = !!src->value[k].u; + break; + default: + FIXME("Cast from %s to %s.\n", debug_hlsl_type(ctx, src->node.data_type), + debug_hlsl_type(ctx, tgt->node.data_type)); + return 0; + } + switch (tgt->node.data_type->base_type) + { + case HLSL_TYPE_FLOAT: + case HLSL_TYPE_HALF: + tgt->value[k].f = f; + break; + case HLSL_TYPE_DOUBLE: + tgt->value[k].d = d; + break; + case HLSL_TYPE_INT: + tgt->value[k].i = i; + break; + case HLSL_TYPE_UINT: + tgt->value[k].u = u; + break; + case HLSL_TYPE_BOOL: + tgt->value[k].u = (!!u) * 0xffffffff; + break; + default: + FIXME("Cast from %s to %s.\n", debug_hlsl_type(ctx, src->node.data_type), + debug_hlsl_type(ctx, tgt->node.data_type)); + return 0; + } + } + return 1; +} + +int constant_op1_neg(struct hlsl_ctx *ctx, struct hlsl_ir_constant *tgt, + struct hlsl_ir_constant *src) +{ + enum hlsl_base_type type = tgt->node.data_type->base_type; + + assert(type == src->node.data_type->base_type); + + for (int k = 0; k < 4; k++) + { + switch (type) + { + case HLSL_TYPE_FLOAT: + case HLSL_TYPE_HALF: + tgt->value[k].f = -src->value[k].f; + break; + case HLSL_TYPE_DOUBLE: + tgt->value[k].d = -src->value[k].d; + break; + case HLSL_TYPE_INT: + tgt->value[k].i = -src->value[k].i; + break; + case HLSL_TYPE_UINT: + tgt->value[k].u = -src->value[k].u; + break; + default: + FIXME("Fold \"%s\" for type %s.", debug_hlsl_expr_op(HLSL_OP1_NEG), + debug_hlsl_type(ctx, tgt->node.data_type)); + return 0; + } + } + return 1; +} + +int constant_op2_add(struct hlsl_ctx *ctx, struct hlsl_ir_constant *tgt, + struct hlsl_ir_constant *src1, struct hlsl_ir_constant *src2) +{ + enum hlsl_base_type type = tgt->node.data_type->base_type; + + assert(type == src1->node.data_type->base_type); + assert(type == src2->node.data_type->base_type); + + for (int k = 0; k < 4; k++) + { + switch (type) + { + case HLSL_TYPE_FLOAT: + case HLSL_TYPE_HALF: + tgt->value[k].f = src1->value[k].f + src2->value[k].f; + break; + case HLSL_TYPE_DOUBLE: + tgt->value[k].d = src1->value[k].d + src2->value[k].d; + break; + case HLSL_TYPE_INT: + tgt->value[k].i = src1->value[k].i + src2->value[k].i; + break; + case HLSL_TYPE_UINT: + tgt->value[k].u = src1->value[k].u + src2->value[k].u; + break; + default: + FIXME("Fold \"%s\" for type %s.", debug_hlsl_expr_op(HLSL_OP2_ADD), + debug_hlsl_type(ctx, tgt->node.data_type)); + return 0; + } + } + return 1; +} + +int constant_op2_mul(struct hlsl_ctx *ctx, struct hlsl_ir_constant *tgt, + struct hlsl_ir_constant *src1, struct hlsl_ir_constant *src2) +{ + enum hlsl_base_type type = tgt->node.data_type->base_type; + + assert(type == src1->node.data_type->base_type); + assert(type == src2->node.data_type->base_type); + + for (int k = 0; k < 4; k++) + { + switch (type) + { + case HLSL_TYPE_FLOAT: + case HLSL_TYPE_HALF: + tgt->value[k].f = src1->value[k].f * src2->value[k].f; + break; + case HLSL_TYPE_DOUBLE: + tgt->value[k].d = src1->value[k].d * src2->value[k].d; + break; + case HLSL_TYPE_INT: + tgt->value[k].i = src1->value[k].i * src2->value[k].i; + break; + case HLSL_TYPE_UINT: + tgt->value[k].u = src1->value[k].u * src2->value[k].u; + break; + default: + FIXME("Fold \"%s\" for type %s.", debug_hlsl_expr_op(HLSL_OP2_MUL), + debug_hlsl_type(ctx, tgt->node.data_type)); + return 0; + } + } + return 1; +} + + + + diff --git a/libs/vkd3d-shader/hlsl_constant_ops.h b/libs/vkd3d-shader/hlsl_constant_ops.h new file mode 100644 index 00000000..17a8a2e8 --- /dev/null +++ b/libs/vkd3d-shader/hlsl_constant_ops.h @@ -0,0 +1,30 @@ +/* + * HLSL constant value operations for constant folding + * + * Copyright 2022 Francisco Casas for CodeWeavers + * + * 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 + */ + +#include "hlsl.h" + +int constant_op1_cast(struct hlsl_ctx *ctx, struct hlsl_ir_constant *tgt, + struct hlsl_ir_constant *src); +int constant_op1_neg(struct hlsl_ctx *ctx, struct hlsl_ir_constant *tgt, + struct hlsl_ir_constant *src); +int constant_op2_add(struct hlsl_ctx *ctx, struct hlsl_ir_constant *tgt, + struct hlsl_ir_constant *src1, struct hlsl_ir_constant *src2); +int constant_op2_mul(struct hlsl_ctx *ctx, struct hlsl_ir_constant *tgt, + struct hlsl_ir_constant *src1, struct hlsl_ir_constant *src2); -- 2.25.1

4 7

[PATCH vkd3d 1/8] vkd3d-shader/hlsl: Use list_empty() instead of comparing list_count() against zero.
by Matteo Bruni Jan. 19, 2022

Jan. 19, 2022

From: Zebediah Figura <zfigura(a)codeweavers.com> Signed-off-by: Zebediah Figura <zfigura(a)codeweavers.com> Signed-off-by: Giovanni Mascellani <gmascellani(a)codeweavers.com> Signed-off-by: Matteo Bruni <mbruni(a)codeweavers.com> --- I'm sorry for the very late signoffs. For some reason, this year resuming work after vacation has been a real struggle. Then my usual workflow of going through all the pending stuff before sending replies didn't help at all. I'm going to resend the first batch of pending patches (that look okay to me) because: - the oldest of the lot just fell out of the patch tracker - I want to make some trivial changes to a few more - it seems nice to clarify order (although the patches look mostly independent) I'm also going to send a bunch of comments to various, more or less old, email threads. Most of which I had written at the time, but then see above. I'll make it a point to change at least that going forward. Zebediah and Giovanni's reviews and comments that have been flowing in the meantime have been very helpful and you can assume that I agree if I don't say otherwise. libs/vkd3d-shader/hlsl.y | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libs/vkd3d-shader/hlsl.y b/libs/vkd3d-shader/hlsl.y index 636882c4..5a17449a 100644 --- a/libs/vkd3d-shader/hlsl.y +++ b/libs/vkd3d-shader/hlsl.y @@ -326,7 +326,7 @@ static bool append_conditional_break(struct hlsl_ctx *ctx, struct list *cond_lis struct hlsl_ir_if *iff; /* E.g. "for (i = 0; ; ++i)". */ - if (!list_count(cond_list)) + if (list_empty(cond_list)) return true; condition = node_from_list(cond_list); -- 2.34.1

2 15

[PATCH vkd3d] vkd3d/tests: Revert "Enable on vkd3d the tests for ID3D12CommandQueue::Wait()."
by Conor McCarthy Jan. 19, 2022

Jan. 19, 2022

This reverts commit 41230b6e763980ad02930213a11336e1358197e3. The tests can hang and the Wait() implementation has issues. Signed-off-by: Conor McCarthy <cmccarthy(a)codeweavers.com> --- tests/d3d12.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/tests/d3d12.c b/tests/d3d12.c index 9d4fb7d4..ebf137a8 100644 --- a/tests/d3d12.c +++ b/tests/d3d12.c @@ -33289,6 +33289,12 @@ static void test_queue_wait(void) check_readback_data_uint(&rb, NULL, 0xff00ff00, 0); release_resource_readback(&rb); + if (!vkd3d_test_platform_is_windows()) + { + skip("Wait() is not implemented yet.\n"); /* FIXME */ + goto skip_tests; + } + /* Wait() before CPU signal */ update_buffer_data(cb, 0, sizeof(blue), &blue); queue_wait(queue, fence, 2); @@ -33364,6 +33370,7 @@ static void test_queue_wait(void) check_readback_data_uint(&rb, NULL, 0xff00ff00, 0); release_resource_readback(&rb); +skip_tests: /* Signal() and Wait() in the same command queue */ update_buffer_data(cb, 0, sizeof(blue), &blue); queue_signal(queue, fence, 7); -- 2.34.1

2 1

[PATCH vkd3d v2 1/5] vkd3d-shader/hlsl: Write SM4 equality instructions.
by Giovanni Mascellani Jan. 19, 2022

Jan. 19, 2022

Signed-off-by: Giovanni Mascellani <gmascellani(a)codeweavers.com> --- v2: Add hlsl_fixme(). --- libs/vkd3d-shader/hlsl_sm4.c | 34 ++++++++++++++++++++++++++++++++++ 1 file changed, 34 insertions(+) diff --git a/libs/vkd3d-shader/hlsl_sm4.c b/libs/vkd3d-shader/hlsl_sm4.c index 2458018f..26ff019c 100644 --- a/libs/vkd3d-shader/hlsl_sm4.c +++ b/libs/vkd3d-shader/hlsl_sm4.c @@ -1489,6 +1489,40 @@ static void write_sm4_expr(struct hlsl_ctx *ctx, break; } + case HLSL_TYPE_BOOL: + { + switch (expr->op) + { + case HLSL_OP2_EQUAL: + { + const struct hlsl_type *src_type = arg1->data_type; + + switch (src_type->base_type) + { + case HLSL_TYPE_FLOAT: + write_sm4_binary_op(buffer, VKD3D_SM4_OP_EQ, &expr->node, arg1, arg2); + break; + + case HLSL_TYPE_BOOL: + case HLSL_TYPE_INT: + case HLSL_TYPE_UINT: + write_sm4_binary_op(buffer, VKD3D_SM4_OP_IEQ, &expr->node, arg1, arg2); + break; + + default: + hlsl_fixme(ctx, &expr->node.loc, "SM4 equality between \"%s\".\n", debug_hlsl_type(ctx, src_type)); + break; + } + break; + } + + default: + hlsl_fixme(ctx, &expr->node.loc, "SM4 bool \"%s\" expression.\n", debug_hlsl_expr_op(expr->op)); + break; + } + break; + } + default: { struct vkd3d_string_buffer *string; -- 2.34.1

3 11

[PATCH vkd3d v2 06/10] vkd3d-shader/hlsl: Properly free parse_variable_def memory in declare_vars().
by Francisco Casas Jan. 19, 2022

Jan. 19, 2022

Signed-off-by: Francisco Casas <fcasas(a)codeweavers.com> --- libs/vkd3d-shader/hlsl.y | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/libs/vkd3d-shader/hlsl.y b/libs/vkd3d-shader/hlsl.y index 636882c4..988e0743 100644 --- a/libs/vkd3d-shader/hlsl.y +++ b/libs/vkd3d-shader/hlsl.y @@ -1424,6 +1424,9 @@ static struct list *declare_vars(struct hlsl_ctx *ctx, struct hlsl_type *basic_t hlsl_error(ctx, &v->loc, VKD3D_SHADER_ERROR_HLSL_MISSING_INITIALIZER, "Const variable \"%s\" is missing an initializer.", var->name); hlsl_free_var(var); + free_parse_initializer(&v->initializer); + if (v->arrays.count) + vkd3d_free(v->arrays.sizes); vkd3d_free(v); continue; } @@ -1436,6 +1439,9 @@ static struct list *declare_vars(struct hlsl_ctx *ctx, struct hlsl_type *basic_t "Variable \"%s\" was already declared in this scope.", var->name); hlsl_note(ctx, &old->loc, VKD3D_SHADER_LOG_ERROR, "\"%s\" was previously declared here.", old->name); hlsl_free_var(var); + free_parse_initializer(&v->initializer); + if (v->arrays.count) + vkd3d_free(v->arrays.sizes); vkd3d_free(v); continue; } @@ -1485,6 +1491,7 @@ static struct list *declare_vars(struct hlsl_ctx *ctx, struct hlsl_type *basic_t { hlsl_fixme(ctx, &v->loc, "Array initializer."); free_parse_initializer(&v->initializer); + vkd3d_free(v->arrays.sizes); vkd3d_free(v); continue; } @@ -1507,6 +1514,9 @@ static struct list *declare_vars(struct hlsl_ctx *ctx, struct hlsl_type *basic_t list_move_tail(statements_list, v->initializer.instrs); vkd3d_free(v->initializer.instrs); } + + if (v->arrays.count) + vkd3d_free(v->arrays.sizes); vkd3d_free(v); } vkd3d_free(var_list); -- 2.25.1

4 6

[PATCH vkd3d v2 07/10] vkd3d-shader/hlsl: Support complex numeric initializers.
by Francisco Casas Jan. 19, 2022

Jan. 19, 2022

Signed-off-by: Francisco Casas <fcasas(a)codeweavers.com> --- Makefile.am | 1 - libs/vkd3d-shader/hlsl.y | 72 ++++++++++++++++++++++++++++++++++------ 2 files changed, 61 insertions(+), 12 deletions(-) diff --git a/Makefile.am b/Makefile.am index d15e50c3..1086d028 100644 --- a/Makefile.am +++ b/Makefile.am @@ -298,7 +298,6 @@ XFAIL_TESTS = \ tests/hlsl-initializer-flattening.shader_test \ tests/hlsl-initializer-invalid-n-args.shader_test \ tests/hlsl-initializer-local-array.shader_test \ - tests/hlsl-initializer-numeric.shader_test \ tests/hlsl-initializer-static-array.shader_test \ tests/hlsl-initializer-struct.shader_test \ tests/hlsl-bool-cast.shader_test \ diff --git a/libs/vkd3d-shader/hlsl.y b/libs/vkd3d-shader/hlsl.y index 988e0743..a89e4432 100644 --- a/libs/vkd3d-shader/hlsl.y +++ b/libs/vkd3d-shader/hlsl.y @@ -1278,6 +1278,54 @@ static bool add_increment(struct hlsl_ctx *ctx, struct list *instrs, bool decrem return true; } +static void numeric_var_initializer(struct hlsl_ctx *ctx, struct hlsl_ir_var *var, + struct parse_variable_def *v, unsigned int reg_offset, struct hlsl_type *type, + unsigned int *initializer_i, struct list *instrs) +{ + unsigned int writemask_offset = 0; + unsigned int components_read = 0; + + if (type->type == HLSL_CLASS_MATRIX) + hlsl_fixme(ctx, &var->loc, "Matrix initializer\n"); + + while (components_read < hlsl_type_component_count(type)) + { + struct hlsl_ir_store *store; + struct hlsl_ir_constant *c; + struct hlsl_ir_node *node; + unsigned int width; + + assert(*initializer_i < v->initializer.args_count); + node = v->initializer.args[*initializer_i]; + *initializer_i += 1; + + width = hlsl_type_component_count(node->data_type); + components_read += width; + + if (components_read > hlsl_type_component_count(type)) + { + hlsl_error(ctx, &node->loc, VKD3D_SHADER_ERROR_HLSL_WRONG_PARAMETER_COUNT, + "Initializer argument has spare components.\n"); + return; + } + + if (!(node = add_implicit_conversion(ctx, instrs, node, + hlsl_get_vector_type(ctx, type->base_type, width), &node->loc))) + return; + + if (!(c = hlsl_new_uint_constant(ctx, reg_offset, node->loc))) + return; + list_add_tail(instrs, &c->node.entry); + + if (!(store = hlsl_new_store(ctx, var, &c->node, node, + ((1u << width) - 1) << writemask_offset, node->loc))) + return; + list_add_tail(instrs, &store->node.entry); + + writemask_offset += width; + } +} + static void struct_var_initializer(struct hlsl_ctx *ctx, struct list *list, struct hlsl_ir_var *var, struct parse_initializer *initializer) { @@ -1449,7 +1497,6 @@ static struct list *declare_vars(struct hlsl_ctx *ctx, struct hlsl_type *basic_t if (v->initializer.args_count) { unsigned int size = initializer_size(&v->initializer); - struct hlsl_ir_load *load; if (type->type <= HLSL_CLASS_LAST_NUMERIC && type->dimx * type->dimy != size && size != 1) @@ -1473,7 +1520,6 @@ static struct list *declare_vars(struct hlsl_ctx *ctx, struct hlsl_type *basic_t vkd3d_free(v); continue; } - if (type->type == HLSL_CLASS_STRUCT) { struct_var_initializer(ctx, statements_list, var, &v->initializer); @@ -1495,23 +1541,27 @@ static struct list *declare_vars(struct hlsl_ctx *ctx, struct hlsl_type *basic_t vkd3d_free(v); continue; } - if (v->initializer.args_count > 1) + + if (type->type <= HLSL_CLASS_LAST_NUMERIC && size == 1) { - hlsl_fixme(ctx, &v->loc, "Complex initializer."); - free_parse_initializer(&v->initializer); - vkd3d_free(v); - continue; + struct hlsl_ir_load *load = hlsl_new_var_load(ctx, var, var->loc); + + list_add_tail(v->initializer.instrs, &load->node.entry); + add_assignment(ctx, v->initializer.instrs, &load->node, ASSIGN_OP_ASSIGN, v->initializer.args[0]); } + else + { + unsigned int initializer_i = 0; - load = hlsl_new_var_load(ctx, var, var->loc); - list_add_tail(v->initializer.instrs, &load->node.entry); - add_assignment(ctx, v->initializer.instrs, &load->node, ASSIGN_OP_ASSIGN, v->initializer.args[0]); - vkd3d_free(v->initializer.args); + numeric_var_initializer(ctx, var, v, 0, type, &initializer_i, v->initializer.instrs); + } if (modifiers & HLSL_STORAGE_STATIC) list_move_tail(&ctx->static_initializers, v->initializer.instrs); else list_move_tail(statements_list, v->initializer.instrs); + + vkd3d_free(v->initializer.args); vkd3d_free(v->initializer.instrs); } -- 2.25.1

4 8

[PATCH vkd3d v4 05/10] vkd3d-shader/hlsl: Add Gather methods with offset and tests.
by Francisco Casas Jan. 19, 2022

Jan. 19, 2022

Signed-off-by: Francisco Casas <fcasas(a)codeweavers.com> --- Makefile.am | 4 + libs/vkd3d-shader/hlsl.c | 4 + libs/vkd3d-shader/hlsl.h | 4 + libs/vkd3d-shader/hlsl.y | 106 ++++++++++++++++++++++++++ libs/vkd3d-shader/hlsl_sm4.c | 61 +++++++++++++++ tests/hlsl-gather-offset.shader_test | 93 +++++++++++++++++++++++ tests/hlsl-gather.shader_test | 107 +++++++++++++++++++++++++++ 7 files changed, 379 insertions(+) create mode 100644 tests/hlsl-gather-offset.shader_test create mode 100644 tests/hlsl-gather.shader_test diff --git a/Makefile.am b/Makefile.am index 16848f5a..749df363 100644 --- a/Makefile.am +++ b/Makefile.am @@ -68,6 +68,8 @@ vkd3d_shader_tests = \ tests/hlsl-duplicate-modifiers.shader_test \ tests/hlsl-for.shader_test \ tests/hlsl-function-overload.shader_test \ + tests/hlsl-gather-offset.shader_test \ + tests/hlsl-gather.shader_test \ tests/hlsl-invalid.shader_test \ tests/hlsl-majority-pragma.shader_test \ tests/hlsl-majority-typedef.shader_test \ @@ -294,6 +296,8 @@ XFAIL_TESTS = \ tests/hlsl-duplicate-modifiers.shader_test \ tests/hlsl-for.shader_test \ tests/hlsl-function-overload.shader_test \ + tests/hlsl-gather-offset.shader_test \ + tests/hlsl-gather.shader_test \ tests/hlsl-majority-pragma.shader_test \ tests/hlsl-majority-typedef.shader_test \ tests/hlsl-nested-arrays.shader_test \ diff --git a/libs/vkd3d-shader/hlsl.c b/libs/vkd3d-shader/hlsl.c index b426e461..3ace7462 100644 --- a/libs/vkd3d-shader/hlsl.c +++ b/libs/vkd3d-shader/hlsl.c @@ -1277,6 +1277,10 @@ static void dump_ir_resource_load(struct vkd3d_string_buffer *buffer, const stru { [HLSL_RESOURCE_LOAD] = "load_resource", [HLSL_RESOURCE_SAMPLE] = "sample", + [HLSL_RESOURCE_GATHER_RED] = "gather_red", + [HLSL_RESOURCE_GATHER_GREEN] = "gather_green", + [HLSL_RESOURCE_GATHER_BLUE] = "gather_blue", + [HLSL_RESOURCE_GATHER_ALPHA] = "gather_alpha", }; vkd3d_string_buffer_printf(buffer, "%s(resource = ", type_names[load->load_type]); diff --git a/libs/vkd3d-shader/hlsl.h b/libs/vkd3d-shader/hlsl.h index 2396adb4..49fa8d9d 100644 --- a/libs/vkd3d-shader/hlsl.h +++ b/libs/vkd3d-shader/hlsl.h @@ -378,6 +378,10 @@ enum hlsl_resource_load_type { HLSL_RESOURCE_LOAD, HLSL_RESOURCE_SAMPLE, + HLSL_RESOURCE_GATHER_RED, + HLSL_RESOURCE_GATHER_GREEN, + HLSL_RESOURCE_GATHER_BLUE, + HLSL_RESOURCE_GATHER_ALPHA, }; struct hlsl_ir_resource_load diff --git a/libs/vkd3d-shader/hlsl.y b/libs/vkd3d-shader/hlsl.y index 3a0ddb93..a321789c 100644 --- a/libs/vkd3d-shader/hlsl.y +++ b/libs/vkd3d-shader/hlsl.y @@ -1930,6 +1930,112 @@ static bool add_method_call(struct hlsl_ctx *ctx, struct list *instrs, struct hl list_add_tail(instrs, &load->node.entry); return true; } + else if (!strcmp(name, "Gather") || !strcmp(name, "GatherRed") || !strcmp(name, "GatherBlue") + || !strcmp(name, "GatherGreen") || !strcmp(name, "GatherAlpha")) + { + const unsigned int sampler_dim = sampler_dim_count(object_type->sampler_dim); + enum hlsl_resource_load_type load_type; + const struct hlsl_type *sampler_type; + struct hlsl_ir_resource_load *load; + struct hlsl_ir_node *offset = NULL; + struct hlsl_ir_load *sampler_load; + struct hlsl_type *result_type; + struct hlsl_ir_node *coords; + int status_out_arg; + int read_channel; + + if (!strcmp(name, "GatherGreen")) + { + load_type = HLSL_RESOURCE_GATHER_GREEN; + read_channel = 1; + } + else if (!strcmp(name, "GatherBlue")) + { + load_type = HLSL_RESOURCE_GATHER_BLUE; + read_channel = 2; + } + else if (!strcmp(name, "GatherAlpha")) + { + load_type = HLSL_RESOURCE_GATHER_ALPHA; + read_channel = 3; + } + else + { + load_type = HLSL_RESOURCE_GATHER_RED; + read_channel = 0; + } + + + if (!strcmp(name, "Gather")) + { + if (params->args_count != 2 && params->args_count != 3) + { + hlsl_error(ctx, loc, VKD3D_SHADER_ERROR_HLSL_WRONG_PARAMETER_COUNT, + "Wrong number of arguments to method 'Gather': expected 2 or 3, but got %u.", params->args_count); + return false; + } + } + else if (params->args_count < 2 || params->args_count == 5 || params->args_count > 7) + { + hlsl_error(ctx, loc, VKD3D_SHADER_ERROR_HLSL_WRONG_PARAMETER_COUNT, + "Wrong number of arguments to method '%s': expected 2,3,4,6 or 7, but got %u.", name, params->args_count); + return false; + } + + status_out_arg = -1; + if (params->args_count == 4) + status_out_arg = 3; + if (params->args_count == 7) + status_out_arg = 6; + if (status_out_arg != -1) + hlsl_fixme(ctx, loc, "Ignoring 'status' output parameter.\n"); + + if (params->args_count == 6 || params->args_count == 7) + hlsl_fixme(ctx, loc, "Ignoring multiple offset parameters.\n"); + + if (params->args_count == 3 || params->args_count == 4) + { + if (!(offset = add_implicit_conversion(ctx, instrs, params->args[2], + hlsl_get_vector_type(ctx, HLSL_TYPE_INT, sampler_dim), loc))) + return false; + } + + sampler_type = params->args[0]->data_type; + if (sampler_type->type != HLSL_CLASS_OBJECT || sampler_type->base_type != HLSL_TYPE_SAMPLER + || sampler_type->sampler_dim != HLSL_SAMPLER_DIM_GENERIC) + { + struct vkd3d_string_buffer *string; + + if ((string = hlsl_type_to_string(ctx, sampler_type))) + hlsl_error(ctx, loc, VKD3D_SHADER_ERROR_HLSL_INVALID_TYPE, + "Wrong type for argument 0 of %s(): expected 'sampler', but got '%s'.", name, string->buffer); + hlsl_release_string_buffer(ctx, string); + return false; + } + + if (read_channel >= object_type->e.resource_format->dimx) + { + hlsl_error(ctx, loc, VKD3D_SHADER_ERROR_HLSL_INVALID_TYPE, + "Method %s() requires at least %d channels.", name, read_channel+1); + return false; + } + + result_type = hlsl_get_vector_type(ctx, object_type->e.resource_format->base_type, 4); + + /* Only HLSL_IR_LOAD can return an object. */ + sampler_load = hlsl_ir_load(params->args[0]); + + if (!(coords = add_implicit_conversion(ctx, instrs, params->args[1], + hlsl_get_vector_type(ctx, HLSL_TYPE_FLOAT, sampler_dim), loc))) + return false; + + if (!(load = hlsl_new_resource_load(ctx, result_type, + load_type, object_load->src.var, object_load->src.offset.node, + sampler_load->src.var, sampler_load->src.offset.node, coords, offset, loc))) + return false; + list_add_tail(instrs, &load->node.entry); + return true; + } else { struct vkd3d_string_buffer *string; diff --git a/libs/vkd3d-shader/hlsl_sm4.c b/libs/vkd3d-shader/hlsl_sm4.c index d55815f1..61247031 100644 --- a/libs/vkd3d-shader/hlsl_sm4.c +++ b/libs/vkd3d-shader/hlsl_sm4.c @@ -1661,6 +1661,50 @@ static void write_sm4_loop(struct hlsl_ctx *ctx, write_sm4_instruction(buffer, &instr); } + +static void write_sm4_gather(struct hlsl_ctx *ctx, struct vkd3d_bytecode_buffer *buffer, + const struct hlsl_type *resource_type, const struct hlsl_ir_node *dst, + const struct hlsl_deref *resource, const struct hlsl_deref *sampler, const struct hlsl_ir_node *coords, + unsigned int swizzle, const struct hlsl_ir_node *texel_offset) +{ + struct sm4_instruction instr; + unsigned int writemask; + + memset(&instr, 0, sizeof(instr)); + + /* TODO: possible optimization, if the offset ranges from -8 to 7, an VKD3D_SM4_OP_GATHER4 with + an aoffimmi modifier can be used. */ + instr.opcode = texel_offset? VKD3D_SM5_OP_GATHER4_PO : VKD3D_SM4_OP_GATHER4; + + sm4_register_from_node(&instr.dsts[0].reg, &instr.dsts[0].writemask, NULL, dst); + instr.dst_count = 1; + + instr.src_count = (instr.opcode == VKD3D_SM5_OP_GATHER4_PO)? 4 : 3; + + sm4_register_from_node(&instr.srcs[0].reg, &writemask, &instr.srcs[0].swizzle_type, coords); + instr.srcs[0].swizzle = hlsl_swizzle_from_writemask(writemask); + + if (texel_offset) + { + sm4_register_from_node(&instr.srcs[1].reg, &writemask, &instr.srcs[1].swizzle_type, texel_offset); + instr.srcs[1].swizzle = hlsl_swizzle_from_writemask(writemask); + } + + sm4_register_from_deref(ctx, &instr.srcs[instr.src_count-2].reg, &writemask, + &instr.srcs[instr.src_count-2].swizzle_type, resource, resource_type); + instr.srcs[instr.src_count-2].swizzle = hlsl_map_swizzle(hlsl_swizzle_from_writemask(writemask), instr.dsts[0].writemask); + + sm4_register_from_deref(ctx, &instr.srcs[instr.src_count-1].reg, &writemask, NULL, + sampler, sampler->var->data_type); + instr.srcs[instr.src_count-1].reg.type = VKD3D_SM4_RT_SAMPLER; + instr.srcs[instr.src_count-1].reg.dim = VKD3D_SM4_DIMENSION_VEC4; + instr.srcs[instr.src_count-1].swizzle_type = VKD3D_SM4_SWIZZLE_SCALAR; + instr.srcs[instr.src_count-1].swizzle = swizzle; + + write_sm4_instruction(buffer, &instr); +} + + static void write_sm4_resource_load(struct hlsl_ctx *ctx, struct vkd3d_bytecode_buffer *buffer, const struct hlsl_ir_resource_load *load) { @@ -1701,6 +1745,23 @@ static void write_sm4_resource_load(struct hlsl_ctx *ctx, write_sm4_sample(ctx, buffer, resource_type, &load->node, &load->resource, &load->sampler, coords, texel_offset); break; + + case HLSL_RESOURCE_GATHER_RED: + write_sm4_gather(ctx, buffer, resource_type, &load->node, &load->resource, + &load->sampler, coords, HLSL_SWIZZLE(X, X, X, X), texel_offset); + break; + case HLSL_RESOURCE_GATHER_GREEN: + write_sm4_gather(ctx, buffer, resource_type, &load->node, &load->resource, + &load->sampler, coords, HLSL_SWIZZLE(Y, Y, Y, Y), texel_offset); + break; + case HLSL_RESOURCE_GATHER_BLUE: + write_sm4_gather(ctx, buffer, resource_type, &load->node, &load->resource, + &load->sampler, coords, HLSL_SWIZZLE(Z, Z, Z, Z), texel_offset); + break; + case HLSL_RESOURCE_GATHER_ALPHA: + write_sm4_gather(ctx, buffer, resource_type, &load->node, &load->resource, + &load->sampler, coords, HLSL_SWIZZLE(W, W, W, W), texel_offset); + break; } } diff --git a/tests/hlsl-gather-offset.shader_test b/tests/hlsl-gather-offset.shader_test new file mode 100644 index 00000000..d7e1a233 --- /dev/null +++ b/tests/hlsl-gather-offset.shader_test @@ -0,0 +1,93 @@ +[sampler 0] +filter linear linear linear +address clamp clamp clamp + +[texture 0] +size (3, 3) +0.0 0.0 0.0 0.4 0.1 0.0 0.5 0.0 0.2 0.0 0.0 0.4 +0.0 0.1 0.5 0.0 0.1 0.1 0.0 0.4 0.2 0.1 0.5 0.0 +0.0 0.2 0.0 0.4 0.1 0.2 0.5 0.0 0.2 0.2 0.0 0.4 + + +[pixel shader] +sampler s; +Texture2D t; + +float4 main() : sv_target +{ + return t.Gather(s, float2(0.2, 0.2), int2(1, 0)); +} + +[test] +draw quad +probe all rgba (0.1, 0.2, 0.2, 0.1) + + +[pixel shader] +sampler s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherRed(s, float2(0.6, 0.6), int2(-1, 0)); +} + +[test] +draw quad +probe all rgba (0.0, 0.1, 0.1, 0.0) + + +[pixel shader] +SamplerState s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherGreen(s, float2(0.2, 0.2), int2(0, 1)); +} + +[test] +draw quad +probe all rgba (0.2, 0.2, 0.1, 0.1) + + +[pixel shader] +SamplerState s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherGreen(s, float2(0.8, 0.8), int2(-1, -1)); +} + +[test] +draw quad +probe all rgba (0.1, 0.1, 0.0, 0.0) + + +[pixel shader] +SamplerState s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherBlue(s, float2(0.2, 0.8), int2(1, 0)); +} + +[test] +draw quad +probe all rgba (0.5, 0.0, 0.5, 0.0) + + +[pixel shader] +SamplerState s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherAlpha(s, float2(0.2, 0.8), int2(0, -1)); +} + +[test] +draw quad +probe all rgba (0.0, 0.4, 0.0, 0.4) diff --git a/tests/hlsl-gather.shader_test b/tests/hlsl-gather.shader_test new file mode 100644 index 00000000..2682c6ef --- /dev/null +++ b/tests/hlsl-gather.shader_test @@ -0,0 +1,107 @@ +[sampler 0] +filter linear linear linear +address clamp clamp clamp + +[texture 0] +size (3, 3) +0.0 0.0 0.0 0.4 0.1 0.0 0.5 0.0 0.2 0.0 0.0 0.4 +0.0 0.1 0.5 0.0 0.1 0.1 0.0 0.4 0.2 0.1 0.5 0.0 +0.0 0.2 0.0 0.4 0.1 0.2 0.5 0.0 0.2 0.2 0.0 0.4 + + +[pixel shader] +sampler s; +Texture2D t; + +float4 main() : sv_target +{ + return t.Gather(s, float2(0.2, 0.2)); +} + +[test] +draw quad +probe all rgba (0.0, 0.1, 0.1, 0.0) + + +[pixel shader] +sampler s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherRed(s, float2(0.6, 0.6), int2(0, 0)); +} + +[test] +draw quad +probe all rgba (0.1, 0.2, 0.2, 0.1) + + +[pixel shader] +SamplerState s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherGreen(s, float2(0.2, 0.2), int2(0, 0)); +} + +[test] +draw quad +probe all rgba (0.1, 0.1, 0.0, 0.0) + + +[pixel shader] +SamplerState s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherGreen(s, float2(0.8, 0.8)); +} + +[test] +draw quad +probe all rgba (0.2, 0.2, 0.1, 0.1) + + +[pixel shader] +SamplerState s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherGreen(s, float2(0.2, 0.2), int2(0, 0), int2(0, 0), int2(0, 0), int2(0, 0)); +} + +[test] +draw quad +probe all rgba (0.1, 0.1, 0.0, 0.0) + + +[pixel shader] +SamplerState s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherBlue(s, float2(0.2, 0.8), int2(0, 0)); +} + +[test] +draw quad +probe all rgba (0.0, 0.5, 0.0, 0.5) + + +[pixel shader] +SamplerState s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherAlpha(s, float2(0.2, 0.8), int2(0, 0)); +} + +[test] +draw quad +probe all rgba (0.4, 0.0, 0.4, 0.0) -- 2.25.1

2 1