[PATCH v7 0/3] MR441: vkd3d-shader/dxil: Implement the DXIL CMP2 instruction.
-- v7: vkd3d-shader/spirv: Introduce orderedness to comparison instructions. vkd3d-shader/dxil: Implement the DXIL CMP2 instruction. vkd3d-shader/spirv: Support bool dst register in spirv_compiler_emit_comparison_instruction(). https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/441
From: Conor McCarthy <cmccarthy(a)codeweavers.com> --- libs/vkd3d-shader/spirv.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/libs/vkd3d-shader/spirv.c b/libs/vkd3d-shader/spirv.c index 35561079a..51cc639e7 100644 --- a/libs/vkd3d-shader/spirv.c +++ b/libs/vkd3d-shader/spirv.c @@ -7387,7 +7387,8 @@ static void spirv_compiler_emit_comparison_instruction(struct spirv_compiler *co result_id = vkd3d_spirv_build_op_tr2(builder, &builder->function_stream, op, type_id, src0_id, src1_id); - result_id = spirv_compiler_emit_bool_to_int(compiler, component_count, result_id, true); + if (dst->reg.data_type != VKD3D_DATA_BOOL) + result_id = spirv_compiler_emit_bool_to_int(compiler, component_count, result_id, true); spirv_compiler_emit_store_reg(compiler, &dst->reg, dst->write_mask, result_id); } -- GitLab https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/441
From: Conor McCarthy <cmccarthy(a)codeweavers.com> --- libs/vkd3d-shader/dxil.c | 170 +++++++++++++++++++++++ libs/vkd3d-shader/vkd3d_shader_private.h | 2 + tests/hlsl/all.shader_test | 14 +- tests/hlsl/any.shader_test | 36 ++--- tests/hlsl/bool-cast.shader_test | 4 +- tests/hlsl/cast-to-float.shader_test | 2 +- tests/hlsl/cast-to-half.shader_test | 2 +- tests/hlsl/cast-to-int.shader_test | 2 +- tests/hlsl/cast-to-uint.shader_test | 2 +- tests/hlsl/sign.shader_test | 20 +-- 10 files changed, 213 insertions(+), 41 deletions(-) diff --git a/libs/vkd3d-shader/dxil.c b/libs/vkd3d-shader/dxil.c index 4614ed18f..1127e9f4d 100644 --- a/libs/vkd3d-shader/dxil.c +++ b/libs/vkd3d-shader/dxil.c @@ -306,6 +306,36 @@ enum dxil_cast_code CAST_ADDRSPACECAST = 12, }; +enum dxil_predicate +{ + FCMP_FALSE = 0, + FCMP_OEQ = 1, + FCMP_OGT = 2, + FCMP_OGE = 3, + FCMP_OLT = 4, + FCMP_OLE = 5, + FCMP_ONE = 6, + FCMP_ORD = 7, + FCMP_UNO = 8, + FCMP_UEQ = 9, + FCMP_UGT = 10, + FCMP_UGE = 11, + FCMP_ULT = 12, + FCMP_ULE = 13, + FCMP_UNE = 14, + FCMP_TRUE = 15, + ICMP_EQ = 32, + ICMP_NE = 33, + ICMP_UGT = 34, + ICMP_UGE = 35, + ICMP_ULT = 36, + ICMP_ULE = 37, + ICMP_SGT = 38, + ICMP_SGE = 39, + ICMP_SLT = 40, + ICMP_SLE = 41, +}; + struct sm6_pointer_info { const struct sm6_type *type; @@ -514,6 +544,7 @@ struct sm6_parser struct sm6_type *types; size_t type_count; + struct sm6_type *bool_type; struct sm6_type *metadata_type; struct sm6_type *handle_type; @@ -1388,6 +1419,8 @@ static enum vkd3d_result sm6_parser_type_table_init(struct sm6_parser *sm6) switch ((width = record->operands[0])) { case 1: + sm6->bool_type = type; + break; case 8: case 16: case 32: @@ -3416,6 +3449,140 @@ static void sm6_parser_emit_cast(struct sm6_parser *sm6, const struct dxil_recor src_param->reg.data_type = dst->u.reg.data_type; } +struct sm6_cmp_info +{ + enum vkd3d_shader_opcode handler_idx; + bool src_swap; +}; + +static const struct sm6_cmp_info *sm6_map_cmp2_op(uint64_t code) +{ + static const struct sm6_cmp_info cmp_op_table[] = + { + [FCMP_FALSE] = {VKD3DSIH_INVALID}, + [FCMP_OEQ] = {VKD3DSIH_EQ}, + [FCMP_OGT] = {VKD3DSIH_LT, true}, + [FCMP_OGE] = {VKD3DSIH_GE}, + [FCMP_OLT] = {VKD3DSIH_LT}, + [FCMP_OLE] = {VKD3DSIH_GE, true}, + [FCMP_ONE] = {VKD3DSIH_NE}, + [FCMP_ORD] = {VKD3DSIH_INVALID}, + [FCMP_UNO] = {VKD3DSIH_INVALID}, + [FCMP_UEQ] = {VKD3DSIH_EQ}, + [FCMP_UGT] = {VKD3DSIH_LT, true}, + [FCMP_UGE] = {VKD3DSIH_GE}, + [FCMP_ULT] = {VKD3DSIH_LT}, + [FCMP_ULE] = {VKD3DSIH_GE, true}, + [FCMP_UNE] = {VKD3DSIH_NE}, + [FCMP_TRUE] = {VKD3DSIH_INVALID}, + + [ICMP_EQ] = {VKD3DSIH_IEQ}, + [ICMP_NE] = {VKD3DSIH_INE}, + [ICMP_UGT] = {VKD3DSIH_ULT, true}, + [ICMP_UGE] = {VKD3DSIH_UGE}, + [ICMP_ULT] = {VKD3DSIH_ULT}, + [ICMP_ULE] = {VKD3DSIH_UGE, true}, + [ICMP_SGT] = {VKD3DSIH_ILT, true}, + [ICMP_SGE] = {VKD3DSIH_IGE}, + [ICMP_SLT] = {VKD3DSIH_ILT}, + [ICMP_SLE] = {VKD3DSIH_IGE, true}, + }; + + return (code < ARRAY_SIZE(cmp_op_table)) ? &cmp_op_table[code] : NULL; +} + +static void sm6_parser_emit_cmp2(struct sm6_parser *sm6, const struct dxil_record *record, + struct vkd3d_shader_instruction *ins, struct sm6_value *dst) +{ + struct vkd3d_shader_src_param *src_params; + const struct sm6_type *type_a, *type_b; + const struct sm6_cmp_info *cmp; + const struct sm6_value *a, *b; + unsigned int i = 0; + bool is_int, is_fp; + uint64_t code; + + if (!(dst->type = sm6->bool_type)) + { + WARN("Bool type not found.\n"); + vkd3d_shader_parser_error(&sm6->p, VKD3D_SHADER_ERROR_DXIL_INVALID_MODULE, + "Module does not define a boolean type for comparison results."); + return; + } + + a = sm6_parser_get_value_by_ref(sm6, record, NULL, &i); + b = sm6_parser_get_value_by_ref(sm6, record, a->type, &i); + if (!a || !b) + return; + + if (!dxil_record_validate_operand_count(record, i + 1, i + 2, sm6)) + return; + + type_a = a->type; + type_b = b->type; + is_int = sm6_type_is_bool_i16_i32_i64(type_a); + is_fp = sm6_type_is_floating_point(type_a); + + code = record->operands[i++]; + + if ((!is_int && !is_fp) || is_int != (code >= ICMP_EQ)) + { + FIXME("Invalid operation %"PRIu64" on type class %u.\n", code, type_a->class); + vkd3d_shader_parser_error(&sm6->p, VKD3D_SHADER_ERROR_DXIL_INVALID_OPERAND, + "Comparison operation %"PRIu64" on type class %u is invalid.", code, type_a->class); + return; + } + + if (type_a != type_b) + { + WARN("Type mismatch, type %u width %u vs type %u width %u.\n", type_a->class, + type_a->u.width, type_b->class, type_b->u.width); + vkd3d_shader_parser_warning(&sm6->p, VKD3D_SHADER_WARNING_DXIL_TYPE_MISMATCH, + "Type mismatch in comparison operation arguments."); + } + + if (!(cmp = sm6_map_cmp2_op(code)) || !cmp->handler_idx || cmp->handler_idx == VKD3DSIH_INVALID) + { + FIXME("Unhandled operation %"PRIu64".\n", code); + vkd3d_shader_parser_error(&sm6->p, VKD3D_SHADER_ERROR_DXIL_INVALID_OPERAND, + "Comparison operation %"PRIu64" is unhandled.", code); + return; + } + + vsir_instruction_init(ins, &sm6->p.location, cmp->handler_idx); + + if (record->operand_count > i) + { + uint64_t flags = record->operands[i]; + bool silence_warning = false; + + if (is_fp) + { + if (!(flags & FP_ALLOW_UNSAFE_ALGEBRA)) + ins->flags |= VKD3DSI_PRECISE_X; + flags &= ~FP_ALLOW_UNSAFE_ALGEBRA; + /* SPIR-V FPFastMathMode is only available in the Kernel executon model. */ + silence_warning = !(flags & ~(FP_NO_NAN | FP_NO_INF | FP_NO_SIGNED_ZEROS | FP_ALLOW_RECIPROCAL)); + } + if (flags && silence_warning) + { + TRACE("Ignoring fast FP modifier %#"PRIx64".\n", flags); + } + else if (flags) + { + WARN("Ignoring flags %#"PRIx64".\n", flags); + vkd3d_shader_parser_warning(&sm6->p, VKD3D_SHADER_WARNING_DXIL_IGNORING_OPERANDS, + "Ignoring flags %#"PRIx64" for a comparison operation.", flags); + } + } + + src_params = instruction_src_params_alloc(ins, 2, sm6); + src_param_init_from_value(&src_params[0 ^ cmp->src_swap], a); + src_param_init_from_value(&src_params[1 ^ cmp->src_swap], b); + + instruction_dst_param_init_ssa_scalar(ins, sm6); +} + static void sm6_parser_emit_extractval(struct sm6_parser *sm6, const struct dxil_record *record, struct vkd3d_shader_instruction *ins, struct sm6_value *dst) { @@ -3635,6 +3802,9 @@ static enum vkd3d_result sm6_parser_function_init(struct sm6_parser *sm6, const case FUNC_CODE_INST_CAST: sm6_parser_emit_cast(sm6, record, ins, dst); break; + case FUNC_CODE_INST_CMP2: + sm6_parser_emit_cmp2(sm6, record, ins, dst); + break; case FUNC_CODE_INST_EXTRACTVAL: sm6_parser_emit_extractval(sm6, record, ins, dst); break; diff --git a/libs/vkd3d-shader/vkd3d_shader_private.h b/libs/vkd3d-shader/vkd3d_shader_private.h index 30fda1362..a38aaf3ef 100644 --- a/libs/vkd3d-shader/vkd3d_shader_private.h +++ b/libs/vkd3d-shader/vkd3d_shader_private.h @@ -1411,6 +1411,8 @@ static inline enum vkd3d_shader_component_type vkd3d_component_type_from_data_ty return VKD3D_SHADER_COMPONENT_INT; case VKD3D_DATA_DOUBLE: return VKD3D_SHADER_COMPONENT_DOUBLE; + case VKD3D_DATA_BOOL: + return VKD3D_SHADER_COMPONENT_BOOL; default: FIXME("Unhandled data type %#x.\n", data_type); /* fall-through */ diff --git a/tests/hlsl/all.shader_test b/tests/hlsl/all.shader_test index 564cc5b00..7bdb0dc82 100644 --- a/tests/hlsl/all.shader_test +++ b/tests/hlsl/all.shader_test @@ -11,17 +11,17 @@ float4 main() : sv_target [test] uniform 0 float4 -1.1 1.6 1.3 0.5 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) [test] uniform 0 float4 0.0 1.6 1.3 0.5 -todo(sm>=6) draw quad +draw quad probe all rgba (0.0, 0.0, 0.0, 0.0) [test] uniform 0 float4 1.0 0.0 1.3 0.5 -todo(sm>=6) draw quad +draw quad probe all rgba (0.0, 0.0, 0.0, 0.0) [pixel shader] @@ -34,12 +34,12 @@ float4 main() : sv_target [test] uniform 0 float4 1.0 0.0 0.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) [test] uniform 0 float4 0.0 0.0 0.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (0.0, 0.0, 0.0, 0.0) [pixel shader] @@ -53,11 +53,11 @@ float4 main() : sv_target [test] uniform 0 float4 1.0 2.0 0.0 0.0 uniform 4 float4 3.0 4.0 0.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) [test] uniform 0 float4 1.0 2.0 0.0 0.0 uniform 4 float4 0.0 4.0 0.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (0.0, 0.0, 0.0, 0.0) diff --git a/tests/hlsl/any.shader_test b/tests/hlsl/any.shader_test index 9b8b922c9..f2298d3a3 100644 --- a/tests/hlsl/any.shader_test +++ b/tests/hlsl/any.shader_test @@ -8,25 +8,25 @@ float4 main() : sv_target [test] uniform 0 float4 1.0 1.0 1.0 1.0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) uniform 0 float4 1.0 0.0 0.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) uniform 0 float4 0.0 1.0 0.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) uniform 0 float4 0.0 0.0 1.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) uniform 0 float4 0.0 0.0 0.0 1.0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) uniform 0 float4 0.0 0.0 0.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (0.0, 0.0, 0.0, 0.0) uniform 0 float4 -1.0 -1.0 -1.0 -1.0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) [pixel shader] @@ -39,13 +39,13 @@ float4 main() : sv_target [test] uniform 0 float4 1.0 0.0 0.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) uniform 0 float4 0.0 0.0 0.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (0.0, 0.0, 0.0, 0.0) uniform 0 float4 -1.0 0.0 0.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) [require] @@ -61,22 +61,22 @@ float4 main() : sv_target [test] uniform 0 uint4 1 1 1 1 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) uniform 0 uint4 1 0 0 0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) uniform 0 uint4 0 1 0 0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) uniform 0 uint4 0 0 1 0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) uniform 0 uint4 0 0 0 1 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) uniform 0 uint4 0 0 0 0 -todo(sm>=6) draw quad +draw quad probe all rgba (0.0, 0.0, 0.0, 0.0) [pixel shader] @@ -89,8 +89,8 @@ float4 main() : sv_target [test] uniform 0 uint4 1 0 0 0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) uniform 0 uint4 0 0 0 0 -todo(sm>=6) draw quad +draw quad probe all rgba (0.0, 0.0, 0.0, 0.0) diff --git a/tests/hlsl/bool-cast.shader_test b/tests/hlsl/bool-cast.shader_test index dc75ea376..09ca12e2b 100644 --- a/tests/hlsl/bool-cast.shader_test +++ b/tests/hlsl/bool-cast.shader_test @@ -30,7 +30,7 @@ float4 main() : SV_TARGET [test] uniform 0 float4 0.0 0.0 2.0 4.0 uniform 4 int4 0 1 0 10 -todo(sm>=6) draw quad +draw quad probe all rgba (0.0, 10.0, 1.0, 11.0) @@ -44,5 +44,5 @@ float4 main() : sv_target [test] uniform 0 uint4 0x00000001 0x00000002 0x80000000 0x00000000 -todo(sm>=6) draw quad +draw quad probe all rgba (2.0, 2.0, 2.0, 0.0) diff --git a/tests/hlsl/cast-to-float.shader_test b/tests/hlsl/cast-to-float.shader_test index 7c32acfa3..caaf98c02 100644 --- a/tests/hlsl/cast-to-float.shader_test +++ b/tests/hlsl/cast-to-float.shader_test @@ -17,7 +17,7 @@ uniform 0 int -1 uniform 1 uint 3 uniform 2 int -2 uniform 3 float 0.5 -todo(sm>=6) draw quad +draw quad probe all rgba (0.5, 0.5, 0.5, 0.5) [pixel shader] diff --git a/tests/hlsl/cast-to-half.shader_test b/tests/hlsl/cast-to-half.shader_test index 1579c63f3..b8feb6760 100644 --- a/tests/hlsl/cast-to-half.shader_test +++ b/tests/hlsl/cast-to-half.shader_test @@ -17,7 +17,7 @@ uniform 0 int -1 uniform 1 uint 3 uniform 2 int -2 uniform 3 float 0.5 -todo(sm>=6) draw quad +draw quad probe all rgba (0.5, 0.5, 0.5, 0.5) [pixel shader] diff --git a/tests/hlsl/cast-to-int.shader_test b/tests/hlsl/cast-to-int.shader_test index ae566a039..3e850fb5b 100644 --- a/tests/hlsl/cast-to-int.shader_test +++ b/tests/hlsl/cast-to-int.shader_test @@ -23,7 +23,7 @@ uniform 0 float 2.6 uniform 1 int -2 uniform 2 int -2 uniform 3 float -3.6 -todo(sm>=6) draw quad +draw quad probe all rgba (0.5, 0.5, 0.5, 0.5) [pixel shader] diff --git a/tests/hlsl/cast-to-uint.shader_test b/tests/hlsl/cast-to-uint.shader_test index dece99b75..07479984a 100644 --- a/tests/hlsl/cast-to-uint.shader_test +++ b/tests/hlsl/cast-to-uint.shader_test @@ -23,7 +23,7 @@ uniform 0 float 2.6 uniform 1 int 2 uniform 2 int -2 uniform 3 float -3.6 -todo(sm>=6) draw quad +draw quad probe all rgba (0.5, 0.5, 0.5, 0.5) [pixel shader] diff --git a/tests/hlsl/sign.shader_test b/tests/hlsl/sign.shader_test index 54b953681..5d8b43168 100644 --- a/tests/hlsl/sign.shader_test +++ b/tests/hlsl/sign.shader_test @@ -8,13 +8,13 @@ float4 main() : sv_target [test] uniform 0 float4 1.0 0.0 0.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) uniform 0 float4 -1.0 0.0 0.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (-1.0, -1.0, -1.0, -1.0) uniform 0 float4 0.0 0.0 0.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (0.0, 0.0, 0.0, 0.0) [pixel shader] @@ -27,7 +27,7 @@ float4 main() : sv_target [test] uniform 0 float4 1.0 2.0 3.0 4.0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) [pixel shader] @@ -41,7 +41,7 @@ float4 main() : sv_target [test] uniform 0 float4 1.0 2.0 0.0 0.0 uniform 4 float4 3.0 4.0 0.0 0.0 -todo(sm>=6) draw quad +draw quad probe all rgba (1.0, 1.0, 1.0, 1.0) [require] @@ -58,13 +58,13 @@ float4 main() : sv_target [test] uniform 0 int4 1 0 0 0 -todo(sm>=6) draw quad +draw quad probe all rgba (1, 1, 1, 1) uniform 0 int4 -1 0 0 0 -todo(sm>=6) draw quad +draw quad probe all rgba (-1, -1, -1, -1) uniform 0 int4 0 0 0 0 -todo(sm>=6) draw quad +draw quad probe all rgba (0, 0, 0, 0) [pixel shader] @@ -77,7 +77,7 @@ float4 main() : sv_target [test] uniform 0 int4 1 2 3 4 -todo(sm>=6) draw quad +draw quad probe all rgba (1, 1, 1, 1) [pixel shader] @@ -91,5 +91,5 @@ float4 main() : sv_target [test] uniform 0 int4 1 2 0 0 uniform 4 int4 3 4 0 0 -todo(sm>=6) draw quad +draw quad probe all rgba (1, 1, 1, 1) -- GitLab https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/441
From: Conor McCarthy <cmccarthy(a)codeweavers.com> --- libs/vkd3d-shader/d3d_asm.c | 16 +++++++----- libs/vkd3d-shader/dxil.c | 24 +++++++++--------- libs/vkd3d-shader/spirv.c | 32 +++++++++++++++--------- libs/vkd3d-shader/tpf.c | 12 ++++----- libs/vkd3d-shader/vkd3d_shader_private.h | 16 +++++++----- 5 files changed, 58 insertions(+), 42 deletions(-) diff --git a/libs/vkd3d-shader/d3d_asm.c b/libs/vkd3d-shader/d3d_asm.c index 0a9561d05..442c1e414 100644 --- a/libs/vkd3d-shader/d3d_asm.c +++ b/libs/vkd3d-shader/d3d_asm.c @@ -106,9 +106,9 @@ static const char * const shader_opcode_names[] = [VKD3DSIH_DEFAULT ] = "default", [VKD3DSIH_DEFB ] = "defb", [VKD3DSIH_DEFI ] = "defi", - [VKD3DSIH_DEQ ] = "deq", + [VKD3DSIH_DEQO ] = "deq", [VKD3DSIH_DFMA ] = "dfma", - [VKD3DSIH_DGE ] = "dge", + [VKD3DSIH_DGEO ] = "dge", [VKD3DSIH_DISCARD ] = "discard", [VKD3DSIH_DIV ] = "div", [VKD3DSIH_DLT ] = "dlt", @@ -140,7 +140,8 @@ static const char * const shader_opcode_names[] = [VKD3DSIH_ENDLOOP ] = "endloop", [VKD3DSIH_ENDREP ] = "endrep", [VKD3DSIH_ENDSWITCH ] = "endswitch", - [VKD3DSIH_EQ ] = "eq", + [VKD3DSIH_EQO ] = "eq", + [VKD3DSIH_EQU ] = "eq_unord", [VKD3DSIH_EVAL_CENTROID ] = "eval_centroid", [VKD3DSIH_EVAL_SAMPLE_INDEX ] = "eval_sample_index", [VKD3DSIH_EXP ] = "exp", @@ -164,7 +165,8 @@ static const char * const shader_opcode_names[] = [VKD3DSIH_GATHER4_PO_C_S ] = "gather4_po_c_s", [VKD3DSIH_GATHER4_PO_S ] = "gather4_po_s", [VKD3DSIH_GATHER4_S ] = "gather4_s", - [VKD3DSIH_GE ] = "ge", + [VKD3DSIH_GEO ] = "ge", + [VKD3DSIH_GEU ] = "ge_unord", [VKD3DSIH_HS_CONTROL_POINT_PHASE ] = "hs_control_point_phase", [VKD3DSIH_HS_DECLS ] = "hs_decls", [VKD3DSIH_HS_FORK_PHASE ] = "hs_fork_phase", @@ -217,7 +219,8 @@ static const char * const shader_opcode_names[] = [VKD3DSIH_LOGP ] = "logp", [VKD3DSIH_LOOP ] = "loop", [VKD3DSIH_LRP ] = "lrp", - [VKD3DSIH_LT ] = "lt", + [VKD3DSIH_LTO ] = "lt", + [VKD3DSIH_LTU ] = "lt_unord", [VKD3DSIH_M3x2 ] = "m3x2", [VKD3DSIH_M3x3 ] = "m3x3", [VKD3DSIH_M3x4 ] = "m3x4", @@ -231,7 +234,8 @@ static const char * const shader_opcode_names[] = [VKD3DSIH_MOVC ] = "movc", [VKD3DSIH_MSAD ] = "msad", [VKD3DSIH_MUL ] = "mul", - [VKD3DSIH_NE ] = "ne", + [VKD3DSIH_NEO ] = "ne_ord", + [VKD3DSIH_NEU ] = "ne", [VKD3DSIH_NOP ] = "nop", [VKD3DSIH_NOT ] = "not", [VKD3DSIH_NRM ] = "nrm", diff --git a/libs/vkd3d-shader/dxil.c b/libs/vkd3d-shader/dxil.c index 1127e9f4d..2174ba52c 100644 --- a/libs/vkd3d-shader/dxil.c +++ b/libs/vkd3d-shader/dxil.c @@ -3460,20 +3460,20 @@ static const struct sm6_cmp_info *sm6_map_cmp2_op(uint64_t code) static const struct sm6_cmp_info cmp_op_table[] = { [FCMP_FALSE] = {VKD3DSIH_INVALID}, - [FCMP_OEQ] = {VKD3DSIH_EQ}, - [FCMP_OGT] = {VKD3DSIH_LT, true}, - [FCMP_OGE] = {VKD3DSIH_GE}, - [FCMP_OLT] = {VKD3DSIH_LT}, - [FCMP_OLE] = {VKD3DSIH_GE, true}, - [FCMP_ONE] = {VKD3DSIH_NE}, + [FCMP_OEQ] = {VKD3DSIH_EQO}, + [FCMP_OGT] = {VKD3DSIH_LTO, true}, + [FCMP_OGE] = {VKD3DSIH_GEO}, + [FCMP_OLT] = {VKD3DSIH_LTO}, + [FCMP_OLE] = {VKD3DSIH_GEO, true}, + [FCMP_ONE] = {VKD3DSIH_NEO}, [FCMP_ORD] = {VKD3DSIH_INVALID}, [FCMP_UNO] = {VKD3DSIH_INVALID}, - [FCMP_UEQ] = {VKD3DSIH_EQ}, - [FCMP_UGT] = {VKD3DSIH_LT, true}, - [FCMP_UGE] = {VKD3DSIH_GE}, - [FCMP_ULT] = {VKD3DSIH_LT}, - [FCMP_ULE] = {VKD3DSIH_GE, true}, - [FCMP_UNE] = {VKD3DSIH_NE}, + [FCMP_UEQ] = {VKD3DSIH_EQU}, + [FCMP_UGT] = {VKD3DSIH_LTU, true}, + [FCMP_UGE] = {VKD3DSIH_GEU}, + [FCMP_ULT] = {VKD3DSIH_LTU}, + [FCMP_ULE] = {VKD3DSIH_GEU, true}, + [FCMP_UNE] = {VKD3DSIH_NEU}, [FCMP_TRUE] = {VKD3DSIH_INVALID}, [ICMP_EQ] = {VKD3DSIH_IEQ}, diff --git a/libs/vkd3d-shader/spirv.c b/libs/vkd3d-shader/spirv.c index 51cc639e7..d2621ffa1 100644 --- a/libs/vkd3d-shader/spirv.c +++ b/libs/vkd3d-shader/spirv.c @@ -7359,18 +7359,22 @@ static void spirv_compiler_emit_comparison_instruction(struct spirv_compiler *co switch (instruction->handler_idx) { - case VKD3DSIH_DEQ: - case VKD3DSIH_EQ: op = SpvOpFOrdEqual; break; - case VKD3DSIH_DGE: - case VKD3DSIH_GE: op = SpvOpFOrdGreaterThanEqual; break; + case VKD3DSIH_DEQO: + case VKD3DSIH_EQO: op = SpvOpFOrdEqual; break; + case VKD3DSIH_EQU: op = SpvOpFUnordEqual; break; + case VKD3DSIH_DGEO: + case VKD3DSIH_GEO: op = SpvOpFOrdGreaterThanEqual; break; + case VKD3DSIH_GEU: op = SpvOpFUnordGreaterThanEqual; break; case VKD3DSIH_IEQ: op = SpvOpIEqual; break; case VKD3DSIH_IGE: op = SpvOpSGreaterThanEqual; break; case VKD3DSIH_ILT: op = SpvOpSLessThan; break; case VKD3DSIH_INE: op = SpvOpINotEqual; break; case VKD3DSIH_DLT: - case VKD3DSIH_LT: op = SpvOpFOrdLessThan; break; + case VKD3DSIH_LTO: op = SpvOpFOrdLessThan; break; + case VKD3DSIH_LTU: op = SpvOpFUnordLessThan; break; + case VKD3DSIH_NEO: op = SpvOpFOrdNotEqual; break; case VKD3DSIH_DNE: - case VKD3DSIH_NE: op = SpvOpFUnordNotEqual; break; + case VKD3DSIH_NEU: op = SpvOpFUnordNotEqual; break; case VKD3DSIH_UGE: op = SpvOpUGreaterThanEqual; break; case VKD3DSIH_ULT: op = SpvOpULessThan; break; default: @@ -9559,18 +9563,22 @@ static int spirv_compiler_handle_instruction(struct spirv_compiler *compiler, case VKD3DSIH_FTOU: spirv_compiler_emit_ftou(compiler, instruction); break; - case VKD3DSIH_DEQ: - case VKD3DSIH_DGE: + case VKD3DSIH_DEQO: + case VKD3DSIH_DGEO: case VKD3DSIH_DLT: case VKD3DSIH_DNE: - case VKD3DSIH_EQ: - case VKD3DSIH_GE: + case VKD3DSIH_EQO: + case VKD3DSIH_EQU: + case VKD3DSIH_GEO: + case VKD3DSIH_GEU: case VKD3DSIH_IEQ: case VKD3DSIH_IGE: case VKD3DSIH_ILT: case VKD3DSIH_INE: - case VKD3DSIH_LT: - case VKD3DSIH_NE: + case VKD3DSIH_LTO: + case VKD3DSIH_LTU: + case VKD3DSIH_NEO: + case VKD3DSIH_NEU: case VKD3DSIH_UGE: case VKD3DSIH_ULT: spirv_compiler_emit_comparison_instruction(compiler, instruction); diff --git a/libs/vkd3d-shader/tpf.c b/libs/vkd3d-shader/tpf.c index 3c8096d59..c7c4a77b8 100644 --- a/libs/vkd3d-shader/tpf.c +++ b/libs/vkd3d-shader/tpf.c @@ -1345,12 +1345,12 @@ static void init_sm4_lookup_tables(struct vkd3d_sm4_lookup_tables *lookup) {VKD3D_SM4_OP_ENDIF, VKD3DSIH_ENDIF, "", ""}, {VKD3D_SM4_OP_ENDLOOP, VKD3DSIH_ENDLOOP, "", ""}, {VKD3D_SM4_OP_ENDSWITCH, VKD3DSIH_ENDSWITCH, "", ""}, - {VKD3D_SM4_OP_EQ, VKD3DSIH_EQ, "u", "ff"}, + {VKD3D_SM4_OP_EQ, VKD3DSIH_EQO, "u", "ff"}, {VKD3D_SM4_OP_EXP, VKD3DSIH_EXP, "f", "f"}, {VKD3D_SM4_OP_FRC, VKD3DSIH_FRC, "f", "f"}, {VKD3D_SM4_OP_FTOI, VKD3DSIH_FTOI, "i", "f"}, {VKD3D_SM4_OP_FTOU, VKD3DSIH_FTOU, "u", "f"}, - {VKD3D_SM4_OP_GE, VKD3DSIH_GE, "u", "ff"}, + {VKD3D_SM4_OP_GE, VKD3DSIH_GEO, "u", "ff"}, {VKD3D_SM4_OP_IADD, VKD3DSIH_IADD, "i", "ii"}, {VKD3D_SM4_OP_IF, VKD3DSIH_IF, "", "u", shader_sm4_read_conditional_op}, @@ -1371,7 +1371,7 @@ static void init_sm4_lookup_tables(struct vkd3d_sm4_lookup_tables *lookup) {VKD3D_SM4_OP_LD2DMS, VKD3DSIH_LD2DMS, "u", "iRi"}, {VKD3D_SM4_OP_LOG, VKD3DSIH_LOG, "f", "f"}, {VKD3D_SM4_OP_LOOP, VKD3DSIH_LOOP, "", ""}, - {VKD3D_SM4_OP_LT, VKD3DSIH_LT, "u", "ff"}, + {VKD3D_SM4_OP_LT, VKD3DSIH_LTO, "u", "ff"}, {VKD3D_SM4_OP_MAD, VKD3DSIH_MAD, "f", "fff"}, {VKD3D_SM4_OP_MIN, VKD3DSIH_MIN, "f", "ff"}, {VKD3D_SM4_OP_MAX, VKD3DSIH_MAX, "f", "ff"}, @@ -1380,7 +1380,7 @@ static void init_sm4_lookup_tables(struct vkd3d_sm4_lookup_tables *lookup) {VKD3D_SM4_OP_MOV, VKD3DSIH_MOV, "f", "f"}, {VKD3D_SM4_OP_MOVC, VKD3DSIH_MOVC, "f", "uff"}, {VKD3D_SM4_OP_MUL, VKD3DSIH_MUL, "f", "ff"}, - {VKD3D_SM4_OP_NE, VKD3DSIH_NE, "u", "ff"}, + {VKD3D_SM4_OP_NE, VKD3DSIH_NEU, "u", "ff"}, {VKD3D_SM4_OP_NOP, VKD3DSIH_NOP, "", ""}, {VKD3D_SM4_OP_NOT, VKD3DSIH_NOT, "u", "u"}, {VKD3D_SM4_OP_OR, VKD3DSIH_OR, "u", "uu"}, @@ -1551,8 +1551,8 @@ static void init_sm4_lookup_tables(struct vkd3d_sm4_lookup_tables *lookup) {VKD3D_SM5_OP_DMAX, VKD3DSIH_DMAX, "d", "dd"}, {VKD3D_SM5_OP_DMIN, VKD3DSIH_DMIN, "d", "dd"}, {VKD3D_SM5_OP_DMUL, VKD3DSIH_DMUL, "d", "dd"}, - {VKD3D_SM5_OP_DEQ, VKD3DSIH_DEQ, "u", "dd"}, - {VKD3D_SM5_OP_DGE, VKD3DSIH_DGE, "u", "dd"}, + {VKD3D_SM5_OP_DEQ, VKD3DSIH_DEQO, "u", "dd"}, + {VKD3D_SM5_OP_DGE, VKD3DSIH_DGEO, "u", "dd"}, {VKD3D_SM5_OP_DLT, VKD3DSIH_DLT, "u", "dd"}, {VKD3D_SM5_OP_DNE, VKD3DSIH_DNE, "u", "dd"}, {VKD3D_SM5_OP_DMOV, VKD3DSIH_DMOV, "d", "d"}, diff --git a/libs/vkd3d-shader/vkd3d_shader_private.h b/libs/vkd3d-shader/vkd3d_shader_private.h index a38aaf3ef..e5d590637 100644 --- a/libs/vkd3d-shader/vkd3d_shader_private.h +++ b/libs/vkd3d-shader/vkd3d_shader_private.h @@ -294,9 +294,9 @@ enum vkd3d_shader_opcode VKD3DSIH_DEFAULT, VKD3DSIH_DEFB, VKD3DSIH_DEFI, - VKD3DSIH_DEQ, + VKD3DSIH_DEQO, VKD3DSIH_DFMA, - VKD3DSIH_DGE, + VKD3DSIH_DGEO, VKD3DSIH_DISCARD, VKD3DSIH_DIV, VKD3DSIH_DLT, @@ -328,7 +328,8 @@ enum vkd3d_shader_opcode VKD3DSIH_ENDLOOP, VKD3DSIH_ENDREP, VKD3DSIH_ENDSWITCH, - VKD3DSIH_EQ, + VKD3DSIH_EQO, + VKD3DSIH_EQU, VKD3DSIH_EVAL_CENTROID, VKD3DSIH_EVAL_SAMPLE_INDEX, VKD3DSIH_EXP, @@ -352,7 +353,8 @@ enum vkd3d_shader_opcode VKD3DSIH_GATHER4_PO_C_S, VKD3DSIH_GATHER4_PO_S, VKD3DSIH_GATHER4_S, - VKD3DSIH_GE, + VKD3DSIH_GEO, + VKD3DSIH_GEU, VKD3DSIH_HS_CONTROL_POINT_PHASE, VKD3DSIH_HS_DECLS, VKD3DSIH_HS_FORK_PHASE, @@ -405,7 +407,8 @@ enum vkd3d_shader_opcode VKD3DSIH_LOGP, VKD3DSIH_LOOP, VKD3DSIH_LRP, - VKD3DSIH_LT, + VKD3DSIH_LTO, + VKD3DSIH_LTU, VKD3DSIH_M3x2, VKD3DSIH_M3x3, VKD3DSIH_M3x4, @@ -419,7 +422,8 @@ enum vkd3d_shader_opcode VKD3DSIH_MOVC, VKD3DSIH_MSAD, VKD3DSIH_MUL, - VKD3DSIH_NE, + VKD3DSIH_NEO, + VKD3DSIH_NEU, VKD3DSIH_NOP, VKD3DSIH_NOT, VKD3DSIH_NRM, -- GitLab https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/441
This merge request was approved by Henri Verbeet. -- https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/441
participants (3)
-
Conor McCarthy -
Conor McCarthy (＠cmccarthy)
-
Henri Verbeet (＠hverbeet)