XenosRecomp/XenosRecomp/shader_recompiler.cpp
Isaac Marovitz 995d2889d4
Fix missing newline
Signed-off-by: Isaac Marovitz <isaacryu@icloud.com>
2025-03-20 22:30:05 -04:00

2026 lines
60 KiB
C++

#include "shader_recompiler.h"
#include "shader_common.h"
static constexpr char SWIZZLES[] =
{
'x',
'y',
'z',
'w',
'0',
'1',
'_',
'_'
};
static constexpr const char* USAGE_TYPES[] =
{
"float4", // POSITION
"float4", // BLENDWEIGHT
"uint4", // BLENDINDICES
"uint4", // NORMAL
"float4", // PSIZE
"float4", // TEXCOORD
"uint4", // TANGENT
"uint4", // BINORMAL
"float4", // TESSFACTOR
"float4", // POSITIONT
"float4", // COLOR
"float4", // FOG
"float4", // DEPTH
"float4", // SAMPLE
};
static constexpr const char* USAGE_VARIABLES[] =
{
"Position",
"BlendWeight",
"BlendIndices",
"Normal",
"PointSize",
"TexCoord",
"Tangent",
"Binormal",
"TessFactor",
"PositionT",
"Color",
"Fog",
"Depth",
"Sample"
};
static constexpr const char* USAGE_SEMANTICS[] =
{
"POSITION",
"BLENDWEIGHT",
"BLENDINDICES",
"NORMAL",
"PSIZE",
"TEXCOORD",
"TANGENT",
"BINORMAL",
"TESSFACTOR",
"POSITIONT",
"COLOR",
"FOG",
"DEPTH",
"SAMPLE"
};
struct DeclUsageLocation
{
DeclUsage usage;
uint32_t usageIndex;
uint32_t location;
};
// NOTE: These are specialized Vulkan locations for Unleashed Recompiled. Change as necessary. Likely not going to work with other games.
static constexpr DeclUsageLocation USAGE_LOCATIONS[] =
{
{ DeclUsage::Position, 0, 0 },
{ DeclUsage::Normal, 0, 1 },
{ DeclUsage::Tangent, 0, 2 },
{ DeclUsage::Binormal, 0, 3 },
{ DeclUsage::TexCoord, 0, 4 },
{ DeclUsage::TexCoord, 1, 5 },
{ DeclUsage::TexCoord, 2, 6 },
{ DeclUsage::TexCoord, 3, 7 },
{ DeclUsage::Color, 0, 8 },
{ DeclUsage::BlendIndices, 0, 9 },
{ DeclUsage::BlendWeight, 0, 10 },
{ DeclUsage::Color, 1, 11 },
{ DeclUsage::TexCoord, 4, 12 },
{ DeclUsage::TexCoord, 5, 13 },
{ DeclUsage::TexCoord, 6, 14 },
{ DeclUsage::TexCoord, 7, 15 },
{ DeclUsage::Position, 1, 15 },
};
static constexpr std::pair<DeclUsage, size_t> INTERPOLATORS[] =
{
{ DeclUsage::TexCoord, 0 },
{ DeclUsage::TexCoord, 1 },
{ DeclUsage::TexCoord, 2 },
{ DeclUsage::TexCoord, 3 },
{ DeclUsage::TexCoord, 4 },
{ DeclUsage::TexCoord, 5 },
{ DeclUsage::TexCoord, 6 },
{ DeclUsage::TexCoord, 7 },
{ DeclUsage::TexCoord, 8 },
{ DeclUsage::TexCoord, 9 },
{ DeclUsage::TexCoord, 10 },
{ DeclUsage::TexCoord, 11 },
{ DeclUsage::TexCoord, 12 },
{ DeclUsage::TexCoord, 13 },
{ DeclUsage::TexCoord, 14 },
{ DeclUsage::TexCoord, 15 },
{ DeclUsage::Color, 0 },
{ DeclUsage::Color, 1 }
};
static constexpr std::string_view TEXTURE_DIMENSIONS[] =
{
"2D",
"3D",
"Cube"
};
static FetchDestinationSwizzle getDestSwizzle(uint32_t dstSwizzle, uint32_t index)
{
return FetchDestinationSwizzle((dstSwizzle >> (index * 3)) & 0x7);
}
void ShaderRecompiler::printDstSwizzle(uint32_t dstSwizzle, bool operand)
{
for (size_t i = 0; i < 4; i++)
{
const auto swizzle = getDestSwizzle(dstSwizzle, i);
if (swizzle >= FetchDestinationSwizzle::X && swizzle <= FetchDestinationSwizzle::W)
out += SWIZZLES[operand ? uint32_t(swizzle) : i];
}
}
void ShaderRecompiler::printDstSwizzle01(uint32_t dstRegister, uint32_t dstSwizzle)
{
for (size_t i = 0; i < 4; i++)
{
const auto swizzle = getDestSwizzle(dstSwizzle, i);
if (swizzle == FetchDestinationSwizzle::Zero)
{
indent();
println("r{}.{} = 0.0;", dstRegister, SWIZZLES[i]);
}
else if (swizzle == FetchDestinationSwizzle::One)
{
indent();
println("r{}.{} = 1.0;", dstRegister, SWIZZLES[i]);
}
}
}
void ShaderRecompiler::recompile(const VertexFetchInstruction& instr, uint32_t address)
{
if (instr.isPredicated)
{
indent();
println("if ({}p0)", instr.predicateCondition ? "" : "!");
indent();
out += "{\n";
++indentation;
}
indent();
print("r{}.", instr.dstRegister);
printDstSwizzle(instr.dstSwizzle, false);
out += " = ";
auto findResult = vertexElements.find(address);
assert(findResult != vertexElements.end());
switch (findResult->second.usage)
{
case DeclUsage::Normal:
case DeclUsage::Tangent:
case DeclUsage::Binormal:
specConstantsMask |= SPEC_CONSTANT_R11G11B10_NORMAL;
print("tfetchR11G11B10(");
break;
case DeclUsage::TexCoord:
print("tfetchTexcoord(g_SwappedTexcoords, ");
break;
}
print("i{}{}", USAGE_VARIABLES[uint32_t(findResult->second.usage)], uint32_t(findResult->second.usageIndex));
switch (findResult->second.usage)
{
case DeclUsage::Normal:
case DeclUsage::Tangent:
case DeclUsage::Binormal:
out += ')';
break;
case DeclUsage::TexCoord:
print(", {})", uint32_t(findResult->second.usageIndex));
break;
}
out += '.';
printDstSwizzle(instr.dstSwizzle, true);
out += ";\n";
printDstSwizzle01(instr.dstRegister, instr.dstSwizzle);
if (instr.isPredicated)
{
--indentation;
indent();
out += "}\n";
}
}
void ShaderRecompiler::recompile(const TextureFetchInstruction& instr, bool bicubic)
{
if (instr.opcode != FetchOpcode::TextureFetch && instr.opcode != FetchOpcode::GetTextureWeights)
return;
if (instr.isPredicated)
{
indent();
println("if ({}p0)", instr.predCondition ? "" : "!");
indent();
out += "{\n";
++indentation;
}
auto printSrcRegister = [&](size_t componentCount)
{
print("r{}.", instr.srcRegister);
for (size_t i = 0; i < componentCount; i++)
out += SWIZZLES[((instr.srcSwizzle >> (i * 2))) & 0x3];
};
std::string constName;
const char* constNamePtr = nullptr;
#ifdef UNLEASHED_RECOMP
bool subtractFromOne = false;
#endif
auto findResult = samplers.find(instr.constIndex);
if (findResult != samplers.end())
{
constNamePtr = findResult->second;
#ifdef UNLEASHED_RECOMP
subtractFromOne = hasMtxPrevInvViewProjection && strcmp(constNamePtr, "sampZBuffer") == 0;
#endif
}
else
{
constName = fmt::format("s{}", instr.constIndex);
constNamePtr = constName.c_str();
}
#ifdef UNLEASHED_RECOMP
if (instr.constIndex == 0 && instr.dimension == TextureDimension::Texture2D)
{
indent();
println("pixelCoord = getPixelCoord(");
println("#ifdef __air__");
indent();
println("g_Texture2DDescriptorHeap,");
println("#endif");
indent();
print("{}_Texture2DDescriptorIndex, ", constNamePtr);
printSrcRegister(2);
out += ");\n";
}
#endif
indent();
print("r{}.", instr.dstRegister);
printDstSwizzle(instr.dstSwizzle, false);
out += " = ";
switch (instr.opcode)
{
case FetchOpcode::TextureFetch:
{
#ifdef UNLEASHED_RECOMP
if (subtractFromOne)
out += "1.0 - ";
#endif
out += "tfetch";
break;
}
case FetchOpcode::GetTextureWeights:
{
out += "getWeights";
break;
}
}
std::string_view dimension;
uint32_t componentCount = 0;
switch (instr.dimension)
{
case TextureDimension::Texture1D:
dimension = "1D";
componentCount = 1;
break;
case TextureDimension::Texture2D:
dimension = "2D";
componentCount = 2;
break;
case TextureDimension::Texture3D:
dimension = "3D";
componentCount = 3;
break;
case TextureDimension::TextureCube:
dimension = "Cube";
componentCount = 3;
break;
}
out += dimension;
#ifdef UNLEASHED_RECOMP
if (bicubic)
out += "Bicubic";
#endif
println("(");
println("#ifdef __air__");
indent();
println("\tg_Texture{}DescriptorHeap,", dimension);
indent();
println("\tg_SamplerDescriptorHeap,");
println("#endif");
indent();
print("\t{0}_Texture{1}DescriptorIndex, {0}_SamplerDescriptorIndex, ", constNamePtr, dimension);
printSrcRegister(componentCount);
switch (instr.dimension)
{
case TextureDimension::Texture2D:
print(", float2({}, {})", instr.offsetX * 0.5f, instr.offsetY * 0.5f);
break;
case TextureDimension::TextureCube:
out += ", cubeMapData";
break;
}
out += ").";
printDstSwizzle(instr.dstSwizzle, true);
out += ";\n";
printDstSwizzle01(instr.dstRegister, instr.dstSwizzle);
if (instr.isPredicated)
{
--indentation;
indent();
out += "}\n";
}
}
void ShaderRecompiler::recompile(const AluInstruction& instr)
{
if (instr.isPredicated)
{
indent();
println("if ({}p0)", instr.predicateCondition ? "" : "!");
indent();
out += "{\n";
++indentation;
}
enum
{
VECTOR_0,
VECTOR_1,
VECTOR_2,
SCALAR_0,
SCALAR_1,
SCALAR_CONSTANT_0,
SCALAR_CONSTANT_1
};
auto op = [&](size_t operand)
{
size_t reg = 0;
size_t swizzle = 0;
bool select = true;
bool negate = false;
bool abs = false;
switch (operand)
{
case SCALAR_CONSTANT_0:
reg = instr.src3Register;
swizzle = instr.src3Swizzle;
select = false;
negate = instr.src3Negate;
abs = instr.absConstants;
break;
case SCALAR_CONSTANT_1:
reg = (uint32_t(instr.scalarOpcode) & 1) | (instr.src3Select << 1) | (instr.src3Swizzle & 0x3C);
swizzle = instr.src3Swizzle;
select = true;
negate = instr.src3Negate;
abs = instr.absConstants;
break;
default:
switch (operand)
{
case VECTOR_0:
reg = instr.src1Register;
swizzle = instr.src1Swizzle;
select = instr.src1Select;
negate = instr.src1Negate;
break;
case VECTOR_1:
reg = instr.src2Register;
swizzle = instr.src2Swizzle;
select = instr.src2Select;
negate = instr.src2Negate;
break;
case VECTOR_2:
case SCALAR_0:
case SCALAR_1:
reg = instr.src3Register;
swizzle = instr.src3Swizzle;
select = instr.src3Select;
negate = instr.src3Negate;
break;
}
if (select)
{
abs = (reg & 0x80) != 0;
reg &= 0x3F;
}
else
{
abs = instr.absConstants;
}
break;
}
std::string regFormatted;
if (select)
{
regFormatted = fmt::format("r{}", reg);
}
else
{
auto findResult = float4Constants.find(reg);
if (findResult != float4Constants.end())
{
const char* constantName = reinterpret_cast<const char*>(constantTableData + findResult->second->name);
if (findResult->second->registerCount > 1)
{
#ifdef UNLEASHED_RECOMP
if (hasMtxProjection && strcmp(constantName, "g_MtxProjection") == 0)
{
regFormatted = fmt::format("(iterationIndex == 0 ? mtxProjectionReverseZ[{0}] : mtxProjection[{0}])",
reg - findResult->second->registerIndex);
}
else
#endif
{
regFormatted = fmt::format("{}({}{})", constantName,
reg - findResult->second->registerIndex, instr.const0Relative ? (instr.constAddressRegisterRelative ? " + a0" : " + aL") : "");
}
}
else
{
assert(!instr.const0Relative && !instr.const1Relative);
regFormatted = constantName;
}
}
else
{
assert(!instr.const0Relative && !instr.const1Relative);
regFormatted = fmt::format("c{}", reg);
}
}
std::string result;
if (negate)
result += '-';
if (abs)
result += "abs(";
result += regFormatted;
result += '.';
switch (operand)
{
case VECTOR_0:
case VECTOR_1:
case VECTOR_2:
{
uint32_t mask;
switch (instr.vectorOpcode)
{
case AluVectorOpcode::Dp2Add:
mask = (operand == VECTOR_2) ? 0b1 : 0b11;
break;
case AluVectorOpcode::Dp3:
mask = 0b111;
break;
case AluVectorOpcode::Dp4:
case AluVectorOpcode::Max4:
mask = 0b1111;
break;
default:
mask = instr.vectorWriteMask != 0 ? instr.vectorWriteMask : 0b1;
break;
}
for (size_t i = 0; i < 4; i++)
{
if ((mask >> i) & 0x1)
result += SWIZZLES[((swizzle >> (i * 2)) + i) & 0x3];
}
break;
}
case SCALAR_0:
case SCALAR_CONSTANT_0:
result += SWIZZLES[((swizzle >> 6) + 3) & 0x3];
break;
case SCALAR_1:
case SCALAR_CONSTANT_1:
result += SWIZZLES[swizzle & 0x3];
break;
}
if (abs)
result += ")";
return result;
};
switch (instr.vectorOpcode)
{
case AluVectorOpcode::KillEq:
indent();
println("clip(any({} == {}) ? -1 : 1);", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::KillGt:
indent();
println("clip(any({} > {}) ? -1 : 1);", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::KillGe:
indent();
println("clip(any({} >= {}) ? -1 : 1);", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::KillNe:
indent();
println("clip(any({} != {}) ? -1 : 1);", op(VECTOR_0), op(VECTOR_1));
break;
}
bool closeIfBracket = false;
std::string_view exportRegister;
if (instr.exportData)
{
if (isPixelShader)
{
switch (ExportRegister(instr.vectorDest))
{
case ExportRegister::PSColor0:
exportRegister = "oC0";
break;
case ExportRegister::PSColor1:
exportRegister = "oC1";
break;
case ExportRegister::PSColor2:
exportRegister = "oC2";
break;
case ExportRegister::PSColor3:
exportRegister = "oC3";
break;
case ExportRegister::PSDepth:
exportRegister = "oDepth";
break;
}
}
else
{
switch (ExportRegister(instr.vectorDest))
{
case ExportRegister::VSPosition:
exportRegister = "oPos";
#ifdef UNLEASHED_RECOMP
if (hasMtxProjection)
{
indent();
out += "if ((g_SpecConstants() & SPEC_CONSTANT_REVERSE_Z) == 0 || iterationIndex == 0)\n";
indent();
out += "{\n";
++indentation;
closeIfBracket = true;
}
#endif
break;
default:
{
auto findResult = interpolators.find(instr.vectorDest);
assert(findResult != interpolators.end());
exportRegister = findResult->second;
break;
}
}
}
}
if (instr.vectorOpcode >= AluVectorOpcode::SetpEqPush && instr.vectorOpcode <= AluVectorOpcode::SetpGePush)
{
indent();
print("p0 = {} == 0.0 && {} ", op(VECTOR_0), op(VECTOR_1));
switch (instr.vectorOpcode)
{
case AluVectorOpcode::SetpEqPush:
out += "==";
break;
case AluVectorOpcode::SetpNePush:
out += "!=";
break;
case AluVectorOpcode::SetpGtPush:
out += ">";
break;
case AluVectorOpcode::SetpGePush:
out += ">=";
break;
}
out += " 0.0;\n";
}
else if (instr.vectorOpcode >= AluVectorOpcode::MaxA)
{
indent();
println("a0 = (int)clamp(floor(({}).w + 0.5), -256.0, 255.0);", op(VECTOR_0));
}
uint32_t vectorWriteMask = instr.vectorWriteMask;
if (instr.exportData)
vectorWriteMask &= ~instr.scalarWriteMask;
if (vectorWriteMask != 0)
{
indent();
if (!exportRegister.empty())
{
out += exportRegister;
out += '.';
}
else
{
print("r{}.", instr.vectorDest);
}
uint32_t vectorWriteSize = 0;
for (size_t i = 0; i < 4; i++)
{
if ((vectorWriteMask >> i) & 0x1) {
out += SWIZZLES[i];
vectorWriteSize++;
}
}
out += " = ";
if (vectorWriteSize > 1) {
print("(float{})(", vectorWriteSize);
} else {
out += "(float)(";
}
if (instr.vectorSaturate)
out += "saturate(";
switch (instr.vectorOpcode)
{
case AluVectorOpcode::Add:
print("{} + {}", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::Mul:
print("{} * {}", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::Max:
case AluVectorOpcode::MaxA:
print("max({}, {})", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::Min:
print("min({}, {})", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::Seq:
print("{} == {}", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::Sgt:
print("{} > {}", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::Sge:
print("{} >= {}", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::Sne:
print("{} != {}", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::Frc:
print("frac({})", op(VECTOR_0));
break;
case AluVectorOpcode::Trunc:
print("trunc({})", op(VECTOR_0));
break;
case AluVectorOpcode::Floor:
print("floor({})", op(VECTOR_0));
break;
case AluVectorOpcode::Mad:
print("{} * {} + {}", op(VECTOR_0), op(VECTOR_1), op(VECTOR_2));
break;
case AluVectorOpcode::CndEq:
print("selectWrapper({} == 0.0, {}, {})", op(VECTOR_0), op(VECTOR_1), op(VECTOR_2));
break;
case AluVectorOpcode::CndGe:
print("selectWrapper({} >= 0.0, {}, {})", op(VECTOR_0), op(VECTOR_1), op(VECTOR_2));
break;
case AluVectorOpcode::CndGt:
print("selectWrapper({} > 0.0, {}, {})", op(VECTOR_0), op(VECTOR_1), op(VECTOR_2));
break;
case AluVectorOpcode::Dp4:
case AluVectorOpcode::Dp3:
print("dot({}, {})", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::Dp2Add:
print("dot({}, {}) + {}", op(VECTOR_0), op(VECTOR_1), op(VECTOR_2));
break;
case AluVectorOpcode::Cube:
print("cube(r{}, cubeMapData)", instr.src1Register);
break;
case AluVectorOpcode::Max4:
print("max4({})", op(VECTOR_0));
break;
case AluVectorOpcode::SetpEqPush:
case AluVectorOpcode::SetpNePush:
case AluVectorOpcode::SetpGtPush:
case AluVectorOpcode::SetpGePush:
print("p0 ? 0.0 : {} + 1.0", op(VECTOR_0));
break;
case AluVectorOpcode::KillEq:
print("any({} == {})", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::KillGt:
print("any({} > {})", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::KillGe:
print("any({} >= {})", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::KillNe:
print("any({} != {})", op(VECTOR_0), op(VECTOR_1));
break;
case AluVectorOpcode::Dst:
print("dst({}, {})", op(VECTOR_0), op(VECTOR_1));
break;
}
if (instr.vectorSaturate)
out += ')';
out += ");\n";
}
if (instr.scalarOpcode != AluScalarOpcode::RetainPrev)
{
if (instr.scalarOpcode >= AluScalarOpcode::SetpEq && instr.scalarOpcode <= AluScalarOpcode::SetpRstr)
{
indent();
out += "p0 = ";
switch (instr.scalarOpcode)
{
case AluScalarOpcode::SetpEq:
print("{} == 0.0", op(SCALAR_0));
break;
case AluScalarOpcode::SetpNe:
print("{} != 0.0", op(SCALAR_0));
break;
case AluScalarOpcode::SetpGt:
print("{} > 0.0", op(SCALAR_0));
break;
case AluScalarOpcode::SetpGe:
print("{} >= 0.0", op(SCALAR_0));
break;
case AluScalarOpcode::SetpInv:
print("{} == 1.0", op(SCALAR_0));
break;
case AluScalarOpcode::SetpPop:
print("{} - 1.0 <= 0.0", op(SCALAR_0));
break;
case AluScalarOpcode::SetpClr:
out += "false";
break;
case AluScalarOpcode::SetpRstr:
print("{} == 0.0", op(SCALAR_0));
break;
}
out += ";\n";
}
indent();
out += "ps = ";
if (instr.scalarSaturate)
out += "saturate(";
switch (instr.scalarOpcode)
{
case AluScalarOpcode::Adds:
print("{} + {}", op(SCALAR_0), op(SCALAR_1));
break;
case AluScalarOpcode::AddsPrev:
print("{} + ps", op(SCALAR_0));
break;
case AluScalarOpcode::Muls:
print("{} * {}", op(SCALAR_0), op(SCALAR_1));
break;
case AluScalarOpcode::MulsPrev:
case AluScalarOpcode::MulsPrev2:
print("{} * ps", op(SCALAR_0));
break;
case AluScalarOpcode::Maxs:
case AluScalarOpcode::MaxAs:
case AluScalarOpcode::MaxAsf:
print("max({}, {})", op(SCALAR_0), op(SCALAR_1));
break;
case AluScalarOpcode::Mins:
print("min({}, {})", op(SCALAR_0), op(SCALAR_1));
break;
case AluScalarOpcode::Seqs:
print("{} == 0.0", op(SCALAR_0));
break;
case AluScalarOpcode::Sgts:
print("{} > 0.0", op(SCALAR_0));
break;
case AluScalarOpcode::Sges:
print("{} >= 0.0", op(SCALAR_0));
break;
case AluScalarOpcode::Snes:
print("{} != 0.0", op(SCALAR_0));
break;
case AluScalarOpcode::Frcs:
print("frac({})", op(SCALAR_0));
break;
case AluScalarOpcode::Truncs:
print("trunc({})", op(SCALAR_0));
break;
case AluScalarOpcode::Floors:
print("floor({})", op(SCALAR_0));
break;
case AluScalarOpcode::Exp:
print("exp2({})", op(SCALAR_0));
break;
case AluScalarOpcode::Logc:
case AluScalarOpcode::Log:
print("clamp(log2({}), FLT_MIN, FLT_MAX)", op(SCALAR_0));
break;
case AluScalarOpcode::Rcpc:
case AluScalarOpcode::Rcpf:
case AluScalarOpcode::Rcp:
print("clamp(rcp({}), FLT_MIN, FLT_MAX)", op(SCALAR_0));
break;
case AluScalarOpcode::Rsqc:
case AluScalarOpcode::Rsqf:
case AluScalarOpcode::Rsq:
print("clamp(rsqrt({}), FLT_MIN, FLT_MAX)", op(SCALAR_0));
break;
case AluScalarOpcode::Subs:
print("{} - {}", op(SCALAR_0), op(SCALAR_1));
break;
case AluScalarOpcode::SubsPrev:
print("{} - ps", op(SCALAR_0));
break;
case AluScalarOpcode::SetpEq:
case AluScalarOpcode::SetpNe:
case AluScalarOpcode::SetpGt:
case AluScalarOpcode::SetpGe:
out += "p0 ? 0.0 : 1.0";
break;
case AluScalarOpcode::SetpInv:
print("{0} == 0.0 ? 1.0 : {0}", op(SCALAR_0));
break;
case AluScalarOpcode::SetpPop:
print("p0 ? 0.0 : ({} - 1.0)", op(SCALAR_0));
break;
case AluScalarOpcode::SetpClr:
out += "FLT_MAX";
break;
case AluScalarOpcode::SetpRstr:
print("p0 ? 0.0 : {}", op(SCALAR_0));
break;
case AluScalarOpcode::KillsEq:
print("{} == 0.0", op(SCALAR_0));
break;
case AluScalarOpcode::KillsGt:
print("{} > 0.0", op(SCALAR_0));
break;
case AluScalarOpcode::KillsGe:
print("{} >= 0.0", op(SCALAR_0));
break;
case AluScalarOpcode::KillsNe:
print("{} != 0.0", op(SCALAR_0));
break;
case AluScalarOpcode::KillsOne:
print("{} == 1.0", op(SCALAR_0));
break;
case AluScalarOpcode::Sqrt:
print("sqrt({})", op(SCALAR_0));
break;
case AluScalarOpcode::Mulsc0:
case AluScalarOpcode::Mulsc1:
print("{} * {}", op(SCALAR_CONSTANT_0), op(SCALAR_CONSTANT_1));
break;
case AluScalarOpcode::Addsc0:
case AluScalarOpcode::Addsc1:
print("{} + {}", op(SCALAR_CONSTANT_0), op(SCALAR_CONSTANT_1));
break;
case AluScalarOpcode::Subsc0:
case AluScalarOpcode::Subsc1:
print("{} - {}", op(SCALAR_CONSTANT_0), op(SCALAR_CONSTANT_1));
break;
case AluScalarOpcode::Sin:
print("sin({})", op(SCALAR_0));
break;
case AluScalarOpcode::Cos:
print("cos({})", op(SCALAR_0));
break;
}
if (instr.scalarSaturate)
out += ')';
out += ";\n";
switch (instr.scalarOpcode)
{
case AluScalarOpcode::MaxAs:
indent();
println("a0 = (int)clamp(floor({} + 0.5), -256.0, 255.0);", op(SCALAR_0));
break;
case AluScalarOpcode::MaxAsf:
indent();
println("a0 = (int)clamp(floor({}), -256.0, 255.0);", op(SCALAR_0));
break;
}
}
uint32_t scalarWriteMask = instr.scalarWriteMask;
if (instr.exportData)
scalarWriteMask &= ~instr.vectorWriteMask;
if (scalarWriteMask != 0)
{
indent();
if (!exportRegister.empty())
{
out += exportRegister;
out += '.';
}
else
{
print("r{}.", instr.scalarDest);
}
for (size_t i = 0; i < 4; i++)
{
if ((scalarWriteMask >> i) & 0x1)
out += SWIZZLES[i];
}
out += " = ps;\n";
}
if (instr.exportData)
{
uint32_t zeroMask = instr.scalarDestRelative ? (0b1111 & ~(instr.vectorWriteMask | instr.scalarWriteMask)) : 0;
uint32_t oneMask = instr.vectorWriteMask & instr.scalarWriteMask;
for (size_t i = 0; i < 4; i++)
{
uint32_t mask = 1 << i;
if (zeroMask & mask)
{
indent();
println("{}.{} = 0.0;", exportRegister, SWIZZLES[i]);
}
else if (oneMask & mask)
{
indent();
println("{}.{} = 1.0;", exportRegister, SWIZZLES[i]);
}
}
}
if (instr.scalarOpcode >= AluScalarOpcode::KillsEq && instr.scalarOpcode <= AluScalarOpcode::KillsOne)
{
indent();
out += "clip(ps != 0.0 ? -1 : 1);\n";
}
if (closeIfBracket)
{
--indentation;
indent();
out += "}\n";
}
if (instr.isPredicated)
{
--indentation;
indent();
out += "}\n";
}
}
void ShaderRecompiler::recompile(const uint8_t* shaderData, const std::string_view& include)
{
const auto shaderContainer = reinterpret_cast<const ShaderContainer*>(shaderData);
assert((shaderContainer->flags & 0xFFFFFF00) == 0x102A1100);
assert(shaderContainer->constantTableOffset != NULL);
out += include;
out += '\n';
isPixelShader = (shaderContainer->flags & 0x1) == 0;
const auto constantTableContainer = reinterpret_cast<const ConstantTableContainer*>(shaderData + shaderContainer->constantTableOffset);
constantTableData = reinterpret_cast<const uint8_t*>(&constantTableContainer->constantTable);
out += "#ifdef __spirv__\n\n";
#ifdef UNLEASHED_RECOMP
bool isMetaInstancer = false;
bool hasIndexCount = false;
#endif
for (uint32_t i = 0; i < constantTableContainer->constantTable.constants; i++)
{
const auto constantInfo = reinterpret_cast<const ConstantInfo*>(
constantTableData + constantTableContainer->constantTable.constantInfo + i * sizeof(ConstantInfo));
const char* constantName = reinterpret_cast<const char*>(constantTableData + constantInfo->name);
#ifdef UNLEASHED_RECOMP
if (!isPixelShader)
{
if (strcmp(constantName, "g_MtxProjection") == 0)
hasMtxProjection = true;
else if (strcmp(constantName, "g_InstanceTypes") == 0)
isMetaInstancer = true;
else if (strcmp(constantName, "g_IndexCount") == 0)
hasIndexCount = true;
}
else
{
if (strcmp(constantName, "g_MtxPrevInvViewProjection") == 0)
hasMtxPrevInvViewProjection = true;
}
#endif
switch (constantInfo->registerSet)
{
case RegisterSet::Float4:
{
const char* shaderName = isPixelShader ? "Pixel" : "Vertex";
if (constantInfo->registerCount > 1)
{
uint32_t tailCount = (isPixelShader ? 224 : 256) - constantInfo->registerIndex;
println("#define {}(INDEX) selectWrapper((INDEX) < {}, vk::RawBufferLoad<float4>(g_PushConstants.{}ShaderConstants + ({} + min(INDEX, {})) * 16, 0x10), 0.0)",
constantName, tailCount, shaderName, constantInfo->registerIndex.get(), tailCount - 1);
}
else
{
println("#define {} vk::RawBufferLoad<float4>(g_PushConstants.{}ShaderConstants + {}, 0x10)",
constantName, shaderName, constantInfo->registerIndex * 16);
}
for (uint16_t j = 0; j < constantInfo->registerCount; j++)
float4Constants.emplace(constantInfo->registerIndex + j, constantInfo);
break;
}
case RegisterSet::Sampler:
{
for (size_t j = 0; j < std::size(TEXTURE_DIMENSIONS); j++)
{
println("#define {}_Texture{}DescriptorIndex vk::RawBufferLoad<uint>(g_PushConstants.SharedConstants + {})",
constantName, TEXTURE_DIMENSIONS[j], j * 64 + constantInfo->registerIndex * 4);
}
println("#define {}_SamplerDescriptorIndex vk::RawBufferLoad<uint>(g_PushConstants.SharedConstants + {})",
constantName, std::size(TEXTURE_DIMENSIONS) * 64 + constantInfo->registerIndex * 4);
samplers.emplace(constantInfo->registerIndex, constantName);
break;
}
}
}
out += "\n#elif __air__\n\n";
for (uint32_t i = 0; i < constantTableContainer->constantTable.constants; i++)
{
const auto constantInfo = reinterpret_cast<const ConstantInfo*>(
constantTableData + constantTableContainer->constantTable.constantInfo + i * sizeof(ConstantInfo));
const char* constantName = reinterpret_cast<const char*>(constantTableData + constantInfo->name);
#ifdef UNLEASHED_RECOMP
if (!isPixelShader)
{
if (strcmp(constantName, "g_MtxProjection") == 0)
hasMtxProjection = true;
else if (strcmp(constantName, "g_InstanceTypes") == 0)
isMetaInstancer = true;
else if (strcmp(constantName, "g_IndexCount") == 0)
hasIndexCount = true;
}
else
{
if (strcmp(constantName, "g_MtxPrevInvViewProjection") == 0)
hasMtxPrevInvViewProjection = true;
}
#endif
switch (constantInfo->registerSet)
{
case RegisterSet::Float4:
{
const char* shaderName = isPixelShader ? "Pixel" : "Vertex";
if (constantInfo->registerCount > 1)
{
uint32_t tailCount = (isPixelShader ? 224 : 256) - constantInfo->registerIndex;
println("#define {}(INDEX) selectWrapper((INDEX) < {}, (*(reinterpret_cast<device float4*>(g_PushConstants.{}ShaderConstants + ({} + min(INDEX, {})) * 16))), 0.0)",
constantName, tailCount, shaderName, constantInfo->registerIndex.get(), tailCount - 1);
}
else
{
println("#define {} (*(reinterpret_cast<device float4*>(g_PushConstants.{}ShaderConstants + {})))",
constantName, shaderName, constantInfo->registerIndex * 16);
}
for (uint16_t j = 0; j < constantInfo->registerCount; j++)
float4Constants.emplace(constantInfo->registerIndex + j, constantInfo);
break;
}
case RegisterSet::Sampler:
{
for (size_t j = 0; j < std::size(TEXTURE_DIMENSIONS); j++)
{
println("#define {}_Texture{}DescriptorIndex (*(reinterpret_cast<device uint*>(g_PushConstants.SharedConstants + {})))",
constantName, TEXTURE_DIMENSIONS[j], j * 64 + constantInfo->registerIndex * 4);
}
println("#define {}_SamplerDescriptorIndex (*(reinterpret_cast<device uint*>(g_PushConstants.SharedConstants + {})))",
constantName, std::size(TEXTURE_DIMENSIONS) * 64 + constantInfo->registerIndex * 4);
samplers.emplace(constantInfo->registerIndex, constantName);
break;
}
}
}
out += "\n#else\n\n";
println("cbuffer {}ShaderConstants : register(b{}, space4)", isPixelShader ? "Pixel" : "Vertex", isPixelShader ? 1 : 0);
out += "{\n";
for (uint32_t i = 0; i < constantTableContainer->constantTable.constants; i++)
{
const auto constantInfo = reinterpret_cast<const ConstantInfo*>(
constantTableData + constantTableContainer->constantTable.constantInfo + i * sizeof(ConstantInfo));
if (constantInfo->registerSet == RegisterSet::Float4)
{
const char* constantName = reinterpret_cast<const char*>(constantTableData + constantInfo->name);
print("\tfloat4 {}", constantName);
if (constantInfo->registerCount > 1)
print("[{}]", constantInfo->registerCount.get());
println(" : packoffset(c{});", constantInfo->registerIndex.get());
if (constantInfo->registerCount > 1)
{
uint32_t tailCount = (isPixelShader ? 224 : 256) - constantInfo->registerIndex;
println("#define {0}(INDEX) selectWrapper((INDEX) < {1}, {0}[min(INDEX, {2})], 0.0)", constantName, tailCount, tailCount - 1);
}
}
}
out += "};\n\n";
out += "cbuffer SharedConstants : register(b2, space4)\n";
out += "{\n";
for (uint32_t i = 0; i < constantTableContainer->constantTable.constants; i++)
{
const auto constantInfo = reinterpret_cast<const ConstantInfo*>(
constantTableData + constantTableContainer->constantTable.constantInfo + i * sizeof(ConstantInfo));
if (constantInfo->registerSet == RegisterSet::Sampler)
{
const char* constantName = reinterpret_cast<const char*>(constantTableData + constantInfo->name);
for (size_t j = 0; j < std::size(TEXTURE_DIMENSIONS); j++)
{
println("\tuint {}_Texture{}DescriptorIndex : packoffset(c{}.{});",
constantName, TEXTURE_DIMENSIONS[j], j * 4 + constantInfo->registerIndex / 4, SWIZZLES[constantInfo->registerIndex % 4]);
}
println("\tuint {}_SamplerDescriptorIndex : packoffset(c{}.{});",
constantName, 4 * std::size(TEXTURE_DIMENSIONS) + constantInfo->registerIndex / 4, SWIZZLES[constantInfo->registerIndex % 4]);
}
}
out += "\tDEFINE_SHARED_CONSTANTS();\n";
out += "};\n\n";
out += "#endif\n";
for (uint32_t i = 0; i < constantTableContainer->constantTable.constants; i++)
{
const auto constantInfo = reinterpret_cast<const ConstantInfo*>(
constantTableData + constantTableContainer->constantTable.constantInfo + i * sizeof(ConstantInfo));
if (constantInfo->registerSet == RegisterSet::Bool)
{
const char* constantName = reinterpret_cast<const char*>(constantTableData + constantInfo->name);
println("\t#define {} (1 << {})", constantName, constantInfo->registerIndex + (isPixelShader ? 16 : 0));
boolConstants.emplace(constantInfo->registerIndex, constantName);
}
}
out += '\n';
const auto shader = reinterpret_cast<const Shader*>(shaderData + shaderContainer->shaderOffset);
out += "#if __air__\n";
out += "struct StageIn\n";
out += "{\n";
for (auto& [usage, usageIndex] : INTERPOLATORS)
println("\tfloat4 i{}{};", USAGE_VARIABLES[uint32_t(usage)], usageIndex);
out += "};\n";
if (isPixelShader)
out += "[[fragment]]\n";
else
out += "[[vertex]]\n";
out += "#elifndef __spirv__\n";
if (isPixelShader)
out += "[shader(\"pixel\")]\n";
else
out += "[shader(\"vertex\")]\n";
out += "#endif\n";
out += "void shaderMain(\n";
if (isPixelShader)
{
out += "#ifdef __air__\n";
out += "\tStageIn iStageIn [[stage_in]],\n";
out += "\tfloat4 iPos [[position]],\n";
out += "\tbool iFace [[front_facing]],\n";
auto pixelShader = reinterpret_cast<const PixelShader*>(shader);
if (pixelShader->outputs & PIXEL_SHADER_OUTPUT_COLOR0)
out += "\tfloat4 oC0 [[color(0)]],\n";
if (pixelShader->outputs & PIXEL_SHADER_OUTPUT_COLOR1)
out += "\tfloat4 oC1 [[color(1)]],\n";
if (pixelShader->outputs & PIXEL_SHADER_OUTPUT_COLOR2)
out += "\tfloat4 oC2 [[color(2)]],\n";
if (pixelShader->outputs & PIXEL_SHADER_OUTPUT_COLOR3)
out += "\tfloat4 oC3 [[color(3)]],\n";
if (pixelShader->outputs & PIXEL_SHADER_OUTPUT_DEPTH)
out += "\tfloat oDepth [[depth(any)]],\n";
out += "\tconstant Texture2DDescriptorHeap& g_Texture2DDescriptorHeap [[buffer(0)]],\n";
out += "\tconstant Texture3DDescriptorHeap& g_Texture3DDescriptorHeap [[buffer(1)]],\n";
out += "\tconstant TextureCubeDescriptorHeap& g_TextureCubeDescriptorHeap [[buffer(2)]],\n";
out += "\tconstant SamplerDescriptorHeap& g_SamplerDescriptorHeap [[buffer(3)]],\n";
out += "\tconstant PushConstants& g_PushConstants [[buffer(4)]]\n";
out += "#else\n";
out += "\tin float4 iPos : SV_Position,\n";
for (auto& [usage, usageIndex] : INTERPOLATORS)
println("\tin float4 i{0}{1} : {2}{1},", USAGE_VARIABLES[uint32_t(usage)], usageIndex, USAGE_SEMANTICS[uint32_t(usage)]);
out += "#ifdef __spirv__\n";
out += "\tin bool iFace : SV_IsFrontFace\n";
out += "#else\n";
out += "\tin uint iFace : SV_IsFrontFace\n";
out += "#endif\n";
if (pixelShader->outputs & PIXEL_SHADER_OUTPUT_COLOR0)
out += ",\n\tout float4 oC0 : SV_Target0";
if (pixelShader->outputs & PIXEL_SHADER_OUTPUT_COLOR1)
out += ",\n\tout float4 oC1 : SV_Target1";
if (pixelShader->outputs & PIXEL_SHADER_OUTPUT_COLOR2)
out += ",\n\tout float4 oC2 : SV_Target2";
if (pixelShader->outputs & PIXEL_SHADER_OUTPUT_COLOR3)
out += ",\n\tout float4 oC3 : SV_Target3";
if (pixelShader->outputs & PIXEL_SHADER_OUTPUT_DEPTH)
out += ",\n\tout float oDepth : SV_Depth";
out += "\n#endif\n";
}
else
{
auto vertexShader = reinterpret_cast<const VertexShader*>(shader);
for (uint32_t i = 0; i < vertexShader->vertexElementCount; i++)
{
union
{
VertexElement vertexElement;
uint32_t value;
};
value = vertexShader->vertexElementsAndInterpolators[vertexShader->field18 + i];
const char* usageType = USAGE_TYPES[uint32_t(vertexElement.usage)];
#ifdef UNLEASHED_RECOMP
if ((vertexElement.usage == DeclUsage::TexCoord && vertexElement.usageIndex == 2 && isMetaInstancer) ||
(vertexElement.usage == DeclUsage::Position && vertexElement.usageIndex == 1))
{
usageType = "uint4";
}
#endif
out += '\t';
for (auto& usageLocation : USAGE_LOCATIONS)
{
if (usageLocation.usage == vertexElement.usage && usageLocation.usageIndex == vertexElement.usageIndex)
{
print("[[vk::location({})]] ", usageLocation.location);
break;
}
}
println("in {0} i{1}{2} : {3}{2},", usageType, USAGE_VARIABLES[uint32_t(vertexElement.usage)],
uint32_t(vertexElement.usageIndex), USAGE_SEMANTICS[uint32_t(vertexElement.usage)]);
vertexElements.emplace(uint32_t(vertexElement.address), vertexElement);
}
#ifdef UNLEASHED_RECOMP
if (hasIndexCount)
{
out += "\tin uint iVertexId : SV_VertexID,\n";
out += "\tin uint iInstanceId : SV_InstanceID,\n";
}
#endif
out += "\tout float4 oPos : SV_Position";
for (auto& [usage, usageIndex] : INTERPOLATORS)
print(",\n\tout float4 o{0}{1} : {2}{1}", USAGE_VARIABLES[uint32_t(usage)], usageIndex, USAGE_SEMANTICS[uint32_t(usage)]);
}
out += ")\n";
out += "{\n";
#ifdef UNLEASHED_RECOMP
if (hasMtxProjection)
{
specConstantsMask |= SPEC_CONSTANT_REVERSE_Z;
out += "\toPos = 0.0;\n";
out += "\tfloat4x4 mtxProjection = float4x4(g_MtxProjection(0), g_MtxProjection(1), g_MtxProjection(2), g_MtxProjection(3));\n";
out += "\tfloat4x4 mtxProjectionReverseZ = mul(mtxProjection, float4x4(1, 0, 0, 0, 0, 1, 0, 0, 0, 0, -1, 0, 0, 0, 1, 1));\n";
out += "\t[unroll] for (int iterationIndex = 0; iterationIndex < 2; iterationIndex++)\n";
out += "\t{\n";
}
#endif
if (shaderContainer->definitionTableOffset != NULL)
{
auto definitionTable = reinterpret_cast<const DefinitionTable*>(shaderData + shaderContainer->definitionTableOffset);
auto definitions = definitionTable->definitions;
while (*definitions != 0)
{
auto definition = reinterpret_cast<const Float4Definition*>(definitions);
auto value = reinterpret_cast<const be<uint32_t>*>(shaderData + shaderContainer->virtualSize + definition->physicalOffset);
for (uint16_t i = 0; i < (definition->count + 3) / 4; i++)
{
println("#ifdef __air__");
println("\tfloat4 c{} = as_type<float4>(uint4(0x{:X}, 0x{:X}, 0x{:X}, 0x{:X}));",
definition->registerIndex + i - (isPixelShader ? 256 : 0), value[0].get(), value[1].get(), value[2].get(), value[3].get());
println("#else");
println("\tfloat4 c{} = asfloat(uint4(0x{:X}, 0x{:X}, 0x{:X}, 0x{:X}));",
definition->registerIndex + i - (isPixelShader ? 256 : 0), value[0].get(), value[1].get(), value[2].get(), value[3].get());
println("#endif");
value += 4;
}
definitions += 2;
}
++definitions;
while (*definitions != 0)
{
auto definition = reinterpret_cast<const Int4Definition*>(definitions);
for (uint16_t i = 0; i < definition->count; i++)
{
union
{
uint32_t value;
struct
{
int8_t x;
int8_t y;
int8_t z;
int8_t w;
};
};
value = definition->values[i].get();
println("\tint4 i{} = int4({}, {}, {}, {});",
(definition->registerIndex - 8992) / 4 + i, x, y, z, w);
}
definitions += 2;
definitions += definition->count;
}
out += "\n";
}
bool printedRegisters[32]{};
uint32_t interpolatorCount = (shader->interpolatorInfo >> 5) & 0x1F;
for (uint32_t i = 0; i < interpolatorCount; i++)
{
union
{
Interpolator interpolator;
uint32_t value;
};
if (isPixelShader)
{
value = reinterpret_cast<const PixelShader*>(shader)->interpolators[i];
println("#ifdef __air__");
println("\tfloat4 r{} = iStageIn.i{}{};", uint32_t(interpolator.reg), USAGE_VARIABLES[uint32_t(interpolator.usage)], uint32_t(interpolator.usageIndex));
println("#else");
println("\tfloat4 r{} = i{}{};", uint32_t(interpolator.reg), USAGE_VARIABLES[uint32_t(interpolator.usage)], uint32_t(interpolator.usageIndex));
println("#endif");
printedRegisters[interpolator.reg] = true;
}
else
{
auto vertexShader = reinterpret_cast<const VertexShader*>(shader);
value = vertexShader->vertexElementsAndInterpolators[vertexShader->field18 + vertexShader->vertexElementCount + i];
interpolators.emplace(i, fmt::format("o{}{}", USAGE_VARIABLES[uint32_t(interpolator.usage)], uint32_t(interpolator.usageIndex)));
}
}
if (!isPixelShader)
{
#ifdef UNLEASHED_RECOMP
if (!hasMtxProjection)
out += "\toPos = 0.0;\n";
#endif
for (auto& [usage, usageIndex] : INTERPOLATORS)
println("\to{}{} = 0.0;", USAGE_VARIABLES[uint32_t(usage)], usageIndex);
out += "\n";
}
for (size_t i = 0; i < 32; i++)
{
if (!printedRegisters[i])
{
print("\tfloat4 r{} = ", i);
if (isPixelShader && i == ((shader->fieldC >> 8) & 0xFF))
{
out += "float4((iPos.xy - 0.5) * float2(iFace ? 1.0 : -1.0, 1.0), 0.0, 0.0);\n";
}
#ifdef UNLEASHED_RECOMP
else if (!isPixelShader && hasIndexCount && i == 0)
{
out += "float4(iVertexId + g_IndexCount.x * iInstanceId, 0.0, 0.0, 0.0);\n";
}
#endif
else
{
out += "0.0;\n";
}
}
}
out += "\tint a0 = 0;\n";
out += "\tint aL = 0;\n";
out += "\tbool p0 = false;\n";
out += "\tfloat ps = 0.0;\n";
if (isPixelShader)
{
#ifdef UNLEASHED_RECOMP
out += "\tfloat2 pixelCoord = 0.0;\n";
#endif
out += "#ifdef __air__\n";
out += "\tCubeMapData cubeMapdata = CubeMapData{};\n";
out += "#else\n";
out += "\tCubeMapData cubeMapData = (CubeMapData)0;\n";
out += "#endif\n";
}
const be<uint32_t>* code = reinterpret_cast<const be<uint32_t>*>(shaderData + shaderContainer->virtualSize + shader->physicalOffset);
union
{
ControlFlowInstruction controlFlow[2];
struct
{
uint32_t code0;
uint32_t code1;
uint32_t code2;
uint32_t code3;
};
};
auto controlFlowCode = code;
uint32_t instrAddress = 0;
uint32_t instrSize = shader->size;
bool simpleControlFlow = true;
while (instrAddress < instrSize)
{
code0 = controlFlowCode[0];
code1 = controlFlowCode[1] & 0xFFFF;
code2 = (controlFlowCode[1] >> 16) | (controlFlowCode[2] << 16);
code3 = controlFlowCode[2] >> 16;
for (auto& cfInstr : controlFlow)
{
uint32_t address = 0;
switch (cfInstr.opcode)
{
case ControlFlowOpcode::Exec:
case ControlFlowOpcode::ExecEnd:
address = cfInstr.exec.address;
break;
case ControlFlowOpcode::CondExec:
case ControlFlowOpcode::CondExecEnd:
case ControlFlowOpcode::CondExecPredClean:
case ControlFlowOpcode::CondExecPredCleanEnd:
address = cfInstr.condExec.address;
break;
case ControlFlowOpcode::CondExecPred:
case ControlFlowOpcode::CondExecPredEnd:
address = cfInstr.condExecPred.address;
break;
case ControlFlowOpcode::CondJmp:
{
if (cfInstr.condJmp.isUnconditional || cfInstr.condJmp.direction)
simpleControlFlow = false;
else
++ifEndLabels[cfInstr.condJmp.address];
break;
}
}
if (address != 0)
instrSize = std::min<uint32_t>(instrSize, address * 12);
}
controlFlowCode += 3;
instrAddress += 12;
}
if (simpleControlFlow)
{
out += '\n';
indentation = 1;
}
else
{
out += "\n\tuint pc = 0;\n";
out += "\twhile (true)\n";
out += "\t{\n";
out += "\t\tswitch (pc)\n";
out += "\t\t{\n";
}
controlFlowCode = code;
instrAddress = 0;
uint32_t pc = 0;
while (instrAddress < instrSize)
{
code0 = controlFlowCode[0];
code1 = controlFlowCode[1] & 0xFFFF;
code2 = (controlFlowCode[1] >> 16) | (controlFlowCode[2] << 16);
code3 = controlFlowCode[2] >> 16;
for (auto& cfInstr : controlFlow)
{
if (!simpleControlFlow)
{
indentation = 3;
println("\t\tcase {}:", pc);
}
else
{
auto findResult = ifEndLabels.find(pc);
if (findResult != ifEndLabels.end())
{
for (uint32_t i = 0; i < findResult->second; i++)
{
--indentation;
indent();
out += "}\n";
}
}
}
++pc;
uint32_t address = 0;
uint32_t count = 0;
uint32_t sequence = 0;
bool shouldReturn = false;
bool shouldCloseCurlyBracket = false;
switch (cfInstr.opcode)
{
case ControlFlowOpcode::Exec:
case ControlFlowOpcode::ExecEnd:
address = cfInstr.exec.address;
count = cfInstr.exec.count;
sequence = cfInstr.exec.sequence;
shouldReturn = (cfInstr.opcode == ControlFlowOpcode::ExecEnd);
break;
case ControlFlowOpcode::CondExec:
case ControlFlowOpcode::CondExecEnd:
case ControlFlowOpcode::CondExecPredClean:
case ControlFlowOpcode::CondExecPredCleanEnd:
address = cfInstr.condExec.address;
count = cfInstr.condExec.count;
sequence = cfInstr.condExec.sequence;
shouldReturn = (cfInstr.opcode == ControlFlowOpcode::CondExecEnd || cfInstr.opcode == ControlFlowOpcode::CondExecEnd);
break;
case ControlFlowOpcode::CondExecPred:
case ControlFlowOpcode::CondExecPredEnd:
address = cfInstr.condExecPred.address;
count = cfInstr.condExecPred.count;
sequence = cfInstr.condExecPred.sequence;
shouldReturn = (cfInstr.opcode == ControlFlowOpcode::CondExecPredEnd);
break;
case ControlFlowOpcode::LoopStart:
if (simpleControlFlow)
{
indent();
#ifdef UNLEASHED_RECOMP
print("UNROLL ");
#endif
println("for (aL = 0; aL < i{}.x; aL++)", uint32_t(cfInstr.loopStart.loopId));
indent();
out += "{\n";
++indentation;
}
else
{
out += "\t\t\taL = 0;\n";
}
break;
case ControlFlowOpcode::LoopEnd:
if (simpleControlFlow)
{
--indentation;
indent();
out += "}\n";
}
else
{
out += "\t\t\t++aL;\n";
println("\t\t\tif (aL < i{}.x)", uint32_t(cfInstr.loopEnd.loopId));
out += "\t\t\t{\n";
println("\t\t\t\tpc = {};", uint32_t(cfInstr.loopEnd.address));
out += "\t\t\t\tcontinue;\n";
out += "\t\t\t}\n";
}
break;
case ControlFlowOpcode::CondJmp:
{
if (cfInstr.condJmp.isUnconditional)
{
assert(!simpleControlFlow);
println("\t\t\tpc = {};", uint32_t(cfInstr.condJmp.address));
out += "\t\t\tcontinue;\n";
}
else
{
indent();
if (cfInstr.condJmp.isPredicated)
{
println("if ({}p0)", cfInstr.condJmp.condition ^ simpleControlFlow ? "" : "!");
}
else
{
auto findResult = boolConstants.find(cfInstr.condJmp.boolAddress);
if (findResult != boolConstants.end())
println("if ((g_Booleans & {}) {}= 0)", findResult->second, cfInstr.condJmp.condition ^ simpleControlFlow ? "!" : "=");
else
println("if (b{} {}= 0)", uint32_t(cfInstr.condJmp.boolAddress), cfInstr.condJmp.condition ^ simpleControlFlow ? "!" : "=");
}
if (simpleControlFlow)
{
indent();
out += "{\n";
++indentation;
}
else
{
out += "\t\t\t{\n";
println("\t\t\t\tpc = {};", uint32_t(cfInstr.condJmp.address));
out += "\t\t\t\tcontinue;\n";
out += "\t\t\t}\n";
}
}
break;
}
}
auto instructionCode = code + address * 3;
for (uint32_t i = 0; i < count; i++)
{
union
{
VertexFetchInstruction vertexFetch;
TextureFetchInstruction textureFetch;
AluInstruction alu;
struct
{
uint32_t code0;
uint32_t code1;
uint32_t code2;
};
};
code0 = instructionCode[0];
code1 = instructionCode[1];
code2 = instructionCode[2];
if ((sequence & 0x1) != 0)
{
if (vertexFetch.opcode == FetchOpcode::VertexFetch)
{
recompile(vertexFetch, address + i);
}
else
{
#ifdef UNLEASHED_RECOMP
if (textureFetch.constIndex == 10) // g_GISampler
{
specConstantsMask |= SPEC_CONSTANT_BICUBIC_GI_FILTER;
indent();
out += "if (g_SpecConstants() & SPEC_CONSTANT_BICUBIC_GI_FILTER)\n";
indent();
out += "{\n";
++indentation;
recompile(textureFetch, true);
--indentation;
indent();
out += "}\n";
indent();
out += "else\n";
indent();
out += "{\n";
++indentation;
recompile(textureFetch, false);
--indentation;
indent();
out += "}\n";
}
else
#endif
{
recompile(textureFetch, false);
}
}
}
else
{
recompile(alu);
}
sequence >>= 2;
instructionCode += 3;
}
if (shouldReturn)
{
if (isPixelShader)
{
specConstantsMask |= SPEC_CONSTANT_ALPHA_TEST;
indent();
out += "BRANCH if (g_SpecConstants() & SPEC_CONSTANT_ALPHA_TEST)\n";
indent();
out += "{\n";
indent();
out += "\tclip(oC0.w - g_AlphaThreshold);\n";
indent();
out += "}\n";
#ifdef UNLEASHED_RECOMP
specConstantsMask |= SPEC_CONSTANT_ALPHA_TO_COVERAGE;
indent();
out += "else if (g_SpecConstants() & SPEC_CONSTANT_ALPHA_TO_COVERAGE)\n";
indent();
out += "{\n";
indent();
out += "\toC0.w *= 1.0 + computeMipLevel(pixelCoord) * 0.25;\n";
indent();
out += "\toC0.w = 0.5 + (oC0.w - g_AlphaThreshold) / max(fwidth(oC0.w), 1e-6);\n";
indent();
out += "}\n";
#endif
}
if (simpleControlFlow)
{
indent();
#ifdef UNLEASHED_RECOMP
if (hasMtxProjection)
{
out += "continue;\n";
}
else
#endif
{
out += "return;\n";
}
}
else
{
out += "\t\t\tbreak;\n";
}
}
if (shouldCloseCurlyBracket)
{
--indentation;
indent();
out += "}\n";
}
}
controlFlowCode += 3;
instrAddress += 12;
}
if (!simpleControlFlow)
{
out += "\t\t\tbreak;\n";
out += "\t\t}\n";
out += "\t\tbreak;\n";
out += "\t}\n";
}
#ifdef UNLEASHED_RECOMP
if (hasMtxProjection)
out += "\t}\n";
#endif
out += "}";
}