mirror of
https://github.com/KytyPS5/KytyPS5.git
synced 2026-10-29 08:41:20 +00:00
refactor(shader): preserve native scalar memory operations
This commit is contained in:
1 parent
c236ba713d
commit
8063822243
10 files changed
+296
-200
No files matched your search
@@ -273,7 +273,7 @@ void ClearRegisterOffsetModifiers(Decoder::Operand& operand) {
|
||||
operand.dpp = false;
|
||||
}
|
||||
|
||||
bool TryGetScalarDestinationCode(const Decoder::Operand& operand, uint32_t& code) {
|
||||
bool TryGetEncodedScalarCode(const Decoder::Operand& operand, uint32_t& code) {
|
||||
switch (operand.kind) {
|
||||
case Decoder::OperandKind::Sgpr: code = operand.reg; return true;
|
||||
case Decoder::OperandKind::VccLo: code = SCALAR_DST_VCC_LO; return true;
|
||||
@@ -290,7 +290,7 @@ bool TryOffsetScalarDestination(const Decoder::Operand& operand, uint32_t index,
|
||||
// aliases such as VCC and EXEC sit beside ordinary SGPRs. Re-decode after adding the
|
||||
// dword index so s_buffer_load_dwordx2 vcc_lo writes vcc_lo, then vcc_hi.
|
||||
uint32_t code = 0;
|
||||
return TryGetScalarDestinationCode(operand, code) &&
|
||||
return TryGetEncodedScalarCode(operand, code) &&
|
||||
Decoder::DecodeScalarDestination(code + index, 0, decoded, nullptr);
|
||||
}
|
||||
|
||||
|
||||
@@ -30,7 +30,7 @@ uint32_t ResourceIndexFromOperand(const Decoder::Operand& operand);
|
||||
MemoryInfo MemoryInfoFromDecoded(const Decoder::Instruction& decoded, ResourceKind kind);
|
||||
uint32_t RawScalarLoadBase(const Decoder::Operand& operand);
|
||||
void ClearRegisterOffsetModifiers(Decoder::Operand& operand);
|
||||
bool TryGetScalarDestinationCode(const Decoder::Operand& operand, uint32_t& code);
|
||||
bool TryGetEncodedScalarCode(const Decoder::Operand& operand, uint32_t& code);
|
||||
bool TryOffsetScalarDestination(const Decoder::Operand& operand, uint32_t index,
|
||||
Decoder::Operand& result);
|
||||
Decoder::Operand OffsetDecodedRegister(const Decoder::Operand& operand, uint32_t index);
|
||||
|
||||
@@ -77,24 +77,40 @@ Opcode LoweredImageOpcode(Decoder::Opcode opcode) {
|
||||
|
||||
bool LowerScalarMemoryLoadDword(const Decoder::Instruction& decoded, BasicBlock& block, Opcode op,
|
||||
std::string* error) {
|
||||
for (uint32_t i = 0; i < decoded.data_dwords; i++) {
|
||||
Instruction inst;
|
||||
inst.pc = decoded.pc;
|
||||
inst.op = op;
|
||||
inst.src_count = 1;
|
||||
const auto kind =
|
||||
op == Opcode::SLoadDword ? ResourceKind::ScalarAddress : ResourceKind::ScalarBuffer;
|
||||
inst.memory = OffsetMemoryInfo(decoded, kind, i);
|
||||
// Raw s_load uses a two-SGPR pointer base; s_buffer_load uses a four-SGPR
|
||||
// descriptor resource index.
|
||||
inst.memory.resource = op == Opcode::SLoadDword ? RawScalarLoadBase(decoded.src0)
|
||||
: ResourceIndexFromOperand(decoded.src0);
|
||||
if (!LowerRegisterOperand(OffsetDecodedRegister(decoded.dst, i), inst.dst, error) ||
|
||||
!LowerSourceOperand(decoded.src1, inst.src[0], error)) {
|
||||
return false;
|
||||
}
|
||||
block.instructions.push_back(inst);
|
||||
uint32_t destination_code = 0;
|
||||
const auto alignment = decoded.data_dwords == 1u ? 1u
|
||||
: decoded.data_dwords == 2u ? 2u
|
||||
: 4u;
|
||||
if (!TryGetEncodedScalarCode(decoded.dst, destination_code) ||
|
||||
destination_code % alignment != 0u ||
|
||||
destination_code + decoded.data_dwords > 108u) {
|
||||
SetError(error, "scalar-memory destination has an invalid aligned SGPR span");
|
||||
return false;
|
||||
}
|
||||
uint32_t base_code = 0;
|
||||
const auto base_alignment = op == Opcode::SLoadDword ? 2u : 4u;
|
||||
const auto base_width = op == Opcode::SLoadDword ? 2u : 4u;
|
||||
if (!TryGetEncodedScalarCode(decoded.src0, base_code) ||
|
||||
base_code % base_alignment != 0u || base_code + base_width > 108u) {
|
||||
SetError(error, "scalar-memory base has an invalid aligned SGPR span");
|
||||
return false;
|
||||
}
|
||||
Instruction inst;
|
||||
inst.pc = decoded.pc;
|
||||
inst.op = op;
|
||||
inst.src_count = 1;
|
||||
const auto kind =
|
||||
op == Opcode::SLoadDword ? ResourceKind::ScalarAddress : ResourceKind::ScalarBuffer;
|
||||
inst.memory = MemoryInfoFromDecoded(decoded, kind);
|
||||
// Raw s_load uses a two-SGPR pointer base; s_buffer_load uses a four-SGPR
|
||||
// descriptor resource index.
|
||||
inst.memory.resource = op == Opcode::SLoadDword ? RawScalarLoadBase(decoded.src0)
|
||||
: ResourceIndexFromOperand(decoded.src0);
|
||||
if (!LowerRegisterOperand(decoded.dst, inst.dst, error) ||
|
||||
!LowerSourceOperand(decoded.src1, inst.src[0], error)) {
|
||||
return false;
|
||||
}
|
||||
block.instructions.push_back(inst);
|
||||
return true;
|
||||
}
|
||||
|
||||
|
||||
@@ -311,6 +311,33 @@ bool ValidateValueProgram(const ValueProgram& program, bool require_ssa, std::st
|
||||
ValueOpcodeName(inst.GetOpcode()), inst.NumArgs(),
|
||||
NumArgsOf(inst.GetOpcode())));
|
||||
}
|
||||
if (IsRuntimeRead(inst.GetOpcode())) {
|
||||
const auto memory_index = inst.Flags<MemoryFlags>().index;
|
||||
if (memory_index >= program.memory_info.size()) {
|
||||
return Fail(error, fmt::format("{} has an invalid memory-info index",
|
||||
ValueOpcodeName(inst.GetOpcode())));
|
||||
}
|
||||
const auto& memory = program.memory_info[memory_index];
|
||||
const bool scalar_buffer = inst.GetOpcode() == ValueOpcode::ReadConstBuffer;
|
||||
const bool scalar_address = memory.kind == ResourceKind::ScalarAddress;
|
||||
if ((scalar_buffer && memory.kind != ResourceKind::ScalarBuffer) ||
|
||||
(!scalar_buffer && !scalar_address && memory.kind != ResourceKind::Flat &&
|
||||
memory.kind != ResourceKind::Global && memory.kind != ResourceKind::Scratch)) {
|
||||
return Fail(error, fmt::format("{} has an invalid scalar-memory resource kind",
|
||||
ValueOpcodeName(inst.GetOpcode())));
|
||||
}
|
||||
const bool scalar_memory = scalar_buffer || scalar_address;
|
||||
const bool valid_group_width =
|
||||
scalar_memory ? memory.component_count == 1u || memory.component_count == 2u ||
|
||||
memory.component_count == 4u || memory.component_count == 8u ||
|
||||
memory.component_count == 16u
|
||||
: memory.component_count >= 1u && memory.component_count <= 4u;
|
||||
if (memory.data_bits != 32u || memory.data_dwords != 1u ||
|
||||
!valid_group_width || memory.component_index >= memory.component_count) {
|
||||
return Fail(error, fmt::format("{} has inconsistent scalar-memory metadata",
|
||||
ValueOpcodeName(inst.GetOpcode())));
|
||||
}
|
||||
}
|
||||
const auto buffer_components = BufferComponentCount(inst.GetOpcode());
|
||||
if (buffer_components != 0u) {
|
||||
const auto memory_index = inst.Flags<MemoryFlags>().index;
|
||||
|
||||
Reference in new issue
Block a user