wangtao
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
llvm-project/llvm/lib/Target/PTX/PTXISelLowering.cpp

523 lines
18 KiB
C++
Raw Blame History

//===-- PTXISelLowering.cpp - PTX DAG Lowering Implementation -------------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file implements the PTXTargetLowering class.
//
//===----------------------------------------------------------------------===//
#include "PTXISelLowering.h"
#include "PTX.h"
#include "PTXMachineFunctionInfo.h"
#include "PTXRegisterInfo.h"
#include "PTXSubtarget.h"
#include "llvm/Function.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/CodeGen/CallingConvLower.h"
#include "llvm/CodeGen/MachineFunction.h"
#include "llvm/CodeGen/MachineFrameInfo.h"
#include "llvm/CodeGen/MachineRegisterInfo.h"
#include "llvm/CodeGen/SelectionDAG.h"
#include "llvm/CodeGen/TargetLoweringObjectFileImpl.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/raw_ostream.h"
using namespace llvm;
//===----------------------------------------------------------------------===//
// TargetLowering Implementation
//===----------------------------------------------------------------------===//
PTXTargetLowering::PTXTargetLowering(TargetMachine &TM)
: TargetLowering(TM, new TargetLoweringObjectFileELF()) {
// Set up the register classes.
addRegisterClass(MVT::i1, PTX::RegPredRegisterClass);
addRegisterClass(MVT::i16, PTX::RegI16RegisterClass);
addRegisterClass(MVT::i32, PTX::RegI32RegisterClass);
addRegisterClass(MVT::i64, PTX::RegI64RegisterClass);
addRegisterClass(MVT::f32, PTX::RegF32RegisterClass);
addRegisterClass(MVT::f64, PTX::RegF64RegisterClass);
setBooleanContents(ZeroOrOneBooleanContent);
setBooleanVectorContents(ZeroOrOneBooleanContent); // FIXME: Is this correct?
setMinFunctionAlignment(2);
// Let LLVM use loads/stores for all mem* operations
maxStoresPerMemcpy = 4096;
maxStoresPerMemmove = 4096;
maxStoresPerMemset = 4096;
////////////////////////////////////
/////////// Expansion //////////////
////////////////////////////////////
// (any/zero/sign) extload => load + (any/zero/sign) extend
setLoadExtAction(ISD::EXTLOAD, MVT::i16, Expand);
setLoadExtAction(ISD::ZEXTLOAD, MVT::i16, Expand);
setLoadExtAction(ISD::SEXTLOAD, MVT::i16, Expand);
// f32 extload => load + fextend
setLoadExtAction(ISD::EXTLOAD, MVT::f32, Expand);
// f64 truncstore => trunc + store
setTruncStoreAction(MVT::f64, MVT::f32, Expand);
// sign_extend_inreg => sign_extend
setOperationAction(ISD::SIGN_EXTEND_INREG, MVT::i1, Expand);
// br_cc => brcond
setOperationAction(ISD::BR_CC, MVT::Other, Expand);
// select_cc => setcc
setOperationAction(ISD::SELECT_CC, MVT::Other, Expand);
setOperationAction(ISD::SELECT_CC, MVT::f32, Expand);
setOperationAction(ISD::SELECT_CC, MVT::f64, Expand);
////////////////////////////////////
//////////// Legal /////////////////
////////////////////////////////////
setOperationAction(ISD::ConstantFP, MVT::f32, Legal);
setOperationAction(ISD::ConstantFP, MVT::f64, Legal);
////////////////////////////////////
//////////// Custom ////////////////
////////////////////////////////////
// customise setcc to use bitwise logic if possible
//setOperationAction(ISD::SETCC, MVT::i1, Custom);
setOperationAction(ISD::SETCC, MVT::i1, Legal);
// customize translation of memory addresses
setOperationAction(ISD::GlobalAddress, MVT::i32, Custom);
setOperationAction(ISD::GlobalAddress, MVT::i64, Custom);
// Compute derived properties from the register classes
computeRegisterProperties();
}
EVT PTXTargetLowering::getSetCCResultType(EVT VT) const {
return MVT::i1;
}
SDValue PTXTargetLowering::LowerOperation(SDValue Op, SelectionDAG &DAG) const {
switch (Op.getOpcode()) {
default:
llvm_unreachable("Unimplemented operand");
case ISD::SETCC:
return LowerSETCC(Op, DAG);
case ISD::GlobalAddress:
return LowerGlobalAddress(Op, DAG);
}
}
const char *PTXTargetLowering::getTargetNodeName(unsigned Opcode) const {
switch (Opcode) {
default:
llvm_unreachable("Unknown opcode");
case PTXISD::COPY_ADDRESS:
return "PTXISD::COPY_ADDRESS";
case PTXISD::LOAD_PARAM:
return "PTXISD::LOAD_PARAM";
case PTXISD::STORE_PARAM:
return "PTXISD::STORE_PARAM";
case PTXISD::READ_PARAM:
return "PTXISD::READ_PARAM";
case PTXISD::WRITE_PARAM:
return "PTXISD::WRITE_PARAM";
case PTXISD::EXIT:
return "PTXISD::EXIT";
case PTXISD::RET:
return "PTXISD::RET";
case PTXISD::CALL:
return "PTXISD::CALL";
}
}
//===----------------------------------------------------------------------===//
// Custom Lower Operation
//===----------------------------------------------------------------------===//
SDValue PTXTargetLowering::LowerSETCC(SDValue Op, SelectionDAG &DAG) const {
assert(Op.getValueType() == MVT::i1 && "SetCC type must be 1-bit integer");
SDValue Op0 = Op.getOperand(0);
SDValue Op1 = Op.getOperand(1);
SDValue Op2 = Op.getOperand(2);
DebugLoc dl = Op.getDebugLoc();
//ISD::CondCode CC = cast<CondCodeSDNode>(Op.getOperand(2))->get();
// Look for X == 0, X == 1, X != 0, or X != 1
// We can simplify these to bitwise logic
//if (Op1.getOpcode() == ISD::Constant &&
// (cast<ConstantSDNode>(Op1)->getZExtValue() == 1 ||
// cast<ConstantSDNode>(Op1)->isNullValue()) &&
// (CC == ISD::SETEQ || CC == ISD::SETNE)) {
//
// return DAG.getNode(ISD::AND, dl, MVT::i1, Op0, Op1);
//}
//ConstantSDNode* COp1 = cast<ConstantSDNode>(Op1);
//if(COp1 && COp1->getZExtValue() == 1) {
// if(CC == ISD::SETNE) {
// return DAG.getNode(PTX::XORripreds, dl, MVT::i1, Op0);
// }
//}
llvm_unreachable("setcc was not matched by a pattern!");
return DAG.getNode(ISD::SETCC, dl, MVT::i1, Op0, Op1, Op2);
}
SDValue PTXTargetLowering::
LowerGlobalAddress(SDValue Op, SelectionDAG &DAG) const {
EVT PtrVT = getPointerTy();
DebugLoc dl = Op.getDebugLoc();
const GlobalValue *GV = cast<GlobalAddressSDNode>(Op)->getGlobal();
assert(PtrVT.isSimple() && "Pointer must be to primitive type.");
SDValue targetGlobal = DAG.getTargetGlobalAddress(GV, dl, PtrVT);
SDValue movInstr = DAG.getNode(PTXISD::COPY_ADDRESS,
dl,
PtrVT.getSimpleVT(),
targetGlobal);
return movInstr;
}
//===----------------------------------------------------------------------===//
// Calling Convention Implementation
//===----------------------------------------------------------------------===//
SDValue PTXTargetLowering::
LowerFormalArguments(SDValue Chain,
CallingConv::ID CallConv,
bool isVarArg,
const SmallVectorImpl<ISD::InputArg> &Ins,
DebugLoc dl,
SelectionDAG &DAG,
SmallVectorImpl<SDValue> &InVals) const {
if (isVarArg) llvm_unreachable("PTX does not support varargs");
MachineFunction &MF = DAG.getMachineFunction();
const PTXSubtarget& ST = getTargetMachine().getSubtarget<PTXSubtarget>();
PTXMachineFunctionInfo *MFI = MF.getInfo<PTXMachineFunctionInfo>();
PTXParamManager &PM = MFI->getParamManager();
switch (CallConv) {
default:
llvm_unreachable("Unsupported calling convention");
case CallingConv::PTX_Kernel:
MFI->setKernel(true);
break;
case CallingConv::PTX_Device:
MFI->setKernel(false);
break;
}
// We do one of two things here:
// IsKernel || SM >= 2.0 -> Use param space for arguments
// SM < 2.0 -> Use registers for arguments
if (MFI->isKernel() || ST.useParamSpaceForDeviceArgs()) {
// We just need to emit the proper LOAD_PARAM ISDs
for (unsigned i = 0, e = Ins.size(); i != e; ++i) {
assert((!MFI->isKernel() || Ins[i].VT != MVT::i1) &&
"Kernels cannot take pred operands");
unsigned ParamSize = Ins[i].VT.getStoreSizeInBits();
unsigned Param = PM.addArgumentParam(ParamSize);
const std::string &ParamName = PM.getParamName(Param);
SDValue ParamValue = DAG.getTargetExternalSymbol(ParamName.c_str(),
MVT::Other);
SDValue ArgValue = DAG.getNode(PTXISD::LOAD_PARAM, dl, Ins[i].VT, Chain,
ParamValue);
InVals.push_back(ArgValue);
}
}
else {
for (unsigned i = 0, e = Ins.size(); i != e; ++i) {
EVT RegVT = Ins[i].VT;
const TargetRegisterClass* TRC = getRegClassFor(RegVT);
unsigned RegType;
// Determine which register class we need
if (RegVT == MVT::i1)
RegType = PTXRegisterType::Pred;
else if (RegVT == MVT::i16)
RegType = PTXRegisterType::B16;
else if (RegVT == MVT::i32)
RegType = PTXRegisterType::B32;
else if (RegVT == MVT::i64)
RegType = PTXRegisterType::B64;
else if (RegVT == MVT::f32)
RegType = PTXRegisterType::F32;
else if (RegVT == MVT::f64)
RegType = PTXRegisterType::F64;
else
llvm_unreachable("Unknown parameter type");
// Use a unique index in the instruction to prevent instruction folding.
// Yes, this is a hack.
SDValue Index = DAG.getTargetConstant(i, MVT::i32);
unsigned Reg = MF.getRegInfo().createVirtualRegister(TRC);
SDValue ArgValue = DAG.getNode(PTXISD::READ_PARAM, dl, RegVT, Chain,
Index);
InVals.push_back(ArgValue);
MFI->addRegister(Reg, RegType, PTXRegisterSpace::Argument);
}
}
return Chain;
}
SDValue PTXTargetLowering::
LowerReturn(SDValue Chain,
CallingConv::ID CallConv,
bool isVarArg,
const SmallVectorImpl<ISD::OutputArg> &Outs,
const SmallVectorImpl<SDValue> &OutVals,
DebugLoc dl,
SelectionDAG &DAG) const {
if (isVarArg) llvm_unreachable("PTX does not support varargs");
switch (CallConv) {
default:
llvm_unreachable("Unsupported calling convention.");
case CallingConv::PTX_Kernel:
assert(Outs.size() == 0 && "Kernel must return void.");
return DAG.getNode(PTXISD::EXIT, dl, MVT::Other, Chain);
case CallingConv::PTX_Device:
assert(Outs.size() <= 1 && "Can at most return one value.");
break;
}
MachineFunction& MF = DAG.getMachineFunction();
PTXMachineFunctionInfo *MFI = MF.getInfo<PTXMachineFunctionInfo>();
PTXParamManager &PM = MFI->getParamManager();
SDValue Flag;
const PTXSubtarget& ST = getTargetMachine().getSubtarget<PTXSubtarget>();
if (ST.useParamSpaceForDeviceArgs()) {
assert(Outs.size() < 2 && "Device functions can return at most one value");
if (Outs.size() == 1) {
unsigned ParamSize = OutVals[0].getValueType().getSizeInBits();
unsigned Param = PM.addReturnParam(ParamSize);
const std::string &ParamName = PM.getParamName(Param);
SDValue ParamValue = DAG.getTargetExternalSymbol(ParamName.c_str(),
MVT::Other);
Chain = DAG.getNode(PTXISD::STORE_PARAM, dl, MVT::Other, Chain,
ParamValue, OutVals[0]);
}
} else {
for (unsigned i = 0, e = Outs.size(); i != e; ++i) {
EVT RegVT = Outs[i].VT;
const TargetRegisterClass* TRC;
unsigned RegType;
// Determine which register class we need
if (RegVT == MVT::i1) {
TRC = PTX::RegPredRegisterClass;
RegType = PTXRegisterType::Pred;
}
else if (RegVT == MVT::i16) {
TRC = PTX::RegI16RegisterClass;
RegType = PTXRegisterType::B16;
}
else if (RegVT == MVT::i32) {
TRC = PTX::RegI32RegisterClass;
RegType = PTXRegisterType::B32;
}
else if (RegVT == MVT::i64) {
TRC = PTX::RegI64RegisterClass;
RegType = PTXRegisterType::B64;
}
else if (RegVT == MVT::f32) {
TRC = PTX::RegF32RegisterClass;
RegType = PTXRegisterType::F32;
}
else if (RegVT == MVT::f64) {
TRC = PTX::RegF64RegisterClass;
RegType = PTXRegisterType::F64;
}
else {
llvm_unreachable("Unknown parameter type");
}
unsigned Reg = MF.getRegInfo().createVirtualRegister(TRC);
SDValue Copy = DAG.getCopyToReg(Chain, dl, Reg, OutVals[i]/*, Flag*/);
SDValue OutReg = DAG.getRegister(Reg, RegVT);
Chain = DAG.getNode(PTXISD::WRITE_PARAM, dl, MVT::Other, Copy, OutReg);
MFI->addRegister(Reg, RegType, PTXRegisterSpace::Return);
}
}
if (Flag.getNode() == 0) {
return DAG.getNode(PTXISD::RET, dl, MVT::Other, Chain);
}
else {
return DAG.getNode(PTXISD::RET, dl, MVT::Other, Chain, Flag);
}
}
SDValue
PTXTargetLowering::LowerCall(SDValue Chain, SDValue Callee,
CallingConv::ID CallConv, bool isVarArg,
bool doesNotRet, bool &isTailCall,
const SmallVectorImpl<ISD::OutputArg> &Outs,
const SmallVectorImpl<SDValue> &OutVals,
const SmallVectorImpl<ISD::InputArg> &Ins,
DebugLoc dl, SelectionDAG &DAG,
SmallVectorImpl<SDValue> &InVals) const {
MachineFunction& MF = DAG.getMachineFunction();
PTXMachineFunctionInfo *PTXMFI = MF.getInfo<PTXMachineFunctionInfo>();
PTXParamManager &PM = PTXMFI->getParamManager();
MachineFrameInfo *MFI = MF.getFrameInfo();
assert(getTargetMachine().getSubtarget<PTXSubtarget>().callsAreHandled() &&
"Calls are not handled for the target device");
// Identify the callee function
const GlobalValue *GV = cast<GlobalAddressSDNode>(Callee)->getGlobal();
const Function *function = cast<Function>(GV);
// allow non-device calls only for printf
bool isPrintf = function->getName() == "printf" || function->getName() == "puts";
assert((isPrintf || function->getCallingConv() == CallingConv::PTX_Device) &&
"PTX function calls must be to PTX device functions");
unsigned outSize = isPrintf ? 2 : Outs.size();
std::vector<SDValue> Ops;
// The layout of the ops will be [Chain, #Ins, Ins, Callee, #Outs, Outs]
Ops.resize(outSize + Ins.size() + 4);
Ops[0] = Chain;
// Identify the callee function
Callee = DAG.getTargetGlobalAddress(GV, dl, getPointerTy());
Ops[Ins.size()+2] = Callee;
// #Outs
Ops[Ins.size()+3] = DAG.getTargetConstant(outSize, MVT::i32);
if (isPrintf) {
// first argument is the address of the global string variable in memory
unsigned Param0 = PM.addLocalParam(getPointerTy().getSizeInBits());
SDValue ParamValue0 = DAG.getTargetExternalSymbol(PM.getParamName(Param0).c_str(),
MVT::Other);
Chain = DAG.getNode(PTXISD::STORE_PARAM, dl, MVT::Other, Chain,
ParamValue0, OutVals[0]);
Ops[Ins.size()+4] = ParamValue0;
// alignment is the maximum size of all the arguments
unsigned alignment = 0;
for (unsigned i = 1; i < OutVals.size(); ++i) {
alignment = std::max(alignment,
OutVals[i].getValueType().getSizeInBits());
}
// size is the alignment multiplied by the number of arguments
unsigned size = alignment * (OutVals.size() - 1);
// second argument is the address of the stack object (unless no arguments)
unsigned Param1 = PM.addLocalParam(getPointerTy().getSizeInBits());
SDValue ParamValue1 = DAG.getTargetExternalSymbol(PM.getParamName(Param1).c_str(),
MVT::Other);
Ops[Ins.size()+5] = ParamValue1;
if (size > 0)
{
// create a local stack object to store the arguments
unsigned StackObject = MFI->CreateStackObject(size / 8, alignment / 8, false);
SDValue FrameIndex = DAG.getFrameIndex(StackObject, getPointerTy());
// store each of the arguments to the stack in turn
for (unsigned int i = 1; i != OutVals.size(); i++) {
SDValue FrameAddr = DAG.getNode(ISD::ADD, dl, getPointerTy(), FrameIndex, DAG.getTargetConstant((i - 1) * 8, getPointerTy()));
Chain = DAG.getStore(Chain, dl, OutVals[i], FrameAddr,
MachinePointerInfo(),
false, false, 0);
}
// copy the address of the local frame index to get the address in non-local space
SDValue genericAddr = DAG.getNode(PTXISD::COPY_ADDRESS, dl, getPointerTy(), FrameIndex);
// store this address in the second argument
Chain = DAG.getNode(PTXISD::STORE_PARAM, dl, MVT::Other, Chain, ParamValue1, genericAddr);
}
}
else
{
// Generate STORE_PARAM nodes for each function argument. In PTX, function
// arguments are explicitly stored into .param variables and passed as
// arguments. There is no register/stack-based calling convention in PTX.
for (unsigned i = 0; i != OutVals.size(); ++i) {
unsigned Size = OutVals[i].getValueType().getSizeInBits();
unsigned Param = PM.addLocalParam(Size);
const std::string &ParamName = PM.getParamName(Param);
SDValue ParamValue = DAG.getTargetExternalSymbol(ParamName.c_str(),
MVT::Other);
Chain = DAG.getNode(PTXISD::STORE_PARAM, dl, MVT::Other, Chain,
ParamValue, OutVals[i]);
Ops[i+Ins.size()+4] = ParamValue;
}
}
std::vector<SDValue> InParams;
// Generate list of .param variables to hold the return value(s).
Ops[1] = DAG.getTargetConstant(Ins.size(), MVT::i32);
for (unsigned i = 0; i < Ins.size(); ++i) {
unsigned Size = Ins[i].VT.getStoreSizeInBits();
unsigned Param = PM.addLocalParam(Size);
const std::string &ParamName = PM.getParamName(Param);
SDValue ParamValue = DAG.getTargetExternalSymbol(ParamName.c_str(),
MVT::Other);
Ops[i+2] = ParamValue;
InParams.push_back(ParamValue);
}
Ops[0] = Chain;
// Create the CALL node.
Chain = DAG.getNode(PTXISD::CALL, dl, MVT::Other, &Ops[0], Ops.size());
// Create the LOAD_PARAM nodes that retrieve the function return value(s).
for (unsigned i = 0; i < Ins.size(); ++i) {
SDValue Load = DAG.getNode(PTXISD::LOAD_PARAM, dl, Ins[i].VT, Chain,
InParams[i]);
InVals.push_back(Load);
}
return Chain;
}
unsigned PTXTargetLowering::getNumRegisters(LLVMContext &Context, EVT VT) {
// All arguments consist of one "register," regardless of the type.
return 1;
}
Reference in New Issue View Git Blame Copy Permalink