llvm-6502/lib/Target/R600/AMDGPUTargetMachine.cpp
Tom Stellard c0b0c677a1 R600: Support for indirect addressing v4
Only implemented for R600 so far.  SI is missing implementations of a
few callbacks used by the Indirect Addressing pass and needs code to
handle frame indices.

At the moment R600 only supports array sizes of 16 dwords or less.
Register packing of vector types is currently disabled, which means that a
vec4 is stored in T0_X, T1_X, T2_X, T3_X, rather than T0_XYZW. In order
to correctly pack registers in all cases, we will need to implement an
analysis pass for R600 that determines the correct vector width for each
array.

v2:
  - Add support for i8 zext load from stack.
  - Coding style fixes

v3:
  - Don't reserve registers for indirect addressing when it isn't
    being used.
  - Fix bug caused by LLVM limiting the number of SubRegIndex
    declarations.

v4:
  - Fix 64-bit defines

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174525 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-06 17:32:29 +00:00

155 lines
4.6 KiB
C++

//===-- AMDGPUTargetMachine.cpp - TargetMachine for hw codegen targets-----===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
/// \file
/// \brief The AMDGPU target machine contains all of the hardware specific
/// information needed to emit code for R600 and SI GPUs.
//
//===----------------------------------------------------------------------===//
#include "AMDGPUTargetMachine.h"
#include "AMDGPU.h"
#include "R600ISelLowering.h"
#include "R600InstrInfo.h"
#include "SIISelLowering.h"
#include "SIInstrInfo.h"
#include "llvm/Analysis/Passes.h"
#include "llvm/Analysis/Verifier.h"
#include "llvm/CodeGen/MachineFunctionAnalysis.h"
#include "llvm/CodeGen/MachineModuleInfo.h"
#include "llvm/CodeGen/Passes.h"
#include "llvm/MC/MCAsmInfo.h"
#include "llvm/PassManager.h"
#include "llvm/Support/TargetRegistry.h"
#include "llvm/Support/raw_os_ostream.h"
#include "llvm/Transforms/IPO.h"
#include "llvm/Transforms/Scalar.h"
#include <llvm/CodeGen/Passes.h>
using namespace llvm;
extern "C" void LLVMInitializeR600Target() {
// Register the target
RegisterTargetMachine<AMDGPUTargetMachine> X(TheAMDGPUTarget);
}
AMDGPUTargetMachine::AMDGPUTargetMachine(const Target &T, StringRef TT,
StringRef CPU, StringRef FS,
TargetOptions Options,
Reloc::Model RM, CodeModel::Model CM,
CodeGenOpt::Level OptLevel
)
:
LLVMTargetMachine(T, TT, CPU, FS, Options, RM, CM, OptLevel),
Subtarget(TT, CPU, FS),
Layout(Subtarget.getDataLayout()),
FrameLowering(TargetFrameLowering::StackGrowsUp,
Subtarget.device()->getStackAlignment(), 0),
IntrinsicInfo(this),
InstrItins(&Subtarget.getInstrItineraryData()) {
// TLInfo uses InstrInfo so it must be initialized after.
if (Subtarget.device()->getGeneration() <= AMDGPUDeviceInfo::HD6XXX) {
InstrInfo = new R600InstrInfo(*this);
TLInfo = new R600TargetLowering(*this);
} else {
InstrInfo = new SIInstrInfo(*this);
TLInfo = new SITargetLowering(*this);
}
}
AMDGPUTargetMachine::~AMDGPUTargetMachine() {
}
namespace {
class AMDGPUPassConfig : public TargetPassConfig {
public:
AMDGPUPassConfig(AMDGPUTargetMachine *TM, PassManagerBase &PM)
: TargetPassConfig(TM, PM) {}
AMDGPUTargetMachine &getAMDGPUTargetMachine() const {
return getTM<AMDGPUTargetMachine>();
}
virtual bool addPreISel();
virtual bool addInstSelector();
virtual bool addPreRegAlloc();
virtual bool addPostRegAlloc();
virtual bool addPreSched2();
virtual bool addPreEmitPass();
};
} // End of anonymous namespace
TargetPassConfig *AMDGPUTargetMachine::createPassConfig(PassManagerBase &PM) {
return new AMDGPUPassConfig(this, PM);
}
bool
AMDGPUPassConfig::addPreISel() {
const AMDGPUSubtarget &ST = TM->getSubtarget<AMDGPUSubtarget>();
if (ST.device()->getGeneration() > AMDGPUDeviceInfo::HD6XXX) {
addPass(createAMDGPUStructurizeCFGPass());
addPass(createSIAnnotateControlFlowPass());
}
return false;
}
bool AMDGPUPassConfig::addInstSelector() {
addPass(createAMDGPUPeepholeOpt(*TM));
addPass(createAMDGPUISelDag(getAMDGPUTargetMachine()));
const AMDGPUSubtarget &ST = TM->getSubtarget<AMDGPUSubtarget>();
if (ST.device()->getGeneration() <= AMDGPUDeviceInfo::HD6XXX) {
// This callbacks this pass uses are not implemented yet on SI.
addPass(createAMDGPUIndirectAddressingPass(*TM));
}
return false;
}
bool AMDGPUPassConfig::addPreRegAlloc() {
const AMDGPUSubtarget &ST = TM->getSubtarget<AMDGPUSubtarget>();
if (ST.device()->getGeneration() > AMDGPUDeviceInfo::HD6XXX) {
addPass(createSIAssignInterpRegsPass(*TM));
}
addPass(createAMDGPUConvertToISAPass(*TM));
return false;
}
bool AMDGPUPassConfig::addPostRegAlloc() {
const AMDGPUSubtarget &ST = TM->getSubtarget<AMDGPUSubtarget>();
if (ST.device()->getGeneration() > AMDGPUDeviceInfo::HD6XXX) {
addPass(createSIInsertWaits(*TM));
}
return false;
}
bool AMDGPUPassConfig::addPreSched2() {
addPass(&IfConverterID);
return false;
}
bool AMDGPUPassConfig::addPreEmitPass() {
const AMDGPUSubtarget &ST = TM->getSubtarget<AMDGPUSubtarget>();
if (ST.device()->getGeneration() <= AMDGPUDeviceInfo::HD6XXX) {
addPass(createAMDGPUCFGPreparationPass(*TM));
addPass(createAMDGPUCFGStructurizerPass(*TM));
addPass(createR600ExpandSpecialInstrsPass(*TM));
addPass(&FinalizeMachineBundlesID);
addPass(createR600LowerConstCopy(*TM));
} else {
addPass(createSILowerLiteralConstantsPass(*TM));
addPass(createSILowerControlFlowPass(*TM));
}
return false;
}