/// however on the GPU, each address space points to
/// a separate piece of memory that is unique from other
/// memory locations.
-struct AMDGPUAS {
- // The following address space values depend on the triple environment.
- unsigned PRIVATE_ADDRESS; ///< Address space for private memory.
- unsigned FLAT_ADDRESS; ///< Address space for flat memory.
- unsigned REGION_ADDRESS; ///< Address space for region memory.
-
+namespace AMDGPUAS {
enum : unsigned {
// The maximum value for flat, generic, local, private, constant and region.
MAX_AMDGPU_ADDRESS = 6,
+ FLAT_ADDRESS = 0, ///< Address space for flat memory.
GLOBAL_ADDRESS = 1, ///< Address space for global memory (RAT0, VTX0).
+ REGION_ADDRESS = 2, ///< Address space for region memory.
+
CONSTANT_ADDRESS = 4, ///< Address space for constant memory (VTX2)
LOCAL_ADDRESS = 3, ///< Address space for local memory.
+ PRIVATE_ADDRESS = 5, ///< Address space for private memory.
CONSTANT_ADDRESS_32BIT = 6, ///< Address space for 32-bit constant memory
};
};
-namespace llvm {
-namespace AMDGPU {
-AMDGPUAS getAMDGPUAS(const Module &M);
-AMDGPUAS getAMDGPUAS(const TargetMachine &TM);
-AMDGPUAS getAMDGPUAS(Triple T);
-} // namespace AMDGPU
-} // namespace llvm
-
#endif
}
// Must match the table in getAliasResult.
-AMDGPUAAResult::ASAliasRulesTy::ASAliasRulesTy(AMDGPUAS AS_, Triple::ArchType Arch_)
- : Arch(Arch_), AS(AS_) {
+AMDGPUAAResult::ASAliasRulesTy::ASAliasRulesTy(Triple::ArchType Arch_)
+ : Arch(Arch_) {
// These arrarys are indexed by address space value
// enum elements 0 ... to 6
- static const AliasResult ASAliasRulesPrivIsZero[7][7] = {
- /* Private Global Constant Group Flat Region Constant 32-bit */
- /* Private */ {MayAlias, NoAlias , NoAlias , NoAlias , MayAlias, NoAlias , NoAlias},
- /* Global */ {NoAlias , MayAlias, MayAlias, NoAlias , MayAlias, NoAlias , MayAlias},
- /* Constant */ {NoAlias , MayAlias, MayAlias, NoAlias , MayAlias, NoAlias , MayAlias},
- /* Group */ {NoAlias , NoAlias , NoAlias , MayAlias, MayAlias, NoAlias , NoAlias},
- /* Flat */ {MayAlias, MayAlias, MayAlias, MayAlias, MayAlias, MayAlias, MayAlias},
- /* Region */ {NoAlias , NoAlias , NoAlias , NoAlias , MayAlias, MayAlias, NoAlias},
- /* Constant 32-bit */ {NoAlias , MayAlias, MayAlias, NoAlias , MayAlias, NoAlias , MayAlias}
- };
static const AliasResult ASAliasRulesGenIsZero[7][7] = {
/* Flat Global Region Group Constant Private Constant 32-bit */
/* Flat */ {MayAlias, MayAlias, MayAlias, MayAlias, MayAlias, MayAlias, MayAlias},
/* Private */ {MayAlias, NoAlias , NoAlias , NoAlias , NoAlias , MayAlias, NoAlias},
/* Constant 32-bit */ {MayAlias, MayAlias, MayAlias, NoAlias , MayAlias, NoAlias , NoAlias}
};
+
static_assert(AMDGPUAS::MAX_AMDGPU_ADDRESS <= 6, "Addr space out of range");
- if (AS.FLAT_ADDRESS == 0) {
- assert(AS.GLOBAL_ADDRESS == 1 &&
- AS.REGION_ADDRESS == 2 &&
- AS.LOCAL_ADDRESS == 3 &&
- AS.CONSTANT_ADDRESS == 4 &&
- AS.PRIVATE_ADDRESS == 5 &&
- AS.CONSTANT_ADDRESS_32BIT == 6);
- ASAliasRules = &ASAliasRulesGenIsZero;
- } else {
- assert(AS.PRIVATE_ADDRESS == 0 &&
- AS.GLOBAL_ADDRESS == 1 &&
- AS.CONSTANT_ADDRESS == 2 &&
- AS.LOCAL_ADDRESS == 3 &&
- AS.FLAT_ADDRESS == 4 &&
- AS.REGION_ADDRESS == 5 &&
- AS.CONSTANT_ADDRESS_32BIT == 6);
- ASAliasRules = &ASAliasRulesPrivIsZero;
- }
+
+ ASAliasRules = &ASAliasRulesGenIsZero;
}
AliasResult AMDGPUAAResult::ASAliasRulesTy::getAliasResult(unsigned AS1,
unsigned AS2) const {
- if (AS1 > AS.MAX_AMDGPU_ADDRESS || AS2 > AS.MAX_AMDGPU_ADDRESS) {
+ if (AS1 > AMDGPUAS::MAX_AMDGPU_ADDRESS || AS2 > AMDGPUAS::MAX_AMDGPU_ADDRESS) {
if (Arch == Triple::amdgcn)
report_fatal_error("Pointer address space out of range");
return AS1 == AS2 ? MayAlias : NoAlias;
bool AMDGPUAAResult::pointsToConstantMemory(const MemoryLocation &Loc,
bool OrLocal) {
const Value *Base = GetUnderlyingObject(Loc.Ptr, DL);
-
- if (Base->getType()->getPointerAddressSpace() == AS.CONSTANT_ADDRESS ||
- Base->getType()->getPointerAddressSpace() == AS.CONSTANT_ADDRESS_32BIT) {
+ unsigned AS = Base->getType()->getPointerAddressSpace();
+ if (AS == AMDGPUAS::CONSTANT_ADDRESS ||
+ AS == AMDGPUAS::CONSTANT_ADDRESS_32BIT) {
return true;
}
friend AAResultBase<AMDGPUAAResult>;
const DataLayout &DL;
- AMDGPUAS AS;
public:
explicit AMDGPUAAResult(const DataLayout &DL, Triple T) : AAResultBase(),
- DL(DL), AS(AMDGPU::getAMDGPUAS(T)), ASAliasRules(AS, T.getArch()) {}
+ DL(DL), ASAliasRules(T.getArch()) {}
AMDGPUAAResult(AMDGPUAAResult &&Arg)
- : AAResultBase(std::move(Arg)), DL(Arg.DL), AS(Arg.AS),
+ : AAResultBase(std::move(Arg)), DL(Arg.DL),
ASAliasRules(Arg.ASAliasRules){}
/// Handle invalidation events from the new pass manager.
class ASAliasRulesTy {
public:
- ASAliasRulesTy(AMDGPUAS AS_, Triple::ArchType Arch_);
+ ASAliasRulesTy(Triple::ArchType Arch_);
AliasResult getAliasResult(unsigned AS1, unsigned AS2) const;
private:
Triple::ArchType Arch;
- AMDGPUAS AS;
const AliasResult (*ASAliasRules)[7][7];
} ASAliasRules;
};
}
bool AMDGPUAlwaysInline::runOnModule(Module &M) {
- AMDGPUAS AMDGPUAS = AMDGPU::getAMDGPUAS(M);
-
std::vector<GlobalAlias*> AliasesToRemove;
SmallPtrSet<Function *, 8> FuncsToAlwaysInline;
for (GlobalVariable &GV : M.globals()) {
// TODO: Region address
unsigned AS = GV.getType()->getAddressSpace();
- if (AS != AMDGPUAS::LOCAL_ADDRESS && AS != AMDGPUAS.REGION_ADDRESS)
+ if (AS != AMDGPUAS::LOCAL_ADDRESS && AS != AMDGPUAS::REGION_ADDRESS)
continue;
recursivelyVisitUsers(GV, FuncsToAlwaysInline);
class AMDGPUAnnotateKernelFeatures : public CallGraphSCCPass {
private:
const TargetMachine *TM = nullptr;
- AMDGPUAS AS;
bool addFeatureAttributes(Function &F);
CallGraphSCCPass::getAnalysisUsage(AU);
}
- static bool visitConstantExpr(const ConstantExpr *CE, AMDGPUAS AS);
+ static bool visitConstantExpr(const ConstantExpr *CE);
static bool visitConstantExprsRecursively(
const Constant *EntryC,
- SmallPtrSet<const Constant *, 8> &ConstantExprVisited,
- AMDGPUAS AS);
+ SmallPtrSet<const Constant *, 8> &ConstantExprVisited);
};
} // end anonymous namespace
// The queue ptr is only needed when casting to flat, not from it.
-static bool castRequiresQueuePtr(unsigned SrcAS, const AMDGPUAS &AS) {
- return SrcAS == AS.LOCAL_ADDRESS || SrcAS == AS.PRIVATE_ADDRESS;
+static bool castRequiresQueuePtr(unsigned SrcAS) {
+ return SrcAS == AMDGPUAS::LOCAL_ADDRESS || SrcAS == AMDGPUAS::PRIVATE_ADDRESS;
}
-static bool castRequiresQueuePtr(const AddrSpaceCastInst *ASC,
- const AMDGPUAS &AS) {
- return castRequiresQueuePtr(ASC->getSrcAddressSpace(), AS);
+static bool castRequiresQueuePtr(const AddrSpaceCastInst *ASC) {
+ return castRequiresQueuePtr(ASC->getSrcAddressSpace());
}
-bool AMDGPUAnnotateKernelFeatures::visitConstantExpr(const ConstantExpr *CE,
- AMDGPUAS AS) {
+bool AMDGPUAnnotateKernelFeatures::visitConstantExpr(const ConstantExpr *CE) {
if (CE->getOpcode() == Instruction::AddrSpaceCast) {
unsigned SrcAS = CE->getOperand(0)->getType()->getPointerAddressSpace();
- return castRequiresQueuePtr(SrcAS, AS);
+ return castRequiresQueuePtr(SrcAS);
}
return false;
bool AMDGPUAnnotateKernelFeatures::visitConstantExprsRecursively(
const Constant *EntryC,
- SmallPtrSet<const Constant *, 8> &ConstantExprVisited,
- AMDGPUAS AS) {
+ SmallPtrSet<const Constant *, 8> &ConstantExprVisited) {
if (!ConstantExprVisited.insert(EntryC).second)
return false;
// Check this constant expression.
if (const auto *CE = dyn_cast<ConstantExpr>(C)) {
- if (visitConstantExpr(CE, AS))
+ if (visitConstantExpr(CE))
return true;
}
continue;
if (const AddrSpaceCastInst *ASC = dyn_cast<AddrSpaceCastInst>(&I)) {
- if (castRequiresQueuePtr(ASC, AS)) {
+ if (castRequiresQueuePtr(ASC)) {
NeedQueuePtr = true;
continue;
}
if (!OpC)
continue;
- if (visitConstantExprsRecursively(OpC, ConstantExprVisited, AS)) {
+ if (visitConstantExprsRecursively(OpC, ConstantExprVisited)) {
NeedQueuePtr = true;
break;
}
if (!TPC)
report_fatal_error("TargetMachine is required");
- AS = AMDGPU::getAMDGPUAS(CG.getModule());
TM = &TPC->getTM<TargetMachine>();
return false;
}
LoopInfo *LI;
DenseMap<Value*, GetElementPtrInst*> noClobberClones;
bool isKernelFunc;
- AMDGPUAS AMDGPUASI;
public:
static char ID;
if (!DA->isUniform(Ptr))
return;
auto isGlobalLoad = [&](LoadInst &Load)->bool {
- return Load.getPointerAddressSpace() == AMDGPUASI.GLOBAL_ADDRESS;
+ return Load.getPointerAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS;
};
// We're tracking up to the Function boundaries
// We cannot go beyond because of FunctionPass restrictions
}
bool AMDGPUAnnotateUniformValues::doInitialization(Module &M) {
- AMDGPUASI = AMDGPU::getAMDGPUAS(M);
return false;
}
AMDGPUAsmPrinter::AMDGPUAsmPrinter(TargetMachine &TM,
std::unique_ptr<MCStreamer> Streamer)
: AsmPrinter(TM, std::move(Streamer)) {
- AMDGPUASI = static_cast<AMDGPUTargetMachine*>(&TM)->getAMDGPUAS();
- }
+}
StringRef AMDGPUAsmPrinter::getPassName() const {
return "AMDGPU Assembly Printer";
protected:
mutable std::vector<std::string> DisasmLines, HexLines;
mutable size_t DisasmLineMaxLen;
- AMDGPUAS AMDGPUASI;
};
} // end namespace llvm
using namespace llvm;
AMDGPUCallLowering::AMDGPUCallLowering(const AMDGPUTargetLowering &TLI)
- : CallLowering(&TLI), AMDGPUASI(TLI.getAMDGPUAS()) {
+ : CallLowering(&TLI) {
}
bool AMDGPUCallLowering::lowerReturn(MachineIRBuilder &MIRBuilder,
MachineRegisterInfo &MRI = MF.getRegInfo();
const Function &F = MF.getFunction();
const DataLayout &DL = F.getParent()->getDataLayout();
- PointerType *PtrTy = PointerType::get(ParamTy, AMDGPUASI.CONSTANT_ADDRESS);
+ PointerType *PtrTy = PointerType::get(ParamTy, AMDGPUAS::CONSTANT_ADDRESS);
LLT PtrType = getLLTForType(*PtrTy, DL);
unsigned DstReg = MRI.createGenericVirtualRegister(PtrType);
unsigned KernArgSegmentPtr =
MachineFunction &MF = MIRBuilder.getMF();
const Function &F = MF.getFunction();
const DataLayout &DL = F.getParent()->getDataLayout();
- PointerType *PtrTy = PointerType::get(ParamTy, AMDGPUASI.CONSTANT_ADDRESS);
+ PointerType *PtrTy = PointerType::get(ParamTy, AMDGPUAS::CONSTANT_ADDRESS);
MachinePointerInfo PtrInfo(UndefValue::get(PtrTy));
unsigned TypeSize = DL.getTypeStoreSize(ParamTy);
unsigned PtrReg = lowerParameterPtr(MIRBuilder, ParamTy, Offset);
class AMDGPUTargetLowering;
class AMDGPUCallLowering: public CallLowering {
- AMDGPUAS AMDGPUASI;
-
unsigned lowerParameterPtr(MachineIRBuilder &MIRBuilder, Type *ParamTy,
uint64_t Offset) const;
LegacyDivergenceAnalysis *DA = nullptr;
Module *Mod = nullptr;
bool HasUnsafeFPMath = false;
- AMDGPUAS AMDGPUASI;
/// Copies exact/nsw/nuw flags (if any) from binary operation \p I to
/// binary operation \p V.
if (!WidenLoads)
return false;
- if ((I.getPointerAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS ||
- I.getPointerAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS_32BIT) &&
+ if ((I.getPointerAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS ||
+ I.getPointerAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS_32BIT) &&
canWidenScalarExtLoad(I)) {
IRBuilder<> Builder(&I);
Builder.SetCurrentDebugLocation(I.getDebugLoc());
AC = &getAnalysis<AssumptionCacheTracker>().getAssumptionCache(F);
DA = &getAnalysis<LegacyDivergenceAnalysis>();
HasUnsafeFPMath = hasUnsafeFPMath(F);
- AMDGPUASI = TM.getAMDGPUAS();
bool MadeChange = false;
AddressSpaceQualifier MetadataStreamer::getAddressSpaceQualifer(
unsigned AddressSpace) const {
- if (AddressSpace == AMDGPUASI.PRIVATE_ADDRESS)
+ if (AddressSpace == AMDGPUAS::PRIVATE_ADDRESS)
return AddressSpaceQualifier::Private;
- if (AddressSpace == AMDGPUASI.GLOBAL_ADDRESS)
+ if (AddressSpace == AMDGPUAS::GLOBAL_ADDRESS)
return AddressSpaceQualifier::Global;
- if (AddressSpace == AMDGPUASI.CONSTANT_ADDRESS)
+ if (AddressSpace == AMDGPUAS::CONSTANT_ADDRESS)
return AddressSpaceQualifier::Constant;
- if (AddressSpace == AMDGPUASI.LOCAL_ADDRESS)
+ if (AddressSpace == AMDGPUAS::LOCAL_ADDRESS)
return AddressSpaceQualifier::Local;
- if (AddressSpace == AMDGPUASI.FLAT_ADDRESS)
+ if (AddressSpace == AMDGPUAS::FLAT_ADDRESS)
return AddressSpaceQualifier::Generic;
- if (AddressSpace == AMDGPUASI.REGION_ADDRESS)
+ if (AddressSpace == AMDGPUAS::REGION_ADDRESS)
return AddressSpaceQualifier::Region;
llvm_unreachable("Unknown address space qualifier");
.Case("queue_t", ValueKind::Queue)
.Default(isa<PointerType>(Ty) ?
(Ty->getPointerAddressSpace() ==
- AMDGPUASI.LOCAL_ADDRESS ?
+ AMDGPUAS::LOCAL_ADDRESS ?
ValueKind::DynamicSharedPointer :
ValueKind::GlobalBuffer) :
ValueKind::ByValue);
unsigned PointeeAlign = 0;
if (auto PtrTy = dyn_cast<PointerType>(Ty)) {
- if (PtrTy->getAddressSpace() == AMDGPUASI.LOCAL_ADDRESS) {
+ if (PtrTy->getAddressSpace() == AMDGPUAS::LOCAL_ADDRESS) {
PointeeAlign = Arg.getParamAlignment();
if (PointeeAlign == 0)
PointeeAlign = DL.getABITypeAlignment(PtrTy->getElementType());
emitKernelArg(DL, Int64Ty, ValueKind::HiddenGlobalOffsetZ);
auto Int8PtrTy = Type::getInt8PtrTy(Func.getContext(),
- AMDGPUASI.GLOBAL_ADDRESS);
+ AMDGPUAS::GLOBAL_ADDRESS);
// Emit "printf buffer" argument if printf is used, otherwise emit dummy
// "none" argument.
}
void MetadataStreamer::begin(const Module &Mod) {
- AMDGPUASI = getAMDGPUAS(Mod);
emitVersion();
emitPrintf(Mod);
}
class MetadataStreamer final {
private:
Metadata HSAMetadata;
- AMDGPUAS AMDGPUASI;
void dump(StringRef HSAMetadataString) const;
// Subtarget - Keep a pointer to the AMDGPU Subtarget around so that we can
// make the right decision when generating code for different targets.
const GCNSubtarget *Subtarget;
- AMDGPUAS AMDGPUASI;
bool EnableLateStructurizeCFG;
public:
explicit AMDGPUDAGToDAGISel(TargetMachine *TM = nullptr,
CodeGenOpt::Level OptLevel = CodeGenOpt::Default)
: SelectionDAGISel(*TM, OptLevel) {
- AMDGPUASI = AMDGPU::getAMDGPUAS(*TM);
EnableLateStructurizeCFG = AMDGPUTargetMachine::EnableLateStructurizeCFG;
}
~AMDGPUDAGToDAGISel() override = default;
class R600DAGToDAGISel : public AMDGPUDAGToDAGISel {
const R600Subtarget *Subtarget;
- AMDGPUAS AMDGPUASI;
bool isConstantLoad(const MemSDNode *N, int cbID) const;
bool SelectGlobalValueConstantOffset(SDValue Addr, SDValue& IntPtr);
SDValue& Offset);
public:
explicit R600DAGToDAGISel(TargetMachine *TM, CodeGenOpt::Level OptLevel) :
- AMDGPUDAGToDAGISel(TM, OptLevel) {
- AMDGPUASI = AMDGPU::getAMDGPUAS(*TM);
- }
+ AMDGPUDAGToDAGISel(TM, OptLevel) {}
void Select(SDNode *N) override;
}
SDNode *AMDGPUDAGToDAGISel::glueCopyToM0(SDNode *N) const {
- if (cast<MemSDNode>(N)->getAddressSpace() != AMDGPUASI.LOCAL_ADDRESS ||
+ if (cast<MemSDNode>(N)->getAddressSpace() != AMDGPUAS::LOCAL_ADDRESS ||
!Subtarget->ldsRequiresM0Init())
return N;
void AMDGPUDAGToDAGISel::SelectATOMIC_CMP_SWAP(SDNode *N) {
MemSDNode *Mem = cast<MemSDNode>(N);
unsigned AS = Mem->getAddressSpace();
- if (AS == AMDGPUASI.FLAT_ADDRESS) {
+ if (AS == AMDGPUAS::FLAT_ADDRESS) {
SelectCode(N);
return;
}
if (!N->readMem())
return false;
if (CbId == -1)
- return N->getAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS ||
- N->getAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS_32BIT;
+ return N->getAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS ||
+ N->getAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS_32BIT;
- return N->getAddressSpace() == AMDGPUASI.CONSTANT_BUFFER_0 + CbId;
+ return N->getAddressSpace() == AMDGPUAS::CONSTANT_BUFFER_0 + CbId;
}
bool R600DAGToDAGISel::SelectGlobalValueConstantOffset(SDValue Addr,
AMDGPUTargetLowering::AMDGPUTargetLowering(const TargetMachine &TM,
const AMDGPUSubtarget &STI)
: TargetLowering(TM), Subtarget(&STI) {
- AMDGPUASI = AMDGPU::getAMDGPUAS(TM);
// Lower floating point store/load to integer store/load to reduce the number
// of patterns in tablegen.
setOperationAction(ISD::LOAD, MVT::f32, Promote);
{
const LoadSDNode * L = dyn_cast<LoadSDNode>(N);
if (L->getMemOperand()->getAddrSpace()
- == AMDGPUASI.CONSTANT_ADDRESS_32BIT)
+ == AMDGPUAS::CONSTANT_ADDRESS_32BIT)
return true;
return false;
}
GlobalAddressSDNode *G = cast<GlobalAddressSDNode>(Op);
const GlobalValue *GV = G->getGlobal();
- if (G->getAddressSpace() == AMDGPUASI.LOCAL_ADDRESS ||
- G->getAddressSpace() == AMDGPUASI.REGION_ADDRESS) {
+ if (G->getAddressSpace() == AMDGPUAS::LOCAL_ADDRESS ||
+ G->getAddressSpace() == AMDGPUAS::REGION_ADDRESS) {
if (!MFI->isEntryFunction()) {
const Function &Fn = DAG.getMachineFunction().getFunction();
DiagnosticInfoUnsupported BadLDSDecl(
static unsigned numBitsSigned(SDValue Op, SelectionDAG &DAG);
protected:
- AMDGPUAS AMDGPUASI;
-
SDValue LowerEXTRACT_SUBVECTOR(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerCONCAT_VECTORS(SDValue Op, SelectionDAG &DAG) const;
/// Split a vector store into multiple scalar stores.
uint32_t getImplicitParameterOffset(const MachineFunction &MF,
const ImplicitParameter Param) const;
- AMDGPUAS getAMDGPUAS() const {
- return AMDGPUASI;
- }
-
MVT getFenceOperandTy(const DataLayout &DL) const override {
return MVT::i32;
}
if (!Callee)
return (unsigned)Thres;
- const AMDGPUAS AS = AMDGPU::getAMDGPUAS(*Caller->getParent());
-
// If we have a pointer to private array passed into a function
// it will not be optimized out, leaving scratch usage.
// Increase the inline threshold to allow inliniting in this case.
for (Value *PtrArg : CS.args()) {
Type *Ty = PtrArg->getType();
if (!Ty->isPointerTy() ||
- Ty->getPointerAddressSpace() != AS.PRIVATE_ADDRESS)
+ Ty->getPointerAddressSpace() != AMDGPUAS::PRIVATE_ADDRESS)
continue;
PtrArg = GetUnderlyingObject(PtrArg, DL);
if (const AllocaInst *AI = dyn_cast<AllocaInst>(PtrArg)) {
#define GET_GLOBALISEL_TEMPORARIES_INIT
#include "AMDGPUGenGlobalISel.inc"
#undef GET_GLOBALISEL_TEMPORARIES_INIT
- ,AMDGPUASI(STI.getAMDGPUAS())
{
}
if (!I.hasOneMemOperand())
return false;
- if ((*I.memoperands_begin())->getAddrSpace() != AMDGPUASI.CONSTANT_ADDRESS &&
- (*I.memoperands_begin())->getAddrSpace() != AMDGPUASI.CONSTANT_ADDRESS_32BIT)
+ if ((*I.memoperands_begin())->getAddrSpace() != AMDGPUAS::CONSTANT_ADDRESS &&
+ (*I.memoperands_begin())->getAddrSpace() != AMDGPUAS::CONSTANT_ADDRESS_32BIT)
return false;
if (!isInstrUniform(I))
#define GET_GLOBALISEL_TEMPORARIES_DECL
#include "AMDGPUGenGlobalISel.inc"
#undef GET_GLOBALISEL_TEMPORARIES_DECL
-
-protected:
- AMDGPUAS AMDGPUASI;
};
} // End llvm namespace.
>;
class PrivateAddress : CodePatPred<[{
- return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUASI.PRIVATE_ADDRESS;
+ return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::PRIVATE_ADDRESS;
}]>;
class ConstantAddress : CodePatPred<[{
- return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS;
+ return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS;
}]>;
class LocalAddress : CodePatPred<[{
- return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUASI.LOCAL_ADDRESS;
+ return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::LOCAL_ADDRESS;
}]>;
class GlobalAddress : CodePatPred<[{
- return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUASI.GLOBAL_ADDRESS;
+ return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS;
}]>;
class GlobalLoadAddress : CodePatPred<[{
auto AS = cast<MemSDNode>(N)->getAddressSpace();
- return AS == AMDGPUASI.GLOBAL_ADDRESS || AS == AMDGPUASI.CONSTANT_ADDRESS;
+ return AS == AMDGPUAS::GLOBAL_ADDRESS || AS == AMDGPUAS::CONSTANT_ADDRESS;
}]>;
class FlatLoadAddress : CodePatPred<[{
const auto AS = cast<MemSDNode>(N)->getAddressSpace();
- return AS == AMDGPUASI.FLAT_ADDRESS ||
- AS == AMDGPUASI.GLOBAL_ADDRESS ||
- AS == AMDGPUASI.CONSTANT_ADDRESS;
+ return AS == AMDGPUAS::FLAT_ADDRESS ||
+ AS == AMDGPUAS::GLOBAL_ADDRESS ||
+ AS == AMDGPUAS::CONSTANT_ADDRESS;
}]>;
class FlatStoreAddress : CodePatPred<[{
const auto AS = cast<MemSDNode>(N)->getAddressSpace();
- return AS == AMDGPUASI.FLAT_ADDRESS ||
- AS == AMDGPUASI.GLOBAL_ADDRESS;
+ return AS == AMDGPUAS::FLAT_ADDRESS ||
+ AS == AMDGPUAS::GLOBAL_ADDRESS;
}]>;
class AZExtLoadBase <SDPatternOperator ld_node>: PatFrag<(ops node:$ptr),
class local_binary_atomic_op<SDNode atomic_op> :
PatFrag<(ops node:$ptr, node:$value),
(atomic_op node:$ptr, node:$value), [{
- return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUASI.LOCAL_ADDRESS;
+ return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::LOCAL_ADDRESS;
}]>;
def atomic_swap_local : local_binary_atomic_op<atomic_swap>;
def mskor_global : PatFrag<(ops node:$val, node:$ptr),
(AMDGPUstore_mskor node:$val, node:$ptr), [{
- return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUASI.GLOBAL_ADDRESS;
+ return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS;
}]>;
class AtomicCmpSwapLocal <SDNode cmp_swap_node> : PatFrag<
(ops node:$ptr, node:$cmp, node:$swap),
(cmp_swap_node node:$ptr, node:$cmp, node:$swap), [{
AtomicSDNode *AN = cast<AtomicSDNode>(N);
- return AN->getAddressSpace() == AMDGPUASI.LOCAL_ADDRESS;
+ return AN->getAddressSpace() == AMDGPUAS::LOCAL_ADDRESS;
}]>;
def atomic_cmp_swap_local : AtomicCmpSwapLocal <atomic_cmp_swap>;
def "" : PatFrag<
(ops node:$ptr, node:$value),
(atomic_op node:$ptr, node:$value),
- [{return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUASI.GLOBAL_ADDRESS;}]>;
+ [{return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS;}]>;
def _noret : PatFrag<
(ops node:$ptr, node:$value),
(atomic_op node:$ptr, node:$value),
- [{return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUASI.GLOBAL_ADDRESS && (SDValue(N, 0).use_empty());}]>;
+ [{return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS && (SDValue(N, 0).use_empty());}]>;
def _ret : PatFrag<
(ops node:$ptr, node:$value),
(atomic_op node:$ptr, node:$value),
- [{return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUASI.GLOBAL_ADDRESS && (!SDValue(N, 0).use_empty());}]>;
+ [{return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS && (!SDValue(N, 0).use_empty());}]>;
}
defm atomic_swap_global : global_binary_atomic_op<atomic_swap>;
def atomic_cmp_swap_global_noret : PatFrag<
(ops node:$ptr, node:$cmp, node:$value),
(atomic_cmp_swap node:$ptr, node:$cmp, node:$value),
- [{return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUASI.GLOBAL_ADDRESS && (SDValue(N, 0).use_empty());}]>;
+ [{return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS && (SDValue(N, 0).use_empty());}]>;
def atomic_cmp_swap_global_ret : PatFrag<
(ops node:$ptr, node:$cmp, node:$value),
(atomic_cmp_swap node:$ptr, node:$cmp, node:$value),
- [{return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUASI.GLOBAL_ADDRESS && (!SDValue(N, 0).use_empty());}]>;
+ [{return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS && (!SDValue(N, 0).use_empty());}]>;
//===----------------------------------------------------------------------===//
// Misc Pattern Fragments
return LLT::pointer(AS, TM.getPointerSizeInBits(AS));
};
- auto AMDGPUAS = ST.getAMDGPUAS();
-
const LLT S1 = LLT::scalar(1);
const LLT V2S16 = LLT::vector(2, 16);
const LLT GlobalPtr = GetAddrSpacePtr(AMDGPUAS::GLOBAL_ADDRESS);
const LLT ConstantPtr = GetAddrSpacePtr(AMDGPUAS::CONSTANT_ADDRESS);
const LLT LocalPtr = GetAddrSpacePtr(AMDGPUAS::LOCAL_ADDRESS);
- const LLT FlatPtr = GetAddrSpacePtr(AMDGPUAS.FLAT_ADDRESS);
- const LLT PrivatePtr = GetAddrSpacePtr(AMDGPUAS.PRIVATE_ADDRESS);
+ const LLT FlatPtr = GetAddrSpacePtr(AMDGPUAS::FLAT_ADDRESS);
+ const LLT PrivatePtr = GetAddrSpacePtr(AMDGPUAS::PRIVATE_ADDRESS);
const LLT AddrSpaces[] = {
GlobalPtr,
// for OpenCL 2.0 we have only generic implementation of sincos
// function.
AMDGPULibFunc nf(AMDGPULibFunc::EI_SINCOS, fInfo);
- const AMDGPUAS AS = AMDGPU::getAMDGPUAS(*M);
- nf.getLeads()[0].PtrKind = AMDGPULibFunc::getEPtrKindFromAddrSpace(AS.FLAT_ADDRESS);
+ nf.getLeads()[0].PtrKind = AMDGPULibFunc::getEPtrKindFromAddrSpace(AMDGPUAS::FLAT_ADDRESS);
Function *Fsincos = dyn_cast_or_null<Function>(getFunction(M, nf));
if (!Fsincos) return false;
// The allocaInst allocates the memory in private address space. This need
// to be bitcasted to point to the address space of cos pointer type.
// In OpenCL 2.0 this is generic, while in 1.2 that is private.
- if (PTy->getPointerAddressSpace() != AS.PRIVATE_ADDRESS)
+ if (PTy->getPointerAddressSpace() != AMDGPUAS::PRIVATE_ADDRESS)
P = B.CreateAddrSpaceCast(Alloc, PTy);
CallInst *Call = CreateCallEx2(B, Fsincos, UI->getArgOperand(0), P);
const DataLayout *DL;
- AMDGPUAS AS;
-
const TargetLowering *TLI;
void visit(const Function &F);
const Module &M = *F.getParent();
DL = &M.getDataLayout();
- AS = AMDGPU::getAMDGPUAS(M);
visit(F);
auto Loc = FIM.find(&F);
if (auto PT = dyn_cast<PointerType>(V->getType())) {
unsigned As = PT->getAddressSpace();
// Flat likely points to global too.
- return As == AS.GLOBAL_ADDRESS || As == AS.FLAT_ADDRESS;
+ return As == AMDGPUAS::GLOBAL_ADDRESS || As == AMDGPUAS::FLAT_ADDRESS;
}
return false;
}
bool AMDGPUPerfHint::isLocalAddr(const Value *V) const {
if (auto PT = dyn_cast<PointerType>(V->getType()))
- return PT->getAddressSpace() == AS.LOCAL_ADDRESS;
+ return PT->getAddressSpace() == AMDGPUAS::LOCAL_ADDRESS;
return false;
}
bool AMDGPUPerfHint::isConstantAddr(const Value *V) const {
if (auto PT = dyn_cast<PointerType>(V->getType())) {
unsigned As = PT->getAddressSpace();
- return As == AS.CONSTANT_ADDRESS || As == AS.CONSTANT_ADDRESS_32BIT;
+ return As == AMDGPUAS::CONSTANT_ADDRESS ||
+ As == AMDGPUAS::CONSTANT_ADDRESS_32BIT;
}
return false;
}
const TargetMachine *TM;
Module *Mod = nullptr;
const DataLayout *DL = nullptr;
- AMDGPUAS AS;
// FIXME: This should be per-kernel.
uint32_t LocalMemLimit = 0;
if (!ST.isPromoteAllocaEnabled())
return false;
- AS = AMDGPU::getAMDGPUAS(*F.getParent());
-
bool SufficientLDS = hasSufficientLocalMem(F);
bool Changed = false;
BasicBlock &EntryBB = *F.begin();
Type *I32Ty = Type::getInt32Ty(Mod->getContext());
Value *CastDispatchPtr = Builder.CreateBitCast(
- DispatchPtr, PointerType::get(I32Ty, AS.CONSTANT_ADDRESS));
+ DispatchPtr, PointerType::get(I32Ty, AMDGPUAS::CONSTANT_ADDRESS));
// We could do a single 64-bit load here, but it's likely that the basic
// 32-bit and extract sequence is already present, and it is probably easier
}
}
-static bool tryPromoteAllocaToVector(AllocaInst *Alloca, AMDGPUAS AS) {
+static bool tryPromoteAllocaToVector(AllocaInst *Alloca) {
if (DisablePromoteAllocaToVector) {
LLVM_DEBUG(dbgs() << " Promotion alloca to vector is disabled\n");
IRBuilder<> Builder(Inst);
switch (Inst->getOpcode()) {
case Instruction::Load: {
- Type *VecPtrTy = VectorTy->getPointerTo(AS.PRIVATE_ADDRESS);
+ Type *VecPtrTy = VectorTy->getPointerTo(AMDGPUAS::PRIVATE_ADDRESS);
Value *Ptr = cast<LoadInst>(Inst)->getPointerOperand();
Value *Index = calculateVectorIndex(Ptr, GEPVectorIdx);
break;
}
case Instruction::Store: {
- Type *VecPtrTy = VectorTy->getPointerTo(AS.PRIVATE_ADDRESS);
+ Type *VecPtrTy = VectorTy->getPointerTo(AMDGPUAS::PRIVATE_ADDRESS);
StoreInst *SI = cast<StoreInst>(Inst);
Value *Ptr = SI->getPointerOperand();
// we cannot use local memory in the pass.
for (Type *ParamTy : FTy->params()) {
PointerType *PtrTy = dyn_cast<PointerType>(ParamTy);
- if (PtrTy && PtrTy->getAddressSpace() == AS.LOCAL_ADDRESS) {
+ if (PtrTy && PtrTy->getAddressSpace() == AMDGPUAS::LOCAL_ADDRESS) {
LocalMemLimit = 0;
LLVM_DEBUG(dbgs() << "Function has local memory argument. Promoting to "
"local memory disabled.\n");
// Check how much local memory is being used by global objects
CurrentLocalMemUsage = 0;
for (GlobalVariable &GV : Mod->globals()) {
- if (GV.getType()->getAddressSpace() != AS.LOCAL_ADDRESS)
+ if (GV.getType()->getAddressSpace() != AMDGPUAS::LOCAL_ADDRESS)
continue;
for (const User *U : GV.users()) {
LLVM_DEBUG(dbgs() << "Trying to promote " << I << '\n');
- if (tryPromoteAllocaToVector(&I, AS))
+ if (tryPromoteAllocaToVector(&I))
return true; // Promoted to vector.
const Function &ContainingFunction = *I.getParent()->getParent();
Twine(F->getName()) + Twine('.') + I.getName(),
nullptr,
GlobalVariable::NotThreadLocal,
- AS.LOCAL_ADDRESS);
+ AMDGPUAS::LOCAL_ADDRESS);
GV->setUnnamedAddr(GlobalValue::UnnamedAddr::Global);
GV->setAlignment(I.getAlignment());
if (ICmpInst *CI = dyn_cast<ICmpInst>(V)) {
Value *Src0 = CI->getOperand(0);
Type *EltTy = Src0->getType()->getPointerElementType();
- PointerType *NewTy = PointerType::get(EltTy, AS.LOCAL_ADDRESS);
+ PointerType *NewTy = PointerType::get(EltTy, AMDGPUAS::LOCAL_ADDRESS);
if (isa<ConstantPointerNull>(CI->getOperand(0)))
CI->setOperand(0, ConstantPointerNull::get(NewTy));
continue;
Type *EltTy = V->getType()->getPointerElementType();
- PointerType *NewTy = PointerType::get(EltTy, AS.LOCAL_ADDRESS);
+ PointerType *NewTy = PointerType::get(EltTy, AMDGPUAS::LOCAL_ADDRESS);
// FIXME: It doesn't really make sense to try to do this for all
// instructions.
Type *SrcTy = Src->getType()->getPointerElementType();
Function *ObjectSize = Intrinsic::getDeclaration(Mod,
Intrinsic::objectsize,
- { Intr->getType(), PointerType::get(SrcTy, AS.LOCAL_ADDRESS) }
+ { Intr->getType(), PointerType::get(SrcTy, AMDGPUAS::LOCAL_ADDRESS) }
);
CallInst *NewCall = Builder.CreateCall(
InstrInfo(initializeSubtargetDependencies(TT, GPU, FS)),
TLInfo(TM, *this),
FrameLowering(TargetFrameLowering::StackGrowsUp, getStackAlignment(), 0) {
- AS = AMDGPU::getAMDGPUAS(TT);
CallLoweringInfo.reset(new AMDGPUCallLowering(*getTargetLowering()));
Legalizer.reset(new AMDGPULegalizerInfo(*this, TM));
RegBankInfo.reset(new AMDGPURegisterBankInfo(*getRegisterInfo()));
TexVTXClauseSize(0),
Gen(R600),
TLInfo(TM, initializeSubtargetDependencies(TT, GPU, FS)),
- InstrItins(getInstrItineraryForCPU(GPU)),
- AS (AMDGPU::getAMDGPUAS(TT)) { }
+ InstrItins(getInstrItineraryForCPU(GPU)) { }
void GCNSubtarget::overrideSchedPolicy(MachineSchedPolicy &Policy,
unsigned NumRegionInstrs) const {
bool FeatureDisable;
SelectionDAGTargetInfo TSInfo;
- AMDGPUAS AS;
private:
SIInstrInfo InstrInfo;
SITargetLowering TLInfo;
return MaxPrivateElementSize;
}
- AMDGPUAS getAMDGPUAS() const {
- return AS;
- }
-
bool hasIntClamp() const {
return HasIntClamp;
}
R600TargetLowering TLInfo;
InstrItineraryData InstrItins;
SelectionDAGTargetInfo TSInfo;
- AMDGPUAS AS;
public:
R600Subtarget(const Triple &TT, StringRef CPU, StringRef FS,
short getTexVTXClauseSize() const { return TexVTXClauseSize; }
- AMDGPUAS getAMDGPUAS() const { return AS; }
-
bool enableMachineScheduler() const override {
return true;
}
FS, Options, getEffectiveRelocModel(RM),
getEffectiveCodeModel(CM), OptLevel),
TLOF(createTLOF(getTargetTriple())) {
- AS = AMDGPU::getAMDGPUAS(TT);
initAsmInfo();
}
class AMDGPUTargetMachine : public LLVMTargetMachine {
protected:
std::unique_ptr<TargetLoweringObjectFile> TLOF;
- AMDGPUAS AS;
StringRef getGPUName(const Function &F) const;
StringRef getFeatureString(const Function &F) const;
TargetLoweringObjectFile *getObjFileLowering() const override {
return TLOF.get();
}
- AMDGPUAS getAMDGPUAS() const {
- return AS;
- }
void adjustPassManager(PassManagerBuilder &) override;
+
/// Get the integer value of a null pointer in the given address space.
uint64_t getNullPointerValue(unsigned AddrSpace) const {
- if (AddrSpace == AS.LOCAL_ADDRESS || AddrSpace == AS.REGION_ADDRESS)
- return -1;
- return 0;
+ return (AddrSpace == AMDGPUAS::LOCAL_ADDRESS ||
+ AddrSpace == AMDGPUAS::REGION_ADDRESS) ? -1 : 0;
}
};
unsigned ThresholdPrivate = UnrollThresholdPrivate;
unsigned ThresholdLocal = UnrollThresholdLocal;
unsigned MaxBoost = std::max(ThresholdPrivate, ThresholdLocal);
- const AMDGPUAS &ASST = AMDGPU::getAMDGPUAS(TargetTriple);
for (const BasicBlock *BB : L->getBlocks()) {
const DataLayout &DL = BB->getModule()->getDataLayout();
unsigned LocalGEPsSeen = 0;
unsigned AS = GEP->getAddressSpace();
unsigned Threshold = 0;
- if (AS == ASST.PRIVATE_ADDRESS)
+ if (AS == AMDGPUAS::PRIVATE_ADDRESS)
Threshold = ThresholdPrivate;
- else if (AS == ASST.LOCAL_ADDRESS)
+ else if (AS == AMDGPUAS::LOCAL_ADDRESS)
Threshold = ThresholdLocal;
else
continue;
if (UP.Threshold >= Threshold)
continue;
- if (AS == ASST.PRIVATE_ADDRESS) {
+ if (AS == AMDGPUAS::PRIVATE_ADDRESS) {
const Value *Ptr = GEP->getPointerOperand();
const AllocaInst *Alloca =
dyn_cast<AllocaInst>(GetUnderlyingObject(Ptr, DL));
unsigned AllocaSize = Ty->isSized() ? DL.getTypeAllocSize(Ty) : 0;
if (AllocaSize > MaxAlloca)
continue;
- } else if (AS == ASST.LOCAL_ADDRESS) {
+ } else if (AS == AMDGPUAS::LOCAL_ADDRESS) {
LocalGEPsSeen++;
// Inhibit unroll for local memory if we have seen addressing not to
// a variable, most likely we will be unable to combine it.
}
unsigned GCNTTIImpl::getLoadStoreVecRegBitWidth(unsigned AddrSpace) const {
- AMDGPUAS AS = ST->getAMDGPUAS();
- if (AddrSpace == AS.GLOBAL_ADDRESS ||
- AddrSpace == AS.CONSTANT_ADDRESS ||
- AddrSpace == AS.CONSTANT_ADDRESS_32BIT) {
+ if (AddrSpace == AMDGPUAS::GLOBAL_ADDRESS ||
+ AddrSpace == AMDGPUAS::CONSTANT_ADDRESS ||
+ AddrSpace == AMDGPUAS::CONSTANT_ADDRESS_32BIT) {
return 512;
}
- if (AddrSpace == AS.FLAT_ADDRESS ||
- AddrSpace == AS.LOCAL_ADDRESS ||
- AddrSpace == AS.REGION_ADDRESS)
+ if (AddrSpace == AMDGPUAS::FLAT_ADDRESS ||
+ AddrSpace == AMDGPUAS::LOCAL_ADDRESS ||
+ AddrSpace == AMDGPUAS::REGION_ADDRESS)
return 128;
- if (AddrSpace == AS.PRIVATE_ADDRESS)
+ if (AddrSpace == AMDGPUAS::PRIVATE_ADDRESS)
return 8 * ST->getMaxPrivateElementSize();
llvm_unreachable("unhandled address space");
// We allow vectorization of flat stores, even though we may need to decompose
// them later if they may access private memory. We don't have enough context
// here, and legalization can handle it.
- if (AddrSpace == ST->getAMDGPUAS().PRIVATE_ADDRESS) {
+ if (AddrSpace == AMDGPUAS::PRIVATE_ADDRESS) {
return (Alignment >= 4 || ST->hasUnalignedScratchAccess()) &&
ChainSizeInBytes <= ST->getMaxPrivateElementSize();
}
// All other loads are not divergent, because if threads issue loads with the
// same arguments, they will always get the same result.
if (const LoadInst *Load = dyn_cast<LoadInst>(V))
- return Load->getPointerAddressSpace() ==
- ST->getAMDGPUAS().PRIVATE_ADDRESS ||
- Load->getPointerAddressSpace() == ST->getAMDGPUAS().FLAT_ADDRESS;
+ return Load->getPointerAddressSpace() == AMDGPUAS::PRIVATE_ADDRESS ||
+ Load->getPointerAddressSpace() == AMDGPUAS::FLAT_ADDRESS;
// Atomics are divergent because they are executed sequentially: when an
// atomic operation refers to the same address in each thread, then each
}
unsigned R600TTIImpl::getLoadStoreVecRegBitWidth(unsigned AddrSpace) const {
- AMDGPUAS AS = ST->getAMDGPUAS();
- if (AddrSpace == AS.GLOBAL_ADDRESS ||
- AddrSpace == AS.CONSTANT_ADDRESS)
+ if (AddrSpace == AMDGPUAS::GLOBAL_ADDRESS ||
+ AddrSpace == AMDGPUAS::CONSTANT_ADDRESS)
return 128;
- if (AddrSpace == AS.LOCAL_ADDRESS ||
- AddrSpace == AS.REGION_ADDRESS)
+ if (AddrSpace == AMDGPUAS::LOCAL_ADDRESS ||
+ AddrSpace == AMDGPUAS::REGION_ADDRESS)
return 64;
- if (AddrSpace == AS.PRIVATE_ADDRESS)
+ if (AddrSpace == AMDGPUAS::PRIVATE_ADDRESS)
return 32;
- if ((AddrSpace == AS.PARAM_D_ADDRESS ||
- AddrSpace == AS.PARAM_I_ADDRESS ||
- (AddrSpace >= AS.CONSTANT_BUFFER_0 &&
- AddrSpace <= AS.CONSTANT_BUFFER_15)))
+ if ((AddrSpace == AMDGPUAS::PARAM_D_ADDRESS ||
+ AddrSpace == AMDGPUAS::PARAM_I_ADDRESS ||
+ (AddrSpace >= AMDGPUAS::CONSTANT_BUFFER_0 &&
+ AddrSpace <= AMDGPUAS::CONSTANT_BUFFER_15)))
return 128;
llvm_unreachable("unhandled address space");
}
// We allow vectorization of flat stores, even though we may need to decompose
// them later if they may access private memory. We don't have enough context
// here, and legalization can handle it.
- if (AddrSpace == ST->getAMDGPUAS().PRIVATE_ADDRESS)
- return false;
- return true;
+ return (AddrSpace != AMDGPUAS::PRIVATE_ADDRESS);
}
bool R600TTIImpl::isLegalToVectorizeLoadChain(unsigned ChainSizeInBytes,
if (IsGraphicsShader)
return -1;
return ST->hasFlatAddressSpace() ?
- ST->getAMDGPUAS().FLAT_ADDRESS : ST->getAMDGPUAS().UNKNOWN_ADDRESS_SPACE;
+ AMDGPUAS::FLAT_ADDRESS : AMDGPUAS::UNKNOWN_ADDRESS_SPACE;
}
unsigned getVectorSplitCost() { return 0; }
class MubufLoad <SDPatternOperator op> : PatFrag <
(ops node:$ptr), (op node:$ptr), [{
auto const AS = cast<MemSDNode>(N)->getAddressSpace();
- return AS == AMDGPUASI.GLOBAL_ADDRESS ||
- AS == AMDGPUASI.CONSTANT_ADDRESS;
+ return AS == AMDGPUAS::GLOBAL_ADDRESS ||
+ AS == AMDGPUAS::CONSTANT_ADDRESS;
}]>;
def mubuf_load : MubufLoad <load>;
class flat_binary_atomic_op<SDNode atomic_op> : PatFrag<
(ops node:$ptr, node:$value),
(atomic_op node:$ptr, node:$value),
- [{return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUASI.FLAT_ADDRESS;}]
+ [{return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::FLAT_ADDRESS;}]
>;
def atomic_cmp_swap_flat : flat_binary_atomic_op<AMDGPUatomic_cmp_swap>;
}
case Intrinsic::r600_implicitarg_ptr: {
- MVT PtrVT = getPointerTy(DAG.getDataLayout(), AMDGPUASI.PARAM_I_ADDRESS);
+ MVT PtrVT = getPointerTy(DAG.getDataLayout(), AMDGPUAS::PARAM_I_ADDRESS);
uint32_t ByteOffset = getImplicitParameterOffset(MF, FIRST_IMPLICIT);
return DAG.getConstant(ByteOffset, DL, PtrVT);
}
SDValue Op,
SelectionDAG &DAG) const {
GlobalAddressSDNode *GSD = cast<GlobalAddressSDNode>(Op);
- if (GSD->getAddressSpace() != AMDGPUASI.CONSTANT_ADDRESS)
+ if (GSD->getAddressSpace() != AMDGPUAS::CONSTANT_ADDRESS)
return AMDGPUTargetLowering::LowerGlobalAddress(MFI, Op, DAG);
const DataLayout &DL = DAG.getDataLayout();
const GlobalValue *GV = GSD->getGlobal();
- MVT ConstPtrVT = getPointerTy(DL, AMDGPUASI.CONSTANT_ADDRESS);
+ MVT ConstPtrVT = getPointerTy(DL, AMDGPUAS::CONSTANT_ADDRESS);
SDValue GA = DAG.getTargetGlobalAddress(GV, SDLoc(GSD), ConstPtrVT);
return DAG.getNode(AMDGPUISD::CONST_DATA_PTR, SDLoc(GSD), ConstPtrVT, GA);
unsigned DwordOffset) const {
unsigned ByteOffset = DwordOffset * 4;
PointerType * PtrType = PointerType::get(VT.getTypeForEVT(*DAG.getContext()),
- AMDGPUASI.PARAM_I_ADDRESS);
+ AMDGPUAS::PARAM_I_ADDRESS);
// We shouldn't be using an offset wider than 16-bits for implicit parameters.
assert(isInt<16>(ByteOffset));
//TODO: Who creates the i8 stores?
assert(Store->isTruncatingStore()
|| Store->getValue().getValueType() == MVT::i8);
- assert(Store->getAddressSpace() == AMDGPUASI.PRIVATE_ADDRESS);
+ assert(Store->getAddressSpace() == AMDGPUAS::PRIVATE_ADDRESS);
SDValue Mask;
if (Store->getMemoryVT() == MVT::i8) {
// Load dword
// TODO: can we be smarter about machine pointer info?
MachinePointerInfo PtrInfo(UndefValue::get(
- Type::getInt32PtrTy(*DAG.getContext(), AMDGPUASI.PRIVATE_ADDRESS)));
+ Type::getInt32PtrTy(*DAG.getContext(), AMDGPUAS::PRIVATE_ADDRESS)));
SDValue Dst = DAG.getLoad(MVT::i32, DL, Chain, Ptr, PtrInfo);
Chain = Dst.getValue(1);
SDLoc DL(Op);
// Neither LOCAL nor PRIVATE can do vectors at the moment
- if ((AS == AMDGPUASI.LOCAL_ADDRESS || AS == AMDGPUASI.PRIVATE_ADDRESS) &&
+ if ((AS == AMDGPUAS::LOCAL_ADDRESS || AS == AMDGPUAS::PRIVATE_ADDRESS) &&
VT.isVector()) {
- if ((AS == AMDGPUASI.PRIVATE_ADDRESS) &&
+ if ((AS == AMDGPUAS::PRIVATE_ADDRESS) &&
StoreNode->isTruncatingStore()) {
// Add an extra level of chain to isolate this vector
SDValue NewChain = DAG.getNode(AMDGPUISD::DUMMY_CHAIN, DL, MVT::Other, Chain);
SDValue DWordAddr = DAG.getNode(ISD::SRL, DL, PtrVT, Ptr,
DAG.getConstant(2, DL, PtrVT));
- if (AS == AMDGPUASI.GLOBAL_ADDRESS) {
+ if (AS == AMDGPUAS::GLOBAL_ADDRESS) {
// It is beneficial to create MSKOR here instead of combiner to avoid
// artificial dependencies introduced by RMW
if (StoreNode->isTruncatingStore()) {
}
// GLOBAL_ADDRESS has been handled above, LOCAL_ADDRESS allows all sizes
- if (AS != AMDGPUASI.PRIVATE_ADDRESS)
+ if (AS != AMDGPUAS::PRIVATE_ADDRESS)
return SDValue();
if (MemVT.bitsLT(MVT::i32))
// Load dword
// TODO: can we be smarter about machine pointer info?
MachinePointerInfo PtrInfo(UndefValue::get(
- Type::getInt32PtrTy(*DAG.getContext(), AMDGPUASI.PRIVATE_ADDRESS)));
+ Type::getInt32PtrTy(*DAG.getContext(), AMDGPUAS::PRIVATE_ADDRESS)));
SDValue Read = DAG.getLoad(MVT::i32, DL, Chain, Ptr, PtrInfo);
// Get offset within the register.
EVT MemVT = LoadNode->getMemoryVT();
ISD::LoadExtType ExtType = LoadNode->getExtensionType();
- if (AS == AMDGPUASI.PRIVATE_ADDRESS &&
+ if (AS == AMDGPUAS::PRIVATE_ADDRESS &&
ExtType != ISD::NON_EXTLOAD && MemVT.bitsLT(MVT::i32)) {
return lowerPrivateExtLoad(Op, DAG);
}
SDValue Chain = LoadNode->getChain();
SDValue Ptr = LoadNode->getBasePtr();
- if ((LoadNode->getAddressSpace() == AMDGPUASI.LOCAL_ADDRESS ||
- LoadNode->getAddressSpace() == AMDGPUASI.PRIVATE_ADDRESS) &&
+ if ((LoadNode->getAddressSpace() == AMDGPUAS::LOCAL_ADDRESS ||
+ LoadNode->getAddressSpace() == AMDGPUAS::PRIVATE_ADDRESS) &&
VT.isVector()) {
return scalarizeVectorLoad(LoadNode, DAG);
}
DAG.getNode(ISD::SRL, DL, MVT::i32, Ptr,
DAG.getConstant(4, DL, MVT::i32)),
DAG.getConstant(LoadNode->getAddressSpace() -
- AMDGPUASI.CONSTANT_BUFFER_0, DL, MVT::i32)
+ AMDGPUAS::CONSTANT_BUFFER_0, DL, MVT::i32)
);
}
return DAG.getMergeValues(MergedValues, DL);
}
- if (LoadNode->getAddressSpace() != AMDGPUASI.PRIVATE_ADDRESS) {
+ if (LoadNode->getAddressSpace() != AMDGPUAS::PRIVATE_ADDRESS) {
return SDValue();
}
}
PointerType *PtrTy = PointerType::get(VT.getTypeForEVT(*DAG.getContext()),
- AMDGPUASI.PARAM_I_ADDRESS);
+ AMDGPUAS::PARAM_I_ADDRESS);
// i64 isn't a legal type, so the register type used ends up as i32, which
// isn't expected here. It attempts to create this sextload, but it ends up
bool R600TargetLowering::canMergeStoresTo(unsigned AS, EVT MemVT,
const SelectionDAG &DAG) const {
// Local and Private addresses do not handle vectors. Limit to i32
- if ((AS == AMDGPUASI.LOCAL_ADDRESS || AS == AMDGPUASI.PRIVATE_ADDRESS)) {
+ if ((AS == AMDGPUAS::LOCAL_ADDRESS || AS == AMDGPUAS::PRIVATE_ADDRESS)) {
return (MemVT.getSizeInBits() <= 32);
}
return true;
switch (Kind) {
case PseudoSourceValue::Stack:
case PseudoSourceValue::FixedStack:
- return ST.getAMDGPUAS().PRIVATE_ADDRESS;
+ return AMDGPUAS::PRIVATE_ADDRESS;
case PseudoSourceValue::ConstantPool:
case PseudoSourceValue::GOT:
case PseudoSourceValue::JumpTable:
case PseudoSourceValue::GlobalValueCallEntry:
case PseudoSourceValue::ExternalSymbolCallEntry:
case PseudoSourceValue::TargetCustom:
- return ST.getAMDGPUAS().CONSTANT_ADDRESS;
+ return AMDGPUAS::CONSTANT_ADDRESS;
}
+
llvm_unreachable("Invalid pseudo source kind");
- return ST.getAMDGPUAS().PRIVATE_ADDRESS;
}
class LoadParamFrag <PatFrag load_type> : PatFrag <
(ops node:$ptr), (load_type node:$ptr),
[{ return isConstantLoad(cast<LoadSDNode>(N), 0) ||
- (cast<LoadSDNode>(N)->getAddressSpace() == AMDGPUASI.PARAM_I_ADDRESS); }]
+ (cast<LoadSDNode>(N)->getAddressSpace() == AMDGPUAS::PARAM_I_ADDRESS); }]
>;
def vtx_id3_az_extloadi8 : LoadParamFrag<az_extloadi8>;
class LoadVtxId1 <PatFrag load> : PatFrag <
(ops node:$ptr), (load node:$ptr), [{
const MemSDNode *LD = cast<MemSDNode>(N);
- return LD->getAddressSpace() == AMDGPUASI.GLOBAL_ADDRESS ||
- (LD->getAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS &&
+ return LD->getAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS ||
+ (LD->getAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS &&
!isa<GlobalValue>(GetUnderlyingObject(
LD->getMemOperand()->getValue(), CurDAG->getDataLayout())));
}]>;
class LoadVtxId2 <PatFrag load> : PatFrag <
(ops node:$ptr), (load node:$ptr), [{
const MemSDNode *LD = cast<MemSDNode>(N);
- return LD->getAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS &&
+ return LD->getAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS &&
isa<GlobalValue>(GetUnderlyingObject(
LD->getMemOperand()->getValue(), CurDAG->getDataLayout()));
}]>;
if (AM.BaseGV)
return false;
- if (AS == AMDGPUASI.GLOBAL_ADDRESS)
+ if (AS == AMDGPUAS::GLOBAL_ADDRESS)
return isLegalGlobalAddressingMode(AM);
- if (AS == AMDGPUASI.CONSTANT_ADDRESS ||
- AS == AMDGPUASI.CONSTANT_ADDRESS_32BIT) {
+ if (AS == AMDGPUAS::CONSTANT_ADDRESS ||
+ AS == AMDGPUAS::CONSTANT_ADDRESS_32BIT) {
// If the offset isn't a multiple of 4, it probably isn't going to be
// correctly aligned.
// FIXME: Can we get the real alignment here?
return false;
- } else if (AS == AMDGPUASI.PRIVATE_ADDRESS) {
+ } else if (AS == AMDGPUAS::PRIVATE_ADDRESS) {
return isLegalMUBUFAddressingMode(AM);
- } else if (AS == AMDGPUASI.LOCAL_ADDRESS ||
- AS == AMDGPUASI.REGION_ADDRESS) {
+ } else if (AS == AMDGPUAS::LOCAL_ADDRESS ||
+ AS == AMDGPUAS::REGION_ADDRESS) {
// Basic, single offset DS instructions allow a 16-bit unsigned immediate
// field.
// XXX - If doing a 4-byte aligned 8-byte type access, we effectively have
return true;
return false;
- } else if (AS == AMDGPUASI.FLAT_ADDRESS ||
- AS == AMDGPUASI.UNKNOWN_ADDRESS_SPACE) {
+ } else if (AS == AMDGPUAS::FLAT_ADDRESS ||
+ AS == AMDGPUAS::UNKNOWN_ADDRESS_SPACE) {
// For an unknown address space, this usually means that this is for some
// reason being used for pure arithmetic, and not based on some addressing
// computation. We don't have instructions that compute pointers with any
bool SITargetLowering::canMergeStoresTo(unsigned AS, EVT MemVT,
const SelectionDAG &DAG) const {
- if (AS == AMDGPUASI.GLOBAL_ADDRESS || AS == AMDGPUASI.FLAT_ADDRESS) {
+ if (AS == AMDGPUAS::GLOBAL_ADDRESS || AS == AMDGPUAS::FLAT_ADDRESS) {
return (MemVT.getSizeInBits() <= 4 * 32);
- } else if (AS == AMDGPUASI.PRIVATE_ADDRESS) {
+ } else if (AS == AMDGPUAS::PRIVATE_ADDRESS) {
unsigned MaxPrivateBits = 8 * getSubtarget()->getMaxPrivateElementSize();
return (MemVT.getSizeInBits() <= MaxPrivateBits);
- } else if (AS == AMDGPUASI.LOCAL_ADDRESS) {
+ } else if (AS == AMDGPUAS::LOCAL_ADDRESS) {
return (MemVT.getSizeInBits() <= 2 * 32);
}
return true;
return false;
}
- if (AddrSpace == AMDGPUASI.LOCAL_ADDRESS ||
- AddrSpace == AMDGPUASI.REGION_ADDRESS) {
+ if (AddrSpace == AMDGPUAS::LOCAL_ADDRESS ||
+ AddrSpace == AMDGPUAS::REGION_ADDRESS) {
// ds_read/write_b64 require 8-byte alignment, but we can do a 4 byte
// aligned, 8 byte access in a single operation using ds_read2/write2_b32
// with adjacent offsets.
// will access scratch. If we had access to the IR function, then we
// could determine if any private memory was used in the function.
if (!Subtarget->hasUnalignedScratchAccess() &&
- (AddrSpace == AMDGPUASI.PRIVATE_ADDRESS ||
- AddrSpace == AMDGPUASI.FLAT_ADDRESS)) {
+ (AddrSpace == AMDGPUAS::PRIVATE_ADDRESS ||
+ AddrSpace == AMDGPUAS::FLAT_ADDRESS)) {
return false;
}
// If we have an uniform constant load, it still requires using a slow
// buffer instruction if unaligned.
if (IsFast) {
- *IsFast = (AddrSpace == AMDGPUASI.CONSTANT_ADDRESS ||
- AddrSpace == AMDGPUASI.CONSTANT_ADDRESS_32BIT) ?
+ *IsFast = (AddrSpace == AMDGPUAS::CONSTANT_ADDRESS ||
+ AddrSpace == AMDGPUAS::CONSTANT_ADDRESS_32BIT) ?
(Align % 4 == 0) : true;
}
return MVT::Other;
}
-static bool isFlatGlobalAddrSpace(unsigned AS, AMDGPUAS AMDGPUASI) {
- return AS == AMDGPUASI.GLOBAL_ADDRESS ||
- AS == AMDGPUASI.FLAT_ADDRESS ||
- AS == AMDGPUASI.CONSTANT_ADDRESS ||
- AS == AMDGPUASI.CONSTANT_ADDRESS_32BIT;
+static bool isFlatGlobalAddrSpace(unsigned AS) {
+ return AS == AMDGPUAS::GLOBAL_ADDRESS ||
+ AS == AMDGPUAS::FLAT_ADDRESS ||
+ AS == AMDGPUAS::CONSTANT_ADDRESS ||
+ AS == AMDGPUAS::CONSTANT_ADDRESS_32BIT;
}
bool SITargetLowering::isNoopAddrSpaceCast(unsigned SrcAS,
unsigned DestAS) const {
- return isFlatGlobalAddrSpace(SrcAS, AMDGPUASI) &&
- isFlatGlobalAddrSpace(DestAS, AMDGPUASI);
+ return isFlatGlobalAddrSpace(SrcAS) && isFlatGlobalAddrSpace(DestAS);
}
bool SITargetLowering::isMemOpHasNoClobberedMemOperand(const SDNode *N) const {
unsigned DestAS) const {
// Flat -> private/local is a simple truncate.
// Flat -> global is no-op
- if (SrcAS == AMDGPUASI.FLAT_ADDRESS)
+ if (SrcAS == AMDGPUAS::FLAT_ADDRESS)
return true;
return isNoopAddrSpaceCast(SrcAS, DestAS);
= Info->getPreloadedValue(AMDGPUFunctionArgInfo::KERNARG_SEGMENT_PTR);
MachineRegisterInfo &MRI = DAG.getMachineFunction().getRegInfo();
- MVT PtrVT = getPointerTy(DL, AMDGPUASI.CONSTANT_ADDRESS);
+ MVT PtrVT = getPointerTy(DL, AMDGPUAS::CONSTANT_ADDRESS);
SDValue BasePtr = DAG.getCopyFromReg(Chain, SL,
MRI.getLiveInVirtReg(InputPtrReg->getRegister()), PtrVT);
uint64_t Offset, unsigned Align, bool Signed,
const ISD::InputArg *Arg) const {
Type *Ty = MemVT.getTypeForEVT(*DAG.getContext());
- PointerType *PtrTy = PointerType::get(Ty, AMDGPUASI.CONSTANT_ADDRESS);
+ PointerType *PtrTy = PointerType::get(Ty, AMDGPUAS::CONSTANT_ADDRESS);
MachinePointerInfo PtrInfo(UndefValue::get(PtrTy));
// Try to avoid using an extload by loading earlier than the argument address,
/*isVol = */ false, /*AlwaysInline = */ true,
/*isTailCall = */ false, DstInfo,
MachinePointerInfo(UndefValue::get(Type::getInt8PtrTy(
- *DAG.getContext(), AMDGPUASI.PRIVATE_ADDRESS))));
+ *DAG.getContext(), AMDGPUAS::PRIVATE_ADDRESS))));
MemOpChains.push_back(Cpy);
} else {
bool SITargetLowering::shouldEmitFixup(const GlobalValue *GV) const {
const Triple &TT = getTargetMachine().getTargetTriple();
- return (GV->getType()->getAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS ||
- GV->getType()->getAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS_32BIT) &&
+ return (GV->getType()->getAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS ||
+ GV->getType()->getAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS_32BIT) &&
AMDGPU::shouldEmitConstantsToTextSection(TT);
}
bool SITargetLowering::shouldEmitGOTReloc(const GlobalValue *GV) const {
- return (GV->getType()->getAddressSpace() == AMDGPUASI.GLOBAL_ADDRESS ||
- GV->getType()->getAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS ||
- GV->getType()->getAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS_32BIT) &&
+ return (GV->getType()->getAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS ||
+ GV->getType()->getAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS ||
+ GV->getType()->getAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS_32BIT) &&
!shouldEmitFixup(GV) &&
!getTargetMachine().shouldAssumeDSOLocal(*GV->getParent(), GV);
}
SelectionDAG &DAG) const {
// FIXME: Use inline constants (src_{shared, private}_base) instead.
if (Subtarget->hasApertureRegs()) {
- unsigned Offset = AS == AMDGPUASI.LOCAL_ADDRESS ?
+ unsigned Offset = AS == AMDGPUAS::LOCAL_ADDRESS ?
AMDGPU::Hwreg::OFFSET_SRC_SHARED_BASE :
AMDGPU::Hwreg::OFFSET_SRC_PRIVATE_BASE;
- unsigned WidthM1 = AS == AMDGPUASI.LOCAL_ADDRESS ?
+ unsigned WidthM1 = AS == AMDGPUAS::LOCAL_ADDRESS ?
AMDGPU::Hwreg::WIDTH_M1_SRC_SHARED_BASE :
AMDGPU::Hwreg::WIDTH_M1_SRC_PRIVATE_BASE;
unsigned Encoding =
// Offset into amd_queue_t for group_segment_aperture_base_hi /
// private_segment_aperture_base_hi.
- uint32_t StructOffset = (AS == AMDGPUASI.LOCAL_ADDRESS) ? 0x40 : 0x44;
+ uint32_t StructOffset = (AS == AMDGPUAS::LOCAL_ADDRESS) ? 0x40 : 0x44;
SDValue Ptr = DAG.getObjectPtrOffset(DL, QueuePtr, StructOffset);
// TODO: We should use the value from the IR intrinsic call, but it might not
// be available and how do we get it?
Value *V = UndefValue::get(PointerType::get(Type::getInt8Ty(*DAG.getContext()),
- AMDGPUASI.CONSTANT_ADDRESS));
+ AMDGPUAS::CONSTANT_ADDRESS));
MachinePointerInfo PtrInfo(V, StructOffset);
return DAG.getLoad(MVT::i32, DL, QueuePtr.getValue(1), Ptr, PtrInfo,
static_cast<const AMDGPUTargetMachine &>(getTargetMachine());
// flat -> local/private
- if (ASC->getSrcAddressSpace() == AMDGPUASI.FLAT_ADDRESS) {
+ if (ASC->getSrcAddressSpace() == AMDGPUAS::FLAT_ADDRESS) {
unsigned DestAS = ASC->getDestAddressSpace();
- if (DestAS == AMDGPUASI.LOCAL_ADDRESS ||
- DestAS == AMDGPUASI.PRIVATE_ADDRESS) {
+ if (DestAS == AMDGPUAS::LOCAL_ADDRESS ||
+ DestAS == AMDGPUAS::PRIVATE_ADDRESS) {
unsigned NullVal = TM.getNullPointerValue(DestAS);
SDValue SegmentNullPtr = DAG.getConstant(NullVal, SL, MVT::i32);
SDValue NonNull = DAG.getSetCC(SL, MVT::i1, Src, FlatNullPtr, ISD::SETNE);
}
// local/private -> flat
- if (ASC->getDestAddressSpace() == AMDGPUASI.FLAT_ADDRESS) {
+ if (ASC->getDestAddressSpace() == AMDGPUAS::FLAT_ADDRESS) {
unsigned SrcAS = ASC->getSrcAddressSpace();
- if (SrcAS == AMDGPUASI.LOCAL_ADDRESS ||
- SrcAS == AMDGPUASI.PRIVATE_ADDRESS) {
+ if (SrcAS == AMDGPUAS::LOCAL_ADDRESS ||
+ SrcAS == AMDGPUAS::PRIVATE_ADDRESS) {
unsigned NullVal = TM.getNullPointerValue(SrcAS);
SDValue SegmentNullPtr = DAG.getConstant(NullVal, SL, MVT::i32);
bool
SITargetLowering::isOffsetFoldingLegal(const GlobalAddressSDNode *GA) const {
// We can fold offsets for anything that doesn't require a GOT relocation.
- return (GA->getAddressSpace() == AMDGPUASI.GLOBAL_ADDRESS ||
- GA->getAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS ||
- GA->getAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS_32BIT) &&
+ return (GA->getAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS ||
+ GA->getAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS ||
+ GA->getAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS_32BIT) &&
!shouldEmitGOTReloc(GA->getGlobal());
}
GlobalAddressSDNode *GSD = cast<GlobalAddressSDNode>(Op);
const GlobalValue *GV = GSD->getGlobal();
- if (GSD->getAddressSpace() != AMDGPUASI.CONSTANT_ADDRESS &&
- GSD->getAddressSpace() != AMDGPUASI.CONSTANT_ADDRESS_32BIT &&
- GSD->getAddressSpace() != AMDGPUASI.GLOBAL_ADDRESS &&
+ if (GSD->getAddressSpace() != AMDGPUAS::CONSTANT_ADDRESS &&
+ GSD->getAddressSpace() != AMDGPUAS::CONSTANT_ADDRESS_32BIT &&
+ GSD->getAddressSpace() != AMDGPUAS::GLOBAL_ADDRESS &&
// FIXME: It isn't correct to rely on the type of the pointer. This should
// be removed when address space 0 is 64-bit.
!GV->getType()->getElementType()->isFunctionTy())
SIInstrInfo::MO_GOTPCREL32);
Type *Ty = PtrVT.getTypeForEVT(*DAG.getContext());
- PointerType *PtrTy = PointerType::get(Ty, AMDGPUASI.CONSTANT_ADDRESS);
+ PointerType *PtrTy = PointerType::get(Ty, AMDGPUAS::CONSTANT_ADDRESS);
const DataLayout &DataLayout = DAG.getDataLayout();
unsigned Align = DataLayout.getABITypeAlignment(PtrTy);
// FIXME: Use a PseudoSourceValue once those can be assigned an address space.
// Dz/dh, dz/dv and the last odd coord are packed with undef. Also,
// in 1D, derivatives dx/dh and dx/dv are packed with undef.
if (((i + 1) >= (AddrIdx + NumMIVAddrs)) ||
- ((NumGradients / 2) % 2 == 1 &&
- (i == DimIdx + (NumGradients / 2) - 1 ||
+ ((NumGradients / 2) % 2 == 1 &&
+ (i == DimIdx + (NumGradients / 2) - 1 ||
i == DimIdx + NumGradients - 1))) {
AddrHi = DAG.getUNDEF(MVT::f16);
} else {
// FIXME: Constant loads should all be marked invariant.
unsigned AS = Ld->getAddressSpace();
- if (AS != AMDGPUASI.CONSTANT_ADDRESS &&
- AS != AMDGPUASI.CONSTANT_ADDRESS_32BIT &&
+ if (AS != AMDGPUAS::CONSTANT_ADDRESS &&
+ AS != AMDGPUAS::CONSTANT_ADDRESS_32BIT &&
(AS != AMDGPUAS::GLOBAL_ADDRESS || !Ld->isInvariant()))
return SDValue();
SIMachineFunctionInfo *MFI = MF.getInfo<SIMachineFunctionInfo>();
// If there is a possibilty that flat instruction access scratch memory
// then we need to use the same legalization rules we use for private.
- if (AS == AMDGPUASI.FLAT_ADDRESS)
+ if (AS == AMDGPUAS::FLAT_ADDRESS)
AS = MFI->hasFlatScratchInit() ?
- AMDGPUASI.PRIVATE_ADDRESS : AMDGPUASI.GLOBAL_ADDRESS;
+ AMDGPUAS::PRIVATE_ADDRESS : AMDGPUAS::GLOBAL_ADDRESS;
unsigned NumElements = MemVT.getVectorNumElements();
- if (AS == AMDGPUASI.CONSTANT_ADDRESS ||
- AS == AMDGPUASI.CONSTANT_ADDRESS_32BIT) {
+ if (AS == AMDGPUAS::CONSTANT_ADDRESS ||
+ AS == AMDGPUAS::CONSTANT_ADDRESS_32BIT) {
if (!Op->isDivergent() && Alignment >= 4)
return SDValue();
// Non-uniform loads will be selected to MUBUF instructions, so they
//
}
- if (AS == AMDGPUASI.CONSTANT_ADDRESS ||
- AS == AMDGPUASI.CONSTANT_ADDRESS_32BIT ||
- AS == AMDGPUASI.GLOBAL_ADDRESS) {
+ if (AS == AMDGPUAS::CONSTANT_ADDRESS ||
+ AS == AMDGPUAS::CONSTANT_ADDRESS_32BIT ||
+ AS == AMDGPUAS::GLOBAL_ADDRESS) {
if (Subtarget->getScalarizeGlobalBehavior() && !Op->isDivergent() &&
!Load->isVolatile() && isMemOpHasNoClobberedMemOperand(Load) &&
Alignment >= 4)
// loads.
//
}
- if (AS == AMDGPUASI.CONSTANT_ADDRESS ||
- AS == AMDGPUASI.CONSTANT_ADDRESS_32BIT ||
- AS == AMDGPUASI.GLOBAL_ADDRESS ||
- AS == AMDGPUASI.FLAT_ADDRESS) {
+ if (AS == AMDGPUAS::CONSTANT_ADDRESS ||
+ AS == AMDGPUAS::CONSTANT_ADDRESS_32BIT ||
+ AS == AMDGPUAS::GLOBAL_ADDRESS ||
+ AS == AMDGPUAS::FLAT_ADDRESS) {
if (NumElements > 4)
return SplitVectorLoad(Op, DAG);
// v4 loads are supported for private and global memory.
return SDValue();
}
- if (AS == AMDGPUASI.PRIVATE_ADDRESS) {
+ if (AS == AMDGPUAS::PRIVATE_ADDRESS) {
// Depending on the setting of the private_element_size field in the
// resource descriptor, we can only make private accesses up to a certain
// size.
default:
llvm_unreachable("unsupported private_element_size");
}
- } else if (AS == AMDGPUASI.LOCAL_ADDRESS) {
+ } else if (AS == AMDGPUAS::LOCAL_ADDRESS) {
// Use ds_read_b128 if possible.
if (Subtarget->useDS128() && Load->getAlignment() >= 16 &&
MemVT.getStoreSize() == 16)
SIMachineFunctionInfo *MFI = MF.getInfo<SIMachineFunctionInfo>();
// If there is a possibilty that flat instruction access scratch memory
// then we need to use the same legalization rules we use for private.
- if (AS == AMDGPUASI.FLAT_ADDRESS)
+ if (AS == AMDGPUAS::FLAT_ADDRESS)
AS = MFI->hasFlatScratchInit() ?
- AMDGPUASI.PRIVATE_ADDRESS : AMDGPUASI.GLOBAL_ADDRESS;
+ AMDGPUAS::PRIVATE_ADDRESS : AMDGPUAS::GLOBAL_ADDRESS;
unsigned NumElements = VT.getVectorNumElements();
- if (AS == AMDGPUASI.GLOBAL_ADDRESS ||
- AS == AMDGPUASI.FLAT_ADDRESS) {
+ if (AS == AMDGPUAS::GLOBAL_ADDRESS ||
+ AS == AMDGPUAS::FLAT_ADDRESS) {
if (NumElements > 4)
return SplitVectorStore(Op, DAG);
return SDValue();
- } else if (AS == AMDGPUASI.PRIVATE_ADDRESS) {
+ } else if (AS == AMDGPUAS::PRIVATE_ADDRESS) {
switch (Subtarget->getMaxPrivateElementSize()) {
case 4:
return scalarizeVectorStore(Store, DAG);
default:
llvm_unreachable("unsupported private_element_size");
}
- } else if (AS == AMDGPUASI.LOCAL_ADDRESS) {
+ } else if (AS == AMDGPUAS::LOCAL_ADDRESS) {
// Use ds_write_b128 if possible.
if (Subtarget->useDS128() && Store->getAlignment() >= 16 &&
VT.getStoreSize() == 16)
unsigned AS = AtomicNode->getAddressSpace();
// No custom lowering required for local address space
- if (!isFlatGlobalAddrSpace(AS, AMDGPUASI))
+ if (!isFlatGlobalAddrSpace(AS))
return Op;
// Non-local address space requires custom lowering for atomic compare
break;
case ISD::LOAD: {
const LoadSDNode *L = dyn_cast<LoadSDNode>(N);
- if (L->getMemOperand()->getAddrSpace() ==
- Subtarget->getAMDGPUAS().PRIVATE_ADDRESS)
+ // FIXME: Also needs to handle flat.
+ if (L->getMemOperand()->getAddrSpace() == AMDGPUAS::PRIVATE_ADDRESS)
return true;
} break;
case ISD::CALLSEQ_END:
const MachineRegisterInfo *MRI = nullptr;
const MachineLoopInfo *MLI = nullptr;
AMDGPU::IsaInfo::IsaVersion IV;
- AMDGPUAS AMDGPUASI;
DenseSet<MachineBasicBlock *> BlockVisitedSet;
DenseSet<MachineInstr *> TrackedWaitcntSet;
// instruction.
for (const MachineMemOperand *Memop : MI.memoperands()) {
unsigned AS = Memop->getAddrSpace();
- if (AS != AMDGPUASI.LOCAL_ADDRESS)
+ if (AS != AMDGPUAS::LOCAL_ADDRESS)
continue;
unsigned RegNo = SQ_MAX_PGM_VGPRS + EXTRA_VGPR_LDS;
// VM_CNT is only relevant to vgpr or LDS.
// FIXME: Should not be relying on memoperands.
for (const MachineMemOperand *Memop : MI.memoperands()) {
unsigned AS = Memop->getAddrSpace();
- if (AS != AMDGPUASI.LOCAL_ADDRESS)
+ if (AS != AMDGPUAS::LOCAL_ADDRESS)
continue;
unsigned RegNo = SQ_MAX_PGM_VGPRS + EXTRA_VGPR_LDS;
EmitWaitcnt |= ScoreBrackets->updateByWait(
for (const MachineMemOperand *Memop : MI.memoperands()) {
unsigned AS = Memop->getAddrSpace();
- if (AS == AMDGPUASI.LOCAL_ADDRESS || AS == AMDGPUASI.FLAT_ADDRESS)
+ if (AS == AMDGPUAS::LOCAL_ADDRESS || AS == AMDGPUAS::FLAT_ADDRESS)
return true;
}
MLI = &getAnalysis<MachineLoopInfo>();
IV = AMDGPU::IsaInfo::getIsaVersion(ST->getFeatureBits());
const SIMachineFunctionInfo *MFI = MF.getInfo<SIMachineFunctionInfo>();
- AMDGPUASI = ST->getAMDGPUAS();
ForceEmitZeroWaitcnts = ForceEmitZeroFlag;
for (enum InstCounterType T = VM_CNT; T < NUM_INST_CNTS;
switch(Kind) {
case PseudoSourceValue::Stack:
case PseudoSourceValue::FixedStack:
- return ST.getAMDGPUAS().PRIVATE_ADDRESS;
+ return AMDGPUAS::PRIVATE_ADDRESS;
case PseudoSourceValue::ConstantPool:
case PseudoSourceValue::GOT:
case PseudoSourceValue::JumpTable:
case PseudoSourceValue::GlobalValueCallEntry:
case PseudoSourceValue::ExternalSymbolCallEntry:
case PseudoSourceValue::TargetCustom:
- return ST.getAMDGPUAS().CONSTANT_ADDRESS;
+ return AMDGPUAS::CONSTANT_ADDRESS;
}
- return ST.getAMDGPUAS().FLAT_ADDRESS;
+ return AMDGPUAS::FLAT_ADDRESS;
}
static void removeModOperands(MachineInstr &MI) {
unsigned Count = 0;
const TargetRegisterClass *DestRC = MRI.getRegClass(Dest.getReg());
const TargetRegisterClass *NewDestRC = RI.getEquivalentVGPRClass(DestRC);
-
+
switch(Opcode) {
default:
return;
// FIXME: Should also attempt to build VAddr and Offset like the non-split
// case (see call site for this function)
-
+
// Create a vector of result registers
SmallVector<unsigned, 8> ResultRegs;
for (unsigned i = 0; i < Count ; ++i) {
return AMDGPU::NoRegister;
assert(!MI.memoperands_empty() &&
- (*MI.memoperands_begin())->getAddrSpace() == ST.getAMDGPUAS().PRIVATE_ADDRESS);
+ (*MI.memoperands_begin())->getAddrSpace() == AMDGPUAS::PRIVATE_ADDRESS);
FrameIndex = Addr->getIndex();
return getNamedOperand(MI, AMDGPU::OpName::vdata)->getReg();
return true;
for (const MachineMemOperand *MMO : MI.memoperands()) {
- if (MMO->getAddrSpace() == ST.getAMDGPUAS().FLAT_ADDRESS)
+ if (MMO->getAddrSpace() == AMDGPUAS::FLAT_ADDRESS)
return true;
}
return false;
class SIMemOpAccess final {
private:
-
- AMDGPUAS SIAddrSpaceInfo;
AMDGPUMachineModuleInfo *MMI = nullptr;
/// Reports unsupported message \p Msg for \p MI to LLVM context.
}
SIAtomicAddrSpace SIMemOpAccess::toSIAtomicAddrSpace(unsigned AS) const {
- if (AS == SIAddrSpaceInfo.FLAT_ADDRESS)
+ if (AS == AMDGPUAS::FLAT_ADDRESS)
return SIAtomicAddrSpace::FLAT;
- if (AS == SIAddrSpaceInfo.GLOBAL_ADDRESS)
+ if (AS == AMDGPUAS::GLOBAL_ADDRESS)
return SIAtomicAddrSpace::GLOBAL;
- if (AS == SIAddrSpaceInfo.LOCAL_ADDRESS)
+ if (AS == AMDGPUAS::LOCAL_ADDRESS)
return SIAtomicAddrSpace::LDS;
- if (AS == SIAddrSpaceInfo.PRIVATE_ADDRESS)
+ if (AS == AMDGPUAS::PRIVATE_ADDRESS)
return SIAtomicAddrSpace::SCRATCH;
- if (AS == SIAddrSpaceInfo.REGION_ADDRESS)
+ if (AS == AMDGPUAS::REGION_ADDRESS)
return SIAtomicAddrSpace::GDS;
return SIAtomicAddrSpace::OTHER;
}
SIMemOpAccess::SIMemOpAccess(MachineFunction &MF) {
- SIAddrSpaceInfo = getAMDGPUAS(MF.getTarget());
MMI = &MF.getMMI().getObjFileInfo<AMDGPUMachineModuleInfo>();
}
def smrd_load : PatFrag <(ops node:$ptr), (load node:$ptr), [{
auto Ld = cast<LoadSDNode>(N);
return Ld->getAlignment() >= 4 &&
- ((((Ld->getAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS) || (Ld->getAddressSpace() == AMDGPUASI.CONSTANT_ADDRESS_32BIT)) && !N->isDivergent()) ||
- (Subtarget->getScalarizeGlobalBehavior() && Ld->getAddressSpace() == AMDGPUASI.GLOBAL_ADDRESS &&
+ ((((Ld->getAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS) || (Ld->getAddressSpace() == AMDGPUAS::CONSTANT_ADDRESS_32BIT)) && !N->isDivergent()) ||
+ (Subtarget->getScalarizeGlobalBehavior() && Ld->getAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS &&
!Ld->isVolatile() && !N->isDivergent() &&
static_cast<const SITargetLowering *>(getTargetLowering())->isMemOpHasNoClobberedMemOperand(N)));
}]>;
return true;
}
-} // end namespace AMDGPU
-
-} // end namespace llvm
-
-namespace llvm {
-namespace AMDGPU {
-
-AMDGPUAS getAMDGPUAS(Triple T) {
- AMDGPUAS AS;
- AS.FLAT_ADDRESS = 0;
- AS.PRIVATE_ADDRESS = 5;
- AS.REGION_ADDRESS = 2;
- return AS;
-}
-
-AMDGPUAS getAMDGPUAS(const TargetMachine &M) {
- return getAMDGPUAS(M.getTargetTriple());
-}
-
-AMDGPUAS getAMDGPUAS(const Module &M) {
- return getAMDGPUAS(Triple(M.getTargetTriple()));
-}
-
namespace {
struct SourceOfDivergence {
; RUN: llvm-as -data-layout=A5 < %s | llvm-dis | FileCheck %s
-; RUN: llc -mtriple amdgcn-amd-amdhsa-amdgiz < %s
-; RUN: llvm-as -data-layout=A5 < %s | llc -mtriple amdgcn-amd-amdhsa-amdgiz
+; RUN: llc -mtriple amdgcn-amd-amdhsa < %s
+; RUN: llvm-as -data-layout=A5 < %s | llc -mtriple amdgcn-amd-amdhsa
; RUN: opt -data-layout=A5 -S < %s
; RUN: llvm-as -data-layout=A5 < %s | opt -S
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgiz -mcpu=fiji -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgiz -mcpu=hawaii -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgiz -mcpu=gfx900 -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=fiji -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=hawaii -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
declare void @external_void_func_void() #0
-; RUN: llc -march=r600 -mtriple=r600---amdgiz -mcpu=cypress < %s | FileCheck -check-prefix=EG %s
+; RUN: llc -march=r600 -mtriple=r600-- -mcpu=cypress < %s | FileCheck -check-prefix=EG %s
;
; kernel void combine_vloads(global char8 addrspace(5)* src, global char8 addrspace(5)* result) {
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgizcl -verify-machineinstrs < %s | FileCheck %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -verify-machineinstrs < %s | FileCheck %s
%struct.wombat = type { [4 x i32], [4 x i32], [4 x i32] }
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgizcl -verify-machineinstrs < %s | FileCheck %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -verify-machineinstrs < %s | FileCheck %s
%struct.ShapeData = type { <4 x float>, <4 x float>, <4 x float>, <4 x float>, <4 x float>, i32, i32, i64, <4 x float>, i32, i8, i8, i16, i32, i32 }
-; RUN: llc -O0 -mtriple=amdgcn--amdhsa-amdgiz -mcpu=fiji -mattr=+amdgpu-debugger-emit-prologue -verify-machineinstrs < %s | FileCheck %s
-; RUN: llc -O0 -mtriple=amdgcn--amdhsa-amdgiz -mcpu=fiji -verify-machineinstrs < %s | FileCheck %s --check-prefix=NOATTR
+; RUN: llc -O0 -mtriple=amdgcn--amdhsa -mcpu=fiji -mattr=+amdgpu-debugger-emit-prologue -verify-machineinstrs < %s | FileCheck %s
+; RUN: llc -O0 -mtriple=amdgcn--amdhsa -mcpu=fiji -verify-machineinstrs < %s | FileCheck %s --check-prefix=NOATTR
target datalayout = "A5"
; CHECK: debug_wavefront_private_segment_offset_sgpr = [[SOFF:[0-9]+]]
-; RUN: llc -O0 -mtriple=amdgcn--amdhsa-amdgiz -mcpu=fiji -mattr=+amdgpu-debugger-insert-nops -verify-machineinstrs < %s | FileCheck %s --check-prefix=CHECK
-; RUN: llc -O0 -mtriple=amdgcn--amdhsa-amdgiz -mcpu=fiji -mattr=+amdgpu-debugger-insert-nops -verify-machineinstrs < %s | FileCheck %s --check-prefix=CHECKNOP
+; RUN: llc -O0 -mtriple=amdgcn--amdhsa -mcpu=fiji -mattr=+amdgpu-debugger-insert-nops -verify-machineinstrs < %s | FileCheck %s --check-prefix=CHECK
+; RUN: llc -O0 -mtriple=amdgcn--amdhsa -mcpu=fiji -mattr=+amdgpu-debugger-insert-nops -verify-machineinstrs < %s | FileCheck %s --check-prefix=CHECKNOP
target datalayout = "A5"
; This test expects that we have one instance for each line in some order with "s_nop 0" instances after each.
-; RUN: not llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tahiti -mattr=+promote-alloca -verify-machineinstrs < %s 2>&1 | FileCheck %s
-; RUN: not llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tahiti -mattr=-promote-alloca -verify-machineinstrs < %s 2>&1 | FileCheck %s
-; RUN: not llc -march=r600 -mtriple=r600---amdgiz -mcpu=cypress < %s 2>&1 | FileCheck %s
+; RUN: not llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tahiti -mattr=+promote-alloca -verify-machineinstrs < %s 2>&1 | FileCheck %s
+; RUN: not llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tahiti -mattr=-promote-alloca -verify-machineinstrs < %s 2>&1 | FileCheck %s
+; RUN: not llc -march=r600 -mtriple=r600-- -mcpu=cypress < %s 2>&1 | FileCheck %s
target datalayout = "A5"
; CHECK: in function test_dynamic_stackalloc{{.*}}: unsupported dynamic alloca
-; RUN: llc -march=amdgcn -mtriple=amdgcn-amd-amdhsa-amdgiz -verify-machineinstrs < %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-amd-amdhsa -verify-machineinstrs < %s
; Just check the target feature and data layout is accepted without error.
target datalayout = "e-p:64:64-p1:64:64-p2:64:64-p3:32:32-p4:64:64-p5:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-A5"
-target triple = "amdgcn-amd-amdhsa-amdgiz"
+target triple = "amdgcn-amd-amdhsa"
define void @foo() {
entry:
-; RUN: llc -march=amdgcn -mtriple=amdgcn-amd-amdhsa-amdgizcl -verify-machineinstrs < %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-amd-amdhsa -verify-machineinstrs < %s
; Just check the target feature and data layout is accepted without error.
target datalayout = "e-p:64:64-p1:64:64-p2:64:64-p3:32:32-p4:64:64-p5:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-A5"
-target triple = "amdgcn-amd-amdhsa-amdgizcl"
+target triple = "amdgcn-amd-amdhsa"
define void @foo() {
entry:
-; RUN: llc -debug-only=machine-scheduler -march=amdgcn -mtriple=amdgcn---amdgiz -verify-machineinstrs %s -o - 2>&1| FileCheck -check-prefix=SI-NOHSA -check-prefix=FUNC -check-prefix=DEBUG %s
+; RUN: llc -debug-only=machine-scheduler -march=amdgcn -mtriple=amdgcn-- -verify-machineinstrs %s -o - 2>&1| FileCheck -check-prefix=SI-NOHSA -check-prefix=FUNC -check-prefix=DEBUG %s
target datalayout = "A5"
; REQUIRES: asserts
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SI %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SI %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
; GCN-LABEL: {{^}}extract_vector_elt_v2f16:
; GCN: s_load_dword [[VEC:s[0-9]+]]
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
; GCN-LABEL: {{^}}extract_vector_elt_v3f64_2:
; GCN: buffer_load_dwordx4
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,SI,SIVI %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,VI,SIVI,GFX89 %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=gfx900 -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX9,GFX89 %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,SI,SIVI %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,VI,SIVI,GFX89 %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX9,GFX89 %s
; GCN-LABEL: {{^}}extract_vector_elt_v2i16:
; GCN: s_load_dword [[VEC:s[0-9]+]]
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
; How the replacement of i64 stores with v2i32 stores resulted in
; breaking other users of the bitcast if they already existed
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgizcl -mcpu=kaveri < %s | FileCheck %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=kaveri < %s | FileCheck %s
target datalayout = "e-p:64:64-p1:64:64-p2:64:64-p3:32:32-p4:32:32-p5:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-A5"
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgiz -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN %s
-; RUN: llvm-as -data-layout=A5 < %s | llc -mtriple=amdgcn-amd-amdhsa-amdgiz -mcpu=gfx803 -verify-machineinstrs | FileCheck --check-prefix=GCN %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN %s
+; RUN: llvm-as -data-layout=A5 < %s | llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 -verify-machineinstrs | FileCheck --check-prefix=GCN %s
declare i8 addrspace(4)* @llvm.amdgcn.dispatch.ptr()
declare i8 addrspace(4)* @llvm.amdgcn.implicitarg.ptr()
-; RUN: llc -mtriple=amdgcn---amdgiz -mcpu=kaveri -verify-machineinstrs < %s | FileCheck %s
+; RUN: llc -mtriple=amdgcn-- -mcpu=kaveri -verify-machineinstrs < %s | FileCheck %s
;
; The original OpenCL kernel:
; kernel void f(global int *a, int i, int j) {
; x[i] = 7;
; a[0] = x[j];
; }
-; clang -cc1 -triple amdgcn---amdgizcl -emit-llvm -o -
+; clang -cc1 -triple amdgcn--cl -emit-llvm -o -
target datalayout = "e-p:64:64-p1:64:64-p2:64:64-p3:32:32-p4:32:32-p5:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-A5"
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=hawaii -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,CI %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=fiji -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX89 %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=gfx900 -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX89,GFX9 %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=hawaii -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,CI %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=fiji -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX89 %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX89,GFX9 %s
; GCN-LABEL: {{^}}i1_func_void:
; GCN: buffer_load_ubyte v0, off
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgiz -mcpu=gfx700 -filetype=obj -o - < %s | llvm-readobj -elf-output-style=GNU -notes | FileCheck --check-prefix=CHECK --check-prefix=GFX700 --check-prefix=NOTES %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgiz -mcpu=gfx802 -filetype=obj -o - < %s | llvm-readobj -elf-output-style=GNU -notes | FileCheck --check-prefix=CHECK --check-prefix=GFX802 --check-prefix=NOTES %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgiz -mcpu=gfx900 -filetype=obj -o - < %s | llvm-readobj -elf-output-style=GNU -notes | FileCheck --check-prefix=CHECK --check-prefix=GFX900 --check-prefix=NOTES %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx700 -filetype=obj -o - < %s | llvm-readobj -elf-output-style=GNU -notes | FileCheck --check-prefix=CHECK --check-prefix=GFX700 --check-prefix=NOTES %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx802 -filetype=obj -o - < %s | llvm-readobj -elf-output-style=GNU -notes | FileCheck --check-prefix=CHECK --check-prefix=GFX802 --check-prefix=NOTES %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -filetype=obj -o - < %s | llvm-readobj -elf-output-style=GNU -notes | FileCheck --check-prefix=CHECK --check-prefix=GFX900 --check-prefix=NOTES %s
target datalayout = "A5"
declare void @llvm.dbg.declare(metadata, metadata, metadata)
; RUN: not llvm-as -data-layout=A5 < %s 2>&1 | FileCheck -check-prefixes=COMMON,AS %s
-; RUN: not llc -mtriple amdgcn-amd-amdhsa-amdgiz < %s 2>&1 | FileCheck -check-prefixes=COMMON,LLC %s
-; RUN: llvm-as < %s | not llc -mtriple amdgcn-amd-amdhsa-amdgiz 2>&1 | FileCheck -check-prefixes=COMMON,LLC %s
+; RUN: not llc -mtriple amdgcn-amd-amdhsa < %s 2>&1 | FileCheck -check-prefixes=COMMON,LLC %s
+; RUN: llvm-as < %s | not llc -mtriple amdgcn-amd-amdhsa 2>&1 | FileCheck -check-prefixes=COMMON,LLC %s
; RUN: not opt -data-layout=A5 -S < %s 2>&1 | FileCheck -check-prefixes=COMMON,LLC %s
; RUN: llvm-as < %s | not opt -data-layout=A5 2>&1 | FileCheck -check-prefixes=COMMON,LLC %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s
-; RUN: llc -march=r600 -mtriple=r600---amdgiz -mcpu=cypress < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s
+; RUN: llc -march=r600 -mtriple=r600-- -mcpu=cypress < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
; FUNC-LABEL: {{^}}constant_load_i1:
; GCN: buffer_load_ubyte
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s
-; RUN: llc -march=r600 -mtriple=r600---amdgiz -mcpu=cypress < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s
+; RUN: llc -march=r600 -mtriple=r600-- -mcpu=cypress < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
; FUNC-LABEL: {{^}}global_load_i1:
; GCN: buffer_load_ubyte
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,SICIVI,FUNC %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,SICIVI,FUNC %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GFX9,FUNC %s
-; RUN: llc -march=r600 -mtriple=r600---amdgiz -mcpu=cypress < %s | FileCheck -check-prefixes=EG,FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,SICIVI,FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,SICIVI,FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GFX9,FUNC %s
+; RUN: llc -march=r600 -mtriple=r600-- -mcpu=cypress < %s | FileCheck -check-prefixes=EG,FUNC %s
; FUNC-LABEL: {{^}}local_load_i1:
; SICIVI: s_mov_b32 m0
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=GCN,SI,SICIVI,FUNC %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=GCN,VI,SICIVI,FUNC %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=GCN,GFX9,FUNC %s
-; RUN: llc -march=r600 -mtriple=r600---amdgiz -mcpu=redwood -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=EG -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=GCN,SI,SICIVI,FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=GCN,VI,SICIVI,FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=GCN,GFX9,FUNC %s
+; RUN: llc -march=r600 -mtriple=r600-- -mcpu=redwood -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=EG -check-prefix=FUNC %s
; Testing for ds_read/write_b128
; RUN: llc -march=amdgcn -mcpu=tonga -mattr=+enable-ds128 < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=CIVI,FUNC %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgiz -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
target datalayout = "e-p:64:64-p1:64:64-p2:64:64-p3:32:32-p4:32:32-p5:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-A5"
-; RUN: llc -mtriple=amdgcn--amdhsa-amdgiz -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -mtriple=amdgcn--amdhsa -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
; Effectively, check that the compile finishes; in the case
; of an infinite loop, llc toggles between merging 2 ST4s
-;RUN: llc < %s -march=amdgcn -mtriple=amdgcn---amdgiz -verify-machineinstrs | FileCheck -check-prefixes=CHECK,GCN %s
-;RUN: llc < %s -march=r600 -mtriple=r600---amdgiz -verify-machineinstrs | FileCheck -check-prefixes=CHECK,R600 %s
+;RUN: llc < %s -march=amdgcn -mtriple=amdgcn-- -verify-machineinstrs | FileCheck -check-prefixes=CHECK,GCN %s
+;RUN: llc < %s -march=r600 -mtriple=r600-- -verify-machineinstrs | FileCheck -check-prefixes=CHECK,R600 %s
%struct.S = type { i32 addrspace(5)*, i32 addrspace(1)*, i32 addrspace(4)*, i32 addrspace(3)*, i32*, i32 addrspace(2)*}
-; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mtriple=amdgcn---amdgiz -mattr=+promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mtriple=amdgcn-- -mattr=+promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
target datalayout = "A5"
declare {}* @llvm.invariant.start.p5i8(i64, i8 addrspace(5)* nocapture) #0
-; RUN: opt -mtriple=r600---amdgiz -O3 -aa-eval -print-all-alias-modref-info -disable-output < %s 2>&1 | FileCheck %s
+; RUN: opt -mtriple=r600-- -O3 -aa-eval -print-all-alias-modref-info -disable-output < %s 2>&1 | FileCheck %s
; CHECK: NoAlias: i8 addrspace(5)* %p, i8 addrspace(7)* %p1
-# RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgizcl -verify-machineinstrs -run-pass=machine-scheduler -o - %s | FileCheck %s
+# RUN: llc -mtriple=amdgcn-amd-amdhsa -verify-machineinstrs -run-pass=machine-scheduler -o - %s | FileCheck %s
--- |
%struct.widget.0 = type { float, i32, i32 }
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=verde -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=SI %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=gfx803 -mattr=-flat-for-global -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=SI %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=gfx900 -mattr=-flat-for-global -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX9 %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=verde -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=SI %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx803 -mattr=-flat-for-global -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=SI %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -mattr=-flat-for-global -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX9 %s
; This used to fail due to a v_add_i32 instruction with an illegal immediate
; operand that was created during Local Stack Slot Allocation. Test case derived
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=GCN -check-prefix=FUNC %s
-; RUN: llc -march=r600 -mtriple=r600---amdgiz -mcpu=redwood -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=R600 -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=GCN -check-prefix=FUNC %s
+; RUN: llc -march=r600 -mtriple=r600-- -mcpu=redwood -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=R600 -check-prefix=FUNC %s
declare i32 @llvm.r600.read.tidig.x() nounwind readnone
-; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mtriple=amdgcn---amdgiz -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=SI -check-prefix=FUNC %s
-; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=VI -check-prefix=GFX89 -check-prefix=FUNC %s
-; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=gfx900 -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=GFX9 -check-prefix=GFX89 -check-prefix=FUNC %s
-; RUN: llc -amdgpu-scalarize-global-loads=false -march=r600 -mtriple=r600---amdgiz -mcpu=cypress < %s | FileCheck -enable-var-scope -check-prefix=EG -check-prefix=FUNC %s
+; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mtriple=amdgcn-- -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=SI -check-prefix=FUNC %s
+; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=VI -check-prefix=GFX89 -check-prefix=FUNC %s
+; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=GFX9 -check-prefix=GFX89 -check-prefix=FUNC %s
+; RUN: llc -amdgpu-scalarize-global-loads=false -march=r600 -mtriple=r600-- -mcpu=cypress < %s | FileCheck -enable-var-scope -check-prefix=EG -check-prefix=FUNC %s
; FIXME: i16 promotion pass ruins the scalar cases when legal.
; FIXME: r600 fails verifier
-; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=verde -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=GCN -check-prefix=SI -check-prefix=FUNC %s
-; XUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI -check-prefix=FUNC %s
-; RUN: llc -amdgpu-scalarize-global-loads=false -march=r600 -mtriple=r600---amdgiz -mcpu=redwood -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=EG -check-prefix=FUNC %s
+; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mtriple=amdgcn-- -mcpu=verde -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=GCN -check-prefix=SI -check-prefix=FUNC %s
+; XUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI -check-prefix=FUNC %s
+; RUN: llc -amdgpu-scalarize-global-loads=false -march=r600 -mtriple=r600-- -mcpu=redwood -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=EG -check-prefix=FUNC %s
declare i32 @llvm.r600.read.tidig.x() #0
-# RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgizcl -run-pass=si-lower-control-flow -verify-machineinstrs %s -o - | FileCheck -check-prefixes=GCN %s
+# RUN: llc -mtriple=amdgcn-amd-amdhsa -run-pass=si-lower-control-flow -verify-machineinstrs %s -o - | FileCheck -check-prefixes=GCN %s
# Check that assert is not triggered
# GCN-LABEL: name: si-lower-control-flow{{$}}
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgiz -mcpu=fiji -mattr=-flat-for-global -enable-ipra=0 -amdgpu-sroa=0 -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,VI,CIVI,MESA %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgiz -mcpu=hawaii -enable-ipra=0 -amdgpu-sroa=0 -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,CI,CIVI,MESA %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgiz -mcpu=gfx900 -mattr=-flat-for-global -enable-ipra=0 -amdgpu-sroa=0 -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX9,MESA %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=fiji -mattr=-flat-for-global -enable-ipra=0 -amdgpu-sroa=0 -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,VI,CIVI,MESA %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=hawaii -enable-ipra=0 -amdgpu-sroa=0 -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,CI,CIVI,MESA %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -mattr=-flat-for-global -enable-ipra=0 -amdgpu-sroa=0 -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX9,MESA %s
target datalayout = "A5"
; FIXME: Why is this commuted only sometimes?
-; RUN: opt -S -O1 -mtriple=amdgcn---amdgiz -amdgpu-simplify-libcall < %s | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=GCN-POSTLINK %s
-; RUN: opt -S -O1 -mtriple=amdgcn---amdgiz -amdgpu-simplify-libcall -amdgpu-prelink <%s | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=GCN-PRELINK %s
-; RUN: opt -S -O1 -mtriple=amdgcn---amdgiz -amdgpu-use-native -amdgpu-prelink < %s | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=GCN-NATIVE %s
+; RUN: opt -S -O1 -mtriple=amdgcn-- -amdgpu-simplify-libcall < %s | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=GCN-POSTLINK %s
+; RUN: opt -S -O1 -mtriple=amdgcn-- -amdgpu-simplify-libcall -amdgpu-prelink <%s | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=GCN-PRELINK %s
+; RUN: opt -S -O1 -mtriple=amdgcn-- -amdgpu-use-native -amdgpu-prelink < %s | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=GCN-NATIVE %s
; GCN-LABEL: {{^}}define amdgpu_kernel void @test_sincos
; GCN-POSTLINK: tail call fast float @_Z3sinf(
-; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=verde -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=GCN -check-prefix=SI -check-prefix=FUNC %s
-; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=GCN -check-prefix=VI -check-prefix=FUNC %s
-; RUN: llc -amdgpu-scalarize-global-loads=false -march=r600 -mtriple=r600---amdgiz -mcpu=redwood -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=EG -check-prefix=FUNC %s
+; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mtriple=amdgcn-- -mcpu=verde -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=GCN -check-prefix=SI -check-prefix=FUNC %s
+; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=GCN -check-prefix=VI -check-prefix=FUNC %s
+; RUN: llc -amdgpu-scalarize-global-loads=false -march=r600 -mtriple=r600-- -mcpu=redwood -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefix=EG -check-prefix=FUNC %s
declare i32 @llvm.r600.read.tidig.x() #0
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=verde -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SIVI -check-prefix=FUNC %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SIVI -check-prefix=FUNC %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GFX9 -check-prefix=FUNC %s
-; RUN: llc -march=r600 -mtriple=r600---amdgiz -mcpu=redwood -verify-machineinstrs < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
-; RUN: llc -march=r600 -mtriple=r600---amdgiz -mcpu=cayman -verify-machineinstrs < %s | FileCheck -check-prefix=CM -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=verde -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SIVI -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SIVI -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GFX9 -check-prefix=FUNC %s
+; RUN: llc -march=r600 -mtriple=r600-- -mcpu=redwood -verify-machineinstrs < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
+; RUN: llc -march=r600 -mtriple=r600-- -mcpu=cayman -verify-machineinstrs < %s | FileCheck -check-prefix=CM -check-prefix=FUNC %s
; FUNC-LABEL: {{^}}store_i1:
; EG: MEM_RAT MSKOR
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=verde -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,SICIVI,FUNC %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,SICIVI,FUNC %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GFX9,FUNC %s
-; RUN: llc -march=r600 -mtriple=r600---amdgiz -mcpu=redwood < %s | FileCheck -check-prefixes=EG,FUNC %s
-; RUN: llc -march=r600 -mtriple=r600---amdgiz -mcpu=cayman < %s | FileCheck -check-prefixes=CM,FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=verde -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,SICIVI,FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,SICIVI,FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GFX9,FUNC %s
+; RUN: llc -march=r600 -mtriple=r600-- -mcpu=redwood < %s | FileCheck -check-prefixes=EG,FUNC %s
+; RUN: llc -march=r600 -mtriple=r600-- -mcpu=cayman < %s | FileCheck -check-prefixes=CM,FUNC %s
; FUNC-LABEL: {{^}}store_local_i1:
; SICIVI: s_mov_b32 m0
-; RUN: llc < %s -march=r600 -mtriple=r600---amdgiz -mcpu=redwood | FileCheck %s
+; RUN: llc < %s -march=r600 -mtriple=r600-- -mcpu=redwood | FileCheck %s
; This tests for a bug in the SelectionDAG where custom lowered truncated
; vector stores at the end of a basic block were not being added to the
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=unknown -verify-machineinstrs < %s 2>&1 | FileCheck -check-prefix=ERROR -check-prefix=GCN %s
-; RUN: llc -march=r600 -mtriple=r600---amdgiz -mcpu=unknown -verify-machineinstrs < %s 2>&1 | FileCheck -check-prefix=ERROR -check-prefix=R600 %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=unknown -verify-machineinstrs < %s 2>&1 | FileCheck -check-prefix=ERROR -check-prefix=GCN %s
+; RUN: llc -march=r600 -mtriple=r600-- -mcpu=unknown -verify-machineinstrs < %s 2>&1 | FileCheck -check-prefix=ERROR -check-prefix=R600 %s
target datalayout = "A5"
; Should not crash when the processor is not recognized and the
-; RUN: not llc -march=amdgcn -mtriple=amdgcn---amdgiz -tailcallopt < %s 2>&1 | FileCheck -check-prefix=GCN %s
-; RUN: not llc -march=r600 -mtriple=r600---amdgiz -mcpu=cypress -tailcallopt < %s 2>&1 | FileCheck -check-prefix=R600 %s
+; RUN: not llc -march=amdgcn -mtriple=amdgcn-- -tailcallopt < %s 2>&1 | FileCheck -check-prefix=GCN %s
+; RUN: not llc -march=r600 -mtriple=r600-- -mcpu=cypress -tailcallopt < %s 2>&1 | FileCheck -check-prefix=R600 %s
declare i32 @external_function(i32) nounwind
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=verde -mattr=-promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=SI-ALLOCA -check-prefix=SI -check-prefix=FUNC %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=verde -mattr=+promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=SI-PROMOTE -check-prefix=SI -check-prefix=FUNC %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -mattr=-promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=SI-ALLOCA -check-prefix=SI -check-prefix=FUNC %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tonga -mattr=+promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=SI-PROMOTE -check-prefix=SI -check-prefix=FUNC %s
-; RUN: llc -march=r600 -mtriple=r600---amdgiz -mcpu=redwood < %s | FileCheck --check-prefix=EG -check-prefix=FUNC %s
-; RUN: opt -S -mtriple=amdgcn---amdgiz -amdgpu-promote-alloca -sroa -instcombine < %s | FileCheck -check-prefix=OPT %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=verde -mattr=-promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=SI-ALLOCA -check-prefix=SI -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=verde -mattr=+promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=SI-PROMOTE -check-prefix=SI -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -mattr=-promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=SI-ALLOCA -check-prefix=SI -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tonga -mattr=+promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=SI-PROMOTE -check-prefix=SI -check-prefix=FUNC %s
+; RUN: llc -march=r600 -mtriple=r600-- -mcpu=redwood < %s | FileCheck --check-prefix=EG -check-prefix=FUNC %s
+; RUN: opt -S -mtriple=amdgcn-- -amdgpu-promote-alloca -sroa -instcombine < %s | FileCheck -check-prefix=OPT %s
target datalayout = "A5"
; OPT-LABEL: @vector_read(
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tahiti -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCNMESA -check-prefix=SIMESA %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=fiji -mattr=+vgpr-spilling,-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCNMESA -check-prefix=VIMESA %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=gfx900 -mattr=+vgpr-spilling,-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCNMESA -check-prefix=GFX9MESA %s
-; RUN: llc -march=amdgcn -mcpu=hawaii -mtriple=amdgcn-unknown-amdhsa-amdgiz -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=CIHSA -check-prefix=HSA %s
-; RUN: llc -march=amdgcn -mcpu=fiji -mtriple=amdgcn-unknown-amdhsa-amdgiz -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VIHSA -check-prefix=HSA %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tahiti -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCNMESA -check-prefix=SIMESA %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=fiji -mattr=+vgpr-spilling,-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCNMESA -check-prefix=VIMESA %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -mattr=+vgpr-spilling,-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCNMESA -check-prefix=GFX9MESA %s
+; RUN: llc -march=amdgcn -mcpu=hawaii -mtriple=amdgcn-unknown-amdhsa -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=CIHSA -check-prefix=HSA %s
+; RUN: llc -march=amdgcn -mcpu=fiji -mtriple=amdgcn-unknown-amdhsa -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VIHSA -check-prefix=HSA %s
; This ends up using all 256 registers and requires register
; scavenging which will fail to find an unsued register.
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=tahiti -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SI %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=fiji -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn---amdgiz -mcpu=gfx900 -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GFX9 %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tahiti -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SI %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=fiji -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GFX9 %s
; This ends up using all 255 registers and requires register
; scavenging which will fail to find an unsued register.
; RUN: sed -e "s,SRC_COMPDIR,%/p/Inputs,g" %s > %t.ll
-; RUN: llc -mtriple=amdgcn-amd-amdhsa-amdgiz -mcpu=gfx802 -filetype=obj -O0 -o %t.o %t.ll
-; RUN: llvm-objdump -triple=amdgcn-amd-amdhsa-amdgiz -mcpu=gfx802 -disassemble -line-numbers %t.o | FileCheck --check-prefix=LINE %t.ll
-; RUN: llvm-objdump -triple=amdgcn-amd-amdhsa-amdgiz -mcpu=gfx802 -disassemble -source %t.o | FileCheck --check-prefix=SOURCE %t.ll
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx802 -filetype=obj -O0 -o %t.o %t.ll
+; RUN: llvm-objdump -triple=amdgcn-amd-amdhsa -mcpu=gfx802 -disassemble -line-numbers %t.o | FileCheck --check-prefix=LINE %t.ll
+; RUN: llvm-objdump -triple=amdgcn-amd-amdhsa -mcpu=gfx802 -disassemble -source %t.o | FileCheck --check-prefix=SOURCE %t.ll
; Prologue.
; LINE: source_lines_test:
; ModuleID = 'source-lines.cl'
source_filename = "source-lines.cl"
target datalayout = "e-p:32:32-p1:64:64-p2:64:64-p3:32:32-p4:64:64-p5:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-A5"
-target triple = "amdgcn-amd-amdhsa-amdgiz"
+target triple = "amdgcn-amd-amdhsa"
; Function Attrs: noinline nounwind
define amdgpu_kernel void @source_lines_test(i32 addrspace(1)* %Out) #0 !dbg !7 !kernel_arg_addr_space !12 !kernel_arg_access_qual !13 !kernel_arg_type !14 !kernel_arg_base_type !14 !kernel_arg_type_qual !15 {