#include "AMDGPUISelLowering.h"
#include "AMDGPU.h"
+#include "AMDGPUDiagnosticInfoUnsupported.h"
#include "AMDGPUFrameLowering.h"
#include "AMDGPUIntrinsicInfo.h"
#include "AMDGPURegisterInfo.h"
#include "llvm/CodeGen/SelectionDAG.h"
#include "llvm/CodeGen/TargetLoweringObjectFileImpl.h"
#include "llvm/IR/DataLayout.h"
-#include "llvm/IR/DiagnosticInfo.h"
-#include "llvm/IR/DiagnosticPrinter.h"
using namespace llvm;
-namespace {
-
-/// Diagnostic information for unimplemented or unsupported feature reporting.
-class DiagnosticInfoUnsupported : public DiagnosticInfo {
-private:
- const Twine &Description;
- const Function &Fn;
-
- static int KindID;
-
- static int getKindID() {
- if (KindID == 0)
- KindID = llvm::getNextAvailablePluginDiagnosticKind();
- return KindID;
- }
-
-public:
- DiagnosticInfoUnsupported(const Function &Fn, const Twine &Desc,
- DiagnosticSeverity Severity = DS_Error)
- : DiagnosticInfo(getKindID(), Severity),
- Description(Desc),
- Fn(Fn) { }
-
- const Function &getFunction() const { return Fn; }
- const Twine &getDescription() const { return Description; }
-
- void print(DiagnosticPrinter &DP) const override {
- DP << "unsupported " << getDescription() << " in " << Fn.getName();
- }
-
- static bool classof(const DiagnosticInfo *DI) {
- return DI->getKind() == getKindID();
- }
-};
-
-int DiagnosticInfoUnsupported::KindID = 0;
-}
-
-
static bool allocateStack(unsigned ValNo, MVT ValVT, MVT LocVT,
CCValAssign::LocInfo LocInfo,
ISD::ArgFlagsTy ArgFlags, CCState &State) {
setOperationAction(ISD::BR_JT, MVT::Other, Expand);
setOperationAction(ISD::BRIND, MVT::Other, Expand);
+ // This is totally unsupported, just custom lower to produce an error.
+ setOperationAction(ISD::DYNAMIC_STACKALLOC, MVT::i32, Custom);
+
// We need to custom lower some of the intrinsics
setOperationAction(ISD::INTRINSIC_WO_CHAIN, MVT::Other, Custom);
setOperationAction(ISD::SMUL_LOHI, VT, Expand);
setOperationAction(ISD::UMUL_LOHI, VT, Expand);
setOperationAction(ISD::SDIVREM, VT, Custom);
- setOperationAction(ISD::UDIVREM, VT, Custom);
+ setOperationAction(ISD::UDIVREM, VT, Expand);
setOperationAction(ISD::ADDC, VT, Expand);
setOperationAction(ISD::SUBC, VT, Expand);
setOperationAction(ISD::ADDE, VT, Expand);
setFsqrtIsCheap(true);
+ // We want to find all load dependencies for long chains of stores to enable
+ // merging into very wide vectors. The problem is with vectors with > 4
+ // elements. MergeConsecutiveStores will attempt to merge these because x8/x16
+ // vectors are a legal type, even though we have to split the loads
+ // usually. When we can more precisely specify load legality per address
+ // space, we should be able to make FindBetterChain/MergeConsecutiveStores
+ // smarter so that they can figure out what to do in 2 iterations without all
+ // N > 4 stores on the same chain.
+ GatherAllAliasesMaxDepth = 16;
+
// FIXME: Need to really handle these.
MaxStoresPerMemcpy = 4096;
MaxStoresPerMemmove = 4096;
return true;
}
+bool AMDGPUTargetLowering::aggressivelyPreferBuildVectorSources(EVT VecVT) const {
+ // There are few operations which truly have vector input operands. Any vector
+ // operation is going to involve operations on each component, and a
+ // build_vector will be a copy per element, so it always makes sense to use a
+ // build_vector input in place of the extracted element to avoid a copy into a
+ // super register.
+ //
+ // We should probably only do this if all users are extracts only, but this
+ // should be the common case.
+ return true;
+}
+
bool AMDGPUTargetLowering::isTruncateFree(EVT Source, EVT Dest) const {
// Truncate is just accessing a subregister.
return Dest.bitsLT(Source) && (Dest.getSizeInBits() % 32 == 0);
return SDValue();
}
+SDValue AMDGPUTargetLowering::LowerDYNAMIC_STACKALLOC(SDValue Op,
+ SelectionDAG &DAG) const {
+ const Function &Fn = *DAG.getMachineFunction().getFunction();
+
+ DiagnosticInfoUnsupported NoDynamicAlloca(Fn, "dynamic alloca");
+ DAG.getContext()->diagnose(NoDynamicAlloca);
+ return SDValue();
+}
+
SDValue AMDGPUTargetLowering::LowerOperation(SDValue Op,
SelectionDAG &DAG) const {
switch (Op.getOpcode()) {
case ISD::UINT_TO_FP: return LowerUINT_TO_FP(Op, DAG);
case ISD::FP_TO_SINT: return LowerFP_TO_SINT(Op, DAG);
case ISD::FP_TO_UINT: return LowerFP_TO_UINT(Op, DAG);
+ case ISD::DYNAMIC_STACKALLOC: return LowerDYNAMIC_STACKALLOC(Op, DAG);
}
return Op;
}
SelectionDAG &DAG) const {
SDLoc DL(Op);
EVT VT = Op.getValueType();
+ // TODO: Should this propagate fast-math-flags?
SDValue OneSubA = DAG.getNode(ISD::FSUB, DL, VT,
DAG.getConstantFP(1.0f, DL, MVT::f32),
Op.getOperand(1));
if ((Store->getAddressSpace() == AMDGPUAS::LOCAL_ADDRESS ||
Store->getAddressSpace() == AMDGPUAS::PRIVATE_ADDRESS) &&
Store->getValue().getValueType().isVector()) {
- return ScalarizeVectorStore(Op, DAG);
+ return SplitVectorStore(Op, DAG);
}
EVT MemVT = Store->getMemoryVT();
// float fb = (float)ib;
SDValue fb = DAG.getNode(ToFp, DL, FltVT, ib);
+ // TODO: Should this propagate fast-math-flags?
// float fq = native_divide(fa, fb);
SDValue fq = DAG.getNode(ISD::FMUL, DL, FltVT,
fa, DAG.getNode(AMDGPUISD::RCP, DL, FltVT, fb));
SDValue X = Op.getOperand(0);
SDValue Y = Op.getOperand(1);
+ // TODO: Should this propagate fast-math-flags?
+
SDValue Div = DAG.getNode(ISD::FDIV, SL, VT, X, Y);
SDValue Floor = DAG.getNode(ISD::FTRUNC, SL, VT, Div);
SDValue Mul = DAG.getNode(ISD::FMUL, SL, VT, Floor, Y);
SDValue And = DAG.getNode(ISD::AND, SL, SetCCVT, Lt0, NeTrunc);
SDValue Add = DAG.getNode(ISD::SELECT, SL, MVT::f64, And, One, Zero);
+ // TODO: Should this propagate fast-math-flags?
return DAG.getNode(ISD::FADD, SL, MVT::f64, Trunc, Add);
}
SDValue C1 = DAG.getConstantFP(C1Val, SL, MVT::f64);
SDValue CopySign = DAG.getNode(ISD::FCOPYSIGN, SL, MVT::f64, C1, Src);
+ // TODO: Should this propagate fast-math-flags?
+
SDValue Tmp1 = DAG.getNode(ISD::FADD, SL, MVT::f64, Src, CopySign);
SDValue Tmp2 = DAG.getNode(ISD::FSUB, SL, MVT::f64, Tmp1, CopySign);
SDValue T = DAG.getNode(ISD::FTRUNC, SL, MVT::f32, X);
+ // TODO: Should this propagate fast-math-flags?
+
SDValue Diff = DAG.getNode(ISD::FSUB, SL, MVT::f32, X, T);
SDValue AbsDiff = DAG.getNode(ISD::FABS, SL, MVT::f32, Diff);
SDValue And = DAG.getNode(ISD::AND, SL, SetCCVT, Lt0, NeTrunc);
SDValue Add = DAG.getNode(ISD::SELECT, SL, MVT::f64, And, NegOne, Zero);
+ // TODO: Should this propagate fast-math-flags?
return DAG.getNode(ISD::FADD, SL, MVT::f64, Trunc, Add);
}
SDValue LdExp = DAG.getNode(AMDGPUISD::LDEXP, SL, MVT::f64, CvtHi,
DAG.getConstant(32, SL, MVT::i32));
-
+ // TODO: Should this propagate fast-math-flags?
return DAG.getNode(ISD::FADD, SL, MVT::f64, LdExp, CvtLo);
}
SDValue Hi = DAG.getNode(ISD::EXTRACT_ELEMENT, DL, MVT::i32, S0,
DAG.getConstant(1, DL, MVT::i32));
SDValue FloatHi = DAG.getNode(ISD::UINT_TO_FP, DL, MVT::f32, Hi);
+ // TODO: Should this propagate fast-math-flags?
FloatHi = DAG.getNode(ISD::FMUL, DL, MVT::f32, FloatHi,
DAG.getConstantFP(4294967296.0f, DL, MVT::f32)); // 2^32
return DAG.getNode(ISD::FADD, DL, MVT::f32, FloatLo, FloatHi);
MVT::f64);
SDValue K1 = DAG.getConstantFP(BitsToDouble(UINT64_C(0xc1f0000000000000)), SL,
MVT::f64);
-
+ // TODO: Should this propagate fast-math-flags?
SDValue Mul = DAG.getNode(ISD::FMUL, SL, MVT::f64, Trunc, K0);
SDValue FloorMul = DAG.getNode(ISD::FFLOOR, SL, MVT::f64, Mul);
if (ConstantFPSDNode * CFP = dyn_cast<ConstantFPSDNode>(Op)) {
return CFP->isExactlyValue(1.0);
}
- if (ConstantSDNode *C = dyn_cast<ConstantSDNode>(Op)) {
- return C->isAllOnesValue();
- }
- return false;
+ return isAllOnesConstant(Op);
}
bool AMDGPUTargetLowering::isHWFalseValue(SDValue Op) const {
if (ConstantFPSDNode * CFP = dyn_cast<ConstantFPSDNode>(Op)) {
return CFP->getValueAPF().isZero();
}
- if (ConstantSDNode *C = dyn_cast<ConstantSDNode>(Op)) {
- return C->isNullValue();
- }
- return false;
+ return isNullConstant(Op);
}
SDValue AMDGPUTargetLowering::CreateLiveInRegister(SelectionDAG &DAG,
return 1;
unsigned SignBits = 32 - Width->getZExtValue() + 1;
- ConstantSDNode *Offset = dyn_cast<ConstantSDNode>(Op.getOperand(1));
- if (!Offset || !Offset->isNullValue())
+ if (!isNullConstant(Op.getOperand(1)))
return SignBits;
// TODO: Could probably figure something out with non-0 offsets.