//
//===----------------------------------------------------------------------===//
-#define DEBUG_TYPE "arm-selectiondag-info"
#include "ARMTargetMachine.h"
+#include "llvm/CodeGen/SelectionDAG.h"
+#include "llvm/IR/DerivedTypes.h"
using namespace llvm;
-ARMSelectionDAGInfo::ARMSelectionDAGInfo(const TargetMachine &TM)
- : TargetSelectionDAGInfo(TM),
- Subtarget(&TM.getSubtarget<ARMSubtarget>()) {
-}
+#define DEBUG_TYPE "arm-selectiondag-info"
+
+ARMSelectionDAGInfo::ARMSelectionDAGInfo(const DataLayout &DL)
+ : TargetSelectionDAGInfo(&DL) {}
ARMSelectionDAGInfo::~ARMSelectionDAGInfo() {
}
SDValue
-ARMSelectionDAGInfo::EmitTargetCodeForMemcpy(SelectionDAG &DAG, DebugLoc dl,
+ARMSelectionDAGInfo::EmitTargetCodeForMemcpy(SelectionDAG &DAG, SDLoc dl,
SDValue Chain,
SDValue Dst, SDValue Src,
SDValue Size, unsigned Align,
bool isVolatile, bool AlwaysInline,
MachinePointerInfo DstPtrInfo,
MachinePointerInfo SrcPtrInfo) const {
+ const ARMSubtarget &Subtarget =
+ DAG.getMachineFunction().getSubtarget<ARMSubtarget>();
// Do repeated 4-byte loads and stores. To be improved.
// This requires 4-byte alignment.
if ((Align & 3) != 0)
if (!ConstantSize)
return SDValue();
uint64_t SizeVal = ConstantSize->getZExtValue();
- if (!AlwaysInline && SizeVal > Subtarget->getMaxInlineSizeThreshold())
+ if (!AlwaysInline && SizeVal > Subtarget.getMaxInlineSizeThreshold())
return SDValue();
unsigned BytesLeft = SizeVal & 3;
EVT VT = MVT::i32;
unsigned VTSize = 4;
unsigned i = 0;
- const unsigned MAX_LOADS_IN_LDM = 6;
- SDValue TFOps[MAX_LOADS_IN_LDM];
- SDValue Loads[MAX_LOADS_IN_LDM];
+ // Emit a maximum of 4 loads in Thumb1 since we have fewer registers
+ const unsigned MAX_LOADS_IN_LDM = Subtarget.isThumb1Only() ? 4 : 6;
+ SDValue TFOps[6];
+ SDValue Loads[6];
uint64_t SrcOff = 0, DstOff = 0;
// Emit up to MAX_LOADS_IN_LDM loads, then a TokenFactor barrier, then the
i < MAX_LOADS_IN_LDM && EmittedNumMemOps + i < NumMemOps; ++i) {
Loads[i] = DAG.getLoad(VT, dl, Chain,
DAG.getNode(ISD::ADD, dl, MVT::i32, Src,
- DAG.getConstant(SrcOff, MVT::i32)),
+ DAG.getConstant(SrcOff, dl, MVT::i32)),
SrcPtrInfo.getWithOffset(SrcOff), isVolatile,
- false, 0);
+ false, false, 0);
TFOps[i] = Loads[i].getValue(1);
SrcOff += VTSize;
}
- Chain = DAG.getNode(ISD::TokenFactor, dl, MVT::Other, &TFOps[0], i);
+ Chain = DAG.getNode(ISD::TokenFactor, dl, MVT::Other,
+ makeArrayRef(TFOps, i));
for (i = 0;
i < MAX_LOADS_IN_LDM && EmittedNumMemOps + i < NumMemOps; ++i) {
TFOps[i] = DAG.getStore(Chain, dl, Loads[i],
DAG.getNode(ISD::ADD, dl, MVT::i32, Dst,
- DAG.getConstant(DstOff, MVT::i32)),
+ DAG.getConstant(DstOff, dl, MVT::i32)),
DstPtrInfo.getWithOffset(DstOff),
isVolatile, false, 0);
DstOff += VTSize;
}
- Chain = DAG.getNode(ISD::TokenFactor, dl, MVT::Other, &TFOps[0], i);
+ Chain = DAG.getNode(ISD::TokenFactor, dl, MVT::Other,
+ makeArrayRef(TFOps, i));
EmittedNumMemOps += i;
}
Loads[i] = DAG.getLoad(VT, dl, Chain,
DAG.getNode(ISD::ADD, dl, MVT::i32, Src,
- DAG.getConstant(SrcOff, MVT::i32)),
- SrcPtrInfo.getWithOffset(SrcOff), false, false, 0);
+ DAG.getConstant(SrcOff, dl, MVT::i32)),
+ SrcPtrInfo.getWithOffset(SrcOff),
+ false, false, false, 0);
TFOps[i] = Loads[i].getValue(1);
++i;
SrcOff += VTSize;
BytesLeft -= VTSize;
}
- Chain = DAG.getNode(ISD::TokenFactor, dl, MVT::Other, &TFOps[0], i);
+ Chain = DAG.getNode(ISD::TokenFactor, dl, MVT::Other,
+ makeArrayRef(TFOps, i));
i = 0;
BytesLeft = BytesLeftSave;
TFOps[i] = DAG.getStore(Chain, dl, Loads[i],
DAG.getNode(ISD::ADD, dl, MVT::i32, Dst,
- DAG.getConstant(DstOff, MVT::i32)),
+ DAG.getConstant(DstOff, dl, MVT::i32)),
DstPtrInfo.getWithOffset(DstOff), false, false, 0);
++i;
DstOff += VTSize;
BytesLeft -= VTSize;
}
- return DAG.getNode(ISD::TokenFactor, dl, MVT::Other, &TFOps[0], i);
+ return DAG.getNode(ISD::TokenFactor, dl, MVT::Other,
+ makeArrayRef(TFOps, i));
+}
+
+// Adjust parameters for memset, EABI uses format (ptr, size, value),
+// GNU library uses (ptr, value, size)
+// See RTABI section 4.3.4
+SDValue ARMSelectionDAGInfo::
+EmitTargetCodeForMemset(SelectionDAG &DAG, SDLoc dl,
+ SDValue Chain, SDValue Dst,
+ SDValue Src, SDValue Size,
+ unsigned Align, bool isVolatile,
+ MachinePointerInfo DstPtrInfo) const {
+ const ARMSubtarget &Subtarget =
+ DAG.getMachineFunction().getSubtarget<ARMSubtarget>();
+ // Use default for non-AAPCS (or MachO) subtargets
+ if (!Subtarget.isAAPCS_ABI() || Subtarget.isTargetMachO() ||
+ Subtarget.isTargetWindows())
+ return SDValue();
+
+ const ARMTargetLowering &TLI = *Subtarget.getTargetLowering();
+ TargetLowering::ArgListTy Args;
+ TargetLowering::ArgListEntry Entry;
+
+ // First argument: data pointer
+ Type *IntPtrTy = TLI.getDataLayout()->getIntPtrType(*DAG.getContext());
+ Entry.Node = Dst;
+ Entry.Ty = IntPtrTy;
+ Args.push_back(Entry);
+
+ // Second argument: buffer size
+ Entry.Node = Size;
+ Entry.Ty = IntPtrTy;
+ Entry.isSExt = false;
+ Args.push_back(Entry);
+
+ // Extend or truncate the argument to be an i32 value for the call.
+ if (Src.getValueType().bitsGT(MVT::i32))
+ Src = DAG.getNode(ISD::TRUNCATE, dl, MVT::i32, Src);
+ else
+ Src = DAG.getNode(ISD::ZERO_EXTEND, dl, MVT::i32, Src);
+
+ // Third argument: value to fill
+ Entry.Node = Src;
+ Entry.Ty = Type::getInt32Ty(*DAG.getContext());
+ Entry.isSExt = true;
+ Args.push_back(Entry);
+
+ // Emit __eabi_memset call
+ TargetLowering::CallLoweringInfo CLI(DAG);
+ CLI.setDebugLoc(dl).setChain(Chain)
+ .setCallee(TLI.getLibcallCallingConv(RTLIB::MEMSET),
+ Type::getVoidTy(*DAG.getContext()),
+ DAG.getExternalSymbol(TLI.getLibcallName(RTLIB::MEMSET),
+ TLI.getPointerTy()), std::move(Args), 0)
+ .setDiscardResult();
+
+ std::pair<SDValue,SDValue> CallResult = TLI.LowerCallTo(CLI);
+ return CallResult.second;
}
projects / oota-llvm.git / blobdiff