[llvm] r256433 - [X86] Fix shuffle decoding for variable VPERMIL to be tolerant of the Constant type not matching due to folding in the constant pool and to get VPERMILPD correct.
Craig Topper via llvm-commits
llvm-commits at lists.llvm.org
Fri Dec 25 20:50:07 PST 2015
Author: ctopper
Date: Fri Dec 25 22:50:07 2015
New Revision: 256433
URL: http://llvm.org/viewvc/llvm-project?rev=256433&view=rev
Log:
[X86] Fix shuffle decoding for variable VPERMIL to be tolerant of the Constant type not matching due to folding in the constant pool and to get VPERMILPD correct.
Modified:
llvm/trunk/lib/Target/X86/Utils/X86ShuffleDecode.cpp
llvm/trunk/lib/Target/X86/Utils/X86ShuffleDecode.h
llvm/trunk/lib/Target/X86/X86MCInstLower.cpp
Modified: llvm/trunk/lib/Target/X86/Utils/X86ShuffleDecode.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/Utils/X86ShuffleDecode.cpp?rev=256433&r1=256432&r2=256433&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/Utils/X86ShuffleDecode.cpp (original)
+++ llvm/trunk/lib/Target/X86/Utils/X86ShuffleDecode.cpp Fri Dec 25 22:50:07 2015
@@ -388,44 +388,66 @@ void DecodeVPERMMask(unsigned Imm, Small
}
}
-void DecodeVPERMILPMask(const Constant *C, SmallVectorImpl<int> &ShuffleMask) {
+void DecodeVPERMILPMask(const Constant *C, unsigned ElSize,
+ SmallVectorImpl<int> &ShuffleMask) {
Type *MaskTy = C->getType();
- assert(MaskTy->isVectorTy() && "Expected a vector constant mask!");
- assert(MaskTy->getVectorElementType()->isIntegerTy() &&
- "Expected integer constant mask elements!");
- int ElementBits = MaskTy->getScalarSizeInBits();
- int NumElements = MaskTy->getVectorNumElements();
+ // It is not an error for the PSHUFB mask to not be a vector of i8 because the
+ // constant pool uniques constants by their bit representation.
+ // e.g. the following take up the same space in the constant pool:
+ // i128 -170141183420855150465331762880109871104
+ //
+ // <2 x i64> <i64 -9223372034707292160, i64 -9223372034707292160>
+ //
+ // <4 x i32> <i32 -2147483648, i32 -2147483648,
+ // i32 -2147483648, i32 -2147483648>
+
+ unsigned MaskTySize = MaskTy->getPrimitiveSizeInBits();
+
+ if (MaskTySize != 128 && MaskTySize != 256) // FIXME: Add support for AVX-512.
+ return;
+
+ // Only support vector types.
+ if (!MaskTy->isVectorTy())
+ return;
+
+ // Make sure its an integer type.
+ Type *VecEltTy = MaskTy->getVectorElementType();
+ if (!VecEltTy->isIntegerTy())
+ return;
+
+ // Support any element type from byte up to element size.
+ // This is necesary primarily because 64-bit elements get split to 32-bit
+ // in the constant pool on 32-bit target.
+ unsigned EltTySize = VecEltTy->getIntegerBitWidth();
+ if (EltTySize < 8 || EltTySize > ElSize)
+ return;
+
+ unsigned NumElements = MaskTySize / ElSize;
assert((NumElements == 2 || NumElements == 4 || NumElements == 8) &&
"Unexpected number of vector elements.");
ShuffleMask.reserve(NumElements);
- if (auto *CDS = dyn_cast<ConstantDataSequential>(C)) {
- assert((unsigned)NumElements == CDS->getNumElements() &&
- "Constant mask has a different number of elements!");
-
- for (int i = 0; i < NumElements; ++i) {
- int Base = (i * ElementBits / 128) * (128 / ElementBits);
- uint64_t Element = CDS->getElementAsInteger(i);
- // Only the least significant 2 bits of the integer are used.
- int Index = Base + (Element & 0x3);
- ShuffleMask.push_back(Index);
- }
- } else if (auto *CV = dyn_cast<ConstantVector>(C)) {
- assert((unsigned)NumElements == C->getNumOperands() &&
- "Constant mask has a different number of elements!");
-
- for (int i = 0; i < NumElements; ++i) {
- int Base = (i * ElementBits / 128) * (128 / ElementBits);
- Constant *COp = CV->getOperand(i);
- if (isa<UndefValue>(COp)) {
- ShuffleMask.push_back(SM_SentinelUndef);
- continue;
- }
- uint64_t Element = cast<ConstantInt>(COp)->getZExtValue();
- // Only the least significant 2 bits of the integer are used.
- int Index = Base + (Element & 0x3);
- ShuffleMask.push_back(Index);
+ unsigned NumElementsPerLane = 128 / ElSize;
+ unsigned Factor = ElSize / EltTySize;
+
+ for (unsigned i = 0; i < NumElements; ++i) {
+ Constant *COp = C->getAggregateElement(i * Factor);
+ if (!COp) {
+ ShuffleMask.clear();
+ return;
+ } else if (isa<UndefValue>(COp)) {
+ ShuffleMask.push_back(SM_SentinelUndef);
+ continue;
}
+ int Index = i & ~(NumElementsPerLane - 1);
+ uint64_t Element = cast<ConstantInt>(COp)->getZExtValue();
+ if (ElSize == 64)
+ Index += (Element >> 1) & 0x1;
+ else
+ Index += Element & 0x3;
+ ShuffleMask.push_back(Index);
}
+
+ // TODO: Handle funny-looking vectors too.
}
void DecodeZeroExtendMask(MVT SrcVT, MVT DstVT, SmallVectorImpl<int> &Mask) {
Modified: llvm/trunk/lib/Target/X86/Utils/X86ShuffleDecode.h
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/Utils/X86ShuffleDecode.h?rev=256433&r1=256432&r2=256433&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/Utils/X86ShuffleDecode.h (original)
+++ llvm/trunk/lib/Target/X86/Utils/X86ShuffleDecode.h Fri Dec 25 22:50:07 2015
@@ -96,7 +96,8 @@ void decodeVSHUF64x2FamilyMask(MVT VT, u
void DecodeVPERMMask(unsigned Imm, SmallVectorImpl<int> &ShuffleMask);
/// \brief Decode a VPERMILP variable mask from an IR-level vector constant.
-void DecodeVPERMILPMask(const Constant *C, SmallVectorImpl<int> &ShuffleMask);
+void DecodeVPERMILPMask(const Constant *C, unsigned ElSize,
+ SmallVectorImpl<int> &ShuffleMask);
/// \brief Decode a zero extension instruction as a shuffle mask.
void DecodeZeroExtendMask(MVT SrcVT, MVT DstVT,
Modified: llvm/trunk/lib/Target/X86/X86MCInstLower.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86MCInstLower.cpp?rev=256433&r1=256432&r2=256433&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86MCInstLower.cpp (original)
+++ llvm/trunk/lib/Target/X86/X86MCInstLower.cpp Fri Dec 25 22:50:07 2015
@@ -1303,9 +1303,16 @@ void X86AsmPrinter::EmitInstruction(cons
const MachineOperand &SrcOp = MI->getOperand(1);
const MachineOperand &MaskOp = MI->getOperand(5);
+ unsigned ElSize;
+ switch (MI->getOpcode()) {
+ default: llvm_unreachable("Invalid opcode");
+ case X86::VPERMILPSrm: case X86::VPERMILPSYrm: ElSize = 32; break;
+ case X86::VPERMILPDrm: case X86::VPERMILPDYrm: ElSize = 64; break;
+ }
+
if (auto *C = getConstantFromPool(*MI, MaskOp)) {
SmallVector<int, 16> Mask;
- DecodeVPERMILPMask(C, Mask);
+ DecodeVPERMILPMask(C, ElSize, Mask);
if (!Mask.empty())
OutStreamer->AddComment(getShuffleComment(DstOp, SrcOp, Mask));
}
More information about the llvm-commits
mailing list