llvm interpreter: select, shuffle and insertelement instructions.
This patch implements vector support for select instruction and adds specific vector instructions : shuffle and insertelement. (tests are also included)
and functions lle_X_memset, lle_X_memcpy added.
Done by Veselov, Yuri (mailto:Yuri.Veselov@intel.com)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189735 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/lib/ExecutionEngine/Interpreter/Execution.cpp b/lib/ExecutionEngine/Interpreter/Execution.cpp
index fc3d579..e02ba15 100644
--- a/lib/ExecutionEngine/Interpreter/Execution.cpp
+++ b/lib/ExecutionEngine/Interpreter/Execution.cpp
@@ -786,20 +786,31 @@
}
static GenericValue executeSelectInst(GenericValue Src1, GenericValue Src2,
- GenericValue Src3) {
- return Src1.IntVal == 0 ? Src3 : Src2;
+ GenericValue Src3, const Type *Ty) {
+ GenericValue Dest;
+ if(Ty->isVectorTy()) {
+ assert(Src1.AggregateVal.size() == Src2.AggregateVal.size());
+ assert(Src2.AggregateVal.size() == Src3.AggregateVal.size());
+ Dest.AggregateVal.resize( Src1.AggregateVal.size() );
+ for (size_t i = 0; i < Src1.AggregateVal.size(); ++i)
+ Dest.AggregateVal[i] = (Src1.AggregateVal[i].IntVal == 0) ?
+ Src3.AggregateVal[i] : Src2.AggregateVal[i];
+ } else {
+ Dest = (Src1.IntVal == 0) ? Src3 : Src2;
+ }
+ return Dest;
}
void Interpreter::visitSelectInst(SelectInst &I) {
ExecutionContext &SF = ECStack.back();
+ const Type * Ty = I.getOperand(0)->getType();
GenericValue Src1 = getOperandValue(I.getOperand(0), SF);
GenericValue Src2 = getOperandValue(I.getOperand(1), SF);
GenericValue Src3 = getOperandValue(I.getOperand(2), SF);
- GenericValue R = executeSelectInst(Src1, Src2, Src3);
+ GenericValue R = executeSelectInst(Src1, Src2, Src3, Ty);
SetValue(&I, R, SF);
}
-
//===----------------------------------------------------------------------===//
// Terminator Instruction Implementations
//===----------------------------------------------------------------------===//
@@ -1793,10 +1804,115 @@
SetValue(&I, Dest, SF);
}
+void Interpreter::visitInsertElementInst(InsertElementInst &I) {
+ ExecutionContext &SF = ECStack.back();
+ Type *Ty = I.getType();
+
+ if(!(Ty->isVectorTy()) )
+ llvm_unreachable("Unhandled dest type for insertelement instruction");
+
+ GenericValue Src1 = getOperandValue(I.getOperand(0), SF);
+ GenericValue Src2 = getOperandValue(I.getOperand(1), SF);
+ GenericValue Src3 = getOperandValue(I.getOperand(2), SF);
+ GenericValue Dest;
+
+ Type *TyContained = Ty->getContainedType(0);
+
+ const unsigned indx = unsigned(Src3.IntVal.getZExtValue());
+ Dest.AggregateVal = Src1.AggregateVal;
+
+ if(Src1.AggregateVal.size() <= indx)
+ llvm_unreachable("Invalid index in insertelement instruction");
+ switch (TyContained->getTypeID()) {
+ default:
+ llvm_unreachable("Unhandled dest type for insertelement instruction");
+ case Type::IntegerTyID:
+ Dest.AggregateVal[indx].IntVal = Src2.IntVal;
+ break;
+ case Type::FloatTyID:
+ Dest.AggregateVal[indx].FloatVal = Src2.FloatVal;
+ break;
+ case Type::DoubleTyID:
+ Dest.AggregateVal[indx].DoubleVal = Src2.DoubleVal;
+ break;
+ }
+ SetValue(&I, Dest, SF);
+}
+
+void Interpreter::visitShuffleVectorInst(ShuffleVectorInst &I){
+ ExecutionContext &SF = ECStack.back();
+
+ Type *Ty = I.getType();
+ if(!(Ty->isVectorTy()))
+ llvm_unreachable("Unhandled dest type for shufflevector instruction");
+
+ GenericValue Src1 = getOperandValue(I.getOperand(0), SF);
+ GenericValue Src2 = getOperandValue(I.getOperand(1), SF);
+ GenericValue Src3 = getOperandValue(I.getOperand(2), SF);
+ GenericValue Dest;
+
+ // There is no need to check types of src1 and src2, because the compiled
+ // bytecode can't contain different types for src1 and src2 for a
+ // shufflevector instruction.
+
+ Type *TyContained = Ty->getContainedType(0);
+ unsigned src1Size = (unsigned)Src1.AggregateVal.size();
+ unsigned src2Size = (unsigned)Src2.AggregateVal.size();
+ unsigned src3Size = (unsigned)Src3.AggregateVal.size();
+
+ Dest.AggregateVal.resize(src3Size);
+
+ switch (TyContained->getTypeID()) {
+ default:
+ llvm_unreachable("Unhandled dest type for insertelement instruction");
+ break;
+ case Type::IntegerTyID:
+ for( unsigned i=0; i<src3Size; i++) {
+ unsigned j = Src3.AggregateVal[i].IntVal.getZExtValue();
+ if(j < src1Size)
+ Dest.AggregateVal[i].IntVal = Src1.AggregateVal[j].IntVal;
+ else if(j < src1Size + src2Size)
+ Dest.AggregateVal[i].IntVal = Src2.AggregateVal[j-src1Size].IntVal;
+ else
+ // The selector may not be greater than sum of lengths of first and
+ // second operands and llasm should not allow situation like
+ // %tmp = shufflevector <2 x i32> <i32 3, i32 4>, <2 x i32> undef,
+ // <2 x i32> < i32 0, i32 5 >,
+ // where i32 5 is invalid, but let it be additional check here:
+ llvm_unreachable("Invalid mask in shufflevector instruction");
+ }
+ break;
+ case Type::FloatTyID:
+ for( unsigned i=0; i<src3Size; i++) {
+ unsigned j = Src3.AggregateVal[i].IntVal.getZExtValue();
+ if(j < src1Size)
+ Dest.AggregateVal[i].FloatVal = Src1.AggregateVal[j].FloatVal;
+ else if(j < src1Size + src2Size)
+ Dest.AggregateVal[i].FloatVal = Src2.AggregateVal[j-src1Size].FloatVal;
+ else
+ llvm_unreachable("Invalid mask in shufflevector instruction");
+ }
+ break;
+ case Type::DoubleTyID:
+ for( unsigned i=0; i<src3Size; i++) {
+ unsigned j = Src3.AggregateVal[i].IntVal.getZExtValue();
+ if(j < src1Size)
+ Dest.AggregateVal[i].DoubleVal = Src1.AggregateVal[j].DoubleVal;
+ else if(j < src1Size + src2Size)
+ Dest.AggregateVal[i].DoubleVal =
+ Src2.AggregateVal[j-src1Size].DoubleVal;
+ else
+ llvm_unreachable("Invalid mask in shufflevector instruction");
+ }
+ break;
+ }
+ SetValue(&I, Dest, SF);
+}
+
GenericValue Interpreter::getConstantExprValue (ConstantExpr *CE,
ExecutionContext &SF) {
switch (CE->getOpcode()) {
- case Instruction::Trunc:
+ case Instruction::Trunc:
return executeTruncInst(CE->getOperand(0), CE->getType(), SF);
case Instruction::ZExt:
return executeZExtInst(CE->getOperand(0), CE->getType(), SF);
@@ -1832,7 +1948,8 @@
case Instruction::Select:
return executeSelectInst(getOperandValue(CE->getOperand(0), SF),
getOperandValue(CE->getOperand(1), SF),
- getOperandValue(CE->getOperand(2), SF));
+ getOperandValue(CE->getOperand(2), SF),
+ CE->getOperand(0)->getType());
default :
break;
}
diff --git a/lib/ExecutionEngine/Interpreter/ExternalFunctions.cpp b/lib/ExecutionEngine/Interpreter/ExternalFunctions.cpp
index bef4bbf..6fb7973 100644
--- a/lib/ExecutionEngine/Interpreter/ExternalFunctions.cpp
+++ b/lib/ExecutionEngine/Interpreter/ExternalFunctions.cpp
@@ -406,6 +406,7 @@
break;
}
}
+ return GV;
}
// int printf(const char *, ...) - a very rough implementation to make output
@@ -434,7 +435,7 @@
GenericValue GV;
GV.IntVal = APInt(32, sscanf(Args[0], Args[1], Args[2], Args[3], Args[4],
- Args[5], Args[6], Args[7], Args[8], Args[9]));
+ Args[5], Args[6], Args[7], Args[8], Args[9]));
return GV;
}
@@ -450,7 +451,7 @@
GenericValue GV;
GV.IntVal = APInt(32, scanf( Args[0], Args[1], Args[2], Args[3], Args[4],
- Args[5], Args[6], Args[7], Args[8], Args[9]));
+ Args[5], Args[6], Args[7], Args[8], Args[9]));
return GV;
}
@@ -470,6 +471,31 @@
return GV;
}
+GenericValue lle_X_memset(FunctionType *FT,
+ const std::vector<GenericValue> &Args) {
+ int val = (int)Args[1].IntVal.getSExtValue();
+ size_t len = (size_t)Args[2].IntVal.getZExtValue();
+ memset((void*)GVTOP(Args[0]),val, len);
+ // llvm.memset.* returns void, lle_X_* returns GenericValue,
+ // so here we return GenericValue with IntVal set to zero
+ GenericValue GV;
+ GV.IntVal = 0;
+ return GV;
+}
+
+GenericValue lle_X_memcpy(FunctionType *FT,
+ const std::vector<GenericValue> &Args) {
+
+ memcpy(GVTOP(Args[0]), GVTOP(Args[1]),
+ (size_t)(Args[2].IntVal.getLimitedValue()));
+
+ // llvm.mecpy* returns void, lle_X_* returns GenericValue,
+ // so here we return GenericValue with IntVal set to zero
+ GenericValue GV;
+ GV.IntVal = 0;
+ return GV;
+}
+
void Interpreter::initializeExternalFunctions() {
sys::ScopedLock Writer(*FunctionsLock);
FuncNames["lle_X_atexit"] = lle_X_atexit;
@@ -481,4 +507,6 @@
FuncNames["lle_X_sscanf"] = lle_X_sscanf;
FuncNames["lle_X_scanf"] = lle_X_scanf;
FuncNames["lle_X_fprintf"] = lle_X_fprintf;
+ FuncNames["lle_X_memset"] = lle_X_memset;
+ FuncNames["lle_X_memcpy"] = lle_X_memcpy;
}
diff --git a/lib/ExecutionEngine/Interpreter/Interpreter.h b/lib/ExecutionEngine/Interpreter/Interpreter.h
index 2952d7e..802b53e 100644
--- a/lib/ExecutionEngine/Interpreter/Interpreter.h
+++ b/lib/ExecutionEngine/Interpreter/Interpreter.h
@@ -179,6 +179,9 @@
void visitVAArgInst(VAArgInst &I);
void visitExtractElementInst(ExtractElementInst &I);
+ void visitInsertElementInst(InsertElementInst &I);
+ void visitShuffleVectorInst(ShuffleVectorInst &I);
+
void visitInstruction(Instruction &I) {
errs() << I << "\n";
llvm_unreachable("Instruction not interpretable yet!");