aboutsummaryrefslogtreecommitdiff
path: root/clang/lib/AST/ByteCode
diff options
context:
space:
mode:
Diffstat (limited to 'clang/lib/AST/ByteCode')
-rw-r--r--clang/lib/AST/ByteCode/Interp.h5
-rw-r--r--clang/lib/AST/ByteCode/InterpBuiltin.cpp128
2 files changed, 99 insertions, 34 deletions
diff --git a/clang/lib/AST/ByteCode/Interp.h b/clang/lib/AST/ByteCode/Interp.h
index bb0c458..a9c71c7 100644
--- a/clang/lib/AST/ByteCode/Interp.h
+++ b/clang/lib/AST/ByteCode/Interp.h
@@ -3183,9 +3183,8 @@ inline bool CopyArray(InterpState &S, CodePtr OpPC, uint32_t SrcIndex,
if (!CheckLoad(S, OpPC, SP))
return false;
- const Pointer &DP = DestPtr.atIndex(DestIndex + I);
- DP.deref<T>() = SP.deref<T>();
- DP.initialize();
+ DestPtr.elem<T>(DestIndex + I) = SrcPtr.elem<T>(SrcIndex + I);
+ DestPtr.initializeElement(DestIndex + I);
}
return true;
}
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index 922d679..84c5ecc 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -1633,8 +1633,8 @@ static bool interp__builtin_elementwise_countzeroes(InterpState &S,
const InterpFrame *Frame,
const CallExpr *Call,
unsigned BuiltinID) {
- const bool HasZeroArg = Call->getNumArgs() == 2;
- const bool IsCTTZ = BuiltinID == Builtin::BI__builtin_elementwise_ctzg;
+ bool HasZeroArg = Call->getNumArgs() == 2;
+ bool IsCTTZ = BuiltinID == Builtin::BI__builtin_elementwise_ctzg;
assert(Call->getNumArgs() == 1 || HasZeroArg);
if (Call->getArg(0)->getType()->isIntegerType()) {
PrimType ArgT = *S.getContext().classify(Call->getArg(0)->getType());
@@ -2447,18 +2447,18 @@ interp__builtin_x86_pack(InterpState &S, CodePtr, const CallExpr *E,
const Pointer &Dst = S.Stk.peek<Pointer>();
const ASTContext &ASTCtx = S.getASTContext();
- const unsigned SrcBits = ASTCtx.getIntWidth(VT0->getElementType());
- const unsigned LHSVecLen = VT0->getNumElements();
- const unsigned SrcPerLane = 128 / SrcBits;
- const unsigned Lanes = LHSVecLen * SrcBits / 128;
+ unsigned SrcBits = ASTCtx.getIntWidth(VT0->getElementType());
+ unsigned LHSVecLen = VT0->getNumElements();
+ unsigned SrcPerLane = 128 / SrcBits;
+ unsigned Lanes = LHSVecLen * SrcBits / 128;
PrimType SrcT = *S.getContext().classify(VT0->getElementType());
PrimType DstT = *S.getContext().classify(getElemType(Dst));
- const bool IsUnsigend = getElemType(Dst)->isUnsignedIntegerType();
+ bool IsUnsigend = getElemType(Dst)->isUnsignedIntegerType();
for (unsigned Lane = 0; Lane != Lanes; ++Lane) {
- const unsigned BaseSrc = Lane * SrcPerLane;
- const unsigned BaseDst = Lane * (2 * SrcPerLane);
+ unsigned BaseSrc = Lane * SrcPerLane;
+ unsigned BaseDst = Lane * (2 * SrcPerLane);
for (unsigned I = 0; I != SrcPerLane; ++I) {
INT_TYPE_SWITCH_NO_BOOL(SrcT, {
@@ -2596,9 +2596,9 @@ static bool interp__builtin_elementwise_triop_fp(
FPOptions FPO = Call->getFPFeaturesInEffect(S.Ctx.getLangOpts());
llvm::RoundingMode RM = getRoundingMode(FPO);
- const QualType Arg1Type = Call->getArg(0)->getType();
- const QualType Arg2Type = Call->getArg(1)->getType();
- const QualType Arg3Type = Call->getArg(2)->getType();
+ QualType Arg1Type = Call->getArg(0)->getType();
+ QualType Arg2Type = Call->getArg(1)->getType();
+ QualType Arg3Type = Call->getArg(2)->getType();
// Non-vector floating point types.
if (!Arg1Type->isVectorType()) {
@@ -2621,16 +2621,16 @@ static bool interp__builtin_elementwise_triop_fp(
assert(Arg1Type->isVectorType() && Arg2Type->isVectorType() &&
Arg3Type->isVectorType());
- const VectorType *VecT = Arg1Type->castAs<VectorType>();
- const QualType ElemT = VecT->getElementType();
- unsigned NumElems = VecT->getNumElements();
+ const VectorType *VecTy = Arg1Type->castAs<VectorType>();
+ QualType ElemQT = VecTy->getElementType();
+ unsigned NumElems = VecTy->getNumElements();
- assert(ElemT == Arg2Type->castAs<VectorType>()->getElementType() &&
- ElemT == Arg3Type->castAs<VectorType>()->getElementType());
+ assert(ElemQT == Arg2Type->castAs<VectorType>()->getElementType() &&
+ ElemQT == Arg3Type->castAs<VectorType>()->getElementType());
assert(NumElems == Arg2Type->castAs<VectorType>()->getNumElements() &&
NumElems == Arg3Type->castAs<VectorType>()->getNumElements());
- assert(ElemT->isRealFloatingType());
- (void)ElemT;
+ assert(ElemQT->isRealFloatingType());
+ (void)ElemQT;
const Pointer &VZ = S.Stk.pop<Pointer>();
const Pointer &VY = S.Stk.pop<Pointer>();
@@ -2756,6 +2756,45 @@ static bool interp__builtin_ia32_pshuf(InterpState &S, CodePtr OpPC,
return true;
}
+static bool interp__builtin_ia32_test_op(
+ InterpState &S, CodePtr OpPC, const CallExpr *Call,
+ llvm::function_ref<bool(const APInt &A, const APInt &B)> Fn) {
+ const Pointer &RHS = S.Stk.pop<Pointer>();
+ const Pointer &LHS = S.Stk.pop<Pointer>();
+
+ assert(LHS.getNumElems() == RHS.getNumElems());
+
+ unsigned SourceLen = LHS.getNumElems();
+ QualType ElemQT = getElemType(LHS);
+ OptPrimType ElemPT = S.getContext().classify(ElemQT);
+ unsigned LaneWidth = S.getASTContext().getTypeSize(ElemQT);
+
+ APInt AWide(LaneWidth * SourceLen, 0);
+ APInt BWide(LaneWidth * SourceLen, 0);
+
+ for (unsigned I = 0; I != SourceLen; ++I) {
+ APInt ALane;
+ APInt BLane;
+
+ if (ElemQT->isIntegerType()) { // Get value.
+ INT_TYPE_SWITCH_NO_BOOL(*ElemPT, {
+ ALane = LHS.elem<T>(I).toAPSInt();
+ BLane = RHS.elem<T>(I).toAPSInt();
+ });
+ } else if (ElemQT->isFloatingType()) { // Get only sign bit.
+ using T = PrimConv<PT_Float>::T;
+ ALane = LHS.elem<T>(I).getAPFloat().bitcastToAPInt().isNegative();
+ BLane = RHS.elem<T>(I).getAPFloat().bitcastToAPInt().isNegative();
+ } else { // Must be integer or floating type.
+ return false;
+ }
+ AWide.insertBits(ALane, I * LaneWidth);
+ BWide.insertBits(BLane, I * LaneWidth);
+ }
+ pushInteger(S, Fn(AWide, BWide), Call->getType());
+ return true;
+}
+
static bool interp__builtin_elementwise_triop(
InterpState &S, CodePtr OpPC, const CallExpr *Call,
llvm::function_ref<APInt(const APSInt &, const APSInt &, const APSInt &)>
@@ -2775,7 +2814,7 @@ static bool interp__builtin_elementwise_triop(
}
const auto *VecT = Arg0Type->castAs<VectorType>();
- const PrimType &ElemT = *S.getContext().classify(VecT->getElementType());
+ PrimType ElemT = *S.getContext().classify(VecT->getElementType());
unsigned NumElems = VecT->getNumElements();
bool DestUnsigned = Call->getType()->isUnsignedIntegerOrEnumerationType();
@@ -2847,9 +2886,9 @@ static bool interp__builtin_x86_insert_subvector(InterpState &S, CodePtr OpPC,
unsigned Lane = static_cast<unsigned>(Index % NumLanes);
unsigned InsertPos = Lane * SubElements;
- PrimType ElemPT = BaseVec.getFieldDesc()->getPrimType();
+ PrimType ElemT = BaseVec.getFieldDesc()->getPrimType();
- TYPE_SWITCH(ElemPT, {
+ TYPE_SWITCH(ElemT, {
for (unsigned I = 0; I != BaseElements; ++I)
Dst.elem<T>(I) = BaseVec.elem<T>(I);
for (unsigned I = 0; I != SubElements; ++I)
@@ -2872,12 +2911,12 @@ static bool interp__builtin_ia32_pternlog(InterpState &S, CodePtr OpPC,
const Pointer &Dst = S.Stk.peek<Pointer>();
unsigned DstLen = A.getNumElems();
- const QualType ElemQT = getElemType(A);
- const OptPrimType ElemPT = S.getContext().classify(ElemQT);
+ QualType ElemQT = getElemType(A);
+ OptPrimType ElemT = S.getContext().classify(ElemQT);
unsigned LaneWidth = S.getASTContext().getTypeSize(ElemQT);
bool DstUnsigned = ElemQT->isUnsignedIntegerOrEnumerationType();
- INT_TYPE_SWITCH_NO_BOOL(*ElemPT, {
+ INT_TYPE_SWITCH_NO_BOOL(*ElemT, {
for (unsigned I = 0; I != DstLen; ++I) {
APInt ALane = A.elem<T>(I).toAPSInt();
APInt BLane = B.elem<T>(I).toAPSInt();
@@ -2916,13 +2955,13 @@ static bool interp__builtin_vec_ext(InterpState &S, CodePtr OpPC,
unsigned Index =
static_cast<unsigned>(ImmAPS.getZExtValue() & (NumElems - 1));
- PrimType ElemPT = Vec.getFieldDesc()->getPrimType();
+ PrimType ElemT = Vec.getFieldDesc()->getPrimType();
// FIXME(#161685): Replace float+int split with a numeric-only type switch
- if (ElemPT == PT_Float) {
+ if (ElemT == PT_Float) {
S.Stk.push<Floating>(Vec.elem<Floating>(Index));
return true;
}
- INT_TYPE_SWITCH_NO_BOOL(ElemPT, {
+ INT_TYPE_SWITCH_NO_BOOL(ElemT, {
APSInt V = Vec.elem<T>(Index).toAPSInt();
pushInteger(S, V, Call->getType());
});
@@ -2947,8 +2986,8 @@ static bool interp__builtin_vec_set(InterpState &S, CodePtr OpPC,
unsigned Index =
static_cast<unsigned>(ImmAPS.getZExtValue() & (NumElems - 1));
- PrimType ElemPT = Base.getFieldDesc()->getPrimType();
- INT_TYPE_SWITCH_NO_BOOL(ElemPT, {
+ PrimType ElemT = Base.getFieldDesc()->getPrimType();
+ INT_TYPE_SWITCH_NO_BOOL(ElemT, {
for (unsigned I = 0; I != NumElems; ++I)
Dst.elem<T>(I) = Base.elem<T>(I);
Dst.elem<T>(Index) = static_cast<T>(ValAPS);
@@ -3712,7 +3751,34 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call,
S, OpPC, Call, [](const APSInt &F, const APSInt &T, const APSInt &C) {
return ((APInt)C).isNegative() ? T : F;
});
-
+ case X86::BI__builtin_ia32_ptestz128:
+ case X86::BI__builtin_ia32_ptestz256:
+ case X86::BI__builtin_ia32_vtestzps:
+ case X86::BI__builtin_ia32_vtestzps256:
+ case X86::BI__builtin_ia32_vtestzpd:
+ case X86::BI__builtin_ia32_vtestzpd256:
+ return interp__builtin_ia32_test_op(
+ S, OpPC, Call,
+ [](const APInt &A, const APInt &B) { return (A & B) == 0; });
+ case X86::BI__builtin_ia32_ptestc128:
+ case X86::BI__builtin_ia32_ptestc256:
+ case X86::BI__builtin_ia32_vtestcps:
+ case X86::BI__builtin_ia32_vtestcps256:
+ case X86::BI__builtin_ia32_vtestcpd:
+ case X86::BI__builtin_ia32_vtestcpd256:
+ return interp__builtin_ia32_test_op(
+ S, OpPC, Call,
+ [](const APInt &A, const APInt &B) { return (~A & B) == 0; });
+ case X86::BI__builtin_ia32_ptestnzc128:
+ case X86::BI__builtin_ia32_ptestnzc256:
+ case X86::BI__builtin_ia32_vtestnzcps:
+ case X86::BI__builtin_ia32_vtestnzcps256:
+ case X86::BI__builtin_ia32_vtestnzcpd:
+ case X86::BI__builtin_ia32_vtestnzcpd256:
+ return interp__builtin_ia32_test_op(
+ S, OpPC, Call, [](const APInt &A, const APInt &B) {
+ return ((A & B) != 0) && ((~A & B) != 0);
+ });
case X86::BI__builtin_ia32_selectb_128:
case X86::BI__builtin_ia32_selectb_256:
case X86::BI__builtin_ia32_selectb_512:
generated by cgit v1.1 at 2025-10-13 17:32:24 +0000