1 files changed, 185 insertions, 3 deletions

diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
index 0cb4910..ff83c52 100644
--- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp
+++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -8,9 +8,10 @@
 #include "../ExprConstShared.h"
 #include "Boolean.h"
 #include "EvalEmitter.h"
-#include "Interp.h"
 #include "InterpBuiltinBitCast.h"
+#include "InterpHelpers.h"
 #include "PrimType.h"
+#include "Program.h"
 #include "clang/AST/OSLog.h"
 #include "clang/AST/RecordLayout.h"
 #include "clang/Basic/Builtins.h"
@@ -2041,10 +2042,16 @@ static bool interp__builtin_memchr(InterpState &S, CodePtr OpPC,
   }
 
   if (ID == Builtin::BIstrchr || ID == Builtin::BI__builtin_strchr) {
+    int64_t DesiredTrunc;
+    if (S.getASTContext().CharTy->isSignedIntegerType())
+      DesiredTrunc =
+          Desired.trunc(S.getASTContext().getCharWidth()).getSExtValue();
+    else
+      DesiredTrunc =
+          Desired.trunc(S.getASTContext().getCharWidth()).getZExtValue();
     // strchr compares directly to the passed integer, and therefore
     // always fails if given an int that is not a char.
-    if (Desired !=
-        Desired.trunc(S.getASTContext().getCharWidth()).getSExtValue()) {
+    if (Desired != DesiredTrunc) {
       S.Stk.push<Pointer>();
       return true;
     }
@@ -2899,6 +2906,35 @@ static bool interp__builtin_ia32_test_op(
   return true;
 }
 
+static bool interp__builtin_ia32_movmsk_op(InterpState &S, CodePtr OpPC,
+                                           const CallExpr *Call) {
+  assert(Call->getNumArgs() == 1);
+
+  const Pointer &Source = S.Stk.pop<Pointer>();
+
+  unsigned SourceLen = Source.getNumElems();
+  QualType ElemQT = getElemType(Source);
+  OptPrimType ElemT = S.getContext().classify(ElemQT);
+  unsigned ResultLen =
+      S.getASTContext().getTypeSize(Call->getType()); // Always 32-bit integer.
+  APInt Result(ResultLen, 0);
+
+  for (unsigned I = 0; I != SourceLen; ++I) {
+    APInt Elem;
+    if (ElemQT->isIntegerType()) {
+      INT_TYPE_SWITCH_NO_BOOL(*ElemT, { Elem = Source.elem<T>(I).toAPSInt(); });
+    } else if (ElemQT->isRealFloatingType()) {
+      using T = PrimConv<PT_Float>::T;
+      Elem = Source.elem<T>(I).getAPFloat().bitcastToAPInt();
+    } else {
+      return false;
+    }
+    Result.setBitVal(I, Elem.isNegative());
+  }
+  pushInteger(S, Result, 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 &)>
@@ -2962,6 +2998,82 @@ static bool interp__builtin_elementwise_triop(
   return true;
 }
 
+static bool interp__builtin_x86_extract_vector(InterpState &S, CodePtr OpPC,
+                                               const CallExpr *Call,
+                                               unsigned ID) {
+  assert(Call->getNumArgs() == 2);
+
+  APSInt ImmAPS = popToAPSInt(S, Call->getArg(1));
+  uint64_t Index = ImmAPS.getZExtValue();
+
+  const Pointer &Src = S.Stk.pop<Pointer>();
+  if (!Src.getFieldDesc()->isPrimitiveArray())
+    return false;
+
+  const Pointer &Dst = S.Stk.peek<Pointer>();
+  if (!Dst.getFieldDesc()->isPrimitiveArray())
+    return false;
+
+  unsigned SrcElems = Src.getNumElems();
+  unsigned DstElems = Dst.getNumElems();
+
+  unsigned NumLanes = SrcElems / DstElems;
+  unsigned Lane = static_cast<unsigned>(Index % NumLanes);
+  unsigned ExtractPos = Lane * DstElems;
+
+  PrimType ElemT = Src.getFieldDesc()->getPrimType();
+
+  TYPE_SWITCH(ElemT, {
+    for (unsigned I = 0; I != DstElems; ++I) {
+      Dst.elem<T>(I) = Src.elem<T>(ExtractPos + I);
+    }
+  });
+
+  Dst.initializeAllElements();
+  return true;
+}
+
+static bool interp__builtin_x86_extract_vector_masked(InterpState &S,
+                                                      CodePtr OpPC,
+                                                      const CallExpr *Call,
+                                                      unsigned ID) {
+  assert(Call->getNumArgs() == 4);
+
+  APSInt MaskAPS = popToAPSInt(S, Call->getArg(3));
+  const Pointer &Merge = S.Stk.pop<Pointer>();
+  APSInt ImmAPS = popToAPSInt(S, Call->getArg(1));
+  const Pointer &Src = S.Stk.pop<Pointer>();
+
+  if (!Src.getFieldDesc()->isPrimitiveArray() ||
+      !Merge.getFieldDesc()->isPrimitiveArray())
+    return false;
+
+  const Pointer &Dst = S.Stk.peek<Pointer>();
+  if (!Dst.getFieldDesc()->isPrimitiveArray())
+    return false;
+
+  unsigned SrcElems = Src.getNumElems();
+  unsigned DstElems = Dst.getNumElems();
+
+  unsigned NumLanes = SrcElems / DstElems;
+  unsigned Lane = static_cast<unsigned>(ImmAPS.getZExtValue() % NumLanes);
+  unsigned Base = Lane * DstElems;
+
+  PrimType ElemT = Src.getFieldDesc()->getPrimType();
+
+  TYPE_SWITCH(ElemT, {
+    for (unsigned I = 0; I != DstElems; ++I) {
+      if (MaskAPS[I])
+        Dst.elem<T>(I) = Src.elem<T>(Base + I);
+      else
+        Dst.elem<T>(I) = Merge.elem<T>(I);
+    }
+  });
+
+  Dst.initializeAllElements();
+  return true;
+}
+
 static bool interp__builtin_x86_insert_subvector(InterpState &S, CodePtr OpPC,
                                                  const CallExpr *Call,
                                                  unsigned ID) {
@@ -3003,6 +3115,45 @@ static bool interp__builtin_x86_insert_subvector(InterpState &S, CodePtr OpPC,
   return true;
 }
 
+static bool interp__builtin_ia32_phminposuw(InterpState &S, CodePtr OpPC,
+                                            const CallExpr *Call) {
+  assert(Call->getNumArgs() == 1);
+
+  const Pointer &Source = S.Stk.pop<Pointer>();
+  const Pointer &Dest = S.Stk.peek<Pointer>();
+
+  unsigned SourceLen = Source.getNumElems();
+  QualType ElemQT = getElemType(Source);
+  OptPrimType ElemT = S.getContext().classify(ElemQT);
+  unsigned ElemBitWidth = S.getASTContext().getTypeSize(ElemQT);
+
+  bool DestUnsigned = Call->getCallReturnType(S.getASTContext())
+                          ->castAs<VectorType>()
+                          ->getElementType()
+                          ->isUnsignedIntegerOrEnumerationType();
+
+  INT_TYPE_SWITCH_NO_BOOL(*ElemT, {
+    APSInt MinIndex(ElemBitWidth, DestUnsigned);
+    APSInt MinVal = Source.elem<T>(0).toAPSInt();
+
+    for (unsigned I = 1; I != SourceLen; ++I) {
+      APSInt Val = Source.elem<T>(I).toAPSInt();
+      if (MinVal.ugt(Val)) {
+        MinVal = Val;
+        MinIndex = I;
+      }
+    }
+
+    Dest.elem<T>(0) = static_cast<T>(MinVal);
+    Dest.elem<T>(1) = static_cast<T>(MinIndex);
+    for (unsigned I = 2; I != SourceLen; ++I) {
+      Dest.elem<T>(I) = static_cast<T>(APSInt(ElemBitWidth, DestUnsigned));
+    }
+  });
+  Dest.initializeAllElements();
+  return true;
+}
+
 static bool interp__builtin_ia32_pternlog(InterpState &S, CodePtr OpPC,
                                           const CallExpr *Call, bool MaskZ) {
   assert(Call->getNumArgs() == 5);
@@ -3620,6 +3771,25 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call,
         S, OpPC, Call, [](const APSInt &LHS, const APSInt &RHS) {
           return LHS.isSigned() ? LHS.ssub_sat(RHS) : LHS.usub_sat(RHS);
         });
+  case X86::BI__builtin_ia32_extract128i256:
+  case X86::BI__builtin_ia32_vextractf128_pd256:
+  case X86::BI__builtin_ia32_vextractf128_ps256:
+  case X86::BI__builtin_ia32_vextractf128_si256:
+    return interp__builtin_x86_extract_vector(S, OpPC, Call, BuiltinID);
+
+  case X86::BI__builtin_ia32_extractf32x4_256_mask:
+  case X86::BI__builtin_ia32_extractf32x4_mask:
+  case X86::BI__builtin_ia32_extractf32x8_mask:
+  case X86::BI__builtin_ia32_extractf64x2_256_mask:
+  case X86::BI__builtin_ia32_extractf64x2_512_mask:
+  case X86::BI__builtin_ia32_extractf64x4_mask:
+  case X86::BI__builtin_ia32_extracti32x4_256_mask:
+  case X86::BI__builtin_ia32_extracti32x4_mask:
+  case X86::BI__builtin_ia32_extracti32x8_mask:
+  case X86::BI__builtin_ia32_extracti64x2_256_mask:
+  case X86::BI__builtin_ia32_extracti64x2_512_mask:
+  case X86::BI__builtin_ia32_extracti64x4_mask:
+    return interp__builtin_x86_extract_vector_masked(S, OpPC, Call, BuiltinID);
 
   case clang::X86::BI__builtin_ia32_pmulhrsw128:
   case clang::X86::BI__builtin_ia32_pmulhrsw256:
@@ -3630,6 +3800,15 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call,
               .extractBits(16, 1);
         });
 
+  case clang::X86::BI__builtin_ia32_movmskps:
+  case clang::X86::BI__builtin_ia32_movmskpd:
+  case clang::X86::BI__builtin_ia32_pmovmskb128:
+  case clang::X86::BI__builtin_ia32_pmovmskb256:
+  case clang::X86::BI__builtin_ia32_movmskps256:
+  case clang::X86::BI__builtin_ia32_movmskpd256: {
+    return interp__builtin_ia32_movmsk_op(S, OpPC, Call);
+  }
+
   case clang::X86::BI__builtin_ia32_pavgb128:
   case clang::X86::BI__builtin_ia32_pavgw128:
   case clang::X86::BI__builtin_ia32_pavgb256:
@@ -4087,6 +4266,9 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call,
         S, OpPC, Call,
         [](const APSInt &LHS, const APSInt &RHS) { return LHS + RHS; });
 
+  case X86::BI__builtin_ia32_phminposuw128:
+    return interp__builtin_ia32_phminposuw(S, OpPC, Call);
+
   case X86::BI__builtin_ia32_pternlogd128_mask:
   case X86::BI__builtin_ia32_pternlogd256_mask:
   case X86::BI__builtin_ia32_pternlogd512_mask: