[Headers][X86] Allow SLLDQ/SRLDQ byte shift intrinsics to be used in constexpr (llvm#164166)

TianYe717 · Lukacma · commit 76a1b3aea423 · 2025-10-29T11:16:31.000Z
Support constexpr usage for SLLDQ/SRLDQ byte shift intrinsics This draft PR adds support for using the following SRLDQ intrinsics in constant expressions: - _mm_srli_si128 - _mm256_srli_si256 - _mm_slli_si128 - _mm256_slli_si256 Relevant tests are included. Fixes llvm#156494
diff --git a/clang/include/clang/Basic/BuiltinsX86.td b/clang/include/clang/Basic/BuiltinsX86.td
@@ -280,8 +280,6 @@ let Features = "sse2", Attributes = [NoThrow, Const, RequiredVectorWidth<128>] i
   def psllw128 : X86Builtin<"_Vector<8, short>(_Vector<8, short>, _Vector<8, short>)">;
   def pslld128 : X86Builtin<"_Vector<4, int>(_Vector<4, int>, _Vector<4, int>)">;
   def psllq128 : X86Builtin<"_Vector<2, long long int>(_Vector<2, long long int>, _Vector<2, long long int>)">;
-  def pslldqi128_byteshift : X86Builtin<"_Vector<16, char>(_Vector<16, char>, _Constant int)">;
-  def psrldqi128_byteshift : X86Builtin<"_Vector<16, char>(_Vector<16, char>, _Constant int)">;
 }
 
 let Features = "sse2",
@@ -300,6 +298,9 @@ let Features = "sse2",
 
   def psrawi128 : X86Builtin<"_Vector<8, short>(_Vector<8, short>, int)">;
   def psradi128 : X86Builtin<"_Vector<4, int>(_Vector<4, int>, int)">;
+
+  def pslldqi128_byteshift : X86Builtin<"_Vector<16, char>(_Vector<16, char>, _Constant int)">;
+  def psrldqi128_byteshift : X86Builtin<"_Vector<16, char>(_Vector<16, char>, _Constant int)">;
 }
 
 let Features = "sse3", Attributes = [NoThrow] in {
@@ -607,12 +608,10 @@ let Features = "avx2", Attributes = [NoThrow, Const, RequiredVectorWidth<256>] i
       : X86Builtin<
             "_Vector<4, long long int>(_Vector<32, char>, _Vector<32, char>)">;
   def psllw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Vector<8, short>)">;
-  def pslldqi256_byteshift : X86Builtin<"_Vector<32, char>(_Vector<32, char>, _Constant int)">;
   def pslld256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>, _Vector<4, int>)">;
   def psllq256 : X86Builtin<"_Vector<4, long long int>(_Vector<4, long long int>, _Vector<2, long long int>)">;
   def psraw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Vector<8, short>)">;
   def psrad256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>, _Vector<4, int>)">;
-  def psrldqi256_byteshift : X86Builtin<"_Vector<32, char>(_Vector<32, char>, _Constant int)">;
   def psrlw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Vector<8, short>)">;
   def psrld256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>, _Vector<4, int>)">;
   def psrlq256 : X86Builtin<"_Vector<4, long long int>(_Vector<4, long long int>, _Vector<2, long long int>)">;
@@ -646,10 +645,12 @@ let Features = "avx2", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWi
   def psllwi256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, int)">;
   def pslldi256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>, int)">;
   def psllqi256 : X86Builtin<"_Vector<4, long long int>(_Vector<4, long long int>, int)">;
+  def pslldqi256_byteshift : X86Builtin<"_Vector<32, char>(_Vector<32, char>, _Constant int)">;
 
   def psrlwi256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, int)">;
   def psrldi256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>, int)">;
   def psrlqi256 : X86Builtin<"_Vector<4, long long int>(_Vector<4, long long int>, int)">;
+  def psrldqi256_byteshift : X86Builtin<"_Vector<32, char>(_Vector<32, char>, _Constant int)">;
 
   def psrawi256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, int)">;
   def psradi256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>, int)">;
@@ -2090,6 +2091,9 @@ let Features = "avx512bw", Attributes = [NoThrow, Const, RequiredVectorWidth<512
       : X86Builtin<"_Vector<32, short>(_Vector<32, short>, _Vector<8, short>)">;
   def psrlw512
       : X86Builtin<"_Vector<32, short>(_Vector<32, short>, _Vector<8, short>)">;
+}
+
+let Features = "avx512bw", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<512>] in {
   def pslldqi512_byteshift : X86Builtin<"_Vector<64, char>(_Vector<64, char>, _Constant int)">;
   def psrldqi512_byteshift : X86Builtin<"_Vector<64, char>(_Vector<64, char>, _Constant int)">;
 }
diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp b/clang/lib/AST/ByteCode/InterpBuiltin.cpp
@@ -3320,6 +3320,38 @@ static bool interp__builtin_ia32_vpconflict(InterpState &S, CodePtr OpPC,
   return true;
 }
 
+static bool interp__builtin_x86_byteshift(
+    InterpState &S, CodePtr OpPC, const CallExpr *Call, unsigned ID,
+    llvm::function_ref<APInt(const Pointer &, unsigned Lane, unsigned I,
+                             unsigned Shift)>
+        Fn) {
+  assert(Call->getNumArgs() == 2);
+
+  APSInt ImmAPS = popToAPSInt(S, Call->getArg(1));
+  uint64_t Shift = ImmAPS.getZExtValue() & 0xff;
+
+  const Pointer &Src = S.Stk.pop<Pointer>();
+  if (!Src.getFieldDesc()->isPrimitiveArray())
+    return false;
+
+  unsigned NumElems = Src.getNumElems();
+  const Pointer &Dst = S.Stk.peek<Pointer>();
+  PrimType ElemT = Src.getFieldDesc()->getPrimType();
+
+  for (unsigned Lane = 0; Lane != NumElems; Lane += 16) {
+    for (unsigned I = 0; I != 16; ++I) {
+      unsigned Base = Lane + I;
+      APSInt Result = APSInt(Fn(Src, Lane, I, Shift));
+      INT_TYPE_SWITCH_NO_BOOL(ElemT,
+                              { Dst.elem<T>(Base) = static_cast<T>(Result); });
+    }
+  }
+
+  Dst.initializeAllElements();
+
+  return true;
+}
+
 bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call,
                       uint32_t BuiltinID) {
   if (!S.getASTContext().BuiltinInfo.isConstantEvaluated(BuiltinID))
@@ -4390,6 +4422,39 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call,
   case X86::BI__builtin_ia32_vec_set_v4di:
     return interp__builtin_vec_set(S, OpPC, Call, BuiltinID);
 
+  case X86::BI__builtin_ia32_pslldqi128_byteshift:
+  case X86::BI__builtin_ia32_pslldqi256_byteshift:
+  case X86::BI__builtin_ia32_pslldqi512_byteshift:
+    // These SLLDQ intrinsics always operate on byte elements (8 bits).
+    // The lane width is hardcoded to 16 to match the SIMD register size,
+    // but the algorithm processes one byte per iteration,
+    // so APInt(8, ...) is correct and intentional.
+    return interp__builtin_x86_byteshift(
+        S, OpPC, Call, BuiltinID,
+        [](const Pointer &Src, unsigned Lane, unsigned I, unsigned Shift) {
+          if (I < Shift) {
+            return APInt(8, 0);
+          }
+          return APInt(8, Src.elem<uint8_t>(Lane + I - Shift));
+        });
+
+  case X86::BI__builtin_ia32_psrldqi128_byteshift:
+  case X86::BI__builtin_ia32_psrldqi256_byteshift:
+  case X86::BI__builtin_ia32_psrldqi512_byteshift:
+    // These SRLDQ intrinsics always operate on byte elements (8 bits).
+    // The lane width is hardcoded to 16 to match the SIMD register size,
+    // but the algorithm processes one byte per iteration,
+    // so APInt(8, ...) is correct and intentional.
+    return interp__builtin_x86_byteshift(
+        S, OpPC, Call, BuiltinID,
+        [](const Pointer &Src, unsigned Lane, unsigned I, unsigned Shift) {
+          if (I + Shift < 16) {
+            return APInt(8, Src.elem<uint8_t>(Lane + I + Shift));
+          }
+
+          return APInt(8, 0);
+        });
+
   default:
     S.FFDiag(S.Current->getLocation(OpPC),
              diag::note_invalid_subexpr_in_const_expr)
diff --git a/clang/lib/AST/ExprConstant.cpp b/clang/lib/AST/ExprConstant.cpp
@@ -12906,6 +12906,66 @@ bool VectorExprEvaluator::VisitCallExpr(const CallExpr *E) {
 
     return Success(APValue(Elems.data(), NumElems), E);
   }
+
+  case X86::BI__builtin_ia32_pslldqi128_byteshift:
+  case X86::BI__builtin_ia32_pslldqi256_byteshift:
+  case X86::BI__builtin_ia32_pslldqi512_byteshift: {
+    assert(E->getNumArgs() == 2);
+
+    APValue Src;
+    APSInt Imm;
+    if (!EvaluateAsRValue(Info, E->getArg(0), Src) ||
+        !EvaluateInteger(E->getArg(1), Imm, Info))
+      return false;
+
+    unsigned VecLen = Src.getVectorLength();
+    unsigned Shift = Imm.getZExtValue() & 0xff;
+
+    SmallVector<APValue> ResultElements;
+    for (unsigned Lane = 0; Lane != VecLen; Lane += 16) {
+      for (unsigned I = 0; I != 16; ++I) {
+        if (I < Shift) {
+          APSInt Zero(8, /*isUnsigned=*/true);
+          Zero = 0;
+          ResultElements.push_back(APValue(Zero));
+        } else {
+          ResultElements.push_back(Src.getVectorElt(Lane + I - Shift));
+        }
+      }
+    }
+
+    return Success(APValue(ResultElements.data(), ResultElements.size()), E);
+  }
+
+  case X86::BI__builtin_ia32_psrldqi128_byteshift:
+  case X86::BI__builtin_ia32_psrldqi256_byteshift:
+  case X86::BI__builtin_ia32_psrldqi512_byteshift: {
+    assert(E->getNumArgs() == 2);
+
+    APValue Src;
+    APSInt Imm;
+    if (!EvaluateAsRValue(Info, E->getArg(0), Src) ||
+        !EvaluateInteger(E->getArg(1), Imm, Info))
+      return false;
+
+    unsigned VecLen = Src.getVectorLength();
+    unsigned Shift = Imm.getZExtValue() & 0xff;
+
+    SmallVector<APValue> ResultElements;
+    for (unsigned Lane = 0; Lane != VecLen; Lane += 16) {
+      for (unsigned I = 0; I != 16; ++I) {
+        if (I + Shift < 16) {
+          ResultElements.push_back(Src.getVectorElt(Lane + I + Shift));
+        } else {
+          APSInt Zero(8, /*isUnsigned=*/true);
+          Zero = 0;
+          ResultElements.push_back(APValue(Zero));
+        }
+      }
+    }
+
+    return Success(APValue(ResultElements.data(), ResultElements.size()), E);
+  }
   }
 }
 
diff --git a/clang/test/CodeGen/X86/avx2-builtins.c b/clang/test/CodeGen/X86/avx2-builtins.c
@@ -1236,6 +1236,8 @@ __m256i test_mm256_slli_si256(__m256i a) {
   // CHECK: shufflevector <32 x i8> zeroinitializer, <32 x i8> %{{.*}}, <32 x i32> <i32 13, i32 14, i32 15, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 29, i32 30, i32 31, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60>
   return _mm256_slli_si256(a, 3);
 }
+TEST_CONSTEXPR(match_v32qi(_mm256_slli_si256(((__m256i)(__v32qi){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32}), 3), 0, 0, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 0, 0, 0, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29));
+TEST_CONSTEXPR(match_v32qi(_mm256_slli_si256(((__m256i)(__v32qi){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32}), 16), 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0));
 
 __m128i test_mm_sllv_epi32(__m128i a, __m128i b) {
   // CHECK-LABEL: test_mm_sllv_epi32
@@ -1379,6 +1381,8 @@ __m256i test_mm256_srli_si256(__m256i a) {
   // CHECK: shufflevector <32 x i8> %{{.*}}, <32 x i8> zeroinitializer, <32 x i32> <i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 32, i32 33, i32 34, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 48, i32 49, i32 50>
   return _mm256_srli_si256(a, 3);
 }
+TEST_CONSTEXPR(match_v32qi(_mm256_srli_si256(((__m256i)(__v32qi){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32}), 3), 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 0, 0, 0, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 0, 0, 0));
+TEST_CONSTEXPR(match_v32qi(_mm256_srli_si256(((__m256i)(__v32qi){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32}), 16), 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0));
 
 __m128i test_mm_srlv_epi32(__m128i a, __m128i b) {
   // CHECK-LABEL: test_mm_srlv_epi32
diff --git a/clang/test/CodeGen/X86/avx512bw-builtins.c b/clang/test/CodeGen/X86/avx512bw-builtins.c
@@ -2045,6 +2045,8 @@ __m512i test_mm512_bslli_epi128(__m512i __A) {
   // CHECK: shufflevector <64 x i8> zeroinitializer, <64 x i8> %{{.*}}, <64 x i32> <i32 11, i32 12, i32 13, i32 14, i32 15, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 27, i32 28, i32 29, i32 30, i32 31, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 43, i32 44, i32 45, i32 46, i32 47, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 59, i32 60, i32 61, i32 62, i32 63, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122>
   return _mm512_bslli_epi128(__A, 5);
 }
+TEST_CONSTEXPR(match_v64qi(_mm512_bslli_epi128(((__m512i)(__v64qi){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64}), 5), 0, 0, 0, 0, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 0, 0, 0, 0, 0, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 0, 0, 0, 0, 0, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 0, 0, 0, 0, 0, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59));
+TEST_CONSTEXPR(match_v64qi(_mm512_bslli_epi128(((__m512i)(__v64qi){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64}), 16), 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0));
 
 __m512i test_mm512_srlv_epi16(__m512i __A, __m512i __B) {
   // CHECK-LABEL: test_mm512_srlv_epi16
@@ -2223,6 +2225,9 @@ __m512i test_mm512_bsrli_epi128(__m512i __A) {
   // CHECK: shufflevector <64 x i8> %{{.*}}, <64 x i8> zeroinitializer, <64 x i32> <i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 64, i32 65, i32 66, i32 67, i32 68, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 80, i32 81, i32 82, i32 83, i32 84, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 96, i32 97, i32 98, i32 99, i32 100, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63, i32 112, i32 113, i32 114, i32 115, i32 116>
   return _mm512_bsrli_epi128(__A, 5);
 }
+TEST_CONSTEXPR(match_v64qi(_mm512_bsrli_epi128(((__m512i)(__v64qi){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64}), 5), 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 0, 0, 0, 0, 0, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 0, 0, 0, 0, 0, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 0, 0, 0, 0, 0, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 0, 0, 0, 0, 0));
+TEST_CONSTEXPR(match_v64qi(_mm512_bsrli_epi128(((__m512i)(__v64qi){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64}), 16), 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0));
+
 __m512i test_mm512_mask_mov_epi16(__m512i __W, __mmask32 __U, __m512i __A) {
   // CHECK-LABEL: test_mm512_mask_mov_epi16
   // CHECK: select <32 x i1> %{{.*}}, <32 x i16> %{{.*}}, <32 x i16> %{{.*}}
diff --git a/clang/test/CodeGen/X86/sse2-builtins.c b/clang/test/CodeGen/X86/sse2-builtins.c
@@ -1418,6 +1418,8 @@ __m128i test_mm_slli_si128(__m128i A) {
   // CHECK: shufflevector <16 x i8> zeroinitializer, <16 x i8> %{{.*}}, <16 x i32> <i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26>
   return _mm_slli_si128(A, 5);
 }
+TEST_CONSTEXPR(match_v16qi(_mm_slli_si128(((__m128i)(__v16qi){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16}), 5), 0, 0, 0, 0, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11));
+TEST_CONSTEXPR(match_v16qi(_mm_slli_si128(((__m128i)(__v16qi){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16}), 16), 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0));
 
 __m128i test_mm_slli_si128_2(__m128i A) {
   // CHECK-LABEL: test_mm_slli_si128_2
@@ -1570,6 +1572,8 @@ __m128i test_mm_srli_si128(__m128i A) {
   // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> zeroinitializer, <16 x i32> <i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20>
   return _mm_srli_si128(A, 5);
 }
+TEST_CONSTEXPR(match_v16qi(_mm_srli_si128(((__m128i)(__v16qi){0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15}), 5), 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 0, 0, 0, 0, 0));
+TEST_CONSTEXPR(match_v16qi(_mm_srli_si128(((__m128i)(__v16qi){0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15}), 16), 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0));
 
 __m128i test_mm_srli_si128_2(__m128i A) {
   // CHECK-LABEL: test_mm_srli_si128_2
