Fix codegen of consteval functions returning an empty class, and related issues (llvm#93115)

Fix codegen of consteval functions returning an empty class, and related
issues

If a class is empty, don't store it to memory: the store might overwrite
useful data. Similarly, if a class has tail padding that might overlap
other fields, don't store the tail padding to memory.

The problem here turned out a bit more general than I initially thought:
basically all uses of EmitAggregateStore were broken. Call lowering had
a method that did mostly the right thing, though: CreateCoercedStore.
Adapt CreateCoercedStore so it always does the conservatively right
thing, and use it for both calls and ConstantExpr.

Also, along the way, fix the "overlap" bit in AggValueSlot: the bit was
set incorrectly for empty classes in some cases.

Fixes llvm#93040.

(cherry picked from commit 1762e01)

Author: efriedma-quic

clang/lib/CodeGen/CGCall.cpp

@@ -1336,75 +1336,50 @@ static llvm::Value *CreateCoercedLoad(Address Src, llvm::Type *Ty,
   return CGF.Builder.CreateLoad(Tmp);
 }
 
-// Function to store a first-class aggregate into memory.  We prefer to
-// store the elements rather than the aggregate to be more friendly to
-// fast-isel.
-// FIXME: Do we need to recurse here?
-void CodeGenFunction::EmitAggregateStore(llvm::Value *Val, Address Dest,
-                                         bool DestIsVolatile) {
-  // Prefer scalar stores to first-class aggregate stores.
-  if (llvm::StructType *STy = dyn_cast<llvm::StructType>(Val->getType())) {
-    for (unsigned i = 0, e = STy->getNumElements(); i != e; ++i) {
-      Address EltPtr = Builder.CreateStructGEP(Dest, i);
-      llvm::Value *Elt = Builder.CreateExtractValue(Val, i);
-      Builder.CreateStore(Elt, EltPtr, DestIsVolatile);
-    }
-  } else {
-    Builder.CreateStore(Val, Dest, DestIsVolatile);
-  }
-}
-
-/// CreateCoercedStore - Create a store to \arg DstPtr from \arg Src,
-/// where the source and destination may have different types.  The
-/// destination is known to be aligned to \arg DstAlign bytes.
-///
-/// This safely handles the case when the src type is larger than the
-/// destination type; the upper bits of the src will be lost.
-static void CreateCoercedStore(llvm::Value *Src,
-                               Address Dst,
-                               bool DstIsVolatile,
-                               CodeGenFunction &CGF) {
-  llvm::Type *SrcTy = Src->getType();
-  llvm::Type *DstTy = Dst.getElementType();
-  if (SrcTy == DstTy) {
-    CGF.Builder.CreateStore(Src, Dst, DstIsVolatile);
-    return;
-  }
-
-  llvm::TypeSize SrcSize = CGF.CGM.getDataLayout().getTypeAllocSize(SrcTy);
-
-  if (llvm::StructType *DstSTy = dyn_cast<llvm::StructType>(DstTy)) {
-    Dst = EnterStructPointerForCoercedAccess(Dst, DstSTy,
-                                             SrcSize.getFixedValue(), CGF);
-    DstTy = Dst.getElementType();
-  }
-
-  llvm::PointerType *SrcPtrTy = llvm::dyn_cast<llvm::PointerType>(SrcTy);
-  llvm::PointerType *DstPtrTy = llvm::dyn_cast<llvm::PointerType>(DstTy);
-  if (SrcPtrTy && DstPtrTy &&
-      SrcPtrTy->getAddressSpace() != DstPtrTy->getAddressSpace()) {
-    Src = CGF.Builder.CreateAddrSpaceCast(Src, DstTy);
-    CGF.Builder.CreateStore(Src, Dst, DstIsVolatile);
+void CodeGenFunction::CreateCoercedStore(llvm::Value *Src, Address Dst,
+                                         llvm::TypeSize DstSize,
+                                         bool DstIsVolatile) {
+  if (!DstSize)
     return;
-  }
 
-  // If the source and destination are integer or pointer types, just do an
-  // extension or truncation to the desired type.
-  if ((isa<llvm::IntegerType>(SrcTy) || isa<llvm::PointerType>(SrcTy)) &&
-      (isa<llvm::IntegerType>(DstTy) || isa<llvm::PointerType>(DstTy))) {
-    Src = CoerceIntOrPtrToIntOrPtr(Src, DstTy, CGF);
-    CGF.Builder.CreateStore(Src, Dst, DstIsVolatile);
-    return;
+  llvm::Type *SrcTy = Src->getType();
+  llvm::TypeSize SrcSize = CGM.getDataLayout().getTypeAllocSize(SrcTy);
+
+  // GEP into structs to try to make types match.
+  // FIXME: This isn't really that useful with opaque types, but it impacts a
+  // lot of regression tests.
+  if (SrcTy != Dst.getElementType()) {
+    if (llvm::StructType *DstSTy =
+            dyn_cast<llvm::StructType>(Dst.getElementType())) {
+      assert(!SrcSize.isScalable());
+      Dst = EnterStructPointerForCoercedAccess(Dst, DstSTy,
+                                               SrcSize.getFixedValue(), *this);
+    }
   }
 
-  llvm::TypeSize DstSize = CGF.CGM.getDataLayout().getTypeAllocSize(DstTy);
-
-  // If store is legal, just bitcast the src pointer.
-  if (isa<llvm::ScalableVectorType>(SrcTy) ||
-      isa<llvm::ScalableVectorType>(DstTy) ||
-      SrcSize.getFixedValue() <= DstSize.getFixedValue()) {
-    Dst = Dst.withElementType(SrcTy);
-    CGF.EmitAggregateStore(Src, Dst, DstIsVolatile);
+  if (SrcSize.isScalable() || SrcSize <= DstSize) {
+    if (SrcTy->isIntegerTy() && Dst.getElementType()->isPointerTy() &&
+        SrcSize == CGM.getDataLayout().getTypeAllocSize(Dst.getElementType())) {
+      // If the value is supposed to be a pointer, convert it before storing it.
+      Src = CoerceIntOrPtrToIntOrPtr(Src, Dst.getElementType(), *this);
+      Builder.CreateStore(Src, Dst, DstIsVolatile);
+    } else if (llvm::StructType *STy =
+                   dyn_cast<llvm::StructType>(Src->getType())) {
+      // Prefer scalar stores to first-class aggregate stores.
+      Dst = Dst.withElementType(SrcTy);
+      for (unsigned i = 0, e = STy->getNumElements(); i != e; ++i) {
+        Address EltPtr = Builder.CreateStructGEP(Dst, i);
+        llvm::Value *Elt = Builder.CreateExtractValue(Src, i);
+        Builder.CreateStore(Elt, EltPtr, DstIsVolatile);
+      }
+    } else {
+      Builder.CreateStore(Src, Dst.withElementType(SrcTy), DstIsVolatile);
+    }
+  } else if (SrcTy->isIntegerTy()) {
+    // If the source is a simple integer, coerce it directly.
+    llvm::Type *DstIntTy = Builder.getIntNTy(DstSize.getFixedValue() * 8);
+    Src = CoerceIntOrPtrToIntOrPtr(Src, DstIntTy, *this);
+    Builder.CreateStore(Src, Dst.withElementType(DstIntTy), DstIsVolatile);
   } else {
     // Otherwise do coercion through memory. This is stupid, but
     // simple.
@@ -1416,12 +1391,12 @@ static void CreateCoercedStore(llvm::Value *Src,
     // FIXME: Assert that we aren't truncating non-padding bits when have access
     // to that information.
     RawAddress Tmp =
-        CreateTempAllocaForCoercion(CGF, SrcTy, Dst.getAlignment());
-    CGF.Builder.CreateStore(Src, Tmp);
-    CGF.Builder.CreateMemCpy(
-        Dst.emitRawPointer(CGF), Dst.getAlignment().getAsAlign(),
-        Tmp.getPointer(), Tmp.getAlignment().getAsAlign(),
-        llvm::ConstantInt::get(CGF.IntPtrTy, DstSize.getFixedValue()));
+        CreateTempAllocaForCoercion(*this, SrcTy, Dst.getAlignment());
+    Builder.CreateStore(Src, Tmp);
+    Builder.CreateMemCpy(Dst.emitRawPointer(*this),
+                         Dst.getAlignment().getAsAlign(), Tmp.getPointer(),
+                         Tmp.getAlignment().getAsAlign(),
+                         Builder.CreateTypeSize(IntPtrTy, DstSize));
   }
 }
 
@@ -3309,7 +3284,12 @@ void CodeGenFunction::EmitFunctionProlog(const CGFunctionInfo &FI,
         assert(NumIRArgs == 1);
         auto AI = Fn->getArg(FirstIRArg);
         AI->setName(Arg->getName() + ".coerce");
-        CreateCoercedStore(AI, Ptr, /*DstIsVolatile=*/false, *this);
+        CreateCoercedStore(
+            AI, Ptr,
+            llvm::TypeSize::getFixed(
+                getContext().getTypeSizeInChars(Ty).getQuantity() -
+                ArgI.getDirectOffset()),
+            /*DstIsVolatile=*/false);
       }
 
       // Match to what EmitParmDecl is expecting for this type.
@@ -5939,17 +5919,8 @@ RValue CodeGenFunction::EmitCall(const CGFunctionInfo &CallInfo,
             llvm::Value *Imag = Builder.CreateExtractValue(CI, 1);
             return RValue::getComplex(std::make_pair(Real, Imag));
           }
-          case TEK_Aggregate: {
-            Address DestPtr = ReturnValue.getAddress();
-            bool DestIsVolatile = ReturnValue.isVolatile();
-
-            if (!DestPtr.isValid()) {
-              DestPtr = CreateMemTemp(RetTy, "agg.tmp");
-              DestIsVolatile = false;
-            }
-            EmitAggregateStore(CI, DestPtr, DestIsVolatile);
-            return RValue::getAggregate(DestPtr);
-          }
+          case TEK_Aggregate:
+            break;
           case TEK_Scalar: {
             // If the argument doesn't match, perform a bitcast to coerce it.
             // This can happen due to trivial type mismatches.
@@ -5959,7 +5930,6 @@ RValue CodeGenFunction::EmitCall(const CGFunctionInfo &CallInfo,
             return RValue::get(V);
           }
           }
-          llvm_unreachable("bad evaluation kind");
         }
 
         // If coercing a fixed vector from a scalable vector for ABI
@@ -5981,10 +5951,13 @@ RValue CodeGenFunction::EmitCall(const CGFunctionInfo &CallInfo,
 
         Address DestPtr = ReturnValue.getValue();
         bool DestIsVolatile = ReturnValue.isVolatile();
+        uint64_t DestSize =
+            getContext().getTypeInfoDataSizeInChars(RetTy).Width.getQuantity();
 
         if (!DestPtr.isValid()) {
           DestPtr = CreateMemTemp(RetTy, "coerce");
           DestIsVolatile = false;
+          DestSize = getContext().getTypeSizeInChars(RetTy).getQuantity();
         }
 
         // An empty record can overlap other data (if declared with
@@ -5993,7 +5966,10 @@ RValue CodeGenFunction::EmitCall(const CGFunctionInfo &CallInfo,
         if (!isEmptyRecord(getContext(), RetTy, true)) {
           // If the value is offset in memory, apply the offset now.
           Address StorePtr = emitAddressAtOffset(*this, DestPtr, RetAI);
-          CreateCoercedStore(CI, StorePtr, DestIsVolatile, *this);
+          CreateCoercedStore(
+              CI, StorePtr,
+              llvm::TypeSize::getFixed(DestSize - RetAI.getDirectOffset()),
+              DestIsVolatile);
         }
 
         return convertTempToRValue(DestPtr, RetTy, SourceLocation());

clang/lib/CodeGen/CGExprAgg.cpp

@@ -131,15 +131,12 @@ class AggExprEmitter : public StmtVisitor<AggExprEmitter> {
     EnsureDest(E->getType());
 
     if (llvm::Value *Result = ConstantEmitter(CGF).tryEmitConstantExpr(E)) {
-      Address StoreDest = Dest.getAddress();
-      // The emitted value is guaranteed to have the same size as the
-      // destination but can have a different type. Just do a bitcast in this
-      // case to avoid incorrect GEPs.
-      if (Result->getType() != StoreDest.getType())
-        StoreDest = StoreDest.withElementType(Result->getType());
-
-      CGF.EmitAggregateStore(Result, StoreDest,
-                             E->getType().isVolatileQualified());
+      CGF.CreateCoercedStore(
+          Result, Dest.getAddress(),
+          llvm::TypeSize::getFixed(
+              Dest.getPreferredSize(CGF.getContext(), E->getType())
+                  .getQuantity()),
+          E->getType().isVolatileQualified());
       return;
     }
     return Visit(E->getSubExpr());
@@ -2050,6 +2047,10 @@ CodeGenFunction::getOverlapForFieldInit(const FieldDecl *FD) {
   if (!FD->hasAttr<NoUniqueAddressAttr>() || !FD->getType()->isRecordType())
     return AggValueSlot::DoesNotOverlap;
 
+  // Empty fields can overlap earlier fields.
+  if (FD->getType()->getAsCXXRecordDecl()->isEmpty())
+    return AggValueSlot::MayOverlap;
+
   // If the field lies entirely within the enclosing class's nvsize, its tail
   // padding cannot overlap any already-initialized object. (The only subobjects
   // with greater addresses that might already be initialized are vbases.)
@@ -2072,6 +2073,10 @@ AggValueSlot::Overlap_t CodeGenFunction::getOverlapForBaseInit(
   if (IsVirtual)
     return AggValueSlot::MayOverlap;
 
+  // Empty bases can overlap earlier bases.
+  if (BaseRD->isEmpty())
+    return AggValueSlot::MayOverlap;
+
   // If the base class is laid out entirely within the nvsize of the derived
   // class, its tail padding cannot yet be initialized, so we can issue
   // stores at the full width of the base class.

clang/lib/CodeGen/CodeGenFunction.h

@@ -4838,9 +4838,10 @@ class CodeGenFunction : public CodeGenTypeCache {
   void EmitAggFinalDestCopy(QualType Type, AggValueSlot Dest, const LValue &Src,
                             ExprValueKind SrcKind);
 
-  /// Build all the stores needed to initialize an aggregate at Dest with the
-  /// value Val.
-  void EmitAggregateStore(llvm::Value *Val, Address Dest, bool DestIsVolatile);
+  /// Create a store to \arg DstPtr from \arg Src, truncating the stored value
+  /// to at most \arg DstSize bytes.
+  void CreateCoercedStore(llvm::Value *Src, Address Dst, llvm::TypeSize DstSize,
+                          bool DstIsVolatile);
 
   /// EmitExtendGCLifetime - Given a pointer to an Objective-C object,
   /// make sure it survives garbage collection until this point.

clang/test/CodeGen/arm-mve-intrinsics/vld24.c

@@ -48,10 +48,13 @@ uint8x16x4_t test_vld4q_u8(const uint8_t *addr)
 
 // CHECK-LABEL: @test_vst2q_u32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[VALUE_COERCE_FCA_0_0_EXTRACT:%.*]] = extractvalue [[STRUCT_UINT32X4X2_T:%.*]] [[VALUE_COERCE:%.*]], 0, 0
-// CHECK-NEXT:    [[VALUE_COERCE_FCA_0_1_EXTRACT:%.*]] = extractvalue [[STRUCT_UINT32X4X2_T]] [[VALUE_COERCE]], 0, 1
-// CHECK-NEXT:    call void @llvm.arm.mve.vst2q.p0.v4i32(ptr [[ADDR:%.*]], <4 x i32> [[VALUE_COERCE_FCA_0_0_EXTRACT]], <4 x i32> [[VALUE_COERCE_FCA_0_1_EXTRACT]], i32 0)
-// CHECK-NEXT:    call void @llvm.arm.mve.vst2q.p0.v4i32(ptr [[ADDR]], <4 x i32> [[VALUE_COERCE_FCA_0_0_EXTRACT]], <4 x i32> [[VALUE_COERCE_FCA_0_1_EXTRACT]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = extractvalue [[STRUCT_UINT32X4X2_T:%.*]] [[VALUE_COERCE:%.*]], 0
+// CHECK-NEXT:    [[DOTFCA_0_EXTRACT:%.*]] = extractvalue [2 x <4 x i32>] [[TMP0]], 0
+// CHECK-NEXT:    [[DOTFCA_1_EXTRACT:%.*]] = extractvalue [2 x <4 x i32>] [[TMP0]], 1
+// CHECK-NEXT:    [[DOTFCA_0_0_INSERT:%.*]] = insertvalue [[STRUCT_UINT32X4X2_T]] poison, <4 x i32> [[DOTFCA_0_EXTRACT]], 0, 0
+// CHECK-NEXT:    [[DOTFCA_0_1_INSERT:%.*]] = insertvalue [[STRUCT_UINT32X4X2_T]] [[DOTFCA_0_0_INSERT]], <4 x i32> [[DOTFCA_1_EXTRACT]], 0, 1
+// CHECK-NEXT:    call void @llvm.arm.mve.vst2q.p0.v4i32(ptr [[ADDR:%.*]], <4 x i32> [[DOTFCA_0_EXTRACT]], <4 x i32> [[DOTFCA_1_EXTRACT]], i32 0)
+// CHECK-NEXT:    call void @llvm.arm.mve.vst2q.p0.v4i32(ptr [[ADDR]], <4 x i32> [[DOTFCA_0_EXTRACT]], <4 x i32> [[DOTFCA_1_EXTRACT]], i32 1)
 // CHECK-NEXT:    ret void
 //
 void test_vst2q_u32(uint32_t *addr, uint32x4x2_t value)
@@ -65,14 +68,19 @@ void test_vst2q_u32(uint32_t *addr, uint32x4x2_t value)
 
 // CHECK-LABEL: @test_vst4q_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[VALUE_COERCE_FCA_0_0_EXTRACT:%.*]] = extractvalue [[STRUCT_INT8X16X4_T:%.*]] [[VALUE_COERCE:%.*]], 0, 0
-// CHECK-NEXT:    [[VALUE_COERCE_FCA_0_1_EXTRACT:%.*]] = extractvalue [[STRUCT_INT8X16X4_T]] [[VALUE_COERCE]], 0, 1
-// CHECK-NEXT:    [[VALUE_COERCE_FCA_0_2_EXTRACT:%.*]] = extractvalue [[STRUCT_INT8X16X4_T]] [[VALUE_COERCE]], 0, 2
-// CHECK-NEXT:    [[VALUE_COERCE_FCA_0_3_EXTRACT:%.*]] = extractvalue [[STRUCT_INT8X16X4_T]] [[VALUE_COERCE]], 0, 3
-// CHECK-NEXT:    call void @llvm.arm.mve.vst4q.p0.v16i8(ptr [[ADDR:%.*]], <16 x i8> [[VALUE_COERCE_FCA_0_0_EXTRACT]], <16 x i8> [[VALUE_COERCE_FCA_0_1_EXTRACT]], <16 x i8> [[VALUE_COERCE_FCA_0_2_EXTRACT]], <16 x i8> [[VALUE_COERCE_FCA_0_3_EXTRACT]], i32 0)
-// CHECK-NEXT:    call void @llvm.arm.mve.vst4q.p0.v16i8(ptr [[ADDR]], <16 x i8> [[VALUE_COERCE_FCA_0_0_EXTRACT]], <16 x i8> [[VALUE_COERCE_FCA_0_1_EXTRACT]], <16 x i8> [[VALUE_COERCE_FCA_0_2_EXTRACT]], <16 x i8> [[VALUE_COERCE_FCA_0_3_EXTRACT]], i32 1)
-// CHECK-NEXT:    call void @llvm.arm.mve.vst4q.p0.v16i8(ptr [[ADDR]], <16 x i8> [[VALUE_COERCE_FCA_0_0_EXTRACT]], <16 x i8> [[VALUE_COERCE_FCA_0_1_EXTRACT]], <16 x i8> [[VALUE_COERCE_FCA_0_2_EXTRACT]], <16 x i8> [[VALUE_COERCE_FCA_0_3_EXTRACT]], i32 2)
-// CHECK-NEXT:    call void @llvm.arm.mve.vst4q.p0.v16i8(ptr [[ADDR]], <16 x i8> [[VALUE_COERCE_FCA_0_0_EXTRACT]], <16 x i8> [[VALUE_COERCE_FCA_0_1_EXTRACT]], <16 x i8> [[VALUE_COERCE_FCA_0_2_EXTRACT]], <16 x i8> [[VALUE_COERCE_FCA_0_3_EXTRACT]], i32 3)
+// CHECK-NEXT:    [[TMP0:%.*]] = extractvalue [[STRUCT_INT8X16X4_T:%.*]] [[VALUE_COERCE:%.*]], 0
+// CHECK-NEXT:    [[DOTFCA_0_EXTRACT:%.*]] = extractvalue [4 x <16 x i8>] [[TMP0]], 0
+// CHECK-NEXT:    [[DOTFCA_1_EXTRACT:%.*]] = extractvalue [4 x <16 x i8>] [[TMP0]], 1
+// CHECK-NEXT:    [[DOTFCA_2_EXTRACT:%.*]] = extractvalue [4 x <16 x i8>] [[TMP0]], 2
+// CHECK-NEXT:    [[DOTFCA_3_EXTRACT:%.*]] = extractvalue [4 x <16 x i8>] [[TMP0]], 3
+// CHECK-NEXT:    [[DOTFCA_0_0_INSERT:%.*]] = insertvalue [[STRUCT_INT8X16X4_T]] poison, <16 x i8> [[DOTFCA_0_EXTRACT]], 0, 0
+// CHECK-NEXT:    [[DOTFCA_0_1_INSERT:%.*]] = insertvalue [[STRUCT_INT8X16X4_T]] [[DOTFCA_0_0_INSERT]], <16 x i8> [[DOTFCA_1_EXTRACT]], 0, 1
+// CHECK-NEXT:    [[DOTFCA_0_2_INSERT:%.*]] = insertvalue [[STRUCT_INT8X16X4_T]] [[DOTFCA_0_1_INSERT]], <16 x i8> [[DOTFCA_2_EXTRACT]], 0, 2
+// CHECK-NEXT:    [[DOTFCA_0_3_INSERT:%.*]] = insertvalue [[STRUCT_INT8X16X4_T]] [[DOTFCA_0_2_INSERT]], <16 x i8> [[DOTFCA_3_EXTRACT]], 0, 3
+// CHECK-NEXT:    call void @llvm.arm.mve.vst4q.p0.v16i8(ptr [[ADDR:%.*]], <16 x i8> [[DOTFCA_0_EXTRACT]], <16 x i8> [[DOTFCA_1_EXTRACT]], <16 x i8> [[DOTFCA_2_EXTRACT]], <16 x i8> [[DOTFCA_3_EXTRACT]], i32 0)
+// CHECK-NEXT:    call void @llvm.arm.mve.vst4q.p0.v16i8(ptr [[ADDR]], <16 x i8> [[DOTFCA_0_EXTRACT]], <16 x i8> [[DOTFCA_1_EXTRACT]], <16 x i8> [[DOTFCA_2_EXTRACT]], <16 x i8> [[DOTFCA_3_EXTRACT]], i32 1)
+// CHECK-NEXT:    call void @llvm.arm.mve.vst4q.p0.v16i8(ptr [[ADDR]], <16 x i8> [[DOTFCA_0_EXTRACT]], <16 x i8> [[DOTFCA_1_EXTRACT]], <16 x i8> [[DOTFCA_2_EXTRACT]], <16 x i8> [[DOTFCA_3_EXTRACT]], i32 2)
+// CHECK-NEXT:    call void @llvm.arm.mve.vst4q.p0.v16i8(ptr [[ADDR]], <16 x i8> [[DOTFCA_0_EXTRACT]], <16 x i8> [[DOTFCA_1_EXTRACT]], <16 x i8> [[DOTFCA_2_EXTRACT]], <16 x i8> [[DOTFCA_3_EXTRACT]], i32 3)
 // CHECK-NEXT:    ret void
 //
 void test_vst4q_s8(int8_t *addr, int8x16x4_t value)
@@ -86,10 +94,13 @@ void test_vst4q_s8(int8_t *addr, int8x16x4_t value)
 
 // CHECK-LABEL: @test_vst2q_f16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[VALUE_COERCE_FCA_0_0_EXTRACT:%.*]] = extractvalue [[STRUCT_FLOAT16X8X2_T:%.*]] [[VALUE_COERCE:%.*]], 0, 0
-// CHECK-NEXT:    [[VALUE_COERCE_FCA_0_1_EXTRACT:%.*]] = extractvalue [[STRUCT_FLOAT16X8X2_T]] [[VALUE_COERCE]], 0, 1
-// CHECK-NEXT:    call void @llvm.arm.mve.vst2q.p0.v8f16(ptr [[ADDR:%.*]], <8 x half> [[VALUE_COERCE_FCA_0_0_EXTRACT]], <8 x half> [[VALUE_COERCE_FCA_0_1_EXTRACT]], i32 0)
-// CHECK-NEXT:    call void @llvm.arm.mve.vst2q.p0.v8f16(ptr [[ADDR]], <8 x half> [[VALUE_COERCE_FCA_0_0_EXTRACT]], <8 x half> [[VALUE_COERCE_FCA_0_1_EXTRACT]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = extractvalue [[STRUCT_FLOAT16X8X2_T:%.*]] [[VALUE_COERCE:%.*]], 0
+// CHECK-NEXT:    [[DOTFCA_0_EXTRACT:%.*]] = extractvalue [2 x <8 x half>] [[TMP0]], 0
+// CHECK-NEXT:    [[DOTFCA_1_EXTRACT:%.*]] = extractvalue [2 x <8 x half>] [[TMP0]], 1
+// CHECK-NEXT:    [[DOTFCA_0_0_INSERT:%.*]] = insertvalue [[STRUCT_FLOAT16X8X2_T]] poison, <8 x half> [[DOTFCA_0_EXTRACT]], 0, 0
+// CHECK-NEXT:    [[DOTFCA_0_1_INSERT:%.*]] = insertvalue [[STRUCT_FLOAT16X8X2_T]] [[DOTFCA_0_0_INSERT]], <8 x half> [[DOTFCA_1_EXTRACT]], 0, 1
+// CHECK-NEXT:    call void @llvm.arm.mve.vst2q.p0.v8f16(ptr [[ADDR:%.*]], <8 x half> [[DOTFCA_0_EXTRACT]], <8 x half> [[DOTFCA_1_EXTRACT]], i32 0)
+// CHECK-NEXT:    call void @llvm.arm.mve.vst2q.p0.v8f16(ptr [[ADDR]], <8 x half> [[DOTFCA_0_EXTRACT]], <8 x half> [[DOTFCA_1_EXTRACT]], i32 1)
 // CHECK-NEXT:    ret void
 //
 void test_vst2q_f16(float16_t *addr, float16x8x2_t value)