diff --git a/llvm/lib/Analysis/VectorUtils.cpp b/llvm/lib/Analysis/VectorUtils.cpp
index f90fca9d937fc..5b57f0a25cec8 100644
--- a/llvm/lib/Analysis/VectorUtils.cpp
+++ b/llvm/lib/Analysis/VectorUtils.cpp
@@ -123,6 +123,8 @@ bool llvm::isVectorIntrinsicWithScalarOpAtArg(Intrinsic::ID ID,
 
 bool llvm::isVectorIntrinsicWithOverloadTypeAtArg(Intrinsic::ID ID,
                                                   int OpdIdx) {
+  assert(ID != Intrinsic::not_intrinsic && "Not an intrinsic!");
+
   switch (ID) {
   case Intrinsic::fptosi_sat:
   case Intrinsic::fptoui_sat:
diff --git a/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp b/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
index ced081d429164..3049915260646 100644
--- a/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
+++ b/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
@@ -11596,10 +11596,6 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E, bool PostponedPHIs) {
       CallInst *CI = cast<CallInst>(VL0);
       setInsertPointAfterBundle(E);
 
-      Intrinsic::ID IID = Intrinsic::not_intrinsic;
-      if (Function *FI = CI->getCalledFunction())
-        IID = FI->getIntrinsicID();
-
       Intrinsic::ID ID = getVectorIntrinsicIDForCall(CI, TLI);
 
       auto VecCallCosts = getVectorCallCosts(CI, VecTy, TTI, TLI);
@@ -11610,18 +11606,18 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E, bool PostponedPHIs) {
       SmallVector<Value *> OpVecs;
       SmallVector<Type *, 2> TysForDecl;
       // Add return type if intrinsic is overloaded on it.
-      if (isVectorIntrinsicWithOverloadTypeAtArg(IID, -1))
+      if (UseIntrinsic && isVectorIntrinsicWithOverloadTypeAtArg(ID, -1))
         TysForDecl.push_back(
             FixedVectorType::get(CI->getType(), E->Scalars.size()));
       for (unsigned I : seq<unsigned>(0, CI->arg_size())) {
         ValueList OpVL;
         // Some intrinsics have scalar arguments. This argument should not be
         // vectorized.
-        if (UseIntrinsic && isVectorIntrinsicWithScalarOpAtArg(IID, I)) {
+        if (UseIntrinsic && isVectorIntrinsicWithScalarOpAtArg(ID, I)) {
           CallInst *CEI = cast<CallInst>(VL0);
           ScalarArg = CEI->getArgOperand(I);
           OpVecs.push_back(CEI->getArgOperand(I));
-          if (isVectorIntrinsicWithOverloadTypeAtArg(IID, I))
+          if (isVectorIntrinsicWithOverloadTypeAtArg(ID, I))
             TysForDecl.push_back(ScalarArg->getType());
           continue;
         }
@@ -11633,7 +11629,7 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E, bool PostponedPHIs) {
         }
         LLVM_DEBUG(dbgs() << "SLP: OpVec[" << I << "]: " << *OpVec << "\n");
         OpVecs.push_back(OpVec);
-        if (isVectorIntrinsicWithOverloadTypeAtArg(IID, I))
+        if (UseIntrinsic && isVectorIntrinsicWithOverloadTypeAtArg(ID, I))
           TysForDecl.push_back(OpVec->getType());
       }
 
diff --git a/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp b/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
index 991491de20892..76961629aeceb 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
@@ -498,16 +498,17 @@ void VPWidenCallRecipe::execute(VPTransformState &State) {
          "DbgInfoIntrinsic should have been dropped during VPlan construction");
   State.setDebugLocFrom(CI.getDebugLoc());
 
+  bool UseIntrinsic = VectorIntrinsicID != Intrinsic::not_intrinsic;
   FunctionType *VFTy = nullptr;
   if (Variant)
     VFTy = Variant->getFunctionType();
   for (unsigned Part = 0; Part < State.UF; ++Part) {
     SmallVector<Type *, 2> TysForDecl;
     // Add return type if intrinsic is overloaded on it.
-    if (isVectorIntrinsicWithOverloadTypeAtArg(VectorIntrinsicID, -1)) {
+    if (UseIntrinsic &&
+        isVectorIntrinsicWithOverloadTypeAtArg(VectorIntrinsicID, -1))
       TysForDecl.push_back(
           VectorType::get(CI.getType()->getScalarType(), State.VF));
-    }
     SmallVector<Value *, 4> Args;
     for (const auto &I : enumerate(operands())) {
       // Some intrinsics have a scalar argument - don't replace it with a
@@ -516,18 +517,19 @@ void VPWidenCallRecipe::execute(VPTransformState &State) {
       // e.g. linear parameters for pointers.
       Value *Arg;
       if ((VFTy && !VFTy->getParamType(I.index())->isVectorTy()) ||
-          (VectorIntrinsicID != Intrinsic::not_intrinsic &&
+          (UseIntrinsic &&
            isVectorIntrinsicWithScalarOpAtArg(VectorIntrinsicID, I.index())))
         Arg = State.get(I.value(), VPIteration(0, 0));
       else
         Arg = State.get(I.value(), Part);
-      if (isVectorIntrinsicWithOverloadTypeAtArg(VectorIntrinsicID, I.index()))
+      if (UseIntrinsic &&
+          isVectorIntrinsicWithOverloadTypeAtArg(VectorIntrinsicID, I.index()))
         TysForDecl.push_back(Arg->getType());
       Args.push_back(Arg);
     }
 
     Function *VectorF;
-    if (VectorIntrinsicID != Intrinsic::not_intrinsic) {
+    if (UseIntrinsic) {
       // Use vector version of the intrinsic.
       Module *M = State.Builder.GetInsertBlock()->getModule();
       VectorF = Intrinsic::getDeclaration(M, VectorIntrinsicID, TysForDecl);
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/vector-call-linear-args.ll b/llvm/test/Transforms/LoopVectorize/AArch64/vector-call-linear-args.ll
index 876d58131bd7a..16506b3a57573 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/vector-call-linear-args.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/vector-call-linear-args.ll
@@ -1,4 +1,4 @@
-; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --filter "call.*(foo|bar|baz|quux)" --version 2
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --filter "call.*(foo|bar|baz|quux|goo)" --version 2
 ; RUN: opt < %s -passes=loop-vectorize -force-vector-interleave=1 -S | FileCheck %s --check-prefixes=NEON
 ; RUN: opt < %s -mattr=+sve -passes=loop-vectorize -force-vector-interleave=1 -S | FileCheck %s --check-prefixes=SVE_OR_NEON
 ; RUN: opt < %s -mattr=+sve -passes=loop-vectorize -force-vector-interleave=1 -S -prefer-predicate-over-epilogue=predicate-dont-vectorize | FileCheck %s --check-prefixes=SVE_TF
@@ -15,13 +15,13 @@ define void @test_linear8(ptr noalias %a, ptr readnone %b, i64 %n) {
 ;
 ; SVE_OR_NEON-LABEL: define void @test_linear8
 ; SVE_OR_NEON-SAME: (ptr noalias [[A:%.*]], ptr readnone [[B:%.*]], i64 [[N:%.*]]) #[[ATTR0:[0-9]+]] {
-; SVE_OR_NEON:    [[TMP13:%.*]] = call <vscale x 2 x i64> @vec_foo_linear8_nomask_sve(ptr [[TMP12:%.*]])
+; SVE_OR_NEON:    [[TMP15:%.*]] = call <vscale x 2 x i64> @vec_foo_linear8_nomask_sve(ptr [[TMP14:%.*]])
 ; SVE_OR_NEON:    [[DATA:%.*]] = call i64 @foo(ptr [[GEPB:%.*]]) #[[ATTR2:[0-9]+]]
 ;
 ; SVE_TF-LABEL: define void @test_linear8
 ; SVE_TF-SAME: (ptr noalias [[A:%.*]], ptr readnone [[B:%.*]], i64 [[N:%.*]]) #[[ATTR0:[0-9]+]] {
-; SVE_TF:    [[TMP19:%.*]] = call <vscale x 2 x i64> @vec_foo_linear8_mask_sve(ptr [[TMP18:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
-; SVE_TF:    [[DATA:%.*]] = call i64 @foo(ptr [[GEPB:%.*]]) #[[ATTR3:[0-9]+]]
+; SVE_TF:    [[TMP21:%.*]] = call <vscale x 2 x i64> @vec_foo_linear8_mask_sve(ptr [[TMP20:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SVE_TF:    [[DATA:%.*]] = call i64 @foo(ptr [[GEPB:%.*]]) #[[ATTR4:[0-9]+]]
 ;
 entry:
   br label %for.body
@@ -48,12 +48,12 @@ define void @test_vector_linear4(ptr noalias %a, ptr readnone %b, ptr readonly %
 ;
 ; SVE_OR_NEON-LABEL: define void @test_vector_linear4
 ; SVE_OR_NEON-SAME: (ptr noalias [[A:%.*]], ptr readnone [[B:%.*]], ptr readonly [[C:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
-; SVE_OR_NEON:    [[TMP15:%.*]] = call <vscale x 4 x i32> @vec_baz_vector_linear4_nomask_sve(<vscale x 4 x i32> [[WIDE_LOAD:%.*]], ptr [[TMP14:%.*]])
+; SVE_OR_NEON:    [[TMP17:%.*]] = call <vscale x 4 x i32> @vec_baz_vector_linear4_nomask_sve(<vscale x 4 x i32> [[WIDE_LOAD:%.*]], ptr [[TMP16:%.*]])
 ; SVE_OR_NEON:    [[DATA:%.*]] = call i32 @baz(i32 [[INPUT:%.*]], ptr [[GEPB:%.*]]) #[[ATTR3:[0-9]+]]
 ;
 ; SVE_TF-LABEL: define void @test_vector_linear4
 ; SVE_TF-SAME: (ptr noalias [[A:%.*]], ptr readnone [[B:%.*]], ptr readonly [[C:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
-; SVE_TF:    [[DATA:%.*]] = call i32 @baz(i32 [[INPUT:%.*]], ptr [[GEPB:%.*]]) #[[ATTR4:[0-9]+]]
+; SVE_TF:    [[DATA:%.*]] = call i32 @baz(i32 [[INPUT:%.*]], ptr [[GEPB:%.*]]) #[[ATTR5:[0-9]+]]
 ;
 entry:
   br label %for.body
@@ -85,7 +85,7 @@ define void @test_linear8_bad_stride(ptr noalias %a, ptr readnone %b, i64 %n) {
 ;
 ; SVE_TF-LABEL: define void @test_linear8_bad_stride
 ; SVE_TF-SAME: (ptr noalias [[A:%.*]], ptr readnone [[B:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
-; SVE_TF:    [[DATA:%.*]] = call i64 @foo(ptr [[GEPB:%.*]]) #[[ATTR5:[0-9]+]]
+; SVE_TF:    [[DATA:%.*]] = call i64 @foo(ptr [[GEPB:%.*]]) #[[ATTR6:[0-9]+]]
 ;
 entry:
   br label %for.body
@@ -112,12 +112,12 @@ define void @test_linear16_wide_stride(ptr noalias %a, ptr readnone %b, i64 %n)
 ;
 ; SVE_OR_NEON-LABEL: define void @test_linear16_wide_stride
 ; SVE_OR_NEON-SAME: (ptr noalias [[A:%.*]], ptr readnone [[B:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
-; SVE_OR_NEON:    [[TMP14:%.*]] = call <vscale x 2 x i64> @vec_foo_linear16_nomask_sve(ptr [[TMP13:%.*]])
+; SVE_OR_NEON:    [[TMP16:%.*]] = call <vscale x 2 x i64> @vec_foo_linear16_nomask_sve(ptr [[TMP15:%.*]])
 ; SVE_OR_NEON:    [[DATA:%.*]] = call i64 @foo(ptr [[GEPB:%.*]]) #[[ATTR4]]
 ;
 ; SVE_TF-LABEL: define void @test_linear16_wide_stride
 ; SVE_TF-SAME: (ptr noalias [[A:%.*]], ptr readnone [[B:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
-; SVE_TF:    [[DATA:%.*]] = call i64 @foo(ptr [[GEPB:%.*]]) #[[ATTR5]]
+; SVE_TF:    [[DATA:%.*]] = call i64 @foo(ptr [[GEPB:%.*]]) #[[ATTR6]]
 ;
 entry:
   br label %for.body
@@ -145,13 +145,13 @@ define void @test_linear4_linear8(ptr noalias %a, ptr readnone %b, ptr readonly
 ;
 ; SVE_OR_NEON-LABEL: define void @test_linear4_linear8
 ; SVE_OR_NEON-SAME: (ptr noalias [[A:%.*]], ptr readnone [[B:%.*]], ptr readonly [[C:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
-; SVE_OR_NEON:    [[TMP15:%.*]] = call <vscale x 4 x i32> @vec_quux_linear4_linear8_mask_sve(ptr [[TMP13:%.*]], ptr [[TMP14:%.*]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer))
+; SVE_OR_NEON:    [[TMP17:%.*]] = call <vscale x 4 x i32> @vec_quux_linear4_linear8_mask_sve(ptr [[TMP15:%.*]], ptr [[TMP16:%.*]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer))
 ; SVE_OR_NEON:    [[DATA:%.*]] = call i32 @quux(ptr [[GEPC:%.*]], ptr [[GEPB:%.*]]) #[[ATTR5:[0-9]+]]
 ;
 ; SVE_TF-LABEL: define void @test_linear4_linear8
 ; SVE_TF-SAME: (ptr noalias [[A:%.*]], ptr readnone [[B:%.*]], ptr readonly [[C:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
-; SVE_TF:    [[TMP21:%.*]] = call <vscale x 4 x i32> @vec_quux_linear4_linear8_mask_sve(ptr [[TMP19:%.*]], ptr [[TMP20:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
-; SVE_TF:    [[DATA:%.*]] = call i32 @quux(ptr [[GEPC:%.*]], ptr [[GEPB:%.*]]) #[[ATTR6:[0-9]+]]
+; SVE_TF:    [[TMP23:%.*]] = call <vscale x 4 x i32> @vec_quux_linear4_linear8_mask_sve(ptr [[TMP21:%.*]], ptr [[TMP22:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SVE_TF:    [[DATA:%.*]] = call i32 @quux(ptr [[GEPC:%.*]], ptr [[GEPB:%.*]]) #[[ATTR7:[0-9]+]]
 ;
 entry:
   br label %for.body
@@ -179,12 +179,12 @@ define void @test_linear3_non_ptr(ptr noalias %a, i64 %n) {
 ;
 ; SVE_OR_NEON-LABEL: define void @test_linear3_non_ptr
 ; SVE_OR_NEON-SAME: (ptr noalias [[A:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
-; SVE_OR_NEON:    [[TMP13:%.*]] = call <vscale x 4 x i32> @vec_bar_linear3_nomask_sve(i32 [[TMP12:%.*]])
+; SVE_OR_NEON:    [[TMP15:%.*]] = call <vscale x 4 x i32> @vec_bar_linear3_nomask_sve(i32 [[TMP14:%.*]])
 ; SVE_OR_NEON:    [[DATA:%.*]] = call i32 @bar(i32 [[TREBLED:%.*]]) #[[ATTR6:[0-9]+]]
 ;
 ; SVE_TF-LABEL: define void @test_linear3_non_ptr
 ; SVE_TF-SAME: (ptr noalias [[A:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
-; SVE_TF:    [[DATA:%.*]] = call i32 @bar(i32 [[TREBLED:%.*]]) #[[ATTR7:[0-9]+]]
+; SVE_TF:    [[DATA:%.*]] = call i32 @bar(i32 [[TREBLED:%.*]]) #[[ATTR8:[0-9]+]]
 ;
 entry:
   br label %for.body
@@ -212,12 +212,12 @@ define void @test_linearn5_non_ptr_neg_stride(ptr noalias %a, i64 %n) {
 ;
 ; SVE_OR_NEON-LABEL: define void @test_linearn5_non_ptr_neg_stride
 ; SVE_OR_NEON-SAME: (ptr noalias [[A:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
-; SVE_OR_NEON:    [[TMP13:%.*]] = call <vscale x 4 x i32> @vec_bar_linearn5_nomask_sve(i32 [[TMP12:%.*]])
+; SVE_OR_NEON:    [[TMP15:%.*]] = call <vscale x 4 x i32> @vec_bar_linearn5_nomask_sve(i32 [[TMP14:%.*]])
 ; SVE_OR_NEON:    [[DATA:%.*]] = call i32 @bar(i32 [[NEGSTRIDE:%.*]]) #[[ATTR7:[0-9]+]]
 ;
 ; SVE_TF-LABEL: define void @test_linearn5_non_ptr_neg_stride
 ; SVE_TF-SAME: (ptr noalias [[A:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
-; SVE_TF:    [[DATA:%.*]] = call i32 @bar(i32 [[NEGSTRIDE:%.*]]) #[[ATTR8:[0-9]+]]
+; SVE_TF:    [[DATA:%.*]] = call i32 @bar(i32 [[NEGSTRIDE:%.*]]) #[[ATTR9:[0-9]+]]
 ;
 entry:
   br label %for.body
@@ -237,10 +237,44 @@ for.cond.cleanup:
   ret void
 }
 
+define void @test_linear8_return_void(ptr noalias %in, ptr noalias %out, i64 %n) {
+; NEON-LABEL: define void @test_linear8_return_void
+; NEON-SAME: (ptr noalias [[IN:%.*]], ptr noalias [[OUT:%.*]], i64 [[N:%.*]]) {
+; NEON:    call void @vec_goo_linear8_nomask_neon(<2 x i64> [[WIDE_LOAD:%.*]], ptr [[TMP4:%.*]])
+; NEON:    call void @goo(i64 [[NUM:%.*]], ptr [[GEP_OUT:%.*]]) #[[ATTR6:[0-9]+]]
+;
+; SVE_OR_NEON-LABEL: define void @test_linear8_return_void
+; SVE_OR_NEON-SAME: (ptr noalias [[IN:%.*]], ptr noalias [[OUT:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
+; SVE_OR_NEON:    call void @vec_goo_linear8_nomask_sve(<vscale x 2 x i64> [[WIDE_LOAD:%.*]], ptr [[TMP16:%.*]])
+; SVE_OR_NEON:    call void @goo(i64 [[NUM:%.*]], ptr [[GEP_OUT:%.*]]) #[[ATTR8:[0-9]+]]
+;
+; SVE_TF-LABEL: define void @test_linear8_return_void
+; SVE_TF-SAME: (ptr noalias [[IN:%.*]], ptr noalias [[OUT:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
+; SVE_TF:    call void @vec_goo_linear8_mask_sve(<vscale x 2 x i64> [[WIDE_MASKED_LOAD:%.*]], ptr [[TMP22:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SVE_TF:    call void @goo(i64 [[NUM:%.*]], ptr [[GEP_OUT:%.*]]) #[[ATTR10:[0-9]+]]
+;
+entry:
+  br label %for.body
+
+for.body:
+  %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
+  %gep.in = getelementptr i64, ptr %in, i64 %indvars.iv
+  %num = load i64, ptr %gep.in, align 8
+  %gep.out = getelementptr i64, ptr %out, i64 %indvars.iv
+  call void @goo(i64 %num, ptr %gep.out) #6
+  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
+  %exitcond = icmp eq i64 %indvars.iv.next, %n
+  br i1 %exitcond, label %for.cond.cleanup, label %for.body
+
+for.cond.cleanup:
+  ret void
+}
+
 declare i64 @foo(ptr)
 declare i32 @baz(i32, ptr)
 declare i32 @quux(ptr, ptr)
 declare i32 @bar(i32)
+declare void @goo(i64, ptr)
 
 ; neon vector variants of foo
 declare <2 x i64> @vec_foo_linear8_nomask_neon(ptr)
@@ -249,6 +283,7 @@ declare <4 x i32> @vec_baz_vector_linear4_nomask_neon(<4 x i32>, ptr)
 declare <4 x i32> @vec_quux_linear4_linear8_nomask_neon(ptr, ptr)
 declare <4 x i32> @vec_bar_linear3_nomask_neon(i32)
 declare <4 x i32> @vec_bar_linearn5_nomask_neon(i32)
+declare void @vec_goo_linear8_nomask_neon(<2 x i64>, ptr)
 
 ; scalable vector variants of foo
 declare <vscale x 2 x i64> @vec_foo_linear8_mask_sve(ptr, <vscale x 2 x i1>)
@@ -258,6 +293,8 @@ declare <vscale x 4 x i32> @vec_baz_vector_linear4_nomask_sve(<vscale x 4 x i32>
 declare <vscale x 4 x i32> @vec_quux_linear4_linear8_mask_sve(ptr, ptr, <vscale x 4 x i1>)
 declare <vscale x 4 x i32> @vec_bar_linear3_nomask_sve(i32)
 declare <vscale x 4 x i32> @vec_bar_linearn5_nomask_sve(i32)
+declare void @vec_goo_linear8_nomask_sve(<vscale x 2 x i64>, ptr)
+declare void @vec_goo_linear8_mask_sve(<vscale x 2 x i64>, ptr, <vscale x 2 x i1>)
 
 attributes #0 = { nounwind "vector-function-abi-variant"="_ZGVsNxl8_foo(vec_foo_linear8_nomask_sve),_ZGVsMxl8_foo(vec_foo_linear8_mask_sve),_ZGVnN2l8_foo(vec_foo_linear8_nomask_neon)" }
 attributes #1 = { nounwind "vector-function-abi-variant"="_ZGVsNxvl4_baz(vec_baz_vector_linear4_nomask_sve),_ZGVnN4vl4_baz(vec_baz_vector_linear4_nomask_neon)" }
@@ -265,3 +302,4 @@ attributes #2 = { nounwind "vector-function-abi-variant"="_ZGVsNxl16_foo(vec_foo
 attributes #3 = { nounwind "vector-function-abi-variant"="_ZGVsMxl4l8_quux(vec_quux_linear4_linear8_mask_sve),_ZGVnN4l4l8_quux(vec_quux_linear4_linear8_nomask_neon)" }
 attributes #4 = { nounwind "vector-function-abi-variant"="_ZGVsNxl3_bar(vec_bar_linear3_nomask_sve),_ZGVnN4l3_bar(vec_bar_linear3_nomask_neon)" }
 attributes #5 = { nounwind "vector-function-abi-variant"="_ZGVsNxln5_bar(vec_bar_linearn5_nomask_sve),_ZGVnN4ln5_bar(vec_bar_linearn5_nomask_neon)" }
+attributes #6 = { nounwind "vector-function-abi-variant"="_ZGVsNxvl8_goo(vec_goo_linear8_nomask_sve),_ZGVsMxvl8_goo(vec_goo_linear8_mask_sve),_ZGVsN2vl8_goo(vec_goo_linear8_nomask_neon)" }