From c83b2793b5d2a67af97995351b0efdc58a1d961e Mon Sep 17 00:00:00 2001
From: AmrDeveloper <amr96@programmer.net>
Date: Thu, 6 Mar 2025 20:44:34 +0100
Subject: [PATCH] [CIR][CIRGen][Builtin][Neon] Lower neon vaddlv_u32

---
 clang/lib/CIR/CodeGen/CIRGenBuiltinAArch64.cpp |  3 ++-
 clang/test/CIR/CodeGen/AArch64/neon.c          | 17 +++++++++++------
 2 files changed, 13 insertions(+), 7 deletions(-)

diff --git a/clang/lib/CIR/CodeGen/CIRGenBuiltinAArch64.cpp b/clang/lib/CIR/CodeGen/CIRGenBuiltinAArch64.cpp
index 96610d2df193..9908dcf02e83 100644
--- a/clang/lib/CIR/CodeGen/CIRGenBuiltinAArch64.cpp
+++ b/clang/lib/CIR/CodeGen/CIRGenBuiltinAArch64.cpp
@@ -2659,7 +2659,8 @@ static mlir::Value emitCommonNeonSISDBuiltinExpr(
   case NEON::BI__builtin_neon_vaddlv_s32:
     llvm_unreachable(" neon_vaddlv_s32 NYI ");
   case NEON::BI__builtin_neon_vaddlv_u32:
-    llvm_unreachable(" neon_vaddlv_u32 NYI ");
+    return emitNeonCall(builder, {argTy}, ops, "aarch64.neon.uaddlv", resultTy,
+                        loc);
   case NEON::BI__builtin_neon_vaddlvq_s32:
     llvm_unreachable(" neon_vaddlvq_s32 NYI ");
   case NEON::BI__builtin_neon_vaddlvq_u32:
diff --git a/clang/test/CIR/CodeGen/AArch64/neon.c b/clang/test/CIR/CodeGen/AArch64/neon.c
index 1d31dc7ab5b4..0d9559d8a22e 100644
--- a/clang/test/CIR/CodeGen/AArch64/neon.c
+++ b/clang/test/CIR/CodeGen/AArch64/neon.c
@@ -19217,12 +19217,17 @@ uint32_t test_vaddv_u32(uint32x2_t a) {
 //   return vaddlv_s32(a);
 // }
 
-// NYI-LABEL: @test_vaddlv_u32(
-// NYI:   [[VADDLV_U32_I:%.*]] = call i64 @llvm.aarch64.neon.uaddlv.i64.v2i32(<2 x i32> %a)
-// NYI:   ret i64 [[VADDLV_U32_I]]
-// uint64_t test_vaddlv_u32(uint32x2_t a) {
-//   return vaddlv_u32(a);
-// }
+uint64_t test_vaddlv_u32(uint32x2_t a) {
+  return vaddlv_u32(a);
+
+  // CIR-LABEL: vaddlv_u32
+  // CIR: cir.llvm.intrinsic "aarch64.neon.uaddlv" {{%.*}} : (!cir.vector<!u32i x 2>) -> !u64i
+
+  // LLVM-LABEL: test_vaddlv_u32
+  // LLVM-SAME: (<2 x i32> [[a:%.*]])
+  // LLVM:   [[VADDLV_U32_I:%.*]] = call i64 @llvm.aarch64.neon.uaddlv.i64.v2i32(<2 x i32> [[a]])
+  // LLVM:   ret i64 [[VADDLV_U32_I]]
+}
 
 uint8x8_t test_vmovn_u16(uint16x8_t a) {
   return vmovn_u16(a);