[CIR][CIRGen][Builtin][Neon] Lower neon_vaddv_f32, neon_vaddvq_f32 and neon_vaddvq_f64 (#1238)

[Neon intrinsic
definition](https://developer.arm.com/architectures/instruction-sets/intrinsics/vaddv_f32).
They are vector across operation which LLVM doesn't currently have a
generic intrinsic about it. As a side note for brainstorm, it might be
worth in the future for CIR to introduce Vector Across type operations
even though LLVM dialect doesn't have it yet. This would help to expose
opt opportunities.
E.g. a very trivial constant fold can happen if we are adding across a
constant vector.

Author: ghehg

clang/lib/CIR/CodeGen/CIRGenBuiltinAArch64.cpp

@@ -2678,16 +2678,15 @@ static mlir::Value emitCommonNeonSISDBuiltinExpr(
   case NEON::BI__builtin_neon_vaddlvq_u32:
     return emitNeonCall(builder, {argTy}, ops, "aarch64.neon.uaddlv", resultTy,
                         loc);
-  case NEON::BI__builtin_neon_vaddv_f32:
-    llvm_unreachable(" neon_vaddv_f32 NYI ");
   case NEON::BI__builtin_neon_vaddv_s32:
     llvm_unreachable(" neon_vaddv_s32 NYI ");
   case NEON::BI__builtin_neon_vaddv_u32:
     llvm_unreachable(" neon_vaddv_u32 NYI ");
+  case NEON::BI__builtin_neon_vaddv_f32:
   case NEON::BI__builtin_neon_vaddvq_f32:
-    llvm_unreachable(" neon_vaddvq_f32 NYI ");
   case NEON::BI__builtin_neon_vaddvq_f64:
-    llvm_unreachable(" neon_vaddvq_f64 NYI ");
+    return emitNeonCall(builder, {argTy}, ops, "aarch64.neon.faddv", resultTy,
+                        loc);
   case NEON::BI__builtin_neon_vaddvq_s32:
     llvm_unreachable(" neon_vaddvq_s32 NYI ");
   case NEON::BI__builtin_neon_vaddvq_s64:

clang/test/CIR/CodeGen/AArch64/neon.c

@@ -18383,26 +18383,41 @@ uint64x1_t test_vrsra_n_u64(uint64x1_t a, uint64x1_t b) {
 //   return vneg_s64(a);
 // }
 
-// NYI-LABEL: @test_vaddv_f32(
-// NYI:   [[VADDV_F32_I:%.*]] = call float @llvm.aarch64.neon.faddv.f32.v2f32(<2 x float> %a)
-// NYI:   ret float [[VADDV_F32_I]]
-// float32_t test_vaddv_f32(float32x2_t a) {
-//   return vaddv_f32(a);
-// }
+float32_t test_vaddv_f32(float32x2_t a) {
+  return vaddv_f32(a);
 
-// NYI-LABEL: @test_vaddvq_f32(
-// NYI:   [[VADDVQ_F32_I:%.*]] = call float @llvm.aarch64.neon.faddv.f32.v4f32(<4 x float> %a)
-// NYI:   ret float [[VADDVQ_F32_I]]
-// float32_t test_vaddvq_f32(float32x4_t a) {
-//   return vaddvq_f32(a);
-// }
+  // CIR-LABEL: vaddv_f32
+  // CIR: cir.llvm.intrinsic "aarch64.neon.faddv" {{%.*}} : (!cir.vector<!cir.float x 2>) -> !cir.float
 
-// NYI-LABEL: @test_vaddvq_f64(
-// NYI:   [[VADDVQ_F64_I:%.*]] = call double @llvm.aarch64.neon.faddv.f64.v2f64(<2 x double> %a)
-// NYI:   ret double [[VADDVQ_F64_I]]
-// float64_t test_vaddvq_f64(float64x2_t a) {
-//   return vaddvq_f64(a);
-// }
+  // LLVM-LABEL: test_vaddv_f32
+  // LLVM-SAME: (<2 x float> [[a:%.*]])
+  // LLVM: [[VADDV_F32_I:%.*]] = call float @llvm.aarch64.neon.faddv.f32.v2f32(<2 x float> [[a]])
+  // LLVM: ret float [[VADDV_F32_I]]
+}
+
+float32_t test_vaddvq_f32(float32x4_t a) {
+  return vaddvq_f32(a);
+
+  // CIR-LABEL: vaddvq_f32
+  // CIR: cir.llvm.intrinsic "aarch64.neon.faddv" {{%.*}} : (!cir.vector<!cir.float x 4>) -> !cir.float
+
+  // LLVM-LABEL: test_vaddvq_f32
+  // LLVM-SAME: (<4 x float> [[a:%.*]])
+  // LLVM: [[VADDVQ_F32_I:%.*]] = call float @llvm.aarch64.neon.faddv.f32.v4f32(<4 x float> [[a]])
+  // LLVM: ret float [[VADDVQ_F32_I]]
+}
+
+float64_t test_vaddvq_f64(float64x2_t a) {
+  return vaddvq_f64(a);
+
+  // CIR-LABEL: vaddvq_f64
+  // CIR: cir.llvm.intrinsic "aarch64.neon.faddv" {{%.*}} : (!cir.vector<!cir.double x 2>) -> !cir.double
+
+  // LLVM-LABEL: test_vaddvq_f64
+  // LLVM-SAME: (<2 x double> [[a:%.*]])
+  // LLVM: [[VADDVQ_F64_I:%.*]] = call double @llvm.aarch64.neon.faddv.f64.v2f64(<2 x double> [[a]])
+  // LLVM: ret double [[VADDVQ_F64_I]]
+}
 
 // NYI-LABEL: @test_vmaxv_f32(
 // NYI:   [[VMAXV_F32_I:%.*]] = call float @llvm.aarch64.neon.fmaxv.f32.v2f32(<2 x float> %a)