[WASM SIMD] Implement horizontal add for float and integer types.

src/compiler/arm/instruction-selector-arm.cc

Index: src/compiler/arm/instruction-selector-arm.cc
diff --git a/src/compiler/arm/instruction-selector-arm.cc b/src/compiler/arm/instruction-selector-arm.cc
index 9850e40cf340f3586ec37c3ea551e6176e216207..e4b489524f3cfbd50b0e38ef0d2f0dc79be2f6d1 100644
--- a/src/compiler/arm/instruction-selector-arm.cc
+++ b/src/compiler/arm/instruction-selector-arm.cc
@@ -2443,6 +2443,7 @@ VISIT_ATOMIC_BINOP(Xor)
 
 #define SIMD_BINOP_LIST(V)                          \
   V(F32x4Add, kArmF32x4Add)                         \
+  V(F32x4AddHoriz, kArmF32x4AddHoriz)               \
   V(F32x4Sub, kArmF32x4Sub)                         \
   V(F32x4Mul, kArmF32x4Mul)                         \
   V(F32x4Min, kArmF32x4Min)                         \
@@ -2454,6 +2455,7 @@ VISIT_ATOMIC_BINOP(Xor)
   V(F32x4Lt, kArmF32x4Lt)                           \
   V(F32x4Le, kArmF32x4Le)                           \
   V(I32x4Add, kArmI32x4Add)                         \
+  V(I32x4AddHoriz, kArmI32x4AddHoriz)               \
   V(I32x4Sub, kArmI32x4Sub)                         \
   V(I32x4Mul, kArmI32x4Mul)                         \
   V(I32x4MinS, kArmI32x4MinS)                       \
@@ -2469,6 +2471,7 @@ VISIT_ATOMIC_BINOP(Xor)
   V(I16x8SConvertI32x4, kArmI16x8SConvertI32x4)     \
   V(I16x8Add, kArmI16x8Add)                         \
   V(I16x8AddSaturateS, kArmI16x8AddSaturateS)       \
+  V(I16x8AddHoriz, kArmI16x8AddHoriz)               \
   V(I16x8Sub, kArmI16x8Sub)                         \
   V(I16x8SubSaturateS, kArmI16x8SubSaturateS)       \
   V(I16x8Mul, kArmI16x8Mul)                         \
@@ -2488,6 +2491,7 @@ VISIT_ATOMIC_BINOP(Xor)
   V(I8x16SConvertI16x8, kArmI8x16SConvertI16x8)     \
   V(I8x16Add, kArmI8x16Add)                         \
   V(I8x16AddSaturateS, kArmI8x16AddSaturateS)       \
+  V(I8x16AddHoriz, kArmI8x16AddHoriz)               \
   V(I8x16Sub, kArmI8x16Sub)                         \
   V(I8x16SubSaturateS, kArmI8x16SubSaturateS)       \
   V(I8x16Mul, kArmI8x16Mul)                         \