[WebAssembly] Codegen for v128.loadX_lane instructions

Replace the experimental clang builtin and LLVM intrinsics for these
instructions with normal codegen patterns. Resolves PR50433.

Differential Revision: https://reviews.llvm.org/D105950

diff --git a/clang/include/clang/Basic/BuiltinsWebAssembly.def b/clang/include/clang/Basic/BuiltinsWebAssembly.def
index 58d10ea..d05f5cd 100644 (file)
--- a/clang/include/clang/Basic/BuiltinsWebAssembly.def
+++ b/clang/include/clang/Basic/BuiltinsWebAssembly.def
@@ -195,10 +195,6 @@ TARGET_BUILTIN(__builtin_wasm_trunc_sat_zero_u_f64x2_i32x4, "V4UiV2d", "nc", "si
 TARGET_BUILTIN(__builtin_wasm_load32_zero, "V4iiC*", "n", "simd128")
 TARGET_BUILTIN(__builtin_wasm_load64_zero, "V2LLiLLiC*", "n", "simd128")
 
-TARGET_BUILTIN(__builtin_wasm_load8_lane, "V16ScScC*V16ScIi", "n", "simd128")
-TARGET_BUILTIN(__builtin_wasm_load16_lane, "V8ssC*V8sIi", "n", "simd128")
-TARGET_BUILTIN(__builtin_wasm_load32_lane, "V4iiC*V4iIi", "n", "simd128")
-TARGET_BUILTIN(__builtin_wasm_load64_lane, "V2LLiLLiC*V2LLiIi", "n", "simd128")
 TARGET_BUILTIN(__builtin_wasm_store8_lane, "vSc*V16ScIi", "n", "simd128")
 TARGET_BUILTIN(__builtin_wasm_store16_lane, "vs*V8sIi", "n", "simd128")
 TARGET_BUILTIN(__builtin_wasm_store32_lane, "vi*V4iIi", "n", "simd128")

diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index baa1436..d4b2414 100644 (file)
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -17776,10 +17776,6 @@ Value *CodeGenFunction::EmitWebAssemblyBuiltinExpr(unsigned BuiltinID,
     Function *Callee = CGM.getIntrinsic(Intrinsic::wasm_load64_zero);
     return Builder.CreateCall(Callee, {Ptr});
   }
-  case WebAssembly::BI__builtin_wasm_load8_lane:
-  case WebAssembly::BI__builtin_wasm_load16_lane:
-  case WebAssembly::BI__builtin_wasm_load32_lane:
-  case WebAssembly::BI__builtin_wasm_load64_lane:
   case WebAssembly::BI__builtin_wasm_store8_lane:
   case WebAssembly::BI__builtin_wasm_store16_lane:
   case WebAssembly::BI__builtin_wasm_store32_lane:
@@ -17792,18 +17788,6 @@ Value *CodeGenFunction::EmitWebAssemblyBuiltinExpr(unsigned BuiltinID,
     Value *LaneIdx = llvm::ConstantInt::get(getLLVMContext(), *LaneIdxConst);
     unsigned IntNo;
     switch (BuiltinID) {
-    case WebAssembly::BI__builtin_wasm_load8_lane:
-      IntNo = Intrinsic::wasm_load8_lane;
-      break;
-    case WebAssembly::BI__builtin_wasm_load16_lane:
-      IntNo = Intrinsic::wasm_load16_lane;
-      break;
-    case WebAssembly::BI__builtin_wasm_load32_lane:
-      IntNo = Intrinsic::wasm_load32_lane;
-      break;
-    case WebAssembly::BI__builtin_wasm_load64_lane:
-      IntNo = Intrinsic::wasm_load64_lane;
-      break;
     case WebAssembly::BI__builtin_wasm_store8_lane:
       IntNo = Intrinsic::wasm_store8_lane;
       break;

diff --git a/clang/lib/Headers/wasm_simd128.h b/clang/lib/Headers/wasm_simd128.h
index dc54bb9..d52cbc0 100644 (file)
--- a/clang/lib/Headers/wasm_simd128.h
+++ b/clang/lib/Headers/wasm_simd128.h
@@ -170,21 +170,49 @@ wasm_v128_load64_zero(const void *__mem) {
   return (v128_t)(__i64x2){__v, 0};
 }
 
-#define wasm_v128_load8_lane(__ptr, __vec, __i)                                \
-  ((v128_t)__builtin_wasm_load8_lane((const signed char *)(__ptr),             \
-                                     (__i8x16)(__vec), (__i)))
+static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_v128_load8_lane(
+    const void *__mem, v128_t __vec, int __i) __REQUIRE_CONSTANT(__i) {
+  struct __wasm_v128_load8_lane_struct {
+    int8_t __v;
+  } __attribute__((__packed__, __may_alias__));
+  int8_t __v = ((const struct __wasm_v128_load8_lane_struct *)__mem)->__v;
+  __i8x16 __ret = (__i8x16)__vec;
+  __ret[__i] = __v;
+  return (v128_t)__ret;
+}
 
-#define wasm_v128_load16_lane(__ptr, __vec, __i)                               \
-  ((v128_t)__builtin_wasm_load16_lane((const short *)(__ptr),                  \
-                                      (__i16x8)(__vec), (__i)))
+static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_v128_load16_lane(
+    const void *__mem, v128_t __vec, int __i) __REQUIRE_CONSTANT(__i) {
+  struct __wasm_v128_load16_lane_struct {
+    int16_t __v;
+  } __attribute__((__packed__, __may_alias__));
+  int16_t __v = ((const struct __wasm_v128_load16_lane_struct *)__mem)->__v;
+  __i16x8 __ret = (__i16x8)__vec;
+  __ret[__i] = __v;
+  return (v128_t)__ret;
+}
 
-#define wasm_v128_load32_lane(__ptr, __vec, __i)                               \
-  ((v128_t)__builtin_wasm_load32_lane((const int *)(__ptr), (__i32x4)(__vec),  \
-                                      (__i)))
+static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_v128_load32_lane(
+    const void *__mem, v128_t __vec, int __i) __REQUIRE_CONSTANT(__i) {
+  struct __wasm_v128_load32_lane_struct {
+    int32_t __v;
+  } __attribute__((__packed__, __may_alias__));
+  int32_t __v = ((const struct __wasm_v128_load32_lane_struct *)__mem)->__v;
+  __i32x4 __ret = (__i32x4)__vec;
+  __ret[__i] = __v;
+  return (v128_t)__ret;
+}
 
-#define wasm_v128_load64_lane(__ptr, __vec, __i)                               \
-  ((v128_t)__builtin_wasm_load64_lane((const long long int *)(__ptr),          \
-                                      (__i64x2)(__vec), (__i)))
+static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_v128_load64_lane(
+    const void *__mem, v128_t __vec, int __i) __REQUIRE_CONSTANT(__i) {
+  struct __wasm_v128_load64_lane_struct {
+    int64_t __v;
+  } __attribute__((__packed__, __may_alias__));
+  int64_t __v = ((const struct __wasm_v128_load64_lane_struct *)__mem)->__v;
+  __i64x2 __ret = (__i64x2)__vec;
+  __ret[__i] = __v;
+  return (v128_t)__ret;
+}
 
 static __inline__ void __DEFAULT_FN_ATTRS wasm_v128_store(void *__mem,
                                                           v128_t __a) {

diff --git a/clang/test/CodeGen/builtins-wasm.c b/clang/test/CodeGen/builtins-wasm.c
index 8999d39..16b19ff 100644 (file)
--- a/clang/test/CodeGen/builtins-wasm.c
+++ b/clang/test/CodeGen/builtins-wasm.c
@@ -284,34 +284,6 @@ f64x2 replace_lane_f64x2(f64x2 v, double x) {
   // WEBASSEMBLY-NEXT: ret
 }
 
-i8x16 load8_lane(const signed char *p, i8x16 v) {
-  return __builtin_wasm_load8_lane(p, v, 0);
-  // WEBASSEMBLY: tail call <16 x i8> @llvm.wasm.load8.lane(
-  // WEBASSEMBLY-SAME: i8* %p, <16 x i8> %v, i32 0)
-  // WEBASSEMBLY-NEXT: ret
-}
-
-i16x8 load16_lane(const short *p, i16x8 v) {
-  return __builtin_wasm_load16_lane(p, v, 0);
-  // WEBASSEMBLY: tail call <8 x i16> @llvm.wasm.load16.lane(
-  // WEBASSEMBLY-SAME: i16* %p, <8 x i16> %v, i32 0)
-  // WEBASSEMBLY-NEXT: ret
-}
-
-i32x4 load32_lane(const int *p, i32x4 v) {
-  return __builtin_wasm_load32_lane(p, v, 0);
-  // WEBASSEMBLY: tail call <4 x i32> @llvm.wasm.load32.lane(
-  // WEBASSEMBLY-SAME: i32* %p, <4 x i32> %v, i32 0)
-  // WEBASSEMBLY-NEXT: ret
-}
-
-i64x2 load64_lane(const long long *p, i64x2 v) {
-  return __builtin_wasm_load64_lane(p, v, 0);
-  // WEBASSEMBLY: tail call <2 x i64> @llvm.wasm.load64.lane(
-  // WEBASSEMBLY-SAME: i64* %p, <2 x i64> %v, i32 0)
-  // WEBASSEMBLY-NEXT: ret
-}
-
 void store8_lane(signed char *p, i8x16 v) {
   __builtin_wasm_store8_lane(p, v, 0);
   // WEBASSEMBLY: call void @llvm.wasm.store8.lane(

diff --git a/clang/test/Headers/wasm.c b/clang/test/Headers/wasm.c
index cc2f2b6..6ac78e1 100644 (file)
--- a/clang/test/Headers/wasm.c
+++ b/clang/test/Headers/wasm.c
@@ -162,9 +162,10 @@ v128_t test_v128_load64_zero(const void *mem) {
 
 // CHECK-LABEL: @test_v128_load8_lane(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[VEC:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <16 x i8> @llvm.wasm.load8.lane(i8* [[PTR:%.*]], <16 x i8> [[TMP0]], i32 15)
-// CHECK-NEXT:    [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <4 x i32>
+// CHECK-NEXT:    [[TMP0:%.*]] = load i8, i8* [[PTR:%.*]], align 1, !tbaa [[TBAA2]]
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[VEC:%.*]] to <16 x i8>
+// CHECK-NEXT:    [[VECINS_I:%.*]] = insertelement <16 x i8> [[TMP1]], i8 [[TMP0]], i32 15
+// CHECK-NEXT:    [[TMP2:%.*]] = bitcast <16 x i8> [[VECINS_I]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
 v128_t test_v128_load8_lane(const uint8_t *ptr, v128_t vec) {
@@ -173,9 +174,10 @@ v128_t test_v128_load8_lane(const uint8_t *ptr, v128_t vec) {
 
 // CHECK-LABEL: @test_v128_load16_lane(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[VEC:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <8 x i16> @llvm.wasm.load16.lane(i16* [[PTR:%.*]], <8 x i16> [[TMP0]], i32 7)
-// CHECK-NEXT:    [[TMP2:%.*]] = bitcast <8 x i16> [[TMP1]] to <4 x i32>
+// CHECK-NEXT:    [[TMP0:%.*]] = load i16, i16* [[PTR:%.*]], align 1, !tbaa [[TBAA2]]
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[VEC:%.*]] to <8 x i16>
+// CHECK-NEXT:    [[VECINS_I:%.*]] = insertelement <8 x i16> [[TMP1]], i16 [[TMP0]], i32 7
+// CHECK-NEXT:    [[TMP2:%.*]] = bitcast <8 x i16> [[VECINS_I]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
 v128_t test_v128_load16_lane(const uint16_t *ptr, v128_t vec) {
@@ -184,8 +186,9 @@ v128_t test_v128_load16_lane(const uint16_t *ptr, v128_t vec) {
 
 // CHECK-LABEL: @test_v128_load32_lane(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <4 x i32> @llvm.wasm.load32.lane(i32* [[PTR:%.*]], <4 x i32> [[VEC:%.*]], i32 3)
-// CHECK-NEXT:    ret <4 x i32> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = load i32, i32* [[PTR:%.*]], align 1, !tbaa [[TBAA2]]
+// CHECK-NEXT:    [[VECINS_I:%.*]] = insertelement <4 x i32> [[VEC:%.*]], i32 [[TMP0]], i32 3
+// CHECK-NEXT:    ret <4 x i32> [[VECINS_I]]
 //
 v128_t test_v128_load32_lane(const uint32_t *ptr, v128_t vec) {
   return wasm_v128_load32_lane(ptr, vec, 3);
@@ -193,9 +196,10 @@ v128_t test_v128_load32_lane(const uint32_t *ptr, v128_t vec) {
 
 // CHECK-LABEL: @test_v128_load64_lane(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[VEC:%.*]] to <2 x i64>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x i64> @llvm.wasm.load64.lane(i64* [[PTR:%.*]], <2 x i64> [[TMP0]], i32 1)
-// CHECK-NEXT:    [[TMP2:%.*]] = bitcast <2 x i64> [[TMP1]] to <4 x i32>
+// CHECK-NEXT:    [[TMP0:%.*]] = load i64, i64* [[PTR:%.*]], align 1, !tbaa [[TBAA2]]
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[VEC:%.*]] to <2 x i64>
+// CHECK-NEXT:    [[VECINS_I:%.*]] = insertelement <2 x i64> [[TMP1]], i64 [[TMP0]], i32 1
+// CHECK-NEXT:    [[TMP2:%.*]] = bitcast <2 x i64> [[VECINS_I]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
 v128_t test_v128_load64_lane(const uint64_t *ptr, v128_t vec) {
@@ -1281,7 +1285,7 @@ v128_t test_v128_andnot(v128_t a, v128_t b) {
 // CHECK-LABEL: @test_v128_any_true(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call i32 @llvm.wasm.anytrue.v16i8(<16 x i8> [[TMP0]]) #[[ATTR10:[0-9]+]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call i32 @llvm.wasm.anytrue.v16i8(<16 x i8> [[TMP0]]) #[[ATTR8:[0-9]+]]
 // CHECK-NEXT:    [[TOBOOL_I:%.*]] = icmp ne i32 [[TMP1]], 0
 // CHECK-NEXT:    ret i1 [[TOBOOL_I]]
 //
@@ -1291,7 +1295,7 @@ bool test_v128_any_true(v128_t a) {
 
 // CHECK-LABEL: @test_v128_bitselect(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <4 x i32> @llvm.wasm.bitselect.v4i32(<4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]], <4 x i32> [[MASK:%.*]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <4 x i32> @llvm.wasm.bitselect.v4i32(<4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]], <4 x i32> [[MASK:%.*]]) #[[ATTR8]]
 // CHECK-NEXT:    ret <4 x i32> [[TMP0]]
 //
 v128_t test_v128_bitselect(v128_t a, v128_t b, v128_t mask) {
@@ -1301,7 +1305,7 @@ v128_t test_v128_bitselect(v128_t a, v128_t b, v128_t mask) {
 // CHECK-LABEL: @test_i8x16_abs(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <16 x i8> @llvm.abs.v16i8(<16 x i8> [[TMP0]], i1 false) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <16 x i8> @llvm.abs.v16i8(<16 x i8> [[TMP0]], i1 false) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -1323,7 +1327,7 @@ v128_t test_i8x16_neg(v128_t a) {
 // CHECK-LABEL: @test_i8x16_all_true(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call i32 @llvm.wasm.alltrue.v16i8(<16 x i8> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call i32 @llvm.wasm.alltrue.v16i8(<16 x i8> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TOBOOL_I:%.*]] = icmp ne i32 [[TMP1]], 0
 // CHECK-NEXT:    ret i1 [[TOBOOL_I]]
 //
@@ -1334,7 +1338,7 @@ bool test_i8x16_all_true(v128_t a) {
 // CHECK-LABEL: @test_i8x16_bitmask(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call i32 @llvm.wasm.bitmask.v16i8(<16 x i8> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call i32 @llvm.wasm.bitmask.v16i8(<16 x i8> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    ret i32 [[TMP1]]
 //
 int32_t test_i8x16_bitmask(v128_t a) {
@@ -1344,7 +1348,7 @@ int32_t test_i8x16_bitmask(v128_t a) {
 // CHECK-LABEL: @test_i8x16_popcnt(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <16 x i8> @llvm.wasm.popcnt(<16 x i8> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <16 x i8> @llvm.wasm.popcnt(<16 x i8> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -1410,7 +1414,7 @@ v128_t test_i8x16_add(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.sadd.sat.v16i8(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.sadd.sat.v16i8(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <16 x i8> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -1422,7 +1426,7 @@ v128_t test_i8x16_add_sat(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.uadd.sat.v16i8(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.uadd.sat.v16i8(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <16 x i8> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -1446,7 +1450,7 @@ v128_t test_i8x16_sub(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.wasm.sub.sat.signed.v16i8(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.wasm.sub.sat.signed.v16i8(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <16 x i8> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -1458,7 +1462,7 @@ v128_t test_i8x16_sub_sat(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.wasm.sub.sat.unsigned.v16i8(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.wasm.sub.sat.unsigned.v16i8(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <16 x i8> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -1522,7 +1526,7 @@ v128_t test_u8x16_max(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.wasm.avgr.unsigned.v16i8(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.wasm.avgr.unsigned.v16i8(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <16 x i8> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -1533,7 +1537,7 @@ v128_t test_u8x16_avgr(v128_t a, v128_t b) {
 // CHECK-LABEL: @test_i16x8_abs(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <8 x i16> @llvm.abs.v8i16(<8 x i16> [[TMP0]], i1 false) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <8 x i16> @llvm.abs.v8i16(<8 x i16> [[TMP0]], i1 false) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <8 x i16> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -1555,7 +1559,7 @@ v128_t test_i16x8_neg(v128_t a) {
 // CHECK-LABEL: @test_i16x8_all_true(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call i32 @llvm.wasm.alltrue.v8i16(<8 x i16> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call i32 @llvm.wasm.alltrue.v8i16(<8 x i16> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TOBOOL_I:%.*]] = icmp ne i32 [[TMP1]], 0
 // CHECK-NEXT:    ret i1 [[TOBOOL_I]]
 //
@@ -1566,7 +1570,7 @@ bool test_i16x8_all_true(v128_t a) {
 // CHECK-LABEL: @test_i16x8_bitmask(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call i32 @llvm.wasm.bitmask.v8i16(<8 x i16> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call i32 @llvm.wasm.bitmask.v8i16(<8 x i16> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    ret i32 [[TMP1]]
 //
 int32_t test_i16x8_bitmask(v128_t a) {
@@ -1631,7 +1635,7 @@ v128_t test_i16x8_add(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.sadd.sat.v8i16(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.sadd.sat.v8i16(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <8 x i16> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -1643,7 +1647,7 @@ v128_t test_i16x8_add_sat(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.uadd.sat.v8i16(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.uadd.sat.v8i16(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <8 x i16> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -1667,7 +1671,7 @@ v128_t test_i16x8_sub(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.sub.sat.signed.v8i16(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.sub.sat.signed.v8i16(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <8 x i16> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -1679,7 +1683,7 @@ v128_t test_i16x8_sub_sat(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.sub.sat.unsigned.v8i16(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.sub.sat.unsigned.v8i16(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <8 x i16> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -1755,7 +1759,7 @@ v128_t test_u16x8_max(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.avgr.unsigned.v8i16(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.avgr.unsigned.v8i16(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <8 x i16> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -1765,7 +1769,7 @@ v128_t test_u16x8_avgr(v128_t a, v128_t b) {
 
 // CHECK-LABEL: @test_i32x4_abs(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <4 x i32> @llvm.abs.v4i32(<4 x i32> [[A:%.*]], i1 false) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <4 x i32> @llvm.abs.v4i32(<4 x i32> [[A:%.*]], i1 false) #[[ATTR8]]
 // CHECK-NEXT:    ret <4 x i32> [[TMP0]]
 //
 v128_t test_i32x4_abs(v128_t a) {
@@ -1783,7 +1787,7 @@ v128_t test_i32x4_neg(v128_t a) {
 
 // CHECK-LABEL: @test_i32x4_all_true(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call i32 @llvm.wasm.alltrue.v4i32(<4 x i32> [[A:%.*]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call i32 @llvm.wasm.alltrue.v4i32(<4 x i32> [[A:%.*]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TOBOOL_I:%.*]] = icmp ne i32 [[TMP0]], 0
 // CHECK-NEXT:    ret i1 [[TOBOOL_I]]
 //
@@ -1793,7 +1797,7 @@ bool test_i32x4_all_true(v128_t a) {
 
 // CHECK-LABEL: @test_i32x4_bitmask(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call i32 @llvm.wasm.bitmask.v4i32(<4 x i32> [[A:%.*]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call i32 @llvm.wasm.bitmask.v4i32(<4 x i32> [[A:%.*]]) #[[ATTR8]]
 // CHECK-NEXT:    ret i32 [[TMP0]]
 //
 int32_t test_i32x4_bitmask(v128_t a) {
@@ -1904,7 +1908,7 @@ v128_t test_u32x4_max(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x i32> @llvm.wasm.dot(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x i32> @llvm.wasm.dot(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
 v128_t test_i32x4_dot_i16x8(v128_t a, v128_t b) {
@@ -1914,7 +1918,7 @@ v128_t test_i32x4_dot_i16x8(v128_t a, v128_t b) {
 // CHECK-LABEL: @test_i64x2_abs(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x i64>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x i64> @llvm.abs.v2i64(<2 x i64> [[TMP0]], i1 false) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x i64> @llvm.abs.v2i64(<2 x i64> [[TMP0]], i1 false) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <2 x i64> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -1936,7 +1940,7 @@ v128_t test_i64x2_neg(v128_t a) {
 // CHECK-LABEL: @test_i64x2_all_true(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x i64>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call i32 @llvm.wasm.alltrue.v2i64(<2 x i64> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call i32 @llvm.wasm.alltrue.v2i64(<2 x i64> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TOBOOL_I:%.*]] = icmp ne i32 [[TMP1]], 0
 // CHECK-NEXT:    ret i1 [[TOBOOL_I]]
 //
@@ -1947,7 +1951,7 @@ bool test_i64x2_all_true(v128_t a) {
 // CHECK-LABEL: @test_i64x2_bitmask(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x i64>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call i32 @llvm.wasm.bitmask.v2i64(<2 x i64> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call i32 @llvm.wasm.bitmask.v2i64(<2 x i64> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    ret i32 [[TMP1]]
 //
 int32_t test_i64x2_bitmask(v128_t a) {
@@ -2035,7 +2039,7 @@ v128_t test_i64x2_mul(v128_t a, v128_t b) {
 // CHECK-LABEL: @test_f32x4_abs(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <4 x float>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.fabs.v4f32(<4 x float> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.fabs.v4f32(<4 x float> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <4 x float> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2057,7 +2061,7 @@ v128_t test_f32x4_neg(v128_t a) {
 // CHECK-LABEL: @test_f32x4_sqrt(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <4 x float>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.sqrt.v4f32(<4 x float> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.sqrt.v4f32(<4 x float> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <4 x float> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2068,7 +2072,7 @@ v128_t test_f32x4_sqrt(v128_t a) {
 // CHECK-LABEL: @test_f32x4_ceil(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <4 x float>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.ceil.v4f32(<4 x float> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.ceil.v4f32(<4 x float> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <4 x float> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2079,7 +2083,7 @@ v128_t test_f32x4_ceil(v128_t a) {
 // CHECK-LABEL: @test_f32x4_floor(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <4 x float>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.floor.v4f32(<4 x float> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.floor.v4f32(<4 x float> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <4 x float> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2090,7 +2094,7 @@ v128_t test_f32x4_floor(v128_t a) {
 // CHECK-LABEL: @test_f32x4_trunc(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <4 x float>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.trunc.v4f32(<4 x float> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.trunc.v4f32(<4 x float> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <4 x float> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2101,7 +2105,7 @@ v128_t test_f32x4_trunc(v128_t a) {
 // CHECK-LABEL: @test_f32x4_nearest(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <4 x float>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.nearbyint.v4f32(<4 x float> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.nearbyint.v4f32(<4 x float> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <4 x float> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2161,7 +2165,7 @@ v128_t test_f32x4_div(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <4 x float>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <4 x float>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x float> @llvm.minimum.v4f32(<4 x float> [[TMP0]], <4 x float> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x float> @llvm.minimum.v4f32(<4 x float> [[TMP0]], <4 x float> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <4 x float> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2173,7 +2177,7 @@ v128_t test_f32x4_min(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <4 x float>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <4 x float>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x float> @llvm.maximum.v4f32(<4 x float> [[TMP0]], <4 x float> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x float> @llvm.maximum.v4f32(<4 x float> [[TMP0]], <4 x float> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <4 x float> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2185,7 +2189,7 @@ v128_t test_f32x4_max(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <4 x float>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <4 x float>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x float> @llvm.wasm.pmin.v4f32(<4 x float> [[TMP0]], <4 x float> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x float> @llvm.wasm.pmin.v4f32(<4 x float> [[TMP0]], <4 x float> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <4 x float> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2197,7 +2201,7 @@ v128_t test_f32x4_pmin(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <4 x float>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <4 x float>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x float> @llvm.wasm.pmax.v4f32(<4 x float> [[TMP0]], <4 x float> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x float> @llvm.wasm.pmax.v4f32(<4 x float> [[TMP0]], <4 x float> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <4 x float> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2208,7 +2212,7 @@ v128_t test_f32x4_pmax(v128_t a, v128_t b) {
 // CHECK-LABEL: @test_f64x2_abs(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x double>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.fabs.v2f64(<2 x double> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.fabs.v2f64(<2 x double> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <2 x double> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2230,7 +2234,7 @@ v128_t test_f64x2_neg(v128_t a) {
 // CHECK-LABEL: @test_f64x2_sqrt(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x double>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.sqrt.v2f64(<2 x double> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.sqrt.v2f64(<2 x double> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <2 x double> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2241,7 +2245,7 @@ v128_t test_f64x2_sqrt(v128_t a) {
 // CHECK-LABEL: @test_f64x2_ceil(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x double>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.ceil.v2f64(<2 x double> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.ceil.v2f64(<2 x double> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <2 x double> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2252,7 +2256,7 @@ v128_t test_f64x2_ceil(v128_t a) {
 // CHECK-LABEL: @test_f64x2_floor(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x double>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.floor.v2f64(<2 x double> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.floor.v2f64(<2 x double> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <2 x double> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2263,7 +2267,7 @@ v128_t test_f64x2_floor(v128_t a) {
 // CHECK-LABEL: @test_f64x2_trunc(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x double>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.trunc.v2f64(<2 x double> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.trunc.v2f64(<2 x double> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <2 x double> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2274,7 +2278,7 @@ v128_t test_f64x2_trunc(v128_t a) {
 // CHECK-LABEL: @test_f64x2_nearest(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x double>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.nearbyint.v2f64(<2 x double> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.nearbyint.v2f64(<2 x double> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <2 x double> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2334,7 +2338,7 @@ v128_t test_f64x2_div(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x double>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <2 x double>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <2 x double> @llvm.minimum.v2f64(<2 x double> [[TMP0]], <2 x double> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <2 x double> @llvm.minimum.v2f64(<2 x double> [[TMP0]], <2 x double> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <2 x double> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2346,7 +2350,7 @@ v128_t test_f64x2_min(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x double>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <2 x double>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <2 x double> @llvm.maximum.v2f64(<2 x double> [[TMP0]], <2 x double> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <2 x double> @llvm.maximum.v2f64(<2 x double> [[TMP0]], <2 x double> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <2 x double> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2358,7 +2362,7 @@ v128_t test_f64x2_max(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x double>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <2 x double>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <2 x double> @llvm.wasm.pmin.v2f64(<2 x double> [[TMP0]], <2 x double> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <2 x double> @llvm.wasm.pmin.v2f64(<2 x double> [[TMP0]], <2 x double> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <2 x double> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2370,7 +2374,7 @@ v128_t test_f64x2_pmin(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x double>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <2 x double>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <2 x double> @llvm.wasm.pmax.v2f64(<2 x double> [[TMP0]], <2 x double> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <2 x double> @llvm.wasm.pmax.v2f64(<2 x double> [[TMP0]], <2 x double> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <2 x double> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2381,7 +2385,7 @@ v128_t test_f64x2_pmax(v128_t a, v128_t b) {
 // CHECK-LABEL: @test_i32x4_trunc_sat_f32x4(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <4 x float>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x i32> @llvm.fptosi.sat.v4i32.v4f32(<4 x float> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x i32> @llvm.fptosi.sat.v4i32.v4f32(<4 x float> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    ret <4 x i32> [[TMP1]]
 //
 v128_t test_i32x4_trunc_sat_f32x4(v128_t a) {
@@ -2391,7 +2395,7 @@ v128_t test_i32x4_trunc_sat_f32x4(v128_t a) {
 // CHECK-LABEL: @test_u32x4_trunc_sat_f32x4(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <4 x float>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x i32> @llvm.fptoui.sat.v4i32.v4f32(<4 x float> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x i32> @llvm.fptoui.sat.v4i32.v4f32(<4 x float> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    ret <4 x i32> [[TMP1]]
 //
 v128_t test_u32x4_trunc_sat_f32x4(v128_t a) {
@@ -2443,7 +2447,7 @@ v128_t test_f64x2_convert_low_u32x4(v128_t a) {
 // CHECK-LABEL: @test_i32x4_trunc_sat_f64x2_zero(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x double>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x i32> @llvm.fptosi.sat.v2i32.v2f64(<2 x double> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x i32> @llvm.fptosi.sat.v2i32.v2f64(<2 x double> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2454,7 +2458,7 @@ v128_t test_i32x4_trunc_sat_f64x2_zero(v128_t a) {
 // CHECK-LABEL: @test_u32x4_trunc_sat_f64x2_zero(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <2 x double>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x i32> @llvm.fptoui.sat.v2i32.v2f64(<2 x double> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x i32> @llvm.fptoui.sat.v2i32.v2f64(<2 x double> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2538,7 +2542,7 @@ v128_t test_i64x2_shuffle(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.wasm.swizzle(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.wasm.swizzle(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <16 x i8> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2550,7 +2554,7 @@ v128_t test_i8x16_swizzle(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.wasm.narrow.signed.v16i8.v8i16(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.wasm.narrow.signed.v16i8.v8i16(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <16 x i8> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2562,7 +2566,7 @@ v128_t test_i8x16_narrow_i16x8(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.wasm.narrow.unsigned.v16i8.v8i16(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i8> @llvm.wasm.narrow.unsigned.v16i8.v8i16(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <16 x i8> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2572,7 +2576,7 @@ v128_t test_u8x16_narrow_i16x8(v128_t a, v128_t b) {
 
 // CHECK-LABEL: @test_i16x8_narrow_i32x4(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <8 x i16> @llvm.wasm.narrow.signed.v8i16.v4i32(<4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <8 x i16> @llvm.wasm.narrow.signed.v8i16.v4i32(<4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <8 x i16> [[TMP0]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP1]]
 //
@@ -2582,7 +2586,7 @@ v128_t test_i16x8_narrow_i32x4(v128_t a, v128_t b) {
 
 // CHECK-LABEL: @test_u16x8_narrow_i32x4(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <8 x i16> @llvm.wasm.narrow.unsigned.v8i16.v4i32(<4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <8 x i16> @llvm.wasm.narrow.unsigned.v8i16.v4i32(<4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <8 x i16> [[TMP0]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP1]]
 //
@@ -2729,7 +2733,7 @@ v128_t test_u64x2_extend_high_u32x4(v128_t a) {
 // CHECK-LABEL: @test_i16x8_extadd_pairwise_i8x16(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <8 x i16> @llvm.wasm.extadd.pairwise.signed.v8i16(<16 x i8> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <8 x i16> @llvm.wasm.extadd.pairwise.signed.v8i16(<16 x i8> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <8 x i16> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2740,7 +2744,7 @@ v128_t test_i16x8_extadd_pairwise_i8x16(v128_t a) {
 // CHECK-LABEL: @test_u16x8_extadd_pairwise_u8x16(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <8 x i16> @llvm.wasm.extadd.pairwise.unsigned.v8i16(<16 x i8> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <8 x i16> @llvm.wasm.extadd.pairwise.unsigned.v8i16(<16 x i8> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP2:%.*]] = bitcast <8 x i16> [[TMP1]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
@@ -2751,7 +2755,7 @@ v128_t test_u16x8_extadd_pairwise_u8x16(v128_t a) {
 // CHECK-LABEL: @test_i32x4_extadd_pairwise_i16x8(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x i32> @llvm.wasm.extadd.pairwise.signed.v4i32(<8 x i16> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x i32> @llvm.wasm.extadd.pairwise.signed.v4i32(<8 x i16> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    ret <4 x i32> [[TMP1]]
 //
 v128_t test_i32x4_extadd_pairwise_i16x8(v128_t a) {
@@ -2761,7 +2765,7 @@ v128_t test_i32x4_extadd_pairwise_i16x8(v128_t a) {
 // CHECK-LABEL: @test_u32x4_extadd_pairwise_u16x8(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x i32> @llvm.wasm.extadd.pairwise.unsigned.v4i32(<8 x i16> [[TMP0]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x i32> @llvm.wasm.extadd.pairwise.unsigned.v4i32(<8 x i16> [[TMP0]]) #[[ATTR8]]
 // CHECK-NEXT:    ret <4 x i32> [[TMP1]]
 //
 v128_t test_u32x4_extadd_pairwise_u16x8(v128_t a) {
@@ -2772,7 +2776,7 @@ v128_t test_u32x4_extadd_pairwise_u16x8(v128_t a) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.extmul.low.signed.v8i16(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.extmul.low.signed.v8i16(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <8 x i16> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2784,7 +2788,7 @@ v128_t test_i16x8_extmul_low_i8x16(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.extmul.high.signed.v8i16(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.extmul.high.signed.v8i16(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <8 x i16> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2796,7 +2800,7 @@ v128_t test_i16x8_extmul_high_i8x16(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.extmul.low.unsigned.v8i16(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.extmul.low.unsigned.v8i16(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <8 x i16> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2808,7 +2812,7 @@ v128_t test_u16x8_extmul_low_u8x16(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <16 x i8>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <16 x i8>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.extmul.high.unsigned.v8i16(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.extmul.high.unsigned.v8i16(<16 x i8> [[TMP0]], <16 x i8> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <8 x i16> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //
@@ -2820,7 +2824,7 @@ v128_t test_u16x8_extmul_high_u8x16(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x i32> @llvm.wasm.extmul.low.signed.v4i32(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x i32> @llvm.wasm.extmul.low.signed.v4i32(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
 v128_t test_i32x4_extmul_low_i16x8(v128_t a, v128_t b) {
@@ -2831,7 +2835,7 @@ v128_t test_i32x4_extmul_low_i16x8(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x i32> @llvm.wasm.extmul.high.signed.v4i32(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x i32> @llvm.wasm.extmul.high.signed.v4i32(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
 v128_t test_i32x4_extmul_high_i16x8(v128_t a, v128_t b) {
@@ -2842,7 +2846,7 @@ v128_t test_i32x4_extmul_high_i16x8(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x i32> @llvm.wasm.extmul.low.unsigned.v4i32(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x i32> @llvm.wasm.extmul.low.unsigned.v4i32(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
 v128_t test_u32x4_extmul_low_u16x8(v128_t a, v128_t b) {
@@ -2853,7 +2857,7 @@ v128_t test_u32x4_extmul_low_u16x8(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x i32> @llvm.wasm.extmul.high.unsigned.v4i32(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <4 x i32> @llvm.wasm.extmul.high.unsigned.v4i32(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    ret <4 x i32> [[TMP2]]
 //
 v128_t test_u32x4_extmul_high_u16x8(v128_t a, v128_t b) {
@@ -2862,7 +2866,7 @@ v128_t test_u32x4_extmul_high_u16x8(v128_t a, v128_t b) {
 
 // CHECK-LABEL: @test_i64x2_extmul_low_i32x4(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <2 x i64> @llvm.wasm.extmul.low.signed.v2i64(<4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <2 x i64> @llvm.wasm.extmul.low.signed.v2i64(<4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <2 x i64> [[TMP0]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP1]]
 //
@@ -2872,7 +2876,7 @@ v128_t test_i64x2_extmul_low_i32x4(v128_t a, v128_t b) {
 
 // CHECK-LABEL: @test_i64x2_extmul_high_i32x4(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <2 x i64> @llvm.wasm.extmul.high.signed.v2i64(<4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <2 x i64> @llvm.wasm.extmul.high.signed.v2i64(<4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <2 x i64> [[TMP0]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP1]]
 //
@@ -2882,7 +2886,7 @@ v128_t test_i64x2_extmul_high_i32x4(v128_t a, v128_t b) {
 
 // CHECK-LABEL: @test_u64x2_extmul_low_u32x4(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <2 x i64> @llvm.wasm.extmul.low.unsigned.v2i64(<4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <2 x i64> @llvm.wasm.extmul.low.unsigned.v2i64(<4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <2 x i64> [[TMP0]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP1]]
 //
@@ -2892,7 +2896,7 @@ v128_t test_u64x2_extmul_low_u32x4(v128_t a, v128_t b) {
 
 // CHECK-LABEL: @test_u64x2_extmul_high_u32x4(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <2 x i64> @llvm.wasm.extmul.high.unsigned.v2i64(<4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <2 x i64> @llvm.wasm.extmul.high.unsigned.v2i64(<4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <2 x i64> [[TMP0]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP1]]
 //
@@ -2904,7 +2908,7 @@ v128_t test_u64x2_extmul_high_u32x4(v128_t a, v128_t b) {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast <4 x i32> [[A:%.*]] to <8 x i16>
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast <4 x i32> [[B:%.*]] to <8 x i16>
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.q15mulr.sat.signed(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR10]]
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <8 x i16> @llvm.wasm.q15mulr.sat.signed(<8 x i16> [[TMP0]], <8 x i16> [[TMP1]]) #[[ATTR8]]
 // CHECK-NEXT:    [[TMP3:%.*]] = bitcast <8 x i16> [[TMP2]] to <4 x i32>
 // CHECK-NEXT:    ret <4 x i32> [[TMP3]]
 //

diff --git a/llvm/include/llvm/IR/IntrinsicsWebAssembly.td b/llvm/include/llvm/IR/IntrinsicsWebAssembly.td
index 17b97b0..d5398a9 100644 (file)
--- a/llvm/include/llvm/IR/IntrinsicsWebAssembly.td
+++ b/llvm/include/llvm/IR/IntrinsicsWebAssembly.td
@@ -191,26 +191,6 @@ def int_wasm_load64_zero :
 // ISD::TargetConstant, which would require extra complications in the ISel
 // tablegen patterns. TODO: Replace these intrinsic with normal ISel patterns
 // once the load_lane instructions are merged to the proposal.
-def int_wasm_load8_lane :
-  Intrinsic<[llvm_v16i8_ty],
-            [LLVMPointerType<llvm_i8_ty>, llvm_v16i8_ty, llvm_i32_ty],
-            [IntrReadMem, IntrArgMemOnly],
-            "", [SDNPMemOperand]>;
-def int_wasm_load16_lane :
-  Intrinsic<[llvm_v8i16_ty],
-            [LLVMPointerType<llvm_i16_ty>, llvm_v8i16_ty, llvm_i32_ty],
-            [IntrReadMem, IntrArgMemOnly],
-            "", [SDNPMemOperand]>;
-def int_wasm_load32_lane :
-  Intrinsic<[llvm_v4i32_ty],
-            [LLVMPointerType<llvm_i32_ty>, llvm_v4i32_ty, llvm_i32_ty],
-            [IntrReadMem, IntrArgMemOnly],
-            "", [SDNPMemOperand]>;
-def int_wasm_load64_lane :
-  Intrinsic<[llvm_v2i64_ty],
-            [LLVMPointerType<llvm_i64_ty>, llvm_v2i64_ty, llvm_i32_ty],
-            [IntrReadMem, IntrArgMemOnly],
-            "", [SDNPMemOperand]>;
 def int_wasm_store8_lane :
   Intrinsic<[],
             [LLVMPointerType<llvm_i8_ty>, llvm_v16i8_ty, llvm_i32_ty],

diff --git a/llvm/lib/Target/WebAssembly/WebAssemblyISelLowering.cpp b/llvm/lib/Target/WebAssembly/WebAssemblyISelLowering.cpp
index 1abbc8d..19dffa8 100644 (file)
--- a/llvm/lib/Target/WebAssembly/WebAssemblyISelLowering.cpp
+++ b/llvm/lib/Target/WebAssembly/WebAssemblyISelLowering.cpp
@@ -767,49 +767,33 @@ bool WebAssemblyTargetLowering::getTgtMemIntrinsic(IntrinsicInfo &Info,
     Info.align = Align(1);
     Info.flags = MachineMemOperand::MOLoad;
     return true;
-  case Intrinsic::wasm_load8_lane:
-  case Intrinsic::wasm_load16_lane:
-  case Intrinsic::wasm_load32_lane:
-  case Intrinsic::wasm_load64_lane:
   case Intrinsic::wasm_store8_lane:
   case Intrinsic::wasm_store16_lane:
   case Intrinsic::wasm_store32_lane:
   case Intrinsic::wasm_store64_lane: {
     MVT MemVT;
     switch (Intrinsic) {
-    case Intrinsic::wasm_load8_lane:
     case Intrinsic::wasm_store8_lane:
       MemVT = MVT::i8;
       break;
-    case Intrinsic::wasm_load16_lane:
     case Intrinsic::wasm_store16_lane:
       MemVT = MVT::i16;
       break;
-    case Intrinsic::wasm_load32_lane:
     case Intrinsic::wasm_store32_lane:
       MemVT = MVT::i32;
       break;
-    case Intrinsic::wasm_load64_lane:
     case Intrinsic::wasm_store64_lane:
       MemVT = MVT::i64;
       break;
     default:
       llvm_unreachable("unexpected intrinsic");
     }
-    if (Intrinsic == Intrinsic::wasm_load8_lane ||
-        Intrinsic == Intrinsic::wasm_load16_lane ||
-        Intrinsic == Intrinsic::wasm_load32_lane ||
-        Intrinsic == Intrinsic::wasm_load64_lane) {
-      Info.opc = ISD::INTRINSIC_W_CHAIN;
-      Info.flags = MachineMemOperand::MOLoad;
-    } else {
-      Info.opc = ISD::INTRINSIC_VOID;
-      Info.flags = MachineMemOperand::MOStore;
-    }
-    Info.ptrVal = I.getArgOperand(0);
+    Info.opc = ISD::INTRINSIC_VOID;
     Info.memVT = MemVT;
+    Info.ptrVal = I.getArgOperand(0);
     Info.offset = 0;
     Info.align = Align(1);
+    Info.flags = MachineMemOperand::MOStore;
     return true;
   }
   default:

diff --git a/llvm/lib/Target/WebAssembly/WebAssemblyInstrSIMD.td b/llvm/lib/Target/WebAssembly/WebAssemblyInstrSIMD.td
index bdffc1a..97c3ece 100644 (file)
--- a/llvm/lib/Target/WebAssembly/WebAssemblyInstrSIMD.td
+++ b/llvm/lib/Target/WebAssembly/WebAssemblyInstrSIMD.td
@@ -300,8 +300,6 @@ multiclass SIMDLoadLane<Vec vec, bits<32> simdop> {
   } // mayLoad = 1, UseNamedOperandTable = 1
 }
 
-// TODO: Also support v4f32 and v2f64 once the instructions are merged
-// to the proposal
 defm "" : SIMDLoadLane<I8x16, 0x54>;
 defm "" : SIMDLoadLane<I16x8, 0x55>;
 defm "" : SIMDLoadLane<I32x4, 0x56>;
@@ -321,10 +319,24 @@ multiclass LoadLanePatNoOffset<Vec vec, SDPatternOperator kind> {
         Requires<[HasAddr64]>;
 }
 
-defm : LoadLanePatNoOffset<I8x16, int_wasm_load8_lane>;
-defm : LoadLanePatNoOffset<I16x8, int_wasm_load16_lane>;
-defm : LoadLanePatNoOffset<I32x4, int_wasm_load32_lane>;
-defm : LoadLanePatNoOffset<I64x2, int_wasm_load64_lane>;
+def load8_lane :
+  PatFrag<(ops node:$ptr, node:$vec, node:$idx),
+          (vector_insert $vec, (i32 (extloadi8 $ptr)), $idx)>;
+def load16_lane :
+  PatFrag<(ops node:$ptr, node:$vec, node:$idx),
+          (vector_insert $vec, (i32 (extloadi16 $ptr)), $idx)>;
+def load32_lane :
+  PatFrag<(ops node:$ptr, node:$vec, node:$idx),
+          (vector_insert $vec, (i32 (load $ptr)), $idx)>;
+def load64_lane :
+  PatFrag<(ops node:$ptr, node:$vec, node:$idx),
+          (vector_insert $vec, (i64 (load $ptr)), $idx)>;
+// TODO: floating point lanes as well
+
+defm : LoadLanePatNoOffset<I8x16, load8_lane>;
+defm : LoadLanePatNoOffset<I16x8, load16_lane>;
+defm : LoadLanePatNoOffset<I32x4, load32_lane>;
+defm : LoadLanePatNoOffset<I64x2, load64_lane>;
 
 // TODO: Also support the other load patterns for load_lane once the instructions
 // are merged to the proposal.

diff --git a/llvm/test/CodeGen/WebAssembly/simd-build-vector.ll b/llvm/test/CodeGen/WebAssembly/simd-build-vector.ll
index 4fc5481..c544594 100644 (file)
--- a/llvm/test/CodeGen/WebAssembly/simd-build-vector.ll
+++ b/llvm/test/CodeGen/WebAssembly/simd-build-vector.ll
@@ -211,7 +211,7 @@ define <16 x i8> @mashup_swizzle_i8x16(<16 x i8> %src, <16 x i8> %mask, i8 %spla
 ; CHECK-LABEL: mashup_const_i8x16:
 ; CHECK-NEXT:  .functype       mashup_const_i8x16 (v128, v128, i32) -> (v128)
 ; CHECK:       v128.const      $push[[L0:[0-9]+]]=, 0, 0, 0, 0, 42, 0, 0, 0, 0, 0, 0, 0, 0, 0, 42, 0
-; CHECK:       i8x16.replace_lane
+; CHECK:       v128.load8_lane
 ; CHECK:       i8x16.replace_lane
 ; CHECK:       i8x16.replace_lane
 ; CHECK:       return
@@ -234,7 +234,7 @@ define <16 x i8> @mashup_const_i8x16(<16 x i8> %src, <16 x i8> %mask, i8 %splatt
 ; CHECK-LABEL: mashup_splat_i8x16:
 ; CHECK-NEXT:  .functype       mashup_splat_i8x16 (v128, v128, i32) -> (v128)
 ; CHECK:       i8x16.splat     $push[[L0:[0-9]+]]=, $2
-; CHECK:       i8x16.replace_lane
+; CHECK:       v128.load8_lane
 ; CHECK:       i8x16.replace_lane
 ; CHECK:       return
 define <16 x i8> @mashup_splat_i8x16(<16 x i8> %src, <16 x i8> %mask, i8 %splatted) {

diff --git a/llvm/test/CodeGen/WebAssembly/simd-load-lane-offset.ll b/llvm/test/CodeGen/WebAssembly/simd-load-lane-offset.ll
index 26ac45e..fdb2328 100644 (file)
--- a/llvm/test/CodeGen/WebAssembly/simd-load-lane-offset.ll
+++ b/llvm/test/CodeGen/WebAssembly/simd-load-lane-offset.ll
@@ -8,11 +8,6 @@
 
 target triple = "wasm32-unknown-unknown"
 
-declare <16 x i8> @llvm.wasm.load8.lane(i8*, <16 x i8>, i32)
-declare <8 x i16> @llvm.wasm.load16.lane(i16*, <8 x i16>, i32)
-declare <4 x i32> @llvm.wasm.load32.lane(i32*, <4 x i32>, i32)
-declare <2 x i64> @llvm.wasm.load64.lane(i64*, <2 x i64>, i32)
-
 declare void @llvm.wasm.store8.lane(i8*, <16 x i8>, i32)
 declare void @llvm.wasm.store16.lane(i16*, <8 x i16>, i32)
 declare void @llvm.wasm.store32.lane(i32*, <4 x i32>, i32)
@@ -30,7 +25,8 @@ define <16 x i8> @load_lane_i8_no_offset(i8* %p, <16 x i8> %v) {
 ; CHECK-NEXT:    local.get 1
 ; CHECK-NEXT:    v128.load8_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
-  %t = tail call <16 x i8> @llvm.wasm.load8.lane(i8* %p, <16 x i8> %v, i32 0)
+  %x = load i8, i8* %p
+  %t = insertelement <16 x i8> %v, i8 %x, i32 0
   ret <16 x i8> %t
 }
 
@@ -47,7 +43,8 @@ define <16 x i8> @load_lane_i8_with_folded_offset(i8* %p, <16 x i8> %v) {
   %q = ptrtoint i8* %p to i32
   %r = add nuw i32 %q, 24
   %s = inttoptr i32 %r to i8*
-  %t = tail call <16 x i8> @llvm.wasm.load8.lane(i8* %s, <16 x i8> %v, i32 0)
+  %x = load i8, i8* %s
+  %t = insertelement <16 x i8> %v, i8 %x, i32 0
   ret <16 x i8> %t
 }
 
@@ -62,7 +59,8 @@ define <16 x i8> @load_lane_i8_with_folded_gep_offset(i8* %p, <16 x i8> %v) {
 ; CHECK-NEXT:    v128.load8_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = getelementptr inbounds i8, i8* %p, i32 6
-  %t = tail call <16 x i8> @llvm.wasm.load8.lane(i8* %s, <16 x i8> %v, i32 0)
+  %x = load i8, i8* %s
+  %t = insertelement <16 x i8> %v, i8 %x, i32 0
   ret <16 x i8> %t
 }
 
@@ -77,7 +75,8 @@ define <16 x i8> @load_lane_i8_with_unfolded_gep_negative_offset(i8* %p, <16 x i
 ; CHECK-NEXT:    v128.load8_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = getelementptr inbounds i8, i8* %p, i32 -6
-  %t = tail call <16 x i8> @llvm.wasm.load8.lane(i8* %s, <16 x i8> %v, i32 0)
+  %x = load i8, i8* %s
+  %t = insertelement <16 x i8> %v, i8 %x, i32 0
   ret <16 x i8> %t
 }
 
@@ -94,7 +93,8 @@ define <16 x i8> @load_lane_i8_with_unfolded_offset(i8* %p, <16 x i8> %v) {
   %q = ptrtoint i8* %p to i32
   %r = add nsw i32 %q, 24
   %s = inttoptr i32 %r to i8*
-  %t = tail call <16 x i8> @llvm.wasm.load8.lane(i8* %s, <16 x i8> %v, i32 0)
+  %x = load i8, i8* %s
+  %t = insertelement <16 x i8> %v, i8 %x, i32 0
   ret <16 x i8> %t
 }
 
@@ -109,7 +109,8 @@ define <16 x i8> @load_lane_i8_with_unfolded_gep_offset(i8* %p, <16 x i8> %v) {
 ; CHECK-NEXT:    v128.load8_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = getelementptr i8, i8* %p, i32 6
-  %t = tail call <16 x i8> @llvm.wasm.load8.lane(i8* %s, <16 x i8> %v, i32 0)
+  %x = load i8, i8* %s
+  %t = insertelement <16 x i8> %v, i8 %x, i32 0
   ret <16 x i8> %t
 }
 
@@ -122,7 +123,8 @@ define <16 x i8> @load_lane_i8_from_numeric_address(<16 x i8> %v) {
 ; CHECK-NEXT:    v128.load8_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = inttoptr i32 42 to i8*
-  %t = tail call <16 x i8> @llvm.wasm.load8.lane(i8* %s, <16 x i8> %v, i32 0)
+  %x = load i8, i8* %s
+  %t = insertelement <16 x i8> %v, i8 %x, i32 0
   ret <16 x i8> %t
 }
 
@@ -135,7 +137,8 @@ define <16 x i8> @load_lane_i8_from_global_address(<16 x i8> %v) {
 ; CHECK-NEXT:    local.get 0
 ; CHECK-NEXT:    v128.load8_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
-  %t = tail call <16 x i8> @llvm.wasm.load8.lane(i8* @gv_i8, <16 x i8> %v, i32 0)
+  %x = load i8, i8* @gv_i8
+  %t = insertelement <16 x i8> %v, i8 %x, i32 0
   ret <16 x i8> %t
 }
 
@@ -265,9 +268,10 @@ define <8 x i16> @load_lane_i16_no_offset(i16* %p, <8 x i16> %v) {
 ; CHECK-NEXT:  # %bb.0:
 ; CHECK-NEXT:    local.get 0
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load16_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load16_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
-  %t = tail call <8 x i16> @llvm.wasm.load16.lane(i16* %p, <8 x i16> %v, i32 0)
+  %x = load i16, i16* %p
+  %t = insertelement <8 x i16> %v, i16 %x, i32 0
   ret <8 x i16> %t
 }
 
@@ -279,12 +283,13 @@ define <8 x i16> @load_lane_i16_with_folded_offset(i16* %p, <8 x i16> %v) {
 ; CHECK-NEXT:    i32.const 24
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load16_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load16_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %q = ptrtoint i16* %p to i32
   %r = add nuw i32 %q, 24
   %s = inttoptr i32 %r to i16*
-  %t = tail call <8 x i16> @llvm.wasm.load16.lane(i16* %s, <8 x i16> %v, i32 0)
+  %x = load i16, i16* %s
+  %t = insertelement <8 x i16> %v, i16 %x, i32 0
   ret <8 x i16> %t
 }
 
@@ -296,10 +301,11 @@ define <8 x i16> @load_lane_i16_with_folded_gep_offset(i16* %p, <8 x i16> %v) {
 ; CHECK-NEXT:    i32.const 12
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load16_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load16_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = getelementptr inbounds i16, i16* %p, i32 6
-  %t = tail call <8 x i16> @llvm.wasm.load16.lane(i16* %s, <8 x i16> %v, i32 0)
+  %x = load i16, i16* %s
+  %t = insertelement <8 x i16> %v, i16 %x, i32 0
   ret <8 x i16> %t
 }
 
@@ -311,10 +317,11 @@ define <8 x i16> @load_lane_i16_with_unfolded_gep_negative_offset(i16* %p, <8 x
 ; CHECK-NEXT:    i32.const -12
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load16_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load16_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = getelementptr inbounds i16, i16* %p, i32 -6
-  %t = tail call <8 x i16> @llvm.wasm.load16.lane(i16* %s, <8 x i16> %v, i32 0)
+  %x = load i16, i16* %s
+  %t = insertelement <8 x i16> %v, i16 %x, i32 0
   ret <8 x i16> %t
 }
 
@@ -326,12 +333,13 @@ define <8 x i16> @load_lane_i16_with_unfolded_offset(i16* %p, <8 x i16> %v) {
 ; CHECK-NEXT:    i32.const 24
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load16_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load16_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %q = ptrtoint i16* %p to i32
   %r = add nsw i32 %q, 24
   %s = inttoptr i32 %r to i16*
-  %t = tail call <8 x i16> @llvm.wasm.load16.lane(i16* %s, <8 x i16> %v, i32 0)
+  %x = load i16, i16* %s
+  %t = insertelement <8 x i16> %v, i16 %x, i32 0
   ret <8 x i16> %t
 }
 
@@ -343,10 +351,11 @@ define <8 x i16> @load_lane_i16_with_unfolded_gep_offset(i16* %p, <8 x i16> %v)
 ; CHECK-NEXT:    i32.const 12
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load16_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load16_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = getelementptr i16, i16* %p, i32 6
-  %t = tail call <8 x i16> @llvm.wasm.load16.lane(i16* %s, <8 x i16> %v, i32 0)
+  %x = load i16, i16* %s
+  %t = insertelement <8 x i16> %v, i16 %x, i32 0
   ret <8 x i16> %t
 }
 
@@ -356,10 +365,11 @@ define <8 x i16> @load_lane_i16_from_numeric_address(<8 x i16> %v) {
 ; CHECK-NEXT:  # %bb.0:
 ; CHECK-NEXT:    i32.const 42
 ; CHECK-NEXT:    local.get 0
-; CHECK-NEXT:    v128.load16_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load16_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = inttoptr i32 42 to i16*
-  %t = tail call <8 x i16> @llvm.wasm.load16.lane(i16* %s, <8 x i16> %v, i32 0)
+  %x = load i16, i16* %s
+  %t = insertelement <8 x i16> %v, i16 %x, i32 0
   ret <8 x i16> %t
 }
 
@@ -370,9 +380,10 @@ define <8 x i16> @load_lane_i16_from_global_address(<8 x i16> %v) {
 ; CHECK-NEXT:  # %bb.0:
 ; CHECK-NEXT:    i32.const gv_i16
 ; CHECK-NEXT:    local.get 0
-; CHECK-NEXT:    v128.load16_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load16_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
-  %t = tail call <8 x i16> @llvm.wasm.load16.lane(i16* @gv_i16, <8 x i16> %v, i32 0)
+  %x = load i16, i16* @gv_i16
+  %t = insertelement <8 x i16> %v, i16 %x, i32 0
   ret <8 x i16> %t
 }
 
@@ -502,9 +513,10 @@ define <4 x i32> @load_lane_i32_no_offset(i32* %p, <4 x i32> %v) {
 ; CHECK-NEXT:  # %bb.0:
 ; CHECK-NEXT:    local.get 0
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load32_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load32_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
-  %t = tail call <4 x i32> @llvm.wasm.load32.lane(i32* %p, <4 x i32> %v, i32 0)
+  %x = load i32, i32* %p
+  %t = insertelement <4 x i32> %v, i32 %x, i32 0
   ret <4 x i32> %t
 }
 
@@ -516,12 +528,13 @@ define <4 x i32> @load_lane_i32_with_folded_offset(i32* %p, <4 x i32> %v) {
 ; CHECK-NEXT:    i32.const 24
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load32_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load32_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %q = ptrtoint i32* %p to i32
   %r = add nuw i32 %q, 24
   %s = inttoptr i32 %r to i32*
-  %t = tail call <4 x i32> @llvm.wasm.load32.lane(i32* %s, <4 x i32> %v, i32 0)
+  %x = load i32, i32* %s
+  %t = insertelement <4 x i32> %v, i32 %x, i32 0
   ret <4 x i32> %t
 }
 
@@ -533,10 +546,11 @@ define <4 x i32> @load_lane_i32_with_folded_gep_offset(i32* %p, <4 x i32> %v) {
 ; CHECK-NEXT:    i32.const 24
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load32_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load32_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = getelementptr inbounds i32, i32* %p, i32 6
-  %t = tail call <4 x i32> @llvm.wasm.load32.lane(i32* %s, <4 x i32> %v, i32 0)
+  %x = load i32, i32* %s
+  %t = insertelement <4 x i32> %v, i32 %x, i32 0
   ret <4 x i32> %t
 }
 
@@ -548,10 +562,11 @@ define <4 x i32> @load_lane_i32_with_unfolded_gep_negative_offset(i32* %p, <4 x
 ; CHECK-NEXT:    i32.const -24
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load32_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load32_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = getelementptr inbounds i32, i32* %p, i32 -6
-  %t = tail call <4 x i32> @llvm.wasm.load32.lane(i32* %s, <4 x i32> %v, i32 0)
+  %x = load i32, i32* %s
+  %t = insertelement <4 x i32> %v, i32 %x, i32 0
   ret <4 x i32> %t
 }
 
@@ -563,12 +578,13 @@ define <4 x i32> @load_lane_i32_with_unfolded_offset(i32* %p, <4 x i32> %v) {
 ; CHECK-NEXT:    i32.const 24
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load32_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load32_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %q = ptrtoint i32* %p to i32
   %r = add nsw i32 %q, 24
   %s = inttoptr i32 %r to i32*
-  %t = tail call <4 x i32> @llvm.wasm.load32.lane(i32* %s, <4 x i32> %v, i32 0)
+  %x = load i32, i32* %s
+  %t = insertelement <4 x i32> %v, i32 %x, i32 0
   ret <4 x i32> %t
 }
 
@@ -580,10 +596,11 @@ define <4 x i32> @load_lane_i32_with_unfolded_gep_offset(i32* %p, <4 x i32> %v)
 ; CHECK-NEXT:    i32.const 24
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load32_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load32_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = getelementptr i32, i32* %p, i32 6
-  %t = tail call <4 x i32> @llvm.wasm.load32.lane(i32* %s, <4 x i32> %v, i32 0)
+  %x = load i32, i32* %s
+  %t = insertelement <4 x i32> %v, i32 %x, i32 0
   ret <4 x i32> %t
 }
 
@@ -593,10 +610,11 @@ define <4 x i32> @load_lane_i32_from_numeric_address(<4 x i32> %v) {
 ; CHECK-NEXT:  # %bb.0:
 ; CHECK-NEXT:    i32.const 42
 ; CHECK-NEXT:    local.get 0
-; CHECK-NEXT:    v128.load32_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load32_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = inttoptr i32 42 to i32*
-  %t = tail call <4 x i32> @llvm.wasm.load32.lane(i32* %s, <4 x i32> %v, i32 0)
+  %x = load i32, i32* %s
+  %t = insertelement <4 x i32> %v, i32 %x, i32 0
   ret <4 x i32> %t
 }
 
@@ -607,9 +625,10 @@ define <4 x i32> @load_lane_i32_from_global_address(<4 x i32> %v) {
 ; CHECK-NEXT:  # %bb.0:
 ; CHECK-NEXT:    i32.const gv_i32
 ; CHECK-NEXT:    local.get 0
-; CHECK-NEXT:    v128.load32_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load32_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
-  %t = tail call <4 x i32> @llvm.wasm.load32.lane(i32* @gv_i32, <4 x i32> %v, i32 0)
+  %x = load i32, i32* @gv_i32
+  %t = insertelement <4 x i32> %v, i32 %x, i32 0
   ret <4 x i32> %t
 }
 
@@ -739,9 +758,10 @@ define <2 x i64> @load_lane_i64_no_offset(i64* %p, <2 x i64> %v) {
 ; CHECK-NEXT:  # %bb.0:
 ; CHECK-NEXT:    local.get 0
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load64_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load64_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
-  %t = tail call <2 x i64> @llvm.wasm.load64.lane(i64* %p, <2 x i64> %v, i32 0)
+  %x = load i64, i64* %p
+  %t = insertelement <2 x i64> %v, i64 %x, i32 0
   ret <2 x i64> %t
 }
 
@@ -753,12 +773,13 @@ define <2 x i64> @load_lane_i64_with_folded_offset(i64* %p, <2 x i64> %v) {
 ; CHECK-NEXT:    i32.const 24
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load64_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load64_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %q = ptrtoint i64* %p to i32
   %r = add nuw i32 %q, 24
   %s = inttoptr i32 %r to i64*
-  %t = tail call <2 x i64> @llvm.wasm.load64.lane(i64* %s, <2 x i64> %v, i32 0)
+  %x = load i64, i64* %s
+  %t = insertelement <2 x i64> %v, i64 %x, i32 0
   ret <2 x i64> %t
 }
 
@@ -770,10 +791,11 @@ define <2 x i64> @load_lane_i64_with_folded_gep_offset(i64* %p, <2 x i64> %v) {
 ; CHECK-NEXT:    i32.const 48
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load64_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load64_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = getelementptr inbounds i64, i64* %p, i32 6
-  %t = tail call <2 x i64> @llvm.wasm.load64.lane(i64* %s, <2 x i64> %v, i32 0)
+  %x = load i64, i64* %s
+  %t = insertelement <2 x i64> %v, i64 %x, i32 0
   ret <2 x i64> %t
 }
 
@@ -785,10 +807,11 @@ define <2 x i64> @load_lane_i64_with_unfolded_gep_negative_offset(i64* %p, <2 x
 ; CHECK-NEXT:    i32.const -48
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load64_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load64_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = getelementptr inbounds i64, i64* %p, i32 -6
-  %t = tail call <2 x i64> @llvm.wasm.load64.lane(i64* %s, <2 x i64> %v, i32 0)
+  %x = load i64, i64* %s
+  %t = insertelement <2 x i64> %v, i64 %x, i32 0
   ret <2 x i64> %t
 }
 
@@ -800,12 +823,13 @@ define <2 x i64> @load_lane_i64_with_unfolded_offset(i64* %p, <2 x i64> %v) {
 ; CHECK-NEXT:    i32.const 24
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load64_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load64_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %q = ptrtoint i64* %p to i32
   %r = add nsw i32 %q, 24
   %s = inttoptr i32 %r to i64*
-  %t = tail call <2 x i64> @llvm.wasm.load64.lane(i64* %s, <2 x i64> %v, i32 0)
+  %x = load i64, i64* %s
+  %t = insertelement <2 x i64> %v, i64 %x, i32 0
   ret <2 x i64> %t
 }
 
@@ -817,10 +841,11 @@ define <2 x i64> @load_lane_i64_with_unfolded_gep_offset(i64* %p, <2 x i64> %v)
 ; CHECK-NEXT:    i32.const 48
 ; CHECK-NEXT:    i32.add
 ; CHECK-NEXT:    local.get 1
-; CHECK-NEXT:    v128.load64_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load64_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = getelementptr i64, i64* %p, i32 6
-  %t = tail call <2 x i64> @llvm.wasm.load64.lane(i64* %s, <2 x i64> %v, i32 0)
+  %x = load i64, i64* %s
+  %t = insertelement <2 x i64> %v, i64 %x, i32 0
   ret <2 x i64> %t
 }
 
@@ -830,10 +855,11 @@ define <2 x i64> @load_lane_i64_from_numeric_address(<2 x i64> %v) {
 ; CHECK-NEXT:  # %bb.0:
 ; CHECK-NEXT:    i32.const 42
 ; CHECK-NEXT:    local.get 0
-; CHECK-NEXT:    v128.load64_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load64_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
   %s = inttoptr i32 42 to i64*
-  %t = tail call <2 x i64> @llvm.wasm.load64.lane(i64* %s, <2 x i64> %v, i32 0)
+  %x = load i64, i64* %s
+  %t = insertelement <2 x i64> %v, i64 %x, i32 0
   ret <2 x i64> %t
 }
 
@@ -844,9 +870,10 @@ define <2 x i64> @load_lane_i64_from_global_address(<2 x i64> %v) {
 ; CHECK-NEXT:  # %bb.0:
 ; CHECK-NEXT:    i32.const gv_i64
 ; CHECK-NEXT:    local.get 0
-; CHECK-NEXT:    v128.load64_lane 0:p2align=0, 0
+; CHECK-NEXT:    v128.load64_lane 0, 0
 ; CHECK-NEXT:    # fallthrough-return
-  %t = tail call <2 x i64> @llvm.wasm.load64.lane(i64* @gv_i64, <2 x i64> %v, i32 0)
+  %x = load i64, i64* @gv_i64
+  %t = insertelement <2 x i64> %v, i64 %x, i32 0
   ret <2 x i64> %t
 }
 

diff --git a/llvm/test/CodeGen/WebAssembly/simd-load-store-alignment.ll b/llvm/test/CodeGen/WebAssembly/simd-load-store-alignment.ll
index 97db928..49f8e7b 100644 (file)
--- a/llvm/test/CodeGen/WebAssembly/simd-load-store-alignment.ll
+++ b/llvm/test/CodeGen/WebAssembly/simd-load-store-alignment.ll
@@ -133,6 +133,34 @@ define <16 x i8> @load_splat_v16i8_a2(i8* %p) {
   ret <16 x i8> %v2
 }
 
+; 1 is the default alignment for v128.load8_lane so no attribute is needed.
+define <16 x i8> @load_lane_i8_a1(i8* %p, <16 x i8> %v) {
+; CHECK-LABEL: load_lane_i8_a1:
+; CHECK:         .functype load_lane_i8_a1 (i32, v128) -> (v128)
+; CHECK-NEXT:  # %bb.0:
+; CHECK-NEXT:    local.get 0
+; CHECK-NEXT:    local.get 1
+; CHECK-NEXT:    v128.load8_lane 0, 0
+; CHECK-NEXT:    # fallthrough-return
+  %e = load i8, i8* %p, align 1
+  %v1 = insertelement <16 x i8> %v, i8 %e, i32 0
+  ret <16 x i8> %v1
+}
+
+; 2 is greater than the default alignment so it is ignored.
+define <16 x i8> @load_lane_i8_a2(i8* %p, <16 x i8> %v) {
+; CHECK-LABEL: load_lane_i8_a2:
+; CHECK:         .functype load_lane_i8_a2 (i32, v128) -> (v128)
+; CHECK-NEXT:  # %bb.0:
+; CHECK-NEXT:    local.get 0
+; CHECK-NEXT:    local.get 1
+; CHECK-NEXT:    v128.load8_lane 0, 0
+; CHECK-NEXT:    # fallthrough-return
+  %e = load i8, i8* %p, align 2
+  %v1 = insertelement <16 x i8> %v, i8 %e, i32 0
+  ret <16 x i8> %v1
+}
+
 ; ==============================================================================
 ; 8 x i16
 ; ==============================================================================
@@ -393,6 +421,47 @@ define <8 x i16> @load_splat_v8i16_a4(i16* %p) {
   ret <8 x i16> %v2
 }
 
+define <8 x i16> @load_lane_i16_a1(i16* %p, <8 x i16> %v) {
+; CHECK-LABEL: load_lane_i16_a1:
+; CHECK:         .functype load_lane_i16_a1 (i32, v128) -> (v128)
+; CHECK-NEXT:  # %bb.0:
+; CHECK-NEXT:    local.get 0
+; CHECK-NEXT:    local.get 1
+; CHECK-NEXT:    v128.load16_lane 0:p2align=0, 0
+; CHECK-NEXT:    # fallthrough-return
+  %e = load i16, i16* %p, align 1
+  %v1 = insertelement <8 x i16> %v, i16 %e, i32 0
+  ret <8 x i16> %v1
+}
+
+; 2 is the default alignment for v128.load16_lane so no attribute is needed.
+define <8 x i16> @load_lane_i16_a2(i16* %p, <8 x i16> %v) {
+; CHECK-LABEL: load_lane_i16_a2:
+; CHECK:         .functype load_lane_i16_a2 (i32, v128) -> (v128)
+; CHECK-NEXT:  # %bb.0:
+; CHECK-NEXT:    local.get 0
+; CHECK-NEXT:    local.get 1
+; CHECK-NEXT:    v128.load16_lane 0, 0
+; CHECK-NEXT:    # fallthrough-return
+  %e = load i16, i16* %p, align 2
+  %v1 = insertelement <8 x i16> %v, i16 %e, i32 0
+  ret <8 x i16> %v1
+}
+
+; 4 is greater than the default alignment so it is ignored.
+define <8 x i16> @load_lane_i16_a4(i16* %p, <8 x i16> %v) {
+; CHECK-LABEL: load_lane_i16_a4:
+; CHECK:         .functype load_lane_i16_a4 (i32, v128) -> (v128)
+; CHECK-NEXT:  # %bb.0:
+; CHECK-NEXT:    local.get 0
+; CHECK-NEXT:    local.get 1
+; CHECK-NEXT:    v128.load16_lane 0, 0
+; CHECK-NEXT:    # fallthrough-return
+  %e = load i16, i16* %p, align 4
+  %v1 = insertelement <8 x i16> %v, i16 %e, i32 0
+  ret <8 x i16> %v1
+}
+
 ; ==============================================================================
 ; 4 x i32
 ; ==============================================================================
@@ -666,6 +735,60 @@ define <4 x i32> @load_splat_v4i32_a8(i32* %addr) {
   ret <4 x i32> %v2
 }
 
+define <4 x i32> @load_lane_i32_a1(i32* %p, <4 x i32> %v) {
+; CHECK-LABEL: load_lane_i32_a1:
+; CHECK:         .functype load_lane_i32_a1 (i32, v128) -> (v128)
+; CHECK-NEXT:  # %bb.0:
+; CHECK-NEXT:    local.get 0
+; CHECK-NEXT:    local.get 1
+; CHECK-NEXT:    v128.load32_lane 0:p2align=0, 0
+; CHECK-NEXT:    # fallthrough-return
+  %e = load i32, i32* %p, align 1
+  %v1 = insertelement <4 x i32> %v, i32 %e, i32 0
+  ret <4 x i32> %v1
+}
+
+define <4 x i32> @load_lane_i32_a2(i32* %p, <4 x i32> %v) {
+; CHECK-LABEL: load_lane_i32_a2:
+; CHECK:         .functype load_lane_i32_a2 (i32, v128) -> (v128)
+; CHECK-NEXT:  # %bb.0:
+; CHECK-NEXT:    local.get 0
+; CHECK-NEXT:    local.get 1
+; CHECK-NEXT:    v128.load32_lane 0:p2align=1, 0
+; CHECK-NEXT:    # fallthrough-return
+  %e = load i32, i32* %p, align 2
+  %v1 = insertelement <4 x i32> %v, i32 %e, i32 0
+  ret <4 x i32> %v1
+}
+
+; 4 is the default alignment for v128.load32_lane so no attribute is needed.
+define <4 x i32> @load_lane_i32_a4(i32* %p, <4 x i32> %v) {
+; CHECK-LABEL: load_lane_i32_a4:
+; CHECK:         .functype load_lane_i32_a4 (i32, v128) -> (v128)
+; CHECK-NEXT:  # %bb.0:
+; CHECK-NEXT:    local.get 0
+; CHECK-NEXT:    local.get 1
+; CHECK-NEXT:    v128.load32_lane 0, 0
+; CHECK-NEXT:    # fallthrough-return
+  %e = load i32, i32* %p, align 4
+  %v1 = insertelement <4 x i32> %v, i32 %e, i32 0
+  ret <4 x i32> %v1
+}
+
+; 8 is greater than the default alignment so it is ignored.
+define <4 x i32> @load_lane_i32_a8(i32* %p, <4 x i32> %v) {
+; CHECK-LABEL: load_lane_i32_a8:
+; CHECK:         .functype load_lane_i32_a8 (i32, v128) -> (v128)
+; CHECK-NEXT:  # %bb.0:
+; CHECK-NEXT:    local.get 0
+; CHECK-NEXT:    local.get 1
+; CHECK-NEXT:    v128.load32_lane 0, 0
+; CHECK-NEXT:    # fallthrough-return
+  %e = load i32, i32* %p, align 8
+  %v1 = insertelement <4 x i32> %v, i32 %e, i32 0
+  ret <4 x i32> %v1
+}
+
 ; ==============================================================================
 ; 2 x i64
 ; ==============================================================================
@@ -833,6 +956,73 @@ define <2 x i64> @load_splat_v2i64_a16(i64* %p) {
   ret <2 x i64> %v2
 }
 
+define <2 x i64> @load_lane_i64_a1(i64* %p, <2 x i64> %v) {
+; CHECK-LABEL: load_lane_i64_a1:
+; CHECK:         .functype load_lane_i64_a1 (i32, v128) -> (v128)
+; CHECK-NEXT:  # %bb.0:
+; CHECK-NEXT:    local.get 0
+; CHECK-NEXT:    local.get 1
+; CHECK-NEXT:    v128.load64_lane 0:p2align=0, 0
+; CHECK-NEXT:    # fallthrough-return
+  %e = load i64, i64* %p, align 1
+  %v1 = insertelement <2 x i64> %v, i64 %e, i32 0
+  ret <2 x i64> %v1
+}
+
+define <2 x i64> @load_lane_i64_a2(i64* %p, <2 x i64> %v) {
+; CHECK-LABEL: load_lane_i64_a2:
+; CHECK:         .functype load_lane_i64_a2 (i32, v128) -> (v128)
+; CHECK-NEXT:  # %bb.0:
+; CHECK-NEXT:    local.get 0
+; CHECK-NEXT:    local.get 1
+; CHECK-NEXT:    v128.load64_lane 0:p2align=1, 0
+; CHECK-NEXT:    # fallthrough-return
+  %e = load i64, i64* %p, align 2
+  %v1 = insertelement <2 x i64> %v, i64 %e, i32 0
+  ret <2 x i64> %v1
+}
+
+define <2 x i64> @load_lane_i64_a4(i64* %p, <2 x i64> %v) {
+; CHECK-LABEL: load_lane_i64_a4:
+; CHECK:         .functype load_lane_i64_a4 (i32, v128) -> (v128)
+; CHECK-NEXT:  # %bb.0:
+; CHECK-NEXT:    local.get 0
+; CHECK-NEXT:    local.get 1
+; CHECK-NEXT:    v128.load64_lane 0:p2align=2, 0
+; CHECK-NEXT:    # fallthrough-return
+  %e = load i64, i64* %p, align 4
+  %v1 = insertelement <2 x i64> %v, i64 %e, i32 0
+  ret <2 x i64> %v1
+}
+
+; 8 is the default alignment for v128.load64_lane so no attribute is needed.
+define <2 x i64> @load_lane_i64_a8(i64* %p, <2 x i64> %v) {
+; CHECK-LABEL: load_lane_i64_a8:
+; CHECK:         .functype load_lane_i64_a8 (i32, v128) -> (v128)
+; CHECK-NEXT:  # %bb.0:
+; CHECK-NEXT:    local.get 0
+; CHECK-NEXT:    local.get 1
+; CHECK-NEXT:    v128.load64_lane 0, 0
+; CHECK-NEXT:    # fallthrough-return
+  %e = load i64, i64* %p, align 8
+  %v1 = insertelement <2 x i64> %v, i64 %e, i32 0
+  ret <2 x i64> %v1
+}
+
+; 16 is greater than the default alignment so it is ignored.
+define <2 x i64> @load_lane_i64_a16(i64* %p, <2 x i64> %v) {
+; CHECK-LABEL: load_lane_i64_a16:
+; CHECK:         .functype load_lane_i64_a16 (i32, v128) -> (v128)
+; CHECK-NEXT:  # %bb.0:
+; CHECK-NEXT:    local.get 0
+; CHECK-NEXT:    local.get 1
+; CHECK-NEXT:    v128.load64_lane 0, 0
+; CHECK-NEXT:    # fallthrough-return
+  %e = load i64, i64* %p, align 16
+  %v1 = insertelement <2 x i64> %v, i64 %e, i32 0
+  ret <2 x i64> %v1
+}
+
 ; ==============================================================================
 ; 4 x float
 ; ==============================================================================