[gcc r12-2949] aarch64: Remove macros for vld2[q]_lane Neon intrinsics

[gcc r12-2949] aarch64: Remove macros for vld2[q]_lane Neon intrinsics

[Jonathan Wright]

https://gcc.gnu.org/g:5ed35a9874ba8c3aa2bbbd720e46783db264b684

commit r12-2949-g5ed35a9874ba8c3aa2bbbd720e46783db264b684
Author: Jonathan Wright <jonathan.wright@arm.com>
Date:   Thu Aug 12 12:27:15 2021 +0100

    aarch64: Remove macros for vld2[q]_lane Neon intrinsics
    
    Remove macros for vld2[q]_lane Neon intrinsics. This is a preparatory
    step before adding new modes for structures of Advanced SIMD vectors.
    
    gcc/ChangeLog:
    
    2021-08-12  Jonathan Wright  <jonathan.wright@arm.com>
    
            * config/aarch64/arm_neon.h (__LD2_LANE_FUNC): Delete.
            (__LD2Q_LANE_FUNC): Likewise.
            (vld2_lane_u8): Define without macro.
            (vld2_lane_u16): Likewise.
            (vld2_lane_u32): Likewise.
            (vld2_lane_u64): Likewise.
            (vld2_lane_s8): Likewise.
            (vld2_lane_s16): Likewise.
            (vld2_lane_s32): Likewise.
            (vld2_lane_s64): Likewise.
            (vld2_lane_f16): Likewise.
            (vld2_lane_f32): Likewise.
            (vld2_lane_f64): Likewise.
            (vld2_lane_p8): Likewise.
            (vld2_lane_p16): Likewise.
            (vld2_lane_p64): Likewise.
            (vld2q_lane_u8): Likewise.
            (vld2q_lane_u16): Likewise.
            (vld2q_lane_u32): Likewise.
            (vld2q_lane_u64): Likewise.
            (vld2q_lane_s8): Likewise.
            (vld2q_lane_s16): Likewise.
            (vld2q_lane_s32): Likewise.
            (vld2q_lane_s64): Likewise.
            (vld2q_lane_f16): Likewise.
            (vld2q_lane_f32): Likewise.
            (vld2q_lane_f64): Likewise.
            (vld2q_lane_p8): Likewise.
            (vld2q_lane_p16): Likewise.
            (vld2q_lane_p64): Likewise.
            (vld2_lane_bf16): Likewise.
            (vld2q_lane_bf16): Likewise.

Diff:
---
 gcc/config/aarch64/arm_neon.h | 558 +++++++++++++++++++++++++++++++++++-------
 1 file changed, 474 insertions(+), 84 deletions(-)

diff --git a/gcc/config/aarch64/arm_neon.h b/gcc/config/aarch64/arm_neon.h
index 390cf9a7743..91c072fe457 100644
--- a/gcc/config/aarch64/arm_neon.h
+++ b/gcc/config/aarch64/arm_neon.h
@@ -19882,92 +19882,455 @@ vld4q_dup_p64 (const poly64_t * __a)
 
 /* vld2_lane */
 
-#define __LD2_LANE_FUNC(intype, vectype, largetype, ptrtype, mode,	   \
-			 qmode, ptrmode, funcsuffix, signedtype)	   \
-__extension__ extern __inline intype \
-__attribute__ ((__always_inline__, __gnu_inline__,__artificial__)) \
-vld2_lane_##funcsuffix (const ptrtype * __ptr, intype __b, const int __c)  \
-{									   \
-  __builtin_aarch64_simd_oi __o;					   \
-  largetype __temp;							   \
-  __temp.val[0] =							   \
-    vcombine_##funcsuffix (__b.val[0], vcreate_##funcsuffix (0));	   \
-  __temp.val[1] =							   \
-    vcombine_##funcsuffix (__b.val[1], vcreate_##funcsuffix (0));	   \
-  __o = __builtin_aarch64_set_qregoi##qmode (__o,			   \
-					    (signedtype) __temp.val[0],	   \
-					    0);				   \
-  __o = __builtin_aarch64_set_qregoi##qmode (__o,			   \
-					    (signedtype) __temp.val[1],	   \
-					    1);				   \
-  __o =	__builtin_aarch64_ld2_lane##mode (				   \
-	  (__builtin_aarch64_simd_##ptrmode *) __ptr, __o, __c);	   \
-  __b.val[0] = (vectype) __builtin_aarch64_get_dregoidi (__o, 0);	   \
-  __b.val[1] = (vectype) __builtin_aarch64_get_dregoidi (__o, 1);	   \
-  return __b;								   \
+__extension__ extern __inline uint8x8x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_u8 (const uint8_t * __ptr, uint8x8x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  uint8x16x2_t __temp;
+  __temp.val[0] = vcombine_u8 (__b.val[0], vcreate_u8 (0));
+  __temp.val[1] = vcombine_u8 (__b.val[1], vcreate_u8 (0));
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanev8qi (
+	  (__builtin_aarch64_simd_qi *) __ptr, __o, __c);
+  __b.val[0] = (uint8x8_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (uint8x8_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
 }
 
-__LD2_LANE_FUNC (float16x4x2_t, float16x4_t, float16x8x2_t, float16_t, v4hf,
-		 v8hf, hf, f16, float16x8_t)
-__LD2_LANE_FUNC (float32x2x2_t, float32x2_t, float32x4x2_t, float32_t, v2sf, v4sf,
-		 sf, f32, float32x4_t)
-__LD2_LANE_FUNC (float64x1x2_t, float64x1_t, float64x2x2_t, float64_t, df, v2df,
-		 df, f64, float64x2_t)
-__LD2_LANE_FUNC (poly8x8x2_t, poly8x8_t, poly8x16x2_t, poly8_t, v8qi, v16qi, qi, p8,
-		 int8x16_t)
-__LD2_LANE_FUNC (poly16x4x2_t, poly16x4_t, poly16x8x2_t, poly16_t, v4hi, v8hi, hi,
-		 p16, int16x8_t)
-__LD2_LANE_FUNC (poly64x1x2_t, poly64x1_t, poly64x2x2_t, poly64_t, di,
-		 v2di_ssps, di, p64, poly64x2_t)
-__LD2_LANE_FUNC (int8x8x2_t, int8x8_t, int8x16x2_t, int8_t, v8qi, v16qi, qi, s8,
-		 int8x16_t)
-__LD2_LANE_FUNC (int16x4x2_t, int16x4_t, int16x8x2_t, int16_t, v4hi, v8hi, hi, s16,
-		 int16x8_t)
-__LD2_LANE_FUNC (int32x2x2_t, int32x2_t, int32x4x2_t, int32_t, v2si, v4si, si, s32,
-		 int32x4_t)
-__LD2_LANE_FUNC (int64x1x2_t, int64x1_t, int64x2x2_t, int64_t, di, v2di, di, s64,
-		 int64x2_t)
-__LD2_LANE_FUNC (uint8x8x2_t, uint8x8_t, uint8x16x2_t, uint8_t, v8qi, v16qi, qi, u8,
-		 int8x16_t)
-__LD2_LANE_FUNC (uint16x4x2_t, uint16x4_t, uint16x8x2_t, uint16_t, v4hi, v8hi, hi,
-		 u16, int16x8_t)
-__LD2_LANE_FUNC (uint32x2x2_t, uint32x2_t, uint32x4x2_t, uint32_t, v2si, v4si, si,
-		 u32, int32x4_t)
-__LD2_LANE_FUNC (uint64x1x2_t, uint64x1_t, uint64x2x2_t, uint64_t, di, v2di, di,
-		 u64, int64x2_t)
+__extension__ extern __inline uint16x4x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_u16 (const uint16_t * __ptr, uint16x4x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  uint16x8x2_t __temp;
+  __temp.val[0] = vcombine_u16 (__b.val[0], vcreate_u16 (0));
+  __temp.val[1] = vcombine_u16 (__b.val[1], vcreate_u16 (0));
+  __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanev4hi (
+	  (__builtin_aarch64_simd_hi *) __ptr, __o, __c);
+  __b.val[0] = (uint16x4_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (uint16x4_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
+}
+
+__extension__ extern __inline uint32x2x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_u32 (const uint32_t * __ptr, uint32x2x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  uint32x4x2_t __temp;
+  __temp.val[0] = vcombine_u32 (__b.val[0], vcreate_u32 (0));
+  __temp.val[1] = vcombine_u32 (__b.val[1], vcreate_u32 (0));
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanev2si (
+	  (__builtin_aarch64_simd_si *) __ptr, __o, __c);
+  __b.val[0] = (uint32x2_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (uint32x2_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
+}
+
+__extension__ extern __inline uint64x1x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_u64 (const uint64_t * __ptr, uint64x1x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  uint64x2x2_t __temp;
+  __temp.val[0] = vcombine_u64 (__b.val[0], vcreate_u64 (0));
+  __temp.val[1] = vcombine_u64 (__b.val[1], vcreate_u64 (0));
+  __o = __builtin_aarch64_set_qregoiv2di (__o, (int64x2_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv2di (__o, (int64x2_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanedi (
+	  (__builtin_aarch64_simd_di *) __ptr, __o, __c);
+  __b.val[0] = (uint64x1_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (uint64x1_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
+}
+
+__extension__ extern __inline int8x8x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_s8 (const int8_t * __ptr, int8x8x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  int8x16x2_t __temp;
+  __temp.val[0] = vcombine_s8 (__b.val[0], vcreate_s8 (0));
+  __temp.val[1] = vcombine_s8 (__b.val[1], vcreate_s8 (0));
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanev8qi (
+	  (__builtin_aarch64_simd_qi *) __ptr, __o, __c);
+  __b.val[0] = (int8x8_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (int8x8_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
+}
+
+__extension__ extern __inline int16x4x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_s16 (const int16_t * __ptr, int16x4x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  int16x8x2_t __temp;
+  __temp.val[0] = vcombine_s16 (__b.val[0], vcreate_s16 (0));
+  __temp.val[1] = vcombine_s16 (__b.val[1], vcreate_s16 (0));
+  __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanev4hi (
+	  (__builtin_aarch64_simd_hi *) __ptr, __o, __c);
+  __b.val[0] = (int16x4_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (int16x4_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
+}
+
+__extension__ extern __inline int32x2x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_s32 (const int32_t * __ptr, int32x2x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  int32x4x2_t __temp;
+  __temp.val[0] = vcombine_s32 (__b.val[0], vcreate_s32 (0));
+  __temp.val[1] = vcombine_s32 (__b.val[1], vcreate_s32 (0));
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanev2si (
+	  (__builtin_aarch64_simd_si *) __ptr, __o, __c);
+  __b.val[0] = (int32x2_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (int32x2_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
+}
+
+__extension__ extern __inline int64x1x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_s64 (const int64_t * __ptr, int64x1x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  int64x2x2_t __temp;
+  __temp.val[0] = vcombine_s64 (__b.val[0], vcreate_s64 (0));
+  __temp.val[1] = vcombine_s64 (__b.val[1], vcreate_s64 (0));
+  __o = __builtin_aarch64_set_qregoiv2di (__o, (int64x2_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv2di (__o, (int64x2_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanedi (
+	  (__builtin_aarch64_simd_di *) __ptr, __o, __c);
+  __b.val[0] = (int64x1_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (int64x1_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
+}
+
+__extension__ extern __inline float16x4x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_f16 (const float16_t * __ptr, float16x4x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  float16x8x2_t __temp;
+  __temp.val[0] = vcombine_f16 (__b.val[0], vcreate_f16 (0));
+  __temp.val[1] = vcombine_f16 (__b.val[1], vcreate_f16 (0));
+  __o = __builtin_aarch64_set_qregoiv8hf (__o, (float16x8_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv8hf (__o, (float16x8_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanev4hf (
+	  (__builtin_aarch64_simd_hf *) __ptr, __o, __c);
+  __b.val[0] = (float16x4_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (float16x4_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
+}
+
+__extension__ extern __inline float32x2x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_f32 (const float32_t * __ptr, float32x2x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  float32x4x2_t __temp;
+  __temp.val[0] = vcombine_f32 (__b.val[0], vcreate_f32 (0));
+  __temp.val[1] = vcombine_f32 (__b.val[1], vcreate_f32 (0));
+  __o = __builtin_aarch64_set_qregoiv4sf (__o, (float32x4_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4sf (__o, (float32x4_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanev2sf (
+	  (__builtin_aarch64_simd_sf *) __ptr, __o, __c);
+  __b.val[0] = (float32x2_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (float32x2_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
+}
+
+__extension__ extern __inline float64x1x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_f64 (const float64_t * __ptr, float64x1x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  float64x2x2_t __temp;
+  __temp.val[0] = vcombine_f64 (__b.val[0], vcreate_f64 (0));
+  __temp.val[1] = vcombine_f64 (__b.val[1], vcreate_f64 (0));
+  __o = __builtin_aarch64_set_qregoiv2df (__o, (float64x2_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv2df (__o, (float64x2_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanedf (
+	  (__builtin_aarch64_simd_df *) __ptr, __o, __c);
+  __b.val[0] = (float64x1_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (float64x1_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
+}
+
+__extension__ extern __inline poly8x8x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_p8 (const poly8_t * __ptr, poly8x8x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  poly8x16x2_t __temp;
+  __temp.val[0] = vcombine_p8 (__b.val[0], vcreate_p8 (0));
+  __temp.val[1] = vcombine_p8 (__b.val[1], vcreate_p8 (0));
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanev8qi (
+	  (__builtin_aarch64_simd_qi *) __ptr, __o, __c);
+  __b.val[0] = (poly8x8_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (poly8x8_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
+}
+
+__extension__ extern __inline poly16x4x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_p16 (const poly16_t * __ptr, poly16x4x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  poly16x8x2_t __temp;
+  __temp.val[0] = vcombine_p16 (__b.val[0], vcreate_p16 (0));
+  __temp.val[1] = vcombine_p16 (__b.val[1], vcreate_p16 (0));
+  __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv8hi (__o, (int16x8_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanev4hi (
+	  (__builtin_aarch64_simd_hi *) __ptr, __o, __c);
+  __b.val[0] = (poly16x4_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (poly16x4_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
+}
+
+__extension__ extern __inline poly64x1x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_p64 (const poly64_t * __ptr, poly64x1x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  poly64x2x2_t __temp;
+  __temp.val[0] = vcombine_p64 (__b.val[0], vcreate_p64 (0));
+  __temp.val[1] = vcombine_p64 (__b.val[1], vcreate_p64 (0));
+  __o = __builtin_aarch64_set_qregoiv2di (__o, (int64x2_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv2di (__o, (int64x2_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanedi (
+	  (__builtin_aarch64_simd_di *) __ptr, __o, __c);
+  __b.val[0] = (poly64x1_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (poly64x1_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
+}
 
 /* vld2q_lane */
 
-#define __LD2Q_LANE_FUNC(intype, vtype, ptrtype, mode, ptrmode, funcsuffix) \
-__extension__ extern __inline intype \
-__attribute__ ((__always_inline__, __gnu_inline__,__artificial__)) \
-vld2q_lane_##funcsuffix (const ptrtype * __ptr, intype __b, const int __c) \
-{									   \
-  __builtin_aarch64_simd_oi __o;					   \
-  intype ret;								   \
-  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0); \
-  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1); \
-  __o = __builtin_aarch64_ld2_lane##mode (				   \
-	(__builtin_aarch64_simd_##ptrmode *) __ptr, __o, __c);		   \
-  ret.val[0] = (vtype) __builtin_aarch64_get_qregoiv4si (__o, 0);	   \
-  ret.val[1] = (vtype) __builtin_aarch64_get_qregoiv4si (__o, 1);	   \
-  return ret;								   \
+__extension__ extern __inline uint8x16x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_u8 (const uint8_t * __ptr, uint8x16x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  uint8x16x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev16qi (
+	(__builtin_aarch64_simd_qi *) __ptr, __o, __c);
+  ret.val[0] = (uint8x16_t) __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = (uint8x16_t) __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
 }
 
-__LD2Q_LANE_FUNC (float16x8x2_t, float16x8_t, float16_t, v8hf, hf, f16)
-__LD2Q_LANE_FUNC (float32x4x2_t, float32x4_t, float32_t, v4sf, sf, f32)
-__LD2Q_LANE_FUNC (float64x2x2_t, float64x2_t, float64_t, v2df, df, f64)
-__LD2Q_LANE_FUNC (poly8x16x2_t, poly8x16_t, poly8_t, v16qi, qi, p8)
-__LD2Q_LANE_FUNC (poly16x8x2_t, poly16x8_t, poly16_t, v8hi, hi, p16)
-__LD2Q_LANE_FUNC (poly64x2x2_t, poly64x2_t, poly64_t, v2di, di, p64)
-__LD2Q_LANE_FUNC (int8x16x2_t, int8x16_t, int8_t, v16qi, qi, s8)
-__LD2Q_LANE_FUNC (int16x8x2_t, int16x8_t, int16_t, v8hi, hi, s16)
-__LD2Q_LANE_FUNC (int32x4x2_t, int32x4_t, int32_t, v4si, si, s32)
-__LD2Q_LANE_FUNC (int64x2x2_t, int64x2_t, int64_t, v2di, di, s64)
-__LD2Q_LANE_FUNC (uint8x16x2_t, uint8x16_t, uint8_t, v16qi, qi, u8)
-__LD2Q_LANE_FUNC (uint16x8x2_t, uint16x8_t, uint16_t, v8hi, hi, u16)
-__LD2Q_LANE_FUNC (uint32x4x2_t, uint32x4_t, uint32_t, v4si, si, u32)
-__LD2Q_LANE_FUNC (uint64x2x2_t, uint64x2_t, uint64_t, v2di, di, u64)
+__extension__ extern __inline uint16x8x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_u16 (const uint16_t * __ptr, uint16x8x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  uint16x8x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev8hi (
+	(__builtin_aarch64_simd_hi *) __ptr, __o, __c);
+  ret.val[0] = (uint16x8_t) __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = (uint16x8_t) __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
+}
+
+__extension__ extern __inline uint32x4x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_u32 (const uint32_t * __ptr, uint32x4x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  uint32x4x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev4si (
+	(__builtin_aarch64_simd_si *) __ptr, __o, __c);
+  ret.val[0] = (uint32x4_t) __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = (uint32x4_t) __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
+}
+
+__extension__ extern __inline uint64x2x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_u64 (const uint64_t * __ptr, uint64x2x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  uint64x2x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev2di (
+	(__builtin_aarch64_simd_di *) __ptr, __o, __c);
+  ret.val[0] = (uint64x2_t) __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = (uint64x2_t) __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
+}
+
+__extension__ extern __inline int8x16x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_s8 (const int8_t * __ptr, int8x16x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  int8x16x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev16qi (
+	(__builtin_aarch64_simd_qi *) __ptr, __o, __c);
+  ret.val[0] = (int8x16_t) __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = (int8x16_t) __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
+}
+
+__extension__ extern __inline int16x8x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_s16 (const int16_t * __ptr, int16x8x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  int16x8x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev8hi (
+	(__builtin_aarch64_simd_hi *) __ptr, __o, __c);
+  ret.val[0] = (int16x8_t) __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = (int16x8_t) __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
+}
+
+__extension__ extern __inline int32x4x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_s32 (const int32_t * __ptr, int32x4x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  int32x4x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev4si (
+	(__builtin_aarch64_simd_si *) __ptr, __o, __c);
+  ret.val[0] = __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
+}
+
+__extension__ extern __inline int64x2x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_s64 (const int64_t * __ptr, int64x2x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  int64x2x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev2di (
+	(__builtin_aarch64_simd_di *) __ptr, __o, __c);
+  ret.val[0] = (int64x2_t) __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = (int64x2_t) __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
+}
+
+__extension__ extern __inline float16x8x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_f16 (const float16_t * __ptr, float16x8x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  float16x8x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev8hf (
+	(__builtin_aarch64_simd_hf *) __ptr, __o, __c);
+  ret.val[0] = (float16x8_t) __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = (float16x8_t) __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
+}
+
+__extension__ extern __inline float32x4x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_f32 (const float32_t * __ptr, float32x4x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  float32x4x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev4sf (
+	(__builtin_aarch64_simd_sf *) __ptr, __o, __c);
+  ret.val[0] = (float32x4_t) __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = (float32x4_t) __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
+}
+
+__extension__ extern __inline float64x2x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_f64 (const float64_t * __ptr, float64x2x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  float64x2x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev2df (
+	(__builtin_aarch64_simd_df *) __ptr, __o, __c);
+  ret.val[0] = (float64x2_t) __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = (float64x2_t) __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
+}
+
+__extension__ extern __inline poly8x16x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_p8 (const poly8_t * __ptr, poly8x16x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  poly8x16x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev16qi (
+	(__builtin_aarch64_simd_qi *) __ptr, __o, __c);
+  ret.val[0] = (poly8x16_t) __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = (poly8x16_t) __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
+}
+
+__extension__ extern __inline poly16x8x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_p16 (const poly16_t * __ptr, poly16x8x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  poly16x8x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev8hi (
+	(__builtin_aarch64_simd_hi *) __ptr, __o, __c);
+  ret.val[0] = (poly16x8_t) __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = (poly16x8_t) __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
+}
+
+__extension__ extern __inline poly64x2x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_p64 (const poly64_t * __ptr, poly64x2x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  poly64x2x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev2di (
+	(__builtin_aarch64_simd_di *) __ptr, __o, __c);
+  ret.val[0] = (poly64x2_t) __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = (poly64x2_t) __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
+}
 
 /* vld3_lane */
 
@@ -34584,9 +34947,38 @@ vcopyq_laneq_bf16 (bfloat16x8_t __a, const int __lane1,
 				  __a, __lane1);
 }
 
-__LD2_LANE_FUNC (bfloat16x4x2_t, bfloat16x4_t, bfloat16x8x2_t, bfloat16_t, v4bf,
-		 v8bf, bf, bf16, bfloat16x8_t)
-__LD2Q_LANE_FUNC (bfloat16x8x2_t, bfloat16x8_t, bfloat16_t, v8bf, bf, bf16)
+__extension__ extern __inline bfloat16x4x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2_lane_bf16 (const bfloat16_t * __ptr, bfloat16x4x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  bfloat16x8x2_t __temp;
+  __temp.val[0] = vcombine_bf16 (__b.val[0], vcreate_bf16 (0));
+  __temp.val[1] = vcombine_bf16 (__b.val[1], vcreate_bf16 (0));
+  __o = __builtin_aarch64_set_qregoiv8bf (__o, (bfloat16x8_t) __temp.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv8bf (__o, (bfloat16x8_t) __temp.val[1], 1);
+  __o =	__builtin_aarch64_ld2_lanev4bf (
+	  (__builtin_aarch64_simd_bf *) __ptr, __o, __c);
+  __b.val[0] = (bfloat16x4_t) __builtin_aarch64_get_dregoidi (__o, 0);
+  __b.val[1] = (bfloat16x4_t) __builtin_aarch64_get_dregoidi (__o, 1);
+  return __b;
+}
+
+__extension__ extern __inline bfloat16x8x2_t
+__attribute__ ((__always_inline__, __gnu_inline__,__artificial__))
+vld2q_lane_bf16 (const bfloat16_t * __ptr, bfloat16x8x2_t __b, const int __c)
+{
+  __builtin_aarch64_simd_oi __o;
+  bfloat16x8x2_t ret;
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv4si (__o, (int32x4_t) __b.val[1], 1);
+  __o = __builtin_aarch64_ld2_lanev8bf (
+	(__builtin_aarch64_simd_bf *) __ptr, __o, __c);
+  ret.val[0] = (bfloat16x8_t) __builtin_aarch64_get_qregoiv4si (__o, 0);
+  ret.val[1] = (bfloat16x8_t) __builtin_aarch64_get_qregoiv4si (__o, 1);
+  return ret;
+}
+
 __LD3_LANE_FUNC (bfloat16x4x3_t, bfloat16x4_t, bfloat16x8x3_t, bfloat16_t, v4bf,
 		 v8bf, bf, bf16, bfloat16x8_t)
 __LD3Q_LANE_FUNC (bfloat16x8x3_t, bfloat16x8_t, bfloat16_t, v8bf, bf, bf16)
@@ -34888,8 +35280,6 @@ vaddq_p128 (poly128_t __a, poly128_t __b)
 #undef __aarch64_vdupq_laneq_u32
 #undef __aarch64_vdupq_laneq_u64
 
-#undef __LD2_LANE_FUNC
-#undef __LD2Q_LANE_FUNC
 #undef __LD3_LANE_FUNC
 #undef __LD3Q_LANE_FUNC
 #undef __LD4_LANE_FUNC