RE: [PATCH 20/23] arm: [MVE intrinsics] rework vqrshrunbq vqrshruntq vqshrunbq vqshruntq

[Kyrylo Tkachov via Gcc-patches]

> -----Original Message-----
> From: Christophe Lyon <christophe.l...@arm.com>
> Sent: Friday, May 5, 2023 9:39 AM
> To: gcc-patches@gcc.gnu.org; Kyrylo Tkachov <kyrylo.tkac...@arm.com>;
> Richard Earnshaw <richard.earns...@arm.com>; Richard Sandiford
> <richard.sandif...@arm.com>
> Cc: Christophe Lyon <christophe.l...@arm.com>
> Subject: [PATCH 20/23] arm: [MVE intrinsics] rework vqrshrunbq vqrshruntq
> vqshrunbq vqshruntq
> 
> Implement vqrshrunbq, vqrshruntq, vqshrunbq, vqshruntq using the new
> MVE builtins framework.

Ok.
Thanks,
Kyrill

> 
> 2022-09-08  Christophe Lyon  <christophe.l...@arm.com>
> 
>       gcc/
>       * config/arm/arm-mve-builtins-base.cc
> (FUNCTION_ONLY_N_NO_U_F): New.
>       (vqshrunbq, vqshruntq, vqrshrunbq, vqrshruntq): New.
>       * config/arm/arm-mve-builtins-base.def (vqshrunbq, vqshruntq)
>       (vqrshrunbq, vqrshruntq): New.
>       * config/arm/arm-mve-builtins-base.h (vqshrunbq, vqshruntq)
>       (vqrshrunbq, vqrshruntq): New.
>       * config/arm/arm-mve-builtins.cc
>       (function_instance::has_inactive_argument): Handle vqshrunbq,
>       vqshruntq, vqrshrunbq, vqrshruntq.
>       * config/arm/arm_mve.h (vqrshrunbq): Remove.
>       (vqrshruntq): Remove.
>       (vqrshrunbq_m): Remove.
>       (vqrshruntq_m): Remove.
>       (vqrshrunbq_n_s16): Remove.
>       (vqrshrunbq_n_s32): Remove.
>       (vqrshruntq_n_s16): Remove.
>       (vqrshruntq_n_s32): Remove.
>       (vqrshrunbq_m_n_s32): Remove.
>       (vqrshrunbq_m_n_s16): Remove.
>       (vqrshruntq_m_n_s32): Remove.
>       (vqrshruntq_m_n_s16): Remove.
>       (__arm_vqrshrunbq_n_s16): Remove.
>       (__arm_vqrshrunbq_n_s32): Remove.
>       (__arm_vqrshruntq_n_s16): Remove.
>       (__arm_vqrshruntq_n_s32): Remove.
>       (__arm_vqrshrunbq_m_n_s32): Remove.
>       (__arm_vqrshrunbq_m_n_s16): Remove.
>       (__arm_vqrshruntq_m_n_s32): Remove.
>       (__arm_vqrshruntq_m_n_s16): Remove.
>       (__arm_vqrshrunbq): Remove.
>       (__arm_vqrshruntq): Remove.
>       (__arm_vqrshrunbq_m): Remove.
>       (__arm_vqrshruntq_m): Remove.
>       (vqshrunbq): Remove.
>       (vqshruntq): Remove.
>       (vqshrunbq_m): Remove.
>       (vqshruntq_m): Remove.
>       (vqshrunbq_n_s16): Remove.
>       (vqshruntq_n_s16): Remove.
>       (vqshrunbq_n_s32): Remove.
>       (vqshruntq_n_s32): Remove.
>       (vqshrunbq_m_n_s32): Remove.
>       (vqshrunbq_m_n_s16): Remove.
>       (vqshruntq_m_n_s32): Remove.
>       (vqshruntq_m_n_s16): Remove.
>       (__arm_vqshrunbq_n_s16): Remove.
>       (__arm_vqshruntq_n_s16): Remove.
>       (__arm_vqshrunbq_n_s32): Remove.
>       (__arm_vqshruntq_n_s32): Remove.
>       (__arm_vqshrunbq_m_n_s32): Remove.
>       (__arm_vqshrunbq_m_n_s16): Remove.
>       (__arm_vqshruntq_m_n_s32): Remove.
>       (__arm_vqshruntq_m_n_s16): Remove.
>       (__arm_vqshrunbq): Remove.
>       (__arm_vqshruntq): Remove.
>       (__arm_vqshrunbq_m): Remove.
>       (__arm_vqshruntq_m): Remove.
> ---
>  gcc/config/arm/arm-mve-builtins-base.cc  |  13 +
>  gcc/config/arm/arm-mve-builtins-base.def |   4 +
>  gcc/config/arm/arm-mve-builtins-base.h   |   4 +
>  gcc/config/arm/arm-mve-builtins.cc       |   4 +
>  gcc/config/arm/arm_mve.h                 | 320 -----------------------
>  5 files changed, 25 insertions(+), 320 deletions(-)
> 
> diff --git a/gcc/config/arm/arm-mve-builtins-base.cc b/gcc/config/arm/arm-
> mve-builtins-base.cc
> index c95abe70239..e7d2e0abffc 100644
> --- a/gcc/config/arm/arm-mve-builtins-base.cc
> +++ b/gcc/config/arm/arm-mve-builtins-base.cc
> @@ -184,6 +184,15 @@ namespace arm_mve {
>      -1, -1, -1,                                                              
> \
>      UNSPEC##_M_N_S, UNSPEC##_M_N_U, -1))
> 
> +  /* Helper for builtins with only unspec codes, _m predicated
> +     overrides, only _n version, no unsigned, no floating-point.  */
> +#define FUNCTION_ONLY_N_NO_U_F(NAME, UNSPEC) FUNCTION
>               \
> +  (NAME, unspec_mve_function_exact_insn,                             \
> +   (-1, -1, -1,                                                              
> \
> +    UNSPEC##_N_S, -1, -1,                                            \
> +    -1, -1, -1,                                                              
> \
> +    UNSPEC##_M_N_S, -1, -1))
> +
>  FUNCTION_WITHOUT_N (vabdq, VABDQ)
>  FUNCTION_WITH_RTX_M_N (vaddq, PLUS, VADDQ)
>  FUNCTION_WITH_RTX_M (vandq, AND, VANDQ)
> @@ -203,8 +212,12 @@ FUNCTION_WITH_M_N_NO_U_F (vqrdmulhq,
> VQRDMULHQ)
>  FUNCTION_WITH_M_N_R (vqshlq, VQSHLQ)
>  FUNCTION_ONLY_N_NO_F (vqrshrnbq, VQRSHRNBQ)
>  FUNCTION_ONLY_N_NO_F (vqrshrntq, VQRSHRNTQ)
> +FUNCTION_ONLY_N_NO_U_F (vqrshrunbq, VQRSHRUNBQ)
> +FUNCTION_ONLY_N_NO_U_F (vqrshruntq, VQRSHRUNTQ)
>  FUNCTION_ONLY_N_NO_F (vqshrnbq, VQSHRNBQ)
>  FUNCTION_ONLY_N_NO_F (vqshrntq, VQSHRNTQ)
> +FUNCTION_ONLY_N_NO_U_F (vqshrunbq, VQSHRUNBQ)
> +FUNCTION_ONLY_N_NO_U_F (vqshruntq, VQSHRUNTQ)
>  FUNCTION_WITH_M_N_NO_F (vqsubq, VQSUBQ)
>  FUNCTION (vreinterpretq, vreinterpretq_impl,)
>  FUNCTION_WITHOUT_N_NO_F (vrhaddq, VRHADDQ)
> diff --git a/gcc/config/arm/arm-mve-builtins-base.def b/gcc/config/arm/arm-
> mve-builtins-base.def
> index 3dd40086663..50cb2d055e9 100644
> --- a/gcc/config/arm/arm-mve-builtins-base.def
> +++ b/gcc/config/arm/arm-mve-builtins-base.def
> @@ -36,10 +36,14 @@ DEF_MVE_FUNCTION (vqrdmulhq, binary_opt_n,
> all_signed, m_or_none)
>  DEF_MVE_FUNCTION (vqrshlq, binary_round_lshift, all_integer, m_or_none)
>  DEF_MVE_FUNCTION (vqrshrnbq, binary_rshift_narrow, integer_16_32,
> m_or_none)
>  DEF_MVE_FUNCTION (vqrshrntq, binary_rshift_narrow, integer_16_32,
> m_or_none)
> +DEF_MVE_FUNCTION (vqrshrunbq, binary_rshift_narrow_unsigned,
> signed_16_32, m_or_none)
> +DEF_MVE_FUNCTION (vqrshruntq, binary_rshift_narrow_unsigned,
> signed_16_32, m_or_none)
>  DEF_MVE_FUNCTION (vqshlq, binary_lshift, all_integer, m_or_none)
>  DEF_MVE_FUNCTION (vqshlq, binary_lshift_r, all_integer, m_or_none)
>  DEF_MVE_FUNCTION (vqshrnbq, binary_rshift_narrow, integer_16_32,
> m_or_none)
>  DEF_MVE_FUNCTION (vqshrntq, binary_rshift_narrow, integer_16_32,
> m_or_none)
> +DEF_MVE_FUNCTION (vqshrunbq, binary_rshift_narrow_unsigned,
> signed_16_32, m_or_none)
> +DEF_MVE_FUNCTION (vqshruntq, binary_rshift_narrow_unsigned,
> signed_16_32, m_or_none)
>  DEF_MVE_FUNCTION (vqsubq, binary_opt_n, all_integer, m_or_none)
>  DEF_MVE_FUNCTION (vreinterpretq, unary_convert, reinterpret_integer,
> none)
>  DEF_MVE_FUNCTION (vrhaddq, binary, all_integer, mx_or_none)
> diff --git a/gcc/config/arm/arm-mve-builtins-base.h b/gcc/config/arm/arm-
> mve-builtins-base.h
> index 9e11ac83681..fcac772bc5b 100644
> --- a/gcc/config/arm/arm-mve-builtins-base.h
> +++ b/gcc/config/arm/arm-mve-builtins-base.h
> @@ -41,9 +41,13 @@ extern const function_base *const vqrdmulhq;
>  extern const function_base *const vqrshlq;
>  extern const function_base *const vqrshrnbq;
>  extern const function_base *const vqrshrntq;
> +extern const function_base *const vqrshrunbq;
> +extern const function_base *const vqrshruntq;
>  extern const function_base *const vqshlq;
>  extern const function_base *const vqshrnbq;
>  extern const function_base *const vqshrntq;
> +extern const function_base *const vqshrunbq;
> +extern const function_base *const vqshruntq;
>  extern const function_base *const vqsubq;
>  extern const function_base *const vreinterpretq;
>  extern const function_base *const vrhaddq;
> diff --git a/gcc/config/arm/arm-mve-builtins.cc b/gcc/config/arm/arm-mve-
> builtins.cc
> index 667bbc58483..4fc6160a794 100644
> --- a/gcc/config/arm/arm-mve-builtins.cc
> +++ b/gcc/config/arm/arm-mve-builtins.cc
> @@ -674,8 +674,12 @@ function_instance::has_inactive_argument () const
>        || (base == functions::vqrshlq && mode_suffix_id == MODE_n)
>        || base == functions::vqrshrnbq
>        || base == functions::vqrshrntq
> +      || base == functions::vqrshrunbq
> +      || base == functions::vqrshruntq
>        || base == functions::vqshrnbq
>        || base == functions::vqshrntq
> +      || base == functions::vqshrunbq
> +      || base == functions::vqshruntq
>        || (base == functions::vrshlq && mode_suffix_id == MODE_n)
>        || base == functions::vrshrnbq
>        || base == functions::vrshrntq
> diff --git a/gcc/config/arm/arm_mve.h b/gcc/config/arm/arm_mve.h
> index ed7852e2460..b2701f1135d 100644
> --- a/gcc/config/arm/arm_mve.h
> +++ b/gcc/config/arm/arm_mve.h
> @@ -113,7 +113,6 @@
>  #define vrmlaldavhxq(__a, __b) __arm_vrmlaldavhxq(__a, __b)
>  #define vabavq(__a, __b, __c) __arm_vabavq(__a, __b, __c)
>  #define vbicq_m_n(__a, __imm, __p) __arm_vbicq_m_n(__a, __imm, __p)
> -#define vqrshrunbq(__a, __b, __imm) __arm_vqrshrunbq(__a, __b, __imm)
>  #define vrmlaldavhaq(__a, __b, __c) __arm_vrmlaldavhaq(__a, __b, __c)
>  #define vshlcq(__a, __b, __imm) __arm_vshlcq(__a, __b, __imm)
>  #define vpselq(__a, __b, __p) __arm_vpselq(__a, __b, __p)
> @@ -190,9 +189,6 @@
>  #define vqmovnbq_m(__a, __b, __p) __arm_vqmovnbq_m(__a, __b, __p)
>  #define vqmovntq_m(__a, __b, __p) __arm_vqmovntq_m(__a, __b, __p)
>  #define vrev32q_m(__inactive, __a, __p) __arm_vrev32q_m(__inactive, __a,
> __p)
> -#define vqrshruntq(__a, __b, __imm) __arm_vqrshruntq(__a, __b, __imm)
> -#define vqshrunbq(__a, __b, __imm) __arm_vqshrunbq(__a, __b, __imm)
> -#define vqshruntq(__a, __b, __imm) __arm_vqshruntq(__a, __b, __imm)
>  #define vqmovunbq_m(__a, __b, __p) __arm_vqmovunbq_m(__a, __b, __p)
>  #define vqmovuntq_m(__a, __b, __p) __arm_vqmovuntq_m(__a, __b, __p)
>  #define vsriq_m(__a, __b, __imm, __p) __arm_vsriq_m(__a, __b, __imm,
> __p)
> @@ -236,10 +232,6 @@
>  #define vmulltq_poly_m(__inactive, __a, __b, __p)
> __arm_vmulltq_poly_m(__inactive, __a, __b, __p)
>  #define vqdmullbq_m(__inactive, __a, __b, __p)
> __arm_vqdmullbq_m(__inactive, __a, __b, __p)
>  #define vqdmulltq_m(__inactive, __a, __b, __p)
> __arm_vqdmulltq_m(__inactive, __a, __b, __p)
> -#define vqrshrunbq_m(__a, __b, __imm, __p) __arm_vqrshrunbq_m(__a,
> __b, __imm, __p)
> -#define vqrshruntq_m(__a, __b, __imm, __p) __arm_vqrshruntq_m(__a,
> __b, __imm, __p)
> -#define vqshrunbq_m(__a, __b, __imm, __p) __arm_vqshrunbq_m(__a,
> __b, __imm, __p)
> -#define vqshruntq_m(__a, __b, __imm, __p) __arm_vqshruntq_m(__a, __b,
> __imm, __p)
>  #define vrmlaldavhaq_p(__a, __b, __c, __p) __arm_vrmlaldavhaq_p(__a,
> __b, __c, __p)
>  #define vrmlaldavhaxq_p(__a, __b, __c, __p) __arm_vrmlaldavhaxq_p(__a,
> __b, __c, __p)
>  #define vrmlsldavhaq_p(__a, __b, __c, __p) __arm_vrmlsldavhaq_p(__a,
> __b, __c, __p)
> @@ -889,8 +881,6 @@
>  #define vcvtq_m_f16_u16(__inactive, __a, __p)
> __arm_vcvtq_m_f16_u16(__inactive, __a, __p)
>  #define vcvtq_m_f32_s32(__inactive, __a, __p)
> __arm_vcvtq_m_f32_s32(__inactive, __a, __p)
>  #define vcvtq_m_f32_u32(__inactive, __a, __p)
> __arm_vcvtq_m_f32_u32(__inactive, __a, __p)
> -#define vqrshrunbq_n_s16(__a, __b,  __imm)
> __arm_vqrshrunbq_n_s16(__a, __b,  __imm)
> -#define vqrshrunbq_n_s32(__a, __b,  __imm)
> __arm_vqrshrunbq_n_s32(__a, __b,  __imm)
>  #define vrmlaldavhaq_s32(__a, __b, __c) __arm_vrmlaldavhaq_s32(__a,
> __b, __c)
>  #define vrmlaldavhaq_u32(__a, __b, __c) __arm_vrmlaldavhaq_u32(__a,
> __b, __c)
>  #define vshlcq_s8(__a,  __b,  __imm) __arm_vshlcq_s8(__a,  __b,  __imm)
> @@ -1203,9 +1193,6 @@
>  #define vcmpneq_m_f16(__a, __b, __p) __arm_vcmpneq_m_f16(__a, __b,
> __p)
>  #define vcmpneq_m_n_f16(__a, __b, __p) __arm_vcmpneq_m_n_f16(__a,
> __b, __p)
>  #define vmvnq_m_n_u16(__inactive,  __imm, __p)
> __arm_vmvnq_m_n_u16(__inactive,  __imm, __p)
> -#define vqrshruntq_n_s16(__a, __b,  __imm) __arm_vqrshruntq_n_s16(__a,
> __b,  __imm)
> -#define vqshrunbq_n_s16(__a, __b,  __imm) __arm_vqshrunbq_n_s16(__a,
> __b,  __imm)
> -#define vqshruntq_n_s16(__a, __b,  __imm) __arm_vqshruntq_n_s16(__a,
> __b,  __imm)
>  #define vcvtmq_m_u16_f16(__inactive, __a, __p)
> __arm_vcvtmq_m_u16_f16(__inactive, __a, __p)
>  #define vcvtnq_m_u16_f16(__inactive, __a, __p)
> __arm_vcvtnq_m_u16_f16(__inactive, __a, __p)
>  #define vcvtpq_m_u16_f16(__inactive, __a, __p)
> __arm_vcvtpq_m_u16_f16(__inactive, __a, __p)
> @@ -1278,9 +1265,6 @@
>  #define vcmpneq_m_f32(__a, __b, __p) __arm_vcmpneq_m_f32(__a, __b,
> __p)
>  #define vcmpneq_m_n_f32(__a, __b, __p) __arm_vcmpneq_m_n_f32(__a,
> __b, __p)
>  #define vmvnq_m_n_u32(__inactive,  __imm, __p)
> __arm_vmvnq_m_n_u32(__inactive,  __imm, __p)
> -#define vqrshruntq_n_s32(__a, __b,  __imm) __arm_vqrshruntq_n_s32(__a,
> __b,  __imm)
> -#define vqshrunbq_n_s32(__a, __b,  __imm) __arm_vqshrunbq_n_s32(__a,
> __b,  __imm)
> -#define vqshruntq_n_s32(__a, __b,  __imm) __arm_vqshruntq_n_s32(__a,
> __b,  __imm)
>  #define vcvtmq_m_u32_f32(__inactive, __a, __p)
> __arm_vcvtmq_m_u32_f32(__inactive, __a, __p)
>  #define vcvtnq_m_u32_f32(__inactive, __a, __p)
> __arm_vcvtnq_m_u32_f32(__inactive, __a, __p)
>  #define vcvtpq_m_u32_f32(__inactive, __a, __p)
> __arm_vcvtpq_m_u32_f32(__inactive, __a, __p)
> @@ -1466,14 +1450,6 @@
>  #define vqdmulltq_m_n_s16(__inactive, __a, __b, __p)
> __arm_vqdmulltq_m_n_s16(__inactive, __a, __b, __p)
>  #define vqdmulltq_m_s32(__inactive, __a, __b, __p)
> __arm_vqdmulltq_m_s32(__inactive, __a, __b, __p)
>  #define vqdmulltq_m_s16(__inactive, __a, __b, __p)
> __arm_vqdmulltq_m_s16(__inactive, __a, __b, __p)
> -#define vqrshrunbq_m_n_s32(__a, __b,  __imm, __p)
> __arm_vqrshrunbq_m_n_s32(__a, __b,  __imm, __p)
> -#define vqrshrunbq_m_n_s16(__a, __b,  __imm, __p)
> __arm_vqrshrunbq_m_n_s16(__a, __b,  __imm, __p)
> -#define vqrshruntq_m_n_s32(__a, __b,  __imm, __p)
> __arm_vqrshruntq_m_n_s32(__a, __b,  __imm, __p)
> -#define vqrshruntq_m_n_s16(__a, __b,  __imm, __p)
> __arm_vqrshruntq_m_n_s16(__a, __b,  __imm, __p)
> -#define vqshrunbq_m_n_s32(__a, __b,  __imm, __p)
> __arm_vqshrunbq_m_n_s32(__a, __b,  __imm, __p)
> -#define vqshrunbq_m_n_s16(__a, __b,  __imm, __p)
> __arm_vqshrunbq_m_n_s16(__a, __b,  __imm, __p)
> -#define vqshruntq_m_n_s32(__a, __b,  __imm, __p)
> __arm_vqshruntq_m_n_s32(__a, __b,  __imm, __p)
> -#define vqshruntq_m_n_s16(__a, __b,  __imm, __p)
> __arm_vqshruntq_m_n_s16(__a, __b,  __imm, __p)
>  #define vrmlaldavhaq_p_s32(__a, __b, __c, __p)
> __arm_vrmlaldavhaq_p_s32(__a, __b, __c, __p)
>  #define vrmlaldavhaq_p_u32(__a, __b, __c, __p)
> __arm_vrmlaldavhaq_p_u32(__a, __b, __c, __p)
>  #define vrmlaldavhaxq_p_s32(__a, __b, __c, __p)
> __arm_vrmlaldavhaxq_p_s32(__a, __b, __c, __p)
> @@ -4445,20 +4421,6 @@ __arm_vbicq_m_n_u32 (uint32x4_t __a, const int
> __imm, mve_pred16_t __p)
>    return __builtin_mve_vbicq_m_n_uv4si (__a, __imm, __p);
>  }
> 
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshrunbq_n_s16 (uint8x16_t __a, int16x8_t __b, const int __imm)
> -{
> -  return __builtin_mve_vqrshrunbq_n_sv8hi (__a, __b, __imm);
> -}
> -
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshrunbq_n_s32 (uint16x8_t __a, int32x4_t __b, const int __imm)
> -{
> -  return __builtin_mve_vqrshrunbq_n_sv4si (__a, __b, __imm);
> -}
> -
>  __extension__ extern __inline int64_t
>  __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
>  __arm_vrmlaldavhaq_s32 (int64_t __a, int32x4_t __b, int32x4_t __c)
> @@ -6320,27 +6282,6 @@ __arm_vmvnq_m_n_u16 (uint16x8_t __inactive,
> const int __imm, mve_pred16_t __p)
>    return __builtin_mve_vmvnq_m_n_uv8hi (__inactive, __imm, __p);
>  }
> 
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshruntq_n_s16 (uint8x16_t __a, int16x8_t __b, const int __imm)
> -{
> -  return __builtin_mve_vqrshruntq_n_sv8hi (__a, __b, __imm);
> -}
> -
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshrunbq_n_s16 (uint8x16_t __a, int16x8_t __b, const int __imm)
> -{
> -  return __builtin_mve_vqshrunbq_n_sv8hi (__a, __b, __imm);
> -}
> -
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshruntq_n_s16 (uint8x16_t __a, int16x8_t __b, const int __imm)
> -{
> -  return __builtin_mve_vqshruntq_n_sv8hi (__a, __b, __imm);
> -}
> -
>  __extension__ extern __inline uint8x16_t
>  __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
>  __arm_vqmovunbq_m_s16 (uint8x16_t __a, int16x8_t __b, mve_pred16_t
> __p)
> @@ -6537,27 +6478,6 @@ __arm_vmvnq_m_n_u32 (uint32x4_t __inactive,
> const int __imm, mve_pred16_t __p)
>    return __builtin_mve_vmvnq_m_n_uv4si (__inactive, __imm, __p);
>  }
> 
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshruntq_n_s32 (uint16x8_t __a, int32x4_t __b, const int __imm)
> -{
> -  return __builtin_mve_vqrshruntq_n_sv4si (__a, __b, __imm);
> -}
> -
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshrunbq_n_s32 (uint16x8_t __a, int32x4_t __b, const int __imm)
> -{
> -  return __builtin_mve_vqshrunbq_n_sv4si (__a, __b, __imm);
> -}
> -
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshruntq_n_s32 (uint16x8_t __a, int32x4_t __b, const int __imm)
> -{
> -  return __builtin_mve_vqshruntq_n_sv4si (__a, __b, __imm);
> -}
> -
>  __extension__ extern __inline uint16x8_t
>  __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
>  __arm_vqmovunbq_m_s32 (uint16x8_t __a, int32x4_t __b, mve_pred16_t
> __p)
> @@ -7797,62 +7717,6 @@ __arm_vqdmulltq_m_s16 (int32x4_t __inactive,
> int16x8_t __a, int16x8_t __b, mve_p
>    return __builtin_mve_vqdmulltq_m_sv8hi (__inactive, __a, __b, __p);
>  }
> 
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshrunbq_m_n_s32 (uint16x8_t __a, int32x4_t __b, const int
> __imm, mve_pred16_t __p)
> -{
> -  return __builtin_mve_vqrshrunbq_m_n_sv4si (__a, __b, __imm, __p);
> -}
> -
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshrunbq_m_n_s16 (uint8x16_t __a, int16x8_t __b, const int
> __imm, mve_pred16_t __p)
> -{
> -  return __builtin_mve_vqrshrunbq_m_n_sv8hi (__a, __b, __imm, __p);
> -}
> -
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshruntq_m_n_s32 (uint16x8_t __a, int32x4_t __b, const int
> __imm, mve_pred16_t __p)
> -{
> -  return __builtin_mve_vqrshruntq_m_n_sv4si (__a, __b, __imm, __p);
> -}
> -
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshruntq_m_n_s16 (uint8x16_t __a, int16x8_t __b, const int
> __imm, mve_pred16_t __p)
> -{
> -  return __builtin_mve_vqrshruntq_m_n_sv8hi (__a, __b, __imm, __p);
> -}
> -
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshrunbq_m_n_s32 (uint16x8_t __a, int32x4_t __b, const int
> __imm, mve_pred16_t __p)
> -{
> -  return __builtin_mve_vqshrunbq_m_n_sv4si (__a, __b, __imm, __p);
> -}
> -
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshrunbq_m_n_s16 (uint8x16_t __a, int16x8_t __b, const int
> __imm, mve_pred16_t __p)
> -{
> -  return __builtin_mve_vqshrunbq_m_n_sv8hi (__a, __b, __imm, __p);
> -}
> -
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshruntq_m_n_s32 (uint16x8_t __a, int32x4_t __b, const int __imm,
> mve_pred16_t __p)
> -{
> -  return __builtin_mve_vqshruntq_m_n_sv4si (__a, __b, __imm, __p);
> -}
> -
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshruntq_m_n_s16 (uint8x16_t __a, int16x8_t __b, const int __imm,
> mve_pred16_t __p)
> -{
> -  return __builtin_mve_vqshruntq_m_n_sv8hi (__a, __b, __imm, __p);
> -}
> -
>  __extension__ extern __inline int64_t
>  __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
>  __arm_vrmlaldavhaq_p_s32 (int64_t __a, int32x4_t __b, int32x4_t __c,
> mve_pred16_t __p)
> @@ -16398,20 +16262,6 @@ __arm_vbicq_m_n (uint32x4_t __a, const int
> __imm, mve_pred16_t __p)
>   return __arm_vbicq_m_n_u32 (__a, __imm, __p);
>  }
> 
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshrunbq (uint8x16_t __a, int16x8_t __b, const int __imm)
> -{
> - return __arm_vqrshrunbq_n_s16 (__a, __b, __imm);
> -}
> -
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshrunbq (uint16x8_t __a, int32x4_t __b, const int __imm)
> -{
> - return __arm_vqrshrunbq_n_s32 (__a, __b, __imm);
> -}
> -
>  __extension__ extern __inline int64_t
>  __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
>  __arm_vrmlaldavhaq (int64_t __a, int32x4_t __b, int32x4_t __c)
> @@ -18260,27 +18110,6 @@ __arm_vmvnq_m (uint16x8_t __inactive, const
> int __imm, mve_pred16_t __p)
>   return __arm_vmvnq_m_n_u16 (__inactive, __imm, __p);
>  }
> 
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshruntq (uint8x16_t __a, int16x8_t __b, const int __imm)
> -{
> - return __arm_vqrshruntq_n_s16 (__a, __b, __imm);
> -}
> -
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshrunbq (uint8x16_t __a, int16x8_t __b, const int __imm)
> -{
> - return __arm_vqshrunbq_n_s16 (__a, __b, __imm);
> -}
> -
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshruntq (uint8x16_t __a, int16x8_t __b, const int __imm)
> -{
> - return __arm_vqshruntq_n_s16 (__a, __b, __imm);
> -}
> -
>  __extension__ extern __inline uint8x16_t
>  __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
>  __arm_vqmovunbq_m (uint8x16_t __a, int16x8_t __b, mve_pred16_t __p)
> @@ -18477,27 +18306,6 @@ __arm_vmvnq_m (uint32x4_t __inactive, const
> int __imm, mve_pred16_t __p)
>   return __arm_vmvnq_m_n_u32 (__inactive, __imm, __p);
>  }
> 
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshruntq (uint16x8_t __a, int32x4_t __b, const int __imm)
> -{
> - return __arm_vqrshruntq_n_s32 (__a, __b, __imm);
> -}
> -
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshrunbq (uint16x8_t __a, int32x4_t __b, const int __imm)
> -{
> - return __arm_vqshrunbq_n_s32 (__a, __b, __imm);
> -}
> -
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshruntq (uint16x8_t __a, int32x4_t __b, const int __imm)
> -{
> - return __arm_vqshruntq_n_s32 (__a, __b, __imm);
> -}
> -
>  __extension__ extern __inline uint16x8_t
>  __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
>  __arm_vqmovunbq_m (uint16x8_t __a, int32x4_t __b, mve_pred16_t __p)
> @@ -19737,62 +19545,6 @@ __arm_vqdmulltq_m (int32x4_t __inactive,
> int16x8_t __a, int16x8_t __b, mve_pred1
>   return __arm_vqdmulltq_m_s16 (__inactive, __a, __b, __p);
>  }
> 
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshrunbq_m (uint16x8_t __a, int32x4_t __b, const int __imm,
> mve_pred16_t __p)
> -{
> - return __arm_vqrshrunbq_m_n_s32 (__a, __b, __imm, __p);
> -}
> -
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshrunbq_m (uint8x16_t __a, int16x8_t __b, const int __imm,
> mve_pred16_t __p)
> -{
> - return __arm_vqrshrunbq_m_n_s16 (__a, __b, __imm, __p);
> -}
> -
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshruntq_m (uint16x8_t __a, int32x4_t __b, const int __imm,
> mve_pred16_t __p)
> -{
> - return __arm_vqrshruntq_m_n_s32 (__a, __b, __imm, __p);
> -}
> -
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqrshruntq_m (uint8x16_t __a, int16x8_t __b, const int __imm,
> mve_pred16_t __p)
> -{
> - return __arm_vqrshruntq_m_n_s16 (__a, __b, __imm, __p);
> -}
> -
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshrunbq_m (uint16x8_t __a, int32x4_t __b, const int __imm,
> mve_pred16_t __p)
> -{
> - return __arm_vqshrunbq_m_n_s32 (__a, __b, __imm, __p);
> -}
> -
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshrunbq_m (uint8x16_t __a, int16x8_t __b, const int __imm,
> mve_pred16_t __p)
> -{
> - return __arm_vqshrunbq_m_n_s16 (__a, __b, __imm, __p);
> -}
> -
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshruntq_m (uint16x8_t __a, int32x4_t __b, const int __imm,
> mve_pred16_t __p)
> -{
> - return __arm_vqshruntq_m_n_s32 (__a, __b, __imm, __p);
> -}
> -
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vqshruntq_m (uint8x16_t __a, int16x8_t __b, const int __imm,
> mve_pred16_t __p)
> -{
> - return __arm_vqshruntq_m_n_s16 (__a, __b, __imm, __p);
> -}
> -
>  __extension__ extern __inline int64_t
>  __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
>  __arm_vrmlaldavhaq_p (int64_t __a, int32x4_t __b, int32x4_t __c,
> mve_pred16_t __p)
> @@ -25799,12 +25551,6 @@ extern void *__ARM_undef;
>    int (*)[__ARM_mve_type_uint16x8_t]: __arm_vbicq_m_n_u16
> (__ARM_mve_coerce(__p0, uint16x8_t), p1, p2), \
>    int (*)[__ARM_mve_type_uint32x4_t]: __arm_vbicq_m_n_u32
> (__ARM_mve_coerce(__p0, uint32x4_t), p1, p2));})
> 
> -#define __arm_vqrshrunbq(p0,p1,p2) ({ __typeof(p0) __p0 = (p0); \
> -  __typeof(p1) __p1 = (p1); \
> -  _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> -  int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int16x8_t]:
> __arm_vqrshrunbq_n_s16 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, int16x8_t), p2), \
> -  int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int32x4_t]:
> __arm_vqrshrunbq_n_s32 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, int32x4_t), p2));})
> -
>  #define __arm_vshlcq(p0,p1,p2) ({ __typeof(p0) __p0 = (p0); \
>    _Generic( (int (*)[__ARM_mve_typeid(__p0)])0, \
>    int (*)[__ARM_mve_type_int8x16_t]: __arm_vshlcq_s8
> (__ARM_mve_coerce(__p0, int8x16_t), p1, p2), \
> @@ -26364,18 +26110,6 @@ extern void *__ARM_undef;
>    int (*)[__ARM_mve_type_int8x16_t][__ARM_mve_type_int8x16_t]:
> __arm_vrev16q_m_s8 (__ARM_mve_coerce(__p0, int8x16_t),
> __ARM_mve_coerce(__p1, int8x16_t), p2), \
>    int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_uint8x16_t]:
> __arm_vrev16q_m_u8 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, uint8x16_t), p2));})
> 
> -#define __arm_vqshruntq(p0,p1,p2) ({ __typeof(p0) __p0 = (p0); \
> -  __typeof(p1) __p1 = (p1); \
> -  _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> -  int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int16x8_t]:
> __arm_vqshruntq_n_s16 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, int16x8_t), p2), \
> -  int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int32x4_t]:
> __arm_vqshruntq_n_s32 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, int32x4_t), p2));})
> -
> -#define __arm_vqrshruntq(p0,p1,p2) ({ __typeof(p0) __p0 = (p0); \
> -  __typeof(p1) __p1 = (p1); \
> -  _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> -  int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int16x8_t]:
> __arm_vqrshruntq_n_s16 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, int16x8_t), p2), \
> -  int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int32x4_t]:
> __arm_vqrshruntq_n_s32 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, int32x4_t), p2));})
> -
>  #define __arm_vqmovnbq_m(p0,p1,p2) ({ __typeof(p0) __p0 = (p0); \
>    __typeof(p1) __p1 = (p1); \
>    _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> @@ -26404,12 +26138,6 @@ extern void *__ARM_undef;
>    int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int16x8_t]:
> __arm_vqmovuntq_m_s16 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, int16x8_t), p2), \
>    int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int32x4_t]:
> __arm_vqmovuntq_m_s32 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, int32x4_t), p2));})
> 
> -#define __arm_vqrshruntq(p0,p1,p2) ({ __typeof(p0) __p0 = (p0); \
> -  __typeof(p1) __p1 = (p1); \
> -  _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> -  int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int16x8_t]:
> __arm_vqrshruntq_n_s16 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, int16x8_t), p2), \
> -  int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int32x4_t]:
> __arm_vqrshruntq_n_s32 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, int32x4_t), p2));})
> -
>  #define __arm_vnegq_m(p0,p1,p2) ({ __typeof(p0) __p0 = (p0); \
>    __typeof(p1) __p1 = (p1); \
>    _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> @@ -27544,12 +27272,6 @@ extern void *__ARM_undef;
>    int (*)[__ARM_mve_type_uint16x8_t]: __arm_vbicq_m_n_u16
> (__ARM_mve_coerce(__p0, uint16x8_t), p1, p2), \
>    int (*)[__ARM_mve_type_uint32x4_t]: __arm_vbicq_m_n_u32
> (__ARM_mve_coerce(__p0, uint32x4_t), p1, p2));})
> 
> -#define __arm_vqrshrunbq(p0,p1,p2) ({ __typeof(p0) __p0 = (p0); \
> -  __typeof(p1) __p1 = (p1); \
> -  _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> -  int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int16x8_t]:
> __arm_vqrshrunbq_n_s16 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, int16x8_t), p2), \
> -  int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int32x4_t]:
> __arm_vqrshrunbq_n_s32 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, int32x4_t), p2));})
> -
>  #define __arm_vqrdmlsdhq(p0,p1,p2) ({ __typeof(p0) __p0 = (p0); \
>    __typeof(p1) __p1 = (p1); \
>    __typeof(p2) __p2 = (p2); \
> @@ -27861,24 +27583,12 @@ extern void *__ARM_undef;
>    int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_uint8x16_t]:
> __arm_vrev32q_m_u8 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, uint8x16_t), p2), \
>    int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_uint16x8_t]:
> __arm_vrev32q_m_u16 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, uint16x8_t), p2));})
> 
> -#define __arm_vqshruntq(p0,p1,p2) ({ __typeof(p0) __p0 = (p0); \
> -  __typeof(p1) __p1 = (p1); \
> -  _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> -  int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int16x8_t]:
> __arm_vqshruntq_n_s16 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, int16x8_t), p2), \
> -  int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int32x4_t]:
> __arm_vqshruntq_n_s32 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, int32x4_t), p2));})
> -
>  #define __arm_vrev16q_m(p0,p1,p2) ({ __typeof(p0) __p0 = (p0); \
>    __typeof(p1) __p1 = (p1); \
>    _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
>    int (*)[__ARM_mve_type_int8x16_t][__ARM_mve_type_int8x16_t]:
> __arm_vrev16q_m_s8 (__ARM_mve_coerce(__p0, int8x16_t),
> __ARM_mve_coerce(__p1, int8x16_t), p2), \
>    int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_uint8x16_t]:
> __arm_vrev16q_m_u8 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, uint8x16_t), p2));})
> 
> -#define __arm_vqrshruntq(p0,p1,p2) ({ __typeof(p0) __p0 = (p0); \
> -  __typeof(p1) __p1 = (p1); \
> -  _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> -  int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int16x8_t]:
> __arm_vqrshruntq_n_s16 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, int16x8_t), p2), \
> -  int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int32x4_t]:
> __arm_vqrshruntq_n_s32 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, int32x4_t), p2));})
> -
>  #define __arm_vqmovuntq_m(p0,p1,p2) ({ __typeof(p0) __p0 = (p0); \
>    __typeof(p1) __p1 = (p1); \
>    _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> @@ -28718,30 +28428,6 @@ extern void *__ARM_undef;
>    int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_uint8x16_t]:
> __arm_vshlltq_m_n_u8 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, uint8x16_t), p2, p3), \
>    int (*)[__ARM_mve_type_uint32x4_t][__ARM_mve_type_uint16x8_t]:
> __arm_vshlltq_m_n_u16 (__ARM_mve_coerce(__p0, uint32x4_t),
> __ARM_mve_coerce(__p1, uint16x8_t), p2, p3));})
> 
> -#define __arm_vqshruntq_m(p0,p1,p2,p3) ({ __typeof(p0) __p0 = (p0); \
> -  __typeof(p1) __p1 = (p1); \
> -  _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> -  int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int16x8_t]:
> __arm_vqshruntq_m_n_s16 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, int16x8_t), p2, p3), \
> -  int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int32x4_t]:
> __arm_vqshruntq_m_n_s32 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, int32x4_t), p2, p3));})
> -
> -#define __arm_vqshrunbq_m(p0,p1,p2,p3) ({ __typeof(p0) __p0 = (p0); \
> -  __typeof(p1) __p1 = (p1); \
> -  _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> -  int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int16x8_t]:
> __arm_vqshrunbq_m_n_s16 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, int16x8_t), p2, p3), \
> -  int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int32x4_t]:
> __arm_vqshrunbq_m_n_s32 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, int32x4_t), p2, p3));})
> -
> -#define __arm_vqrshrunbq_m(p0,p1,p2,p3) ({ __typeof(p0) __p0 = (p0); \
> -  __typeof(p1) __p1 = (p1); \
> -  _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> -  int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int16x8_t]:
> __arm_vqrshrunbq_m_n_s16 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, int16x8_t), p2, p3), \
> -  int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int32x4_t]:
> __arm_vqrshrunbq_m_n_s32 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, int32x4_t), p2, p3));})
> -
> -#define __arm_vqrshruntq_m(p0,p1,p2,p3) ({ __typeof(p0) __p0 = (p0); \
> -  __typeof(p1) __p1 = (p1); \
> -  _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> -  int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int16x8_t]:
> __arm_vqrshruntq_m_n_s16 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, int16x8_t), p2, p3), \
> -  int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int32x4_t]:
> __arm_vqrshruntq_m_n_s32 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, int32x4_t), p2, p3));})
> -
>  #define __arm_vmlaldavaq_p(p0,p1,p2,p3) ({ __typeof(p0) __p0 = (p0); \
>    __typeof(p1) __p1 = (p1); \
>    __typeof(p2) __p2 = (p2); \
> @@ -28831,12 +28517,6 @@ extern void *__ARM_undef;
>    int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int_n]:
> __arm_vmvnq_m_n_u16 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce1(__p1, int) , p2), \
>    int (*)[__ARM_mve_type_uint32x4_t][__ARM_mve_type_int_n]:
> __arm_vmvnq_m_n_u32 (__ARM_mve_coerce(__p0, uint32x4_t),
> __ARM_mve_coerce1(__p1, int) , p2));})
> 
> -#define __arm_vqshrunbq(p0,p1,p2) ({ __typeof(p0) __p0 = (p0); \
> -  __typeof(p1) __p1 = (p1); \
> -  _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> -  int (*)[__ARM_mve_type_uint8x16_t][__ARM_mve_type_int16x8_t]:
> __arm_vqshrunbq_n_s16 (__ARM_mve_coerce(__p0, uint8x16_t),
> __ARM_mve_coerce(__p1, int16x8_t), p2), \
> -  int (*)[__ARM_mve_type_uint16x8_t][__ARM_mve_type_int32x4_t]:
> __arm_vqshrunbq_n_s32 (__ARM_mve_coerce(__p0, uint16x8_t),
> __ARM_mve_coerce(__p1, int32x4_t), p2));})
> -
>  #define __arm_vqshluq_m(p0,p1,p2,p3) ({ __typeof(p0) __p0 = (p0); \
>    __typeof(p1) __p1 = (p1); \
>    _Generic( (int (*)[__ARM_mve_typeid(__p0)][__ARM_mve_typeid(__p1)])0,
> \
> --
> 2.34.1