> -----Original Message-----
> From: Andrea Corallo <andrea.cora...@arm.com>
> Sent: Thursday, November 17, 2022 4:38 PM
> To: gcc-patches@gcc.gnu.org
> Cc: Kyrylo Tkachov <kyrylo.tkac...@arm.com>; Richard Earnshaw
> <richard.earns...@arm.com>; Andrea Corallo <andrea.cora...@arm.com>
> Subject: [PATCH 03/35] arm: improve tests and fix vddupq*
> 
> gcc/ChangeLog:
> 
>       * config/arm/mve.md (mve_vddupq_u<mode>_insn): Fix 'vddup.u'
>       spacing.
>       (mve_vddupq_m_wb_u<mode>_insn): Likewise.
> 
> gcc/testsuite/ChangeLog:
> 
>       * gcc.target/arm/mve/intrinsics/vddupq_m_n_u16.c: Improve test.
>       * gcc.target/arm/mve/intrinsics/vddupq_m_n_u32.c : Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_m_n_u8.c : Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_m_wb_u16.c: Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_m_wb_u32.c : Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_m_wb_u8.c : Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_n_u16.c : Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_n_u32.c: Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_n_u8.c: Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_wb_u16.c : Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_wb_u32.c: Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_wb_u8.c: Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_x_n_u16.c: Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_x_n_u32.c : Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_x_n_u8.c : Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_x_wb_u16.c : Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_x_wb_u32.c : Likewise.
>       * gcc.target/arm/mve/intrinsics/vddupq_x_wb_u8.c : Likewise.

Ok.
Thanks,
Kyrill

> ---
>  gcc/config/arm/mve.md                         |  4 +-
>  .../arm/mve/intrinsics/vddupq_m_n_u16.c       | 42 +++++++++++++--
>  .../arm/mve/intrinsics/vddupq_m_n_u32.c       | 46 +++++++++++++---
>  .../arm/mve/intrinsics/vddupq_m_n_u8.c        | 46 +++++++++++++---
>  .../arm/mve/intrinsics/vddupq_m_wb_u16.c      | 42 +++++++++++++--
>  .../arm/mve/intrinsics/vddupq_m_wb_u32.c      | 46 +++++++++++++---
>  .../arm/mve/intrinsics/vddupq_m_wb_u8.c       | 46 +++++++++++++---
>  .../arm/mve/intrinsics/vddupq_n_u16.c         | 32 ++++++++++--
>  .../arm/mve/intrinsics/vddupq_n_u32.c         | 28 +++++++++-
>  .../arm/mve/intrinsics/vddupq_n_u8.c          | 28 +++++++++-
>  .../arm/mve/intrinsics/vddupq_wb_u16.c        | 32 ++++++++++--
>  .../arm/mve/intrinsics/vddupq_wb_u32.c        | 28 +++++++++-
>  .../arm/mve/intrinsics/vddupq_wb_u8.c         | 28 +++++++++-
>  .../arm/mve/intrinsics/vddupq_x_n_u16.c       | 42 +++++++++++++--
>  .../arm/mve/intrinsics/vddupq_x_n_u32.c       | 46 +++++++++++++---
>  .../arm/mve/intrinsics/vddupq_x_n_u8.c        | 46 +++++++++++++---
>  .../arm/mve/intrinsics/vddupq_x_wb_u16.c      | 52 +++++++++++++++----
>  .../arm/mve/intrinsics/vddupq_x_wb_u32.c      | 52 +++++++++++++++----
>  .../arm/mve/intrinsics/vddupq_x_wb_u8.c       | 52 +++++++++++++++----
>  19 files changed, 642 insertions(+), 96 deletions(-)
> 
> diff --git a/gcc/config/arm/mve.md b/gcc/config/arm/mve.md
> index 62186f124da..1215f845388 100644
> --- a/gcc/config/arm/mve.md
> +++ b/gcc/config/arm/mve.md
> @@ -9043,7 +9043,7 @@ (define_insn "mve_vddupq_u<mode>_insn"
>         (minus:SI (match_dup 2)
>                (match_operand:SI 4 "immediate_operand" "i")))]
>   "TARGET_HAVE_MVE"
> - "vddup.u%#<V_sz_elem>  %q0, %1, %3")
> + "vddup.u%#<V_sz_elem>\t%q0, %1, %3")
> 
>  ;;
>  ;; [vddupq_m_n_u])
> @@ -9079,7 +9079,7 @@ (define_insn
> "mve_vddupq_m_wb_u<mode>_insn"
>         (minus:SI (match_dup 3)
>                (match_operand:SI 6 "immediate_operand" "i")))]
>   "TARGET_HAVE_MVE"
> - "vpst\;\tvddupt.u%#<V_sz_elem>\t%q0, %2, %4"
> + "vpst\;vddupt.u%#<V_sz_elem>\t%q0, %2, %4"
>   [(set_attr "length""8")])
> 
>  ;;
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_n_u16.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_n_u16.c
> index 7332711f6a7..7c8b0152763 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_n_u16.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_n_u16.c
> @@ -1,23 +1,57 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u16      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint16x8_t
>  foo (uint16x8_t inactive, uint32_t a, mve_pred16_t p)
>  {
>    return vddupq_m_n_u16 (inactive, a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u16"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u16      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint16x8_t
>  foo1 (uint16x8_t inactive, uint32_t a, mve_pred16_t p)
>  {
>    return vddupq_m (inactive, a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u16"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u16      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint16x8_t
> +foo2 (uint16x8_t inactive, mve_pred16_t p)
> +{
> +  return vddupq_m (inactive, 1, 1, p);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_n_u32.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_n_u32.c
> index 54ad91f2803..810a1a7e21b 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_n_u32.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_n_u32.c
> @@ -1,23 +1,57 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u32      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint32x4_t
>  foo (uint32x4_t inactive, uint32_t a, mve_pred16_t p)
>  {
> -  return vddupq_m_n_u32 (inactive, a, 4, p);
> +  return vddupq_m_n_u32 (inactive, a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u32"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u32      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint32x4_t
>  foo1 (uint32x4_t inactive, uint32_t a, mve_pred16_t p)
>  {
> -  return vddupq_m (inactive, a, 4, p);
> +  return vddupq_m (inactive, a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u32"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u32      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint32x4_t
> +foo2 (uint32x4_t inactive, mve_pred16_t p)
> +{
> +  return vddupq_m (inactive, 1, 1, p);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_n_u8.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_n_u8.c
> index 3746b5db6e5..6642b9f4b88 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_n_u8.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_n_u8.c
> @@ -1,23 +1,57 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u8       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint8x16_t
>  foo (uint8x16_t inactive, uint32_t a, mve_pred16_t p)
>  {
> -  return vddupq_m_n_u8 (inactive, a, 4, p);
> +  return vddupq_m_n_u8 (inactive, a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u8"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u8       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint8x16_t
>  foo1 (uint8x16_t inactive, uint32_t a, mve_pred16_t p)
>  {
> -  return vddupq_m (inactive, a, 4, p);
> +  return vddupq_m (inactive, a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u8"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u8       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint8x16_t
> +foo2 (uint8x16_t inactive, mve_pred16_t p)
> +{
> +  return vddupq_m (inactive, 1, 1, p);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_wb_u16.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_wb_u16.c
> index 8b5d9e86469..cc6a19516d9 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_wb_u16.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_wb_u16.c
> @@ -1,23 +1,57 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u16      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint16x8_t
>  foo (uint16x8_t inactive, uint32_t *a, mve_pred16_t p)
>  {
>    return vddupq_m_wb_u16 (inactive, a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u16"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u16      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint16x8_t
>  foo1 (uint16x8_t inactive, uint32_t *a, mve_pred16_t p)
>  {
>    return vddupq_m (inactive, a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u16"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u16      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint16x8_t
> +foo2 (uint16x8_t inactive, mve_pred16_t p)
> +{
> +  return vddupq_m (inactive, 1, 1, p);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_wb_u32.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_wb_u32.c
> index 7a8c363ac70..cd6c6f86eea 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_wb_u32.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_wb_u32.c
> @@ -1,23 +1,57 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u32      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint32x4_t
>  foo (uint32x4_t inactive, uint32_t *a, mve_pred16_t p)
>  {
> -  return vddupq_m_wb_u32 (inactive, a, 4, p);
> +  return vddupq_m_wb_u32 (inactive, a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u32"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u32      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint32x4_t
>  foo1 (uint32x4_t inactive, uint32_t *a, mve_pred16_t p)
>  {
> -  return vddupq_m (inactive, a, 4, p);
> +  return vddupq_m (inactive, a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u32"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u32      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint32x4_t
> +foo2 (uint32x4_t inactive, mve_pred16_t p)
> +{
> +  return vddupq_m (inactive, 1, 1, p);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_wb_u8.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_wb_u8.c
> index 45784a5c9cd..fe186e743da 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_wb_u8.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_m_wb_u8.c
> @@ -1,23 +1,57 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u8       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint8x16_t
>  foo (uint8x16_t inactive, uint32_t *a, mve_pred16_t p)
>  {
> -  return vddupq_m_wb_u8 (inactive, a, 4, p);
> +  return vddupq_m_wb_u8 (inactive, a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u8"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u8       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint8x16_t
>  foo1 (uint8x16_t inactive, uint32_t *a, mve_pred16_t p)
>  {
> -  return vddupq_m (inactive, a, 4, p);
> +  return vddupq_m (inactive, a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u8"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u8       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint8x16_t
> +foo2 (uint8x16_t inactive, mve_pred16_t p)
> +{
> +  return vddupq_m (inactive, 1, 1, p);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_n_u16.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_n_u16.c
> index 4684e2af553..2dba2d74b61 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_n_u16.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_n_u16.c
> @@ -1,21 +1,45 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vddup.u16       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint16x8_t
>  foo (uint32_t a)
>  {
> -  return vddupq_n_u16 (a, 4);
> +  return vddupq_n_u16 (a, 1);
>  }
> 
> -/* { dg-final { scan-assembler "vddup.u16"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vddup.u16       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint16x8_t
>  foo1 (uint32_t a)
>  {
> -  return vddupq_u16 (a, 4);
> +  return vddupq_u16 (a, 1);
>  }
> 
> -/* { dg-final { scan-assembler "vddup.u16"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vddup.u16       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint16x8_t
> +foo2 ()
> +{
> +  return vddupq_u16 (1, 1);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_n_u32.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_n_u32.c
> index aeaa83eb6bc..6b5cf6c75b0 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_n_u32.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_n_u32.c
> @@ -1,21 +1,45 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vddup.u32       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint32x4_t
>  foo (uint32_t a)
>  {
>    return vddupq_n_u32 (a, 1);
>  }
> 
> -/* { dg-final { scan-assembler "vddup.u32"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vddup.u32       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint32x4_t
>  foo1 (uint32_t a)
>  {
>    return vddupq_u32 (a, 1);
>  }
> 
> -/* { dg-final { scan-assembler "vddup.u32"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vddup.u32       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint32x4_t
> +foo2 ()
> +{
> +  return vddupq_u32 (1, 1);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_n_u8.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_n_u8.c
> index 255a9f80b6b..174e422f4ef 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_n_u8.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_n_u8.c
> @@ -1,21 +1,45 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vddup.u8        q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint8x16_t
>  foo (uint32_t a)
>  {
>    return vddupq_n_u8 (a, 1);
>  }
> 
> -/* { dg-final { scan-assembler "vddup.u8"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vddup.u8        q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint8x16_t
>  foo1 (uint32_t a)
>  {
>    return vddupq_u8 (a, 1);
>  }
> 
> -/* { dg-final { scan-assembler "vddup.u8"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vddup.u8        q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint8x16_t
> +foo2 ()
> +{
> +  return vddupq_u8 (1, 1);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_wb_u16.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_wb_u16.c
> index 40fc6cf2197..6a471a7f72f 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_wb_u16.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_wb_u16.c
> @@ -1,21 +1,45 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vddup.u16       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint16x8_t
>  foo (uint32_t *a)
>  {
> -  return vddupq_wb_u16 (a, 4);
> +  return vddupq_wb_u16 (a, 1);
>  }
> 
> -/* { dg-final { scan-assembler "vddup.u16"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vddup.u16       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint16x8_t
>  foo1 (uint32_t *a)
>  {
> -  return vddupq_u16 (a, 4);
> +  return vddupq_u16 (a, 1);
>  }
> 
> -/* { dg-final { scan-assembler "vddup.u16"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vddup.u16       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint16x8_t
> +foo2 ()
> +{
> +  return vddupq_u16 (1, 1);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_wb_u32.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_wb_u32.c
> index 09b5b1f2f80..debf420d3e8 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_wb_u32.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_wb_u32.c
> @@ -1,21 +1,45 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vddup.u32       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint32x4_t
>  foo (uint32_t *a)
>  {
>    return vddupq_wb_u32 (a, 1);
>  }
> 
> -/* { dg-final { scan-assembler "vddup.u32"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vddup.u32       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint32x4_t
>  foo1 (uint32_t *a)
>  {
>    return vddupq_u32 (a, 1);
>  }
> 
> -/* { dg-final { scan-assembler "vddup.u32"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vddup.u32       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint32x4_t
> +foo2 ()
> +{
> +  return vddupq_u32 (1, 1);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_wb_u8.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_wb_u8.c
> index 00dfa906748..8e6ef8adccd 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_wb_u8.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_wb_u8.c
> @@ -1,21 +1,45 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vddup.u8        q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint8x16_t
>  foo (uint32_t *a)
>  {
>    return vddupq_wb_u8 (a, 1);
>  }
> 
> -/* { dg-final { scan-assembler "vddup.u8"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vddup.u8        q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint8x16_t
>  foo1 (uint32_t *a)
>  {
>    return vddupq_u8 (a, 1);
>  }
> 
> -/* { dg-final { scan-assembler "vddup.u8"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vddup.u8        q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint8x16_t
> +foo2 ()
> +{
> +  return vddupq_u8 (1, 1);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_n_u16.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_n_u16.c
> index 5b0fc0b6340..1aafaf87b82 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_n_u16.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_n_u16.c
> @@ -1,23 +1,57 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u16      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint16x8_t
>  foo (uint32_t a, mve_pred16_t p)
>  {
>    return vddupq_x_n_u16 (a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u16"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u16      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint16x8_t
>  foo1 (uint32_t a, mve_pred16_t p)
>  {
>    return vddupq_x_u16 (a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u16"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u16      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint16x8_t
> +foo2 (mve_pred16_t p)
> +{
> +  return vddupq_x_u16 (1, 1, p);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_n_u32.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_n_u32.c
> index 66def991b65..2e3e268dbee 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_n_u32.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_n_u32.c
> @@ -1,23 +1,57 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u32      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint32x4_t
>  foo (uint32_t a, mve_pred16_t p)
>  {
> -  return vddupq_x_n_u32 (a, 4, p);
> +  return vddupq_x_n_u32 (a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u32"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u32      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint32x4_t
>  foo1 (uint32_t a, mve_pred16_t p)
>  {
> -  return vddupq_x_u32 (a, 4, p);
> +  return vddupq_x_u32 (a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u32"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u32      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint32x4_t
> +foo2 (mve_pred16_t p)
> +{
> +  return vddupq_x_u32 (1, 1, p);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_n_u8.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_n_u8.c
> index 8ac322ed52d..bdf563a8074 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_n_u8.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_n_u8.c
> @@ -1,23 +1,57 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> +/*
> +**foo:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u8       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint8x16_t
>  foo (uint32_t a, mve_pred16_t p)
>  {
> -  return vddupq_x_n_u8 (a, 4, p);
> +  return vddupq_x_n_u8 (a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u8"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u8       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint8x16_t
>  foo1 (uint32_t a, mve_pred16_t p)
>  {
> -  return vddupq_x_u8 (a, 4, p);
> +  return vddupq_x_u8 (a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u8"  }  } */
> +/*
> +**foo2:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u8       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint8x16_t
> +foo2 (mve_pred16_t p)
> +{
> +  return vddupq_x_u8 (1, 1, p);
> +}
> +
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_wb_u16.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_wb_u16.c
> index 030048f840a..713d8b731c8 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_wb_u16.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_wb_u16.c
> @@ -1,25 +1,57 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> -uint32_t *a;
> -
> +/*
> +**foo:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u16      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint16x8_t
> -foo (mve_pred16_t p)
> +foo (uint32_t *a, mve_pred16_t p)
>  {
> -  return vddupq_x_wb_u16 (a, 2, p);
> +  return vddupq_x_wb_u16 (a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u16"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u16      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint16x8_t
> +foo1 (uint32_t *a, mve_pred16_t p)
> +{
> +  return vddupq_x_u16 (a, 1, p);
> +}
> +
> +/*
> +**foo2:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u16      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint16x8_t
> -foo1 (mve_pred16_t p)
> +foo2 (mve_pred16_t p)
>  {
> -  return vddupq_x_u16 (a, 2, p);
> +  return vddupq_x_u16 (1, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u16"  }  } */
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_wb_u32.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_wb_u32.c
> index 95bf28e4052..9f484b3b8fb 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_wb_u32.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_wb_u32.c
> @@ -1,25 +1,57 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> -uint32_t *a;
> -
> +/*
> +**foo:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u32      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint32x4_t
> -foo (mve_pred16_t p)
> +foo (uint32_t *a, mve_pred16_t p)
>  {
> -  return vddupq_x_wb_u32 (a, 8, p);
> +  return vddupq_x_wb_u32 (a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u32"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u32      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint32x4_t
> +foo1 (uint32_t *a, mve_pred16_t p)
> +{
> +  return vddupq_x_u32 (a, 1, p);
> +}
> +
> +/*
> +**foo2:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u32      q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint32x4_t
> -foo1 (mve_pred16_t p)
> +foo2 (mve_pred16_t p)
>  {
> -  return vddupq_x_u32 (a, 8, p);
> +  return vddupq_x_u32 (1, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u32"  }  } */
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_wb_u8.c
> b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_wb_u8.c
> index 2fe81dded55..aa83bfed125 100644
> --- a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_wb_u8.c
> +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vddupq_x_wb_u8.c
> @@ -1,25 +1,57 @@
>  /* { dg-require-effective-target arm_v8_1m_mve_ok } */
>  /* { dg-add-options arm_v8_1m_mve } */
>  /* { dg-additional-options "-O2" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> 
>  #include "arm_mve.h"
> 
> -uint32_t *a;
> -
> +/*
> +**foo:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u8       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint8x16_t
> -foo (mve_pred16_t p)
> +foo (uint32_t *a, mve_pred16_t p)
>  {
> -  return vddupq_x_wb_u8 (a, 8, p);
> +  return vddupq_x_wb_u8 (a, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u8"  }  } */
> 
> +/*
> +**foo1:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u8       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
> +uint8x16_t
> +foo1 (uint32_t *a, mve_pred16_t p)
> +{
> +  return vddupq_x_u8 (a, 1, p);
> +}
> +
> +/*
> +**foo2:
> +**   ...
> +**   vmsr    p0, (?:ip|fp|r[0-9]+)(?:        @.*|)
> +**   ...
> +**   vpst(?: @.*|)
> +**   ...
> +**   vddupt.u8       q[0-9]+, (?:ip|fp|r[0-9]+), #[0-9]+(?:  @.*|)
> +**   ...
> +*/
>  uint8x16_t
> -foo1 (mve_pred16_t p)
> +foo2 (mve_pred16_t p)
>  {
> -  return vddupq_x_u8 (a, 8, p);
> +  return vddupq_x_u8 (1, 1, p);
>  }
> 
> -/* { dg-final { scan-assembler "vpst" } } */
> -/* { dg-final { scan-assembler "vddupt.u8"  }  } */
> +/* { dg-final { scan-assembler-not "__ARM_undef" } } */
> \ No newline at end of file
> --
> 2.25.1

Reply via email to