[PATCH, GCC/ARM, 5/10] Clear VFP registers with VSCCLRM

Kyrill Tkachov kyrylo.tkachov@foss.arm.com
Tue Nov 12 10:13:00 GMT 2019 

Hi Mihail,

On 10/23/19 10:26 AM, Mihail Ionescu wrote:
> [PATCH, GCC/ARM, 5/10] Clear VFP registers with VSCCLRM
>
> Hi,
>
> === Context ===
>
> This patch is part of a patch series to add support for Armv8.1-M
> Mainline Security Extensions architecture. Its purpose is to improve
> code density of functions with the cmse_nonsecure_entry attribute and
> when calling function with the cmse_nonsecure_call attribute by using
> VSCCLRM to do all the VFP register clearing as well as clearing the VFP
> register.
>
> === Patch description ===
>
> This patch adds a new pattern for the VSCCLRM instruction.
> cmse_clear_registers () is then modified to use the new VSCCLRM
> instruction when targeting Armv8.1-M Mainline, thus, making the Armv8-M
> register clearing code specific to Armv8-M.
>
> Since the VSCCLRM instruction mandates VPR in the register list, the
> pattern is encoded with a parallel which only requires an unspecified
> VUNSPEC_CLRM_VPR constant modelling the APSR clearing. Other expression
> in the parallel are expected to be set expression for clearing the VFP
> registers.
>
I see we don't represent the VPR here as a register and use an UNSPEC to 
represent its clearing.

That's okay for now but when we do add support for it for MVE we'll need 
to adjust the RTL representation here to show its clobbering.


> ChangeLog entry is as follows:
>
> *** gcc/ChangeLog ***
>
> 2019-10-23  Mihail-Calin Ionescu <mihail.ionescu@arm.com>
> 2019-10-23  Thomas Preud'homme <thomas.preudhomme@arm.com>
>
>         * config/arm/arm-protos.h (clear_operation_p): Adapt prototype.
>         * config/arm/arm.c (clear_operation_p): Extend to be able to 
> check a
>         clear_vfp_multiple pattern based on a new vfp parameter.
>         (cmse_clear_registers): Generate VSCCLRM to clear VFP 
> registers when
>         targeting Armv8.1-M Mainline.
>         (cmse_nonsecure_entry_clear_before_return): Clear VFP registers
>         unconditionally when targeting Armv8.1-M Mainline 
> architecture.  Check
>         whether VFP registers are available before looking 
> call_used_regs for a
>         VFP register.
>         * config/arm/predicates.md (clear_multiple_operation): Adapt 
> to change
>         of prototype of clear_operation_p.
>         (clear_vfp_multiple_operation): New predicate.
>         * config/arm/unspecs.md (VUNSPEC_VSCCLRM_VPR): New volatile 
> unspec.
>         * config/arm/vfp.md (clear_vfp_multiple): New define_insn.
>
Ok.

Thanks,

kyrill



> *** gcc/testsuite/ChangeLog ***
>
> 2019-10-23  Mihail-Calin Ionescu <mihail.ionescu@arm.com>
> 2019-10-23  Thomas Preud'homme <thomas.preudhomme@arm.com>
>
>         * gcc.target/arm/cmse/bitfield-1.c: Add check for VSCCLRM.
>         * gcc.target/arm/cmse/bitfield-2.c: Likewise.
>         * gcc.target/arm/cmse/bitfield-3.c: Likewise.
>         * gcc.target/arm/cmse/cmse-1.c: Likewise.
>         * gcc.target/arm/cmse/struct-1.c: Likewise.
>         * gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-13.c: Likewise.
>         * gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-5.c: Likewise.
>         * gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-7.c: Likewise.
>         * gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-8.c: Likewise.
>         * gcc.target/arm/cmse/mainline/8_1m/hard/cmse-13.c: Likewise.
>         * gcc.target/arm/cmse/mainline/8_1m/hard/cmse-5.c: Likewise.
>         * gcc.target/arm/cmse/mainline/8_1m/hard/cmse-7.c: Likewise.
>         * gcc.target/arm/cmse/mainline/8_1m/hard/cmse-8.c: Likewise.
>         * gcc.target/arm/cmse/mainline/8_1m/soft/cmse-5.c: Likewise.
>         * gcc.target/arm/cmse/mainline/8_1m/softfp-sp/cmse-5.c: Likewise.
>         * gcc.target/arm/cmse/mainline/8_1m/softfp/cmse-5.c: Likewise.
>
> Testing: Bootstrapped on arm-linux-gnueabihf and testsuite shows no
> regression.
>
> Is this ok for trunk?
>
> Best regards,
>
> Mihail
>
>
> ###############     Attachment also inlined for ease of reply    
> ###############
>
>
> diff --git a/gcc/config/arm/arm-protos.h b/gcc/config/arm/arm-protos.h
> index 
> 1a948d2c97526ad7e67e8d4a610ac74cfdb13882..37a46982bbc1a8f17abe2fc76ba3cb7d65257c0d 
> 100644
> --- a/gcc/config/arm/arm-protos.h
> +++ b/gcc/config/arm/arm-protos.h
> @@ -77,7 +77,7 @@ extern int thumb_legitimate_offset_p (machine_mode, 
> HOST_WIDE_INT);
>  extern int thumb1_legitimate_address_p (machine_mode, rtx, int);
>  extern bool ldm_stm_operation_p (rtx, bool, machine_mode mode,
>                                   bool, bool);
> -extern bool clear_operation_p (rtx);
> +extern bool clear_operation_p (rtx, bool);
>  extern int arm_const_double_rtx (rtx);
>  extern int vfp3_const_double_rtx (rtx);
>  extern int neon_immediate_valid_for_move (rtx, machine_mode, rtx *, 
> int *);
> diff --git a/gcc/config/arm/arm.c b/gcc/config/arm/arm.c
> index 
> f1f730cecff0fb3da7115ea1147dc8b9ab7076b7..5f3ce5c4605f609d1a0e31c0f697871266bdf835 
> 100644
> --- a/gcc/config/arm/arm.c
> +++ b/gcc/config/arm/arm.c
> @@ -13499,8 +13499,9 @@ ldm_stm_operation_p (rtx op, bool load, 
> machine_mode mode,
>    return true;
>  }
>
> -/* Checks whether OP is a valid parallel pattern for a CLRM insn.  To 
> be a
> -   valid CLRM pattern, OP must have the following form:
> +/* Checks whether OP is a valid parallel pattern for a CLRM (if VFP 
> is false)
> +   or VSCCLRM (otherwise) insn.  To be a valid CLRM pattern, OP must 
> have the
> +   following form:
>
>     [(set (reg:SI <N>) (const_int 0))
>      (set (reg:SI <M>) (const_int 0))
> @@ -13511,22 +13512,35 @@ ldm_stm_operation_p (rtx op, bool load, 
> machine_mode mode,
>
>     Any number (including 0) of set expressions is valid, the volatile 
> unspec is
>     optional.  All registers but SP and PC are allowed and registers 
> must be in
> -   strict increasing order.  */
> +   strict increasing order.
> +
> +   To be a valid VSCCLRM pattern, OP must have the following form:
> +
> +   [(unspec_volatile [(const_int 0)]
> +                    VUNSPEC_VSCCLRM_VPR)
> +    (set (reg:SF <N>) (const_int 0))
> +    (set (reg:SF <M>) (const_int 0))
> +    ...
> +   ]
> +
> +   As with CLRM, any number (including 0) of set expressions is 
> valid, however
> +   the volatile unspec is mandatory here.  Any VFP single-precision 
> register is
> +   accepted but all registers must be consecutive and in increasing 
> order.  */
>
>  bool
> -clear_operation_p (rtx op)
> +clear_operation_p (rtx op, bool vfp)
>  {
> -  HOST_WIDE_INT i;
>    unsigned regno, last_regno;
>    rtx elt, reg, zero;
> -  machine_mode mode;
>    HOST_WIDE_INT count = XVECLEN (op, 0);
> +  HOST_WIDE_INT i, first_set = vfp ? 1 : 0;
> +  machine_mode expected_mode = vfp ? E_SFmode : E_SImode;
>
> -  for (i = 0; i < count; i++)
> +  for (i = first_set; i < count; i++)
>      {
>        elt = XVECEXP (op, 0, i);
>
> -      if (GET_CODE (elt) == UNSPEC_VOLATILE)
> +      if (!vfp && GET_CODE (elt) == UNSPEC_VOLATILE)
>          {
>            if (XINT (elt, 1) != VUNSPEC_CLRM_APSR
>                || XVECLEN (elt, 0) != 1
> @@ -13542,17 +13556,26 @@ clear_operation_p (rtx op)
>
>        reg = SET_DEST (elt);
>        regno = REGNO (reg);
> -      mode = GET_MODE (reg);
>        zero = SET_SRC (elt);
>
>        if (!REG_P (reg)
> -         || GET_MODE (reg) != SImode
> -         || regno == SP_REGNUM
> -         || regno == PC_REGNUM
> -         || (i != 0 && regno <= last_regno)
> +         || GET_MODE (reg) != expected_mode
>            || zero != CONST0_RTX (SImode))
>          return false;
>
> +      if (vfp)
> +       {
> +         if (i != 1 && regno != last_regno + 1)
> +           return false;
> +       }
> +      else
> +       {
> +         if (regno == SP_REGNUM || regno == PC_REGNUM)
> +           return false;
> +         if (i != 0 && regno <= last_regno)
> +           return false;
> +       }
> +
>        last_regno = REGNO (reg);
>      }
>
> @@ -17665,6 +17688,43 @@ cmse_clear_registers (sbitmap 
> to_clear_bitmap, uint32_t *padding_bits_to_clear,
>        auto_sbitmap core_regs_bitmap (to_clear_bitmap_size);
>        auto_sbitmap to_clear_core_bitmap (to_clear_bitmap_size);
>
> +      for (i = FIRST_VFP_REGNUM; i <= maxregno; i += nb_regs)
> +       {
> +         /* Find next register to clear and exit if none. */
> +         for (; i <= maxregno && !bitmap_bit_p (to_clear_bitmap, i); 
> i++);
> +         if (i > maxregno)
> +           break;
> +
> +         /* Compute number of consecutive registers to clear.  */
> +         for (j = i; j <= maxregno && bitmap_bit_p (to_clear_bitmap, j);
> +              j++);
> +         nb_regs = j - i;
> +
> +         /* Create VSCCLRM RTX pattern.  */
> +         par = gen_rtx_PARALLEL (VOIDmode, rtvec_alloc (nb_regs + 1));
> +         vunspec_vec = gen_rtvec (1, gen_int_mode (0, SImode));
> +         vunspec = gen_rtx_UNSPEC_VOLATILE (SImode, vunspec_vec,
> + VUNSPEC_VSCCLRM_VPR);
> +         XVECEXP (par, 0, 0) = vunspec;
> +
> +         /* Insert VFP register clearing RTX in the pattern.  */
> +         start_sequence ();
> +         for (k = 1, j = i; j <= maxregno && k < nb_regs + 1; j++)
> +           {
> +             if (!bitmap_bit_p (to_clear_bitmap, j))
> +               continue;
> +
> +             reg = gen_rtx_REG (SFmode, j);
> +             set = gen_rtx_SET (reg, const0_rtx);
> +             XVECEXP (par, 0, k++) = set;
> +             emit_use (reg);
> +           }
> +         use_seq = get_insns ();
> +         end_sequence ();
> +
> +         emit_insn_after (use_seq, emit_insn (par));
> +       }
> +
>        /* Get set of core registers to clear.  */
>        bitmap_clear (core_regs_bitmap);
>        bitmap_set_range (core_regs_bitmap, R0_REGNUM,
> @@ -17703,49 +17763,50 @@ cmse_clear_registers (sbitmap 
> to_clear_bitmap, uint32_t *padding_bits_to_clear,
>        end_sequence ();
>
>        emit_insn_after (use_seq, emit_insn (par));
> -      minregno = FIRST_VFP_REGNUM;
> -    }
> -
> -  /* If not marked for clearing, clearing_reg already does not contain
> -     any secret.  */
> -  if (clearing_regno <= maxregno
> -      && bitmap_bit_p (to_clear_bitmap, clearing_regno))
> -    {
> -      emit_move_insn (clearing_reg, const0_rtx);
> -      emit_use (clearing_reg);
> -      bitmap_clear_bit (to_clear_bitmap, clearing_regno);
>      }
> -
> -  for (regno = minregno; regno <= maxregno; regno++)
> +  else
>      {
> -      if (!bitmap_bit_p (to_clear_bitmap, regno))
> -       continue;
> +      /* If not marked for clearing, clearing_reg already does not 
> contain
> +        any secret.  */
> +      if (clearing_regno <= maxregno
> +         && bitmap_bit_p (to_clear_bitmap, clearing_regno))
> +       {
> +         emit_move_insn (clearing_reg, const0_rtx);
> +         emit_use (clearing_reg);
> +         bitmap_clear_bit (to_clear_bitmap, clearing_regno);
> +       }
>
> -      if (IS_VFP_REGNUM (regno))
> +      for (regno = minregno; regno <= maxregno; regno++)
>          {
> -         /* If regno is an even vfp register and its successor is also to
> -            be cleared, use vmov.  */
> -         if (TARGET_VFP_DOUBLE
> -             && VFP_REGNO_OK_FOR_DOUBLE (regno)
> -             && bitmap_bit_p (to_clear_bitmap, regno + 1))
> +         if (!bitmap_bit_p (to_clear_bitmap, regno))
> +           continue;
> +
> +         if (IS_VFP_REGNUM (regno))
>              {
> -             emit_move_insn (gen_rtx_REG (DFmode, regno),
> -                             CONST1_RTX (DFmode));
> -             emit_use (gen_rtx_REG (DFmode, regno));
> -             regno++;
> +             /* If regno is an even vfp register and its successor is 
> also to
> +                be cleared, use vmov.  */
> +             if (TARGET_VFP_DOUBLE
> +                 && VFP_REGNO_OK_FOR_DOUBLE (regno)
> +                 && bitmap_bit_p (to_clear_bitmap, regno + 1))
> +               {
> +                 emit_move_insn (gen_rtx_REG (DFmode, regno),
> +                                 CONST1_RTX (DFmode));
> +                 emit_use (gen_rtx_REG (DFmode, regno));
> +                 regno++;
> +               }
> +             else
> +               {
> +                 emit_move_insn (gen_rtx_REG (SFmode, regno),
> +                                 CONST1_RTX (SFmode));
> +                 emit_use (gen_rtx_REG (SFmode, regno));
> +               }
>              }
>            else
>              {
> -             emit_move_insn (gen_rtx_REG (SFmode, regno),
> -                             CONST1_RTX (SFmode));
> -             emit_use (gen_rtx_REG (SFmode, regno));
> +             emit_move_insn (gen_rtx_REG (SImode, regno), clearing_reg);
> +             emit_use (gen_rtx_REG (SImode, regno));
>              }
>          }
> -      else
> -       {
> -         emit_move_insn (gen_rtx_REG (SImode, regno), clearing_reg);
> -         emit_use (gen_rtx_REG (SImode, regno));
> -       }
>      }
>  }
>
> @@ -25948,7 +26009,8 @@ thumb1_expand_prologue (void)
>  void
>  cmse_nonsecure_entry_clear_before_return (void)
>  {
> -  int regno, maxregno = TARGET_HARD_FLOAT ? LAST_VFP_REGNUM : IP_REGNUM;
> +  bool clear_vfpregs = TARGET_HARD_FLOAT || TARGET_HAVE_FPCTX_CMSE;
> +  int regno, maxregno = clear_vfpregs ? LAST_VFP_REGNUM : IP_REGNUM;
>    uint32_t padding_bits_to_clear = 0;
>    auto_sbitmap to_clear_bitmap (maxregno + 1);
>    rtx r1_reg, result_rtl, clearing_reg = NULL_RTX;
> @@ -25960,7 +26022,7 @@ cmse_nonsecure_entry_clear_before_return (void)
>
>    /* If we are not dealing with -mfloat-abi=soft we will need to 
> clear VFP
>       registers.  */
> -  if (TARGET_HARD_FLOAT)
> +  if (clear_vfpregs)
>      {
>        int float_bits = D7_VFP_REGNUM - FIRST_VFP_REGNUM + 1;
>
> @@ -25989,7 +26051,9 @@ cmse_nonsecure_entry_clear_before_return (void)
>          continue;
>        if (IN_RANGE (regno, IP_REGNUM, PC_REGNUM))
>          continue;
> -      if (call_used_or_fixed_reg_p (regno))
> +      if (call_used_or_fixed_reg_p (regno)
> +         && (!IN_RANGE (regno, FIRST_VFP_REGNUM, LAST_VFP_REGNUM)
> +             || TARGET_HARD_FLOAT))
>          bitmap_set_bit (to_clear_bitmap, regno);
>      }
>
> diff --git a/gcc/config/arm/predicates.md b/gcc/config/arm/predicates.md
> index 
> e5c583ef3d167194e7a061d7c3e98d3b4bb5269c..69c10c06ff405e19efa172217a08a512c66cb902 
> 100644
> --- a/gcc/config/arm/predicates.md
> +++ b/gcc/config/arm/predicates.md
> @@ -513,7 +513,13 @@
>  (define_special_predicate "clear_multiple_operation"
>    (match_code "parallel")
>  {
> - return clear_operation_p (op);
> + return clear_operation_p (op, /*vfp*/false);
> +})
> +
> +(define_special_predicate "clear_vfp_multiple_operation"
> +  (match_code "parallel")
> +{
> + return clear_operation_p (op, /*vfp*/true);
>  })
>
>  (define_special_predicate "load_multiple_operation"
> diff --git a/gcc/config/arm/unspecs.md b/gcc/config/arm/unspecs.md
> index 
> 498bc0798dbaaa3ee73815ba27864ae92a2fd08e..ad072705525cc3c4035a22b5f99546bcd0e79ec5 
> 100644
> --- a/gcc/config/arm/unspecs.md
> +++ b/gcc/config/arm/unspecs.md
> @@ -175,6 +175,8 @@
>    VUNSPEC_SPECULATION_BARRIER ; Represents an unconditional 
> speculation barrier.
>    VUNSPEC_VSTR_VLDR    ; Represent the vstr/vldr instruction.
>    VUNSPEC_CLRM_APSR    ; Represent the clearing of APSR with clrm 
> instruction.
> +  VUNSPEC_VSCCLRM_VPR  ; Represent the clearing of VPR with vscclrm
> +                       ; instruction.
>  ])
>
>  ;; Enumerators for NEON unspecs.
> diff --git a/gcc/config/arm/vfp.md b/gcc/config/arm/vfp.md
> index 
> 027f0524430fd3af9e4d10cd98bd177141407841..2bf928f32626ce2731ebdfedffbbf64d806eff53 
> 100644
> --- a/gcc/config/arm/vfp.md
> +++ b/gcc/config/arm/vfp.md
> @@ -1641,6 +1641,42 @@
>     (set_attr "type" "load_4")]
>  )
>
> +;; The operands are validated through the clear_multiple_operation
> +;; match_parallel predicate rather than through constraints so enable 
> it only
> +;; after reload.
> +(define_insn "*clear_vfp_multiple"
> +  [(match_parallel 0 "clear_vfp_multiple_operation"
> +     [(unspec_volatile [(const_int 0)]
> +                      VUNSPEC_VSCCLRM_VPR)])]
> +  "TARGET_HAVE_FPCTX_CMSE && use_cmse && reload_completed"
> +  {
> +    int num_regs = XVECLEN (operands[0], 0);
> +    char pattern[30];
> +    const char *regname;
> +    rtx reg;
> +
> +    strcpy (pattern, \"vscclrm%?\\t{%|\");
> +    if (num_regs > 1)
> +      {
> +       reg = XEXP (XVECEXP (operands[0], 0, 1), 0);
> +       strcat (pattern, reg_names[REGNO (reg)]);
> +       if (num_regs > 2)
> +         {
> +           strcat (pattern, \"-%|\");
> +           reg = XEXP (XVECEXP (operands[0], 0, num_regs - 1), 0);
> +           strcat (pattern, reg_names[REGNO (reg)]);
> +         }
> +       strcat (pattern, \", \");
> +      }
> +
> +    strcat (pattern, \"VPR}\");
> +    output_asm_insn (pattern, operands);
> +    return \"\";
> +  }
> +  [(set_attr "predicable" "yes")
> +   (set_attr "type" "mov_reg")]
> +)
> +
>  (define_insn_and_split "*cmpsf_split_vfp"
>    [(set (reg:CCFP CC_REGNUM)
>          (compare:CCFP (match_operand:SF 0 "s_register_operand"  "t")
> diff --git a/gcc/testsuite/gcc.target/arm/cmse/bitfield-1.c 
> b/gcc/testsuite/gcc.target/arm/cmse/bitfield-1.c
> index 
> 7036cb9508c27d56c4b2c01a81c44bf6f1f9c781..cae2f06140766176c9af6e4b2b4a3e722fbad67c 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/bitfield-1.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/bitfield-1.c
> @@ -36,6 +36,7 @@ main (void)
>  /* { dg-final { scan-assembler "movw\tr1, #1855" } } */
>  /* { dg-final { scan-assembler "movt\tr1, 65535" } } */
>  /* { dg-final { scan-assembler "ands\tr0(, r0)?, r1" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s0-s15, VPR\}" { target 
> arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "clrm\t\{r1, r2, r3, ip, APSR\}" { 
> target arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "vldr\tFPCXTNS, \\\[sp\\\], #4" { 
> target arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "bxns" } } */
> diff --git a/gcc/testsuite/gcc.target/arm/cmse/bitfield-2.c 
> b/gcc/testsuite/gcc.target/arm/cmse/bitfield-2.c
> index 
> 50d4979470dd21738453e0d70c7a69ee0752ac41..21a53cd85792e439314f5c88080d7f1c6fcdec66 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/bitfield-2.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/bitfield-2.c
> @@ -33,6 +33,7 @@ main (void)
>  /* { dg-final { scan-assembler "movw\tr1, #1919" } } */
>  /* { dg-final { scan-assembler "movt\tr1, 2047" } } */
>  /* { dg-final { scan-assembler "ands\tr0(, r0)?, r1" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s0-s15, VPR\}" { target 
> arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "clrm\t\{r1, r2, r3, ip, APSR\}" { 
> target arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "vldr\tFPCXTNS, \\\[sp\\\], #4" { 
> target arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "bxns" } } */
> diff --git a/gcc/testsuite/gcc.target/arm/cmse/bitfield-3.c 
> b/gcc/testsuite/gcc.target/arm/cmse/bitfield-3.c
> index 
> 2b7507c900ab18705083ba1d86359e1ae36a50a2..e8f973e3f06881d33d661f81c0eb6671e48a0d01 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/bitfield-3.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/bitfield-3.c
> @@ -34,6 +34,7 @@ main (void)
>  /* { dg-final { scan-assembler "movw\tr1, #65535" } } */
>  /* { dg-final { scan-assembler "movt\tr1, 63" } } */
>  /* { dg-final { scan-assembler "ands\tr0(, r0)?, r1" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s0-s15, VPR\}" { target 
> arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "clrm\t\{r1, r2, r3, ip, APSR\}" { 
> target arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "vldr\tFPCXTNS, \\\[sp\\\], #4" { 
> target arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "bxns" } } */
> diff --git a/gcc/testsuite/gcc.target/arm/cmse/cmse-1.c 
> b/gcc/testsuite/gcc.target/arm/cmse/cmse-1.c
> index 
> 35cab1f3233daac9fba50d25dac23364c798fb9c..29d78ddd6166321bdf0acabf011138b276151e85 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/cmse-1.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/cmse-1.c
> @@ -105,6 +105,7 @@ qux (int_nsfunc_t * callback)
>  /* { dg-final { scan-assembler "bic" } } */
>  /* { dg-final { scan-assembler "push\t\{r4, r5, r6" } } */
>  /* { dg-final { scan-assembler "vstr\tFPCXTNS, \\\[sp, #-4\\\]!" { 
> target arm_cmse_clear_ok } } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s0-s15, VPR\}" { target 
> arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "clrm\t\{r1, r2, r3, ip, APSR\}" { 
> target arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "vldr\tFPCXTNS, \\\[sp\\\], #4" { 
> target arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "msr\tAPSR_nzcvq" { target { ! 
> arm_cmse_clear_ok } } } } */
> diff --git 
> a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-13.c 
> b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-13.c
> index 
> 7b2df1faff5ea6ce6e7a58a37cd23eaf9616ee97..9719f799229e245264d377c3726190ca768ca1dd 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-13.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-13.c
> @@ -10,21 +10,10 @@
>  /* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
>  /* { dg-final { scan-assembler "clrm\t\{r0, r1, r2, r3, APSR\}" } } */
>  /* { dg-final { scan-assembler-not "vmov\.f32\ts0, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts1, #1\.0" } } */
>  /* { dg-final { scan-assembler-not "vmov\.f32\ts2, #1\.0" } } */
>  /* { dg-final { scan-assembler-not "vmov\.f32\ts3, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts4, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts5, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts6, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts7, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts8, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts9, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts10, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts11, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts12, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts13, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts14, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts15, #1\.0" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s1, VPR\}" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s4-s15, VPR\}" } } */
>
>  /* Now we check that we use the correct intrinsic to call. */
>  /* { dg-final { scan-assembler "bl\t__gnu_cmse_nonsecure_call" } } */
> diff --git 
> a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-5.c 
> b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-5.c
> index 
> 638643d0a6772d43f440df3942c6c723f1eff3ef..dc62b742ba5ab9f7480ca93396e5eb354c5e64dd 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-5.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-5.c
> @@ -6,22 +6,7 @@
>  #include "../../../cmse-5.x"
>
>  /* { dg-final { scan-assembler "vstr\tFPCXTNS, \\\[sp, #-4\\\]!" } } */
> -/* { dg-final { scan-assembler-not "vmov\.f32\ts0, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts1, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts2, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts3, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts4, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts5, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts6, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts7, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts8, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts9, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts10, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts11, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts12, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts13, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts14, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts15, #1\.0" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s1-s15, VPR\}" } } */
>  /* { dg-final { scan-assembler "clrm\t\{r0, r1, r2, r3, ip, APSR\}" } 
> } */
>  /* { dg-final { scan-assembler "vldr\tFPCXTNS, \\\[sp\\\], #4" } } */
>  /* { dg-final { scan-assembler "bxns" } } */
> diff --git 
> a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-7.c 
> b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-7.c
> index 
> 9ec14d1a707839e4858bac72bc292fd0984bf86c..a6951d34afeed54c0ac2e937345cbfbbfb2c4b2d 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-7.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-7.c
> @@ -9,22 +9,7 @@
>  /* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
>  /* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
>  /* { dg-final { scan-assembler "clrm\t\{r0, r1, r2, r3, APSR\}" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts0, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts1, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts2, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts3, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts4, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts5, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts6, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts7, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts8, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts9, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts10, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts11, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts12, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts13, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts14, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts15, #1\.0" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s0-s15, VPR\}" } } */
>
>  /* Now we check that we use the correct intrinsic to call. */
>  /* { dg-final { scan-assembler "bl\t__gnu_cmse_nonsecure_call" } } */
> diff --git 
> a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-8.c 
> b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-8.c
> index 
> d375879ed02483c05d7853c95f9b9e15383ff321..23db88dc66d584acbba64a10d3f6fd06f066b321 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-8.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard-sp/cmse-8.c
> @@ -11,20 +11,7 @@
>  /* { dg-final { scan-assembler "clrm\t\{r0, r1, r2, r3, APSR\}" } } */
>  /* { dg-final { scan-assembler-not "vmov\.f32\ts0, #1\.0" } } */
>  /* { dg-final { scan-assembler-not "vmov\.f32\ts1, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts2, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts3, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts4, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts5, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts6, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts7, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts8, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts9, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts10, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts11, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts12, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts13, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts14, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts15, #1\.0" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s2-s15, VPR\}" } } */
>
>  /* Now we check that we use the correct intrinsic to call. */
>  /* { dg-final { scan-assembler "bl\t__gnu_cmse_nonsecure_call" } } */
> diff --git 
> a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-13.c 
> b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-13.c
> index 
> e5cb0fabd4ce20ef8e57f25778b81a3ed3908af0..2898efabb2b37e3c8dacaf029b5a40be8edcc624 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-13.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-13.c
> @@ -9,18 +9,13 @@
>  /* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
>  /* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
>  /* { dg-final { scan-assembler "clrm\t\{r0, r1, r2, r3, APSR\}" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts1, #1\.0" } } */
>  /* { dg-final { scan-assembler-not "vmov\.f32\ts0, #1\.0" } } */
>  /* { dg-final { scan-assembler-not "vmov\.f64\td0, #1\.0" } } */
>  /* { dg-final { scan-assembler-not "vmov\.f64\td1, #1\.0" } } */
>  /* { dg-final { scan-assembler-not "vmov\.f32\ts2, #1\.0" } } */
>  /* { dg-final { scan-assembler-not "vmov\.f32\ts3, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td2, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td3, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td4, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td5, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td6, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td7, #1\.0" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s1, VPR\}" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s4-s15, VPR\}" } } */
>
>  /* Now we check that we use the correct intrinsic to call. */
>  /* { dg-final { scan-assembler "bl\t__gnu_cmse_nonsecure_call" } } */
> diff --git 
> a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-5.c 
> b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-5.c
> index 
> 7e3b7609dfa5c8c1df77e413fa1bb85aa80ea9f6..55e7a4f7ae0c4420a40a250f246a769eff28e6c7 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-5.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-5.c
> @@ -6,15 +6,7 @@
>  #include "../../../cmse-5.x"
>
>  /* { dg-final { scan-assembler "vstr\tFPCXTNS, \\\[sp, #-4\\\]!" } } */
> -/* { dg-final { scan-assembler-not "vmov\.f32\ts0, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts1, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td1, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td2, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td3, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td4, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td5, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td6, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td7, #1\.0" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s1-s15, VPR\}" } } */
>  /* { dg-final { scan-assembler "clrm\t\{r0, r1, r2, r3, ip, APSR\}" } 
> } */
>  /* { dg-final { scan-assembler "vldr\tFPCXTNS, \\\[sp\\\], #4" } } */
>  /* { dg-final { scan-assembler "bxns" } } */
> diff --git 
> a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-7.c 
> b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-7.c
> index 
> d998b43a148250a15eb4aae9fb0ef6e4bf51203b..a4520a9166e36b1d3220befb72a1863799b2f424 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-7.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-7.c
> @@ -9,14 +9,7 @@
>  /* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
>  /* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
>  /* { dg-final { scan-assembler "clrm\t\{r0, r1, r2, r3, APSR\}" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td0, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td1, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td2, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td3, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td4, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td5, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td6, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td7, #1\.0" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s0-s15, VPR\}" } } */
>
>  /* Now we check that we use the correct intrinsic to call. */
>  /* { dg-final { scan-assembler "bl\t__gnu_cmse_nonsecure_call" } } */
> diff --git 
> a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-8.c 
> b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-8.c
> index 
> e416bef2cb9fabd9cb33e3c1c87057ebdd3d6daf..c79d3188026d9248abde7967671c05a09b75c49f 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-8.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/hard/cmse-8.c
> @@ -10,13 +10,7 @@
>  /* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
>  /* { dg-final { scan-assembler "clrm\t\{r0, r1, r2, r3, APSR\}" } } */
>  /* { dg-final { scan-assembler-not "vmov\.f64\td0, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td1, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td2, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td3, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td4, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td5, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td6, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td7, #1\.0" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s2-s15, VPR\}" } } */
>
>  /* Now we check that we use the correct intrinsic to call. */
>  /* { dg-final { scan-assembler "bl\t__gnu_cmse_nonsecure_call" } } */
> diff --git 
> a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/soft/cmse-5.c 
> b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/soft/cmse-5.c
> index 
> 157bccb9ff3256056d496a0d0770374315b04d87..d3a3a742b8b4ff99c973a8c9edd77297c7f37ae2 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/soft/cmse-5.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/soft/cmse-5.c
> @@ -7,6 +7,7 @@
>  /* { dg-final { scan-assembler "vstr\tFPCXTNS, \\\[sp, #-4\\\]!" } } */
>  /* { dg-final { scan-assembler-not "vmov" } } */
>  /* { dg-final { scan-assembler-not "vmsr" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s0-s15, VPR\}" } } */
>  /* { dg-final { scan-assembler "clrm\t\{r1, r2, r3, ip, APSR\}" } } */
>  /* { dg-final { scan-assembler "vldr\tFPCXTNS, \\\[sp\\\], #4" } } */
>  /* { dg-final { scan-assembler "bxns" } } */
> diff --git 
> a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/softfp-sp/cmse-5.c 
> b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/softfp-sp/cmse-5.c
> index 
> 2522a17a6316d76a21d8d241d4c44cdddf1981e0..bbe16cee9c2c29e0701f54c3e393ad3c7efb8e87 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/softfp-sp/cmse-5.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/softfp-sp/cmse-5.c
> @@ -8,22 +8,7 @@
>  /* { dg-final { scan-assembler "__acle_se_foo:" } } */
>  /* { dg-final { scan-assembler "vstr\tFPCXTNS, \\\[sp, #-4\\\]!" } } */
>  /* { dg-final { scan-assembler-not "mov\tr0, lr" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts0, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts1, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts2, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts3, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts4, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts5, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts6, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts7, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts8, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts9, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts10, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts11, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts12, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts13, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts14, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f32\ts15, #1\.0" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s0-s15, VPR\}" } } */
>  /* { dg-final { scan-assembler "clrm\t\{r1, r2, r3, ip, APSR\}" } } */
>  /* { dg-final { scan-assembler "vldr\tFPCXTNS, \\\[sp\\\], #4" } } */
>  /* { dg-final { scan-assembler "bxns" } } */
> diff --git 
> a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/softfp/cmse-5.c 
> b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/softfp/cmse-5.c
> index 
> eb7561f2a2629f3c153afe3a1a5c58d8abb9d07b..c7dda6a3d6088366358ff4bc1e6d83ad9b03ea6e 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/softfp/cmse-5.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/8_1m/softfp/cmse-5.c
> @@ -7,14 +7,7 @@
>
>  /* { dg-final { scan-assembler "__acle_se_foo:" } } */
>  /* { dg-final { scan-assembler "vstr\tFPCXTNS, \\\[sp, #-4\\\]!" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td0, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td1, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td2, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td3, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td4, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td5, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td6, #1\.0" } } */
> -/* { dg-final { scan-assembler "vmov\.f64\td7, #1\.0" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s0-s15, VPR\}" } } */
>  /* { dg-final { scan-assembler "clrm\t\{r1, r2, r3, ip, APSR\}" } } */
>  /* { dg-final { scan-assembler "vldr\tFPCXTNS, \\\[sp\\\], #4" } } */
>  /* { dg-final { scan-assembler "bxns" } } */
> diff --git a/gcc/testsuite/gcc.target/arm/cmse/struct-1.c 
> b/gcc/testsuite/gcc.target/arm/cmse/struct-1.c
> index 
> 90fdac18e30424edca60b6f884227adbf716899d..755564369ae7deecfb038e67591e679bb06f62b1 
> 100644
> --- a/gcc/testsuite/gcc.target/arm/cmse/struct-1.c
> +++ b/gcc/testsuite/gcc.target/arm/cmse/struct-1.c
> @@ -29,6 +29,7 @@ main (void)
>  /* { dg-final { scan-assembler "movs\tr1, #255" } } */
>  /* { dg-final { scan-assembler "movt\tr1, 65535" } } */
>  /* { dg-final { scan-assembler "ands\tr0(, r0)?, r1" } } */
> +/* { dg-final { scan-assembler "vscclrm\t\{s0-s15, VPR\}" { target 
> arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "clrm\t\{r1, r2, r3, ip, APSR\}" { 
> target arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "vldr\tFPCXTNS, \\\[sp\\\], #4" { 
> target arm_cmse_clear_ok } } } */
>  /* { dg-final { scan-assembler "bxns" } } */
>

More information about the Gcc-patches mailing list