[RFC,PATCHv3,ARM,7/8] ARMv8-M Security Extension's cmse_nonsecure_call: use __gnu_cmse_nonsecure_call]

diff --git a/gcc/config/arm/arm-protos.h b/gcc/config/arm/arm-protos.h
index cc8218ac2a286ae936a0b239b5faf5e6aa3bbe87..2f1fd16fd96e060b9d122dac286d0ad421bcad15 100644
--- a/gcc/config/arm/arm-protos.h
+++ b/gcc/config/arm/arm-protos.h
@@ -131,6 +131,7 @@  extern int arm_const_double_inline_cost (rtx);
 extern bool arm_const_double_by_parts (rtx);
 extern bool arm_const_double_by_immediates (rtx);
 extern void arm_emit_call_insn (rtx, rtx, bool);
+bool detect_cmse_nonsecure_call (tree);
 extern const char *output_call (rtx *);
 void arm_emit_movpair (rtx, rtx);
 extern const char *output_mov_long_double_arm_from_arm (rtx *);
diff --git a/gcc/config/arm/arm.c b/gcc/config/arm/arm.c
index 973b97b6e597f613b7ed78c5e4b2036e987cd4bc..e5945941ab3b3f5868f135727c2f5a12f7017145 100644
--- a/gcc/config/arm/arm.c
+++ b/gcc/config/arm/arm.c
@@ -17450,6 +17450,129 @@  note_invalid_constants (rtx_insn *insn, HOST_WIDE_INT address, int do_pushes)
   return;
 }
 
+/* Saves callee saved registers, clears callee saved registers and caller saved
+   registers not used to pass arguments before a cmse_nonsecure_call.  And
+   restores the callee saved registers after.  */
+
+static void
+cmse_nonsecure_call_clear_caller_saved (void)
+{
+  basic_block bb;
+
+  FOR_EACH_BB_FN (bb, cfun)
+    {
+      rtx_insn *insn;
+
+      FOR_BB_INSNS (bb, insn)
+	{
+	  uint64_t to_clear_mask, float_mask;
+	  rtx_insn *seq;
+	  rtx pat, call, unspec, link, reg, cleared_reg, tmp;
+	  unsigned int regno, maxregno;
+	  rtx address;
+
+	  if (!NONDEBUG_INSN_P (insn))
+	    continue;
+
+	  if (!CALL_P (insn))
+	    continue;
+
+	  pat = PATTERN (insn);
+	  gcc_assert (GET_CODE (pat) == PARALLEL && XVECLEN (pat, 0) > 0);
+	  call = XVECEXP (pat, 0, 0);
+
+	  /* Get the real call RTX if the insn sets a value, ie. returns.  */
+	  if (GET_CODE (call) == SET)
+	      call = SET_SRC (call);
+
+	  /* Check if it is a cmse_nonsecure_call.  */
+	  unspec = XEXP (call, 0);
+	  if (GET_CODE (unspec) != UNSPEC
+	      || XINT (unspec, 1) != UNSPEC_NONSECURE_MEM)
+	    continue;
+
+	  /* Determine the caller-saved registers we need to clear.  */
+	  to_clear_mask = (1LL << (NUM_ARG_REGS)) - 1;
+	  maxregno = NUM_ARG_REGS - 1;
+	  if (TARGET_HARD_FLOAT && TARGET_VFP)
+	    {
+	      float_mask = (1LL << (D7_VFP_REGNUM + 1)) - 1;
+	      float_mask &= ~((1LL << FIRST_VFP_REGNUM) - 1);
+	      to_clear_mask |= float_mask;
+	      maxregno = D7_VFP_REGNUM;
+	    }
+
+	  /* Make sure the register used to hold the function address is not
+	     cleared.  */
+	  address = RTVEC_ELT (XVEC (unspec, 0), 0);
+	  gcc_assert (MEM_P (address));
+	  gcc_assert (REG_P (XEXP (address, 0)));
+	  to_clear_mask &= ~(1LL << REGNO (XEXP (address, 0)));
+
+	  /* Set basic block of call insn so that df rescan is performed on
+	     insns inserted here.  */
+	  set_block_for_insn (insn, bb);
+	  df_set_flags (DF_DEFER_INSN_RESCAN);
+	  start_sequence ();
+
+	  /* Make sure the scheduler doesn't schedule other insns beyond here.  */
+	  emit_insn (gen_blockage ());
+
+	  /* Get the list of registers used by the function call.  */
+	  link = CALL_INSN_FUNCTION_USAGE (insn);
+	  for (; link != NULL; link = XEXP (link, 1))
+	    {
+	      pat = XEXP (link, 0);
+	      if (GET_CODE (pat) != USE)
+		continue;
+
+	      reg = XEXP (pat, 0);
+	      to_clear_mask &= ~(1LL << REGNO (reg));
+	      if (ARM_NUM_REGS (GET_MODE (reg)) > 1)
+		to_clear_mask &= ~(1LL << (REGNO (reg) + 1));
+	    }
+
+	  /* We use right shift and left shift to clear the LSB of the address
+	     we jump to instead of using bic, to avoid having to use an extra
+	     register on thumb1.  */
+	  cleared_reg = XEXP (address, 0);
+	  tmp = gen_rtx_LSHIFTRT (SImode, cleared_reg, const1_rtx);
+	  emit_insn (gen_rtx_SET (cleared_reg, tmp));
+	  tmp = gen_rtx_ASHIFT (SImode, cleared_reg, const1_rtx);
+	  emit_insn (gen_rtx_SET (cleared_reg, tmp));
+
+	  /* Clearing all registers that leak before doing a non-secure
+	     call.  */
+	  for (regno = 0; regno <= maxregno; regno++)
+	    {
+	      if (!(to_clear_mask & (1LL << regno)))
+		continue;
+
+	      /* If regno is a vfp register, even and its successor is also to
+		 be cleared, use vmov.  */
+	      if (IS_VFP_REGNUM (regno))
+		{
+		  if (TARGET_VFP_DOUBLE
+		      && VFP_REGNO_OK_FOR_DOUBLE (regno)
+		      && to_clear_mask & (1LL << (regno + 1)))
+		    emit_move_insn (gen_rtx_REG (DFmode, regno++),
+				    CONST1_RTX (DFmode));
+		  else
+		    emit_move_insn (gen_rtx_REG (SFmode, regno),
+				    CONST1_RTX (SFmode));
+		}
+	      else
+		emit_move_insn (gen_rtx_REG (SImode, regno), cleared_reg);
+	    }
+
+	  seq = get_insns ();
+	  end_sequence ();
+	  emit_insn_before (seq, insn);
+
+	}
+    }
+}
+
 /* Rewrite move insn into subtract of 0 if the condition codes will
    be useful in next conditional jump insn.  */
 
@@ -17750,6 +17873,8 @@  arm_reorg (void)
   HOST_WIDE_INT address = 0;
   Mfix * fix;
 
+  if (use_cmse)
+    cmse_nonsecure_call_clear_caller_saved ();
   if (TARGET_THUMB1)
     thumb1_reorg ();
   else if (TARGET_THUMB2)
@@ -18125,6 +18250,23 @@  vfp_emit_fstmd (int base_reg, int count)
   return count * 8;
 }
 
+/* Returns true if -mcmse has been passed and the function pointed to by 'addr'
+   has the cmse_nonsecure_call attribute and returns false otherwise.  */
+
+bool
+detect_cmse_nonsecure_call (tree addr)
+{
+  if (!addr)
+    return FALSE;
+
+  tree fntype = TREE_TYPE (addr);
+  if (use_cmse && lookup_attribute ("cmse_nonsecure_call",
+				    TYPE_ATTRIBUTES (fntype)))
+    return TRUE;
+  return FALSE;
+}
+
+
 /* Emit a call instruction with pattern PAT.  ADDR is the address of
    the call target.  */
 
diff --git a/gcc/config/arm/arm.md b/gcc/config/arm/arm.md
index 665aecf2fe746ca17e28396ca77f27890c7d4891..3942e8840897a50ffa11c91f61cdebba36a7af18 100644
--- a/gcc/config/arm/arm.md
+++ b/gcc/config/arm/arm.md
@@ -7628,6 +7628,7 @@ 
   "
   {
     rtx callee, pat;
+    tree addr = MEM_EXPR (operands[0]);
     
     /* In an untyped call, we can get NULL for operand 2.  */
     if (operands[2] == NULL_RTX)
@@ -7642,8 +7643,17 @@ 
 	: !REG_P (callee))
       XEXP (operands[0], 0) = force_reg (Pmode, callee);
 
-    pat = gen_call_internal (operands[0], operands[1], operands[2]);
-    arm_emit_call_insn (pat, XEXP (operands[0], 0), false);
+    if (detect_cmse_nonsecure_call (addr))
+      {
+	pat = gen_nonsecure_call_internal (operands[0], operands[1],
+					   operands[2]);
+	emit_call_insn (pat);
+      }
+    else
+      {
+	pat = gen_call_internal (operands[0], operands[1], operands[2]);
+	arm_emit_call_insn (pat, XEXP (operands[0], 0), false);
+      }
     DONE;
   }"
 )
@@ -7654,6 +7664,24 @@ 
 	      (use (match_operand 2 "" ""))
 	      (clobber (reg:SI LR_REGNUM))])])
 
+(define_expand "nonsecure_call_internal"
+  [(parallel [(call (unspec:SI [(match_operand 0 "memory_operand" "")]
+			       UNSPEC_NONSECURE_MEM)
+		    (match_operand 1 "general_operand" ""))
+	      (use (match_operand 2 "" ""))
+	      (clobber (reg:SI LR_REGNUM))
+	      (clobber (reg:SI 4))])]
+  ""
+  "
+  {
+    rtx tmp;
+    tmp = copy_to_suggested_reg (XEXP (operands[0], 0),
+				 gen_rtx_REG (SImode, 4),
+				 SImode);
+
+    operands[0] = replace_equiv_address (operands[0], tmp);
+  }")
+
 (define_insn "*call_reg_armv5"
   [(call (mem:SI (match_operand:SI 0 "s_register_operand" "r"))
          (match_operand 1 "" ""))
@@ -7689,6 +7717,7 @@ 
   "
   {
     rtx pat, callee;
+    tree addr = MEM_EXPR (operands[1]);
     
     /* In an untyped call, we can get NULL for operand 2.  */
     if (operands[3] == 0)
@@ -7703,9 +7732,18 @@ 
 	: !REG_P (callee))
       XEXP (operands[1], 0) = force_reg (Pmode, callee);
 
-    pat = gen_call_value_internal (operands[0], operands[1],
-				   operands[2], operands[3]);
-    arm_emit_call_insn (pat, XEXP (operands[1], 0), false);
+    if (detect_cmse_nonsecure_call (addr))
+      {
+	pat = gen_nonsecure_call_value_internal (operands[0], operands[1],
+						 operands[2], operands[3]);
+	emit_call_insn (pat);
+      }
+    else
+      {
+	pat = gen_call_value_internal (operands[0], operands[1],
+				       operands[2], operands[3]);
+	arm_emit_call_insn (pat, XEXP (operands[1], 0), false);
+      }
     DONE;
   }"
 )
@@ -7717,6 +7755,25 @@ 
 	      (use (match_operand 3 "" ""))
 	      (clobber (reg:SI LR_REGNUM))])])
 
+(define_expand "nonsecure_call_value_internal"
+  [(parallel [(set (match_operand       0 "" "")
+		   (call (unspec:SI [(match_operand 1 "memory_operand" "")]
+				    UNSPEC_NONSECURE_MEM)
+			 (match_operand 2 "general_operand" "")))
+	      (use (match_operand 3 "" ""))
+	      (clobber (reg:SI LR_REGNUM))
+	      (clobber (reg:SI 4))])]
+  ""
+  "
+  {
+    rtx tmp;
+    tmp = copy_to_suggested_reg (XEXP (operands[1], 0),
+				 gen_rtx_REG (SImode, 4),
+				 SImode);
+
+    operands[1] = replace_equiv_address (operands[1], tmp);
+  }")
+
 (define_insn "*call_value_reg_armv5"
   [(set (match_operand 0 "" "")
         (call (mem:SI (match_operand:SI 1 "s_register_operand" "r"))
diff --git a/gcc/config/arm/thumb1.md b/gcc/config/arm/thumb1.md
index e895fc77d5dc68cc723f1daf7e80b8b83ce3b20b..b5a6dd152d2480e4ae1eb22cc244840905d4c132 100644
--- a/gcc/config/arm/thumb1.md
+++ b/gcc/config/arm/thumb1.md
@@ -1705,6 +1705,19 @@ 
    (set_attr "type" "call")]
 )
 
+(define_insn "*nonsecure_call_reg_thumb1_v5"
+  [(call (unspec:SI [(mem:SI (match_operand:SI 0 "register_operand" "l*r"))]
+		    UNSPEC_NONSECURE_MEM)
+	 (match_operand 1 "" ""))
+   (use (match_operand 2 "" ""))
+   (clobber (reg:SI LR_REGNUM))
+   (clobber (match_dup 0))]
+  "TARGET_THUMB1 && use_cmse && !SIBLING_CALL_P (insn)"
+  "bl\\t__gnu_cmse_nonsecure_call"
+  [(set_attr "length" "4")
+   (set_attr "type" "call")]
+)
+
 (define_insn "*call_reg_thumb1"
   [(call (mem:SI (match_operand:SI 0 "register_operand" "l*r"))
 	 (match_operand 1 "" ""))
@@ -1737,6 +1750,21 @@ 
    (set_attr "type" "call")]
 )
 
+(define_insn "*nonsecure_call_value_reg_thumb1_v5"
+  [(set (match_operand 0 "" "")
+	(call (unspec:SI
+	       [(mem:SI (match_operand:SI 1 "register_operand" "l*r"))]
+	       UNSPEC_NONSECURE_MEM)
+	      (match_operand 2 "" "")))
+   (use (match_operand 3 "" ""))
+   (clobber (reg:SI LR_REGNUM))
+   (clobber (match_dup 1))]
+  "TARGET_THUMB1 && use_cmse"
+  "bl\\t__gnu_cmse_nonsecure_call"
+  [(set_attr "length" "4")
+   (set_attr "type" "call")]
+)
+
 (define_insn "*call_value_reg_thumb1"
   [(set (match_operand 0 "" "")
 	(call (mem:SI (match_operand:SI 1 "register_operand" "l*r"))
diff --git a/gcc/config/arm/thumb2.md b/gcc/config/arm/thumb2.md
index 7161571fbbd08af739c64d78119ee28032aaa412..c1c8213275738c5008b69587f764b92ead7772b8 100644
--- a/gcc/config/arm/thumb2.md
+++ b/gcc/config/arm/thumb2.md
@@ -581,6 +581,19 @@ 
   [(set_attr "type" "call")]
 )
 
+(define_insn "*nonsecure_call_reg_thumb2"
+  [(call (unspec:SI [(mem:SI (match_operand:SI 0 "s_register_operand" "r"))]
+		    UNSPEC_NONSECURE_MEM)
+	 (match_operand 1 "" ""))
+   (use (match_operand 2 "" ""))
+   (clobber (reg:SI LR_REGNUM))
+   (clobber (match_dup 0))]
+  "TARGET_THUMB2 && use_cmse"
+  "bl\\t__gnu_cmse_nonsecure_call"
+  [(set_attr "length" "4")
+   (set_attr "type" "call")]
+)
+
 (define_insn "*call_value_reg_thumb2"
   [(set (match_operand 0 "" "")
 	(call (mem:SI (match_operand:SI 1 "register_operand" "l*r"))
@@ -592,6 +605,21 @@ 
   [(set_attr "type" "call")]
 )
 
+(define_insn "*nonsecure_call_value_reg_thumb2"
+  [(set (match_operand 0 "" "")
+	(call
+	 (unspec:SI [(mem:SI (match_operand:SI 1 "register_operand" "l*r"))]
+		    UNSPEC_NONSECURE_MEM)
+	 (match_operand 2 "" "")))
+   (use (match_operand 3 "" ""))
+   (clobber (reg:SI LR_REGNUM))
+   (clobber (match_dup 1))]
+  "TARGET_THUMB2 && use_cmse"
+  "bl\t__gnu_cmse_nonsecure_call"
+  [(set_attr "length" "4")
+   (set_attr "type" "call")]
+)
+
 (define_insn "*thumb2_indirect_jump"
   [(set (pc)
 	(match_operand:SI 0 "register_operand" "l*r"))]
diff --git a/gcc/config/arm/unspecs.md b/gcc/config/arm/unspecs.md
index 5744c62cf5ab3b3711fef2b442c62cd3eadf0793..c2f5e5b52cd84380c009af4077bd6effda923701 100644
--- a/gcc/config/arm/unspecs.md
+++ b/gcc/config/arm/unspecs.md
@@ -84,6 +84,8 @@ 
   UNSPEC_VRINTA         ; Represent a float to integral float rounding
                         ; towards nearest, ties away from zero.
   UNSPEC_PROBE_STACK    ; Probe stack memory reference
+  UNSPEC_NONSECURE_MEM	; Represent non-secure memory in ARMv8-M with
+			; security extension
 ])
 
 (define_c_enum "unspec" [
diff --git a/gcc/testsuite/gcc.target/arm/cmse/baseline/cmse-11.c b/gcc/testsuite/gcc.target/arm/cmse/baseline/cmse-11.c
new file mode 100644
index 0000000000000000000000000000000000000000..3007409ad88b9d2312ece4d55f3695787a1aa566
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/baseline/cmse-11.c
@@ -0,0 +1,22 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_base_ok } */
+/* { dg-add-options arm_arch_v8m_base } */
+/* { dg-options "-mcmse" }  */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (int);
+
+int
+foo (int a)
+{
+  return bar (bar (a + 1));
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "movs\tr1, r4" } } */
+/* { dg-final { scan-assembler "movs\tr2, r4" } } */
+/* { dg-final { scan-assembler "movs\tr3, r4" } } */
+
+/* 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/baseline/cmse-13.c b/gcc/testsuite/gcc.target/arm/cmse/baseline/cmse-13.c
new file mode 100644
index 0000000000000000000000000000000000000000..f2b931be5912b421f8427535ded3298883bec639
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/baseline/cmse-13.c
@@ -0,0 +1,25 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_base_ok } */
+/* { dg-add-options arm_arch_v8m_base } */
+/* { dg-options "-mcmse" } */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (float, double);
+
+int
+foo (int a)
+{
+  return bar (1.0f, 2.0) + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler-not "movs\tr0, r4" } } */
+/* { dg-final { scan-assembler "\n\tmovs\tr1, r4" } } */
+/* { dg-final { scan-assembler-not "\n\tmovs\tr2, r4\n\tmovs\tr3, r4" } } */
+/* { dg-final { scan-assembler-not "vmov" } } */
+/* { dg-final { scan-assembler-not "vmsr" } } */
+
+/* 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/baseline/cmse-6.c b/gcc/testsuite/gcc.target/arm/cmse/baseline/cmse-6.c
new file mode 100644
index 0000000000000000000000000000000000000000..95da045690ae170a101949134482e821f27fe3aa
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/baseline/cmse-6.c
@@ -0,0 +1,21 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_base_ok } */
+/* { dg-add-options arm_arch_v8m_base } */
+/* { dg-options "-mcmse" }  */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (double);
+
+int
+foo (int a)
+{
+  return bar (2.0) + a + 1;
+}
+
+/* Remember dont clear r0 and r1, because we are passing the double parameter
+ * for bar in them.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "movs\tr2, r4" } } */
+
+/* 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/cmse-9.c b/gcc/testsuite/gcc.target/arm/cmse/cmse-9.c
index 2846ce6cc203f1810977656bdc09ff86419b7007..1dd82b5cee735025054e517d1bbd9c6ab1ad462a 100644
--- a/gcc/testsuite/gcc.target/arm/cmse/cmse-9.c
+++ b/gcc/testsuite/gcc.target/arm/cmse/cmse-9.c
@@ -3,11 +3,19 @@ 
 /* { dg-skip-if "Testing exclusion of -mcmse" { arm-*-* } { "-mcmse" } { "" } }  */
 
 
+void __attribute__ ((cmse_nonsecure_call)) (*bar) (int);
+typedef void __attribute__ ((cmse_nonsecure_call)) baz (int);
+
 int __attribute__ ((cmse_nonsecure_entry))
-foo (int a)
+foo (int a, baz b)
 {
+  bar (a);
+  b (a);
   return a + 1;
 }
 
+/* { dg-final { scan-assembler-not "bxns" } } */
+/* { dg-final { scan-assembler-not "blxns" } } */
+/* { dg-final { scan-assembler-not "bl\t__gnu_cmse_nonsecure_call" } } */
 /* { dg-final { scan-assembler "foo:" } } */
 /* { dg-final { scan-assembler-not "__acle_se_foo:" } } */
diff --git a/gcc/testsuite/gcc.target/arm/cmse/mainline/hard-sp/cmse-13.c b/gcc/testsuite/gcc.target/arm/cmse/mainline/hard-sp/cmse-13.c
new file mode 100644
index 0000000000000000000000000000000000000000..285fa854b3d61c4d856e9be863ee73a87f1b11a9
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/hard-sp/cmse-13.c
@@ -0,0 +1,43 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_main_ok } */
+/* { dg-add-options arm_arch_v8m_main } */
+/* { dg-skip-if "Do not combine float-abi= hard | soft | softfp" {*-*-*} {"-mfloat-abi=soft" -mfloat-abi=softfp } {""} } */
+/* { dg-skip-if "Skip these if testing double precision" {*-*-*} {"-mfpu=fpv[4-5]-d16"} {""} } */
+/* { dg-options "-mcmse -mfloat-abi=hard -mfpu=fpv5-sp-d16" }  */
+
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (float, double);
+
+int
+foo (int a)
+{
+  return bar (3.0f, 2.0) + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "mov\tr0, r4" } } */
+/* { dg-final { scan-assembler "mov\tr1, r4" } } */
+/* { dg-final { scan-assembler "mov\tr2, r4" } } */
+/* { dg-final { scan-assembler "mov\tr3, r4" } } */
+/* { 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" } } */
+
+/* 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/hard-sp/cmse-7.c b/gcc/testsuite/gcc.target/arm/cmse/mainline/hard-sp/cmse-7.c
new file mode 100644
index 0000000000000000000000000000000000000000..8f56aaf2a97e740a77473cab245614b6fa696c04
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/hard-sp/cmse-7.c
@@ -0,0 +1,42 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_main_ok } */
+/* { dg-add-options arm_arch_v8m_main } */
+/* { dg-skip-if "Do not combine float-abi= hard | soft | softfp" {*-*-*} {"-mfloat-abi=soft" -mfloat-abi=softfp } {""} } */
+/* { dg-skip-if "Skip these if testing double precision" {*-*-*} {"-mfpu=fpv[4-5]-d16"} {""} } */
+/* { dg-options "-mcmse -mfloat-abi=hard -mfpu=fpv5-sp-d16" }  */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (void);
+
+int
+foo (int a)
+{
+  return bar () + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "mov\tr0, r4" } } */
+/* { dg-final { scan-assembler "mov\tr1, r4" } } */
+/* { dg-final { scan-assembler "mov\tr2, r4" } } */
+/* { dg-final { scan-assembler "mov\tr3, r4" } } */
+/* { 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" } } */
+
+/* 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/hard-sp/cmse-8.c b/gcc/testsuite/gcc.target/arm/cmse/mainline/hard-sp/cmse-8.c
new file mode 100644
index 0000000000000000000000000000000000000000..749f8a7278cb4ea9fd4462887ce7f37b57f83248
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/hard-sp/cmse-8.c
@@ -0,0 +1,41 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_main_ok } */
+/* { dg-add-options arm_arch_v8m_main } */
+/* { dg-skip-if "Do not combine float-abi= hard | soft | softfp" {*-*-*} {"-mfloat-abi=soft" -mfloat-abi=softfp } {""} } */
+/* { dg-skip-if "Skip these if testing double precision" {*-*-*} {"-mfpu=fpv[4-5]-d16"} {""} } */
+/* { dg-options "-mcmse -mfloat-abi=hard -mfpu=fpv5-sp-d16" }  */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (double);
+
+int
+foo (int a)
+{
+  return bar (2.0) + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "mov\tr0, r4" } } */
+/* { dg-final { scan-assembler "mov\tr1, r4" } } */
+/* { dg-final { scan-assembler "mov\tr2, r4" } } */
+/* { dg-final { scan-assembler "mov\tr3, r4" } } */
+/* { 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" } } */
+
+/* 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/hard/cmse-13.c b/gcc/testsuite/gcc.target/arm/cmse/mainline/hard/cmse-13.c
new file mode 100644
index 0000000000000000000000000000000000000000..4ad5068aac263f1c2edb17b8707fc9d1c0af6959
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/hard/cmse-13.c
@@ -0,0 +1,35 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_main_ok } */
+/* { dg-add-options arm_arch_v8m_main } */
+/* { dg-skip-if "Do not combine float-abi= hard | soft | softfp" {*-*-*} {"-mfloat-abi=soft" -mfloat-abi=softfp } {""} } */
+/* { dg-skip-if "Skip these if testing single precision" {*-*-*} {"-mfpu=*-sp-*"} {""} } */
+/* { dg-options "-mcmse -mfloat-abi=hard -mfpu=fpv5-d16" }  */
+
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (float, double);
+
+int
+foo (int a)
+{
+  return bar (3.0f, 2.0) + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "mov\tr0, r4" } } */
+/* { dg-final { scan-assembler "mov\tr1, r4" } } */
+/* { dg-final { scan-assembler "mov\tr2, r4" } } */
+/* { dg-final { scan-assembler "mov\tr3, r4" } } */
+/* { 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\.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" } } */
+
+/* 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/hard/cmse-7.c b/gcc/testsuite/gcc.target/arm/cmse/mainline/hard/cmse-7.c
new file mode 100644
index 0000000000000000000000000000000000000000..b18259eaef6da1933ef2d58d63b97d4bb4e450f4
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/hard/cmse-7.c
@@ -0,0 +1,34 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_main_ok } */
+/* { dg-add-options arm_arch_v8m_main } */
+/* { dg-skip-if "Do not combine float-abi= hard | soft | softfp" {*-*-*} {"-mfloat-abi=soft" -mfloat-abi=softfp } {""} } */
+/* { dg-skip-if "Skip these if testing single precision" {*-*-*} {"-mfpu=*-sp-*"} {""} } */
+/* { dg-options "-mcmse -mfloat-abi=hard -mfpu=fpv5-d16" }  */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (void);
+
+int
+foo (int a)
+{
+  return bar () + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "mov\tr0, r4" } } */
+/* { dg-final { scan-assembler "mov\tr1, r4" } } */
+/* { dg-final { scan-assembler "mov\tr2, r4" } } */
+/* { dg-final { scan-assembler "mov\tr3, r4" } } */
+/* { 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" } } */
+
+/* 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/hard/cmse-8.c b/gcc/testsuite/gcc.target/arm/cmse/mainline/hard/cmse-8.c
new file mode 100644
index 0000000000000000000000000000000000000000..182ddf9540374f30ae3694259f1edb62161455e1
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/hard/cmse-8.c
@@ -0,0 +1,33 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_main_ok } */
+/* { dg-add-options arm_arch_v8m_main } */
+/* { dg-skip-if "Do not combine float-abi= hard | soft | softfp" {*-*-*} {"-mfloat-abi=soft" -mfloat-abi=softfp } {""} } */
+/* { dg-skip-if "Skip these if testing single precision" {*-*-*} {"-mfpu=*-sp-*"} {""} } */
+/* { dg-options "-mcmse -mfloat-abi=hard -mfpu=fpv5-d16" }  */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (double);
+
+int
+foo (int a)
+{
+  return bar (2.0) + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "mov\tr0, r4" } } */
+/* { dg-final { scan-assembler "mov\tr1, r4" } } */
+/* { dg-final { scan-assembler "mov\tr2, r4" } } */
+/* { dg-final { scan-assembler "mov\tr3, r4" } } */
+/* { 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" } } */
+
+/* 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/soft/cmse-13.c b/gcc/testsuite/gcc.target/arm/cmse/mainline/soft/cmse-13.c
new file mode 100644
index 0000000000000000000000000000000000000000..dbbd262c89085eb0dc8737ad7955c3a12531492b
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/soft/cmse-13.c
@@ -0,0 +1,27 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_main_ok } */
+/* { dg-add-options arm_arch_v8m_main } */
+/* { dg-skip-if "Do not combine float-abi= hard | soft | softfp" {*-*-*} {"-mfloat-abi=hard" -mfloat-abi=softfp } {""} } */
+/* { dg-options "-mcmse -mfloat-abi=soft" }  */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (float, double);
+
+int
+foo (int a)
+{
+  return bar (1.0f, 2.0) + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler-not "mov\tr0, r4" } } */
+/* { dg-final { scan-assembler "mov\tr1, r4" } } */
+/* { dg-final { scan-assembler-not "mov\tr2, r4" } } */
+/* { dg-final { scan-assembler-not "mov\tr3, r4" } } */
+/* { dg-final { scan-assembler-not "vmov" } } */
+/* { dg-final { scan-assembler-not "vmsr" } } */
+
+/* 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/soft/cmse-7.c b/gcc/testsuite/gcc.target/arm/cmse/mainline/soft/cmse-7.c
new file mode 100644
index 0000000000000000000000000000000000000000..e33568400ef830abaa019b5c2727a9f3b5d5704b
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/soft/cmse-7.c
@@ -0,0 +1,27 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_main_ok } */
+/* { dg-add-options arm_arch_v8m_main } */
+/* { dg-skip-if "Do not combine float-abi= hard | soft | softfp" {*-*-*} {"-mfloat-abi=hard" -mfloat-abi=softfp } {""} } */
+/* { dg-options "-mcmse -mfloat-abi=soft" }  */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (void);
+
+int
+foo (int a)
+{
+  return bar () + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "mov\tr0, r4" } } */
+/* { dg-final { scan-assembler "mov\tr1, r4" } } */
+/* { dg-final { scan-assembler "mov\tr2, r4" } } */
+/* { dg-final { scan-assembler "mov\tr3, r4" } } */
+/* { dg-final { scan-assembler-not "vmov" } } */
+/* { dg-final { scan-assembler-not "vmsr" } } */
+
+/* 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/soft/cmse-8.c b/gcc/testsuite/gcc.target/arm/cmse/mainline/soft/cmse-8.c
new file mode 100644
index 0000000000000000000000000000000000000000..024a12e0a414730eb1cbebfca8574d548238ab28
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/soft/cmse-8.c
@@ -0,0 +1,26 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_main_ok } */
+/* { dg-add-options arm_arch_v8m_main } */
+/* { dg-skip-if "Do not combine float-abi= hard | soft | softfp" {*-*-*} {"-mfloat-abi=hard" -mfloat-abi=softfp } {""} } */
+/* { dg-options "-mcmse -mfloat-abi=soft" }  */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (double);
+
+int
+foo (int a)
+{
+  return bar (2.0) + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler-not "mov\tr0, r4" } } */
+/* { dg-final { scan-assembler-not "mov\tr1, r4" } } */
+/* { dg-final { scan-assembler "mov\tr2, r4" } } */
+/* { dg-final { scan-assembler "mov\tr3, r4" } } */
+/* { dg-final { scan-assembler-not "vmov" } } */
+/* { dg-final { scan-assembler-not "vmsr" } } */
+
+/* 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/softfp-sp/cmse-7.c b/gcc/testsuite/gcc.target/arm/cmse/mainline/softfp-sp/cmse-7.c
new file mode 100644
index 0000000000000000000000000000000000000000..0cf1a68267ee1b3b0207e149e17b3d3fbcc57e40
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/softfp-sp/cmse-7.c
@@ -0,0 +1,42 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_main_ok } */
+/* { dg-add-options arm_arch_v8m_main } */
+/* { dg-skip-if "Do not combine float-abi= hard | soft | softfp" {*-*-*} {"-mfloat-abi=soft" -mfloat-abi=hard } {""} } */
+/* { dg-skip-if "Skip these if testing double precision" {*-*-*} {"-mfpu=fpv[4-5]-d16"} {""} } */
+/* { dg-options "-mcmse -mfloat-abi=softfp -mfpu=fpv5-sp-d16" }  */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (void);
+
+int
+foo (int a)
+{
+  return bar () + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "mov\tr0, r4" } } */
+/* { dg-final { scan-assembler "mov\tr1, r4" } } */
+/* { dg-final { scan-assembler "mov\tr2, r4" } } */
+/* { dg-final { scan-assembler "mov\tr3, r4" } } */
+/* { 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" } } */
+
+/* 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/softfp-sp/cmse-8.c b/gcc/testsuite/gcc.target/arm/cmse/mainline/softfp-sp/cmse-8.c
new file mode 100644
index 0000000000000000000000000000000000000000..e9be53a42f1d62929327a531dda9544bd4c2df62
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/softfp-sp/cmse-8.c
@@ -0,0 +1,41 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_main_ok } */
+/* { dg-add-options arm_arch_v8m_main } */
+/* { dg-skip-if "Do not combine float-abi= hard | soft | softfp" {*-*-*} {"-mfloat-abi=soft" -mfloat-abi=hard } {""} } */
+/* { dg-skip-if "Skip these if testing double precision" {*-*-*} {"-mfpu=fpv[4-5]-d16"} {""} } */
+/* { dg-options "-mcmse -mfloat-abi=softfp -mfpu=fpv5-sp-d16" }  */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (double);
+
+int
+foo (int a)
+{
+  return bar (2.0) + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler-not "mov\tr0, r4" } } */
+/* { dg-final { scan-assembler-not "mov\tr1, r4" } } */
+/* { dg-final { scan-assembler "mov\tr2, r4" } } */
+/* { dg-final { scan-assembler "mov\tr3, r4" } } */
+/* { 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\ts3, #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" } } */
+
+/* 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/softfp/cmse-13.c b/gcc/testsuite/gcc.target/arm/cmse/mainline/softfp/cmse-13.c
new file mode 100644
index 0000000000000000000000000000000000000000..515fd22b9da2216d3f27ed4b0535dc3ab88e0eea
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/softfp/cmse-13.c
@@ -0,0 +1,34 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_main_ok } */
+/* { dg-add-options arm_arch_v8m_main } */
+/* { dg-skip-if "Do not combine float-abi= hard | soft | softfp" {*-*-*} {"-mfloat-abi=soft" -mfloat-abi=hard } {""} } */
+/* { dg-skip-if "Skip these if testing single precision" {*-*-*} {"-mfpu=*-sp-*"} {""} } */
+/* { dg-options "-mcmse -mfloat-abi=softfp -mfpu=fpv5-d16" }  */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (float, double);
+
+int
+foo (int a)
+{
+  return bar (1.0f, 2.0) + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler-not "mov\tr0, r4" } } */
+/* { dg-final { scan-assembler "\n\tmov\tr1, r4" } } */
+/* { dg-final { scan-assembler-not "\n\tmov\tr2, r4\n\tmov\tr3, r4" } } */
+
+/* { 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" } } */
+
+/* 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/softfp/cmse-7.c b/gcc/testsuite/gcc.target/arm/cmse/mainline/softfp/cmse-7.c
new file mode 100644
index 0000000000000000000000000000000000000000..fbe3f6d4fd10700269995cde4aab0f105165c56d
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/softfp/cmse-7.c
@@ -0,0 +1,34 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_main_ok } */
+/* { dg-add-options arm_arch_v8m_main } */
+/* { dg-skip-if "Do not combine float-abi= hard | soft | softfp" {*-*-*} {"-mfloat-abi=soft" -mfloat-abi=hard } {""} } */
+/* { dg-skip-if "Skip these if testing single precision" {*-*-*} {"-mfpu=*-sp-*"} {""} } */
+/* { dg-options "-mcmse -mfloat-abi=softfp -mfpu=fpv5-d16" }  */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (void);
+
+int
+foo (int a)
+{
+  return bar () + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "mov\tr0, r4" } } */
+/* { dg-final { scan-assembler "mov\tr1, r4" } } */
+/* { dg-final { scan-assembler "mov\tr2, r4" } } */
+/* { dg-final { scan-assembler "mov\tr3, r4" } } */
+/* { 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" } } */
+
+/* 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/softfp/cmse-8.c b/gcc/testsuite/gcc.target/arm/cmse/mainline/softfp/cmse-8.c
new file mode 100644
index 0000000000000000000000000000000000000000..4ae1b2f90f6164b4252029fbdac2e9f820d0741c
--- /dev/null
+++ b/gcc/testsuite/gcc.target/arm/cmse/mainline/softfp/cmse-8.c
@@ -0,0 +1,33 @@ 
+/* { dg-do compile } */
+/* { dg-require-effective-target arm_arch_v8m_main_ok } */
+/* { dg-add-options arm_arch_v8m_main } */
+/* { dg-skip-if "Do not combine float-abi= hard | soft | softfp" {*-*-*} {"-mfloat-abi=soft" -mfloat-abi=hard } {""} } */
+/* { dg-skip-if "Skip these if testing single precision" {*-*-*} {"-mfpu=*-sp-*"} {""} } */
+/* { dg-options "-mcmse -mfloat-abi=softfp -mfpu=fpv5-d16" }  */
+
+int __attribute__ ((cmse_nonsecure_call)) (*bar) (double);
+
+int
+foo (int a)
+{
+  return bar (2.0) + a + 1;
+}
+
+/* Checks for saving and clearing prior to function call.  */
+/* { dg-final { scan-assembler "lsrs\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler "lsls\tr4, r4, #1" } } */
+/* { dg-final { scan-assembler-not "mov\tr0, r4" } } */
+/* { dg-final { scan-assembler-not "mov\tr1, r4" } } */
+/* { dg-final { scan-assembler "mov\tr2, r4" } } */
+/* { dg-final { scan-assembler "mov\tr3, r4" } } */
+/* { 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" } } */
+
+/* Now we check that we use the correct intrinsic to call.  */
+/* { dg-final { scan-assembler "bl\t__gnu_cmse_nonsecure_call" } } */
diff --git a/libgcc/config/arm/cmse_nonsecure_call.S b/libgcc/config/arm/cmse_nonsecure_call.S
new file mode 100644
index 0000000000000000000000000000000000000000..a14b583486b2462cd91f42ab46eb80d1f04caf8a
--- /dev/null
+++ b/libgcc/config/arm/cmse_nonsecure_call.S
@@ -0,0 +1,113 @@ 
+/* CMSE wrapper function used to save, clear and restore callee saved registers
+   for cmse_nonsecure_call's.
+
+   Copyright (C) 2016 Free Software Foundation, Inc.
+   Contributed by ARM Ltd.
+
+   This file is free software; you can redistribute it and/or modify it
+   under the terms of the GNU General Public License as published by the
+   Free Software Foundation; either version 3, or (at your option) any
+   later version.
+
+   This file is distributed in the hope that it will be useful, but
+   WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   General Public License for more details.
+
+   Under Section 7 of GPL version 3, you are granted additional
+   permissions described in the GCC Runtime Library Exception, version
+   3.1, as published by the Free Software Foundation.
+
+   You should have received a copy of the GNU General Public License and
+   a copy of the GCC Runtime Library Exception along with this program;
+   see the files COPYING3 and COPYING.RUNTIME respectively.  If not, see
+   <http://www.gnu.org/licenses/>.  */
+
+.syntax unified
+.thumb
+.global __gnu_cmse_nonsecure_call
+__gnu_cmse_nonsecure_call:
+#if defined(__ARM_ARCH_8M_MAIN__)
+push	    {r5-r11,lr}
+mov	    r5, r4
+mov	    r6, r4
+mov	    r7, r4
+mov	    r8, r4
+mov	    r9, r4
+mov	    r10, r4
+mov	    r11, r4
+mov	    ip, r4
+#if !defined(__SOFTFP__)
+vpush.f64   {d8-d15}
+#if __ARM_FP & 0x04
+vmov.f32    s16, #1.00000
+vmov.f32    s17, #1.00000
+vmov.f32    s18, #1.00000
+vmov.f32    s19, #1.00000
+vmov.f32    s20, #1.00000
+vmov.f32    s21, #1.00000
+vmov.f32    s22, #1.00000
+vmov.f32    s23, #1.00000
+vmov.f32    s24, #1.00000
+vmov.f32    s25, #1.00000
+vmov.f32    s26, #1.00000
+vmov.f32    s27, #1.00000
+vmov.f32    s28, #1.00000
+vmov.f32    s29, #1.00000
+vmov.f32    s30, #1.00000
+vmov.f32    s31, #1.00000
+#elif __ARM_FP & 0x08
+vmov.f64    d8, #1.00000
+vmov.f64    d9, #1.00000
+vmov.f64    d10, #1.00000
+vmov.f64    d11, #1.00000
+vmov.f64    d12, #1.00000
+vmov.f64    d13, #1.00000
+vmov.f64    d14, #1.00000
+vmov.f64    d15, #1.00000
+#else
+#error "Half precision implementation not supported."
+#endif
+vmsr	    fpscr, r4
+#endif
+#if defined(__ARM_FEATURE_DSP)
+msr	    APSR_nzcvqg, r4
+#else
+msr	    APSR_nzcvq, r4
+#endif
+blxns	    r4
+#if !defined(__SOFTFP__)
+vpop.f64    {d8-d15}
+#endif
+pop	    {r5-r11}
+#elif defined (__ARM_ARCH_8M_BASE__)
+push	    {r5-r7}
+mov	    r5, r8
+mov	    r6, r9
+mov	    r7, r10
+push	    {r5-r7}
+mov	    r5, r11
+mov	    r6, lr
+push	    {r5,r6}
+mov	    r5, r4
+mov	    r6, r4
+mov	    r7, r4
+mov	    r8, r4
+mov	    r9, r4
+mov	    r10, r4
+mov	    r11, r4
+mov	    ip, r4
+msr	    APSR_nzcvq, r4
+blxns	    r4
+pop	    {r5, r6}
+mov	    lr, r6
+mov	    r11, r5
+pop	    {r5-r7}
+mov	    r10, r7
+mov	    r9, r6
+mov	    r8, r5
+pop	    {r5-r7}
+blx	    lr
+#else
+#error "This should only be used for armv8-m base- and mainline."
+#endif
diff --git a/libgcc/config/arm/t-arm b/libgcc/config/arm/t-arm
index 5618143bfd0f02b170db3f9e4c0a15cecb403cec..9e85ac06b146feaa14ab83374a0156737b870310 100644
--- a/libgcc/config/arm/t-arm
+++ b/libgcc/config/arm/t-arm
@@ -12,4 +12,6 @@  libgcc-objects += cmse.o cmse_nonsecure_call.o
 
 cmse.o: $(srcdir)/config/arm/cmse.c
 	$(gcc_compile) -c $(CMSE_OPTS) $<
+cmse_nonsecure_call.o: $(srcdir)/config/arm/cmse_nonsecure_call.S
+		       $(gcc_compile) -c $<
 endif