diff mbox series

[2/2] Arm: Add NEON and MVE RTL patterns for Complex Addition.

Message ID patch-13917-tamar@arm.com
State New
Headers show
Series None | expand

Commit Message

Tamar Christina Dec. 16, 2020, 8:37 p.m. UTC 
Hi All,

This adds implementation for the optabs for complex additions.  With this the
following C code:

  void f90 (float complex a[restrict N], float complex b[restrict N],
	    float complex c[restrict N])
  {
    for (int i=0; i < N; i++)
      c[i] = a[i] + (b[i] * I);
  }

generates

  f90:
	  add     r3, r2, #1600
  .L2:
	  vld1.32 {q8}, [r0]!
	  vld1.32 {q9}, [r1]!
	  vcadd.f32       q8, q8, q9, #90
	  vst1.32 {q8}, [r2]!
	  cmp     r3, r2
	  bne     .L2
	  bx      lr


instead of

  f90:
	  add     r3, r2, #1600
  .L2:
	  vld2.32 {d24-d27}, [r0]!
	  vld2.32 {d20-d23}, [r1]!
	  vsub.f32	q8, q12, q11
	  vadd.f32	q9, q13, q10
	  vst2.32 {d16-d19}, [r2]!
	  cmp     r3, r2
	  bne     .L2
	  bx      lr

Bootstrapped Regtested on arm-none-linux-gnueabihf and no issues.
Codegen tested for -march=armv8.1-m.main+mve.fp -mfloat-abi=hard -mfpu=auto
and no issues.

This is just a splitting of a previously approved patch due to it having a
dependency on the AArch64 bits which have been requested to be reworked.

Will commit under the previous approval.

Thanks,
Tamar

gcc/ChangeLog:

	* config/arm/arm_mve.h (__arm_vcaddq_rot90_u8, __arm_vcaddq_rot270_u8,
	__arm_vcaddq_rot90_s8, __arm_vcaddq_rot270_s8,
	__arm_vcaddq_rot90_u16, __arm_vcaddq_rot270_u16,
	__arm_vcaddq_rot90_s16, __arm_vcaddq_rot270_s16,
	__arm_vcaddq_rot90_u32, __arm_vcaddq_rot270_u32,
	__arm_vcaddq_rot90_s32, __arm_vcaddq_rot270_s32,
	__arm_vcaddq_rot90_f16, __arm_vcaddq_rot270_f16,
	__arm_vcaddq_rot90_f32, __arm_vcaddq_rot270_f32):  Update builtin calls.
	* config/arm/arm_mve_builtins.def (vcaddq_rot90_u, vcaddq_rot270_u,
	vcaddq_rot90_s, vcaddq_rot270_s, vcaddq_rot90_f, vcaddq_rot270_f):
	Removed.
	(vcaddq_rot90, vcaddq_rot270): New.
	* config/arm/constraints.md (Dz): Include MVE.
	* config/arm/iterators.md (mve_rot): New.
	(supf): Remove VCADDQ_ROT270_S, VCADDQ_ROT270_U, VCADDQ_ROT90_S,
	VCADDQ_ROT90_U.
	(VCADDQ_ROT270, VCADDQ_ROT90): Removed.
	* config/arm/mve.md (mve_vcaddq_rot270_<supf><mode,
	mve_vcaddq_rot90_<supf><mode>, mve_vcaddq_rot270_f<mode>,
	mve_vcaddq_rot90_f<mode>): Removed.
	(mve_vcaddq<mve_rot><mode>, mve_vcaddq<mve_rot><mode>): New.
	* config/arm/unspecs.md (VCADDQ_ROT270_S, VCADDQ_ROT90_S,
	VCADDQ_ROT270_U, VCADDQ_ROT90_U, VCADDQ_ROT270_F,
	VCADDQ_ROT90_F): Removed.
	* config/arm/vec-common.md (cadd<rot><mode>3): New.

--- inline copy of patch -- 
diff --git a/gcc/config/arm/arm_mve.h b/gcc/config/arm/arm_mve.h
index 6c0d1e2e634a32196eb31079166a7733dcd3a4b6..987495dd234ad96ba1163a1f482fe183a46ff437 100644


--
diff mbox series

Patch

diff --git a/gcc/config/arm/arm_mve.h b/gcc/config/arm/arm_mve.h
index 6c0d1e2e634a32196eb31079166a7733dcd3a4b6..987495dd234ad96ba1163a1f482fe183a46ff437 100644
--- a/gcc/config/arm/arm_mve.h
+++ b/gcc/config/arm/arm_mve.h
@@ -3981,14 +3981,16 @@  __extension__ extern __inline uint8x16_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot90_u8 (uint8x16_t __a, uint8x16_t __b)
 {
-  return __builtin_mve_vcaddq_rot90_uv16qi (__a, __b);
+  return (uint8x16_t)
+    __builtin_mve_vcaddq_rot90v16qi ((int8x16_t)__a, (int8x16_t)__b);
 }
 
 __extension__ extern __inline uint8x16_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot270_u8 (uint8x16_t __a, uint8x16_t __b)
 {
-  return __builtin_mve_vcaddq_rot270_uv16qi (__a, __b);
+  return (uint8x16_t)
+    __builtin_mve_vcaddq_rot270v16qi ((int8x16_t)__a, (int8x16_t)__b);
 }
 
 __extension__ extern __inline uint8x16_t
@@ -4520,14 +4522,14 @@  __extension__ extern __inline int8x16_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot90_s8 (int8x16_t __a, int8x16_t __b)
 {
-  return __builtin_mve_vcaddq_rot90_sv16qi (__a, __b);
+  return __builtin_mve_vcaddq_rot90v16qi (__a, __b);
 }
 
 __extension__ extern __inline int8x16_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot270_s8 (int8x16_t __a, int8x16_t __b)
 {
-  return __builtin_mve_vcaddq_rot270_sv16qi (__a, __b);
+  return __builtin_mve_vcaddq_rot270v16qi (__a, __b);
 }
 
 __extension__ extern __inline int8x16_t
@@ -4821,14 +4823,16 @@  __extension__ extern __inline uint16x8_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot90_u16 (uint16x8_t __a, uint16x8_t __b)
 {
-  return __builtin_mve_vcaddq_rot90_uv8hi (__a, __b);
+  return (uint16x8_t)
+    __builtin_mve_vcaddq_rot90v8hi ((int16x8_t)__a, (int16x8_t)__b);
 }
 
 __extension__ extern __inline uint16x8_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot270_u16 (uint16x8_t __a, uint16x8_t __b)
 {
-  return __builtin_mve_vcaddq_rot270_uv8hi (__a, __b);
+  return (uint16x8_t)
+    __builtin_mve_vcaddq_rot270v8hi ((int16x8_t)__a, (int16x8_t)__b);
 }
 
 __extension__ extern __inline uint16x8_t
@@ -5360,14 +5364,14 @@  __extension__ extern __inline int16x8_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot90_s16 (int16x8_t __a, int16x8_t __b)
 {
-  return __builtin_mve_vcaddq_rot90_sv8hi (__a, __b);
+  return __builtin_mve_vcaddq_rot90v8hi (__a, __b);
 }
 
 __extension__ extern __inline int16x8_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot270_s16 (int16x8_t __a, int16x8_t __b)
 {
-  return __builtin_mve_vcaddq_rot270_sv8hi (__a, __b);
+  return __builtin_mve_vcaddq_rot270v8hi (__a, __b);
 }
 
 __extension__ extern __inline int16x8_t
@@ -5661,14 +5665,16 @@  __extension__ extern __inline uint32x4_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot90_u32 (uint32x4_t __a, uint32x4_t __b)
 {
-  return __builtin_mve_vcaddq_rot90_uv4si (__a, __b);
+  return (uint32x4_t)
+    __builtin_mve_vcaddq_rot90v4si ((int32x4_t)__a, (int32x4_t)__b);
 }
 
 __extension__ extern __inline uint32x4_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot270_u32 (uint32x4_t __a, uint32x4_t __b)
 {
-  return __builtin_mve_vcaddq_rot270_uv4si (__a, __b);
+  return (uint32x4_t)
+    __builtin_mve_vcaddq_rot270v4si ((int32x4_t)__a, (int32x4_t)__b);
 }
 
 __extension__ extern __inline uint32x4_t
@@ -6200,14 +6206,14 @@  __extension__ extern __inline int32x4_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot90_s32 (int32x4_t __a, int32x4_t __b)
 {
-  return __builtin_mve_vcaddq_rot90_sv4si (__a, __b);
+  return __builtin_mve_vcaddq_rot90v4si (__a, __b);
 }
 
 __extension__ extern __inline int32x4_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot270_s32 (int32x4_t __a, int32x4_t __b)
 {
-  return __builtin_mve_vcaddq_rot270_sv4si (__a, __b);
+  return __builtin_mve_vcaddq_rot270v4si (__a, __b);
 }
 
 __extension__ extern __inline int32x4_t
@@ -17370,14 +17376,14 @@  __extension__ extern __inline float16x8_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot90_f16 (float16x8_t __a, float16x8_t __b)
 {
-  return __builtin_mve_vcaddq_rot90_fv8hf (__a, __b);
+  return __builtin_mve_vcaddq_rot90v8hf (__a, __b);
 }
 
 __extension__ extern __inline float16x8_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot270_f16 (float16x8_t __a, float16x8_t __b)
 {
-  return __builtin_mve_vcaddq_rot270_fv8hf (__a, __b);
+  return __builtin_mve_vcaddq_rot270v8hf (__a, __b);
 }
 
 __extension__ extern __inline float16x8_t
@@ -17622,14 +17628,14 @@  __extension__ extern __inline float32x4_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot90_f32 (float32x4_t __a, float32x4_t __b)
 {
-  return __builtin_mve_vcaddq_rot90_fv4sf (__a, __b);
+  return __builtin_mve_vcaddq_rot90v4sf (__a, __b);
 }
 
 __extension__ extern __inline float32x4_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 __arm_vcaddq_rot270_f32 (float32x4_t __a, float32x4_t __b)
 {
-  return __builtin_mve_vcaddq_rot270_fv4sf (__a, __b);
+  return __builtin_mve_vcaddq_rot270v4sf (__a, __b);
 }
 
 __extension__ extern __inline float32x4_t
diff --git a/gcc/config/arm/arm_mve_builtins.def b/gcc/config/arm/arm_mve_builtins.def
index f38926ffd8e44f63d25a8fb9bf8f7d8680570ef0..b86e1793e686b07367732c4ffd0603deb78830bd 100644
--- a/gcc/config/arm/arm_mve_builtins.def
+++ b/gcc/config/arm/arm_mve_builtins.def
@@ -125,8 +125,6 @@  VAR3 (BINOP_UNONE_UNONE_UNONE, vcmpeqq_u, v16qi, v8hi, v4si)
 VAR3 (BINOP_UNONE_UNONE_UNONE, vcmpeqq_n_u, v16qi, v8hi, v4si)
 VAR3 (BINOP_UNONE_UNONE_UNONE, vcmpcsq_u, v16qi, v8hi, v4si)
 VAR3 (BINOP_UNONE_UNONE_UNONE, vcmpcsq_n_u, v16qi, v8hi, v4si)
-VAR3 (BINOP_UNONE_UNONE_UNONE, vcaddq_rot90_u, v16qi, v8hi, v4si)
-VAR3 (BINOP_UNONE_UNONE_UNONE, vcaddq_rot270_u, v16qi, v8hi, v4si)
 VAR3 (BINOP_UNONE_UNONE_UNONE, vbicq_u, v16qi, v8hi, v4si)
 VAR3 (BINOP_UNONE_UNONE_UNONE, vandq_u, v16qi, v8hi, v4si)
 VAR3 (BINOP_UNONE_UNONE_UNONE, vaddvq_p_u, v16qi, v8hi, v4si)
@@ -202,8 +200,6 @@  VAR3 (BINOP_NONE_NONE_NONE, vhcaddq_rot270_s, v16qi, v8hi, v4si)
 VAR3 (BINOP_NONE_NONE_NONE, vhaddq_s, v16qi, v8hi, v4si)
 VAR3 (BINOP_NONE_NONE_NONE, vhaddq_n_s, v16qi, v8hi, v4si)
 VAR3 (BINOP_NONE_NONE_NONE, veorq_s, v16qi, v8hi, v4si)
-VAR3 (BINOP_NONE_NONE_NONE, vcaddq_rot90_s, v16qi, v8hi, v4si)
-VAR3 (BINOP_NONE_NONE_NONE, vcaddq_rot270_s, v16qi, v8hi, v4si)
 VAR3 (BINOP_NONE_NONE_NONE, vbrsrq_n_s, v16qi, v8hi, v4si)
 VAR3 (BINOP_NONE_NONE_NONE, vbicq_s, v16qi, v8hi, v4si)
 VAR3 (BINOP_NONE_NONE_NONE, vandq_s, v16qi, v8hi, v4si)
@@ -268,8 +264,6 @@  VAR2 (BINOP_NONE_NONE_NONE, vcmulq_rot90_f, v8hf, v4sf)
 VAR2 (BINOP_NONE_NONE_NONE, vcmulq_rot270_f, v8hf, v4sf)
 VAR2 (BINOP_NONE_NONE_NONE, vcmulq_rot180_f, v8hf, v4sf)
 VAR2 (BINOP_NONE_NONE_NONE, vcmulq_f, v8hf, v4sf)
-VAR2 (BINOP_NONE_NONE_NONE, vcaddq_rot90_f, v8hf, v4sf)
-VAR2 (BINOP_NONE_NONE_NONE, vcaddq_rot270_f, v8hf, v4sf)
 VAR2 (BINOP_NONE_NONE_NONE, vbicq_f, v8hf, v4sf)
 VAR2 (BINOP_NONE_NONE_NONE, vandq_f, v8hf, v4sf)
 VAR2 (BINOP_NONE_NONE_NONE, vaddq_n_f, v8hf, v4sf)
@@ -892,3 +886,7 @@  VAR3 (QUADOP_NONE_NONE_UNONE_IMM_UNONE, vshlcq_m_vec_s, v16qi, v8hi, v4si)
 VAR3 (QUADOP_NONE_NONE_UNONE_IMM_UNONE, vshlcq_m_carry_s, v16qi, v8hi, v4si)
 VAR3 (QUADOP_UNONE_UNONE_UNONE_IMM_UNONE, vshlcq_m_vec_u, v16qi, v8hi, v4si)
 VAR3 (QUADOP_UNONE_UNONE_UNONE_IMM_UNONE, vshlcq_m_carry_u, v16qi, v8hi, v4si)
+
+/* optabs without any suffixes.  */
+VAR5 (BINOP_NONE_NONE_NONE, vcaddq_rot90, v16qi, v8hi, v4si, v8hf, v4sf)
+VAR5 (BINOP_NONE_NONE_NONE, vcaddq_rot270, v16qi, v8hi, v4si, v8hf, v4sf)
diff --git a/gcc/config/arm/constraints.md b/gcc/config/arm/constraints.md
index 789e3332abb7495b308509d03ed241d39498a8b6..6ebddb95b4f9c835f10f5265573f27a06ccbd11f 100644
--- a/gcc/config/arm/constraints.md
+++ b/gcc/config/arm/constraints.md
@@ -310,7 +310,7 @@  (define_constraint "Dz"
  "@internal
   In ARM/Thumb-2 state a vector of constant zeros."
  (and (match_code "const_vector")
-      (match_test "TARGET_NEON && op == CONST0_RTX (mode)")))
+      (match_test "(TARGET_NEON || TARGET_HAVE_MVE) && op == CONST0_RTX (mode)")))
 
 (define_constraint "Da"
  "@internal
diff --git a/gcc/config/arm/iterators.md b/gcc/config/arm/iterators.md
index 036a939e2ee758abede76485cdd4946894993111..adf4c017735812fdb318f615ede1407298377519 100644
--- a/gcc/config/arm/iterators.md
+++ b/gcc/config/arm/iterators.md
@@ -1182,6 +1182,9 @@  (define_int_attr rot [(UNSPEC_VCADD90 "90")
 		      (UNSPEC_VCMLA180 "180")
 		      (UNSPEC_VCMLA270 "270")])
 
+(define_int_attr mve_rot [(UNSPEC_VCADD90 "_rot90")
+			  (UNSPEC_VCADD270 "_rot270")])
+
 (define_int_attr simd32_op [(UNSPEC_QADD8 "qadd8") (UNSPEC_QSUB8 "qsub8")
 			    (UNSPEC_SHADD8 "shadd8") (UNSPEC_SHSUB8 "shsub8")
 			    (UNSPEC_UHADD8 "uhadd8") (UNSPEC_UHSUB8 "uhsub8")
@@ -1232,10 +1235,8 @@  (define_int_attr supf [(VCVTQ_TO_F_S "s") (VCVTQ_TO_F_U "u") (VREV16Q_S "s")
 		       (VADDLVQ_P_U "u") (VCMPNEQ_U "u") (VCMPNEQ_S "s")
 		       (VABDQ_M_S "s") (VABDQ_M_U "u") (VABDQ_S "s")
 		       (VABDQ_U "u") (VADDQ_N_S "s") (VADDQ_N_U "u")
-		       (VADDVQ_P_S "s")	(VADDVQ_P_U "u")
-		       (VBRSRQ_N_S "s") (VBRSRQ_N_U "u") (VCADDQ_ROT270_S "s")
-		       (VCADDQ_ROT270_U "u") (VCADDQ_ROT90_S "s")
-		       (VCMPEQQ_S "s") (VCMPEQQ_U "u") (VCADDQ_ROT90_U "u")
+		       (VADDVQ_P_S "s")	(VADDVQ_P_U "u") (VBRSRQ_N_S "s")
+		       (VBRSRQ_N_U "u") (VCMPEQQ_S "s") (VCMPEQQ_U "u")
 		       (VCMPEQQ_N_S "s") (VCMPEQQ_N_U "u") (VCMPNEQ_N_S "s")
 		       (VCMPNEQ_N_U "u")
 		       (VHADDQ_N_S "s") (VHADDQ_N_U "u") (VHADDQ_S "s")
@@ -1500,8 +1501,6 @@  (define_int_iterator VADDQ_N [VADDQ_N_S VADDQ_N_U])
 (define_int_iterator VADDVAQ [VADDVAQ_S VADDVAQ_U])
 (define_int_iterator VADDVQ_P [VADDVQ_P_U VADDVQ_P_S])
 (define_int_iterator VBRSRQ_N [VBRSRQ_N_U VBRSRQ_N_S])
-(define_int_iterator VCADDQ_ROT270 [VCADDQ_ROT270_S VCADDQ_ROT270_U])
-(define_int_iterator VCADDQ_ROT90 [VCADDQ_ROT90_U VCADDQ_ROT90_S])
 (define_int_iterator VCMPEQQ [VCMPEQQ_U VCMPEQQ_S])
 (define_int_iterator VCMPEQQ_N [VCMPEQQ_N_S VCMPEQQ_N_U])
 (define_int_iterator VCMPNEQ_N [VCMPNEQ_N_U VCMPNEQ_N_S])
diff --git a/gcc/config/arm/mve.md b/gcc/config/arm/mve.md
index b4c5a1e27c41e2270e05f9f4da1c055457a20ad3..516d0a3172e9cbb9a7fe0e9a1cd45ba5d935344b 100644
--- a/gcc/config/arm/mve.md
+++ b/gcc/config/arm/mve.md
@@ -962,34 +962,28 @@  (define_insn "mve_vbrsrq_n_<supf><mode>"
 ])
 
 ;;
-;; [vcaddq_rot270_s, vcaddq_rot270_u])
+;; [vcaddq, vcaddq_rot90, vcadd_rot180, vcadd_rot270])
 ;;
-(define_insn "mve_vcaddq_rot270_<supf><mode>"
+(define_insn "mve_vcaddq<mve_rot><mode>"
   [
    (set (match_operand:MVE_2 0 "s_register_operand" "<earlyclobber_32>")
 	(unspec:MVE_2 [(match_operand:MVE_2 1 "s_register_operand" "w")
 		       (match_operand:MVE_2 2 "s_register_operand" "w")]
-	 VCADDQ_ROT270))
+	 VCADD))
   ]
   "TARGET_HAVE_MVE"
-  "vcadd.i%#<V_sz_elem>	%q0, %q1, %q2, #270"
+  "vcadd.i%#<V_sz_elem>	%q0, %q1, %q2, #<rot>"
   [(set_attr "type" "mve_move")
 ])
 
-;;
-;; [vcaddq_rot90_u, vcaddq_rot90_s])
-;;
-(define_insn "mve_vcaddq_rot90_<supf><mode>"
-  [
-   (set (match_operand:MVE_2 0 "s_register_operand" "<earlyclobber_32>")
-	(unspec:MVE_2 [(match_operand:MVE_2 1 "s_register_operand" "w")
-		       (match_operand:MVE_2 2 "s_register_operand" "w")]
-	 VCADDQ_ROT90))
-  ]
-  "TARGET_HAVE_MVE"
-  "vcadd.i%#<V_sz_elem>	%q0, %q1, %q2, #90"
-  [(set_attr "type" "mve_move")
-])
+;; Auto vectorizer pattern for int vcadd
+(define_expand "cadd<rot><mode>3"
+  [(set (match_operand:MVE_2 0 "register_operand")
+	(unspec:MVE_2 [(match_operand:MVE_2 1 "register_operand")
+		       (match_operand:MVE_2 2 "register_operand")]
+	  VCADD))]
+  "TARGET_HAVE_MVE && !BYTES_BIG_ENDIAN"
+)
 
 ;;
 ;; [vcmpcsq_n_u])
@@ -2102,32 +2096,17 @@  (define_insn "mve_vbicq_n_<supf><mode>"
 ])
 
 ;;
-;; [vcaddq_rot270_f])
-;;
-(define_insn "mve_vcaddq_rot270_f<mode>"
-  [
-   (set (match_operand:MVE_0 0 "s_register_operand" "<earlyclobber_32>")
-	(unspec:MVE_0 [(match_operand:MVE_0 1 "s_register_operand" "w")
-		       (match_operand:MVE_0 2 "s_register_operand" "w")]
-	 VCADDQ_ROT270_F))
-  ]
-  "TARGET_HAVE_MVE && TARGET_HAVE_MVE_FLOAT"
-  "vcadd.f%#<V_sz_elem>	%q0, %q1, %q2, #270"
-  [(set_attr "type" "mve_move")
-])
-
-;;
-;; [vcaddq_rot90_f])
+;; [vcaddq, vcaddq_rot90, vcadd_rot180, vcadd_rot270])
 ;;
-(define_insn "mve_vcaddq_rot90_f<mode>"
+(define_insn "mve_vcaddq<mve_rot><mode>"
   [
    (set (match_operand:MVE_0 0 "s_register_operand" "<earlyclobber_32>")
 	(unspec:MVE_0 [(match_operand:MVE_0 1 "s_register_operand" "w")
 		       (match_operand:MVE_0 2 "s_register_operand" "w")]
-	 VCADDQ_ROT90_F))
+	 VCADD))
   ]
   "TARGET_HAVE_MVE && TARGET_HAVE_MVE_FLOAT"
-  "vcadd.f%#<V_sz_elem>	%q0, %q1, %q2, #90"
+  "vcadd.f%#<V_sz_elem>	%q0, %q1, %q2, #<rot>"
   [(set_attr "type" "mve_move")
 ])
 
diff --git a/gcc/config/arm/unspecs.md b/gcc/config/arm/unspecs.md
index ef64989600dcad642af54d7bf0250728a9fb7502..3f9ebe7b1753045e53044324cc7302f51d0eed21 100644
--- a/gcc/config/arm/unspecs.md
+++ b/gcc/config/arm/unspecs.md
@@ -598,8 +598,6 @@  (define_c_enum "unspec" [
   VADDVAQ_S
   VADDVQ_P_S
   VBRSRQ_N_S
-  VCADDQ_ROT270_S
-  VCADDQ_ROT90_S
   VCMPEQQ_S
   VCMPEQQ_N_S
   VCMPNEQ_N_S
@@ -641,8 +639,6 @@  (define_c_enum "unspec" [
   VADDVAQ_U
   VADDVQ_P_U
   VBRSRQ_N_U
-  VCADDQ_ROT270_U
-  VCADDQ_ROT90_U
   VCMPEQQ_U
   VCMPEQQ_N_U
   VCMPNEQ_N_U
@@ -709,8 +705,6 @@  (define_c_enum "unspec" [
   VABDQ_M_U
   VABDQ_F
   VADDQ_N_F
-  VCADDQ_ROT270_F
-  VCADDQ_ROT90_F
   VCMPEQQ_F
   VCMPEQQ_N_F
   VCMPGEQ_F
diff --git a/gcc/config/arm/vec-common.md b/gcc/config/arm/vec-common.md
index 2d0932b95a1615e94aa6572a593fc9849c9b36db..ce1ea960161164098a6512cf58843bb64251fb05 100644
--- a/gcc/config/arm/vec-common.md
+++ b/gcc/config/arm/vec-common.md
@@ -205,3 +205,13 @@  (define_expand "neg<mode>2"
 	(neg:VDQWH (match_operand:VDQWH 1 "s_register_operand" "")))]
   "ARM_HAVE_<MODE>_ARITH"
 )
+
+(define_expand "cadd<rot><mode>3"
+  [(set (match_operand:VF 0 "register_operand")
+	(unspec:VF [(match_operand:VF 1 "register_operand")
+		    (match_operand:VF 2 "register_operand")]
+		   VCADD))]
+  "(TARGET_COMPLEX || (TARGET_HAVE_MVE && TARGET_HAVE_MVE_FLOAT
+		      && ARM_HAVE_<MODE>_ARITH)) && !BYTES_BIG_ENDIAN"
+)
+