aarch64: Use RTL builtins for [su]mlal_high_n intrinsics

Rewrite [su]mlal_high_n Neon intrinsics to use RTL builtins rather
than inline assembly code, allowing for better scheduling and
optimization.

gcc/ChangeLog:

2021-01-27  Jonathan Wright  <jonathan.wright@arm.com>

	* config/aarch64/aarch64-simd-builtins.def: Add [su]mlal_hi_n
	builtin generator macros.
	* config/aarch64/aarch64-simd.md (aarch64_<su>mlal_hi_n<mode>_insn):
	Define.
	(aarch64_<su>mlal_hi_n<mode>): Define.
	* config/aarch64/arm_neon.h (vmlal_high_n_s16): Use RTL builtin
	instead of inline asm.
	(vmlal_high_n_s32): Likewise.
	(vmlal_high_n_u16): Likewise.
	(vmlal_high_n_u32): Likewise.

gcc/config/aarch64/aarch64-simd-builtins.def

@@ -244,6 +244,10 @@
   BUILTIN_VQW (TERNOP, smlal_hi, 0, NONE)
   BUILTIN_VQW (TERNOPU, umlal_hi, 0, NONE)
 
+  /* Implemented by aarch64_<su>mlal_hi_n<mode>.  */
+  BUILTIN_VQ_HSI (TERNOP, smlal_hi_n, 0, NONE)
+  BUILTIN_VQ_HSI (TERNOPU, umlal_hi_n, 0, NONE)
+
   BUILTIN_VSQN_HSDI (UNOPUS, sqmovun, 0, NONE)
 
   /* Implemented by aarch64_sqxtun2<mode>.  */

gcc/config/aarch64/aarch64-simd.md

@@ -2006,6 +2006,35 @@
 }
 )
 
+(define_insn "aarch64_<su>mlal_hi_n<mode>_insn"
+  [(set (match_operand:<VWIDE> 0 "register_operand" "=w")
+        (plus:<VWIDE>
+          (mult:<VWIDE>
+              (ANY_EXTEND:<VWIDE> (vec_select:<VHALF>
+                 (match_operand:VQ_HSI 2 "register_operand" "w")
+                 (match_operand:VQ_HSI 3 "vect_par_cnst_hi_half" "")))
+              (ANY_EXTEND:<VWIDE> (vec_duplicate:<VCOND>
+	               (match_operand:<VEL> 4 "register_operand" "<h_con>"))))
+          (match_operand:<VWIDE> 1 "register_operand" "0")))]
+  "TARGET_SIMD"
+  "<su>mlal2\t%0.<Vwtype>, %2.<Vtype>, %4.<Vetype>[0]"
+  [(set_attr "type" "neon_mla_<Vetype>_long")]
+)
+
+(define_expand "aarch64_<su>mlal_hi_n<mode>"
+  [(match_operand:<VWIDE> 0 "register_operand")
+   (match_operand:<VWIDE> 1 "register_operand")
+   (ANY_EXTEND:<VWIDE>(match_operand:VQ_HSI 2 "register_operand"))
+   (match_operand:<VEL> 3 "register_operand")]
+  "TARGET_SIMD"
+{
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
+  emit_insn (gen_aarch64_<su>mlal_hi_n<mode>_insn (operands[0],
+             operands[1], operands[2], p, operands[3]));
+  DONE;
+}
+)
+
 (define_insn "*aarch64_<su>mlsl_lo<mode>"
   [(set (match_operand:<VWIDE> 0 "register_operand" "=w")
         (minus:<VWIDE>

gcc/config/aarch64/arm_neon.h

@@ -7268,48 +7268,28 @@ __extension__ extern __inline int32x4_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vmlal_high_n_s16 (int32x4_t __a, int16x8_t __b, int16_t __c)
 {
-  int32x4_t __result;
+  return __builtin_aarch64_smlal_hi_nv8hi (__a, __b, __c);
-  __asm__ ("smlal2 %0.4s,%2.8h,%3.h[0]"
-           : "=w"(__result)
-           : "0"(__a), "w"(__b), "x"(__c)
-           : /* No clobbers */);
-  return __result;
 }
 
 __extension__ extern __inline int64x2_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vmlal_high_n_s32 (int64x2_t __a, int32x4_t __b, int32_t __c)
 {
-  int64x2_t __result;
+  return __builtin_aarch64_smlal_hi_nv4si (__a, __b, __c);
-  __asm__ ("smlal2 %0.2d,%2.4s,%3.s[0]"
-           : "=w"(__result)
-           : "0"(__a), "w"(__b), "w"(__c)
-           : /* No clobbers */);
-  return __result;
 }
 
 __extension__ extern __inline uint32x4_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vmlal_high_n_u16 (uint32x4_t __a, uint16x8_t __b, uint16_t __c)
 {
-  uint32x4_t __result;
+  return __builtin_aarch64_umlal_hi_nv8hi_uuuu (__a, __b, __c);
-  __asm__ ("umlal2 %0.4s,%2.8h,%3.h[0]"
-           : "=w"(__result)
-           : "0"(__a), "w"(__b), "x"(__c)
-           : /* No clobbers */);
-  return __result;
 }
 
 __extension__ extern __inline uint64x2_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vmlal_high_n_u32 (uint64x2_t __a, uint32x4_t __b, uint32_t __c)
 {
-  uint64x2_t __result;
+  return __builtin_aarch64_umlal_hi_nv4si_uuuu (__a, __b, __c);
-  __asm__ ("umlal2 %0.2d,%2.4s,%3.s[0]"
-           : "=w"(__result)
-           : "0"(__a), "w"(__b), "w"(__c)
-           : /* No clobbers */);
-  return __result;
 }
 
 __extension__ extern __inline int16x8_t