arm: Add vstN_lane_bf16 + vstNq_lane_bf16 intrisics

gcc/ChangeLog

2020-10-29  Andrea Corallo  <andrea.corallo@arm.com>

	* config/arm/arm_neon.h (vst2_lane_bf16, vst2q_lane_bf16)
	(vst3_lane_bf16, vst3q_lane_bf16, vst4_lane_bf16)
	(vst4q_lane_bf16): New intrinsics.
	* config/arm/arm_neon_builtins.def: Touch it for:
	__builtin_neon_vst2_lanev4bf, __builtin_neon_vst2_lanev8bf,
	__builtin_neon_vst3_lanev4bf, __builtin_neon_vst3_lanev8bf,
	__builtin_neon_vst4_lanev4bf,__builtin_neon_vst4_lanev8bf.

gcc/testsuite/ChangeLog

2020-10-29  Andrea Corallo  <andrea.corallo@arm.com>

	* gcc.target/aarch64/advsimd-intrinsics/vst2_lane_bf16_indices_1.c:
	Run it also for arm-*-*.
	* gcc.target/aarch64/advsimd-intrinsics/vst2q_lane_bf16_indices_1.c:
	Likewise.
	* gcc.target/aarch64/advsimd-intrinsics/vst3_lane_bf16_indices_1.c:
	Likewise.
	* gcc.target/aarch64/advsimd-intrinsics/vst3q_lane_bf16_indices_1.c:
	Likewise.
	* gcc.target/aarch64/advsimd-intrinsics/vst4_lane_bf16_indices_1.c:
	Likewise.
	* gcc.target/aarch64/advsimd-intrinsics/vst4q_lane_bf16_indices_1.c:
	Likewise.
	* gcc.target/arm/simd/vstn_lane_bf16_1.c: New test.

gcc/config/arm/arm_neon.h

@@ -19783,6 +19783,54 @@ vld4q_lane_bf16 (const bfloat16_t * __a, bfloat16x8x4_t __b, const int __c)
   return __rv.__i;
 }
 
+__extension__ extern __inline void
+__attribute__  ((__always_inline__, __gnu_inline__, __artificial__))
+vst2_lane_bf16 (bfloat16_t * __a, bfloat16x4x2_t __b, const int __c)
+{
+  union { bfloat16x4x2_t __i; __builtin_neon_ti __o; } __bu = { __b };
+  __builtin_neon_vst2_lanev4bf (__a, __bu.__o, __c);
+}
+
+__extension__ extern __inline void
+__attribute__  ((__always_inline__, __gnu_inline__, __artificial__))
+vst2q_lane_bf16 (bfloat16_t * __a, bfloat16x8x2_t __b, const int __c)
+{
+  union { bfloat16x8x2_t __i; __builtin_neon_oi __o; } __bu = { __b };
+  __builtin_neon_vst2_lanev8bf (__a, __bu.__o, __c);
+}
+
+__extension__ extern __inline void
+__attribute__  ((__always_inline__, __gnu_inline__, __artificial__))
+vst3_lane_bf16 (bfloat16_t * __a, bfloat16x4x3_t __b, const int __c)
+{
+  union { bfloat16x4x3_t __i; __builtin_neon_ei __o; } __bu = { __b };
+  __builtin_neon_vst3_lanev4bf (__a, __bu.__o, __c);
+}
+
+__extension__ extern __inline void
+__attribute__  ((__always_inline__, __gnu_inline__, __artificial__))
+vst3q_lane_bf16 (bfloat16_t * __a, bfloat16x8x3_t __b, const int __c)
+{
+  union { bfloat16x8x3_t __i; __builtin_neon_ci __o; } __bu = { __b };
+  __builtin_neon_vst3_lanev8bf (__a, __bu.__o, __c);
+}
+
+__extension__ extern __inline void
+__attribute__  ((__always_inline__, __gnu_inline__, __artificial__))
+vst4_lane_bf16 (bfloat16_t * __a, bfloat16x4x4_t __b, const int __c)
+{
+  union { bfloat16x4x4_t __i; __builtin_neon_oi __o; } __bu = { __b };
+  __builtin_neon_vst4_lanev4bf (__a, __bu.__o, __c);
+}
+
+__extension__ extern __inline void
+__attribute__  ((__always_inline__, __gnu_inline__, __artificial__))
+vst4q_lane_bf16 (bfloat16_t * __a, bfloat16x8x4_t __b, const int __c)
+{
+  union { bfloat16x8x4_t __i; __builtin_neon_xi __o; } __bu = { __b };
+  __builtin_neon_vst4_lanev8bf (__a, __bu.__o, __c);
+}
+
 #pragma GCC pop_options
 
 #ifdef __cplusplus

gcc/config/arm/arm_neon_builtins.def

@@ -329,8 +329,8 @@ VAR11 (LOAD1LANE, vld2_lane,
 VAR8 (LOAD1, vld2_dup, v8qi, v4hi, v4hf, v2si, v2sf, di, v4bf, v8bf)
 VAR13 (STORE1, vst2,
 	v8qi, v4hi, v4hf, v4bf, v2si, v2sf, di, v16qi, v8hi, v8hf, v8bf, v4si, v4sf)
-VAR9 (STORE1LANE, vst2_lane,
-	v8qi, v4hi, v4hf, v2si, v2sf, v8hi, v8hf, v4si, v4sf)
+VAR11 (STORE1LANE, vst2_lane,
+        v8qi, v4hi, v4hf, v2si, v2sf, v8hi, v8hf, v4si, v4sf, v4bf, v8bf)
 VAR13 (LOAD1, vld3,
 	v8qi, v4hi, v4hf, v2si, v2sf, di, v16qi, v8hi, v8hf, v4si, v4sf, v4bf, v8bf)
 VAR11 (LOAD1LANE, vld3_lane,
@@ -338,8 +338,8 @@ VAR11 (LOAD1LANE, vld3_lane,
 VAR8 (LOAD1, vld3_dup, v8qi, v4hi, v4hf, v2si, v2sf, di, v4bf, v8bf)
 VAR13 (STORE1, vst3,
 	v8qi, v4hi, v4hf, v4bf, v2si, v2sf, di, v16qi, v8hi, v8hf, v8bf, v4si, v4sf)
-VAR9 (STORE1LANE, vst3_lane,
-	v8qi, v4hi, v4hf, v2si, v2sf, v8hi, v8hf, v4si, v4sf)
+VAR11 (STORE1LANE, vst3_lane,
+	v8qi, v4hi, v4hf, v2si, v2sf, v8hi, v8hf, v4si, v4sf, v4bf, v8bf)
 VAR13 (LOAD1, vld4,
 	v8qi, v4hi, v4hf, v2si, v2sf, di, v16qi, v8hi, v8hf, v4si, v4sf, v4bf, v8bf)
 VAR11 (LOAD1LANE, vld4_lane,
@@ -347,8 +347,8 @@ VAR11 (LOAD1LANE, vld4_lane,
 VAR8 (LOAD1, vld4_dup, v8qi, v4hi, v4hf, v2si, v2sf, di, v4bf, v8bf)
 VAR13 (STORE1, vst4,
 	v8qi, v4hi, v4hf, v4bf, v2si, v2sf, di, v16qi, v8hi, v8hf, v8bf, v4si, v4sf)
-VAR9 (STORE1LANE, vst4_lane,
-	v8qi, v4hi, v4hf, v2si, v2sf, v8hi, v8hf, v4si, v4sf)
+VAR11 (STORE1LANE, vst4_lane,
+	v8qi, v4hi, v4hf, v2si, v2sf, v8hi, v8hf, v4si, v4sf, v4bf, v8bf)
 VAR2 (TERNOP, sdot, v8qi, v16qi)
 VAR2 (UTERNOP, udot, v8qi, v16qi)
 VAR2 (MAC_LANE, sdot_lane, v8qi, v16qi)

gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vst2_lane_bf16_indices_1.c

@@ -1,4 +1,4 @@
-/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-do compile } */
 /* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
 /* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
 /* { dg-add-options arm_v8_2a_bf16_neon }  */

gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vst2q_lane_bf16_indices_1.c

@@ -1,4 +1,4 @@
-/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-do compile } */
 /* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
 /* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
 /* { dg-add-options arm_v8_2a_bf16_neon }  */

gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vst3_lane_bf16_indices_1.c

@@ -1,4 +1,4 @@
-/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-do compile } */
 /* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
 /* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
 /* { dg-add-options arm_v8_2a_bf16_neon }  */

gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vst3q_lane_bf16_indices_1.c

@@ -1,4 +1,4 @@
-/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-do compile } */
 /* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
 /* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
 /* { dg-add-options arm_v8_2a_bf16_neon }  */

gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vst4_lane_bf16_indices_1.c

@@ -1,4 +1,4 @@
-/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-do compile } */
 /* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
 /* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
 /* { dg-add-options arm_v8_2a_bf16_neon }  */

gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vst4q_lane_bf16_indices_1.c

@@ -1,4 +1,4 @@
-/* { dg-do compile { target { aarch64*-*-* } } } */
+/* { dg-do compile } */
 /* { dg-skip-if "" { *-*-* } { "-fno-fat-lto-objects" } } */
 /* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
 /* { dg-add-options arm_v8_2a_bf16_neon }  */

gcc/testsuite/gcc.target/arm/simd/vstn_lane_bf16_1.c

@@ -0,0 +1,73 @@
+/* { dg-do assemble } */
+/* { dg-require-effective-target arm_v8_2a_bf16_neon_ok } */
+/* { dg-add-options arm_v8_2a_bf16_neon } */
+/* { dg-additional-options "-save-temps -O2 -mfloat-abi=hard" }  */
+/* { dg-final { check-function-bodies "**" "" } } */
+
+#include "arm_neon.h"
+
+/*
+**test_vst2_lane_bf16:
+**	vst2.16	{d0\[2\], d1\[2\]}, \[r0\]
+**	bx	lr
+*/
+void
+test_vst2_lane_bf16 (bfloat16_t *a, bfloat16x4x2_t b)
+{
+  return vst2_lane_bf16 (a, b, 2);
+}
+
+/*
+**test_vst2q_lane_bf16:
+**	vst2.16	{d0\[2\], d2\[2\]}, \[r0\]
+**	bx	lr
+*/
+void
+test_vst2q_lane_bf16 (bfloat16_t *a, bfloat16x8x2_t b)
+{
+  return vst2q_lane_bf16 (a, b, 2);
+}
+
+/*
+**test_vst3_lane_bf16:
+**	vst3.16	{d0\[2\], d1\[2\], d2\[2\]}, \[r0\]
+**	bx	lr
+*/
+void
+test_vst3_lane_bf16 (bfloat16_t *a, bfloat16x4x3_t b)
+{
+  return vst3_lane_bf16 (a, b, 2);
+}
+
+/*
+**test_vst3q_lane_bf16:
+**	vst3.16	{d0\[2\], d2\[2\], d4\[2\]}, \[r0\]
+**	bx	lr
+*/
+void
+test_vst3q_lane_bf16 (bfloat16_t *a, bfloat16x8x3_t b)
+{
+  return vst3q_lane_bf16 (a, b, 2);
+}
+
+/*
+**test_vst4_lane_bf16:
+**	vst4.16	{d0\[2\], d1\[2\], d2\[2\], d3\[2\]}, \[r0\]
+**	bx	lr
+*/
+void
+test_vst4_lane_bf16 (bfloat16_t *a, bfloat16x4x4_t b)
+{
+  return vst4_lane_bf16 (a, b, 2);
+}
+
+/*
+**test_vst4q_lane_bf16:
+**	vst4.16	{d0\[2\], d2\[2\], d4\[2\], d6\[2\]}, \[r0\]
+**	bx	lr
+*/
+void
+test_vst4q_lane_bf16 (bfloat16_t *a, bfloat16x8x4_t b)
+{
+  return vst4q_lane_bf16 (a, b, 2);
+}