aarch64: Handle ASHIFTRT in patterns for shrn2

Similar to the low-half patterns, we want to match both ashiftrt and
lshiftrt with the truncate for SHRN2.  We reuse the SHIFTRT iterator
and the AARCH64_VALID_SHRN_OP check to help, but because we expand the
high-half patterns by their gen_* names we need to disambiguate all the
different trunc+shift combinations in the pattern name, which leads to a
slight renaming of the builtins.  The AARCH64_VALID_SHRN_OP check on the
expander and the define_insns ensures that no invalid combination ends
up getting matched.

Bootstrapped and tested on aarch64-none-linux-gnu and
aarch64_be-none-elf.

gcc/ChangeLog:

	* config/aarch64/aarch64-simd-builtins.def (shrn2_n): Rename builtins to...
	(ushrn2_n): ... This.
	(sqshrn2_n): Rename builtins to...
	(ssqshrn2_n): ... This.
	(uqshrn2_n): Rename builtins to...
	(uqushrn2_n): ... This.
	* config/aarch64/arm_neon.h (vqshrn_high_n_s16): Adjust for the above.
	(vqshrn_high_n_s32): Likewise.
	(vqshrn_high_n_s64): Likewise.
	(vqshrn_high_n_u16): Likewise.
	(vqshrn_high_n_u32): Likewise.
	(vqshrn_high_n_u64): Likewise.
	(vshrn_high_n_s16): Likewise.
	(vshrn_high_n_s32): Likewise.
	(vshrn_high_n_s64): Likewise.
	(vshrn_high_n_u16): Likewise.
	(vshrn_high_n_u32): Likewise.
	(vshrn_high_n_u64): Likewise.
	* config/aarch64/aarch64-simd.md (aarch64_<shrn_op>shrn2_n<mode>_insn_le):
	Rename to...
	(aarch64_<shrn_op><sra_op>shrn2_n<mode>_insn_le): ... This.
	Use SHIFTRT iterator and AARCH64_VALID_SHRN_OP check.
	(aarch64_<shrn_op>shrn2_n<mode>_insn_be): Rename to...
	(aarch64_<shrn_op><sra_op>shrn2_n<mode>_insn_be): ... This.
	Use SHIFTRT iterator and AARCH64_VALID_SHRN_OP check.
	(aarch64_<shrn_op>shrn2_n<mode>): Rename to...
	(aarch64_<shrn_op><sra_op>shrn2_n<mode>): ... This.
	Update expander for the above.

3 files changed, 31 insertions, 29 deletions

diff --git a/gcc/config/aarch64/aarch64-simd-builtins.def b/gcc/config/aarch64/aarch64-simd-builtins.def
index 01cd85d..e2b94ad 100644
--- a/gcc/config/aarch64/aarch64-simd-builtins.def
+++ b/gcc/config/aarch64/aarch64-simd-builtins.def
@@ -263,8 +263,8 @@
   BUILTIN_VQN (SHIFTIMM, shrn_n, 0, NONE)
   BUILTIN_VQN (USHIFTIMM, shrn_n, 0, NONE)
 
-  BUILTIN_VQN (SHIFT2IMM, shrn2_n, 0, NONE)
-  BUILTIN_VQN (USHIFT2IMM, shrn2_n, 0, NONE)
+  BUILTIN_VQN (SHIFT2IMM, ushrn2_n, 0, NONE)
+  BUILTIN_VQN (USHIFT2IMM, ushrn2_n, 0, NONE)
 
   BUILTIN_VQN (SHIFTIMM, rshrn_n, 0, NONE)
   BUILTIN_VQN (USHIFTIMM, rshrn_n, 0, NONE)
@@ -480,8 +480,8 @@
   BUILTIN_SD_HSDI (USHIFTIMM, uqrshrn_n, 0, NONE)
   BUILTIN_VQN (SHIFT2IMM_UUSS, sqshrun2_n, 0, NONE)
   BUILTIN_VQN (SHIFT2IMM_UUSS, sqrshrun2_n, 0, NONE)
-  BUILTIN_VQN (SHIFT2IMM, sqshrn2_n, 0, NONE)
-  BUILTIN_VQN (USHIFT2IMM, uqshrn2_n, 0, NONE)
+  BUILTIN_VQN (SHIFT2IMM, sqsshrn2_n, 0, NONE)
+  BUILTIN_VQN (USHIFT2IMM, uqushrn2_n, 0, NONE)
   BUILTIN_VQN (SHIFT2IMM, sqrshrn2_n, 0, NONE)
   BUILTIN_VQN (USHIFT2IMM, uqrshrn2_n, 0, NONE)
   /* Implemented by aarch64_<sur>s<lr>i_n<mode>.  */
diff --git a/gcc/config/aarch64/aarch64-simd.md b/gcc/config/aarch64/aarch64-simd.md
index b31c713..cd04cbd 100644
--- a/gcc/config/aarch64/aarch64-simd.md
+++ b/gcc/config/aarch64/aarch64-simd.md
@@ -6773,49 +6773,51 @@
   }
 )
 
-(define_insn "aarch64_<shrn_op>shrn2_n<mode>_insn_le"
+(define_insn "aarch64_<shrn_op><sra_op>shrn2_n<mode>_insn_le"
   [(set (match_operand:<VNARROWQ2> 0 "register_operand" "=w")
 	(vec_concat:<VNARROWQ2>
 	  (match_operand:<VNARROWQ> 1 "register_operand" "0")
 	  (ALL_TRUNC:<VNARROWQ>
-	    (<TRUNC_SHIFT>:VQN
+	    (SHIFTRT:VQN
 	      (match_operand:VQN 2 "register_operand" "w")
 	      (match_operand:VQN 3 "aarch64_simd_shift_imm_vec_<vn_mode>")))))]
-  "TARGET_SIMD && !BYTES_BIG_ENDIAN"
+  "TARGET_SIMD && !BYTES_BIG_ENDIAN
+   && AARCH64_VALID_SHRN_OP (<ALL_TRUNC:CODE>, <SHIFTRT:CODE>)"
   "<shrn_op>shrn2\t%<vn2>0.<V2ntype>, %<v>2.<Vtype>, %3"
   [(set_attr "type" "neon_shift_imm_narrow_q")]
 )
 
-(define_insn "aarch64_<shrn_op>shrn2_n<mode>_insn_be"
+(define_insn "aarch64_<shrn_op><sra_op>shrn2_n<mode>_insn_be"
   [(set (match_operand:<VNARROWQ2> 0 "register_operand" "=w")
 	(vec_concat:<VNARROWQ2>
 	  (ALL_TRUNC:<VNARROWQ>
-	    (<TRUNC_SHIFT>:VQN
+	    (SHIFTRT:VQN
 	      (match_operand:VQN 2 "register_operand" "w")
 	      (match_operand:VQN 3 "aarch64_simd_shift_imm_vec_<vn_mode>")))
 	  (match_operand:<VNARROWQ> 1 "register_operand" "0")))]
-  "TARGET_SIMD && BYTES_BIG_ENDIAN"
+  "TARGET_SIMD && BYTES_BIG_ENDIAN
+   && AARCH64_VALID_SHRN_OP (<ALL_TRUNC:CODE>, <SHIFTRT:CODE>)"
   "<shrn_op>shrn2\t%<vn2>0.<V2ntype>, %<v>2.<Vtype>, %3"
   [(set_attr "type" "neon_shift_imm_narrow_q")]
 )
 
-(define_expand "aarch64_<shrn_op>shrn2_n<mode>"
+(define_expand "aarch64_<shrn_op><sra_op>shrn2_n<mode>"
   [(match_operand:<VNARROWQ2> 0 "register_operand")
    (match_operand:<VNARROWQ> 1 "register_operand")
    (ALL_TRUNC:<VNARROWQ>
-     (match_operand:VQN 2 "register_operand"))
+     (SHIFTRT:VQN (match_operand:VQN 2 "register_operand")))
    (match_operand:SI 3 "aarch64_simd_shift_imm_offset_<vn_mode>")]
-  "TARGET_SIMD"
+  "TARGET_SIMD && AARCH64_VALID_SHRN_OP (<ALL_TRUNC:CODE>, <SHIFTRT:CODE>)"
   {
     operands[3] = aarch64_simd_gen_const_vector_dup (<MODE>mode,
 						 INTVAL (operands[3]));
 
     if (BYTES_BIG_ENDIAN)
-      emit_insn (gen_aarch64_<shrn_op>shrn2_n<mode>_insn_be (operands[0],
-				operands[1], operands[2], operands[3]));
+      emit_insn (gen_aarch64_<shrn_op><sra_op>shrn2_n<mode>_insn_be (
+		operands[0], operands[1], operands[2], operands[3]));
     else
-      emit_insn (gen_aarch64_<shrn_op>shrn2_n<mode>_insn_le (operands[0],
-				operands[1], operands[2], operands[3]));
+      emit_insn (gen_aarch64_<shrn_op><sra_op>shrn2_n<mode>_insn_le (
+		operands[0], operands[1], operands[2], operands[3]));
     DONE;
   }
 )
diff --git a/gcc/config/aarch64/arm_neon.h b/gcc/config/aarch64/arm_neon.h
index d350d9e..0ace1ee 100644
--- a/gcc/config/aarch64/arm_neon.h
+++ b/gcc/config/aarch64/arm_neon.h
@@ -5469,42 +5469,42 @@ __extension__ extern __inline int8x16_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vqshrn_high_n_s16 (int8x8_t __a, int16x8_t __b, const int __c)
 {
-  return __builtin_aarch64_sqshrn2_nv8hi (__a, __b, __c);
+  return __builtin_aarch64_sqsshrn2_nv8hi (__a, __b, __c);
 }
 
 __extension__ extern __inline int16x8_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vqshrn_high_n_s32 (int16x4_t __a, int32x4_t __b, const int __c)
 {
-  return __builtin_aarch64_sqshrn2_nv4si (__a, __b, __c);
+  return __builtin_aarch64_sqsshrn2_nv4si (__a, __b, __c);
 }
 
 __extension__ extern __inline int32x4_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vqshrn_high_n_s64 (int32x2_t __a, int64x2_t __b, const int __c)
 {
-  return __builtin_aarch64_sqshrn2_nv2di (__a, __b, __c);
+  return __builtin_aarch64_sqsshrn2_nv2di (__a, __b, __c);
 }
 
 __extension__ extern __inline uint8x16_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vqshrn_high_n_u16 (uint8x8_t __a, uint16x8_t __b, const int __c)
 {
-  return __builtin_aarch64_uqshrn2_nv8hi_uuus (__a, __b, __c);
+  return __builtin_aarch64_uqushrn2_nv8hi_uuus (__a, __b, __c);
 }
 
 __extension__ extern __inline uint16x8_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vqshrn_high_n_u32 (uint16x4_t __a, uint32x4_t __b, const int __c)
 {
-  return __builtin_aarch64_uqshrn2_nv4si_uuus (__a, __b, __c);
+  return __builtin_aarch64_uqushrn2_nv4si_uuus (__a, __b, __c);
 }
 
 __extension__ extern __inline uint32x4_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vqshrn_high_n_u64 (uint32x2_t __a, uint64x2_t __b, const int __c)
 {
-  return __builtin_aarch64_uqshrn2_nv2di_uuus (__a, __b, __c);
+  return __builtin_aarch64_uqushrn2_nv2di_uuus (__a, __b, __c);
 }
 
 __extension__ extern __inline uint8x16_t
@@ -5630,42 +5630,42 @@ __extension__ extern __inline int8x16_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vshrn_high_n_s16 (int8x8_t __a, int16x8_t __b, const int __c)
 {
-  return __builtin_aarch64_shrn2_nv8hi (__a, __b, __c);
+  return __builtin_aarch64_ushrn2_nv8hi (__a, __b, __c);
 }
 
 __extension__ extern __inline int16x8_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vshrn_high_n_s32 (int16x4_t __a, int32x4_t __b, const int __c)
 {
-  return __builtin_aarch64_shrn2_nv4si (__a, __b, __c);
+  return __builtin_aarch64_ushrn2_nv4si (__a, __b, __c);
 }
 
 __extension__ extern __inline int32x4_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vshrn_high_n_s64 (int32x2_t __a, int64x2_t __b, const int __c)
 {
-  return __builtin_aarch64_shrn2_nv2di (__a, __b, __c);
+  return __builtin_aarch64_ushrn2_nv2di (__a, __b, __c);
 }
 
 __extension__ extern __inline uint8x16_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vshrn_high_n_u16 (uint8x8_t __a, uint16x8_t __b, const int __c)
 {
-  return __builtin_aarch64_shrn2_nv8hi_uuus (__a, __b, __c);
+  return __builtin_aarch64_ushrn2_nv8hi_uuus (__a, __b, __c);
 }
 
 __extension__ extern __inline uint16x8_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vshrn_high_n_u32 (uint16x4_t __a, uint32x4_t __b, const int __c)
 {
-  return __builtin_aarch64_shrn2_nv4si_uuus (__a, __b, __c);
+  return __builtin_aarch64_ushrn2_nv4si_uuus (__a, __b, __c);
 }
 
 __extension__ extern __inline uint32x4_t
 __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
 vshrn_high_n_u64 (uint32x2_t __a, uint64x2_t __b, const int __c)
 {
-  return __builtin_aarch64_shrn2_nv2di_uuus (__a, __b, __c);
+  return __builtin_aarch64_ushrn2_nv2di_uuus (__a, __b, __c);
 }
 
 __extension__ extern __inline poly8x8_t