[AArch64] Fix vqtb[lx][234] on big-endian

2015-11-06  Christophe Lyon  <christophe.lyon@linaro.org>

	gcc/
	* config/aarch64/aarch64-simd-builtins.def: Update builtins
	tables: add tbl3v16qi, qtbl[34]*, tbx4v16qi, qtbx[34]*.
	* config/aarch64/aarch64-simd.md (aarch64_tbl3v8qi): Rename to...
	(aarch64_tbl3<mode>) ... this, which supports v16qi too.
	(aarch64_tbx4v8qi): Rename to...
	aarch64_tbx4<mode>): ... this.
	(aarch64_qtbl3<mode>): New pattern.
	(aarch64_qtbx3<mode>): New pattern.
	(aarch64_qtbl4<mode>): New pattern.
	(aarch64_qtbx4<mode>): New pattern.
	* config/aarch64/arm_neon.h (vqtbl2_s8, vqtbl2_u8, vqtbl2_p8)
	(vqtbl2q_s8, vqtbl2q_u8, vqtbl2q_p8, vqtbl3_s8, vqtbl3_u8)
	(vqtbl3_p8, vqtbl3q_s8, vqtbl3q_u8, vqtbl3q_p8, vqtbl4_s8)
	(vqtbl4_u8, vqtbl4_p8, vqtbl4q_s8, vqtbl4q_u8, vqtbl4q_p8)
	(vqtbx2_s8, vqtbx2_u8, vqtbx2_p8, vqtbx2q_s8, vqtbx2q_u8)
	(vqtbx2q_p8, vqtbx3_s8, vqtbx3_u8, vqtbx3_p8, vqtbx3q_s8)
	(vqtbx3q_u8, vqtbx3q_p8, vqtbx4_s8, vqtbx4_u8, vqtbx4_p8)
	(vqtbx4q_s8, vqtbx4q_u8, vqtbx4q_p8): Rewrite using builtin
	functions.

	gcc/testsuite/
	* gcc.target/aarch64/advsimd-intrinsics/vqtbX.c: New test.

From-SVN: r229886

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index 6756e356cf95b961dc74211b96491d4980ceb43b..3fe6742545bfb844fd151ff74e231e4e5ddb3072 100644 (file)
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,25 @@
+2015-11-06  Christophe Lyon  <christophe.lyon@linaro.org>
+
+       * config/aarch64/aarch64-simd-builtins.def: Update builtins
+       tables: add tbl3v16qi, qtbl[34]*, tbx4v16qi, qtbx[34]*.
+       * config/aarch64/aarch64-simd.md (aarch64_tbl3v8qi): Rename to...
+       (aarch64_tbl3<mode>) ... this, which supports v16qi too.
+       (aarch64_tbx4v8qi): Rename to...
+       aarch64_tbx4<mode>): ... this.
+       (aarch64_qtbl3<mode>): New pattern.
+       (aarch64_qtbx3<mode>): New pattern.
+       (aarch64_qtbl4<mode>): New pattern.
+       (aarch64_qtbx4<mode>): New pattern.
+       * config/aarch64/arm_neon.h (vqtbl2_s8, vqtbl2_u8, vqtbl2_p8)
+       (vqtbl2q_s8, vqtbl2q_u8, vqtbl2q_p8, vqtbl3_s8, vqtbl3_u8)
+       (vqtbl3_p8, vqtbl3q_s8, vqtbl3q_u8, vqtbl3q_p8, vqtbl4_s8)
+       (vqtbl4_u8, vqtbl4_p8, vqtbl4q_s8, vqtbl4q_u8, vqtbl4q_p8)
+       (vqtbx2_s8, vqtbx2_u8, vqtbx2_p8, vqtbx2q_s8, vqtbx2q_u8)
+       (vqtbx2q_p8, vqtbx3_s8, vqtbx3_u8, vqtbx3_p8, vqtbx3q_s8)
+       (vqtbx3q_u8, vqtbx3q_p8, vqtbx4_s8, vqtbx4_u8, vqtbx4_p8)
+       (vqtbx4q_s8, vqtbx4q_u8, vqtbx4q_p8): Rewrite using builtin
+       functions.
+
 2015-11-06  Mike Stump  <mikestump@comcast.net>
 
        PR debug/66728

diff --git a/gcc/config/aarch64/aarch64-simd-builtins.def b/gcc/config/aarch64/aarch64-simd-builtins.def
index 04dac6f84db2393ef5057b7ddb04b7ed026c8a72..aad66b134ef740e8d5372a850ec2a7601454bc61 100644 (file)
--- a/gcc/config/aarch64/aarch64-simd-builtins.def
+++ b/gcc/config/aarch64/aarch64-simd-builtins.def
@@ -408,8 +408,26 @@
   VAR1 (BINOPP, crypto_pmull, 0, di)
   VAR1 (BINOPP, crypto_pmull, 0, v2di)
 
-  /* Implemented by aarch64_tbl3v8qi.  */
+  /* Implemented by aarch64_tbl3<mode>.  */
   VAR1 (BINOP, tbl3, 0, v8qi)
+  VAR1 (BINOP, tbl3, 0, v16qi)
 
-  /* Implemented by aarch64_tbx4v8qi.  */
+  /* Implemented by aarch64_qtbl3<mode>.  */
+  VAR1 (BINOP, qtbl3, 0, v8qi)
+  VAR1 (BINOP, qtbl3, 0, v16qi)
+
+  /* Implemented by aarch64_qtbl4<mode>.  */
+  VAR1 (BINOP, qtbl4, 0, v8qi)
+  VAR1 (BINOP, qtbl4, 0, v16qi)
+
+  /* Implemented by aarch64_tbx4<mode>.  */
   VAR1 (TERNOP, tbx4, 0, v8qi)
+  VAR1 (TERNOP, tbx4, 0, v16qi)
+
+  /* Implemented by aarch64_qtbx3<mode>.  */
+  VAR1 (TERNOP, qtbx3, 0, v8qi)
+  VAR1 (TERNOP, qtbx3, 0, v16qi)
+
+  /* Implemented by aarch64_qtbx4<mode>.  */
+  VAR1 (TERNOP, qtbx4, 0, v8qi)
+  VAR1 (TERNOP, qtbx4, 0, v16qi)

diff --git a/gcc/config/aarch64/aarch64-simd.md b/gcc/config/aarch64/aarch64-simd.md
index 55974e611181fcf176ad2405962e843057d82cd3..0fb813421283e9ab5fc2bf3cf2e63ccfc4a96c03 100644 (file)
--- a/gcc/config/aarch64/aarch64-simd.md
+++ b/gcc/config/aarch64/aarch64-simd.md
@@ -4816,24 +4816,70 @@
   [(set_attr "type" "neon_tbl2_q")]
 )
 
-(define_insn "aarch64_tbl3v8qi"
-  [(set (match_operand:V8QI 0 "register_operand" "=w")
-       (unspec:V8QI [(match_operand:OI 1 "register_operand" "w")
-                     (match_operand:V8QI 2 "register_operand" "w")]
+(define_insn "aarch64_tbl3<mode>"
+  [(set (match_operand:VB 0 "register_operand" "=w")
+       (unspec:VB [(match_operand:OI 1 "register_operand" "w")
+                     (match_operand:VB 2 "register_operand" "w")]
                      UNSPEC_TBL))]
   "TARGET_SIMD"
-  "tbl\\t%S0.8b, {%S1.16b - %T1.16b}, %S2.8b"
+  "tbl\\t%S0.<Vbtype>, {%S1.16b - %T1.16b}, %S2.<Vbtype>"
   [(set_attr "type" "neon_tbl3")]
 )
 
-(define_insn "aarch64_tbx4v8qi"
-  [(set (match_operand:V8QI 0 "register_operand" "=w")
-       (unspec:V8QI [(match_operand:V8QI 1 "register_operand" "0")
+(define_insn "aarch64_tbx4<mode>"
+  [(set (match_operand:VB 0 "register_operand" "=w")
+       (unspec:VB [(match_operand:VB 1 "register_operand" "0")
                      (match_operand:OI 2 "register_operand" "w")
-                     (match_operand:V8QI 3 "register_operand" "w")]
+                     (match_operand:VB 3 "register_operand" "w")]
+                     UNSPEC_TBX))]
+  "TARGET_SIMD"
+  "tbx\\t%S0.<Vbtype>, {%S2.16b - %T2.16b}, %S3.<Vbtype>"
+  [(set_attr "type" "neon_tbl4")]
+)
+
+;; Three source registers.
+
+(define_insn "aarch64_qtbl3<mode>"
+  [(set (match_operand:VB 0 "register_operand" "=w")
+       (unspec:VB [(match_operand:CI 1 "register_operand" "w")
+                     (match_operand:VB 2 "register_operand" "w")]
+                     UNSPEC_TBL))]
+  "TARGET_SIMD"
+  "tbl\\t%S0.<Vbtype>, {%S1.16b - %U1.16b}, %S2.<Vbtype>"
+  [(set_attr "type" "neon_tbl3")]
+)
+
+(define_insn "aarch64_qtbx3<mode>"
+  [(set (match_operand:VB 0 "register_operand" "=w")
+       (unspec:VB [(match_operand:VB 1 "register_operand" "0")
+                     (match_operand:CI 2 "register_operand" "w")
+                     (match_operand:VB 3 "register_operand" "w")]
+                     UNSPEC_TBX))]
+  "TARGET_SIMD"
+  "tbx\\t%S0.<Vbtype>, {%S2.16b - %U2.16b}, %S3.<Vbtype>"
+  [(set_attr "type" "neon_tbl3")]
+)
+
+;; Four source registers.
+
+(define_insn "aarch64_qtbl4<mode>"
+  [(set (match_operand:VB 0 "register_operand" "=w")
+       (unspec:VB [(match_operand:XI 1 "register_operand" "w")
+                     (match_operand:VB 2 "register_operand" "w")]
+                     UNSPEC_TBL))]
+  "TARGET_SIMD"
+  "tbl\\t%S0.<Vbtype>, {%S1.16b - %V1.16b}, %S2.<Vbtype>"
+  [(set_attr "type" "neon_tbl4")]
+)
+
+(define_insn "aarch64_qtbx4<mode>"
+  [(set (match_operand:VB 0 "register_operand" "=w")
+       (unspec:VB [(match_operand:VB 1 "register_operand" "0")
+                     (match_operand:XI 2 "register_operand" "w")
+                     (match_operand:VB 3 "register_operand" "w")]
                      UNSPEC_TBX))]
   "TARGET_SIMD"
-  "tbx\\t%S0.8b, {%S2.16b - %T2.16b}, %S3.8b"
+  "tbx\\t%S0.<Vbtype>, {%S2.16b - %V2.16b}, %S3.<Vbtype>"
   [(set_attr "type" "neon_tbl4")]
 )
 

diff --git a/gcc/config/aarch64/arm_neon.h b/gcc/config/aarch64/arm_neon.h
index 265c26679089f5a1b30a41360687fb1e3c2a8e50..7f69043d8a759cbb5b08ccb0eb98f110b9856a10 100644 (file)
--- a/gcc/config/aarch64/arm_neon.h
+++ b/gcc/config/aarch64/arm_neon.h
@@ -10622,224 +10622,6 @@ vqtbl1q_u8 (uint8x16_t a, uint8x16_t b)
   return result;
 }
 
-__extension__ static __inline int8x8_t __attribute__ ((__always_inline__))
-vqtbl2_s8 (int8x16x2_t tab, uint8x8_t idx)
-{
-  int8x8_t result;
-  __asm__ ("ld1 {v16.16b, v17.16b}, %1\n\t"
-          "tbl %0.8b, {v16.16b, v17.16b}, %2.8b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17");
-  return result;
-}
-
-__extension__ static __inline uint8x8_t __attribute__ ((__always_inline__))
-vqtbl2_u8 (uint8x16x2_t tab, uint8x8_t idx)
-{
-  uint8x8_t result;
-  __asm__ ("ld1 {v16.16b, v17.16b}, %1\n\t"
-          "tbl %0.8b, {v16.16b, v17.16b}, %2.8b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17");
-  return result;
-}
-
-__extension__ static __inline poly8x8_t __attribute__ ((__always_inline__))
-vqtbl2_p8 (poly8x16x2_t tab, uint8x8_t idx)
-{
-  poly8x8_t result;
-  __asm__ ("ld1 {v16.16b, v17.16b}, %1\n\t"
-          "tbl %0.8b, {v16.16b, v17.16b}, %2.8b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17");
-  return result;
-}
-
-__extension__ static __inline int8x16_t __attribute__ ((__always_inline__))
-vqtbl2q_s8 (int8x16x2_t tab, uint8x16_t idx)
-{
-  int8x16_t result;
-  __asm__ ("ld1 {v16.16b, v17.16b}, %1\n\t"
-          "tbl %0.16b, {v16.16b, v17.16b}, %2.16b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17");
-  return result;
-}
-
-__extension__ static __inline uint8x16_t __attribute__ ((__always_inline__))
-vqtbl2q_u8 (uint8x16x2_t tab, uint8x16_t idx)
-{
-  uint8x16_t result;
-  __asm__ ("ld1 {v16.16b, v17.16b}, %1\n\t"
-          "tbl %0.16b, {v16.16b, v17.16b}, %2.16b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17");
-  return result;
-}
-
-__extension__ static __inline poly8x16_t __attribute__ ((__always_inline__))
-vqtbl2q_p8 (poly8x16x2_t tab, uint8x16_t idx)
-{
-  poly8x16_t result;
-  __asm__ ("ld1 {v16.16b, v17.16b}, %1\n\t"
-          "tbl %0.16b, {v16.16b, v17.16b}, %2.16b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17");
-  return result;
-}
-
-__extension__ static __inline int8x8_t __attribute__ ((__always_inline__))
-vqtbl3_s8 (int8x16x3_t tab, uint8x8_t idx)
-{
-  int8x8_t result;
-  __asm__ ("ld1 {v16.16b - v18.16b}, %1\n\t"
-          "tbl %0.8b, {v16.16b - v18.16b}, %2.8b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18");
-  return result;
-}
-
-__extension__ static __inline uint8x8_t __attribute__ ((__always_inline__))
-vqtbl3_u8 (uint8x16x3_t tab, uint8x8_t idx)
-{
-  uint8x8_t result;
-  __asm__ ("ld1 {v16.16b - v18.16b}, %1\n\t"
-          "tbl %0.8b, {v16.16b - v18.16b}, %2.8b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18");
-  return result;
-}
-
-__extension__ static __inline poly8x8_t __attribute__ ((__always_inline__))
-vqtbl3_p8 (poly8x16x3_t tab, uint8x8_t idx)
-{
-  poly8x8_t result;
-  __asm__ ("ld1 {v16.16b - v18.16b}, %1\n\t"
-          "tbl %0.8b, {v16.16b - v18.16b}, %2.8b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18");
-  return result;
-}
-
-__extension__ static __inline int8x16_t __attribute__ ((__always_inline__))
-vqtbl3q_s8 (int8x16x3_t tab, uint8x16_t idx)
-{
-  int8x16_t result;
-  __asm__ ("ld1 {v16.16b - v18.16b}, %1\n\t"
-          "tbl %0.16b, {v16.16b - v18.16b}, %2.16b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18");
-  return result;
-}
-
-__extension__ static __inline uint8x16_t __attribute__ ((__always_inline__))
-vqtbl3q_u8 (uint8x16x3_t tab, uint8x16_t idx)
-{
-  uint8x16_t result;
-  __asm__ ("ld1 {v16.16b - v18.16b}, %1\n\t"
-          "tbl %0.16b, {v16.16b - v18.16b}, %2.16b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18");
-  return result;
-}
-
-__extension__ static __inline poly8x16_t __attribute__ ((__always_inline__))
-vqtbl3q_p8 (poly8x16x3_t tab, uint8x16_t idx)
-{
-  poly8x16_t result;
-  __asm__ ("ld1 {v16.16b - v18.16b}, %1\n\t"
-          "tbl %0.16b, {v16.16b - v18.16b}, %2.16b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18");
-  return result;
-}
-
-__extension__ static __inline int8x8_t __attribute__ ((__always_inline__))
-vqtbl4_s8 (int8x16x4_t tab, uint8x8_t idx)
-{
-  int8x8_t result;
-  __asm__ ("ld1 {v16.16b - v19.16b}, %1\n\t"
-          "tbl %0.8b, {v16.16b - v19.16b}, %2.8b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18", "v19");
-  return result;
-}
-
-__extension__ static __inline uint8x8_t __attribute__ ((__always_inline__))
-vqtbl4_u8 (uint8x16x4_t tab, uint8x8_t idx)
-{
-  uint8x8_t result;
-  __asm__ ("ld1 {v16.16b - v19.16b}, %1\n\t"
-          "tbl %0.8b, {v16.16b - v19.16b}, %2.8b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18", "v19");
-  return result;
-}
-
-__extension__ static __inline poly8x8_t __attribute__ ((__always_inline__))
-vqtbl4_p8 (poly8x16x4_t tab, uint8x8_t idx)
-{
-  poly8x8_t result;
-  __asm__ ("ld1 {v16.16b - v19.16b}, %1\n\t"
-          "tbl %0.8b, {v16.16b - v19.16b}, %2.8b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18", "v19");
-  return result;
-}
-
-
-__extension__ static __inline int8x16_t __attribute__ ((__always_inline__))
-vqtbl4q_s8 (int8x16x4_t tab, uint8x16_t idx)
-{
-  int8x16_t result;
-  __asm__ ("ld1 {v16.16b - v19.16b}, %1\n\t"
-          "tbl %0.16b, {v16.16b - v19.16b}, %2.16b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18", "v19");
-  return result;
-}
-
-__extension__ static __inline uint8x16_t __attribute__ ((__always_inline__))
-vqtbl4q_u8 (uint8x16x4_t tab, uint8x16_t idx)
-{
-  uint8x16_t result;
-  __asm__ ("ld1 {v16.16b - v19.16b}, %1\n\t"
-          "tbl %0.16b, {v16.16b - v19.16b}, %2.16b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18", "v19");
-  return result;
-}
-
-__extension__ static __inline poly8x16_t __attribute__ ((__always_inline__))
-vqtbl4q_p8 (poly8x16x4_t tab, uint8x16_t idx)
-{
-  poly8x16_t result;
-  __asm__ ("ld1 {v16.16b - v19.16b}, %1\n\t"
-          "tbl %0.16b, {v16.16b - v19.16b}, %2.16b\n\t"
-          :"=w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18", "v19");
-  return result;
-}
-
-
 __extension__ static __inline int8x8_t __attribute__ ((__always_inline__))
 vqtbx1_s8 (int8x8_t r, int8x16_t tab, uint8x8_t idx)
 {
@@ -10906,227 +10688,6 @@ vqtbx1q_p8 (poly8x16_t r, poly8x16_t tab, uint8x16_t idx)
   return result;
 }
 
-__extension__ static __inline int8x8_t __attribute__ ((__always_inline__))
-vqtbx2_s8 (int8x8_t r, int8x16x2_t tab, uint8x8_t idx)
-{
-  int8x8_t result = r;
-  __asm__ ("ld1 {v16.16b, v17.16b}, %1\n\t"
-          "tbx %0.8b, {v16.16b, v17.16b}, %2.8b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17");
-  return result;
-}
-
-__extension__ static __inline uint8x8_t __attribute__ ((__always_inline__))
-vqtbx2_u8 (uint8x8_t r, uint8x16x2_t tab, uint8x8_t idx)
-{
-  uint8x8_t result = r;
-  __asm__ ("ld1 {v16.16b, v17.16b}, %1\n\t"
-          "tbx %0.8b, {v16.16b, v17.16b}, %2.8b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17");
-  return result;
-}
-
-__extension__ static __inline poly8x8_t __attribute__ ((__always_inline__))
-vqtbx2_p8 (poly8x8_t r, poly8x16x2_t tab, uint8x8_t idx)
-{
-  poly8x8_t result = r;
-  __asm__ ("ld1 {v16.16b, v17.16b}, %1\n\t"
-          "tbx %0.8b, {v16.16b, v17.16b}, %2.8b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17");
-  return result;
-}
-
-
-__extension__ static __inline int8x16_t __attribute__ ((__always_inline__))
-vqtbx2q_s8 (int8x16_t r, int8x16x2_t tab, uint8x16_t idx)
-{
-  int8x16_t result = r;
-  __asm__ ("ld1 {v16.16b, v17.16b}, %1\n\t"
-          "tbx %0.16b, {v16.16b, v17.16b}, %2.16b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17");
-  return result;
-}
-
-__extension__ static __inline uint8x16_t __attribute__ ((__always_inline__))
-vqtbx2q_u8 (uint8x16_t r, uint8x16x2_t tab, uint8x16_t idx)
-{
-  uint8x16_t result = r;
-  __asm__ ("ld1 {v16.16b, v17.16b}, %1\n\t"
-          "tbx %0.16b, {v16.16b, v17.16b}, %2.16b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17");
-  return result;
-}
-
-__extension__ static __inline poly8x16_t __attribute__ ((__always_inline__))
-vqtbx2q_p8 (poly8x16_t r, poly8x16x2_t tab, uint8x16_t idx)
-{
-  poly8x16_t result = r;
-  __asm__ ("ld1 {v16.16b, v17.16b}, %1\n\t"
-          "tbx %0.16b, {v16.16b, v17.16b}, %2.16b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17");
-  return result;
-}
-
-
-__extension__ static __inline int8x8_t __attribute__ ((__always_inline__))
-vqtbx3_s8 (int8x8_t r, int8x16x3_t tab, uint8x8_t idx)
-{
-  int8x8_t result = r;
-  __asm__ ("ld1 {v16.16b - v18.16b}, %1\n\t"
-          "tbx %0.8b, {v16.16b - v18.16b}, %2.8b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18");
-  return result;
-}
-
-__extension__ static __inline uint8x8_t __attribute__ ((__always_inline__))
-vqtbx3_u8 (uint8x8_t r, uint8x16x3_t tab, uint8x8_t idx)
-{
-  uint8x8_t result = r;
-  __asm__ ("ld1 {v16.16b - v18.16b}, %1\n\t"
-          "tbx %0.8b, {v16.16b - v18.16b}, %2.8b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18");
-  return result;
-}
-
-__extension__ static __inline poly8x8_t __attribute__ ((__always_inline__))
-vqtbx3_p8 (poly8x8_t r, poly8x16x3_t tab, uint8x8_t idx)
-{
-  poly8x8_t result = r;
-  __asm__ ("ld1 {v16.16b - v18.16b}, %1\n\t"
-          "tbx %0.8b, {v16.16b - v18.16b}, %2.8b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18");
-  return result;
-}
-
-
-__extension__ static __inline int8x16_t __attribute__ ((__always_inline__))
-vqtbx3q_s8 (int8x16_t r, int8x16x3_t tab, uint8x16_t idx)
-{
-  int8x16_t result = r;
-  __asm__ ("ld1 {v16.16b - v18.16b}, %1\n\t"
-          "tbx %0.16b, {v16.16b - v18.16b}, %2.16b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18");
-  return result;
-}
-
-__extension__ static __inline uint8x16_t __attribute__ ((__always_inline__))
-vqtbx3q_u8 (uint8x16_t r, uint8x16x3_t tab, uint8x16_t idx)
-{
-  uint8x16_t result = r;
-  __asm__ ("ld1 {v16.16b - v18.16b}, %1\n\t"
-          "tbx %0.16b, {v16.16b - v18.16b}, %2.16b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18");
-  return result;
-}
-
-__extension__ static __inline poly8x16_t __attribute__ ((__always_inline__))
-vqtbx3q_p8 (poly8x16_t r, poly8x16x3_t tab, uint8x16_t idx)
-{
-  poly8x16_t result = r;
-  __asm__ ("ld1 {v16.16b - v18.16b}, %1\n\t"
-          "tbx %0.16b, {v16.16b - v18.16b}, %2.16b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18");
-  return result;
-}
-
-
-__extension__ static __inline int8x8_t __attribute__ ((__always_inline__))
-vqtbx4_s8 (int8x8_t r, int8x16x4_t tab, uint8x8_t idx)
-{
-  int8x8_t result = r;
-  __asm__ ("ld1 {v16.16b - v19.16b}, %1\n\t"
-          "tbx %0.8b, {v16.16b - v19.16b}, %2.8b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18", "v19");
-  return result;
-}
-
-__extension__ static __inline uint8x8_t __attribute__ ((__always_inline__))
-vqtbx4_u8 (uint8x8_t r, uint8x16x4_t tab, uint8x8_t idx)
-{
-  uint8x8_t result = r;
-  __asm__ ("ld1 {v16.16b - v19.16b}, %1\n\t"
-          "tbx %0.8b, {v16.16b - v19.16b}, %2.8b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18", "v19");
-  return result;
-}
-
-__extension__ static __inline poly8x8_t __attribute__ ((__always_inline__))
-vqtbx4_p8 (poly8x8_t r, poly8x16x4_t tab, uint8x8_t idx)
-{
-  poly8x8_t result = r;
-  __asm__ ("ld1 {v16.16b - v19.16b}, %1\n\t"
-          "tbx %0.8b, {v16.16b - v19.16b}, %2.8b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18", "v19");
-  return result;
-}
-
-
-__extension__ static __inline int8x16_t __attribute__ ((__always_inline__))
-vqtbx4q_s8 (int8x16_t r, int8x16x4_t tab, uint8x16_t idx)
-{
-  int8x16_t result = r;
-  __asm__ ("ld1 {v16.16b - v19.16b}, %1\n\t"
-          "tbx %0.16b, {v16.16b - v19.16b}, %2.16b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18", "v19");
-  return result;
-}
-
-__extension__ static __inline uint8x16_t __attribute__ ((__always_inline__))
-vqtbx4q_u8 (uint8x16_t r, uint8x16x4_t tab, uint8x16_t idx)
-{
-  uint8x16_t result = r;
-  __asm__ ("ld1 {v16.16b - v19.16b}, %1\n\t"
-          "tbx %0.16b, {v16.16b - v19.16b}, %2.16b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18", "v19");
-  return result;
-}
-
-__extension__ static __inline poly8x16_t __attribute__ ((__always_inline__))
-vqtbx4q_p8 (poly8x16_t r, poly8x16x4_t tab, uint8x16_t idx)
-{
-  poly8x16_t result = r;
-  __asm__ ("ld1 {v16.16b - v19.16b}, %1\n\t"
-          "tbx %0.16b, {v16.16b - v19.16b}, %2.16b\n\t"
-          :"+w"(result)
-          :"Q"(tab),"w"(idx)
-          :"memory", "v16", "v17", "v18", "v19");
-  return result;
-}
-
 /* V7 legacy table intrinsics.  */
 
 __extension__ static __inline int8x8_t __attribute__ ((__always_inline__))
@@ -20714,6 +20275,389 @@ vqsubd_u64 (uint64_t __a, uint64_t __b)
   return __builtin_aarch64_uqsubdi_uuu (__a, __b);
 }
 
+/* vqtbl2 */
+
+__extension__ static __inline int8x8_t __attribute__ ((__always_inline__))
+vqtbl2_s8 (int8x16x2_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_oi __o;
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, tab.val[1], 1);
+  return __builtin_aarch64_tbl3v8qi (__o, (int8x8_t)idx);
+}
+
+__extension__ static __inline uint8x8_t __attribute__ ((__always_inline__))
+vqtbl2_u8 (uint8x16x2_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_oi __o;
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  return (uint8x8_t)__builtin_aarch64_tbl3v8qi (__o, (int8x8_t)idx);
+}
+
+__extension__ static __inline poly8x8_t __attribute__ ((__always_inline__))
+vqtbl2_p8 (poly8x16x2_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_oi __o;
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  return (poly8x8_t)__builtin_aarch64_tbl3v8qi (__o, (int8x8_t)idx);
+}
+
+__extension__ static __inline int8x16_t __attribute__ ((__always_inline__))
+vqtbl2q_s8 (int8x16x2_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_oi __o;
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  return __builtin_aarch64_tbl3v16qi (__o, (int8x16_t)idx);
+}
+
+__extension__ static __inline uint8x16_t __attribute__ ((__always_inline__))
+vqtbl2q_u8 (uint8x16x2_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_oi __o;
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  return (uint8x16_t)__builtin_aarch64_tbl3v16qi (__o, (int8x16_t)idx);
+}
+
+__extension__ static __inline poly8x16_t __attribute__ ((__always_inline__))
+vqtbl2q_p8 (poly8x16x2_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_oi __o;
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  return (poly8x16_t)__builtin_aarch64_tbl3v16qi (__o, (int8x16_t)idx);
+}
+
+/* vqtbl3 */
+
+__extension__ static __inline int8x8_t __attribute__ ((__always_inline__))
+vqtbl3_s8 (int8x16x3_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_ci __o;
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[2], 2);
+  return __builtin_aarch64_qtbl3v8qi (__o, (int8x8_t)idx);
+}
+
+__extension__ static __inline uint8x8_t __attribute__ ((__always_inline__))
+vqtbl3_u8 (uint8x16x3_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_ci __o;
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[2], 2);
+  return (uint8x8_t)__builtin_aarch64_qtbl3v8qi (__o, (int8x8_t)idx);
+}
+
+__extension__ static __inline poly8x8_t __attribute__ ((__always_inline__))
+vqtbl3_p8 (poly8x16x3_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_ci __o;
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[2], 2);
+  return (poly8x8_t)__builtin_aarch64_qtbl3v8qi (__o, (int8x8_t)idx);
+}
+
+__extension__ static __inline int8x16_t __attribute__ ((__always_inline__))
+vqtbl3q_s8 (int8x16x3_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_ci __o;
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[2], 2);
+  return __builtin_aarch64_qtbl3v16qi (__o, (int8x16_t)idx);
+}
+
+__extension__ static __inline uint8x16_t __attribute__ ((__always_inline__))
+vqtbl3q_u8 (uint8x16x3_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_ci __o;
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[2], 2);
+  return (uint8x16_t)__builtin_aarch64_qtbl3v16qi (__o, (int8x16_t)idx);
+}
+
+__extension__ static __inline poly8x16_t __attribute__ ((__always_inline__))
+vqtbl3q_p8 (poly8x16x3_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_ci __o;
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[2], 2);
+  return (poly8x16_t)__builtin_aarch64_qtbl3v16qi (__o, (int8x16_t)idx);
+}
+
+/* vqtbl4 */
+
+__extension__ static __inline int8x8_t __attribute__ ((__always_inline__))
+vqtbl4_s8 (int8x16x4_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_xi __o;
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[2], 2);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[3], 3);
+  return __builtin_aarch64_qtbl4v8qi (__o, (int8x8_t)idx);
+}
+
+__extension__ static __inline uint8x8_t __attribute__ ((__always_inline__))
+vqtbl4_u8 (uint8x16x4_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_xi __o;
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[2], 2);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[3], 3);
+  return (uint8x8_t)__builtin_aarch64_qtbl4v8qi (__o, (int8x8_t)idx);
+}
+
+__extension__ static __inline poly8x8_t __attribute__ ((__always_inline__))
+vqtbl4_p8 (poly8x16x4_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_xi __o;
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[2], 2);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[3], 3);
+  return (poly8x8_t)__builtin_aarch64_qtbl4v8qi (__o, (int8x8_t)idx);
+}
+
+__extension__ static __inline int8x16_t __attribute__ ((__always_inline__))
+vqtbl4q_s8 (int8x16x4_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_xi __o;
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[2], 2);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[3], 3);
+  return __builtin_aarch64_qtbl4v16qi (__o, (int8x16_t)idx);
+}
+
+__extension__ static __inline uint8x16_t __attribute__ ((__always_inline__))
+vqtbl4q_u8 (uint8x16x4_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_xi __o;
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[2], 2);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[3], 3);
+  return (uint8x16_t)__builtin_aarch64_qtbl4v16qi (__o, (int8x16_t)idx);
+}
+
+__extension__ static __inline poly8x16_t __attribute__ ((__always_inline__))
+vqtbl4q_p8 (poly8x16x4_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_xi __o;
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[2], 2);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[3], 3);
+  return (poly8x16_t)__builtin_aarch64_qtbl4v16qi (__o, (int8x16_t)idx);
+}
+
+
+/* vqtbx2 */
+__extension__ static __inline int8x8_t __attribute__ ((__always_inline__))
+vqtbx2_s8 (int8x8_t r, int8x16x2_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_oi __o;
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, tab.val[1], 1);
+  return __builtin_aarch64_tbx4v8qi (r, __o, (int8x8_t)idx);
+}
+
+__extension__ static __inline uint8x8_t __attribute__ ((__always_inline__))
+vqtbx2_u8 (uint8x8_t r, uint8x16x2_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_oi __o;
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  return (uint8x8_t)__builtin_aarch64_tbx4v8qi ((int8x8_t)r, __o,
+                                               (int8x8_t)idx);
+}
+
+__extension__ static __inline poly8x8_t __attribute__ ((__always_inline__))
+vqtbx2_p8 (poly8x8_t r, poly8x16x2_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_oi __o;
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  return (poly8x8_t)__builtin_aarch64_tbx4v8qi ((int8x8_t)r, __o,
+                                               (int8x8_t)idx);
+}
+
+__extension__ static __inline int8x16_t __attribute__ ((__always_inline__))
+vqtbx2q_s8 (int8x16_t r, int8x16x2_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_oi __o;
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, tab.val[1], 1);
+  return __builtin_aarch64_tbx4v16qi (r, __o, (int8x16_t)idx);
+}
+
+__extension__ static __inline uint8x16_t __attribute__ ((__always_inline__))
+vqtbx2q_u8 (uint8x16_t r, uint8x16x2_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_oi __o;
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  return (uint8x16_t)__builtin_aarch64_tbx4v16qi ((int8x16_t)r, __o,
+                                                 (int8x16_t)idx);
+}
+
+__extension__ static __inline poly8x16_t __attribute__ ((__always_inline__))
+vqtbx2q_p8 (poly8x16_t r, poly8x16x2_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_oi __o;
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregoiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  return (poly8x16_t)__builtin_aarch64_tbx4v16qi ((int8x16_t)r, __o,
+                                                 (int8x16_t)idx);
+}
+
+/* vqtbx3 */
+__extension__ static __inline int8x8_t __attribute__ ((__always_inline__))
+vqtbx3_s8 (int8x8_t r, int8x16x3_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_ci __o;
+  __o = __builtin_aarch64_set_qregciv16qi (__o, tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, tab.val[2], 2);
+  return __builtin_aarch64_qtbx3v8qi (r, __o, (int8x8_t)idx);
+}
+
+__extension__ static __inline uint8x8_t __attribute__ ((__always_inline__))
+vqtbx3_u8 (uint8x8_t r, uint8x16x3_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_ci __o;
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[2], 2);
+  return (uint8x8_t)__builtin_aarch64_qtbx3v8qi ((int8x8_t)r, __o,
+                                                (int8x8_t)idx);
+}
+
+__extension__ static __inline poly8x8_t __attribute__ ((__always_inline__))
+vqtbx3_p8 (poly8x8_t r, poly8x16x3_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_ci __o;
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[2], 2);
+  return (poly8x8_t)__builtin_aarch64_qtbx3v8qi ((int8x8_t)r, __o,
+                                                (int8x8_t)idx);
+}
+
+__extension__ static __inline int8x16_t __attribute__ ((__always_inline__))
+vqtbx3q_s8 (int8x16_t r, int8x16x3_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_ci __o;
+  __o = __builtin_aarch64_set_qregciv16qi (__o, tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, tab.val[2], 2);
+  return __builtin_aarch64_qtbx3v16qi (r, __o, (int8x16_t)idx);
+}
+
+__extension__ static __inline uint8x16_t __attribute__ ((__always_inline__))
+vqtbx3q_u8 (uint8x16_t r, uint8x16x3_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_ci __o;
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[2], 2);
+  return (uint8x16_t)__builtin_aarch64_qtbx3v16qi ((int8x16_t)r, __o,
+                                                  (int8x16_t)idx);
+}
+
+__extension__ static __inline poly8x16_t __attribute__ ((__always_inline__))
+vqtbx3q_p8 (poly8x16_t r, poly8x16x3_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_ci __o;
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregciv16qi (__o, (int8x16_t)tab.val[2], 2);
+  return (poly8x16_t)__builtin_aarch64_qtbx3v16qi ((int8x16_t)r, __o,
+                                                  (int8x16_t)idx);
+}
+
+/* vqtbx4 */
+
+__extension__ static __inline int8x8_t __attribute__ ((__always_inline__))
+vqtbx4_s8 (int8x8_t r, int8x16x4_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_xi __o;
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, tab.val[2], 2);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, tab.val[3], 3);
+  return __builtin_aarch64_qtbx4v8qi (r, __o, (int8x8_t)idx);
+}
+
+__extension__ static __inline uint8x8_t __attribute__ ((__always_inline__))
+vqtbx4_u8 (uint8x8_t r, uint8x16x4_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_xi __o;
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[2], 2);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[3], 3);
+  return (uint8x8_t)__builtin_aarch64_qtbx4v8qi ((int8x8_t)r, __o,
+                                                (int8x8_t)idx);
+}
+
+__extension__ static __inline poly8x8_t __attribute__ ((__always_inline__))
+vqtbx4_p8 (poly8x8_t r, poly8x16x4_t tab, uint8x8_t idx)
+{
+  __builtin_aarch64_simd_xi __o;
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[2], 2);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[3], 3);
+  return (poly8x8_t)__builtin_aarch64_qtbx4v8qi ((int8x8_t)r, __o,
+                                                (int8x8_t)idx);
+}
+
+__extension__ static __inline int8x16_t __attribute__ ((__always_inline__))
+vqtbx4q_s8 (int8x16_t r, int8x16x4_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_xi __o;
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, tab.val[2], 2);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, tab.val[3], 3);
+  return __builtin_aarch64_qtbx4v16qi (r, __o, (int8x16_t)idx);
+}
+
+__extension__ static __inline uint8x16_t __attribute__ ((__always_inline__))
+vqtbx4q_u8 (uint8x16_t r, uint8x16x4_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_xi __o;
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[2], 2);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[3], 3);
+  return (uint8x16_t)__builtin_aarch64_qtbx4v16qi ((int8x16_t)r, __o,
+                                                  (int8x16_t)idx);
+}
+
+__extension__ static __inline poly8x16_t __attribute__ ((__always_inline__))
+vqtbx4q_p8 (poly8x16_t r, poly8x16x4_t tab, uint8x16_t idx)
+{
+  __builtin_aarch64_simd_xi __o;
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[0], 0);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[1], 1);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[2], 2);
+  __o = __builtin_aarch64_set_qregxiv16qi (__o, (int8x16_t)tab.val[3], 3);
+  return (poly8x16_t)__builtin_aarch64_qtbx4v16qi ((int8x16_t)r, __o,
+                                                  (int8x16_t)idx);
+}
+
 /* vrbit  */
 
 __extension__ static __inline poly8x8_t __attribute__ ((__always_inline__))

diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index 9a0d1aa22f8e5b9bda20e5a2d6ffee9fe06fc02c..564b45a9d8309a8e854306e78b87e1c424fc88f1 100644 (file)
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -1,3 +1,7 @@
+2015-11-06  Christophe Lyon  <christophe.lyon@linaro.org>
+
+       * gcc.target/aarch64/advsimd-intrinsics/vqtbX.c: New test.
+
 2015-11-06  David Malcolm  <dmalcolm@redhat.com>
 
        * gcc.dg/plugin/diagnostic-test-show-locus-bw.c: New file.
@@ -17,7 +21,7 @@
 
        * gfortran.dg/goacc/combined_loop.f90: XFAIL.
 
-2015-11-07  Jan Hubicka  <hubicka@ucw.cz>
+2015-11-06  Jan Hubicka  <hubicka@ucw.cz>
 
        PR ipa/68057
        PR ipa/68220