[AArch64] Pass number of units to aarch64_simd_vect_par_cnst_half

This patch passes the number of units to aarch64_simd_vect_par_cnst_half,
which avoids a to_constant () once GET_MODE_NUNITS is variable.

2017-11-06  Richard Sandiford  <richard.sandiford@linaro.org>
	    Alan Hayward  <alan.hayward@arm.com>
	    David Sherwood  <david.sherwood@arm.com>

gcc/
	* config/aarch64/aarch64-protos.h (aarch64_simd_vect_par_cnst_half):
	Take the number of units too.
	* config/aarch64/aarch64.c (aarch64_simd_vect_par_cnst_half): Likewise.
	(aarch64_simd_check_vect_par_cnst_half): Update call accordingly,
	but check for a vector mode before rather than after the call.
	* config/aarch64/aarch64-simd.md (aarch64_split_simd_mov<mode>)
	(move_hi_quad_<mode>, vec_unpack<su>_hi_<mode>)
	(vec_unpack<su>_lo_<mode, vec_widen_<su>mult_lo_<mode>)
	(vec_widen_<su>mult_hi_<mode>, vec_unpacks_lo_<mode>)
	(vec_unpacks_hi_<mode>, aarch64_saddl2<mode>, aarch64_uaddl2<mode>)
	(aarch64_ssubl2<mode>, aarch64_usubl2<mode>, widen_ssum<mode>3)
	(widen_usum<mode>3, aarch64_saddw2<mode>, aarch64_uaddw2<mode>)
	(aarch64_ssubw2<mode>, aarch64_usubw2<mode>, aarch64_sqdmlal2<mode>)
	(aarch64_sqdmlsl2<mode>, aarch64_sqdmlal2_lane<mode>)
	(aarch64_sqdmlal2_laneq<mode>, aarch64_sqdmlsl2_lane<mode>)
	(aarch64_sqdmlsl2_laneq<mode>, aarch64_sqdmlal2_n<mode>)
	(aarch64_sqdmlsl2_n<mode>, aarch64_sqdmull2<mode>)
	(aarch64_sqdmull2_lane<mode>, aarch64_sqdmull2_laneq<mode>)
	(aarch64_sqdmull2_n<mode>): Update accordingly.

Reviewed-by: James Greenhalgh <james.greenhalgh@arm.com>

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>

From-SVN: r254468

4 files changed, 64 insertions, 39 deletions

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index ad78a25..5d11224 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -2,6 +2,30 @@
 	    Alan Hayward  <alan.hayward@arm.com>
 	    David Sherwood  <david.sherwood@arm.com>
 
+	* config/aarch64/aarch64-protos.h (aarch64_simd_vect_par_cnst_half):
+	Take the number of units too.
+	* config/aarch64/aarch64.c (aarch64_simd_vect_par_cnst_half): Likewise.
+	(aarch64_simd_check_vect_par_cnst_half): Update call accordingly,
+	but check for a vector mode before rather than after the call.
+	* config/aarch64/aarch64-simd.md (aarch64_split_simd_mov<mode>)
+	(move_hi_quad_<mode>, vec_unpack<su>_hi_<mode>)
+	(vec_unpack<su>_lo_<mode, vec_widen_<su>mult_lo_<mode>)
+	(vec_widen_<su>mult_hi_<mode>, vec_unpacks_lo_<mode>)
+	(vec_unpacks_hi_<mode>, aarch64_saddl2<mode>, aarch64_uaddl2<mode>)
+	(aarch64_ssubl2<mode>, aarch64_usubl2<mode>, widen_ssum<mode>3)
+	(widen_usum<mode>3, aarch64_saddw2<mode>, aarch64_uaddw2<mode>)
+	(aarch64_ssubw2<mode>, aarch64_usubw2<mode>, aarch64_sqdmlal2<mode>)
+	(aarch64_sqdmlsl2<mode>, aarch64_sqdmlal2_lane<mode>)
+	(aarch64_sqdmlal2_laneq<mode>, aarch64_sqdmlsl2_lane<mode>)
+	(aarch64_sqdmlsl2_laneq<mode>, aarch64_sqdmlal2_n<mode>)
+	(aarch64_sqdmlsl2_n<mode>, aarch64_sqdmull2<mode>)
+	(aarch64_sqdmull2_lane<mode>, aarch64_sqdmull2_laneq<mode>)
+	(aarch64_sqdmull2_n<mode>): Update accordingly.
+
+2017-11-06  Richard Sandiford  <richard.sandiford@linaro.org>
+	    Alan Hayward  <alan.hayward@arm.com>
+	    David Sherwood  <david.sherwood@arm.com>
+
 	* config/aarch64/aarch64-protos.h (aarch64_reverse_mask): Take
 	the number of units too.
 	* config/aarch64/aarch64.c (aarch64_reverse_mask): Likewise.
diff --git a/gcc/config/aarch64/aarch64-protos.h b/gcc/config/aarch64/aarch64-protos.h
index 3969115..4fdded7 100644
--- a/gcc/config/aarch64/aarch64-protos.h
+++ b/gcc/config/aarch64/aarch64-protos.h
@@ -391,7 +391,7 @@ const char *aarch64_output_move_struct (rtx *operands);
 rtx aarch64_return_addr (int, rtx);
 rtx aarch64_simd_gen_const_vector_dup (machine_mode, HOST_WIDE_INT);
 bool aarch64_simd_mem_operand_p (rtx);
-rtx aarch64_simd_vect_par_cnst_half (machine_mode, bool);
+rtx aarch64_simd_vect_par_cnst_half (machine_mode, int, bool);
 rtx aarch64_tls_get_addr (void);
 tree aarch64_fold_builtin (tree, int, tree *, bool);
 unsigned aarch64_dbx_register_number (unsigned);
diff --git a/gcc/config/aarch64/aarch64-simd.md b/gcc/config/aarch64/aarch64-simd.md
index 642f4b1..0699e56 100644
--- a/gcc/config/aarch64/aarch64-simd.md
+++ b/gcc/config/aarch64/aarch64-simd.md
@@ -253,8 +253,8 @@
       {
         rtx dst_low_part = gen_lowpart (<VHALF>mode, dst);
         rtx dst_high_part = gen_highpart (<VHALF>mode, dst);
-        rtx lo = aarch64_simd_vect_par_cnst_half (<MODE>mode, false);
-        rtx hi = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+	rtx lo = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, false);
+	rtx hi = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
 
         emit_insn
           (gen_aarch64_simd_mov_from_<mode>low (dst_low_part, src, lo));
@@ -1437,7 +1437,7 @@
   (match_operand:<VHALF> 1 "register_operand" "")]
  "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, false);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, false);
   if (BYTES_BIG_ENDIAN)
     emit_insn (gen_aarch64_simd_move_hi_quad_be_<mode> (operands[0],
 		    operands[1], p));
@@ -1521,7 +1521,7 @@
    (ANY_EXTEND:<VWIDE> (match_operand:VQW 1 "register_operand"))]
   "TARGET_SIMD"
   {
-    rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+    rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
     emit_insn (gen_aarch64_simd_vec_unpack<su>_hi_<mode> (operands[0],
 							  operands[1], p));
     DONE;
@@ -1533,7 +1533,7 @@
    (ANY_EXTEND:<VWIDE> (match_operand:VQW 1 "register_operand" ""))]
   "TARGET_SIMD"
   {
-    rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, false);
+    rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, false);
     emit_insn (gen_aarch64_simd_vec_unpack<su>_lo_<mode> (operands[0],
 							  operands[1], p));
     DONE;
@@ -1653,7 +1653,7 @@
    (ANY_EXTEND:<VWIDE> (match_operand:VQW 2 "register_operand" ""))]
  "TARGET_SIMD"
  {
-   rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, false);
+   rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, false);
    emit_insn (gen_aarch64_simd_vec_<su>mult_lo_<mode> (operands[0],
 						       operands[1],
 						       operands[2], p));
@@ -1680,7 +1680,7 @@
    (ANY_EXTEND:<VWIDE> (match_operand:VQW 2 "register_operand" ""))]
  "TARGET_SIMD"
  {
-   rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+   rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
    emit_insn (gen_aarch64_simd_vec_<su>mult_hi_<mode> (operands[0],
 						       operands[1],
 						       operands[2], p));
@@ -2084,7 +2084,7 @@
    (match_operand:VQ_HSF 1 "register_operand" "")]
   "TARGET_SIMD"
   {
-    rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, false);
+    rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, false);
     emit_insn (gen_aarch64_simd_vec_unpacks_lo_<mode> (operands[0],
 						       operands[1], p));
     DONE;
@@ -2107,7 +2107,7 @@
    (match_operand:VQ_HSF 1 "register_operand" "")]
   "TARGET_SIMD"
   {
-    rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+    rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
     emit_insn (gen_aarch64_simd_vec_unpacks_lo_<mode> (operands[0],
 						       operands[1], p));
     DONE;
@@ -3028,7 +3028,7 @@
    (match_operand:VQW 2 "register_operand" "w")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_saddl<mode>_hi_internal (operands[0], operands[1],
                                                   operands[2], p));
   DONE;
@@ -3040,7 +3040,7 @@
    (match_operand:VQW 2 "register_operand" "w")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_uaddl<mode>_hi_internal (operands[0], operands[1],
                                                   operands[2], p));
   DONE;
@@ -3052,7 +3052,7 @@
    (match_operand:VQW 2 "register_operand" "w")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_ssubl<mode>_hi_internal (operands[0], operands[1],
 						operands[2], p));
   DONE;
@@ -3064,7 +3064,7 @@
    (match_operand:VQW 2 "register_operand" "w")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_usubl<mode>_hi_internal (operands[0], operands[1],
 						operands[2], p));
   DONE;
@@ -3090,7 +3090,7 @@
 		      (match_operand:<VDBLW> 2 "register_operand" "")))]
   "TARGET_SIMD"
   {
-    rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, false);
+    rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, false);
     rtx temp = gen_reg_rtx (GET_MODE (operands[0]));
 
     emit_insn (gen_aarch64_saddw<mode>_internal (temp, operands[2],
@@ -3118,7 +3118,7 @@
 		      (match_operand:<VDBLW> 2 "register_operand" "")))]
   "TARGET_SIMD"
   {
-    rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, false);
+    rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, false);
     rtx temp = gen_reg_rtx (GET_MODE (operands[0]));
 
     emit_insn (gen_aarch64_uaddw<mode>_internal (temp, operands[2],
@@ -3179,7 +3179,7 @@
    (match_operand:VQW 2 "register_operand" "w")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_saddw2<mode>_internal (operands[0], operands[1],
 						operands[2], p));
   DONE;
@@ -3191,7 +3191,7 @@
    (match_operand:VQW 2 "register_operand" "w")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_uaddw2<mode>_internal (operands[0], operands[1],
 						operands[2], p));
   DONE;
@@ -3204,7 +3204,7 @@
    (match_operand:VQW 2 "register_operand" "w")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_ssubw2<mode>_internal (operands[0], operands[1],
 						operands[2], p));
   DONE;
@@ -3216,7 +3216,7 @@
    (match_operand:VQW 2 "register_operand" "w")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_usubw2<mode>_internal (operands[0], operands[1],
 						operands[2], p));
   DONE;
@@ -3736,7 +3736,7 @@
    (match_operand:VQ_HSI 3 "register_operand" "w")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_sqdmlal2<mode>_internal (operands[0], operands[1],
 						  operands[2], operands[3], p));
   DONE;
@@ -3749,7 +3749,7 @@
    (match_operand:VQ_HSI 3 "register_operand" "w")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_sqdmlsl2<mode>_internal (operands[0], operands[1],
 						  operands[2], operands[3], p));
   DONE;
@@ -3817,7 +3817,7 @@
    (match_operand:SI 4 "immediate_operand" "i")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_sqdmlal2_lane<mode>_internal (operands[0], operands[1],
 						       operands[2], operands[3],
 						       operands[4], p));
@@ -3832,7 +3832,7 @@
    (match_operand:SI 4 "immediate_operand" "i")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_sqdmlal2_laneq<mode>_internal (operands[0], operands[1],
 						       operands[2], operands[3],
 						       operands[4], p));
@@ -3847,7 +3847,7 @@
    (match_operand:SI 4 "immediate_operand" "i")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_sqdmlsl2_lane<mode>_internal (operands[0], operands[1],
 						       operands[2], operands[3],
 						       operands[4], p));
@@ -3862,7 +3862,7 @@
    (match_operand:SI 4 "immediate_operand" "i")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_sqdmlsl2_laneq<mode>_internal (operands[0], operands[1],
 						       operands[2], operands[3],
 						       operands[4], p));
@@ -3895,7 +3895,7 @@
    (match_operand:<VEL> 3 "register_operand" "w")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_sqdmlal2_n<mode>_internal (operands[0], operands[1],
 						    operands[2], operands[3],
 						    p));
@@ -3909,7 +3909,7 @@
    (match_operand:<VEL> 3 "register_operand" "w")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_sqdmlsl2_n<mode>_internal (operands[0], operands[1],
 						    operands[2], operands[3],
 						    p));
@@ -4063,7 +4063,7 @@
    (match_operand:VQ_HSI 2 "register_operand" "w")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_sqdmull2<mode>_internal (operands[0], operands[1],
 						  operands[2], p));
   DONE;
@@ -4124,7 +4124,7 @@
    (match_operand:SI 3 "immediate_operand" "i")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_sqdmull2_lane<mode>_internal (operands[0], operands[1],
 						       operands[2], operands[3],
 						       p));
@@ -4138,7 +4138,7 @@
    (match_operand:SI 3 "immediate_operand" "i")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_sqdmull2_laneq<mode>_internal (operands[0], operands[1],
 						       operands[2], operands[3],
 						       p));
@@ -4171,7 +4171,7 @@
    (match_operand:<VEL> 2 "register_operand" "w")]
   "TARGET_SIMD"
 {
-  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, true);
+  rtx p = aarch64_simd_vect_par_cnst_half (<MODE>mode, <nunits>, true);
   emit_insn (gen_aarch64_sqdmull2_n<mode>_internal (operands[0], operands[1],
 						    operands[2], p));
   DONE;
diff --git a/gcc/config/aarch64/aarch64.c b/gcc/config/aarch64/aarch64.c
index b3ce7f6..d209f81 100644
--- a/gcc/config/aarch64/aarch64.c
+++ b/gcc/config/aarch64/aarch64.c
@@ -11757,12 +11757,12 @@ Architecture    3   2   1   0           3   2   1   0
 
 Low Mask:         { 2, 3 }                { 0, 1 }
 High Mask:        { 0, 1 }                { 2, 3 }
-*/
+
+   MODE Is the mode of the vector and NUNITS is the number of units in it.  */
 
 rtx
-aarch64_simd_vect_par_cnst_half (machine_mode mode, bool high)
+aarch64_simd_vect_par_cnst_half (machine_mode mode, int nunits, bool high)
 {
-  int nunits = GET_MODE_NUNITS (mode);
   rtvec v = rtvec_alloc (nunits / 2);
   int high_base = nunits / 2;
   int low_base = 0;
@@ -11791,14 +11791,15 @@ bool
 aarch64_simd_check_vect_par_cnst_half (rtx op, machine_mode mode,
 				       bool high)
 {
-  rtx ideal = aarch64_simd_vect_par_cnst_half (mode, high);
+  if (!VECTOR_MODE_P (mode))
+    return false;
+
+  rtx ideal = aarch64_simd_vect_par_cnst_half (mode, GET_MODE_NUNITS (mode),
+					       high);
   HOST_WIDE_INT count_op = XVECLEN (op, 0);
   HOST_WIDE_INT count_ideal = XVECLEN (ideal, 0);
   int i = 0;
 
-  if (!VECTOR_MODE_P (mode))
-    return false;
-
   if (count_op != count_ideal)
     return false;