[gcc(refs/users/meissner/heads/work046)] Add XXSPLTI32DX support for V4SI/V4SF.

[gcc(refs/users/meissner/heads/work046)] Add XXSPLTI32DX support for V4SI/V4SF.

[Michael Meissner]

https://gcc.gnu.org/g:9685d713ac6a6a1aebe3c1db7902215983e09366

commit 9685d713ac6a6a1aebe3c1db7902215983e09366
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Mon Apr 12 17:07:58 2021 -0400

    Add XXSPLTI32DX support for V4SI/V4SF.
    
    V4SImode and V4SFmode vectors that are initialized as:
    
            (const_vector [const1, const2, const1, const2])
    
    can be initialized with 2 sets of XXSPLTI32DX instructions.
    
    gcc/
    2021-04-12  Michael Meissner  <meissner@linux.ibm.com>
    
            * config/rs6000/rs6000.c (xxsplti32dx_constant_p): Add support for
            V4SImode and V4SFmode.
            * config/rs6000/vsx.md (XXSPLTI32DX): Extend iterator to include
            V4SImode and V4SFmode.
            (xxsplti32dx_<mode>): Properly handle optimizing the second word
            being all 1's.

Diff:
---
 gcc/config/rs6000/rs6000.c | 36 ++++++++++++++++++++++++++++++++++++
 gcc/config/rs6000/vsx.md   | 13 +++++++------
 2 files changed, 43 insertions(+), 6 deletions(-)

diff --git a/gcc/config/rs6000/rs6000.c b/gcc/config/rs6000/rs6000.c
index 7d125d6b65a..db0cbe1edad 100644
--- a/gcc/config/rs6000/rs6000.c
+++ b/gcc/config/rs6000/rs6000.c
@@ -6657,6 +6657,42 @@ xxsplti32dx_constant_p (rtx op,
       mode = GET_MODE_INNER (mode);
     }
 
+  else if (mode == V4SImode || mode == V4SFmode)
+    {
+      /* For V4SI/V4SF, the XXSPLTI32DX instruction pair can represent vectors
+	 where the two even elements are equal and the two odd elements are
+	 equal.  */
+      if (GET_CODE (op) != CONST_VECTOR)
+	return false;
+
+      rtx op0 = CONST_VECTOR_ELT (op, 0);
+      if (!rtx_equal_p (op0, CONST_VECTOR_ELT (op, 2)))
+	return false;
+
+      rtx op1 = CONST_VECTOR_ELT (op, 1);
+      if (!rtx_equal_p (op1, CONST_VECTOR_ELT (op, 3)))
+	return false;
+
+      if (TARGET_XXSPLTIW && rtx_equal_p (op0, op1))
+	return false;
+
+      long op0_value;
+      long op1_value;
+      if (mode == V4SImode)
+	{
+	  op0_value = INTVAL (op0);
+	  op1_value = INTVAL (op1);
+	}
+      else
+	{
+	  op0_value = rs6000_const_f32_to_i32 (op0);
+	  op1_value = rs6000_const_f32_to_i32 (op1);
+	}
+
+      *constant_ptr = (op0_value << 32) | (op1_value & 0xffffffff);
+      return true;
+    }
+
   if (GET_MODE (element) != mode)
     return false;
 
diff --git a/gcc/config/rs6000/vsx.md b/gcc/config/rs6000/vsx.md
index ec2c148fb4d..a5bdcc6ead8 100644
--- a/gcc/config/rs6000/vsx.md
+++ b/gcc/config/rs6000/vsx.md
@@ -6348,8 +6348,9 @@
   DONE;
 })
 
-;; XXSPLTI32DX used to create 64-bit constants
-(define_mode_iterator XXSPLTI32DX [SF DF V2DF V2DI])
+;; XXSPLTI32DX used to create 64-bit constants or 32-bit vector constants where
+;; the even elements match and the odd elements match.
+(define_mode_iterator XXSPLTI32DX [SF DF V4SF V4SI V2DF V2DI])
 
 (define_insn_and_split "*xxsplti32dx_<mode>"
   [(set (match_operand:XXSPLTI32DX 0 "vsx_register_operand" "=wa")
@@ -6371,11 +6372,11 @@
     gcc_unreachable ();
 
   HOST_WIDE_INT high = value >> 32;
-  HOST_WIDE_INT low = value & 0xffffffff;
+  HOST_WIDE_INT low = ((value & 0xffffffff) ^ 0x80000000) - 0x80000000;
 
-  /* If the low bits are 0/-1, initialize that word first.  This way we can
-     use a smaller XXSPLTIB instruction instead the first XXSPLTI32DX.  */
-  if (low == 0 || low == -1)
+  /* If the low bits are 0 or all 1s, initialize that word first.  This way we
+     can use a smaller XXSPLTIB instruction instead the first XXSPLTI32DX.  */
+  if (low == 0 || low ==  -1)
     {
       operands[2] = const1_rtx;
       operands[3] = GEN_INT (low);