From a46bca36b7b3a8a7e15b04225fb2b4f9b1bed62c Mon Sep 17 00:00:00 2001
From: Roger Sayle <roger@nextmovesoftware.com>
Date: Sun, 7 Aug 2022 08:49:48 +0100
Subject: [PATCH] Allow any immediate constant in *cmp<dwi>_doubleword splitter
 on x86_64.

This patch tweaks i386.md's *cmp<dwi>_doubleword splitter's predicate to
allow general_operand, not just x86_64_hilo_general_operand, to improve
code generation.  As a general rule, i386.md's _doubleword splitters should
be post-reload splitters that require integer immediate operands to be
x86_64_hilo_int_operand, so that each part is a valid word mode immediate
constant.  As an exception to this rule, doubleword patterns that must be
split before reload, because they require additional scratch registers,
can use take advantage of this ability to create new pseudos, to accept
any immediate constant, and call force_reg on the high and/or low parts
if they are not suitable immediate operands in word mode.

The benefit is shown in the new cmpti3.c test case below.

__int128 x;
int foo()
{
    __int128 t = 0x1234567890abcdefLL;
    return x == t;
}

where GCC with -O2 currently generates:

        movabsq $1311768467294899695, %rax
        xorl    %edx, %edx
        xorq    x(%rip), %rax
        xorq    x+8(%rip), %rdx
        orq     %rdx, %rax
        sete    %al
        movzbl  %al, %eax
        ret

but with this patch now generates:

        movabsq $1311768467294899695, %rax
        xorq    x(%rip), %rax
        orq     x+8(%rip), %rax
        sete    %al
        movzbl  %al, %eax
        ret

2022-08-07  Roger Sayle  <roger@nextmovesoftware.com>

gcc/ChangeLog
	* config/i386/i386.md (*cmp<dwi>_doubleword): Change predicate
	for x86_64_hilo_general_operand to general operand.  Call
	force_reg on parts that are not x86_64_immediate_operand.

gcc/testsuite/ChangeLog
	* gcc.target/i386/cmpti1.c: New test case.
	* gcc.target/i386/cmpti2.c: Likewise.
	* gcc.target/i386/cmpti3.c: Likewise.
---
 gcc/config/i386/i386.md                | 16 +++++++++++++---
 gcc/testsuite/gcc.target/i386/cmpti1.c |  8 ++++++++
 gcc/testsuite/gcc.target/i386/cmpti2.c | 12 ++++++++++++
 gcc/testsuite/gcc.target/i386/cmpti3.c | 13 +++++++++++++
 4 files changed, 46 insertions(+), 3 deletions(-)
 create mode 100644 gcc/testsuite/gcc.target/i386/cmpti1.c
 create mode 100644 gcc/testsuite/gcc.target/i386/cmpti2.c
 create mode 100644 gcc/testsuite/gcc.target/i386/cmpti3.c

diff --git a/gcc/config/i386/i386.md b/gcc/config/i386/i386.md
index 298e4b303486..fd30c573c270 100644
--- a/gcc/config/i386/i386.md
+++ b/gcc/config/i386/i386.md
@@ -1510,7 +1510,7 @@
 (define_insn_and_split "*cmp<dwi>_doubleword"
   [(set (reg:CCZ FLAGS_REG)
 	(compare:CCZ (match_operand:<DWI> 0 "nonimmediate_operand")
-		     (match_operand:<DWI> 1 "x86_64_hilo_general_operand")))]
+		     (match_operand:<DWI> 1 "general_operand")))]
   "ix86_pre_reload_split ()"
   "#"
   "&& 1"
@@ -1544,7 +1544,12 @@
   else if (operands[0] == constm1_rtx)
     emit_insn (gen_one_cmpl<mode>2 (operands[4], operands[1]));
   else
-    emit_insn (gen_xor<mode>3 (operands[4], operands[0], operands[1]));
+    {
+      if (CONST_SCALAR_INT_P (operands[1])
+	  && !x86_64_immediate_operand (operands[1], <MODE>mode))
+	operands[1] = force_reg (<MODE>mode, operands[1]);
+      emit_insn (gen_xor<mode>3 (operands[4], operands[0], operands[1]));
+    }
 
   if (operands[3] == const0_rtx)
     operands[5] = operands[2];
@@ -1558,7 +1563,12 @@
       else if (operands[2] == constm1_rtx)
 	emit_insn (gen_one_cmpl<mode>2 (operands[5], operands[3]));
       else
-	emit_insn (gen_xor<mode>3 (operands[5], operands[2], operands[3]));
+	{
+	  if (CONST_SCALAR_INT_P (operands[3])
+	      && !x86_64_immediate_operand (operands[3], <MODE>mode))
+	    operands[3] = force_reg (<MODE>mode, operands[3]);
+	  emit_insn (gen_xor<mode>3 (operands[5], operands[2], operands[3]));
+	}
     }
 })
 
diff --git a/gcc/testsuite/gcc.target/i386/cmpti1.c b/gcc/testsuite/gcc.target/i386/cmpti1.c
new file mode 100644
index 000000000000..1c5f121e5443
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/cmpti1.c
@@ -0,0 +1,8 @@
+/* { dg-do compile { target int128 } } */
+/* { dg-options "-O2" } */
+int eq(__int128 x, __int128 y) { return x == y; }
+int ne(__int128 x, __int128 y) { return x != y; }
+/* { dg-final { scan-assembler-times "xorq" 4 } } */
+/* { dg-final { scan-assembler-times "setne" 1 } } */
+/* { dg-final { scan-assembler-times "sete" 1 } } */
+
diff --git a/gcc/testsuite/gcc.target/i386/cmpti2.c b/gcc/testsuite/gcc.target/i386/cmpti2.c
new file mode 100644
index 000000000000..ad9572901ce5
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/cmpti2.c
@@ -0,0 +1,12 @@
+/* { dg-do compile { target int128 } } */
+/* { dg-options "-O2" } */
+
+__int128 x;
+__int128 y;
+
+int eq() { return x == y; }
+int ne() { return x != y; }
+
+/* { dg-final { scan-assembler-times "xorq" 4 } } */
+/* { dg-final { scan-assembler-times "setne" 1 } } */
+/* { dg-final { scan-assembler-times "sete" 1 } } */
diff --git a/gcc/testsuite/gcc.target/i386/cmpti3.c b/gcc/testsuite/gcc.target/i386/cmpti3.c
new file mode 100644
index 000000000000..302efd2ad59e
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/cmpti3.c
@@ -0,0 +1,13 @@
+/* { dg-do compile { target int128 } } */
+/* { dg-options "-O2" } */
+
+__int128 x;
+int foo()
+{
+    __int128 t = 0x1234567890abcdefLL;
+    return x == t;
+}
+
+/* { dg-final { scan-assembler-times "movabsq" 1 } } */
+/* { dg-final { scan-assembler-times "xorq" 1 } } */
+/* { dg-final { scan-assembler-not "xorl" } } */
-- 
GitLab