ix86: Add peephole2 for *add<mode>3_cc_overflow_1 followed by matching memory store...

The following peephole2 changes:
-	addl	(%rdi), %esi
+	xorl	%eax, %eax
+	addl	%esi, (%rdi)
 	setc	%al
-	movl	%esi, (%rdi)
-	movzbl	%al, %eax
 	ret
on the testcase.  *add<mode>3_cc_overflow_1, being an add{l,q} insn, is
commutative, so if TARGET_READ_MODIFY_WRITE we can replace
addl (%rdi), %esi; movl %esi, (%rdi)
with
addl %esi, (%rdi)
if %esi is dead after those two insns.

2020-05-08  Jakub Jelinek  <jakub@redhat.com>

	PR target/94857
	* config/i386/i386.md (peephole2 after *add<mode>3_cc_overflow_1): New
	define_peephole2.

	* gcc.target/i386/pr94857.c: New test.

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index 2e5a051..9a380a5 100644 (file)
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,5 +1,9 @@
 2020-05-08  Jakub Jelinek  <jakub@redhat.com>
 
+       PR target/94857
+       * config/i386/i386.md (peephole2 after *add<mode>3_cc_overflow_1): New
+       define_peephole2.
+
        PR middle-end/94724
        * tree.c (get_narrower): Reuse the op temporary instead of
        shadowing it.

diff --git a/gcc/config/i386/i386.md b/gcc/config/i386/i386.md
index 5fe851e..8bfc9cb 100644 (file)
--- a/gcc/config/i386/i386.md
+++ b/gcc/config/i386/i386.md
@@ -6992,6 +6992,23 @@
   [(set_attr "type" "alu")
    (set_attr "mode" "<MODE>")])
 
+(define_peephole2
+  [(parallel [(set (reg:CCC FLAGS_REG)
+                  (compare:CCC
+                    (plus:SWI (match_operand:SWI 0 "general_reg_operand")
+                              (match_operand:SWI 1 "memory_operand"))
+                    (match_dup 0)))
+             (set (match_dup 0) (plus:SWI (match_dup 0) (match_dup 1)))])
+   (set (match_dup 1) (match_dup 0))]
+  "(TARGET_READ_MODIFY_WRITE || optimize_insn_for_size_p ())
+   && peep2_reg_dead_p (2, operands[0])
+   && !reg_overlap_mentioned_p (operands[0], operands[1])"
+  [(parallel [(set (reg:CCC FLAGS_REG)
+                  (compare:CCC
+                    (plus:SWI (match_dup 1) (match_dup 0))
+                    (match_dup 1)))
+             (set (match_dup 1) (plus:SWI (match_dup 1) (match_dup 0)))])])
+
 (define_insn "*addsi3_zext_cc_overflow_1"
   [(set (reg:CCC FLAGS_REG)
        (compare:CCC

diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index 174198f..db0a837 100644 (file)
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -1,5 +1,8 @@
 2020-05-08  Jakub Jelinek  <jakub@redhat.com>
 
+       PR target/94857
+       * gcc.target/i386/pr94857.c: New test.
+
        PR tree-optimization/94783
        * gcc.dg/tree-ssa/pr94783.c: New test.
 

diff --git a/gcc/testsuite/gcc.target/i386/pr94857.c b/gcc/testsuite/gcc.target/i386/pr94857.c
new file mode 100644 (file)
index 0000000..f84ee22
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/pr94857.c
@@ -0,0 +1,13 @@
+/* PR target/94857 */
+/* { dg-do compile } */
+/* { dg-options "-O2 -mtune=skylake -masm=att" } */
+/* { dg-additional-options "-mregparm=2" { target ia32 } } */
+/* { dg-final { scan-assembler "\taddl\t%\[a-z0-9]\*, \\\(" } } */
+
+int
+foo (unsigned *p, unsigned x)
+{
+  unsigned u = *p;
+  *p += x;
+  return u > *p;
+}