re PR tree-optimization/53081 (memcpy/memset loop recognition)

2012-06-06  Richard Guenther  <rguenther@suse.de>

	PR tree-optimization/53081
	* tree-data-ref.h (adjacent_store_dr_p): Rename to ...
	(adjacent_dr_p): ... this and make it work for reads, too.
	* tree-loop-distribution.c (enum partition_kind): Add PKIND_MEMCPY.
	(struct partition_s): Change main_stmt to main_dr, add
	secondary_dr member.
	(build_size_arg_loc): Change to date data-reference and not
	gimplify here.
	(build_addr_arg_loc): New function split out from ...
	(generate_memset_builtin): ... here.  Use it and simplify.
	(generate_memcpy_builtin): New function.
	(generate_code_for_partition): Adjust.
	(classify_partition): Streamline pattern detection.  Detect
	memcpy.
	(ldist_gen): Adjust.
	(tree_loop_distribution): Adjust seed statements for memcpy
	recognition.

	* gcc.dg/tree-ssa/ldist-20.c: New testcase.
	* gcc.dg/tree-ssa/loop-19.c: Add -fno-tree-loop-distribute-patterns.

From-SVN: r188261

gcc/ChangeLog

+2012-06-06  Richard Guenther  <rguenther@suse.de>
+
+	PR tree-optimization/53081
+	* tree-data-ref.h (adjacent_store_dr_p): Rename to ...
+	(adjacent_dr_p): ... this and make it work for reads, too.
+	* tree-loop-distribution.c (enum partition_kind): Add PKIND_MEMCPY.
+	(struct partition_s): Change main_stmt to main_dr, add
+	secondary_dr member.
+	(build_size_arg_loc): Change to date data-reference and not
+	gimplify here.
+	(build_addr_arg_loc): New function split out from ...
+	(generate_memset_builtin): ... here.  Use it and simplify.
+	(generate_memcpy_builtin): New function.
+	(generate_code_for_partition): Adjust.
+	(classify_partition): Streamline pattern detection.  Detect
+	memcpy.
+	(ldist_gen): Adjust.
+	(tree_loop_distribution): Adjust seed statements for memcpy
+	recognition.
+
 2012-06-06  Matt Turner  <mattst88@gmail.com>
 
 	* config/arm/mmintrin.h (_mm_empty): New.

gcc/testsuite/ChangeLog

+2012-06-06  Richard Guenther  <rguenther@suse.de>
+
+	PR tree-optimization/53081
+	* gcc.dg/tree-ssa/ldist-20.c: New testcase.
+	* gcc.dg/tree-ssa/loop-19.c: Add -fno-tree-loop-distribute-patterns.
+
 2012-06-05  Michael Meissner  <meissner@linux.vnet.ibm.com>
 
 	* gcc.target/powerpc/pr53487.c: New test.

gcc/testsuite/gcc.dg/tree-ssa/ldist-20.c

+/* { dg-do compile } */
+/* { dg-options "-O2 -ftree-loop-distribute-patterns -fdump-tree-ldist-details" } */
+
+void foo(char *);
+void my_memcpy (void *q, unsigned int n)
+{
+  unsigned i;
+  char p[1024];
+  for (i = 0; i < n; ++i)
+    ((char *)p)[i] = ((char *)q)[i];
+  foo(p);
+}
+
+struct S { int i; int j; };
+
+void my_memcpy2 (void *q, unsigned int n)
+{
+  unsigned i;
+  char p[1024];
+  for (i = 0; i < n; ++i)
+    ((struct S *)p)[i] = ((struct S *)q)[i];
+  foo(p);
+}
+
+char p[1024];
+void my_memmove (unsigned int n)
+{
+  unsigned i;
+  for (i = 0; i < n; ++i)
+    p[i] = p[i+1];
+  foo(p);
+}
+
+
+/* { dg-final { scan-tree-dump-times "generated memcpy" 2 "ldist" } } */
+/* { dg-final { scan-tree-dump-times "generated memmove" 1 "ldist" } } */
+/* { dg-final { cleanup-tree-dump "ldist" } } */

gcc/testsuite/gcc.dg/tree-ssa/loop-19.c

@@ -6,7 +6,7 @@
 
 /* { dg-do compile { target { i?86-*-* || { x86_64-*-* || powerpc_hard_double } } } } */
 /* { dg-require-effective-target nonpic } */
-/* { dg-options "-O3 -fno-prefetch-loop-arrays -fdump-tree-optimized" } */
+/* { dg-options "-O3 -fno-tree-loop-distribute-patterns -fno-prefetch-loop-arrays -fdump-tree-optimized" } */
 
 # define N      2000000
 double   a[N],c[N];

gcc/tree-data-ref.h

@@ -615,11 +615,8 @@ bool rdg_defs_used_in_other_loops_p (struct graph *, int);
    with a stride equal to its unit type size.  */
 
 static inline bool
-adjacent_store_dr_p (struct data_reference *dr)
+adjacent_dr_p (struct data_reference *dr)
 {
-  if (!DR_IS_WRITE (dr))
-    return false;
-
   /* If this is a bitfield store bail out.  */
   if (TREE_CODE (DR_REF (dr)) == COMPONENT_REF
       && DECL_BIT_FIELD (TREE_OPERAND (DR_REF (dr), 1)))