Re: [PATCH, middle-end] Fix PR 37861 - Bogus array bounds warning

[Martin Jambor <mjambor at suse dot cz>]

Hi,

On Wed, Nov 05, 2008 at 10:56:27AM +0100, Richard Guenther wrote:
> On Wed, 5 Nov 2008, Martin Jambor wrote:
> > the following patch gets rid of the bogus warning.
> > 
> > The problem is, that fwprop currently generates array_refs of the form
> > &(*formatstr_3(D))[cumulative_offset] for  multidimensional arrays and
> > the cumulative offset can then naturally be bigger than the respective
> > dimension of the array.  
> > 
> > This  happens because, when  multidimensional arrays  are passed  as a
> > formal   parameter,  they   are  gimplified   into  a   pointer,  thus
> > formatstr[10][100] becomes  char[100] * formatstr.  We do  not want to
> > be doing the  transformation on such variables and  the patch inhibits
> > them.
> > 
> > I have bootstrapped and regression tested it on x86_64-suse-linux
> > (revision 141561).  I hope this won't cause any performance problems.
> > 
> > OK for trunk?
> 
> This is ok if you add both testcases from the PR and add a brief
> comment before the INDIRECT_REF check in the patch.
> 

I originally intended to include the testcases but somehow forgot in
the end.  This is the updated patch, I will commit it in a few hours,
unless someone objects.

Thanks,

Martin

2008-11-05  Martin Jambor  <mjambor@suse.cz>

	PR middle-end/37861
	* tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't turn
	pointer arithmetics into array_ref if the array is accessed
	through an indirect_ref.
	* testsuite/gcc.dg/Warray-bounds-5.c: New file.
	* testsuite/gcc.dg/Warray-bounds-6.c: New file.
	
Index: gcc/testsuite/gcc.dg/Warray-bounds-6.c
===================================================================
--- gcc/testsuite/gcc.dg/Warray-bounds-6.c	(revision 0)
+++ gcc/testsuite/gcc.dg/Warray-bounds-6.c	(revision 0)
@@ -0,0 +1,29 @@
+/* { dg-do compile } */
+/* { dg-options "-O3 -Wall" } */
+/* based on PR 37861 */
+
+extern int printf (__const char *__restrict __format, ...);
+
+static int f3(int v)
+{
+  int i,j = 0;
+  for (i = 0; i <= v; i++)
+    j++;
+  return j;
+}
+
+static int f2(char formatstr[10][100]) {
+  printf( "%d %s\n", 0, formatstr[f3(0)] );
+  printf( "%d %s\n", 1, formatstr[f3(1)] );
+  printf( "%d %s\n", 2, formatstr[f3(2)] );
+  printf( "%d %s\n", 3, formatstr[f3(3)] );
+  return 3;
+}
+
+static   char formatstr[10][100];
+int main( void ) {
+  int anz;
+  anz = f2(formatstr);
+  printf( "   %d\n",anz);
+  return 0;
+}
Index: gcc/testsuite/gcc.dg/Warray-bounds-5.c
===================================================================
--- gcc/testsuite/gcc.dg/Warray-bounds-5.c	(revision 0)
+++ gcc/testsuite/gcc.dg/Warray-bounds-5.c	(revision 0)
@@ -0,0 +1,24 @@
+/* { dg-do compile } */
+/* { dg-options "-O3 -Wall" } */
+/* based on PR 37861 */
+
+extern int printf (__const char *__restrict __format, ...);
+
+static int f2(char formatstr[10][100])  
+{
+  int anz;
+  for( anz = 0; anz < 10; ++anz ) {
+    printf( "%d %s\n", anz, formatstr[anz] );
+  } 
+  return anz;
+}
+
+
+static   char formatstr[10][100];
+int main( void ) 
+{
+  int anz;
+  anz = f2(formatstr);
+  printf( "   %d\n",anz);
+  return 0;
+}
Index: gcc/tree-ssa-forwprop.c
===================================================================
--- gcc/tree-ssa-forwprop.c	(revision 141546)
+++ gcc/tree-ssa-forwprop.c	(working copy)
@@ -812,6 +812,9 @@ forward_propagate_addr_expr_1 (tree name
   array_ref = TREE_OPERAND (def_rhs, 0);
   if (TREE_CODE (array_ref) != ARRAY_REF
       || TREE_CODE (TREE_TYPE (TREE_OPERAND (array_ref, 0))) != ARRAY_TYPE
+      /* Avoid accessing hidden multidimensional arrays in this way or VRP
+	 might give out bogus warnings (see PR 37861) */
+      || TREE_CODE (TREE_OPERAND (array_ref, 0)) == INDIRECT_REF
       || !integer_zerop (TREE_OPERAND (array_ref, 1)))
     return false;