re PR middle-end/67442 (GCC 5.2.0 on x86_64 creates invalid address on specific…

re PR middle-end/67442 (GCC 5.2.0 on x86_64 creates invalid address on specific array index calculation through pointer)

2015-09-16  Richard Biener  <rguenther@suse.de>

	PR middle-end/67442
	* fold-const.c (extract_muldiv_1): Properly extend multiplication
	result before builting a tree via wide_int_to_tree.

	* gcc.dg/torture/pr67442.c: New testcase.

From-SVN: r227818

gcc/ChangeLog

+2015-09-16  Richard Biener  <rguenther@suse.de>
+
+	PR middle-end/67442
+	* fold-const.c (extract_muldiv_1): Properly extend multiplication
+	result before builting a tree via wide_int_to_tree.
+
 2015-09-16  Mikhail Maltsev  <maltsevm@gmail.com>
 
 	* Makefile.in: Add memory-block.cc

gcc/fold-const.c

@@ -6166,8 +6166,12 @@ extract_muldiv_1 (tree t, tree c, enum tree_code code, tree wide_type,
 	      && ((sign == UNSIGNED && tcode != MULT_EXPR) || sign == SIGNED))
 	    overflow_p = true;
 	  if (!overflow_p)
-	    return fold_build2 (tcode, ctype, fold_convert (ctype, op0),
-				wide_int_to_tree (ctype, mul));
+	    {
+	      mul = wide_int::from (mul, TYPE_PRECISION (ctype),
+				    TYPE_SIGN (TREE_TYPE (op1)));
+	      return fold_build2 (tcode, ctype, fold_convert (ctype, op0),
+				  wide_int_to_tree (ctype, mul));
+	    }
 	}
 
       /* If these operations "cancel" each other, we have the main

gcc/testsuite/ChangeLog

+2015-09-16  Richard Biener  <rguenther@suse.de>
+
+	PR middle-end/67442
+	* gcc.dg/torture/pr67442.c: New testcase.
+
 2015-09-15  Eric Botcazou  <ebotcazou@adacore.com>
 
 	* gnat.dg/array24.adb: New test.

gcc/testsuite/gcc.dg/torture/pr67442.c

+/* { dg-do run } */
+
+short foo[100];
+
+int main()
+{
+  short* bar = &foo[50];
+  short i = 1;
+  short j = 1;
+  short value = bar[8 - i * 2 * j];
+  return value;
+}