public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* [PATCH] Fix PR81065
@ 2017-06-12 12:45 Richard Biener
  0 siblings, 0 replies; only message in thread
From: Richard Biener @ 2017-06-12 12:45 UTC (permalink / raw)
  To: gcc-patches


I am re-testing the following patch to fix PR81065 with the
fold_addr_of_array_ref_difference hunk added which was the only
case causing the gcc_unreachable to trigger in a all languages
bootstrap and regtest on x86_64-unknown-linux-gnu.  The patch
to commit will omit this case completely.

Bootstrap and regtest running on x86_64-unknown-linux-gnu.

Richard.

2017-06-12  Richard Biener  <rguenther@suse.de>

	PR middle-end/81065
	* fold-const.c (extract_muldiv_1): Remove bogus distribution
	case of C * (x * C2 + C3).
	(fold_addr_of_array_ref_difference): Properly fold index difference.

	* c-c++-common/ubsan/pr81065.c: New testcase.

Index: gcc/testsuite/c-c++-common/ubsan/pr81065.c
===================================================================
--- gcc/testsuite/c-c++-common/ubsan/pr81065.c	(nonexistent)
+++ gcc/testsuite/c-c++-common/ubsan/pr81065.c	(working copy)
@@ -0,0 +1,13 @@
+/* { dg-do run } */
+/* { dg-options "-fsanitize=undefined -fsanitize-undefined-trap-on-error" } */
+
+unsigned char x = 154;
+int foo() {
+  // 8575 * (254408 - 9057) = 8575 * 245351 = 2103884825 = 0x7d66bc19
+  return 8575 * (1652 * x - 9057);
+}
+
+int main() {
+  foo();
+  return 0;
+}
Index: gcc/fold-const.c
===================================================================
--- gcc/fold-const.c	(revision 249112)
+++ gcc/fold-const.c	(working copy)
@@ -6243,8 +6243,11 @@ extract_muldiv_1 (tree t, tree c, enum t
 
       /* If we were able to eliminate our operation from the first side,
 	 apply our operation to the second side and reform the PLUS.  */
-      if (t1 != 0 && (TREE_CODE (t1) != code || code == MULT_EXPR))
+      if (t1 != 0 && TREE_CODE (t1) != code)
+	{
+	  gcc_unreachable ();
 	return fold_build2 (tcode, ctype, fold_convert (ctype, t1), op1);
+	}
 
       /* The last case is if we are a multiply.  In that case, we can
 	 apply the distributive law to commute the multiply and addition
@@ -8908,7 +8911,7 @@ fold_addr_of_array_ref_difference (locat
       tree op0 = fold_convert_loc (loc, type, TREE_OPERAND (aref0, 1));
       tree op1 = fold_convert_loc (loc, type, TREE_OPERAND (aref1, 1));
       tree esz = fold_convert_loc (loc, type, array_ref_element_size (aref0));
-      tree diff = build2 (MINUS_EXPR, type, op0, op1);
+      tree diff = fold_build2_loc (loc, MINUS_EXPR, type, op0, op1);
       return fold_build2_loc (loc, PLUS_EXPR, type,
 			      base_offset,
 			      fold_build2_loc (loc, MULT_EXPR, type,

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2017-06-12 12:45 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-06-12 12:45 [PATCH] Fix PR81065 Richard Biener

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).