public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc r9-8901] fold-const: Fix A <= 0 ? A : -A folding [PR95810]
@ 2020-09-16 19:22 Jakub Jelinek
  0 siblings, 0 replies; only message in thread
From: Jakub Jelinek @ 2020-09-16 19:22 UTC (permalink / raw)
  To: gcc-cvs

https://gcc.gnu.org/g:6ff6c02695c9b6ae6e840422080f6d10449577b8

commit r9-8901-g6ff6c02695c9b6ae6e840422080f6d10449577b8
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Wed Jun 24 10:40:02 2020 +0200

    fold-const: Fix A <= 0 ? A : -A folding [PR95810]
    
    We folded A <= 0 ? A : -A into -ABS (A), which is for signed integral types
    incorrect - can invoke on INT_MIN UB twice, once on ABS and once on its
    negation.
    
    The following patch fixes it by instead folding it to (type)-ABSU (A).
    
    2020-06-24  Jakub Jelinek  <jakub@redhat.com>
    
            PR middle-end/95810
            * fold-const.c (fold_cond_expr_with_comparison): Optimize
            A <= 0 ? A : -A into (type)-absu(A) rather than -abs(A).
    
            * gcc.dg/ubsan/pr95810.c: New test.
    
    (cherry picked from commit 01e10b0ee77e82cb331414c569e02dc7a2c4999e)

Diff:
---
 gcc/fold-const.c                     | 18 ++++++++++++++++--
 gcc/testsuite/gcc.dg/ubsan/pr95810.c | 13 +++++++++++++
 2 files changed, 29 insertions(+), 2 deletions(-)

diff --git a/gcc/fold-const.c b/gcc/fold-const.c
index 64273c1c6bc..8388bde0434 100644
--- a/gcc/fold-const.c
+++ b/gcc/fold-const.c
@@ -5423,8 +5423,22 @@ fold_cond_expr_with_comparison (location_t loc, tree type,
       case LT_EXPR:
 	if (TYPE_UNSIGNED (TREE_TYPE (arg1)))
 	  break;
-	tem = fold_build1_loc (loc, ABS_EXPR, TREE_TYPE (arg1), arg1);
-	return negate_expr (fold_convert_loc (loc, type, tem));
+	if (ANY_INTEGRAL_TYPE_P (TREE_TYPE (arg1))
+	    && !TYPE_OVERFLOW_WRAPS (TREE_TYPE (arg1)))
+	  {
+	    /* A <= 0 ? A : -A for A INT_MIN is valid, but -abs(INT_MIN)
+	       is not, invokes UB both in abs and in the negation of it.
+	       So, use ABSU_EXPR instead.  */
+	    tree utype = unsigned_type_for (TREE_TYPE (arg1));
+	    tem = fold_build1_loc (loc, ABSU_EXPR, utype, arg1);
+	    tem = negate_expr (tem);
+	    return fold_convert_loc (loc, type, tem);
+	  }
+	else
+	  {
+	    tem = fold_build1_loc (loc, ABS_EXPR, TREE_TYPE (arg1), arg1);
+	    return negate_expr (fold_convert_loc (loc, type, tem));
+	  }
       default:
 	gcc_assert (TREE_CODE_CLASS (comp_code) == tcc_comparison);
 	break;
diff --git a/gcc/testsuite/gcc.dg/ubsan/pr95810.c b/gcc/testsuite/gcc.dg/ubsan/pr95810.c
new file mode 100644
index 00000000000..535ace69089
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/ubsan/pr95810.c
@@ -0,0 +1,13 @@
+/* PR middle-end/95810 */
+/* { dg-do run } */
+/* { dg-options "-fsanitize=undefined -fno-sanitize-recover=undefined" } */
+
+int
+main ()
+{
+  int x = -__INT_MAX__ - 1;
+  x = (x <= 0 ? x : -x);
+  if (x != -__INT_MAX__ - 1)
+    __builtin_abort ();
+  return 0;
+}


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

only message in thread, other threads:[~2020-09-16 19:22 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-09-16 19:22 [gcc r9-8901] fold-const: Fix A <= 0 ? A : -A folding [PR95810] Jakub Jelinek

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).