[Bug tree-optimization/47677] New: [4.6 Regression] Hang in VRP

[Bug tree-optimization/47677] New: [4.6 Regression] Hang in VRP

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47677

           Summary: [4.6 Regression] Hang in VRP
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: jakub@gcc.gnu.org


struct S { int s; int u; };
extern int bar (void);
extern struct S *baz (void);

void
foo (int x, struct S *y, char z, int *v, struct S *s)
{
  int c, d;
  if (x & 2)
    d = 1;
  else
    {
      d = (x & 1) ? 11 : 0;
      while ((c = (s->s != (s->u & 1) ? s->s : bar ())) != '\0');
      c = (s->s != '\\' && (s->u & 4) ? s->s : bar ());
      if (c == '<')
        goto lab;
    }
  while ((c = ((s->u & 1) ? s->s : bar ())) != 0
         && ((d != 11 && d != 17) || (v[c] & 1) == 0))
    {
    lab:;
      switch (d)
        {
        case 14:
          if (c == '}')
            y = baz ();
          d = y->s = z == '<' ? 17 : 11;
        }
    }
}

at -O2 hangs in VRP pass, starting with
http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=163724
(at least, I've killed it after 7 minutes).

[Bug tree-optimization/47677] [4.6 Regression] Hang in VRP

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47677

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.6.0

[Bug tree-optimization/47677] [4.6 Regression] Hang in VRP

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47677

--- Comment #1 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-02-10 13:08:28 UTC ---
I'll investigate.

[Bug tree-optimization/47677] [4.6 Regression] Hang in VRP

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47677

Richard Guenther <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2011.02.10 13:18:04
         AssignedTo|unassigned at gcc dot       |rguenth at gcc dot gnu.org
                   |gnu.org                     |
     Ever Confirmed|0                           |1

--- Comment #2 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-02-10 13:18:04 UTC ---
Easy.

[Bug tree-optimization/47677] [4.6 Regression] Hang in VRP

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47677

--- Comment #3 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-02-10 15:28:10 UTC ---
Author: rguenth
Date: Thu Feb 10 15:28:07 2011
New Revision: 170002

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=170002
Log:
2011-02-10  Richard Guenther  <rguenther@suse.de>

    PR tree-optimization/47677
    * tree-vrp.c (vrp_bitmap_equal_p): Fix comparison of empty bitmaps.

    * gcc.dg/torture/pr47677.c: New testcase.

Added:
    trunk/gcc/testsuite/gcc.dg/torture/pr47677.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/tree-vrp.c

[Bug tree-optimization/47677] [4.6 Regression] Hang in VRP

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47677

Richard Guenther <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #4 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-02-10 16:15:59 UTC ---
Fixed.

[Bug tree-optimization/47677] [4.6 Regression] Hang in VRP

[tg at mirbsd dot org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47677

Thorsten Glaser <tg at mirbsd dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tg at mirbsd dot org

--- Comment #5 from Thorsten Glaser <tg at mirbsd dot org> 2011-02-10 17:57:31 UTC ---
Wow, that was quick, thanks. (I think mksh ought to become a compiler
testcase… found a lot of bugs in pcc, one in SUNWcc, and a few in gcc,
including when -fwhole-program --combine – now gone – broke twice.)