[Bug tree-optimization/48189] New: [4.3/4.4/4.5/4.6/4.7 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[Bug tree-optimization/48189] New: [4.3/4.4/4.5/4.6/4.7 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[zsojka at seznam dot cz]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

           Summary: [4.3/4.4/4.5/4.6/4.7 Regression] ICE: SIGFPE (division
                    by zero) in in predict_loops () at predict.c:991 with
                    --param max-predicted-iterations=0
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: zsojka@seznam.cz
              Host: x86_64-pc-linux-gnu
            Target: x86_64-pc-linux-gnu


Created attachment 23708
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23708
reduced testcase

Compiler output:
$ gcc -O --param max-predicted-iterations=0 testcase.c
testcase.c: In function 'foo':
testcase.c:11:1: internal compiler error: Floating point exception
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

Program received signal SIGFPE, Arithmetic exception.
0x000000000080aab2 in predict_loops () at /mnt/svn/gcc-trunk/gcc/predict.c:991
991               probability = ((REG_BR_PROB_BASE + nitercst / 2) / nitercst);

[Bug tree-optimization/48189] [4.3/4.4/4.5/4.6/4.7 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2011.03.21 12:46:09
                 CC|                            |hubicka at gcc dot gnu.org
   Target Milestone|---                         |4.3.6
     Ever Confirmed|0                           |1

--- Comment #1 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-03-21 12:46:09 UTC ---
Honza, if the loop is predicted to run zero times the exit edge isn't executed.
What's the probability value for this?  I'd say it's NaN (aka, not relevant
as it isn't executed as all).

It doesn't seem to be correct to limit max-predicted-iterations to >= 1
as estimated_loop_iterations_int could return zero as well.

[Bug tree-optimization/48189] [4.3/4.4/4.5/4.6/4.7 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2

--- Comment #2 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-04-10 10:44:21 UTC ---
Honza, ping?

[Bug tree-optimization/48189] [4.3/4.4/4.5/4.6/4.7 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[steven at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

Steven Bosscher <steven at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |steven at gcc dot gnu.org

--- Comment #3 from Steven Bosscher <steven at gcc dot gnu.org> 2011-04-10 13:33:30 UTC ---
I'd say probability is zero i.e. never executed, as a special case. Or if you
consider the edge always taken, then the probability should be
REG_BR_PROB_BASE. 

But zero seems more consistent:

Index: predict.c
===================================================================
--- predict.c   (revision 172225)
+++ predict.c   (working copy)
@@ -988,7 +988,14 @@
          else
            continue;

-         probability = ((REG_BR_PROB_BASE + nitercst / 2) / nitercst);
+         if (nitercst == 0 && predictor == PRED_LOOP_ITERATIONS)
+           /* If the prediction for number of iterations is zero, the exits
+              will never be taken.  Only make this prediction if we are quite
+              sure that the loop will never roll, from analysis.  */
+           probability = 0;
+         else
+           probability = ((REG_BR_PROB_BASE + nitercst / 2) / nitercst);
+
          predict_edge (ex, predictor, probability);
        }
       VEC_free (edge, heap, exits);


Or maybe just not predict at all:

Index: predict.c
===================================================================
--- predict.c   (revision 172225)
+++ predict.c   (working copy)
@@ -988,6 +988,11 @@
          else
            continue;

+         /* If the prediction for number of iterations is zero, do not
+            predict the exit edges.  */
+         if (nitercst == 0)
+           continue;
+
          probability = ((REG_BR_PROB_BASE + nitercst / 2) / nitercst);
          predict_edge (ex, predictor, probability);
        }

[Bug tree-optimization/48189] [4.3/4.4/4.5/4.6/4.7 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.3.6                       |4.4.7

--- Comment #4 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-06-27 12:14:49 UTC ---
4.3 branch is being closed, moving to 4.4.7 target.

[Bug tree-optimization/48189] [4.4/4.5/4.6/4.7 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-12-20 13:10:15 UTC ---
Already the:
              if (host_integerp (niter, 1)
                  && compare_tree_int (niter, max-1) == -1)
                nitercst = tree_low_cst (niter, 1) + 1;
looks wrong, if max is 0, then we are calling compare_tree_int with UHWI
maximum
and even huge nitercst values will make through.
IMHO we should just guard the above with max != 0 as well.
I think max_stmt_executions_int will never return 0, so both patches probably
work fine.  Honza?

[Bug tree-optimization/48189] [4.5/4.6/4.7/4.8 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.4.7                       |4.5.4

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> 2012-03-13 12:48:25 UTC ---
4.4 branch is being closed, moving to 4.5.4 target.

[Bug tree-optimization/48189] [4.5/4.6/4.7/4.8 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

Jan Hubicka <hubicka at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
         AssignedTo|unassigned at gcc dot       |hubicka at gcc dot gnu.org
                   |gnu.org                     |

[Bug tree-optimization/48189] [4.5/4.6/4.7/4.8 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.5.4                       |4.6.4

--- Comment #7 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-07-02 12:20:06 UTC ---
The 4.5 branch is being closed, adjusting target milestone.

[Bug tree-optimization/48189] [4.6/4.7/4.8 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

--- Comment #8 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-01-08 15:02:52 UTC ---
Created attachment 29105
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=29105
gcc48-pr48189.patch

I'd say this isn't difficult to solve enough to warrant having it open for
years.

[Bug tree-optimization/48189] [4.6/4.7/4.8 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

--- Comment #9 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-01-09 09:00:32 UTC ---
Author: jakub
Date: Wed Jan  9 09:00:22 2013
New Revision: 195046

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=195046
Log:
    PR tree-optimization/48189
    * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
    If nitercst is 0, don't predict the exit edge.

    * gcc.dg/pr48189.c: New test.

Added:
    trunk/gcc/testsuite/gcc.dg/pr48189.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/predict.c
    trunk/gcc/testsuite/ChangeLog

[Bug tree-optimization/48189] [4.6/4.7 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |4.8.0
            Summary|[4.6/4.7/4.8 Regression]    |[4.6/4.7 Regression] ICE:
                   |ICE: SIGFPE (division by    |SIGFPE (division by zero)
                   |zero) in in predict_loops   |in in predict_loops () at
                   |() at predict.c:991 with    |predict.c:991 with --param
                   |--param                     |max-predicted-iterations=0
                   |max-predicted-iterations=0  |

--- Comment #10 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-01-09 09:03:38 UTC ---
Fixed on the trunk so far.

[Bug tree-optimization/48189] [4.6/4.7 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[mpolacek at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mpolacek at gcc dot gnu.org
         AssignedTo|hubicka at gcc dot gnu.org  |mpolacek at gcc dot gnu.org

--- Comment #11 from Marek Polacek <mpolacek at gcc dot gnu.org> 2013-04-04 09:15:55 UTC ---
I'll do the backports.

[Bug tree-optimization/48189] [4.6/4.7 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[mpolacek at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

--- Comment #12 from Marek Polacek <mpolacek at gcc dot gnu.org> 2013-04-04 11:05:08 UTC ---
Author: mpolacek
Date: Thu Apr  4 11:03:11 2013
New Revision: 197478

URL: http://gcc.gnu.org/viewcvs?rev=197478&root=gcc&view=rev
Log:
        Backported from mainline
        2013-01-09  Steven Bosscher  <steven@gcc.gnu.org>
                    Jakub Jelinek  <jakub@redhat.com>

        PR tree-optimization/48189
        * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
        If nitercst is 0, don't predict the exit edge.

Added:
    branches/gcc-4_6-branch/gcc/testsuite/gcc.dg/pr48189.c
Modified:
    branches/gcc-4_6-branch/gcc/ChangeLog
    branches/gcc-4_6-branch/gcc/predict.c
    branches/gcc-4_6-branch/gcc/testsuite/ChangeLog

[Bug tree-optimization/48189] [4.7 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.6.4                       |4.7.4

--- Comment #13 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-04-12 15:17:11 UTC ---
GCC 4.6.4 has been released and the branch has been closed.

[Bug tree-optimization/48189] [4.7 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[mpolacek at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

--- Comment #14 from Marek Polacek <mpolacek at gcc dot gnu.org> 2013-04-19 08:20:11 UTC ---
Author: mpolacek
Date: Fri Apr 19 08:18:29 2013
New Revision: 198085

URL: http://gcc.gnu.org/viewcvs?rev=198085&root=gcc&view=rev
Log:
    Backport from mainline
    2013-01-08  Steven Bosscher  <steven@gcc.gnu.org>
            Jakub Jelinek  <jakub@redhat.com>

    PR tree-optimization/48189
    * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
    If nitercst is 0, don't predict the exit edge.

    * gcc.dg/pr48189.c: New test.


Added:
    branches/gcc-4_7-branch/gcc/testsuite/gcc.dg/pr48189.c
Modified:
    branches/gcc-4_7-branch/gcc/ChangeLog
    branches/gcc-4_7-branch/gcc/predict.c
    branches/gcc-4_7-branch/gcc/testsuite/ChangeLog

[Bug tree-optimization/48189] [4.7 Regression] ICE: SIGFPE (division by zero) in in predict_loops () at predict.c:991 with --param max-predicted-iterations=0

[mpolacek at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48189

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

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

--- Comment #15 from Marek Polacek <mpolacek at gcc dot gnu.org> 2013-04-19 08:20:34 UTC ---
Fixed.