[Bug c/26859] New: ICE Segmentation Fault

[Bug c/26859] New: ICE Segmentation Fault

[malitzke at metronets dot com]

dunno
see complete test case


-- 
           Summary: ICE Segmentation Fault
           Product: gcc
           Version: 4.2.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: malitzke at metronets dot com
 GCC build triplet: i686-slackware-linux
  GCC host triplet: i686-slackware-linux
GCC target triplet: i686-slackware-linux


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26859

[Bug c/26859] ICE Segmentation Fault

[malitzke at metronets dot com]

------- Comment #1 from malitzke at metronets dot com  2006-03-25 00:45 -------
Created an attachment (id=11118)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11118&action=view)
complete test case (command, console out, *.e)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26859

[Bug tree-optimization/26859] ICE Segmentation Fault

[pinskia at gcc dot gnu dot org]

------- Comment #2 from pinskia at gcc dot gnu dot org  2006-03-25 00:59 -------
Reducing.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26859

[Bug tree-optimization/26859] ICE Segmentation Fault

[pinskia at gcc dot gnu dot org]

------- Comment #3 from pinskia at gcc dot gnu dot org  2006-03-25 01:00 -------
Oh, by the way please next time put the command and console out not in the
attachment but in the comments.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26859

[Bug tree-optimization/26859] [4.2 Regression] ICE Segmentation Fault

[pinskia at gcc dot gnu dot org]

------- Comment #4 from pinskia at gcc dot gnu dot org  2006-03-25 02:40 -------
Reduced testcase:
void f(long nb)
{
    long i, j, d, d2;
    for (i=1; i<=nb; i++)
        if (i>1)
        {
            d = i-1;
            d2 = 2*i-3;
            for (j=d; j>0; j--)
            {
                divrs( d2 );
                d2-=2;
            }
        }
}

----
The segfault is in fold-const but the backtrace comes from scev.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |spop at gcc dot gnu dot org
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
  GCC build triplet|i686-slackware-linux        |
   GCC host triplet|i686-slackware-linux        |
 GCC target triplet|i686-slackware-linux        |
   Last reconfirmed|0000-00-00 00:00:00         |2006-03-25 02:40:09
               date|                            |
            Summary|ICE Segmentation Fault      |[4.2 Regression] ICE
                   |                            |Segmentation Fault
   Target Milestone|---                         |4.2.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26859

[Bug tree-optimization/26859] [4.2 Regression] ICE Segmentation Fault

[malitzke at metronets dot com]

------- Comment #5 from malitzke at metronets dot com  2006-03-27 15:17 -------
The fairly extensive changes submitted by Mr Sebastian Pop solved the PR 26859
as originally reported by myself. 


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26859

[Bug tree-optimization/26859] [4.2 Regression] ICE Segmentation Fault

[malitzke at metronets dot com]

------- Comment #6 from malitzke at metronets dot com  2006-03-28 14:45 -------
As no further comments are forthcoming I am taking the liberty to mark this PR
as resolved.


-- 

malitzke at metronets dot com changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26859

[Bug tree-optimization/26859] [4.2 Regression] ICE Segmentation Fault

[reichelt at gcc dot gnu dot org]

------- Comment #7 from reichelt at gcc dot gnu dot org  2006-03-28 14:50 -------
The problem is not fixed.
The reduced testcase from comment #4 still crashes.


-- 

reichelt at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26859

[Bug tree-optimization/26859] [4.2 Regression] ICE Segmentation Fault

[malitzke at metronets dot com]

------- Comment #8 from malitzke at metronets dot com  2006-03-28 17:07 -------
You are correct. At least as far as the second batch from S.Pop (dated
2006-03-27)
goes. As I did not keep my gcc-4.2.0 binaries relative to the first batch from
S.Pop (dated 2006-03-26) I can not reproduce the non ICE on trans2.c for
pari-2.1.7 (from the PARI group Bordeaux Fr). Plus ca change; plus ca reste la
meme chose.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26859

[Bug tree-optimization/26859] [4.2 Regression] ICE Segmentation Fault

[sebastian dot pop at cri dot ensmp dot fr]

------- Comment #9 from sebastian dot pop at cri dot ensmp dot fr  2006-03-29 00:27 -------
Created an attachment (id=11147)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11147&action=view)
proposed fix

this patch (not tested yet) fixes the problem: it avoids a division by zero.
Part of the problem is that convert returns a step for which it sets the
overflow flag when converting unsigned ffffffe to int -2.  The patch resets
the overflow flag that has no sense for the step of a sequence.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26859

[Bug tree-optimization/26859] [4.2 Regression] ICE Segmentation Fault

[spop at gcc dot gnu dot org]

------- Comment #10 from spop at gcc dot gnu dot org  2006-03-29 17:20 -------
Subject: Bug 26859

Author: spop
Date: Wed Mar 29 17:20:24 2006
New Revision: 112502

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=112502
Log:
        PR tree-optimization/26859
        * tree-ssa-loop-niter.c (infer_loop_bounds_from_undefined): Avoid
        division by zero.
        (convert_step): Remove TREE_OVERFLOW and TREE_CONSTANT_OVERFLOW flags
        for the step after fold_convert.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/tree-ssa-loop-niter.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26859

[Bug tree-optimization/26859] [4.2 Regression] ICE Segmentation Fault

[malitzke at metronets dot com]

------- Comment #11 from malitzke at metronets dot com  2006-03-29 22:24 -------
Maybe I should keep my mouth shut.
However, gcc-4.2.0 2006329 again compiles pari-2.1.7 OK. 2nd However,
pari-2.2.12.beta uses a different approch, which does not give any problems
with various gcc-4.2.0. At least the problem surfaced in PR26859 seems pretty
infrequent and is now _apparently_ solved. Regards


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26859

[Bug tree-optimization/26859] [4.2 Regression] ICE Segmentation Fault

[reichelt at gcc dot gnu dot org]

------- Comment #12 from reichelt at gcc dot gnu dot org  2006-03-31 07:16 -------
Fixed on mainline.


-- 

reichelt at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26859