[Bug c++/57000] New: 4.8 Regression?, ICE with -Ofast and -frounding-math

[Bug c++/57000] New: 4.8 Regression?, ICE with -Ofast and -frounding-math

[vladimir.fuka at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57000

             Bug #: 57000
           Summary: 4.8 Regression?, ICE with -Ofast and -frounding-math
    Classification: Unclassified
           Product: gcc
           Version: 4.8.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: vladimir.fuka@gmail.com


The attached source file produces ICE for compilation command

g++ -c  -Ofast -frounding-math -std=c++11 polyhedron-preprocessed.cc -v         
Using built-in specs.                                                           
COLLECT_GCC=g++                                                                 
Target: x86_64-unknown-linux-gnu                                                
Configured with: ../gcc-4.8-20130411/configure --enable-languages=fortran
--prefix=/usr/local/gcc-4.8                                                     
Thread model: posix                                                             
gcc version 4.8.1 20130411 (prerelease) (GCC)                                   
COLLECT_GCC_OPTIONS='-c' '-Ofast' '-frounding-math' '-std=c++11' '-v'
'-shared-libgcc' '-mtune=generic' '-march=x86-64'                               
 /usr/local/gcc-4.8/libexec/gcc/x86_64-unknown-linux-gnu/4.8.1/cc1plus -quiet
-v -D_GNU_SOURCE polyhedron-preprocessed.cc -quiet -dumpbase
polyhedron-preprocessed.cc -mtune=generic -march=x86-64 -auxbase
polyhedron-preprocessed -Ofast -std=c++11 -version -frounding-math -o
/tmp/ccDOqrHe.s                               
GNU C++ (GCC) version 4.8.1 20130411 (prerelease) (x86_64-unknown-linux-gnu)    
        compiled by GNU C version 4.8.1 20130411 (prerelease), GMP version
4.3.1, MPFR version 2.4.1-p5, MPC version 0.8.1                                 
GGC heuristics: --param ggc-min-expand=100 --param ggc-min-heapsize=131072      
ignoring nonexistent directory
"/usr/local/gcc-4.8/lib/gcc/x86_64-unknown-linux-gnu/4.8.1/../../../../x86_64-unknown-linux-gnu/include" 
#include "..." search starts here:                                              
#include <...> search starts here:                                              

/usr/local/gcc-4.8/lib/gcc/x86_64-unknown-linux-gnu/4.8.1/../../../../include/c++/4.8.1 

/usr/local/gcc-4.8/lib/gcc/x86_64-unknown-linux-gnu/4.8.1/../../../../include/c++/4.8.1/x86_64-unknown-linux-gnu 

/usr/local/gcc-4.8/lib/gcc/x86_64-unknown-linux-gnu/4.8.1/../../../../include/c++/4.8.1/backward 
 /usr/local/gcc-4.8/lib/gcc/x86_64-unknown-linux-gnu/4.8.1/include              
 /usr/local/include                                                             
 /usr/local/gcc-4.8/include                                                     
 /usr/local/gcc-4.8/lib/gcc/x86_64-unknown-linux-gnu/4.8.1/include-fixed        
 /usr/include                                                                   
End of search list.                                                             
GNU C++ (GCC) version 4.8.1 20130411 (prerelease) (x86_64-unknown-linux-gnu)    
        compiled by GNU C version 4.8.1 20130411 (prerelease), GMP version
4.3.1, MPFR version 2.4.1-p5, MPC version 0.8.1                                 
GGC heuristics: --param ggc-min-expand=100 --param ggc-min-heapsize=131072      
Compiler executable checksum: 56fe9a649d5acfd8c75015f1b8b7dca4                  
In file included from /usr/local/include/CGAL/AABB_tree.h:26:0,                 
                 from cgal_polyhedra.cc:2:                                      
/usr/local/include/CGAL/internal/AABB_tree/AABB_traversal_traits.h: In member
function ‘void
CGAL::internal::AABB_tree::Projection_traits<AABBTraits>::intersection(const
Point&, const Primitive&) [with AABBTraits =
CGAL::AABB_traits<CGAL::Simple_cartesian<double>,
CGAL::AABB_polyhedron_triangle_primitive<CGAL::Simple_cartesian<double>,
CGAL::Polyhedron_3<CGAL::Simple_cartesian<double> > > >;
CGAL::internal::AABB_tree::Projection_traits<AABBTraits>::Point =
CGAL::Point_3<CGAL::Simple_cartesian<double> >;
CGAL::internal::AABB_tree::Projection_traits<AABBTraits>::Primitive =
CGAL::AABB_polyhedron_triangle_primitive<CGAL::Simple_cartesian<double>,
CGAL::Polyhedron_3<CGAL::Simple_cartesian<double> > >]’:
/usr/local/include/CGAL/internal/AABB_tree/AABB_traversal_traits.h:299:8:
internal compiler error: Neoprávněný přístup do paměti (SIGSEGV)
   void intersection(const Point& query, const Primitive& primitive)
        ^
0x8b40bf crash_signal
        ../../gcc-4.8-20130411/gcc/toplev.c:332
0x92c8b2 gimple_code
        ../../gcc-4.8-20130411/gcc/gimple.h:1149
0x92c8b2 gimple_nop_p
        ../../gcc-4.8-20130411/gcc/gimple.h:4875
0x92c8b2 walk_aliased_vdefs_1
        ../../gcc-4.8-20130411/gcc/tree-ssa-alias.c:2202
0x92d32d walk_aliased_vdefs(ao_ref_s*, tree_node*, bool (*)(ao_ref_s*,
tree_node*, void*), void*, bitmap_head_def**)
        ../../gcc-4.8-20130411/gcc/tree-ssa-alias.c:2238
0x93bdfb propagate_necessity
        ../../gcc-4.8-20130411/gcc/tree-ssa-dce.c:908
0x93cb14 perform_tree_ssa_dce
        ../../gcc-4.8-20130411/gcc/tree-ssa-dce.c:1583
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <http://gcc.gnu.org/bugs.html> for instructions.





It works on another machine with GCC 4.7.2.
>From gcc-bugs-return-420560-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Thu Apr 18 19:20:41 2013
Return-Path: <gcc-bugs-return-420560-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 25608 invoked by alias); 18 Apr 2013 19:20:41 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 25592 invoked by uid 48); 18 Apr 2013 19:20:38 -0000
From: "vladimir.fuka at gmail dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/57000] 4.8 Regression?, ICE with -Ofast and -frounding-math
Date: Thu, 18 Apr 2013 19:20:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: vladimir.fuka at gmail dot com
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Changed-Fields:
Message-ID: <bug-57000-4-XirQaZMtSo@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-57000-4@http.gcc.gnu.org/bugzilla/>
References: <bug-57000-4@http.gcc.gnu.org/bugzilla/>
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
Content-Type: text/plain; charset="UTF-8"
MIME-Version: 1.0
X-SW-Source: 2013-04/txt/msg01705.txt.bz2
Content-length: 446


http://gcc.gnu.org/bugzilla/show_bug.cgi?idW000

--- Comment #1 from Vladimir Fuka <vladimir.fuka at gmail dot com> 2013-04-18 19:20:38 UTC ---
Created attachment 29899
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id)899
Preprocessed was too long, is at
http://artax.karlin.mff.cuni.cz/~fukav1am/polyhedron-preprocessed.cc

Preprocessed was too long, is at
http://artax.karlin.mff.cuni.cz/~fukav1am/polyhedron-preprocessed.cc

[Bug c++/57000] 4.8 Regression?, ICE with -Ofast and -frounding-math

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57000

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2013-04-18
                 CC|                            |jakub at gcc dot gnu.org
     Ever Confirmed|0                           |1

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-04-18 20:14:42 UTC ---
You could bzip2, gzip or xz it before attaching.

Anyway, bisecting and delta reducing now (the latter probably overnight, as
this is using boost).

[Bug c++/57000] 4.8 Regression?, ICE with -Ofast and -frounding-math

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57000

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-04-18 20:37:17 UTC ---
Started with http://gcc.gnu.org/r191925 .

[Bug c++/57000] [4.8/4.9 Regression] ICE with -Ofast and -frounding-math

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57000

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.8.1
            Summary|4.8 Regression?, ICE with   |[4.8/4.9 Regression] ICE
                   |-Ofast and -frounding-math  |with -Ofast and
                   |                            |-frounding-math

[Bug c++/57000] [4.8/4.9 Regression] ICE with -Ofast and -frounding-math

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57000

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-04-19 07:29:35 UTC ---
Created attachment 29901
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=29901
pr57000.ii

Somewhat reduced testcase (but sitll quite large).

[Bug c++/57000] [4.8/4.9 Regression] ICE with -Ofast and -frounding-math

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57000

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-04-19 07:44:54 UTC ---
reassoc clearly doesn't adjust the virtual ops, not sure if just on the pow*
handling, or everywhere, it expects the math builtin calls to have no vops.
As the vops are missing just with -ffast-math, guess this is just about the
flag_unsafe_math_operations guarded code, i.e. pow and powi handling.

So, either we should give up also if the builtin pow/powi etc. has vops, or
request updating of vops if we see any pow* with vops and modify it, or update
it ourselves.

[Bug c++/57000] [4.8/4.9 Regression] ICE with -Ofast and -frounding-math

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57000

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

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

--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> 2013-04-19 08:30:19 UTC ---
Mine.

[Bug tree-optimization/57000] [4.8 Regression] ICE with -Ofast and -frounding-math

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57000

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c++                         |tree-optimization
      Known to work|                            |4.7.3, 4.9.0
            Summary|[4.8/4.9 Regression] ICE    |[4.8 Regression] ICE with
                   |with -Ofast and             |-Ofast and -frounding-math
                   |-frounding-math             |
      Known to fail|                            |4.8.0

--- Comment #8 from Richard Biener <rguenth at gcc dot gnu.org> 2013-04-19 10:16:09 UTC ---
Author: rguenth
Date: Fri Apr 19 10:15:15 2013
New Revision: 198087

URL: http://gcc.gnu.org/viewcvs?rev=198087&root=gcc&view=rev
Log:
2013-04-19  Richard Biener  <rguenther@suse.de>

    PR tree-optimization/57000
    * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.

    * gcc.dg/tree-ssa/reassoc-27.c: New testcase.

Added:
    trunk/gcc/testsuite/gcc.dg/tree-ssa/reassoc-27.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/tree-ssa-reassoc.c

[Bug tree-optimization/57000] [4.8 Regression] ICE with -Ofast and -frounding-math

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57000

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
      Known to work|                            |4.8.1
         Resolution|                            |FIXED

--- Comment #9 from Richard Biener <rguenth at gcc dot gnu.org> 2013-05-07 08:12:17 UTC ---
Author: rguenth
Date: Tue May  7 08:11:46 2013
New Revision: 198665

URL: http://gcc.gnu.org/viewcvs?rev=198665&root=gcc&view=rev
Log:
2013-05-07  Richard Biener  <rguenther@suse.de>

    Backport from mainline
    2013-05-06  Richard Biener  <rguenther@suse.de>

    PR tree-optimization/57185
    * tree-parloops.c (add_field_for_reduction): Handle anonymous
    SSA names properly.

    * gcc.dg/autopar/pr57185.c: New testcase.

    2013-04-19  Richard Biener  <rguenther@suse.de>

    PR tree-optimization/57000
    * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.

    * gcc.dg/tree-ssa/reassoc-27.c: New testcase.

Added:
    branches/gcc-4_8-branch/gcc/testsuite/gcc.dg/autopar/pr57185.c
    branches/gcc-4_8-branch/gcc/testsuite/gcc.dg/tree-ssa/reassoc-27.c
Modified:
    branches/gcc-4_8-branch/gcc/ChangeLog
    branches/gcc-4_8-branch/gcc/testsuite/ChangeLog
    branches/gcc-4_8-branch/gcc/tree-parloops.c