[Bug c++/25411] New: Optimization with -O2 -finline-functions gives wrong code

[Bug c++/25411] New: Optimization with -O2 -finline-functions gives wrong code

[hans dot ekkehard dot plesser at umb dot no]

A test for functor expressions in the Vigra library testsuite gives erroneous
results if compiled with -O2 -finline-functions.  No problem occurs if either
option is dropped.  The error occurs as well with g++ 4.0.0, but not with
3.4.3. The error occurs under Tru64 V5.1B on an AlphaServer, but not under
Linux on i686 (g++ 4.0.0).

The program in question is

#include <iostream>
#include "vigra/functorexpression.hxx"
#include "vigra/rgbvalue.hxx"

using namespace vigra::functor;

int main()
{
  vigra::RGBValue<int> data[] = { vigra::RGBValue<int>(0),
vigra::RGBValue<int>(1) }; 
  vigra::RGBValue<int> sum(0);
  int count = 0;

  for ( int j = 0 ; j < 2 ; ++j )
    std::cerr << "data[" << j << "]: " << data[j] << std::endl;

  std::for_each(data, data+2, 
                (
                 Var(count) += Param(1) ,
                 Var(sum) += Arg1() 
                 ));

  std::cerr << "count: " << count 
            << "\nsum: " << sum 
            << std::endl;

  return 0;
}

The expected output is

data[0]: (0, 0, 0)
data[1]: (1, 1, 1)
count: 2
sum: (1, 1, 1)

If compiled with -O2 -finline-functions, the following is produced:

data[0]: (0, 0, 0)
data[1]: (1, 1, 1)
count: 0
sum: (3, 1, 1)

If the two items in the comma-expression that forms the third argument of
std::for_each() are exchanged, correct code is produced with the
-finline-functions option in place.

I will be happy to supply header files or the preprocessor output on request.

Hans E Plesser


-- 
           Summary: Optimization with -O2 -finline-functions gives wrong
                    code
           Product: gcc
           Version: 4.0.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: hans dot ekkehard dot plesser at umb dot no
 GCC build triplet: alphaev7-dec-osf5.1b
  GCC host triplet: alphaev7-dec-osf5.1b
GCC target triplet: alphaev7-dec-osf5.1b


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25411

[Bug c++/25411] Optimization with -O2 -finline-functions gives wrong code

[hans dot ekkehard dot plesser at umb dot no]

------- Comment #1 from hans dot ekkehard dot plesser at umb dot no  2005-12-14 13:45 -------
Created an attachment (id=10485)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=10485&action=view)
Preprocessed reproducer for optimize bug

This is the complete preprocessed source code for bug 25411.

g++ -O2 -finline-functions Bug25411.ii -o Bug25411

yields an executable that causes the erroneous output reported under Tru64.

Hans E Plesser


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25411

[Bug c++/25411] Optimization with -O2 -finline-functions gives wrong code

[rguenth at gcc dot gnu dot org]

------- Comment #2 from rguenth at gcc dot gnu dot org  2005-12-14 15:54 -------
Cannot reproduce with gcc (GCC) 4.1.0 20050916 (experimental) on x86_64-linux.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25411

[Bug c++/25411] Optimization with -O2 -finline-functions gives wrong code

[falk at debian dot org]

------- Comment #3 from falk at debian dot org  2005-12-15 08:41 -------
I cannot reproduce this on alphaev68-linux with 4.0.3 20051201 (prerelease),
4.1.0 20051124 (prerelease), or 4.2.0 20051124 (experimental). Could you
perhaps try a newer version (4.0.2) or even better a recent snapshot?


-- 

falk at debian dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25411

[Bug c++/25411] Optimization with -O2 -finline-functions gives wrong code

[hans dot ekkehard dot plesser at umb dot no]

------- Comment #4 from hans dot ekkehard dot plesser at umb dot no  2005-12-16 11:50 -------
(In reply to comment #3)
> I cannot reproduce this on alphaev68-linux with 4.0.3 20051201 (prerelease),
> 4.1.0 20051124 (prerelease), or 4.2.0 20051124 (experimental). Could you
> perhaps try a newer version (4.0.2) or even better a recent snapshot?
> 

The bug occurs under release version 4.0.2 on alphaev7-dec-osf5.1b.

It does NOT occur under 
gcc version 4.0.3 20051208 (prerelease)
gcc version 4.1.0 20051209 (prerelease)

I have not managed a complete build of 4.2.0.

Hans


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25411

[Bug c++/25411] Optimization with -O2 -finline-functions gives wrong code

[falk at debian dot org]

------- Comment #5 from falk at debian dot org  2005-12-16 21:34 -------
(In reply to comment #4)
> It does NOT occur under 
> gcc version 4.0.3 20051208 (prerelease)

Okay, so let's close it, then.


-- 

falk at debian dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25411

[Bug c++/25411] Optimization with -O2 -finline-functions gives wrong code

[pinskia at gcc dot gnu dot org]

-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.0.3


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25411