[Bug fortran/55789] New: Needless realloc

[Bug fortran/55789] New: Needless realloc

[Joost.VandeVondele at mat dot ethz.ch]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

             Bug #: 55789
           Summary: Needless realloc
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: Joost.VandeVondele@mat.ethz.ch


In trying to understand some memory leak reported by mtrace (which turns out, I
think, to be a bug in glibc 
http://sourceware.org/bugzilla/show_bug.cgi?id=14981), I noticed that the
following code

MODULE M1
CONTAINS
  SUBROUTINE cp_1d_i4_sort(arr)
      INTEGER(kind=4), DIMENSION(:), &
        INTENT(inout)                          :: arr
      arr = (/ (i, i = 1, SIZE(arr)) /)
  END SUBROUTINE
END MODULE M1

PROGRAM TEST
  USE M1
  INTEGER :: arr(1)
  INTERFACE
    SUBROUTINE mtrace() BIND(C,name="mtrace")
    END SUBROUTINE
  END INTERFACE
  INTERFACE
    SUBROUTINE muntrace() BIND(C,name="muntrace")
    END SUBROUTINE
  END INTERFACE
  CALL mtrace()
  CALL cp_1d_i4_sort(arr)
  CALL muntrace()
END

performs both an allocate and a reallocate in cp_1d_i4_sort.

While I understand that a temporary is generated... something seems
inconsistent with the size computed, and the reallocate triggers. This seems
needlessly inefficient.

There seem to be some vaguely related issues.

1) a glibc mtrace bug is triggered by having a zero sized array as an argument.
In this case the realloc is zero sized, which might be equivalent to a free
(also that appears to be not un-controversial, seemingly some difference betwee
n C99 and glibc)

2) there is http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49241 which is a memory
leak in a similar situation.

[Bug fortran/55789] Needless realloc

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

Tobias Burnus <burnus at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2012-12-22
                 CC|                            |burnus at gcc dot gnu.org
     Ever Confirmed|0                           |1

--- Comment #1 from Tobias Burnus <burnus at gcc dot gnu.org> 2012-12-22 18:24:45 UTC ---
(In reply to comment #0)
> In trying to understand some memory leak reported by mtrace (which turns out, 
> I think, to be a bug in glibc 
> http://sourceware.org/bugzilla/show_bug.cgi?id=14981)

Using GLIBC 2.17, no memleak is reported here for neither the C program nor
this PR's Fortran code.


> I noticed that the following code
> performs both an allocate and a reallocate in cp_1d_i4_sort.


Confirmed. With GCC 4.5, only a simple malloc is done; ltrace shows:
  malloc(4)

but starting from GCC 4.6 both malloc and realloc are invoked; ltrace has:
  malloc(4)
  realloc(0xafbf90, 8)


With 4.8, one has (slightly simplified):

    ubound.0 = (arr->dim[0].ubound - arr->dim[0].lbound) + 1;
    atmp.4.dim[0].lbound = 0;
    atmp.4.dim[0].ubound = ubound.0-1;
    D.1907 = ubound.0 <= 0 ? 0 : ubound.0 * 4;
    atmp.4.data = __builtin_malloc (MAX_EXPR <D.1907, 1>);
...
    D.1919 = MAX_EXPR<(D.1918->dim[0].ubound - D.1918->dim[0].lbound) + 1, 0>;
      shadow_loopvar.7 = 1;
      atmp.4.dim[0].ubound = atmp.4.dim[0].ubound
              + MAX_EXPR <(D.1919 - shadow_loopvar.7) + 1, 0>;
      atmp.4.data = __builtin_realloc (atmp.4.data,
                                       (atmp.4.dim[0].ubound + 1) * 4);
      while (1)
        {
          if (shadow_loopvar.7 > D.1919) goto L.1;
          (*atmp.4.data)[offset.5] = shadow_loopvar.7;
          offset.5 = offset.5 + 1;
          shadow_loopvar.7 = shadow_loopvar.7 + 1;
        }
      L.1:;

[Bug fortran/55789] [4.6/4.7/4.8 Regression] Needless realloc with array constructor.

[tkoenig at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

Thomas Koenig <tkoenig at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tkoenig at gcc dot gnu.org
   Target Milestone|---                         |4.6.4
            Summary|Needless realloc            |[4.6/4.7/4.8 Regression]
                   |                            |Needless realloc with array
                   |                            |constructor.

--- Comment #2 from Thomas Koenig <tkoenig at gcc dot gnu.org> 2013-01-01 14:25:27 UTC ---
Looks like a regression, then.

[Bug fortran/55789] [4.6/4.7/4.8 Regression] Needless realloc with array constructor.

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P4

[Bug fortran/55789] [4.6/4.7/4.8 Regression] Needless realloc with array constructor.

[pault at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

Paul Thomas <pault at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pault at gcc dot gnu.org
         AssignedTo|unassigned at gcc dot       |pault at gcc dot gnu.org
                   |gnu.org                     |

--- Comment #3 from Paul Thomas <pault at gcc dot gnu.org> 2013-01-12 16:05:27 UTC ---
I'll take this one on - see today's posting to the list.

Cheers

Paul

[Bug fortran/55789] [4.6/4.7/4.8 Regression] Needless realloc with array constructor.

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

--- Comment #4 from Tobias Burnus <burnus at gcc dot gnu.org> 2013-01-16 22:06:37 UTC ---
(In reply to comment #3)
> I'll take this one on - see today's posting to the list.

Approved patch: http://gcc.gnu.org/ml/fortran/2013-01/msg00102.html

[Bug fortran/55789] [4.6/4.7/4.8 Regression] Needless realloc with array constructor.

[pault at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

--- Comment #5 from Paul Thomas <pault at gcc dot gnu.org> 2013-01-27 07:09:12 UTC ---
Author: pault
Date: Sun Jan 27 07:09:06 2013
New Revision: 195492

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=195492
Log:
2013-01-27 Paul Thomas  <pault@gcc.gnu.org>

    PR fortran/55789
    PR fortran/56047
    * gfortran.h : Add associate_var to symbol_attr.
    * resolve.c (resolve_assoc_var): Set associate_var attribute.
    If the target class_ok is set, set it for the associate
    variable.
    * check.c (allocatable_check): Associate variables should not
    have the allocatable attribute even if their symbols do.
    * class.c (gfc_build_class_symbol): Symbols with associate_var
    set will always have a good class container.

2013-01-27  Paul Thomas  <pault@gcc.gnu.org>

    PR fortran/55789
    * gfortran.dg/associate_14.f90: New test.

    PR fortran/56047
    * gfortran.dg/associate_13.f90: New test.

Added:
    trunk/gcc/testsuite/gfortran.dg/associate_13.f90
    trunk/gcc/testsuite/gfortran.dg/associate_14.f90
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/check.c
    trunk/gcc/fortran/class.c
    trunk/gcc/fortran/gfortran.h
    trunk/gcc/fortran/resolve.c
    trunk/gcc/testsuite/ChangeLog

[Bug fortran/55789] [4.6/4.7/4.8 Regression] Needless realloc with array constructor.

[pault at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

--- Comment #6 from Paul Thomas <pault at gcc dot gnu.org> 2013-01-27 07:37:40 UTC ---
Sorry, the apparent fix in comment#5 is just noise.  The real fix is on its way
today.

Paul

[Bug fortran/55789] [4.6/4.7/4.8 Regression] Needless realloc with array constructor.

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

--- Comment #7 from Jan Hubicka <hubicka at gcc dot gnu.org> 2013-02-05 09:13:59 UTC ---
Author: hubicka
Date: Tue Feb  5 09:13:48 2013
New Revision: 195751

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=195751
Log:
    PR tree-optimization/55789
    * g++.dg/tree-ssa/inline-1.C: Update max-inliner-iterations.
    * g++.dg/tree-ssa/inline-2.C: Update max-inliner-iterations.
    * g++.dg/tree-ssa/inline-3.C: Update max-inliner-iterations.
    * g++.dg/ipa/inline-1.C: New testcase.
    * g++.dg/ipa/inline-2.C: New testcase.
    * g++.dg/ipa/inline-3.C: New testcase.
    * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.


Added:
    trunk/gcc/testsuite/g++.dg/ipa/inline-1.C
    trunk/gcc/testsuite/g++.dg/ipa/inline-2.C
    trunk/gcc/testsuite/g++.dg/ipa/inline-3.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/params.def
    trunk/gcc/testsuite/g++.dg/tree-ssa/inline-1.C
    trunk/gcc/testsuite/g++.dg/tree-ssa/inline-2.C
    trunk/gcc/testsuite/g++.dg/tree-ssa/inline-3.C

[Bug fortran/55789] [4.6/4.7/4.8 Regression] Needless realloc with array constructor.

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

--- Comment #8 from Jan Hubicka <hubicka at gcc dot gnu.org> 2013-02-05 15:24:03 UTC ---
Author: hubicka
Date: Tue Feb  5 15:23:56 2013
New Revision: 195758

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=195758
Log:

    PR tree-optimization/55789
    * g++.dg/tree-ssa/inline-1.C: Update max-inliner-iterations.
    * g++.dg/tree-ssa/inline-2.C: Update max-inliner-iterations.
    * g++.dg/tree-ssa/inline-3.C: Update max-inliner-iterations.
    * g++.dg/ipa/inline-1.C: New testcase.
    * g++.dg/ipa/inline-2.C: New testcase.
    * g++.dg/ipa/inline-3.C: New testcase.

Modified:
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/g++.dg/ipa/inline-1.C
    trunk/gcc/testsuite/g++.dg/tree-ssa/inline-3.C

[Bug fortran/55789] [4.6/4.7/4.8 Regression] Needless realloc with array constructor.

[pault at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

--- Comment #9 from Paul Thomas <pault at gcc dot gnu.org> 2013-02-06 20:20:24 UTC ---
Author: pault
Date: Wed Feb  6 20:20:08 2013
New Revision: 195815

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=195815
Log:
2013-02-06 Paul Thomas  <pault@gcc.gnu.org>

    PR fortran/55789
    * trans-array.c (trans_array_constructor): Remove condition
    'dynamic' = true if the loop ubound is a VAR_DECL.

2013-02-06 Paul Thomas  <pault@gcc.gnu.org>

    PR fortran/55789
    * gfortran.dg/array_constructor_41.f90: New test.

Added:
    trunk/gcc/testsuite/gfortran.dg/array_constructor_41.f90
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/trans-array.c
    trunk/gcc/testsuite/ChangeLog

[Bug fortran/55789] [4.6/4.7 Regression] Needless realloc with array constructor.

[pault at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

Paul Thomas <pault at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[4.6/4.7/4.8 Regression]    |[4.6/4.7 Regression]
                   |Needless realloc with array |Needless realloc with array
                   |constructor.                |constructor.

--- Comment #10 from Paul Thomas <pault at gcc dot gnu.org> 2013-02-06 20:29:13 UTC ---
Fixed on trunk.

I will have to do a bit more work to determine if the same fix works correctly
for 4.6 and 4.7.

Paul

[Bug fortran/55789] [4.6/4.7 Regression] Needless realloc with array constructor.

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

--- Comment #11 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-02-07 10:45:20 UTC ---
Author: jakub
Date: Thu Feb  7 10:45:12 2013
New Revision: 195844

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=195844
Log:
    PR tree-optimization/55789
    * g++.dg/ipa/inline-3.C: Use cleanup-ipa-dump instead of
    cleanup-tree-dump.
    * gcc.dg/tree-ssa/inline-3.c: Add
    --param max-early-inliner-iterations=2 option.

Modified:
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/g++.dg/ipa/inline-3.C
    trunk/gcc/testsuite/gcc.dg/tree-ssa/inline-3.c

[Bug fortran/55789] [4.6/4.7 Regression] Needless realloc with array constructor.

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

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

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

--- Comment #12 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-02-08 11:30:36 UTC ---
The #c7, #c8 and #c11 commits really belong to PR55797.

[Bug fortran/55789] [4.6/4.7 Regression] Needless realloc with array constructor.

[mikael at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

Mikael Morin <mikael at gcc dot gnu.org> changed:

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

--- Comment #13 from Mikael Morin <mikael at gcc dot gnu.org> 2013-03-02 17:22:28 UTC ---
(In reply to comment #10)
> Fixed on trunk.
> 
> I will have to do a bit more work to determine if the same fix works correctly
> for 4.6 and 4.7.
> 
Any news?

[Bug fortran/55789] [4.7 Regression] Needless realloc with array constructor.

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

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

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

--- Comment #14 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-04-12 15:16:44 UTC ---
GCC 4.6.4 has been released and the branch has been closed.

[Bug fortran/55789] [4.7 Regression] Needless realloc with array constructor.

[pault at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

--- Comment #15 from Paul Thomas <pault at gcc dot gnu.org> ---
No, it does not work on 4.7.

I am inclined to say that it should be a WONTFIX.

Cheers

Paul

[Bug fortran/55789] [4.7 Regression] Needless realloc with array constructor.

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

--- Comment #16 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
> No, it does not work on 4.7.

Paul,

Do you remember what was not working. I have applied the patches of r195492 and
r195815 to 4.7.4 r209162 without regression (see
http://gcc.gnu.org/ml/gcc-testresults/2014-04/msg00483.html).

If deemed suitable I can do the back port (I think it is now or never: 4.7.4
will likely be the last 4.7 release).

[Bug fortran/55789] [4.7 Regression] Needless realloc with array constructor.

[dominiq at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

--- Comment #18 from dominiq at gcc dot gnu.org ---
Author: dominiq
Date: Sat May 24 15:45:02 2014
New Revision: 210894

URL: http://gcc.gnu.org/viewcvs?rev=210894&root=gcc&view=rev
Log:
2014-05-24  Dominique d'Humieres <dominiq@lps.ens.fr>

    Backport r195492 and r195815
    2013-01-27  Paul Thomas  <pault@gcc.gnu.org>

    PR fortran/55789
    * gfortran.dg/associate_14.f90: New test.

    PR fortran/56047
    * gfortran.dg/associate_13.f90: New test.

    2013-02-06  Paul Thomas <pault@gcc.gnu.org>

    PR fortran/55789          
    * gfortran.dg/array_constructor_41.f90: New test.


Added:
    branches/gcc-4_7-branch/gcc/testsuite/gfortran.dg/array_constructor_41.f90
    branches/gcc-4_7-branch/gcc/testsuite/gfortran.dg/associate_13.f90
    branches/gcc-4_7-branch/gcc/testsuite/gfortran.dg/associate_14.f90
Modified:
    branches/gcc-4_7-branch/gcc/testsuite/ChangeLog

[Bug fortran/55789] [4.7 Regression] Needless realloc with array constructor.

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55789

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

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

--- Comment #19 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
Test results at r210894 posted at
https://gcc.gnu.org/ml/gcc-testresults/2014-05/msg02224.html. Closing as FIXED.