[Bug fortran/108923] New: memory leak of get_intrinsic_dummy_arg result

[Bug fortran/108923] New: memory leak of get_intrinsic_dummy_arg result

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108923

            Bug ID: 108923
           Summary: memory leak of get_intrinsic_dummy_arg result
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: rguenth at gcc dot gnu.org
  Target Milestone: ---

valgrind leak check complains quite often (when building SPEC 2017) like

==8707== 128 bytes in 8 blocks are definitely lost in loss record 750 of 935
==8707==    at 0x4C39571: calloc (in
/usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==8707==    by 0x1C25680: xcalloc (xmalloc.c:164)
==8707==    by 0x73C891: get_intrinsic_dummy_arg (intrinsic.cc:4265)
==8707==    by 0x73C891: sort_actual(char const*, gfc_actual_arglist**,
gfc_intrinsic_arg*, locus*) (intrinsic.cc:4433)
==8707==    by 0x73CD54: check_specific(gfc_intrinsic_sym*, gfc_expr*, int)
(intrinsic.cc:4774)
==8707==    by 0x744F21: gfc_intrinsic_func_interface(gfc_expr*, int)
(intrinsic.cc:5032)
==8707==    by 0x79BD57: resolve_unknown_f (resolve.cc:2998)
==8707==    by 0x79BD57: resolve_function (resolve.cc:3355)
==8707==    by 0x79BD57: gfc_resolve_expr(gfc_expr*) [clone .part.56]
(resolve.cc:7204)
==8707==    by 0x79D469: gfc_resolve_expr (resolve.cc:7171)
==8707==    by 0x79D469: resolve_operator(gfc_expr*) (resolve.cc:4101)
==8707==    by 0x79990F: gfc_resolve_expr(gfc_expr*) [clone .part.56]
(resolve.cc:7197)
==8707==    by 0x7A2274: gfc_resolve_expr (resolve.cc:11893)
==8707==    by 0x7A2274: gfc_resolve_code(gfc_code*, gfc_namespace*)
(resolve.cc:12004)
==8707==    by 0x7A0D4F: gfc_resolve_blocks(gfc_code*, gfc_namespace*)
(resolve.cc:10989)
==8707==    by 0x7A10A8: gfc_resolve_code(gfc_code*, gfc_namespace*)
(resolve.cc:11990)
==8707==    by 0x7A0D4F: gfc_resolve_blocks(gfc_code*, gfc_namespace*)
(resolve.cc:10989)

[Bug fortran/108923] memory leak of get_intrinsic_dummy_arg result

[mikael at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108923

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |mikael at gcc dot gnu.org
   Last reconfirmed|                            |2023-02-24
                 CC|                            |mikael at gcc dot gnu.org
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |ASSIGNED

--- Comment #1 from Mikael Morin <mikael at gcc dot gnu.org> ---
I'm the culprit I guess.
Should be easy to fix anyway.
Taking.

[Bug fortran/108923] memory leak of get_intrinsic_dummy_arg result

[mikael at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108923

--- Comment #2 from Mikael Morin <mikael at gcc dot gnu.org> ---
(In reply to Mikael Morin from comment #1)
> Should be easy to fix anyway.

Not that easy after all.
The following (obvious) fix regresses heavily.

diff --git a/gcc/fortran/expr.cc b/gcc/fortran/expr.cc
index c295721b9d6..73ee50e385c 100644
--- a/gcc/fortran/expr.cc
+++ b/gcc/fortran/expr.cc
@@ -545,6 +545,7 @@ gfc_free_actual_arglist (gfc_actual_arglist *a1)
       a2 = a1->next;
       if (a1->expr)
        gfc_free_expr (a1->expr);
+      free (a1->associated_dummy);
       free (a1);
       a1 = a2;
     }

[Bug fortran/108923] memory leak of get_intrinsic_dummy_arg result

[mikael at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108923

--- Comment #3 from Mikael Morin <mikael at gcc dot gnu.org> ---
Here is a small reproducer by the way:

program p
  implicit none
  call s(0)
contains
  subroutine s(i)
    integer :: i
  end subroutine s
end program p

[Bug fortran/108923] memory leak of get_intrinsic_dummy_arg result

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108923

--- Comment #4 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Mikael Morin <mikael@gcc.gnu.org>:

https://gcc.gnu.org/g:24c9edfa73632276d7698c103f35833f29804d98

commit r13-6337-g24c9edfa73632276d7698c103f35833f29804d98
Author: Mikael Morin <mikael@gcc.gnu.org>
Date:   Fri Feb 24 22:11:17 2023 +0100

    fortran: Plug leak of associated_dummy memory. [PR108923]

    This fixes a memory leak by accompanying the release of
    gfc_actual_arglist elements' memory with a release of the
    associated_dummy field memory (if allocated).
    Actual argument copy is adjusted as well so that each copy can free
    its field independently.

            PR fortran/108923

    gcc/fortran/ChangeLog:

            * expr.cc (gfc_free_actual_arglist): Free associated_dummy
            memory.
            (gfc_copy_actual_arglist): Make a copy of the associated_dummy
            field if it is set in the original element.

[Bug fortran/108923] memory leak of get_intrinsic_dummy_arg result

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108923

anlauf at gcc dot gnu.org changed:

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

--- Comment #5 from anlauf at gcc dot gnu.org ---
At r13-6337-g24c9edfa73632276d7698c103f35833f29804d98 is still get for
testcase

  gcc/testsuite/gfortran.dg/predcom-1.f


==2885== 32 bytes in 2 blocks are definitely lost in loss record 8 of 354
==2885==    at 0x4C39571: calloc (in
/usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==2885==    by 0x2077B44: xcalloc (xmalloc.c:164)
==2885==    by 0x90FE52: get_intrinsic_dummy_arg (intrinsic.cc:4265)
==2885==    by 0x90FE52: sort_actual(char const*, gfc_actual_arglist**,
gfc_intrinsic_arg*, locus*) (intrinsic.cc:4433)
==2885==    by 0x9100E8: check_specific(gfc_intrinsic_sym*, gfc_expr*, int)
(intrinsic.cc:4774)
==2885==    by 0x918D29: gfc_intrinsic_func_interface(gfc_expr*, int)
(intrinsic.cc:5032)
==2885==    by 0x97DB9D: resolve_generic_f0 (resolve.cc:2772)
==2885==    by 0x97DB9D: resolve_generic_f (resolve.cc:2789)
==2885==    by 0x97DB9D: resolve_function (resolve.cc:3347)
==2885==    by 0x97DB9D: gfc_resolve_expr(gfc_expr*) [clone .part.0]
(resolve.cc:7211)
==2885==    by 0x980B4F: gfc_resolve_expr (resolve.cc:7177)
==2885==    by 0x980B4F: resolve_operator(gfc_expr*) (resolve.cc:4092)
==2885==    by 0x97BC57: gfc_resolve_expr(gfc_expr*) [clone .part.0]
(resolve.cc:7204)
==2885==    by 0x985AF8: gfc_resolve_expr (resolve.cc:7177)
==2885==    by 0x985AF8: gfc_resolve_code(gfc_code*, gfc_namespace*)
(resolve.cc:12011)
==2885==    by 0x989425: gfc_resolve_blocks(gfc_code*, gfc_namespace*)
(resolve.cc:10996)
==2885==    by 0x985AAD: gfc_resolve_code(gfc_code*, gfc_namespace*)
(resolve.cc:11997)
==2885==    by 0x98921D: resolve_codes(gfc_namespace*) (resolve.cc:17670)

[Bug fortran/108923] memory leak of get_intrinsic_dummy_arg result

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108923

--- Comment #6 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Mikael Morin <mikael@gcc.gnu.org>:

https://gcc.gnu.org/g:5c638095e7e0fa4de4e4f7326384a86830b25732

commit r13-6346-g5c638095e7e0fa4de4e4f7326384a86830b25732
Author: Mikael Morin <mikael@gcc.gnu.org>
Date:   Sat Feb 25 21:37:46 2023 +0100

    fortran: Reuse associated_dummy memory if previously allocated [PR108923]

    This avoids making the associted_dummy field point to a new memory chunk
    if it's already pointing somewhere, in which case doing so would leak the
    previously allocated chunk.

            PR fortran/108923

    gcc/fortran/ChangeLog:

            * intrinsic.cc (get_intrinsic_dummy_arg,
            set_intrinsic_dummy_arg): Rename the former to the latter.
            Remove the return value, add a reference to the lhs as argument,
            and do the pointer assignment inside the function.  Don't do
            it if the pointer is already non-NULL.
            (sort_actual): Update caller.

[Bug fortran/108923] memory leak of get_intrinsic_dummy_arg result

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108923

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

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

--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> ---
Building the benchmark set fprate_any_fortran of SPEC CPU 2017 there's no leak
from an allocation in get_intrinsic_dummy_arg anymore.

[Bug fortran/108923] memory leak of get_intrinsic_dummy_arg result

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108923

--- Comment #8 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Mikael Morin
<mikael@gcc.gnu.org>:

https://gcc.gnu.org/g:cd65c727fd6d5a252964dbeb3a735daa0a521a89

commit r12-9243-gcd65c727fd6d5a252964dbeb3a735daa0a521a89
Author: Mikael Morin <mikael@gcc.gnu.org>
Date:   Fri Feb 24 22:11:17 2023 +0100

    fortran: Plug leak of associated_dummy memory. [PR108923]

    This fixes a memory leak by accompanying the release of
    gfc_actual_arglist elements' memory with a release of the
    associated_dummy field memory (if allocated).
    Actual argument copy is adjusted as well so that each copy can free
    its field independently.

            PR fortran/108923

    gcc/fortran/ChangeLog:

            * expr.cc (gfc_free_actual_arglist): Free associated_dummy
            memory.
            (gfc_copy_actual_arglist): Make a copy of the associated_dummy
            field if it is set in the original element.

    (cherry picked from commit 24c9edfa73632276d7698c103f35833f29804d98)

[Bug fortran/108923] memory leak of get_intrinsic_dummy_arg result

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108923

--- Comment #9 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Mikael Morin
<mikael@gcc.gnu.org>:

https://gcc.gnu.org/g:7b0c9631ccb492c4457a91dbb91e7a5bcd0bc2a6

commit r12-9244-g7b0c9631ccb492c4457a91dbb91e7a5bcd0bc2a6
Author: Mikael Morin <mikael@gcc.gnu.org>
Date:   Sat Feb 25 21:37:46 2023 +0100

    fortran: Reuse associated_dummy memory if previously allocated [PR108923]

    This avoids making the associted_dummy field point to a new memory chunk
    if it's already pointing somewhere, in which case doing so would leak the
    previously allocated chunk.

            PR fortran/108923

    gcc/fortran/ChangeLog:

            * intrinsic.cc (get_intrinsic_dummy_arg,
            set_intrinsic_dummy_arg): Rename the former to the latter.
            Remove the return value, add a reference to the lhs as argument,
            and do the pointer assignment inside the function.  Don't do
            it if the pointer is already non-NULL.
            (sort_actual): Update caller.

    (cherry picked from commit 5c638095e7e0fa4de4e4f7326384a86830b25732)