[Bug c++/67699] Segfault compiling a constexpr array with -fmerge-all-constants -flto

[Bug c++/67699] Segfault compiling a constexpr array with -fmerge-all-constants -flto

[gcc at norgg dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67699

--- Comment #1 from Norgg <gcc at norgg dot org> ---
Created attachment 36381
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36381&action=edit
Reduced testcase with const instead of constexpr

Futher testing by slowriot found that you can change the constexpr to const and
drop the --std switch.

[Bug c++/67699] Segfault compiling a constexpr array with -fmerge-all-constants -flto

[gcc at norgg dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67699

--- Comment #2 from Norgg <gcc at norgg dot org> ---
It seems the -fmerge-all-constants isn't necessary to cause the segfault either
any more, although it was in an earlier larger test case.

Minimal command line to reproduce is now:
$ g++ const_crash.cpp -flto
const_crash.cpp:11:24: internal compiler error: Segmentation fault
 template class foo<int>;
                        ^

[Bug c++/67699] Segfault compiling a const array with -flto

[gcc at norgg dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67699

--- Comment #3 from Norgg <gcc at norgg dot org> ---
Created attachment 36382
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36382&action=edit
Preprocessed file

[Bug lto/67699] ICE (segfault) compiling a const array with -flto

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67699

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
#2  0x000000000070544c in symtab_node::get (decl=0x7ffff68d9b40)
    at /space/rguenther/src/svn/gcc-5-branch/gcc/cgraph.h:365
365         gcc_checking_assert (TREE_CODE (decl) == FUNCTION_DECL
(gdb) l
360       static inline symtab_node *get (const_tree decl)
361       {
362     #ifdef ENABLE_CHECKING
363         /* Check that we are called for sane type of object - functions
364            and static or external variables.  */
365         gcc_checking_assert (TREE_CODE (decl) == FUNCTION_DECL
366                              || (TREE_CODE (decl) == VAR_DECL
367                                  && (TREE_STATIC (decl) || DECL_EXTERNAL
(decl)
368                                      || in_lto_p)));
(gdb) p debug_tree (decl)
 <var_decl 0x7ffff68d9b40 bar
    type <array_type 0x7ffff6a405e8
...
    readonly decl_1 decl_5 DI file t.ii line 2 col 23 size <integer_cst
0x7ffff68cce58 64> unit size <integer_cst 0x7ffff68cce70 8>
    align 32 context <function_decl 0x7ffff6a411b0 foo>>

and we are not in_lto_p (this is cc1plus).  And we're coming from (tataa)

940           create_references (encoder, vnode);
941           /* For proper debug info, we need to ship the origins, too.  */
942           if (DECL_ABSTRACT_ORIGIN (vnode->decl))
943             {
944               varpool_node *origin_node
945                 = varpool_node::get (DECL_ABSTRACT_ORIGIN (vnode->decl));

that's on the GCC 5 branch.  ISTR fixing this on trunk - ah, but only for
the early LTO debug work.

Hmm.  Let's see if we can survive removing that code - it got added during
GCC 4.9 development.  But we never actually stream DECL_ABSTRACT_ORIGIN...

[Bug lto/67699] ICE (segfault) compiling a const array with -flto

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67699

--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
Author: rguenth
Date: Thu Sep 24 11:01:14 2015
New Revision: 228084

URL: https://gcc.gnu.org/viewcvs?rev=228084&root=gcc&view=rev
Log:
2015-09-24  Richard Biener  <rguenther@suse.de>

        PR lto/67699
        * lto-cgraph.c (compute_ltrans_boundary): Do not stream
        abstract origins.

        * g++.dg/pr67699.C: New testcase.

Added:
    trunk/gcc/testsuite/g++.dg/pr67699.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/lto-cgraph.c
    trunk/gcc/testsuite/ChangeLog

[Bug lto/67699] [4.9/5 Regression] ICE (segfault) compiling a const array with -flto

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67699

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
                 CC|                            |hubicka at gcc dot gnu.org
      Known to work|                            |4.8.5, 6.0
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot gnu.org
   Target Milestone|---                         |4.9.4
            Summary|ICE (segfault) compiling a  |[4.9/5 Regression] ICE
                   |const array with -flto      |(segfault) compiling a
                   |                            |const array with -flto
      Known to fail|6.0                         |

--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> ---
Fixed on trunk.  Will backport to GCC 5 if no issues pop up.

[Bug lto/67699] [4.9/5 Regression] ICE (segfault) compiling a const array with -flto

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67699

--- Comment #8 from Richard Biener <rguenth at gcc dot gnu.org> ---
Author: rguenth
Date: Tue Oct  6 10:19:54 2015
New Revision: 228517

URL: https://gcc.gnu.org/viewcvs?rev=228517&root=gcc&view=rev
Log:
2015-10-06  Richard Biener  <rguenther@suse.de>

        Backport from mainline
        2015-09-15  Richard Biener  <rguenther@suse.de>

        PR middle-end/67563
        * gimple-fold.c (gimplify_and_update_call_from_tree): Do not
        transfer EH info from old to new stmt.
        (replace_call_with_value): Likewise.
        (replace_call_with_call_and_fold): Likewise.
        (gimple_fold_builtin_memory_op): Likewise.
        (gimple_fold_builtin_memset): Likewise.
        (gimple_fold_builtin_stpcpy): Likewise.
        (gimple_fold_call): Likewise.

        * gcc.dg/pr67563.c: New testcase.

        2015-09-24  Richard Biener  <rguenther@suse.de>

        PR lto/67699
        * lto-cgraph.c (compute_ltrans_boundary): Do not stream
        abstract origins.

        * g++.dg/pr67699.C: New testcase.

Added:
    branches/gcc-5-branch/gcc/testsuite/g++.dg/pr67699.C
    branches/gcc-5-branch/gcc/testsuite/gcc.dg/pr67563.c
Modified:
    branches/gcc-5-branch/gcc/ChangeLog
    branches/gcc-5-branch/gcc/gimple-fold.c
    branches/gcc-5-branch/gcc/lto-cgraph.c
    branches/gcc-5-branch/gcc/testsuite/ChangeLog