[Bug c++/106925] New: internal compiler error: Segmentation fault

[Bug c++/106925] New: internal compiler error: Segmentation fault

[carlosgalvezp at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106925

            Bug ID: 106925
           Summary: internal compiler error: Segmentation fault
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: carlosgalvezp at gmail dot com
  Target Milestone: ---

The following minimal example leads to a segfault of GCC on latest trunk:

#include <array>

template <typename T, std::size_t size>
struct MyVector
{
    std::array<T, size> data{};
};

template <typename T>
struct Foo final
{
    T a{0};
};

constexpr std::size_t kSize = 1;

void foo(MyVector<Foo<float>, kSize> input = MyVector<Foo<float>, kSize>())
{

}


<source>: In instantiation of 'constexpr MyVector<Foo<float>, 1>::MyVector()':
<source>:17:74:   required from here
<source>:6:25: internal compiler error: Segmentation fault
    6 |     std::array<T, size> data{};
      |                         ^~~~
0x23428ae internal_error(char const*, ...)
        ???:0
0xb9cae7 maybe_splice_retval_cleanup(tree_node*)
        ???:0
0xce4fc2 do_poplevel(tree_node*)
        ???:0
0xce7973 finish_for_stmt(tree_node*)
        ???:0
0xba5019 build_vec_init(tree_node*, tree_node*, tree_node*, bool, int, int,
vec<tree_node*, va_gc, vl_embed>**)
        ???:0
0xd0ee2e expand_vec_init_expr(tree_node*, tree_node*, int, vec<tree_node*,
va_gc, vl_embed>**)
        ???:0
0xd494d1 digest_nsdmi_init(tree_node*, tree_node*, int)
        ???:0
0xbaabe5 get_nsdmi(tree_node*, bool, int)
        ???:0
0xbd282d get_defaulted_eh_spec(tree_node*, int)
        ???:0
0xc9d42a maybe_instantiate_noexcept(tree_node*, int)
        ???:0
0xc9d23a maybe_instantiate_noexcept(tree_node*, int)
        ???:0
0xb85043 mark_used(tree_node*, int)
        ???:0
0xac7bfb build_new_method_call(tree_node*, tree_node*, vec<tree_node*, va_gc,
vl_embed>**, tree_node*, int, tree_node**, int)
        ???:0
0xac9027 build_special_member_call(tree_node*, tree_node*, vec<tree_node*,
va_gc, vl_embed>**, tree_node*, int, int)
        ???:0
0xba3b2c build_value_init(tree_node*, int)
        ???:0
0xd4a73a build_functional_cast(unsigned int, tree_node*, tree_node*, int)
        ???:0
0xc743a7 c_parse_file()
        ???:0
0xdae9f9 c_common_parse_file()
        ???:0
Please submit a full bug report, with preprocessed source (by using
-freport-bug).
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.
Compiler returned: 1

Can be reproduced on Compiler Explorer:
https://godbolt.org/z/fconh9nKa

[Bug c++/106925] [12/13 Regression] ICE in maybe_splice_retval_cleanup at gcc/cp/except.cc:1330 since r12-8066-g4822108e61ab8790

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106925

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|internal compiler error:    |[12/13 Regression] ICE in
                   |Segmentation fault          |maybe_splice_retval_cleanup
                   |                            |at gcc/cp/except.cc:1330
                   |                            |since
                   |                            |r12-8066-g4822108e61ab8790
   Target Milestone|---                         |12.3
   Last reconfirmed|                            |2022-09-13
                 CC|                            |jason at gcc dot gnu.org,
                   |                            |marxin at gcc dot gnu.org
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1

--- Comment #1 from Martin Liška <marxin at gcc dot gnu.org> ---
Since r12-8066-g4822108e61ab8790, reduced test-case:

$ cat pr106925.ii
struct Foo;
template <int _Nm> struct __array_traits { typedef Foo _Type[_Nm]; };
template <int _Nm> struct array {
  typename __array_traits<_Nm>::_Type _M_elems;
};
template <int size> struct MyVector { array<size> data{}; };
struct Foo {
  float a{0};
};
void foo(MyVector<1> = MyVector<1>());

[Bug c++/106925] [12/13 Regression] ICE in maybe_splice_retval_cleanup at gcc/cp/except.cc:1330 since r12-8066-g4822108e61ab8790

[carlosgalvezp at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106925

--- Comment #2 from Carlos Galvez <carlosgalvezp at gmail dot com> ---
Hi, is there any progress on the issue?

[Bug c++/106925] [12/13 Regression] ICE in maybe_splice_retval_cleanup at gcc/cp/except.cc:1330 since r12-8066-g4822108e61ab8790

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106925

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

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

--- Comment #3 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
I think the problem is that since r12-8066, in cxx_eval_vec_init we perform
expand_vec_init_expr while processing the default argument, at which point
start_preparsed_function hasn't yet set current_function_decl. 
expand_vec_init_expr then leads to maybe_splice_retval_cleanup which checks
DECL_CONSTRUCTOR_P (current_function_decl) without checking that
current_function_decl is non-null first.  It seems correct that
current_function_decl is null at that point, so the fix should be just:

--- a/gcc/cp/except.cc
+++ b/gcc/cp/except.cc
@@ -1327,6 +1327,7 @@ maybe_splice_retval_cleanup (tree compound_stmt)
        && current_binding_level->level_chain->kind == sk_function_parms);

   if ((function_body || current_binding_level->kind == sk_try)
+      && current_function_decl
       && !DECL_CONSTRUCTOR_P (current_function_decl)
       && !DECL_DESTRUCTOR_P (current_function_decl)
       && current_retval_sentinel)

[Bug c++/106925] [12/13 Regression] ICE in maybe_splice_retval_cleanup at gcc/cp/except.cc:1330 since r12-8066-g4822108e61ab8790

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106925

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Priority|P3                          |P2
           Assignee|unassigned at gcc dot gnu.org      |mpolacek at gcc dot gnu.org

--- Comment #4 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Well, why don't I go ahead and post it.

[Bug c++/106925] [12/13 Regression] ICE in maybe_splice_retval_cleanup at gcc/cp/except.cc:1330 since r12-8066-g4822108e61ab8790

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106925

--- Comment #5 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The trunk branch has been updated by Marek Polacek <mpolacek@gcc.gnu.org>:

https://gcc.gnu.org/g:3130e70dab1e64a7b014391fe941090d5f3b6b7d

commit r13-3277-g3130e70dab1e64a7b014391fe941090d5f3b6b7d
Author: Marek Polacek <polacek@redhat.com>
Date:   Tue Oct 11 14:16:54 2022 -0400

    c++: ICE with VEC_INIT_EXPR and defarg [PR106925]

    Since r12-8066, in cxx_eval_vec_init we perform expand_vec_init_expr
    while processing the default argument in this test.  At this point
    start_preparsed_function hasn't yet set current_function_decl.
    expand_vec_init_expr then leads to maybe_splice_retval_cleanup which
    checks DECL_CONSTRUCTOR_P (current_function_decl) without checking that
    c_f_d is non-null first.  It seems correct that c_f_d is null here, so
    it seems to me that maybe_splice_retval_cleanup should check c_f_d as
    in the following patch.

            PR c++/106925

    gcc/cp/ChangeLog:

            * except.cc (maybe_splice_retval_cleanup): Check
current_function_decl.
            Make the bool const.

    gcc/testsuite/ChangeLog:

            * g++.dg/cpp0x/initlist-defarg3.C: New test.

[Bug c++/106925] [12/13 Regression] ICE in maybe_splice_retval_cleanup at gcc/cp/except.cc:1330 since r12-8066-g4822108e61ab8790

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106925

--- Comment #6 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Marek Polacek
<mpolacek@gcc.gnu.org>:

https://gcc.gnu.org/g:8794633003318113e03147a727e63f5d55b350ab

commit r12-8829-g8794633003318113e03147a727e63f5d55b350ab
Author: Marek Polacek <polacek@redhat.com>
Date:   Tue Oct 11 14:16:54 2022 -0400

    c++: ICE with VEC_INIT_EXPR and defarg [PR106925]

    Since r12-8066, in cxx_eval_vec_init we perform expand_vec_init_expr
    while processing the default argument in this test.  At this point
    start_preparsed_function hasn't yet set current_function_decl.
    expand_vec_init_expr then leads to maybe_splice_retval_cleanup which
    checks DECL_CONSTRUCTOR_P (current_function_decl) without checking that
    c_f_d is non-null first.  It seems correct that c_f_d is null here, so
    it seems to me that maybe_splice_retval_cleanup should check c_f_d as
    in the following patch.

            PR c++/106925

    gcc/cp/ChangeLog:

            * except.cc (maybe_splice_retval_cleanup): Check
current_function_decl.
            Make the bool const.

    gcc/testsuite/ChangeLog:

            * g++.dg/cpp0x/initlist-defarg3.C: New test.

    (cherry picked from commit 3130e70dab1e64a7b014391fe941090d5f3b6b7d)

[Bug c++/106925] [12/13 Regression] ICE in maybe_splice_retval_cleanup at gcc/cp/except.cc:1330 since r12-8066-g4822108e61ab8790

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106925

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

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

--- Comment #7 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Fixed.

[Bug c++/106925] [12/13 Regression] ICE in maybe_splice_retval_cleanup at gcc/cp/except.cc:1330 since r12-8066-g4822108e61ab8790

[carlosgalvezp at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106925

--- Comment #8 from Carlos Galvez <carlosgalvezp at gmail dot com> ---
Awesome, thanks a lot for the quick fix!