[Bug tui/30646] New: [gdb/tui] intrusive_list.h:458: internal-error: erase_element: Assertion `elem_node->prev != INTRUSIVE_LIST_UNLINKED_VALUE' failed.

[Bug tui/30646] New: [gdb/tui] intrusive_list.h:458: internal-error: erase_element: Assertion `elem_node->prev != INTRUSIVE_LIST_UNLINKED_VALUE' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30646

            Bug ID: 30646
           Summary: [gdb/tui] intrusive_list.h:458: internal-error:
                    erase_element: Assertion `elem_node->prev !=
                    INTRUSIVE_LIST_UNLINKED_VALUE' failed.
           Product: gdb
           Version: HEAD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: tui
          Assignee: unassigned at sourceware dot org
          Reporter: vries at gcc dot gnu.org
  Target Milestone: ---

I'm running into:
...
FAIL: gdb.python/tui-window-factory.exp: call __del__ at exit: quit (GDB
internal error)
ERROR: Could not resync from internal error (eof)
...

In more detail:
...
(gdb) PASS: gdb.python/tui-window-factory.exp: call __del__ at exit: python
register_window_factory('msg_2')
quit^M
Entering TestWindowFactory.__del__: msg_2^M
/data/vries/gdb/src/gdb/../gdbsupport/intrusive_list.h:458: internal-error:
erase_element: Assertion `elem_node->prev != INTRUSIVE_LIST_UNLINKED_VALUE'
failed.^M
...

This is with an -O2 -flto build, not sure yet if that matters.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug tui/30646] [gdb/tui] intrusive_list.h:458: internal-error: erase_element: Assertion `elem_node->prev != INTRUSIVE_LIST_UNLINKED_VALUE' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30646

--- Comment #1 from Tom de Vries <vries at gcc dot gnu.org> ---
Created attachment 14977
  --> https://sourceware.org/bugzilla/attachment.cgi?id=14977&action=edit
gdb.log

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug tui/30646] [gdb/tui] intrusive_list.h:458: internal-error: erase_element: Assertion `elem_node->prev != INTRUSIVE_LIST_UNLINKED_VALUE' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30646

--- Comment #2 from Tom de Vries <vries at gcc dot gnu.org> ---
Backtrace with c++filt:
----- Backtrace -----
FAIL: gdb.python/tui-window-factory.exp: call __del__ at exit: quit (GDB
internal error)
Resyncing due to internal error.
0x90e1e2 gdb_internal_backtrace_1
        /data/vries/gdb/src/gdb/bt-utils.c:122
0x90e1e2 gdb_internal_backtrace() [clone .part.5]
        /data/vries/gdb/src/gdb/bt-utils.c:168
0xde05c2 gdb_internal_backtrace
        /data/vries/gdb/src/gdb/bt-utils.c:156
0xde05c2 internal_vproblem(internal_problem*, char const*, int, char const*,
__va_list_tag*) [clone .lto_priv.3035]
        /data/vries/gdb/src/gdb/utils.c:396
0xe98d2d internal_verror
        /data/vries/gdb/src/gdb/utils.c:476
0xe98d2d internal_error_loc(char const*, int, char const*, ...)
        /data/vries/gdb/src/gdbsupport/errors.cc:58
0x9c8f78 erase_element
        /data/vries/gdb/src/gdb/../gdbsupport/intrusive_list.h:458
0x9c8f78 intrusive_list<gdbpy_tui_window_maker,
intrusive_base_node<gdbpy_tui_window_maker>
>::erase(intrusive_list_iterator<gdbpy_tui_window_maker,
intrusive_base_node<gdbpy_tui_window_maker> >)
        /data/vries/gdb/src/gdb/../gdbsupport/intrusive_list.h:497
0x9c8f78 __base_dtor 
        /data/vries/gdb/src/gdb/python/py-tui.c:342
0x9d8235
std::_Function_base::_Base_manager<gdbpy_tui_window_maker>::_M_destroy(std::_Any_data&,
std::integral_constant<bool, false>) [clone .isra.37]
        /usr/include/c++/7/bits/std_function.h:207
0x9d8235 _M_manager
        /usr/include/c++/7/bits/std_function.h:231
0xb12859 std::_Function_base::~_Function_base()
        /usr/include/c++/7/bits/std_function.h:276
0xb12859 std::function<void ()>::~function()
        /usr/include/c++/7/bits/std_function.h:389
0xb12859 selftests::selftest::~selftest() [clone .lto_priv.3137]
        /data/vries/gdb/src/gdb/../gdbsupport/selftest.h:36
0xb12859 void
__gnu_cxx::new_allocator<std::pair<std::__cxx11::basic_string<char,
std::char_traits<char>, std::allocator<char> > const,
std::function<tui_win_info* (char const*)> >
>::destroy<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>,
std::allocator<char> > const, std::function<tui_win_info* (char const*)> >
>(std::pair<std::__cxx11::basic_string<char, std::char_traits<char>,
std::allocator<char> > const, std::function<tui_win_info* (char const*)> >*)
        /usr/include/c++/7/ext/new_allocator.h:140
0xb12859 void
std::allocator_traits<std::allocator<std::pair<std::__cxx11::basic_string<char,
std::char_traits<char>, std::allocator<char> > const,
std::function<tui_win_info* (char const*)> > >
>::destroy<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>,
std::allocator<char> > const, std::function<tui_win_info* (char const*)> >
>(std::allocator<std::pair<std::__cxx11::basic_string<char,
std::char_traits<char>, std::allocator<char> > const,
std::function<tui_win_info* (char const*)> > >&,
std::pair<std::__cxx11::basic_string<char, std::char_traits<char>,
std::allocator<char> > const, std::function<tui_win_info* (char const*)> >*)
        /usr/include/c++/7/bits/alloc_traits.h:487
0xb12859
std::__detail::_Hashtable_alloc<std::allocator<std::__detail::_Hash_node<std::pair<std::__cxx11::basic_string<char,
std::char_traits<char>, std::allocator<char> > const,
std::function<tui_win_info* (char const*)> >, true> >
>::_M_deallocate_node(std::__detail::_Hash_node<std::pair<std::__cxx11::basic_string<char,
std::char_traits<char>, std::allocator<char> > const,
std::function<tui_win_info* (char const*)> >, true>*)
        /usr/include/c++/7/bits/hashtable_policy.h:2084
0xb12859
std::__detail::_Hashtable_alloc<std::allocator<std::__detail::_Hash_node<std::pair<std::__cxx11::basic_string<char,
std::char_traits<char>, std::allocator<char> > const,
std::function<tui_win_info* (char const*)> >, true> >
>::_M_deallocate_nodes(std::__detail::_Hash_node<std::pair<std::__cxx11::basic_string<char,
std::char_traits<char>, std::allocator<char> > const,
std::function<tui_win_info* (char const*)> >, true>*) [clone .isra.137]
        /usr/include/c++/7/bits/hashtable_policy.h:2097
0xb12859 clear
        /usr/include/c++/7/bits/hashtable.h:2032
0xb12859 __base_dtor 
        /usr/include/c++/7/bits/hashtable.h:1358
0xb12859 __base_dtor 
        /usr/include/c++/7/bits/unordered_map.h:101
0x7f8ce4982a98 __run_exit_handlers
        /usr/src/debug/glibc-2.31-150300.46.1.x86_64/stdlib/exit.c:108
0x7f8ce4982c29 __GI_exit
        /usr/src/debug/glibc-2.31-150300.46.1.x86_64/stdlib/exit.c:139
0x8f26d2 quit_force(int*, int)
        /data/vries/gdb/src/gdb/top.c:1814
0x905e59 quit_command(char const*, int)
        /data/vries/gdb/src/gdb/cli/cli-cmds.c:508
0x906eb7 cmd_func(cmd_list_element*, char const*, int)
        /data/vries/gdb/src/gdb/cli/cli-decode.c:2735
0x8db933 execute_command(char const*, int)
        /data/vries/gdb/src/gdb/top.c:574
0xcbf09e command_handler(char const*)
        /data/vries/gdb/src/gdb/event-top.c:552
0xcbf13f command_line_handler(std::unique_ptr<char, gdb::xfree_deleter<char>
>&&)
        /data/vries/gdb/src/gdb/event-top.c:788
0xc8b7db gdb_rl_callback_handler
        /data/vries/gdb/src/gdb/event-top.c:259
0xa9fced rl_callback_read_char
        /data/vries/gdb/src/readline/readline/callback.c:290
0xca5662 gdb_rl_callback_read_char_wrapper_noexcept
        /data/vries/gdb/src/gdb/event-top.c:195
0xca576f gdb_rl_callback_read_char_wrapper(void*) [clone .lto_priv.11376]
        /data/vries/gdb/src/gdb/event-top.c:234
0xb159af stdin_event_handler
        /data/vries/gdb/src/gdb/ui.c:155
0xea6704 gdb_wait_for_event(int) [clone .lto_priv.2883]
        /data/vries/gdb/src/gdbsupport/event-loop.cc:694
0xec2b4b gdb_do_one_event(int) [clone .constprop.2391]
        /data/vries/gdb/src/gdbsupport/event-loop.cc:264
0xda5114 start_event_loop
        /data/vries/gdb/src/gdb/main.c:412
0xda5114 captured_command_loop() [clone .lto_priv.13401]
        /data/vries/gdb/src/gdb/main.c:476
0x431274 captured_main
        /data/vries/gdb/src/gdb/main.c:1320
0x431274 gdb_main
        /data/vries/gdb/src/gdb/main.c:1339
0x431274 main
        /data/vries/gdb/src/gdb/gdb.c:32
---------------------

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug tui/30646] [gdb/tui] intrusive_list.h:458: internal-error: erase_element: Assertion `elem_node->prev != INTRUSIVE_LIST_UNLINKED_VALUE' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30646

--- Comment #3 from Tom de Vries <vries at gcc dot gnu.org> ---
(In reply to Tom de Vries from comment #0)
> This is with an -O2 -flto build, not sure yet if that matters.

More specifically -O2 -flto=auto.

Does not reproduce with -O0 or -O2 build.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug tui/30646] [gdb/tui] intrusive_list.h:458: internal-error: erase_element: Assertion `elem_node->prev != INTRUSIVE_LIST_UNLINKED_VALUE' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30646

--- Comment #4 from Tom de Vries <vries at gcc dot gnu.org> ---
Maybe similar to the issue handled by commit 995a34b1772 ("Guard against
frame.c destructors running before frame-info.c's").

In that case, maybe likewise adding an is_linked test in
gdbpy_tui_window_maker::~gdbpy_tui_window_maker () could work.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug tui/30646] [gdb/tui] intrusive_list.h:458: internal-error: erase_element: Assertion `elem_node->prev != INTRUSIVE_LIST_UNLINKED_VALUE' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30646

Tom de Vries <vries at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |aburgess at redhat dot com

--- Comment #5 from Tom de Vries <vries at gcc dot gnu.org> ---
Bisects to commit d27ff777c65 ("gdb/tui: don't leak the known_window_types
map").

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug tui/30646] [gdb/tui] intrusive_list.h:458: internal-error: erase_element: Assertion `elem_node->prev != INTRUSIVE_LIST_UNLINKED_VALUE' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30646

--- Comment #6 from Tom de Vries <vries at gcc dot gnu.org> ---
(In reply to Tom de Vries from comment #4)
> Maybe similar to the issue handled by commit 995a34b1772 ("Guard against
> frame.c destructors running before frame-info.c's").
> 
> In that case, maybe likewise adding an is_linked test in
> gdbpy_tui_window_maker::~gdbpy_tui_window_maker () could work.

Confirmed, this fixes it:
...
diff --git a/gdb/python/py-tui.c b/gdb/python/py-tui.c
index 9ce76659052..3bc169c686f 100644
--- a/gdb/python/py-tui.c
+++ b/gdb/python/py-tui.c
@@ -339,7 +339,8 @@ intrusive_list<gdbpy_tui_window_maker>
 gdbpy_tui_window_maker::~gdbpy_tui_window_maker ()
 {
   /* Remove this gdbpy_tui_window_maker from the global list.  */
-  m_window_maker_list.erase (m_window_maker_list.iterator_to (*this));
+  if (is_linked ())
+    m_window_maker_list.erase (m_window_maker_list.iterator_to (*this));

   if (m_constr != nullptr)
     {
...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug tui/30646] [gdb/tui] intrusive_list.h:458: internal-error: erase_element: Assertion `elem_node->prev != INTRUSIVE_LIST_UNLINKED_VALUE' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30646

--- Comment #7 from Tom de Vries <vries at gcc dot gnu.org> ---
Created attachment 14978
  --> https://sourceware.org/bugzilla/attachment.cgi?id=14978&action=edit
tentative patch

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug tui/30646] [gdb/tui] intrusive_list.h:458: internal-error: erase_element: Assertion `elem_node->prev != INTRUSIVE_LIST_UNLINKED_VALUE' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30646

--- Comment #8 from Tom de Vries <vries at gcc dot gnu.org> ---
https://sourceware.org/pipermail/gdb-patches/2023-July/200956.html

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug tui/30646] [gdb/tui] intrusive_list.h:458: internal-error: erase_element: Assertion `elem_node->prev != INTRUSIVE_LIST_UNLINKED_VALUE' failed.

[cvs-commit at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30646

--- Comment #9 from cvs-commit at gcc dot gnu.org <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Tom de Vries <vries@sourceware.org>:

https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;h=c0afd99439fea90c1e93e5add7eac06cf533bb3e

commit c0afd99439fea90c1e93e5add7eac06cf533bb3e
Author: Tom de Vries <tdevries@suse.de>
Date:   Wed Jul 26 12:29:28 2023 +0200

    [gdb/tui] Fix assert in ~gdbpy_tui_window_maker

    In gdb/tui/tui-layout.c, we have:
    ...
    static window_types_map known_window_types;
    ...
    and in gdb/python/py-tui.c:
    ...
      /* A global list of all gdbpy_tui_window_maker objects.  */
      static intrusive_list<gdbpy_tui_window_maker> m_window_maker_list;
    };

    /* See comment in class declaration above.  */

    intrusive_list<gdbpy_tui_window_maker>
      gdbpy_tui_window_maker::m_window_maker_list;
    ...

    With a gdb build with -O0 or -O2, the static destructor calling order seems
to be:
    - first gdb/tui/tui-layout.c,
    - then gdb/python/py-tui.c.

    So when running test-case gdb.python/tui-window-factory.exp, we see the
    following order of events:
    - the destructor for known_window_types is called, which triggers calling
the
      destructor for the only element E of m_window_maker_list.  The destructor
      destroys E, and also removes E from m_window_maker_list, leaving it
empty.
    - the destructor for m_window_maker_list is called.  It's empty, so it's a
nop.

    However, when building gdb with -O2 -flto=auto, the static destructor
calling
    order seems to be reversed.

    Instead, we have these events:
    - the destructor for m_window_maker_list is called.  This doesn't destroy
it's
      only element E, but it does make m_window_maker_list empty.
    - the destructor for known_window_types is called, which triggers calling
the
      destructor for E.  An attempt is done to remove E from
m_window_maker_list,
      but we run into an assertion failure, because the list is empty.

    Fix this by checking is_linked () before attempting to remove from
    m_window_maker_list, similar to how things were addressed in commit
995a34b1772
    ("Guard against frame.c destructors running before frame-info.c's").

    Tested on x86_64-linux.

    PR tui/30646
    Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30646

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug tui/30646] [gdb/tui] intrusive_list.h:458: internal-error: erase_element: Assertion `elem_node->prev != INTRUSIVE_LIST_UNLINKED_VALUE' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30646

Tom de Vries <vries at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
   Target Milestone|---                         |14.1
         Resolution|---                         |FIXED

--- Comment #10 from Tom de Vries <vries at gcc dot gnu.org> ---
Fixed.

-- 
You are receiving this mail because:
You are on the CC list for the bug.