[PATCH v6 0/4] [gdb/tui] Fix crash with debuginfod query

[PATCH v6 0/4] [gdb/tui] Fix crash with debuginfod query

[Tom de Vries]

This patch series fixes two PRs, PR31449 and PR33794.

An initial patch for PR31449 was submitted here [1].  A review comment
reported a problem when using C-x C-a to enable TUI instead of "tui enable",
not fixed by the patch.  This was filed as PR33794.

The initial patch fixes PR31449 by calling a function that will force the
debuginfod query to happen before enabling the TUI.

Changes in v1 [2]:
- change approach:
  - add a layout with only the command window, and activate that layout first,
    making sure that we finalize the initialization of the command window
    before we activate the initial layout

Changes in v2 [3]:
- change approach:
  - add a facility defer_tui_rerender that allows deferral of calling the
    rerender methods (which draws everything but the window borders), and
  - use that facility to postpone rerendering in tui_enable until after io has
    been initialized.

Changes in v3 [4]:
- change approach:
  - add a check at the start of tui_inject_newline_into_command_window that
    avoids the SIGFPE.

Changes in v4 [5]:
- change approach:
  - bail out if debuginfod enabled is set to ask when entering TUI.  This is
    very similar to [1].  Unlike the previous approaches, this does fix
    PR33794 as well.

Changes in v5 [6]:
- fixed some trivial errors in the v4.
- split the v4 patch up into three patches,
- added a patch introducing defaulted_query_auto_answers_p,
- used defaulted_query_auto_answers_p to make the fix less intrusive

Changes in v6:
- committed first patch of v5, a refactor patch
- change approach:
  - revert to v3 approach for PR31449.  The error thrown in tui_enable also
    affected "gdb -tui", and for that scenario could only be avoided with
    "set debuginfod enabled on/off" in .gdbearlyinit.
- describe entire version history in series description, instead of partly in
  the series description and partly in one of the patches

Tested on x86_64-linux.

Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=31449
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=33794

[1] https://sourceware.org/pipermail/gdb-patches/2024-March/207224.html
[2] v1 https://sourceware.org/pipermail/gdb-patches/2026-January/224038.html
[3] v2 https://sourceware.org/pipermail/gdb-patches/2026-January/224103.html
[4] v3 https://sourceware.org/pipermail/gdb-patches/2026-February/225210.html
[5] v4 https://sourceware.org/pipermail/gdb-patches/2026-February/225351.html
[6] v5 https://sourceware.org/pipermail/gdb-patches/2026-March/226009.html

Tom de Vries (4):
  [gdb/tui] Fix crash with debuginfod query (tui enable)
  [gdb] Add debuginfod_enabled_ask_p
  [gdb] Add defaulted_query_auto_answers_p
  [gdb/tui] Fix crash with debuginfod query (C-x C-a)

 gdb/debuginfod-support.c          | 40 ++++++++++---
 gdb/debuginfod-support.h          |  5 ++
 gdb/testsuite/gdb.tui/query-foo.c | 24 ++++++++
 gdb/testsuite/gdb.tui/query.exp   | 98 +++++++++++++++++++++++++++++++
 gdb/tui/tui-io.c                  |  3 +
 gdb/tui/tui.c                     | 10 +++-
 gdb/tui/tui.h                     |  2 +-
 gdb/utils.c                       | 55 +++++++++++++++--
 gdb/utils.h                       |  4 ++
 9 files changed, 224 insertions(+), 17 deletions(-)
 create mode 100644 gdb/testsuite/gdb.tui/query-foo.c
 create mode 100644 gdb/testsuite/gdb.tui/query.exp


base-commit: 19bcbe810d8c49eb05324624c0af72091d38c4a7
-- 
2.51.0

[PATCH v6 1/4] [gdb/tui] Fix crash with debuginfod query (tui enable)

[Tom de Vries]

PR tui/31449 reports a SIGFPE when the debuginfod query happens while enabling
TUI using the "tui enable" command:
...
Thread 1 "gdb" received signal SIGFPE, Arithmetic exception.
0x0000000001021084 in tui_inject_newline_into_command_window ()
    at /data/vries/gdb/src/gdb/tui/tui-io.c:1096
1096          py += px / tui_cmd_win ()->width;
...
due to divide-by-zero because tui_cmd_win ()->width == 0.

The corresponding backtrace is:
...
 (gdb) bt
 #0  0x0000000001021084 in tui_inject_newline_into_command_window ()
     at gdb/tui/tui-io.c:1096
 #1  0x0000000000fe65fd in gdb_readline_wrapper_line (line=...) at gdb/top.c:939
 #2  0x0000000000944eef in gdb_rl_callback_handler (rl=0x2cc865a0 "n")
     at gdb/event-top.c:288
 #3  0x0000000001175779 in rl_callback_read_char ()
     at readline/readline/callback.c:302
 #4  0x0000000000944bc3 in gdb_rl_callback_read_char_wrapper_sjlj ()
     at gdb/event-top.c:197
 #5  0x0000000000944cd4 in gdb_rl_callback_read_char_wrapper_noexcept ()
     at gdb/event-top.c:240
 #6  0x0000000000944d52 in gdb_rl_callback_read_char_wrapper (...)
     at gdb/event-top.c:252
 #7  0x0000000001062352 in stdin_event_handler (error=0, client_data=0x2c865150)
     at gdb/ui.c:154
 #8  0x0000000001a04edf in handle_file_event (file_ptr=0x2ccf8850, ready_mask=1)
     at gdbsupport/event-loop.cc:551
 #9  0x0000000001a05522 in gdb_wait_for_event (block=1)
     at gdbsupport/event-loop.cc:672
 #10 0x0000000001a043ff in gdb_do_one_event (mstimeout=-1)
     at gdbsupport/event-loop.cc:263
 #11 0x00000000006d5480 in interp::do_one_event (this=0x2cc2af20, mstimeout=-1)
     at gdb/interps.h:93
 #12 0x0000000000fe670d in gdb_readline_wrapper (
     prompt=0x2ccca4e0 "Enable debuginfod for this session? (y or [n]) ")
     at gdb/top.c:1033
 #13 0x00000000010c6853 in defaulted_query(...) (...)
     at gdb/utils.c:844
 #14 0x00000000010c6b8a in nquery (...)
     at gdb/utils.c:901
 #15 0x00000000007a9324 in debuginfod_is_enabled ()
     at gdb/debuginfod-support.c:268
 #16 0x00000000007a950d in debuginfod_source_query (...)
     at gdb/debuginfod-support.c:311
 #17 0x0000000000efc2c7 in open_source_file (s=0x2cc8f4b0) at gdb/source.c:1152
 #18 0x0000000000efc619 in symtab_to_fullname (...) at gdb/source.c:1214
 #19 0x0000000000f5ebb3 in find_line_symtab (...)
     at gdb/symtab.c:3287
 #20 0x0000000000f5f0e5 in find_pc_for_line (...)
     at gdb/symtab.c:3391
 #21 0x0000000001011f54 in tui_get_begin_asm_address (...)
     at gdb/tui/tui-disasm.c:404
 #22 0x000000000104888d in tui_source_window_base::rerender (this=0x2cbdc570)
     at gdb/tui/tui-winsource.c:474
 #23 0x0000000001028e81 in tui_win_info::resize (this=0x2cbdc570,
     height_=21, width_=127, origin_x_=0,
     origin_y_=0) at gdb/tui/tui-layout.c:299
 #24 0x00000000010297d0 in tui_layout_window::apply (this=0x2cc50350,
     x_=0, y_=0, width_=127,
     height_=21, preserve_cmd_win_size_p=false) at gdb/tui/tui-layout.c:432
 #25 0x000000000102bfea in tui_layout_split::apply (this=0x2caea920,
     x_=0, y_=0, width_=127,
     height_=33, preserve_cmd_win_size_p=false) at gdb/tui/tui-layout.c:1026
 #26 0x0000000001028267 in tui_apply_current_layout (...)
     at gdb/tui/tui-layout.c:68
 #27 0x0000000001028737 in tui_set_layout (layout=0x2c9b9e90)
     at gdb/tui/tui-layout.c:133
 #28 0x0000000001028af5 in tui_set_initial_layout () at gdb/tui/tui-layout.c:209
 #29 0x000000000104b795 in tui_enable () at gdb/tui/tui.c:496
 #30 0x000000000104bab3 in tui_enable_command (args=0x0, from_tty=1)
     at gdb/tui/tui.c:591
 #31 0x00000000006c5ffe in do_simple_func (args=0x0, from_tty=1, c=0x2c9bb2f0)
     at gdb/cli/cli-decode.c:94
 #32 0x00000000006cc94f in cmd_func (cmd=0x2c9bb2f0, args=0x0, from_tty=1)
     at gdb/cli/cli-decode.c:2831
 #33 0x0000000000fe53ad in execute_command (p=0x2c86699a "", from_tty=1)
     at gdb/top.c:563
 #34 0x000000000094584d in command_handler (command=0x2c866990 "tui enable")
     at gdb/event-top.c:611
 #35 0x0000000000945dfe in command_line_handler (rl=...) at gdb/event-top.c:844
 #36 0x000000000101e916 in tui_command_line_handler (rl=...)
     at gdb/tui/tui-interp.c:101
 #37 0x0000000000944eef in gdb_rl_callback_handler (rl=0x2cc86a30 "tui enable")
     at gdb/event-top.c:288
 #38 0x0000000001175779 in rl_callback_read_char ()
     at readline/readline/callback.c:302
 #39 0x0000000000944bc3 in gdb_rl_callback_read_char_wrapper_sjlj ()
     at gdb/event-top.c:197
 #40 0x0000000000944cd4 in gdb_rl_callback_read_char_wrapper_noexcept ()
     at gdb/event-top.c:240
 #41 0x0000000000944d52 in gdb_rl_callback_read_char_wrapper (...)
     at gdb/event-top.c:252
 #42 0x0000000001062352 in stdin_event_handler (error=0, client_data=0x2c865150)
     at gdb/ui.c:154
 #43 0x0000000001a04edf in handle_file_event (file_ptr=0x2ccf8850, ready_mask=1)
     at gdbsupport/event-loop.cc:551
 #44 0x0000000001a05522 in gdb_wait_for_event (block=1)
     at gdbsupport/event-loop.cc:672
 #45 0x0000000001a043ff in gdb_do_one_event (mstimeout=-1)
     at gdbsupport/event-loop.cc:263
 #46 0x00000000006d5480 in interp::do_one_event (this=0x2cc2af20, mstimeout=-1)
     at gdb/interps.h:93
 #47 0x0000000000b77f25 in start_event_loop () at gdb/main.c:403
 #48 0x0000000000b78113 in captured_command_loop () at gdb/main.c:468
 #49 0x0000000000b7a07c in captured_main (context=0x7fff660b9e60)
     at gdb/main.c:1381
 #50 0x0000000000b7a178 in gdb_main (args=0x7fff660b9e60) at gdb/main.c:1400
 #51 0x0000000000419705 in main (argc=5, argv=0x7fff660b9f98) at gdb/gdb.c:38
 (gdb)
...

The problem is that the command window is used (during initialization of the
source window) before it's properly initialized.

Fix this by adding a check at the start of
tui_inject_newline_into_command_window that avoids the SIGFPE.

This makes the behavior on x86_64-linux (where we ran into the SIGFPE)
equivalent to the behavior on aarch64-linux, where a divide-by-zero just
returns zero.

Note that this does not address the fact that the query happens somewhere
during rather than before or after TUI initialization, so the query itself
continues to look odd:
...
This GDB supports auto-downloading debuginfo from the following URLs:
                                                                       <foo>
                                                                            Enable debuginfod for this session? (y or [n])
...

The fix is minimal, and consequently low-risk, and applies cleanly to the
gdb-17-branch.

Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=31449
---
 gdb/testsuite/gdb.tui/query-foo.c | 24 ++++++++
 gdb/testsuite/gdb.tui/query.exp   | 99 +++++++++++++++++++++++++++++++
 gdb/tui/tui-io.c                  |  3 +
 3 files changed, 126 insertions(+)
 create mode 100644 gdb/testsuite/gdb.tui/query-foo.c
 create mode 100644 gdb/testsuite/gdb.tui/query.exp

diff --git a/gdb/testsuite/gdb.tui/query-foo.c b/gdb/testsuite/gdb.tui/query-foo.c
new file mode 100644
index 00000000000..bf4a8766dcd
--- /dev/null
+++ b/gdb/testsuite/gdb.tui/query-foo.c
@@ -0,0 +1,24 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+   Copyright 2026 Free Software Foundation, Inc.
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
+
+extern int foo (void);
+
+int
+foo (void)
+{
+  return 0;
+}
diff --git a/gdb/testsuite/gdb.tui/query.exp b/gdb/testsuite/gdb.tui/query.exp
new file mode 100644
index 00000000000..acf0db63ebb
--- /dev/null
+++ b/gdb/testsuite/gdb.tui/query.exp
@@ -0,0 +1,99 @@
+# Copyright 2026 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+# Check that a debuginfod query during TUI initialization doesn't cause a
+# crash.
+
+require allow_tui_tests
+
+load_lib debuginfod-support.exp
+require allow_debuginfod_tests
+
+tuiterm_env
+
+standard_testfile -main.c -foo.c
+
+set tmpfile [standard_output_file $srcfile]
+set srcfiles [list $tmpfile $srcfile2]
+
+set fd [open $tmpfile w]
+puts $fd {
+    extern int foo (void);
+    int
+    main (void)
+    {
+	return foo ();
+    }
+}
+close $fd
+
+if { [build_executable "failed to prepare" $testfile $srcfiles] == -1 } {
+    return
+}
+
+# Delete query-main.c to trigger the debuginfod query during TUI
+# initialization.
+file delete $tmpfile
+
+set l {}
+lappend l "command"
+# Disabled due to PR33794.
+#lappend l "keys"
+
+foreach_with_prefix how $l {
+    save_vars { env(DEBUGINFOD_URLS) } {
+	setenv DEBUGINFOD_URLS "foo"
+	Term::clean_restart 24 80 $testfile
+    }
+
+    if {![Term::prepare_for_tui]} {
+	return 0
+    }
+
+    Term::gen_prompt
+
+    if { $how == "command" } {
+	send_gdb "tui enable\n"
+	set re \
+	    [string_to_regexp \
+		 {(y or [n]) }]
+	gdb_assert { [Term::wait_for_region_contents 0 0 80 24 $re] }
+	send_gdb "n\n"
+	Term::wait_for ""
+    } else {
+	send_gdb "\030\001"
+	set re \
+	    "Please set debuginfod enabled to on or off before enabling TUI"
+	gdb_assert { [Term::wait_for_region_contents 0 0 80 24 $re] }
+    }
+
+    # Check that prompt is responsive.
+    gdb_assert { [Term::command "print 1"] } "responsive prompt"
+
+    set cmd "set debuginfod enabled off"
+    gdb_assert { [Term::command "$cmd"] } $cmd
+
+    with_test_prefix tui {
+	if { $how == "command" } {
+	    send_gdb "tui enable\n"
+	} else {
+	    send_gdb "\030\001"
+	}
+	gdb_assert { [Term::wait_for ""] } "entered"
+
+	# Check that prompt is responsive.
+	gdb_assert { [Term::command "print 1"] } "responsive prompt"
+    }
+}
diff --git a/gdb/tui/tui-io.c b/gdb/tui/tui-io.c
index 642b88ead0c..503c7edf4d6 100644
--- a/gdb/tui/tui-io.c
+++ b/gdb/tui/tui-io.c
@@ -1080,6 +1080,9 @@ void
 tui_inject_newline_into_command_window ()
 {
   gdb_assert (tui_active);
+  if (tui_cmd_win () == nullptr || tui_cmd_win ()->width == 0
+      || tui_cmd_win ()->height == 0)
+    return;
 
   WINDOW *w = tui_cmd_win ()->handle.get ();
 
-- 
2.51.0

[PATCH v6 2/4] [gdb] Add debuginfod_enabled_ask_p

[Tom de Vries]

Factor out debuginfod_is_enabled_1 out of debuginfod_is_enabled, and use it to
implement debuginfod_enabled_ask_p.
---
 gdb/debuginfod-support.c | 40 +++++++++++++++++++++++++++++++---------
 gdb/debuginfod-support.h |  5 +++++
 2 files changed, 36 insertions(+), 9 deletions(-)

diff --git a/gdb/debuginfod-support.c b/gdb/debuginfod-support.c
index 0d666692ff8..56e284fbb7b 100644
--- a/gdb/debuginfod-support.c
+++ b/gdb/debuginfod-support.c
@@ -224,20 +224,20 @@ get_debuginfod_client ()
   return global_client;
 }
 
-/* Check if debuginfod is enabled.  If configured to do so, ask the user
-   whether to enable debuginfod.  */
+/* Check if debuginfod is enabled.  If configured to do so and ALLOW_ASK, ask
+   the user whether to enable debuginfod.  */
 
-static bool
-debuginfod_is_enabled ()
+static const char *
+debuginfod_is_enabled_1 (bool allow_ask)
 {
   const char *urls = skip_spaces (getenv (DEBUGINFOD_URLS_ENV_VAR));
 
   if (urls == nullptr
       || *urls == '\0')
-    return false;
+    return debuginfod_off;
 
-  if (debuginfod_enabled != debuginfod_ask)
-    return debuginfod_enabled == debuginfod_on;
+  if (debuginfod_enabled != debuginfod_ask || !allow_ask)
+    return debuginfod_enabled;
 
   gdb_printf (_("\nThis GDB supports auto-downloading debuginfo " \
 		"from the following URLs:\n"));
@@ -285,8 +285,18 @@ debuginfod_is_enabled ()
       debuginfod_enabled = debuginfod_on;
     }
 
-  gdb_assert (debuginfod_enabled != debuginfod_ask);
-  return debuginfod_enabled == debuginfod_on;
+  return debuginfod_enabled;
+}
+
+/* Check if debuginfod is enabled.  If configured to do so, ask the user
+   whether to enable debuginfod.  */
+
+static bool
+debuginfod_is_enabled ()
+{
+  const char *res = debuginfod_is_enabled_1 (true);
+  gdb_assert (res != debuginfod_ask);
+  return res == debuginfod_on;
 }
 
 /* Print the result of the most recent attempted download.  */
@@ -489,6 +499,18 @@ debuginfod_section_query (const unsigned char *build_id,
 
 #endif
 
+/* See debuginfod-support.h.  */
+
+bool
+debuginfod_enabled_ask_p ()
+{
+#if defined(HAVE_LIBDEBUGINFOD)
+  return debuginfod_is_enabled_1 (false) == debuginfod_ask;
+#else
+  return false;
+#endif
+}
+
 /* Set callback for "set debuginfod enabled".  */
 
 static void
diff --git a/gdb/debuginfod-support.h b/gdb/debuginfod-support.h
index cc5bf331812..7e7b5965c3b 100644
--- a/gdb/debuginfod-support.h
+++ b/gdb/debuginfod-support.h
@@ -105,4 +105,9 @@ extern scoped_fd debuginfod_section_query (const unsigned char *build_id,
 					   const char *section_name,
 					   gdb::unique_xmalloc_ptr<char>
 					     *destname);
+
+/* Check if debuginfod enabled is set to ask.  */
+
+extern bool debuginfod_enabled_ask_p ();
+
 #endif /* GDB_DEBUGINFOD_SUPPORT_H */
-- 
2.51.0

[PATCH v6 3/4] [gdb] Add defaulted_query_auto_answers_p

[Tom de Vries]

Factor out "static bool defaulted_query_auto_answers_p (bool)" out of
defaulted_query, and use it to implement
"extern tribool defaulted_query_auto_answers_p (void)".
---
 gdb/utils.c | 55 ++++++++++++++++++++++++++++++++++++++++++++++++-----
 gdb/utils.h |  4 ++++
 2 files changed, 54 insertions(+), 5 deletions(-)

diff --git a/gdb/utils.c b/gdb/utils.c
index f5f19301460..c521a2299bd 100644
--- a/gdb/utils.c
+++ b/gdb/utils.c
@@ -744,6 +744,54 @@ class scoped_input_handler
 };
 
 \f
+/* Return true if defaulted queries should be answered:
+   - if VERBOSELY, automatically and verbosely
+   - otherwise, automatically and quietly.  */
+
+static bool
+defaulted_query_auto_answers_p (bool verbosely)
+{
+  if (verbosely)
+    {
+      /* Automatically answer verbosely if input isn't coming from the user
+	 directly.  */
+      return (current_ui->instream != current_ui->stdin_stream
+	      || !current_ui->input_interactive_p ()
+	      /* Restrict queries to the main UI.  */
+	      || current_ui != main_ui);
+    }
+  else
+    {
+      /* Automatically answer quietly if:
+	 - the user did not want prompts, or
+	 - the command was issued with the server prefix.  */
+      return !confirm || server_command;
+    }
+}
+
+/* See utils.h.  */
+
+tribool
+defaulted_query_auto_answers_p ()
+{
+  if (defaulted_query_auto_answers_p (false)
+      || defaulted_query_auto_answers_p (true))
+    return TRIBOOL_TRUE;
+
+  if (deprecated_query_hook)
+    {
+      /* The only user of deprecated_query_hook in core gdb seems to be
+	 mi_interp_query_hook, for which we could return TRIBOOL_TRUE here.
+	 If that turns out to be required, we could accomplish this by
+	 changing the hook interface to say:
+	   int (*deprecated_query_hook) (const char *, va_list,
+					 bool auto_answers_p = false)
+	 and calling it here with auto_answers_p == true.  */
+      return TRIBOOL_UNKNOWN;
+    }
+
+  return TRIBOOL_FALSE;
+}
 
 /* This function supports the query, nquery, and yquery functions.
    Ask user a y-or-n question and return 0 if answer is no, 1 if
@@ -791,17 +839,14 @@ defaulted_query (const char *ctlstr, const char defchar, va_list args)
 
   /* Automatically answer the default value if the user did not want
      prompts or the command was issued with the server prefix.  */
-  if (!confirm || server_command)
+  if (defaulted_query_auto_answers_p (false))
     return def_value;
 
   /* If input isn't coming from the user directly, just say what
      question we're asking, and then answer the default automatically.  This
      way, important error messages don't get lost when talking to GDB
      over a pipe.  */
-  if (current_ui->instream != current_ui->stdin_stream
-      || !current_ui->input_interactive_p ()
-      /* Restrict queries to the main UI.  */
-      || current_ui != main_ui)
+  if (defaulted_query_auto_answers_p (true))
     {
       target_terminal::scoped_restore_terminal_state term_state;
       target_terminal::ours_for_output ();
diff --git a/gdb/utils.h b/gdb/utils.h
index ecef7d13c20..55fabd297fd 100644
--- a/gdb/utils.h
+++ b/gdb/utils.h
@@ -523,4 +523,8 @@ struct deferred_warnings final : public warning_hook_handler_type
 extern unsigned char c_ctrl (unsigned char c);
 extern unsigned char c_unctrl (unsigned char c);
 
+/* Whether defaulted queries are answered automatically.  */
+
+extern tribool defaulted_query_auto_answers_p ();
+
 #endif /* GDB_UTILS_H */
-- 
2.51.0

[PATCH v6 4/4] [gdb/tui] Fix crash with debuginfod query (C-x C-a)

[Tom de Vries]

PR tui/33794 reports an abort when the debuginfod query happens while enabling
TUI using the "C-x C-a" key combination.

The abort triggering is this one in rl_callback_read_char:
...
  if (rl_linefunc == NULL)
    {
      _rl_errmsg ("readline_callback_read_char() called with no handler!");
      abort ();
    }
...

This is similar to PR python/24796, and can be fixed by applying a similar fix
in tui_enable:
...
 void
 tui_enable (void)
 {
+  SCOPE_EXIT { reinstall_readline_callback_handler_cleanup (); };
...

After doing so, things seem to be fixed, but when compiling gdb with Asan, we
run into a heap-use-after-free.

The problem here is that readline is used recursively:
- readline is entered when handling the C-x C-a key combination
- readline is entered again when handling the debuginfod query

This would have to be addressed in readline, but in the mean time we need a
fix in gdb.

An RFC [1] was posted exploring the option of avoiding recursive readline use
by rather than handling the key combination directly, queueing it in an event
queue and handling the events later.

This patch takes a conservative approach: bail out if debuginfod enabled is
set to ask when entering TUI.

The fix is minimal, and consequently low-risk, and applies cleanly to the
gdb-17-branch (and that holds more or less as well for the patches this patch
is dependent on).

Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=33794

[1] https://sourceware.org/pipermail/gdb-patches/2026-February/225289.html
---
 gdb/testsuite/gdb.tui/query.exp |  3 +--
 gdb/tui/tui.c                   | 10 ++++++++--
 gdb/tui/tui.h                   |  2 +-
 3 files changed, 10 insertions(+), 5 deletions(-)

diff --git a/gdb/testsuite/gdb.tui/query.exp b/gdb/testsuite/gdb.tui/query.exp
index acf0db63ebb..eb0bee49aec 100644
--- a/gdb/testsuite/gdb.tui/query.exp
+++ b/gdb/testsuite/gdb.tui/query.exp
@@ -49,8 +49,7 @@ file delete $tmpfile
 
 set l {}
 lappend l "command"
-# Disabled due to PR33794.
-#lappend l "keys"
+lappend l "keys"
 
 foreach_with_prefix how $l {
     save_vars { env(DEBUGINFOD_URLS) } {
diff --git a/gdb/tui/tui.c b/gdb/tui/tui.c
index 9cf21f390c8..d2f1b79616c 100644
--- a/gdb/tui/tui.c
+++ b/gdb/tui/tui.c
@@ -41,6 +41,7 @@
 #include "top.h"
 #include "ui.h"
 #include "observable.h"
+#include "debuginfod-support.h"
 
 #include <fcntl.h>
 
@@ -140,7 +141,7 @@ tui_rl_switch_mode (int notused1 = 0, int notused2 = 0)
 
 	  /* If tui_enable throws, we'll re-prep below.  */
 	  rl_deprep_terminal ();
-	  tui_enable ();
+	  tui_enable (false);
 	}
     }
   catch (const gdb_exception_forced_quit &ex)
@@ -399,13 +400,18 @@ gdb_getenv_term (void)
    the gdb output, configures the readline to work in tui mode.
    When in curses mode, it does nothing.  */
 void
-tui_enable (void)
+tui_enable (bool queries_allowed)
 {
   TUI_SCOPED_DEBUG_ENTER_EXIT;
 
   if (tui_active)
     return;
 
+  if (!queries_allowed
+      && debuginfod_enabled_ask_p ()
+      && defaulted_query_auto_answers_p () != TRIBOOL_TRUE)
+    error (_("Please set debuginfod enabled to on or off before enabling TUI"));
+
   tui_batch_rendering defer;
 
   /* To avoid to initialize curses when gdb starts, there is a deferred
diff --git a/gdb/tui/tui.h b/gdb/tui/tui.h
index 7daefe2c9bb..e90c40b562e 100644
--- a/gdb/tui/tui.h
+++ b/gdb/tui/tui.h
@@ -70,7 +70,7 @@ extern bool tui_get_command_dimension (unsigned int *width,
 extern void tui_ensure_readline_initialized ();
 
 /* Enter in the tui mode (curses).  */
-extern void tui_enable (void);
+extern void tui_enable (bool queries_allowed = true);
 
 /* Leave the tui mode.  */
 extern void tui_disable (void);
-- 
2.51.0

Re: [PATCH v6 0/4] [gdb/tui] Fix crash with debuginfod query

[Andrew Burgess]

Tom de Vries <tdevries@suse.de> writes:

> This patch series fixes two PRs, PR31449 and PR33794.
>
> An initial patch for PR31449 was submitted here [1].  A review comment
> reported a problem when using C-x C-a to enable TUI instead of "tui enable",
> not fixed by the patch.  This was filed as PR33794.
>
> The initial patch fixes PR31449 by calling a function that will force the
> debuginfod query to happen before enabling the TUI.

Hi Tom,

Just FYI, I'm not forgotten this patch.  I'm in the process of reviewing
it, and will have some feedback next week.

Thanks,
Andrew


>
> Changes in v1 [2]:
> - change approach:
>   - add a layout with only the command window, and activate that layout first,
>     making sure that we finalize the initialization of the command window
>     before we activate the initial layout
>
> Changes in v2 [3]:
> - change approach:
>   - add a facility defer_tui_rerender that allows deferral of calling the
>     rerender methods (which draws everything but the window borders), and
>   - use that facility to postpone rerendering in tui_enable until after io has
>     been initialized.
>
> Changes in v3 [4]:
> - change approach:
>   - add a check at the start of tui_inject_newline_into_command_window that
>     avoids the SIGFPE.
>
> Changes in v4 [5]:
> - change approach:
>   - bail out if debuginfod enabled is set to ask when entering TUI.  This is
>     very similar to [1].  Unlike the previous approaches, this does fix
>     PR33794 as well.
>
> Changes in v5 [6]:
> - fixed some trivial errors in the v4.
> - split the v4 patch up into three patches,
> - added a patch introducing defaulted_query_auto_answers_p,
> - used defaulted_query_auto_answers_p to make the fix less intrusive
>
> Changes in v6:
> - committed first patch of v5, a refactor patch
> - change approach:
>   - revert to v3 approach for PR31449.  The error thrown in tui_enable also
>     affected "gdb -tui", and for that scenario could only be avoided with
>     "set debuginfod enabled on/off" in .gdbearlyinit.
> - describe entire version history in series description, instead of partly in
>   the series description and partly in one of the patches
>
> Tested on x86_64-linux.
>
> Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=31449
> Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=33794
>
> [1] https://sourceware.org/pipermail/gdb-patches/2024-March/207224.html
> [2] v1 https://sourceware.org/pipermail/gdb-patches/2026-January/224038.html
> [3] v2 https://sourceware.org/pipermail/gdb-patches/2026-January/224103.html
> [4] v3 https://sourceware.org/pipermail/gdb-patches/2026-February/225210.html
> [5] v4 https://sourceware.org/pipermail/gdb-patches/2026-February/225351.html
> [6] v5 https://sourceware.org/pipermail/gdb-patches/2026-March/226009.html
>
> Tom de Vries (4):
>   [gdb/tui] Fix crash with debuginfod query (tui enable)
>   [gdb] Add debuginfod_enabled_ask_p
>   [gdb] Add defaulted_query_auto_answers_p
>   [gdb/tui] Fix crash with debuginfod query (C-x C-a)
>
>  gdb/debuginfod-support.c          | 40 ++++++++++---
>  gdb/debuginfod-support.h          |  5 ++
>  gdb/testsuite/gdb.tui/query-foo.c | 24 ++++++++
>  gdb/testsuite/gdb.tui/query.exp   | 98 +++++++++++++++++++++++++++++++
>  gdb/tui/tui-io.c                  |  3 +
>  gdb/tui/tui.c                     | 10 +++-
>  gdb/tui/tui.h                     |  2 +-
>  gdb/utils.c                       | 55 +++++++++++++++--
>  gdb/utils.h                       |  4 ++
>  9 files changed, 224 insertions(+), 17 deletions(-)
>  create mode 100644 gdb/testsuite/gdb.tui/query-foo.c
>  create mode 100644 gdb/testsuite/gdb.tui/query.exp
>
>
> base-commit: 19bcbe810d8c49eb05324624c0af72091d38c4a7
> -- 
> 2.51.0

Re: [PATCH v6 0/4] [gdb/tui] Fix crash with debuginfod query

[Andrew Burgess]

Andrew Burgess <aburgess@redhat.com> writes:

> Tom de Vries <tdevries@suse.de> writes:
>
>> This patch series fixes two PRs, PR31449 and PR33794.
>>
>> An initial patch for PR31449 was submitted here [1].  A review comment
>> reported a problem when using C-x C-a to enable TUI instead of "tui enable",
>> not fixed by the patch.  This was filed as PR33794.
>>
>> The initial patch fixes PR31449 by calling a function that will force the
>> debuginfod query to happen before enabling the TUI.
>
> Hi Tom,
>
> Just FYI, I'm not forgotten this patch.  I'm in the process of reviewing
> it, and will have some feedback next week.

I finished looking through this series, and I couldn't shake the feeling
that it should be possible to allow the prompt to be displayed correctly
during the switch to TUI mode.

Fixing the 'tui enable' case is reasonably straight forward, but getting
'C-x C-a' to work correctly was a little harder, but I think I managed
it in the end.

I've posted these patches as my take on this problem:

  https://inbox.sourceware.org/gdb-patches/cover.1777645161.git.aburgess@redhat.com

It would be great to hear what you think.

Thanks,
Andrew

Re: [PATCH v6 0/4] [gdb/tui] Fix crash with debuginfod query

[Tom de Vries]

On 5/1/26 4:26 PM, Andrew Burgess wrote:
> Andrew Burgess <aburgess@redhat.com> writes:
> 
>> Tom de Vries <tdevries@suse.de> writes:
>>
>>> This patch series fixes two PRs, PR31449 and PR33794.
>>>
>>> An initial patch for PR31449 was submitted here [1].  A review comment
>>> reported a problem when using C-x C-a to enable TUI instead of "tui enable",
>>> not fixed by the patch.  This was filed as PR33794.
>>>
>>> The initial patch fixes PR31449 by calling a function that will force the
>>> debuginfod query to happen before enabling the TUI.
>>
>> Hi Tom,
>>
>> Just FYI, I'm not forgotten this patch.  I'm in the process of reviewing
>> it, and will have some feedback next week.
> 
> I finished looking through this series, and I couldn't shake the feeling
> that it should be possible to allow the prompt to be displayed correctly
> during the switch to TUI mode.
> 
> Fixing the 'tui enable' case is reasonably straight forward, but getting
> 'C-x C-a' to work correctly was a little harder, but I think I managed
> it in the end.
> 
> I've posted these patches as my take on this problem:
> 
>    https://inbox.sourceware.org/gdb-patches/cover.1777645161.git.aburgess@redhat.com
> 
> It would be great to hear what you think.
> 

Hi Andrew,

thanks for identifying that additional problem with 'C-x C-a' during a 
secondary prompt (patch #3).  The fix you came up with looks reasonably 
simple, and a low-risk backport to gdb-17-branch.

Patch #5 is not as trivial as the solution I proposed in this series, 
but I still think it's trivial enough to backport.

I do wonder whether patch #6 is backportable to gdb-17-branch, that 
looks a bit risky to me.

Unfortunately this is all the feedback I can provide, I'm on vacation 
for a week starting now.

Thanks,
- Tom

> Thanks,
> Andrew
>