From: Kevin Buettner <kevinb@redhat.com>
To: Klaus Gerlicher <klaus.gerlicher@intel.com>
Cc: gdb-patches@sourceware.org, aburgess@redhat.com
Subject: Re: [PATCH v3 1/1] gdb: remember previously selected thread per inferior
Date: Tue, 8 Sep 2026 13:01:02 -0700 [thread overview]
Message-ID: <20260908130102.7ccb3698@f44-mesa-1> (raw)
In-Reply-To: <20260908133426.444197-2-klaus.gerlicher@intel.com>
Hi Klaus,
I had my AI reviewer look this over. It found some problems. In
addition to needing a rebase w/ resultant build problems fixed, the
test case needs some rework.
I'll point out what it found below...
On Tue, 8 Sep 2026 13:34:26 +0000
Klaus Gerlicher <klaus.gerlicher@intel.com> wrote:
[...]
> diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
> index 09080ccdad9..d4a59ff6445 100644
> --- a/gdb/doc/gdb.texinfo
> +++ b/gdb/doc/gdb.texinfo
> @@ -3518,6 +3518,26 @@ To switch focus between inferiors, use the @code{inferior} command:
> Make inferior number @var{infno} the current inferior. The argument
> @var{infno} is the inferior number assigned by @value{GDBN}, as shown
> in the first field of the @samp{info inferiors} display.
> +
> +By default, when switching to an inferior, @value{GDBN} selects the first
> +available thread. Optionally, @value{GDBN} can remember and restore the
> +thread that was previously selected in that inferior (if it still exists).
> +This behavior can be controlled with the @code{set remember-threads-per-inferior}
> +command, described below.
> +@end table
> +
> +@cindex remember threads per inferior
> +@kindex set remember-threads-per-inferior
> +@kindex show remember-threads-per-inferior
> +@table @code
> +@item set remember-threads-per-inferior on|off
> +Control whether @value{GDBN} remembers the selected thread for each inferior.
> +When enabled, switching to a previously used inferior will restore the thread
> +that was active when you switched away. When disabled (the default),
I believe that Eli asked for the word "selected" to be used instead
of "active" here.
[...]
> diff --git a/gdb/inferior.c b/gdb/inferior.c
> index 8c619ffec5a..9d8ea91f29e 100644
> --- a/gdb/inferior.c
> +++ b/gdb/inferior.c
> @@ -46,6 +46,9 @@ static int highest_inferior_num;
> /* See inferior.h. */
> bool print_inferior_events = true;
>
> +/* Control whether to remember selected threads per inferior. */
> +bool remember_inferior_threads = false;
> +
> /* The Current Inferior. This is a strong reference. I.e., whenever
> an inferior is the current inferior, its refcount is
> incremented. */
> @@ -508,6 +511,32 @@ inferior_pid_to_str (int pid)
>
> /* See inferior.h. */
>
> +void
> +save_inferior_last_thread ()
> +{
> + if (has_current_thread ())
> + current_inferior ()->last_user_thread
> + = thread_info_ref::new_reference (inferior_thread ());
> + else
> + current_inferior ()->last_user_thread = nullptr;
> +}
> +
> +/* Observer callback for explicit user-initiated thread switches.
> + This captures switches from commands like "thread X" or "inferior Y".
> + Note: This does NOT fire for automatic event-driven switches (e.g.,
> + breakpoint hits in schedule-multiple mode). Those are handled by a
> + direct call in fetch_inferior_event. Both mechanisms are necessary
> + to capture all thread switch scenarios. */
> +
> +static void
> +inferiors_on_user_selected_context_changed (user_selected_what selection)
> +{
> + if (remember_inferior_threads && (selection & USER_SELECTED_THREAD))
> + save_inferior_last_thread ();
> +}
> +
> +/* See inferior.h. */
> +
> void
> print_selected_inferior (struct ui_out *uiout)
> {
> @@ -771,8 +800,23 @@ inferior_command (const char *args, int from_tty)
> {
> if (inf != current_inferior ())
> {
> - thread_info *tp = any_non_exited_thread_of_inferior (inf);
> - if (tp == NULL)
> + thread_info *tp = nullptr;
> +
> + if (remember_inferior_threads)
> + {
> + thread_info *stored_tp = inf->last_user_thread.get ();
> +
> + /* Fallback to selecting any non-exited thread of
> + inferior. */
> + tp = (stored_tp == nullptr
> + || stored_tp->state == THREAD_EXITED)
thread_info::state is now a member function. That line should
now be:
|| stored_tp->state () == THREAD_EXITED)
> + ? any_non_exited_thread_of_inferior (inf)
> + : stored_tp;
> + }
> + else
> + tp = any_non_exited_thread_of_inferior (inf);
> +
> + if (tp == nullptr)
> error (_("Inferior has no threads."));
>
> switch_to_thread (tp);
[...]
> diff --git a/gdb/testsuite/gdb.base/inferior-switch.c b/gdb/testsuite/gdb.base/inferior-switch.c
> new file mode 100644
> index 00000000000..3b94713ff43
> --- /dev/null
> +++ b/gdb/testsuite/gdb.base/inferior-switch.c
> @@ -0,0 +1,42 @@
> +/* This testcase is part of GDB, the GNU debugger.
> + Copyright 2024 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/>. */
> +
> +#include <unistd.h>
> +#include <pthread.h>
> +
> +static void *
> +task (void *arg)
> +{
> + volatile unsigned int duration = 0;
> +
> + int a = 1; /* worker thread break 1. */
> +
> + sleep (duration);
> + int b = 2; /* worker thread break 2. */
> +
> + sleep (duration);
> + return NULL;
> +}
> +
> +int
> +main (void)
> +{
> + pthread_t th;
> +
> + alarm (30);
> +
> + pthread_create (&th, NULL, task, NULL);
> + pthread_join (&th, NULL);
I think this should be:
pthread_join (th, NULL);
[...]
> diff --git a/gdb/testsuite/gdb.base/inferior-switch.exp b/gdb/testsuite/gdb.base/inferior-switch.exp
> new file mode 100644
> index 00000000000..5cf7ee98fb1
> --- /dev/null
> +++ b/gdb/testsuite/gdb.base/inferior-switch.exp
[...]
FYI, for this test, I'm seeing the failures below. I think this is
likely due to the change in default behavior.
FAIL: gdb.base/inferior-switch.exp: test_disabled_behavior: feature on again, restores previously selected thread 1.2
FAIL: gdb.base/inferior-switch.exp: test_explicit: again back to inferior 2 thread 2.2
FAIL: gdb.base/inferior-switch.exp: test_explicit: back to inferior 1 thread 1.2
FAIL: gdb.base/inferior-switch.exp: test_explicit: switch to inferior 2 thread 2.2
FAIL: gdb.base/inferior-switch.exp: test_implicit: switch inferior 1 stopped at 2nd BP
FAIL: gdb.base/inferior-switch.exp: test_implicit: switch inferior 2 stopped at 2nd BP
Kevin
prev parent reply other threads:[~2026-09-08 20:01 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-08 13:34 [PATCH v3 0/1] " Klaus Gerlicher
2026-09-08 13:34 ` [PATCH v3 1/1] " Klaus Gerlicher
2026-09-08 20:01 ` Kevin Buettner [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260908130102.7ccb3698@f44-mesa-1 \
--to=kevinb@redhat.com \
--cc=aburgess@redhat.com \
--cc=gdb-patches@sourceware.org \
--cc=klaus.gerlicher@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox