The patch titled
move the related code from exit_notify() to exit_signals()
has been added to the -mm tree. Its filename is
move-the-related-code-from-exit_notify-to-exit_signals.patch
*** Remember to use Documentation/SubmitChecklist when testing your code ***
See http://www.zip.com.au/~akpm/linux/patches/stuff/added-to-mm.txt to find
out what to do about this
------------------------------------------------------
Subject: move the related code from exit_notify() to exit_signals()
From: Oleg Nesterov <oleg@xxxxxxxxxx>
The previous bugfix was not optimal, we shouldn't care about group stop
when we are the only thread or the group stop is in progress. In that case
nothing special is needed, just set PF_EXITING and return.
Also, take the related "TIF_SIGPENDING re-targeting" code from exit_notify().
So, from the performance POV the only difference is that we don't trust
!signal_pending() until we take ->siglock. But this in fact fixes another
___pure___ theoretical minor race. __group_complete_signal() finds the
task without PF_EXITING and chooses it as the target for signal_wake_up().
But nothing prevents this task from exiting in between without noticing the
pending signal and thus unpredictably delaying the actual delivery.
Signed-off-by: Oleg Nesterov <oleg@xxxxxxxxxx>
Cc: Davide Libenzi <davidel@xxxxxxxxxxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxx>
Cc: Roland McGrath <roland@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
kernel/exit.c | 18 ------------------
kernel/signal.c | 27 ++++++++++++++++++++++-----
2 files changed, 22 insertions(+), 23 deletions(-)
diff -puN kernel/exit.c~move-the-related-code-from-exit_notify-to-exit_signals kernel/exit.c
--- a/kernel/exit.c~move-the-related-code-from-exit_notify-to-exit_signals
+++ a/kernel/exit.c
@@ -741,24 +741,6 @@ static void exit_notify(struct task_stru
struct task_struct *t;
struct pid *pgrp;
- if (signal_pending(tsk) && !(tsk->signal->flags & SIGNAL_GROUP_EXIT)
- && !thread_group_empty(tsk)) {
- /*
- * This occurs when there was a race between our exit
- * syscall and a group signal choosing us as the one to
- * wake up. It could be that we are the only thread
- * alerted to check for pending signals, but another thread
- * should be woken now to take the signal since we will not.
- * Now we'll wake all the threads in the group just to make
- * sure someone gets all the pending signals.
- */
- spin_lock_irq(&tsk->sighand->siglock);
- for (t = next_thread(tsk); t != tsk; t = next_thread(t))
- if (!signal_pending(t) && !(t->flags & PF_EXITING))
- recalc_sigpending_and_wake(t);
- spin_unlock_irq(&tsk->sighand->siglock);
- }
-
/*
* This does two things:
*
diff -puN kernel/signal.c~move-the-related-code-from-exit_notify-to-exit_signals kernel/signal.c
--- a/kernel/signal.c~move-the-related-code-from-exit_notify-to-exit_signals
+++ a/kernel/signal.c
@@ -1871,19 +1871,36 @@ relock:
void exit_signals(struct task_struct *tsk)
{
int group_stop = 0;
+ struct task_struct *t;
- spin_lock_irq(&tsk->sighand->siglock);
- if (unlikely(tsk->signal->group_stop_count) &&
- !--tsk->signal->group_stop_count) {
- tsk->signal->flags = SIGNAL_STOP_STOPPED;
- group_stop = 1;
+ if (thread_group_empty(tsk) || signal_group_exit(tsk->signal)) {
+ tsk->flags |= PF_EXITING;
+ return;
}
+ spin_lock_irq(&tsk->sighand->siglock);
/*
* From now this task is not visible for group-wide signals,
* see wants_signal(), do_signal_stop().
*/
tsk->flags |= PF_EXITING;
+ if (!signal_pending(tsk))
+ goto out;
+
+ /* It could be that __group_complete_signal() choose us to
+ * notify about group-wide signal. Another thread should be
+ * woken now to take the signal since we will not.
+ */
+ for (t = tsk; (t = next_thread(t)) != tsk; )
+ if (!signal_pending(t) && !(t->flags & PF_EXITING))
+ recalc_sigpending_and_wake(t);
+
+ if (unlikely(tsk->signal->group_stop_count) &&
+ !--tsk->signal->group_stop_count) {
+ tsk->signal->flags = SIGNAL_STOP_STOPPED;
+ group_stop = 1;
+ }
+out:
spin_unlock_irq(&tsk->sighand->siglock);
if (unlikely(group_stop)) {
_
Patches currently in -mm which might be from oleg@xxxxxxxxxx are
__group_complete_signal-fix-coredump-with-group-stop-race.patch
remove-handle_group_stop-in-favor-of-do_signal_stop.patch
exec-rework-the-group-exit-and-fix-the-race-with-kill.patch
use-__set_task_state-for-traced-stopped-tasks.patch
add-task_wakekill.patch
do_wait-remove-one-else-if-branch.patch
proc-implement-proc_single_file_operations.patch
proc-rewrite-do_task_stat-to-correctly-handle-pid-namespaces.patch
proc-seqfile-convert-proc_pid_statm.patch
proc-proper-pidns-handling-for-proc-self.patch
proc-fix-the-threaded-proc-self.patch
kill-pt_attached.patch
kill-my_ptrace_child.patch
ptrace_check_attach-remove-unneeded-signal-=-null-check.patch
ptrace_stop-fix-the-race-with-ptrace-detachattach.patch
wait_task_stopped-simplify-and-fix-races-with-sigcont-sigkill-untrace.patch
do_wait-factor-out-retval-=-0-checks.patch
ptrace_stop-fix-racy-nonstop_code-setting.patch
wait_task_stopped-remove-unneeded-delay_group_leader-check.patch
do_wait-cleanup-delay_group_leader-usage.patch
do_wait-fix-security-checks.patch
do_wait-fix-security-checks-fix.patch
wait_task_continued-zombie-dont-use-task_pid_nr_ns-lockless.patch
wait_task_zombie-remove-exit_state-exit_signal-checks-for-wnowait.patch
sys_setpgid-simplify-pid-ns-interaction.patch
fix-setsid-for-sub-namespace-sbin-init.patch
teach-set_special_pids-to-use-struct-pid.patch
move-daemonized-kernel-threads-into-the-swappers-session.patch
start-the-global-sbin-init-with-00-special-pids.patch
fix-group-stop-with-exit-race.patch
sys_setsid-remove-now-unneeded-session-=-1-check.patch
move-the-related-code-from-exit_notify-to-exit_signals.patch
-
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
