The patch titled signals: re-assign CLD_CONTINUED notification from the sender to reciever has been added to the -mm tree. Its filename is signals-re-assign-cld_continued-notification-from-the-sender-to-reciever.patch Before you just go and hit "reply", please: a) Consider who else should be cc'ed b) Prefer to cc a suitable mailing list as well c) Ideally: find the original patch on the mailing list and do a reply-to-all to that, adding suitable additional cc's *** 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 The current -mm tree may be found at http://userweb.kernel.org/~akpm/mmotm/ ------------------------------------------------------ Subject: signals: re-assign CLD_CONTINUED notification from the sender to reciever From: Oleg Nesterov <oleg@xxxxxxxxxx> Based on discussion with Jiri and Roland. In short: currently handle_stop_signal(SIGCONT, p) sends the notification to p->parent, with this patch p itself notifies its parent when it becomes running. handle_stop_signal(SIGCONT) has to drop ->siglock temporary in order to notify the parent with do_notify_parent_cldstop(). This leads to multiple problems: - as Jiri Kosina pointed out, the stopped task can resume without actually seeing SIGCONT which may have a handler. - we race with another sig_kernel_stop() signal which may come in that window. - we race with sig_fatal() signals which may set SIGNAL_GROUP_EXIT in that window. - we can't avoid taking tasklist_lock() while sending SIGCONT. With this patch handle_stop_signal() just sets the new SIGNAL_CLD_CONTINUED flag in p->signal->flags and returns. The notification is sent by the first task which returns from finish_stop() (there should be at least one) or any other signalled thread from get_signal_to_deliver(). This is a user-visible change. Say, currently kill(SIGCONT, stopped_child) can't return without seeing SIGCHLD, with this patch SIGCHLD can be delayed unpredictably. Another difference is that if the child is ptraced by another process, CLD_CONTINUED may be delivered to ->real_parent after ptrace_detach() while currently it always goes to the tracer which doesn't actually need this notification. Hopefully not a problem. The patch asks for the futher obvious cleanups, I'll send them separately. Signed-off-by: Oleg Nesterov <oleg@xxxxxxxxxx> Cc: Roland McGrath <roland@xxxxxxxxxx> Cc: Jiri Kosina <jkosina@xxxxxxx> Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> --- include/linux/sched.h | 6 ++++++ kernel/signal.c | 29 +++++++++++++++++++---------- 2 files changed, 25 insertions(+), 10 deletions(-) diff -puN include/linux/sched.h~signals-re-assign-cld_continued-notification-from-the-sender-to-reciever include/linux/sched.h --- a/include/linux/sched.h~signals-re-assign-cld_continued-notification-from-the-sender-to-reciever +++ a/include/linux/sched.h @@ -554,6 +554,12 @@ struct signal_struct { #define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */ #define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */ #define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */ +/* + * Pending notifications to parent. + */ +#define SIGNAL_CLD_STOPPED 0x00000010 +#define SIGNAL_CLD_CONTINUED 0x00000020 +#define SIGNAL_CLD_MASK (SIGNAL_CLD_STOPPED|SIGNAL_CLD_CONTINUED) /* If true, all threads except ->group_exit_task have pending SIGKILL */ static inline int signal_group_exit(const struct signal_struct *sig) diff -puN kernel/signal.c~signals-re-assign-cld_continued-notification-from-the-sender-to-reciever kernel/signal.c --- a/kernel/signal.c~signals-re-assign-cld_continued-notification-from-the-sender-to-reciever +++ a/kernel/signal.c @@ -603,10 +603,8 @@ static void handle_stop_signal(int sig, * the SIGCHLD was pending on entry to this kill. */ p->signal->group_stop_count = 0; - p->signal->flags = SIGNAL_STOP_CONTINUED; - spin_unlock(&p->sighand->siglock); - do_notify_parent_cldstop(p, CLD_STOPPED); - spin_lock(&p->sighand->siglock); + p->signal->flags = SIGNAL_STOP_CONTINUED | + SIGNAL_CLD_STOPPED; } rm_from_queue(SIG_KERNEL_STOP_MASK, &p->signal->shared_pending); t = p; @@ -643,25 +641,23 @@ static void handle_stop_signal(int sig, * We were in fact stopped, and are now continued. * Notify the parent with CLD_CONTINUED. */ - p->signal->flags = SIGNAL_STOP_CONTINUED; + p->signal->flags = SIGNAL_STOP_CONTINUED | + SIGNAL_CLD_CONTINUED; p->signal->group_exit_code = 0; - spin_unlock(&p->sighand->siglock); - do_notify_parent_cldstop(p, CLD_CONTINUED); - spin_lock(&p->sighand->siglock); } else { /* * We are not stopped, but there could be a stop * signal in the middle of being processed after * being removed from the queue. Clear that too. */ - p->signal->flags = 0; + p->signal->flags &= ~SIGNAL_STOP_DEQUEUED; } } else if (sig == SIGKILL) { /* * Make sure that any pending stop signal already dequeued * is undone by the wakeup for SIGKILL. */ - p->signal->flags = 0; + p->signal->flags &= ~SIGNAL_STOP_DEQUEUED; } } @@ -1745,6 +1741,19 @@ relock: try_to_freeze(); spin_lock_irq(¤t->sighand->siglock); + + if (unlikely(current->signal->flags & SIGNAL_CLD_MASK)) { + int why = (current->signal->flags & SIGNAL_STOP_CONTINUED) + ? CLD_CONTINUED : CLD_STOPPED; + current->signal->flags &= ~SIGNAL_CLD_MASK; + spin_unlock_irq(¤t->sighand->siglock); + + read_lock(&tasklist_lock); + do_notify_parent_cldstop(current->group_leader, why); + read_unlock(&tasklist_lock); + goto relock; + } + for (;;) { struct k_sigaction *ka; _ Patches currently in -mm which might be from oleg@xxxxxxxxxx are git-hrt.patch kthread-call-wake_up_process-without-the-lock-being-held.patch mmap_region-cleanup-the-final-vma_merge-related-code.patch remove-unused-variable-from-send_signal.patch turn-legacy_queue-macro-into-static-inline-function.patch consolidate-checking-for-ignored-legacy-signals.patch consolidate-checking-for-ignored-legacy-signals-simplify.patch signals-do_signal_stop-use-signal_group_exit.patch signals-do_group_exit-use-signal_group_exit-more-consistently.patch lock_task_sighand-add-rcu-lock-unlock.patch k_getrusage-dont-take-rcu_read_lock.patch do_task_stat-dont-take-rcu_read_lock.patch signals-consolidate-checks-for-whether-or-not-to-ignore-a-signal.patch signals-clean-dequeue_signal-from-excess-checks-and-assignments.patch signals-consolidate-send_sigqueue-and-send_group_sigqueue.patch signals-cleanup-security_task_kill-usage-implementation.patch signals-re-assign-cld_continued-notification-from-the-sender-to-reciever.patch ptrace-introduce-task_detached-helper.patch ptrace-introduce-ptrace_reparented-helper.patch reparent_thread-use-same_thread_group.patch document-de_thread-with-exit_notify-connection.patch workqueues-shrink-cpu_populated_map-when-cpu-dies.patch workqueues-shrink-cpu_populated_map-when-cpu-dies-fix.patch cleanup_workqueue_thread-remove-the-unneeded-cpu-parameter.patch simplify-cpu_hotplug_begin-put_online_cpus.patch redo-locking-of-tty-pgrp.patch resume-tty-on-susp-and-fix-crnl-order-in-n_tty-line-discipline.patch tty_io-fix-remaining-pid-struct-locking.patch tty_io-fix-remaining-pid-struct-locking-small-cleanup.patch procfs-task-exe-symlink.patch procfs-task-exe-symlink-fix.patch procfs-task-exe-symlink-fix-2.patch free_pidmap-turn-it-into-free_pidmapstruct-upid.patch put_pid-make-sure-we-dont-free-the-live-pid.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