exec: rework the group exit and fix the race with kill

As Roland pointed out, we have the very old problem with exec.  de_thread()
sets SIGNAL_GROUP_EXIT, kills other threads, changes ->group_leader and then
clears signal->flags.  All signals (even fatal ones) sent in this window
(which is not too small) will be lost.

With this patch exec doesn't abuse SIGNAL_GROUP_EXIT.  signal_group_exit(),
the new helper, should be used to detect exit_group() or exec() in progress.
It can have more users, but this patch does only strictly necessary changes.

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Robin Holt <holt@sgi.com>
Cc: Roland McGrath <roland@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

1 files changed, 2 insertions, 2 deletions

diff --git a/kernel/signal.c b/kernel/signal.c
index 1117b28488c2..6a5f97cd337a 100644
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -957,7 +957,6 @@ void zap_other_threads(struct task_struct *p)
 {
         struct task_struct *t;
 
-        p->signal->flags = SIGNAL_GROUP_EXIT;
         p->signal->group_stop_count = 0;
 
         for (t = next_thread(p); t != p; t = next_thread(t)) {
@@ -1697,7 +1696,8 @@ static int do_signal_stop(int signr)
         } else {
                 struct task_struct *t;
 
-                if (!likely(sig->flags & SIGNAL_STOP_DEQUEUED))
+                if (!likely(sig->flags & SIGNAL_STOP_DEQUEUED) ||
+                    unlikely(sig->group_exit_task))
                         return 0;
                 /*
                  * There is no group stop already in progress.