Nowadays thread_group_empty() and next_thread() are simple list operations,
this optimization doesn't make sense: we are doing exactly same check one
line below.
Signed-off-by: Oleg Nesterov <[email protected]>
--- t/kernel/signal.c~4_ZOT 2007-08-03 21:11:59.000000000 +0400
+++ t/kernel/signal.c 2007-08-05 19:28:13.000000000 +0400
@@ -987,9 +987,6 @@ void zap_other_threads(struct task_struc
p->signal->flags = SIGNAL_GROUP_EXIT;
p->signal->group_stop_count = 0;
- if (thread_group_empty(p))
- return;
-
for (t = next_thread(p); t != p; t = next_thread(t)) {
/*
* Don't bother with already dead threads
Looks fine to me.
Thanks,
Roland