[PATCH rc1-mm 3/3] coredump: copy_process: don't check SIGNAL_GROUP_EXIT

[PATCH rc1-mm 3/3] coredump: copy_process: don't check SIGNAL_GROUP_EXIT

[Oleg Nesterov]

After the previous patch SIGNAL_GROUP_EXIT implies a pending SIGKILL,
we can remove this check from copy_process() because we already checked
!signal_pending().

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>

--- MM/kernel/fork.c~3_fork	2006-03-23 22:59:33.000000000 +0300
+++ MM/kernel/fork.c	2006-04-09 03:17:12.000000000 +0400
@@ -1157,18 +1157,6 @@ static task_t *copy_process(unsigned lon
 	}
 
 	if (clone_flags & CLONE_THREAD) {
-		/*
-		 * Important: if an exit-all has been started then
-		 * do not create this new thread - the whole thread
-		 * group is supposed to exit anyway.
-		 */
-		if (current->signal->flags & SIGNAL_GROUP_EXIT) {
-			spin_unlock(&current->sighand->siglock);
-			write_unlock_irq(&tasklist_lock);
-			retval = -EAGAIN;
-			goto bad_fork_cleanup_namespace;
-		}
-
 		p->group_leader = current->group_leader;
 		list_add_tail_rcu(&p->thread_group, &p->group_leader->thread_group);
 

Re: [PATCH rc1-mm 3/3] coredump: copy_process: don't check SIGNAL_GROUP_EXIT

[Roland McGrath]

> After the previous patch SIGNAL_GROUP_EXIT implies a pending SIGKILL,
> we can remove this check from copy_process() because we already checked
> !signal_pending().

My SIGNAL_GROUP_EXEC change made this spot check for either flag, but this
already holds true for the exec case and so my change to this check here
was not needed.  With the coredump change done, I concur that the exit
cases are covered too and so this check can go.  For paranoia's sake,
perhaps we should leave behind a BUG_ON.


Thanks,
Roland