[PATCH 4/4] signal: zap_pid_ns_processes: s/SEND_SIG_NOINFO/SEND_SIG_FORCED/

[Oleg Nesterov <oleg@redhat.com>]

Change zap_pid_ns_processes() to use SEND_SIG_FORCED, it looks more
clear compared to SEND_SIG_NOINFO which relies on from_ancestor_ns
logic send_signal().

It is also more effecient if we need to kill a lot of tasks because
it doesn't alloc sigqueue.

While at it, add the __fatal_signal_pending(task) check as a minor
optimization.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
---
 kernel/pid_namespace.c |    8 ++------
 1 files changed, 2 insertions(+), 6 deletions(-)

diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c
index a896839..17b2328 100644
--- a/kernel/pid_namespace.c
+++ b/kernel/pid_namespace.c
@@ -168,13 +168,9 @@ void zap_pid_ns_processes(struct pid_namespace *pid_ns)
 	while (nr > 0) {
 		rcu_read_lock();
 
-		/*
-		 * Any nested-container's init processes won't ignore the
-		 * SEND_SIG_NOINFO signal, see send_signal()->si_fromuser().
-		 */
 		task = pid_task(find_vpid(nr), PIDTYPE_PID);
-		if (task)
-			send_sig_info(SIGKILL, SEND_SIG_NOINFO, task);
+		if (task && !__fatal_signal_pending(task))
+			send_sig_info(SIGKILL, SEND_SIG_FORCED, task);
 
 		rcu_read_unlock();
 
-- 
1.5.5.1