redefine thread_group_leader() as exit_signal >= 0
Change de_thread() to set old_leader->exit_signal = -1. This is
good for the consistency, it is no longer the leader and all
sub-threads have exit_signal = -1 set by copy_process(CLONE_THREAD).
And this allows us to micro-optimize thread_group_leader(), it can
simply check exit_signal >= 0. This also makes sense because we
should move ->group_leader from task_struct to signal_struct.
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Reviewed-by: Tejun Heo <tj@kernel.org>
diff --git a/fs/exec.c b/fs/exec.c
index 8dca45b..c3d517b 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -963,6 +963,7 @@
leader->group_leader = tsk;
tsk->exit_signal = SIGCHLD;
+ leader->exit_signal = -1;
BUG_ON(leader->exit_state != EXIT_ZOMBIE);
leader->exit_state = EXIT_DEAD;
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 39acee2..b38ed51 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -2284,8 +2284,10 @@
return tsk->signal->nr_threads;
}
-/* de_thread depends on thread_group_leader not being a pid based check */
-#define thread_group_leader(p) (p == p->group_leader)
+static inline bool thread_group_leader(struct task_struct *p)
+{
+ return p->exit_signal >= 0;
+}
/* Do to the insanities of de_thread it is possible for a process
* to have the pid of the thread group leader without actually being