handle the multi-threaded init's exit() properly
authorOleg Nesterov <oleg@tv-sign.ru>
Wed, 17 Oct 2007 06:26:49 +0000 (23:26 -0700)
committerLinus Torvalds <torvalds@woody.linux-foundation.org>
Wed, 17 Oct 2007 15:42:51 +0000 (08:42 -0700)
commit84eb646b6eabcd82ec563f30d2d9d40c2054a9c9
tree96219c5844c80c40cbc9814a13a9a72de7430fc7
parent8e9073ed027771bcdee4033eb900a3c09ac90a19
handle the multi-threaded init's exit() properly

With or without this patch, multi-threaded init's are not fully supported,
but do_exit() is completely wrong.  This becomes a real problem when we
support pid namespaces.

1. do_exit() panics when the main thread of /sbin/init exits. It should not
   until the whole thread group exits. Move the code below, under the
   "if (group_dead)" check.

   Note: this means that forget_original_parent() can use an already dead
   child_reaper()'s task_struct. This is OK for /sbin/init because

    - do_wait() from alive sub-thread still can reap a zombie, we iterate
      over all sub-thread's ->children lists

    - do_notify_parent() will wakeup some alive sub-thread because it sends
      the group-wide signal

   However, we should remove choose_new_parent()->BUG_ON(reaper->exit_state)
   for this.

2. We are playing games with ->nsproxy->pid_ns. This code is bogus today, and
   it has to be changed anyway when we really support pid namespaces, just
   remove it.

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Roland McGrath <roland@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Sukadev Bhattiprolu <sukadev@us.ibm.com>
Cc: Serge Hallyn <serue@us.ibm.com>
Cc: Cedric Le Goater <clg@fr.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
kernel/exit.c