pids: de_thread: don't clear session/pgrp pids for the old leader
authorOleg Nesterov <oleg@tv-sign.ru>
Wed, 30 Apr 2008 07:54:25 +0000 (00:54 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Wed, 30 Apr 2008 15:29:48 +0000 (08:29 -0700)
commit65450cebc6a2efde80ed45514f727e6e4dc1eafd
tree26d3cdb75b4e94eed773c71c4bcb3c5d5843eed3
parent5cd204550b1a006f2b0c986b0e0f53220ebfd391
pids: de_thread: don't clear session/pgrp pids for the old leader

Based on Eric W. Biederman's idea.

Unless task == current, without tasklist_lock held task_session()/task_pgrp()
can return NULL if the caller races with de_thread() which switches the group
leader.

Change transfer_pid() to not clear old->pids[type].pid for the old leader.
This means that its .pid can point to "nowhere", but this is already true for
sub-threads, and the old leader is not group_leader() any longer.  IOW, with
or without this change we can't trust task's special pids unless it is the
group leader.

With this change the following code

rcu_read_lock();
task = find_task_by_xxx();
do_something(task_pgrp(task), task_session(task));
rcu_read_unlock();

can't race with exec and hit the NULL pid.

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Pavel Emelyanov <xemul@openvz.org>
Cc: Roland McGrath <roland@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
kernel/pid.c